article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
there are many systems of particles interacting with long - range interactions in nature : self - gravitating bodies in astrophysics and cosmology @xcite , two - dimensional fluid dynamics @xcite , cold atoms @xcite , etc .
considering , for simplicity , @xmath10dimensional particle systems which interact through an isotropic pair potential @xmath11 , long - range systems are usually defined as those for which [ pot ] v(r)~ , where @xmath12 , and @xmath13 is a coupling constant .
this characterization of interactions as long - range arises in equilibrium statistical mechanics @xcite : in a system of @xmath5 particles in a volume @xmath14 , the average energy of a particle is , for @xmath15 , independent of the size of the system in the usual "
thermodynamic limit @xmath16 , @xmath17 at fixed density @xmath18 .
for @xmath12 a different thermodynamic limit must be taken in order to recover extensivity of the thermodynamic potentials , and @xmath5 independent intensive properties of the system , as @xmath19 .
more specifically , the potential energy @xmath20 of a particle scales as @xmath21 and @xmath13 and @xmath14 must be scaled appropriately with @xmath5 so that @xmath20 is constant .
this is usually called the mean - field thermodynamic limit ( or the vlasov limit when is is taken at fixed system size ) . using this scaling ,
the total energy becomes extensive and it is possible to compute thermal equilibrium properties . for the class of systems we consider here , with attractive power law interactions at large scales in three dimensions , such a treatment has been given in @xcite .
for @xmath22 they present unusual features compared to short range systems : inhomogeneous spatial distributions , inequivalence of the statistical ensembles , negative specific heat in the microcanonical ensemble etc .. for the case of gravity it was understood decades ago , however , in the context of astrophysics ( through the seminal works of chandrasekhar , lynden - bell and others ) that such considerations based on equilibrium statistical mechanics are only relevant physically on time scales very long compared to those on which such systems evolve dynamically ( e.g. the formation and evolution of galaxies ) , and that the scenario of the dynamics of such systems is completely different to that of short range systems : on a timescale @xmath23 characteristic of the mean field dynamics ( and independent of @xmath5 in the mean field limit described above ) one observes the formation , under the effect of a mean field global interaction through so - called mean - field relaxation , of very slowly evolving macroscopic states ( e.g. galaxies ) which are far from thermal equilibrium .
for gravity in @xmath7 dimensions , the time scale for evolution towards equilibrium was first estimated by chandrasekhar @xcite to be @xmath24 .
thus as @xmath19 in the mean field limit , the system remains trapped in such states and never evolves towards thermodynamic equilibrium .
a similar phenomenology has been established in the last years in the study of various other systems with long - range interactions ( see e.g. @xcite ) : relaxation on a mean field time scale to a quasi - stationary state " ( qss ) followed by a relaxation towards thermodynamic equilibrium on a time scale which diverges with the particle number @xmath5 .
this scenario has thus been proposed as a kind of paradigm for the dynamics of this class of interactions ( e.g. @xcite ) .
more formally the evolution of a system of @xmath5 particles interacting through the pair potential can be described by the equation [ vlasov ] + ( , t)+=c_n , where @xmath25 is the mean phase space density function , i.e. , the density of particles at the position @xmath26 with velocity @xmath27 at time @xmath28 , and @xmath29 is called the collision term " . in general
the latter is a functional of the @xmath30 distribution functions .
the term @xmath31 $ ] is the _
mean field force _ which can be written in terms of the pair potential @xmath11 as [ force - vlasov ] = -f(,,t ) _ v(|-|)dd .
a mean - field dynamical description is valid if , in the mean - field ( or vlasov ) limit , we have that [ coll_limit ] _
n c_n=0 in which case the dynamics is described by the vlasov equation , known as the collisionless boltzmann equation " in the astrophysical literature ( e.g. @xcite ) .
qss are understood as stable stationary solutions of these equations , and mean - field relaxation as the evolution towards such states in the same mean - field framework ( on timescales of order @xmath23 ) .
correspondingly , in any finite ( but large ) @xmath5 system , the term @xmath29 then describes the collisional " corrections to the mean - field dynamics . for long - range interactions , therefore , to show that qss should exist one should analyze these collision terms , and determine firstly that they do indeed satisfy the condition ( [ coll_limit ] ) . further in order to understand their evolution away from qss at large but finite @xmath5 , and ( possibly ) towards thermal equilibrium , one needs to derive a suitable kinetic theory , which should allow one to infer the scalings of the time scale ( or scales ) characterizing such evolution as a function of @xmath5 .
concerning the first step rigorous results have been obtained showing that the limit does exist in the gravitational case @xcite and for any potential with @xmath32 @xcite ( both in @xmath7 dimensions ) and provided a suitable regularization ( i.e. softening ) of the potential is imposed at small separations ( see also @xcite ) .
however these provide only rigorous lower bounds ( @xmath33 ) to the time scales on which the vlasov dynamics is valid .
they do not allow us to calculate in any practical manner the time scales for collisional relaxation , nor even to determine their parametric scalings .
many attempts have been made in this direction through the construction of explicit kinetic theories @xcite but , in practice it is difficult to apply these methods to realistic systems to establish the relevant time scales , and in particular their parametric scalings .
moreover , these theories do not take into account strong collisions .
often ( e.g. @xcite ) it is argued , using such approaches , that the characteristic time scale for collisional relaxation has a generic scaling @xmath34 , except for the special case of homogeneous qss in one dimension . in this paper
we explore the conditions under which the limit is satisfied for the generic power law interaction .
to do so we use a non - rigorous ( but well defined ) approach to the problem : we generalize the simple method initiated by chandrasekhar for the case of gravity @xcite .
this amounts to assuming that the dominant contribution to the collisionality , described by the term @xmath29 , comes from _ two body collisions_. for the gravitational interaction this simple approach has turned out to account remarkably well for the observed time scales of collisional relaxation ( in numerical simulations ) .
we generalize this approach to a generic power - law interaction ; and compare the results obtained to the results of numerical simulations of several such systems .
several important results emerge from this analysis .
firstly , it becomes evident through this approach that , in general , the characteristic time @xmath35 for collisional relaxation scales with the particle number @xmath5 and may depend on the properties of the two body potential at small distances .
our results for the two body collisional relaxation lead to the conclusion that , in this respect , an important qualitative distinction can be made between the cases @xmath6 and @xmath36 : in both cases , for unsoftened potentials , @xmath37 where @xmath38 is a constant depending on @xmath8 and the dimension of space @xmath39
. however the sign of @xmath38 is positive only if @xmath6 .
this means that , when the size of the core is sent to zero , the condition eq .
( [ coll_limit ] ) can be satisfied only for @xmath6 .
the existence of qss requires the satisfaction of this condition , and therefore such states can exist for @xmath40 only if the rate of collisionality is reduced through the introduction of a sufficiently soft core . in other words , for @xmath6 qss can be considered to occur simply because of the large distance behavior of the potential , while for @xmath9 their existence depends on the details of the short - distance behavior .
this leads to what we call a _ dynamical _ ( rather than thermodynamical ) classification of the range of interactions , which has been proposed also using different analyses in @xcite . the essential result above has already been reported in @xcite . in this paper
we present a more detailed and more extended study of collisional relaxation in these systems , both for the analytical and numerical parts . in the analytical part we present both a new quantitative treatment of the two body relaxation including the contribution from hard collisions , and also of the case of different specified soft core regularizations . in the numerical part we present much more extensive results and detailed analysis , including notably potentials which decrease more slowly than the gravitational potential , and a full quantitative exploration of the role of softening .
the paper is organized as follows : in the next section we give a brief review on the literature of the collisional relaxation in the context of gravitational systems and detail our generalization of chandrasekhar calculation of the two body collisional relaxation rate for the pair potentials , with soft or hard regularizations at small distances .
this leads us to write parametric scalings which allow us to infer our classification of the range of pair interactions . in the following section we describe the numerical simulations we use to explore the validity of our analytical results , their initial conditions and the macroscopic quantities we measure to characterize collisional relaxation . in the next section we present our numerical results , first for the previously studied case of gravity , and then for several cases with @xmath41 and @xmath42
we compare then quantitatively the relaxation time obtained theoretically with our simulations and , in the next section , we give numerical evidence indicating that the maximum impact parameter scales with the size of the system . in the final section
we draw our conclusions .
the parametric dependence of the characteristic time @xmath23 for mean field evolution is given by that of the typical time a particle needs to cross the system , of size @xmath43 , under the mean field force : [ tau_mf ] _ dyn , where @xmath44 is the mass of each particle .
the determination of the parametric dependence of the characteristic time of collisionality @xmath45 and , as expected , of relaxation towards thermodynamic equilibrium is much less evident . for the case of gravity ( @xmath46 ) in three dimensions , chandrasekhar gave the first estimates in 1943 @xcite , through a calculation of a diffusion coefficient in velocity space for an infinite homogeneous self - gravitating distribution of particles .
the central hypothesis , as for short - ranged systems , was to suppose that the main contribution to the collisional relaxation process arises from _ two - body encounters_. he calculated the variation of velocity of a test particle undergoing a `` collision '' with a particle of the homogeneous distribution , the global relaxation process being the cumulative effect of such `` collisions '' .
as we will see in the next subsection the standard notion of _ impact parameter _ appears in the calculations .
due to the assumption of an infinite homogeneous distribution and to the long - range nature of gravity , chandrasekhar had to cut - off the maximum impact parameter allowed at some scale , which he chose to be given by the typical inter - particle separation .
more than twenty years after the paper of chandrasekhar , hnon @xcite did a new calculation following the hypothesis of chandrasekhar , but considering that all the particles in the system would contribute to the relaxation .
there is then no need to introduce artificially an upper cutoff in the impact parameter , as it is naturally fixed by the size of the system .
more recent theoretical approaches , like e.g. @xcite ( and references therein ) , have followed a more complete approach , linearizing the boltzmann equation .
this approach makes possible to take into account not only local but also collective effects .
this approach is , however , very cumbersome analytically and does not lead in practice to definite conclusions about the issues we address here . on the other hand , @xmath5-body computer simulations of the relaxation problem
have been performed to test the analytical predictions . in three dimensions
such studies have been developed only for the case of gravitational interaction .
we note , amongst others , numerical studies focusing on the cosmological aspect @xcite , others focusing on the maximum relevant impact parameter in the relaxation process @xcite .
after some controversy , it seems that the appropriate maximal impact parameter is the size of the system ( rather than the inter - particle distance as postulated initially by chandrasekhar ) .
the study of the relaxation in softened potentials ( see e.g. @xcite ) give more indications in this direction .
this is a result we will confirm and provide new evidence for in this paper . in the rest of this section we present our generalization of the two body collisional relaxation time for any attractive power law pair potential of the form ( [ pot ] ) , with @xmath47 and a soft or hard core regularization at @xmath48 , and any spatial dimension @xmath49 .
the reasons for these restrictions on @xmath8 and @xmath39 become evident in the calculation below .
these calculations give us the parametric dependence for the relaxation rate via two body collisions , @xmath50 , in a virialized system .
as discussed in the introduction , if we assume that these processes are the dominant ones in the collisional dynamics , we can then write the condition for the existence of a regime in which a mean - field ( vlasov ) description of the dynamics is valid as @xcite _ dyn 0 n , [ criterion - qss ] since qss corresponds to the stationary ( and thus virialized ) states of the vlasov equation , condition ( [ criterion - qss ] ) is also a necessary one for the existence of such states .
we consider two particles of equal mass @xmath44 , position vectors @xmath51 and @xmath52 , and velocity vectors @xmath53 and @xmath54 .
their relative position vector is denoted [ relative - vector ] = _ 1-_2 and their relative velocity @xmath55 . in their center of mass frame ,
the velocities of the two particles are given by @xmath56 .
thus if @xmath57 is the change in the relative velocity of the particles in the two body encounter , the changes in velocity of the two particles in the laboratory frame , @xmath58 and @xmath59 , ( which are equal to those in the cent - re of mass frame ) are @xmath60 the equations of the relative motion are those of a single particle of mass @xmath61 with position vector @xmath62 subject to the central potential .
we decompose @xmath57 as = v _ e _ + v _ e _ , where @xmath63 is a unit vector defined parallel to the initial axis of motion , and @xmath64 a unit vector orthogonal to it , in the plane of the motion ( see fig . [ coll ] ) . in the center of mass frame
, the collision occurs as depicted in fig .
[ coll ] , which shows the definition of the impact factor @xmath65 , and the deflection angle @xmath66 . as energy
is conserved in the collision , the magnitudes of the initial and final relative velocity , @xmath67 , are equal .
it follows that [ v_perp - and - v_par ] @xmath68 the angle @xmath69 can be calculated , as a function of the impact factor @xmath65 , using the classic formula @xcite [ phi - def ] ( b)=_r_min^ , where @xmath70 is the positive root of the denominator .
we consider now the case of a pure decaying power law pair potential , v(r)=- [ power - law - potential ] and @xmath47 .
for @xmath71 the corresponding force is attractive , while @xmath72 it is repulsive . in what follows
we will consider the attractive case , but we will discuss below also the repulsive case . indeed it turns out that our essential results hold in both cases .
the integral leads naturally to the definition of the characteristic length scale [ b0gamma ] b_0=()^1/. considering the attractive case , eq . may then be rewritten as [ phi2 ] ( b)=_r_min^ .
changing to the variable @xmath73 , we obtain [ phi2-x ] ( b / b_0)=_0^x_max , where now @xmath74 is the positive root of the denominator .
since @xmath75 , for given @xmath76 , is a function of @xmath77 only , it follows that @xmath69 is also a function of @xmath78 only .
equation can be solved analytically only in a few cases , and notably for the case @xmath46 which corresponds to gravity in @xmath7 . for the general ( @xmath79 ) case ,
the integral can easily be computed numerically , and @xmath80 and @xmath81 can then be calculated .
figure [ ruth ] displays the results for a few chosen cases . in order to derive analytically the parametric dependences of the two body relaxation rate
, it suffices , as we will see , to have analytical approximations in the two asymptotic regimes of soft ( @xmath82 ) and hard ( @xmath83 ) collisions .
the corresponding expressions have been derived in a separate article @xcite by one of us ( bm ) and another collaborator . in what follows
we make use of the relevant results of @xcite , where the full details of their derivations may be found . when @xmath84 the particle trajectories are weakly perturbed , and the collision is said to be _
it is shown in @xcite that , in this region , one has [ douxmoins ] ( b / b_0 ) = 2a ( ) ( b_0/ b ) ^ + ( ( b_0/b)^2 ) , where [ definition - a ] a ( ) = , with @xmath85 being the euler gamma function .
as the angle of deflection @xmath86 , it follows that [ change - v - soft ] @xmath87 in appendix [ alternative - derivation ] an alternative derivation of eq .
is presented .
it is shown in @xcite that , in this asymptotic regime , [ durplus1 ] ( b / b_0 ) = + ( ( b / b_0)^ ) , where @xmath89 for @xmath90 , @xmath91 for @xmath92 and @xmath93 for @xmath94 . if @xmath95 , collisions are well defined with an asymptotic free state @xcite only if [ condition - binary ] b > b_0 , where = ^1/(1-)^. for @xmath96 , on the other hand , there is a finite time singularity , i.e. , the relative distance of the particles vanishes at a finite time . the first term in the asymptotic expansion eq .
gives the angle of deflection in the limit of arbitrarily small impact factors , and shows that it depends on @xmath8 . while for the case @xmath46 ( i.e. gravity in @xmath7 ) each particle velocity is exactly reversed in the center of mass frame ( @xmath97 ) , the general result for the deflection angle is different , and it increases to infinity as @xmath98 from below . at @xmath99
each particle performs one full loop around the center of mass and escapes asymptotically in the same direction it arrived in , at @xmath100 each particle performs two full loops etc . , and as @xmath98 from below the number of such loops diverges . for @xmath101 , as noted , there is in fact a singularity , with the particles running into one another at a finite time . to include this case in our treatment
we must therefore assume that the pair potential eq .
( [ power - law - potential ] ) is regularized at @xmath48 , so that there is a well defined collision for any impact factor .
it follows from our analysis that this means that the asymptotic behavior below some arbitrarily small scale must be either repulsive , or , if attractive , diverging more slowly that @xmath102 . in what follows
this assumption will suffice to extend our results to the range @xmath101 .
following chandrasekhar we assume that thermal relaxation is induced by the randomization of particles velocity by two body collisions . in order to estimate the accumulated effect of two body collisions on a particle as it crosses the whole system
, we estimate first the number of encounters per unit of time with impact parameter @xmath65 . in doing so we make the following approximations : 1 .
the system is treated as a homogeneous random distribution of particles in a @xmath39 dimensional sphere of radius @xmath43 , 2 .
the initial squared relative velocity of colliding particles is given by the variance of the particle velocities in the system .
each particle is then assumed to perform a simple homogeneous random walk in velocity space , with zero mean change in velocity ( because the deflections due to each encounter have no preferred direction ) , and a positive mean squared velocity which we determine below . in this approximation
, we assume that the particles have rectilinear trajectories .
this approximation clearly breaks down in the case of hard collisions , in which the trajectory is strongly perturbed .
we expect however the estimation of the number of collisions per unit of time to remain correct in this case , because encounters modify only the direction of the velocity , and not its modulus .
as illustrated schematically in fig . , we now divide the system in disks of thickness @xmath103 , and write the average number of encounters with impact parameter between @xmath65 and @xmath104 of a particle crossing this disk as n= b^d-2 db dz [ eq22 ] where @xmath105 is a numerical factor which depends on the spatial dimension @xmath39 ( e.g. @xmath106 , @xmath107 ) .
multiplying eq . by the square of eq . with the condition , and integrating from @xmath108 to @xmath109 and from @xmath110 to @xmath111
, we then estimate the average change in the velocity _ during one crossing _ of the system , for the perpendicular and parallel components of the velocity respectively , as : [ total - change - v ] = 2b_d n ( ) ^d-1 i _ , ( ) where [ i - powerlaw ] i _ , ( x_r)= _ 0^x_r dx x^d-2 _ , ( x ) where @xmath112 , @xmath113 and [ theta ] @xmath114 ^ 2.\end{aligned}\ ] ] writing the expression for @xmath115 in this way allows a simple and useful comparison with the case of particles interacting by an exact repulsive hard core potential .
indeed it is straightforward to show ( see e.g. @xcite ) that for ( infinitely ) hard particles with a diameter @xmath116 , one has [ deflection - hardcore ] ( b ) = \ { ll 2 ( ) & b + 0 & . calculating @xmath115 for this case using exactly the same approach used above , one obtains , for the case @xmath117 , exactly eq . with @xmath118
let us return now to the expressions eq . for the case of ( attractive ) power law interactions .
given that @xmath119 we can make the approximation [ int - divided ] i _ , ( x_r ) & & _ 0 ^ 1 dx x^d-2 _ , ( x ) + & + & _ 1^x_r dx x^d-2 _ , ( x ) . the first integral gives the contribution due to hard collisions ( @xmath120 ) .
it is finite provided only that the deflection angle is well defined , i.e. , provided only that the two body collisions is well defined . as we have discussed above this
is true for any @xmath121 , and for @xmath122 if we assume the singularity at @xmath48 to be appropriately regularized .
thus this term gives a contribution to @xmath115 which has precisely the parametric dependences of an exact repulsive hard core , differing only by an overall numerical factor .
considering now the second term , giving the contribution from soft collisions ( @xmath123 ) , we see that there are two different cases according to the large @xmath124 behavior of @xmath125 : the integral is convergent as @xmath126 if and only if @xmath127 as @xmath128 .
we thus infer from eqs .
( [ change - v - soft ] ) the following : * for @xmath129 , [ i - vlr ] i _ ( x_r ) & & 4a^2 ( ) _ 0^x_r dx x^d-2 - 2 + & = & a^2 ( ) x_r^d-1 - 2 and @xmath130 .
thus the integral is dominated by the contribution of soft scatterings , for which the change in the relative velocity is predominantly orthogonal to the initial relative velocity .
replacing eq . in eq .
, we obtain the scaling [ scaling - slr ] ~n()^2 . where = + . * for @xmath131 , which corresponds to gravity in @xmath7 , the contribution from all impact factors from the scale @xmath132 must be included and _ ( x_r ) 4a^2 ( ) x_r . as in the previous case , @xmath130 .
note that , given @xmath119 this result for @xmath133 is very insensitive to precisely where the lower cut - off at @xmath134 is chosen .
we obtain therefore [ scaling - grav - like ] ~n()^d-1 ( ) . * for @xmath135
, we have _ , ( x_r ) _ , ( ) _ 0^ dx x^d-2 _ , ( x ) which is a constant that can be numerically calculated in a straightforward way for any given pair potential in this class .
we obtain therefore [ scaling - mlr ] ~n()^d-1 . in the last case , for sufficiently rapidly decaying potentials ,
we obtain therefore the same scaling as for the case of hard core particles of diameter @xmath132 . using these results ,
we now determine how the relaxation rate scales with the parameters of the system . assuming the system to be in a qss we can then obtain its scaling as a function of @xmath5 alone . for clarity we drop irrelevant numerical prefactors , but these will be analyzed further in sect .
[ quantitative ] .
we define the relaxation rate @xmath136 as the inverse of the time scale at which the normalized average change in velocity squared due to collisions is equal to one .
given that the estimated @xmath137 is the average change in a crossing time @xmath23 , we have therefore [ def_gamma ] _ dyn . in order to obtain the scaling with @xmath5 from the above results , we need to determine how the ratio @xmath138 scales with @xmath5 .
using the definition eq . and
assuming , as stated above , that the modulus of the relative velocity of colliding particles can be taken to be of the same order as the typical velocity of a single particle @xmath139 , we have ( ) ^~ ~ ~ [ scaling - virialization ] where @xmath140 is the total kinetic energy and @xmath141 the total potential energy of the system .
if we now assume the system to be in a qss , i.e. in virial equilibrium , the _ virial theorem _ gives that [ virial ] 2k + u
= 3 pv , where @xmath142 is the pressure of the particles on the boundaries if the system is enclosed , and @xmath143 if the system is open
. by definition the mean - field scaling with @xmath5 makes each term in eq .
scale in the same way with @xmath5 so that the relation remains valid independently of @xmath5 ( up to finite @xmath5 fluctuations ) . thus using this scaling
we can infer that [ scaling - bmin ] b_0~r n^-1/. using eqs .
, and , we then infer the following behaviors : * for @xmath144 , [ rate_llr ] _ dyn ~n ( b_0/r)^2~n^-1 . * for @xmath145 [ rate_gr ] _ dyn ~n^-1 ( n ) . * for @xmath135 [ rate_slr ] _
dyn~n^-(d-1-)/. it follows that that the condition eq
. only holds for potentials with @xmath6 . only in this case
therefore can the qss be supposed to exist as we have assumed . for @xmath9 , on the other hand , the relaxation induced by two body collisionality occurs on a time scale which is short compared to a particle crossing time , and a stationary non - thermal state can not exist on the latter time scale , i.e. , a qss can not exist .
we consider now the case in which the power - law potential is softened " at short distances , i.e. , regulated so that the modulus of the force between two particles is bounded above at some finite value .
the principle motivation for considering this case here is that , in practice , even for @xmath146 , we are unable numerically to test directly the validity of the scaling predictions eqs . - for the exact ( singular ) potentials : the numerical cost of integrating sufficiently accurately hard two body scatterings over the long time scales required is prohibitive .
instead we will consider power - law potentials softened at a scale @xmath147 , and study the scaling with both @xmath5 and @xmath147 of the relaxation rates in the numerically accessible range for these parameters .
a detailed analysis of the two body scattering for such softened power law potentials has been given also in @xcite .
we again use the results of this paper to infer , using eqs .
- above , the parametric scalings of the relaxation rate . as in the previous section ,
we defer until later a discussion of the exact numerical factors , for the specific smoothing functions used in our numerical simulations . as softening modifies the force below a characteristic scale @xmath147 ,
its effect is to modify the deflection angles for impact factor @xmath65 below a scale of the same order . from the considerations above it
is then evident that , for @xmath148 , such a softening does not change the parametric scalings : it can only change the numerical value of the ( finite ) first integral in eq . .
for @xmath149 , on the other hand , the second integral in eq . is modified because the functions @xmath125 are modified up to @xmath150 .
assuming that @xmath151 , this will lead to a modification of the parametric scaling of the full expressions for @xmath137 when @xmath152 . in @xcite
it is shown that , when @xmath153 , the deflection angle can be approximated as [ phi - approx - large - eps ] \ { ll 2b ( ) ( ) ^ ( ) & b < ^ * + 2a ( ) ( ) ^ & b > ^ * , . where @xmath154 is a finite constant the exact value of which depends on the functional form of softening used ( and @xmath155 is as defined in eq . ) .
the scale @xmath156 is of the same order as @xmath147 ( from continuity of eq . at @xmath157 ,
their ratio is given by @xmath158 ) .
we can now calculate approximately the second integral in eq . for the cases in which the parametric dependence of their values are modified by the smoothing ( with @xmath149 ) : * for @xmath159 ( taking @xmath160 ) : @xmath161\left(\frac{\ep}{b_0}\right)^{d-1 - 2\ga } \label{i - smoothed - perp } \\
{ \mathcal i}_{\parallel}&\simeq \left[\frac{b^4(\gamma)}{4(d+3)}+\frac{a^4(\ga)}{4\ga+1-d}\right ] \left(\frac{\ep}{b_0}\right)^{d-1 - 4\ga } \label{i - smoothed - par}.\end{aligned}\ ] ] and therefore @xmath162 if @xmath163 . * for @xmath145 , assuming @xmath164 ( i.e. @xmath151 ) , we obtain _ a^2 ( ) ( ) , while @xmath165 is given as eq . , and
@xmath162 if @xmath163 . using these results
we infer finally that the scalings of the relaxation rates of a qss ( with @xmath132 scaling as in eq . ) in the large @xmath5 limit are the following : * if @xmath166 , [ rate_llr - soft ] _ dyn ~n^-1 ,
i.e. the same as in the absence of smoothing ; * if @xmath135 , then [ rate_slr - soft ] _ dyn~n^-1 ( ) ^d-1 - 2 * if @xmath145 , then [ rate_gr - soft ] _ dyn ~n^-1 ( ) . in summary , the correct parametric scaling for the two body relaxation rates of a qss , in the case of a power - law potential softened at a scale @xmath149 ,
are well approximated by simply introducing a cutoff at an impact factor of order @xmath147 ( and therefore considering only the contribution from soft collisions ) . for what concerns the existence of qss , we thus conclude that , with a softened power law potential , one can satisfy the condition eq . even for any @xmath167 .
indeed , taking @xmath168 to be independent of @xmath5 ( i.e. scaling the softening with the system size ) , we obtain in all cases that @xmath169 .
more generally , it is straightforward to deduce what scaling of @xmath147 with @xmath5 is required to satisfy the condition eq . in the mean - field limit .
we have performed numerical simulations in @xmath7 of the evolution of @xmath5 particle systems , extending to sufficiently long times to observe their collisional evolution . as we have discussed in the previous section , exact power law interactions with @xmath170 lead to strong collisions at impact factors @xmath171 .
indeed , as we have seen , when @xmath8 increases much above unity particles can even make multiple loops around one another during collisions ( cf .
the smaller is @xmath65 , the shorter is the characteristic time for a collision compared to the mean field time and therefore the greater is temporal resolution required for an accurate integration ( and , in particular , conservation of the energy ) .
this means it is too expensive numerically , even for a few thousand particles , to accurately simulate such a system for times long enough to be comparable to the predicted relaxation times .
indeed we have seen that the calculation we have done predicts that , even for @xmath172 ( i.e. @xmath173 in @xmath7 ) , relaxation should be dominated by strong collisions with @xmath134 but nevertheless @xmath174 diverges in the mean field limit . for these reasons , we employ a potential with a softening which is sufficiently large to suppress strong collisions .
the predicted scalings we can test are thus those given in sect .
[ theo - soft ] , rather than the ones corresponding to pure power - law potentials given in sect .
[ rel_rate ] . by studying also the scalings with the softening @xmath147 at fixed @xmath5 ,
however , we can indirectly test in this way the extrapolation to the scalings in sect .
[ rel_rate ] .
we use a modification of the publicly available gravity code gadget2@xcite .
the force is computed using a modified barnes and hut tree algorithm , and we have modified the code in order to treat pair potentials of the form eq . and softened versions of them ( which are those we use in practice ) .
we use an opening angle @xmath175 , which ensures a very accurate computation of the force .
the evolution of the system is computed using a verlet - type drift - kick - drift symplectic integration scheme .
the simulations are checked using simple convergence tests on the numerical parameters , and their accuracy is monitored using energy conservation . for the time - steps used here it is typically conserved to within @xmath176 over the whole run , orders of magnitude smaller than the typical variation of the kinetic or potential energy over the same time . as initial conditions we take the @xmath5 particles randomly distributed in a sphere of radius @xmath177 , and ascribe velocities to particles so that each component is an independent uniformly distributed variable in an interval @xmath178 $ ] ( i.e. waterbag " type initial conditions in phase space ) .
the parameter @xmath179 is chosen so that initial virial ratio is unity , i.e. , @xmath180 .
we make this choice of initial conditions because it is expected to be close to a qss , to which ( collisionless ) relaxation should occur gently " , and this is indeed what we observe .
we have chosen to enclose the system in a cubic box of size @xmath181 , in order to avoid the complexities associated with particle evaporation .
this constraint is imposed in practice using _ soft _ boundary conditions , which are implemented by changing the sign of the @xmath182 component of the velocity when the @xmath182 component of the position lies outside the simulation box .
we use a time step of the order of @xmath183 ( which provides well converged results ) , where @xmath23 is defined precisely below .
we have performed simulations using two different softening schemes : a `` compact '' softening and a `` plummer '' softening .
the former corresponds to a two body potential [ compact - potential ] v^(r , ) = \ { ll - & r + - ( r/ ) & 0 r , . where @xmath184 is a polynomial ,
of which the exact expression is given in app .
[ appendix_potential ] .
it is chosen so that the potential and its first two derivatives are continuous at @xmath185 , and it interpolates to a force which vanishes at @xmath48 via a region in which the force becomes repulsive .
the plummer smoothing corresponds to the simple potential v_^p(r)=- , which is everywhere attractive .
as we have noted it is straightforward to calculate numerically the relaxation rates for these softened potentials , using eq .
and eq . .
we show in fig .
[ rutherford_softening3 ] the ratio of the resultant @xmath186 compared to its value for the exact power law , for @xmath187 and @xmath188 , as a function of the ratio @xmath189 . as described in the previous section we observe that , for @xmath190 , the effect of the softening is negligible , while for @xmath163 , we recover a simple power law scaling with @xmath147 which agrees with that derived above for this regime , cf .
we note that in fig .
[ rutherford_softening3 ] the normalization for the asymptotic plummer curves is greater that for the compact softening .
performing simulations with these two different softening schemes allows us to test not just the robustness of the agreement with the theoretical scalings derived above , which should not depend on the details of the softening scheme .
it also allows to test more quantitatively for the correctness of the theoretical predictions for the relaxation rates , which predicts also the relative amplitude of the relaxation rate in the regime @xmath163 . to facilitate this comparison
it is convenient to define an effective softening @xmath191 obtained by assuming that all the collisions are soft , i.e. , [ phi - approx - large - eps - estimate ] _ \ { ll 0 & b < _ eff + 2a()()^ & b _ eff . .
computing the same quantity as in fig . , we can determine , by matching with the result for any other softening scheme , a value of @xmath191 in units of @xmath147 .
we can compute therefore an effective softening using [ alpha - soft ] _
eff= , where the values of @xmath192 are given in tab .
[ table - soft ] for our two softening schemes , for the values of @xmath8 we explore here ( in the range @xmath193 where the softening plays a role ) .
the result for the case of gravity and the plummer softening is in agreement with that derived in @xcite ( see also @xcite ) .
thus our analytical calculations predict that the relaxation rates of qss measured with the different softening schemes should not only scale in the same way as a function of @xmath147 ( for @xmath163 ) but also they should be equal at values of @xmath147 corresponding to the same @xmath194 .
.factor @xmath192 ( see eq . ) to compute the effective softening @xmath194 ( see text ) in units of @xmath147 , for the two different softening schemed used in this work . [
cols="^,^,^",options="header " , ] going back to the original derivation of the two body relaxation rate by chandrasekhar there has been a debate about the correct choice of the _ maximum _ impact parameter . in section [ chandrasekhar ]
we have argued that it should be assumed to be of the order of the size of the system , and we have obtained our results making this hypothesis . for the case
@xmath195 , which is dominated by the largest impact factors , we can in principle test this hypothesis .
if , instead of eq . , we fix an arbitrary maximum parameter @xmath196 , it is straightforward to show that we obtain [ result_bmax ] _ dyn = c n^-1()^2-d+1 , where @xmath197 is a numerical coefficient ( depending only on @xmath8 and @xmath39 ) .
if we now assume that @xmath198 we obtain [ result_bmax - mod ] _ dyn ~n^ , where @xmath199 .
the case @xmath200 corresponds to the assumption we have made up to now , and the result .
the case @xmath201 corresponds , on the other hand , to the assumption that @xmath196 scales in proportion to the inter - particle distance ( as originally assumed by chandrasekhar @xcite ) .
now we have seen in section [ results - gne1 ] that the scaling of the relaxation rate for the simulated cases @xmath202 and @xmath203 are @xmath204 , which are in agreement with @xmath205 and hence @xmath206 . in the specific case @xmath207 , i.e. gravity in @xmath7 , it is in fact possible to _ quantify _ the maximum impact factor rather than just its scaling . instead of eq .
( replacing @xmath147 by @xmath194 following the discussion in sect . [ softenings ] )
we have [ result_bmax_grav ] _ dyn = d n^-1 ( ) , where @xmath208 is a ( calculable ) numerical coefficient . using the simulations presented in sect .
[ simu - grav ] , we can fit very well the relaxation rate with [ scaling_grav_eps ] _ dyn = ( ) . comparing these last two equations , we have that @xmath209 , and , further , that @xmath210 .
this size corresponds with the sharp fall - off of the density profile shown in the inset of fig .
[ n1_rho_n ] . to check that @xmath196 does not depend on @xmath5
, we did another set of simulations with the same parameters but @xmath211 particles . from these we obtained the scaling of the relaxation rate as a function of @xmath147 plotted in fig .
[ slope_n1_eps ] , in which , according to eq .
, the relaxation rate has been multiplied by a factor of eight .
we thus obtain very good agreement with the predicted scaling .
our findings confirm therefore the results of farouki & salpeter @xcite , who found that the maximum impact parameter should be taken of order of the size of the system .
in this paper , we have studied collisional relaxation in systems of particles interacting with a power - law potential @xmath212 , introducing a regularization of the singularity in the force as @xmath213 when necessary . in our analytical calculations
we have generalized the chandrasekhar approach " in the case of gravity to such potentials .
we have also included the contribution of hard collisions rather than just weak collisions , in which the mean field trajectories of the particles are weakly perturbed , which is the approximation usually found in the literature , see e.g. @xcite .
we have found that the collisional dynamics is dominated by * weak collisions , if @xmath214 , and * hard collisions , if @xmath159 , while the case @xmath215 , which corresponds to gravity in @xmath7 , is at the threshold .
moreover we considered the large @xmath5 , mean field ( or vlasov ) limit scaling of the two body relaxation rate , assuming the considered particle system to be in viral equilibrium . _ in absence of force regularization _
( other than an infinitesimal one assumed implicitly to make two body collisions defined for @xmath216 ) , we found that this rate , expressed in units of the characteristic time for mean - field dynamics @xmath23 , vanishes in the large @xmath5 for @xmath217 , and diverges in this limit for @xmath218 .
this means that only in the former case does the mean - field limit of the dynamics exist for a virialized system ; in the latter case it does not because the collisional relaxation completely dominates the mean field dynamics . only in the former case , therefore , can a qss be expected to exist on a physically relevant time scale .
this leads to the following _ dynamical _ classification of interactions : 1 .
power - law interactions is _ dynamically _ long - range if @xmath219 for a sufficiently large number of particles , and in particular @xmath220 , which occurs for @xmath217 .
the interaction is _ dynamically _ short range if @xmath221 for a sufficiently large number of particles , and in particular @xmath222 , which occurs for @xmath217 .
this classification was proposed initially @xcite on the basis of a formal analysis of convergence properties of the force on a particle in the thermodynamic limit , and subsequently in @xcite on the basis of the analysis detailed here .
it has also been justified using different analytical approaches to the full kinetic theory of such systems @xcite .
as noted in the introduction , this classification differs from the usual one used to distinguish long range from short range interactions , according to the thermal equilibrium of the system , in which the important feature is the integrability of the _ potential_. there is therefore a range of @xmath8 , @xmath223 , in which the interaction is _ dynamically _ short range , but long - range according to its thermal equilibrium properties . in this case , if the number of particles is sufficiently large , there will be no qss ( as in short range systems ) , but the thermal equilibrium state will present the typical features of a long - range system , i.e. , spatial inhomogeneity , inequivalence of ensembles etc .. we have also generalized these scalings when the inter particle potential is regularized ( `` softened '' ) at small scales . with this regularization the case @xmath224 ( in which the potential barrier can not prevent the particles to collide for pure power
law potentials ) becomes well defined . in this case , the relaxation rate depends on the value of the softening length @xmath147 for interactions in which small impact factors play a predominant role , i.e. , @xmath195 .
we have presented , for @xmath7 , detailed numerical results which support our theoretical findings . we have confirmed previous results in the literature for the gravitational case @xmath46 , notably for the scaling relations satisfied by the relaxation rate as function of the softening @xmath147 and the number of particles @xmath5 .
furthermore , using the scaling of the relaxation rate with @xmath147 , we have found very strong numerical evidence that the maximum impact parameter is related with the size of the system and not microscopic scales such as the inter - particle distance .
we have simulated also dynamically long - range cases @xmath225 and @xmath188 , in which the collisional relaxation is dominated by collisions around the minimum impact parameter , obtaining again very good agreement with the theoretical scalings . for
dynamically long - range systems dominated in our calculations by collisions with the largest impact parameter , we have found find , as predicted , that a softening in the potential does not affect the relaxation rate .
the natural extension of this work is the numerical study of collisional relaxation allowing strong collisions , in order to check the scalings of this regime derived in this paper . for such study
, it is necessary to develop very refined integration schemes in order to integrate properly such collisions .
another interesting perspective is to study the problem with a more rigorous approach using the angle - action variables ( with probably also many approximations because it is a very complicated formalism ) in order to describe more precisely the relaxation dynamics , and in particular study more precisely the validity of the chandrasekhar approximation as a function of the range of the interaction @xmath8 .
we acknowledge many useful discussions with j. morand , f. sicard and p. viot .
this work was partly supported by the anr 09-jcjc-009401 interlop project and the cnpq ( national council for scientific development , brazil ) .
numerical simulations have been performed at the cluster of the sigamm hosted at `` observatoire de cte dazur '' , universit de nice
sophia antipolis and the hpc and visualization resources of the centre de calcul interactif hosted by universit nice sophia antipolis .
it is interesting to derive eq . simpler method which can give more physical insight .
we can compute the change in perpendicular velocity integrating the perpendicular component of the force for all the duration of the collision , assuming that the relative trajectories is unperturbed with constant relative velocity @xmath14 : f_= ^-(+1 ) , the change in the perpendicular component of the velocity in a time @xmath226 is thus [ integral0 ] |_| & = & _ -t_c^t_c dt ^-(+1 ) + & = & _ ^- ds ( 1+s^2)^-(+1 ) + & & ( ) ^_-^ ds ( 1+s^2)^-(+1 ) ; [ deflection ] taking the limit @xmath227 and performing the integral we obtain exactly .
the potential @xmath228 is , for @xmath229 , exactly v(r,)=. we define @xmath230 .
for @xmath231 we use the following form of the potential for _ soft _ core softenings : * @xmath202 : v(u,1)^1/2=15.75 u^2 - 22.5u^3 + 8.75u^4 , * @xmath232 : v(u,1)^3/4= 11.875 u^2 -17.4167 u^3 + 6.875 u^4 , * @xmath233 : v(u,1)= 10 u^2 - 15 u^3 + 6 u^4 , * @xmath187 : v(u,1)^5/4= 8.925 u^2 - 13.65u^3 + 5.525 u^4 , * @xmath234 : v(u,1)^3/2= 8.25 u^2 - 12.8333333 u^3 + 5.25 u^4 . | * abstract * we explore the formation and relaxation of so - called quasi - stationary states ( qss ) for particle distributions in three dimensions interacting via an attractive radial pair potential @xmath0 with @xmath1 , and either a soft - core or hard - core regularization at small @xmath2 . in the first part of the paper we generalize , for any spatial dimension @xmath3 , chandrasekhar s approach for the case of gravity to obtain analytic estimates of the rate of collisional relaxation due to two body collisions .
the resultant relaxation rates indicate an essential qualitative difference depending on the integrability of the pair force at large distances : for @xmath4 the rate diverges in the large particle number @xmath5 ( mean field ) limit , unless a sufficiently large soft core is present ; for @xmath6 , on the other hand , the rate vanishes in the same limit even in the absence of any regularization . in the second part of the paper we compare our analytical predictions with the results of extensive parallel numerical simulations in @xmath7 performed with an appropriate modification of the gadget code , for a range of different exponents @xmath8 and soft cores leading to the formation of qss .
we find , just as for the previously well studied case of gravity ( which we also revisit ) , excellent agreement between the parametric dependence of the observed relaxation times and our analytic predictions .
further , as in the case of gravity , we find that the results indicate that , when large impact factors dominate , the appropriate cut - off is the size of the system ( rather than , for example , the mean inter - particle distance ) .
our results provide strong evidence that the existence of qss is robust only for long - range interactions with a large distance behavior @xmath6 ; for @xmath9 the existence of such states will be conditioned strongly on the short range properties of the interaction . |
among the radio pulsars that are found in binary systems , four belong to a subgroup where the pulsar has the characteristics of a young , isolated pulsar , but is orbiting a massive companion ( @xmath5m@xmath1 ) in a rather wide ( 502000day orbital period ) and eccentric ( @xmath6 ) orbit @xcite . the companion to the first of these systems ,
psrb1259@xmath263 @xcite was identified as a @xmath7 early - type b2e star .
psrj0045@xmath27319 , located in the smc , is the second system and has a @xmath8 b1v companion @xcite .
no information on a possible counterpart has been published for the fourth system , psrj1638@xmath24725 @xcite .
the third system , psr j1740@xmath23052 , was discovered by @xcite in the parkes multibeam survey @xcite as a young ( 350kyr ) pulsar .
subsequent timing observations showed that the pulsar is part of a binary system with an orbital period of 231d and an eccentricity of 0.57 .
the mass function of the pulsar yielded a lower limit to the mass of the companion of 11m@xmath1 .
a candidate counterpart to psrj1740@xmath23052 was identified in near - infrared observations . from @xmath9-band spectra ,
the counterpart was classified as late - type , having a spectral type between k5 and m3 @xcite .
this identification is at odds with the binary parameters which set the roche lobe radius significantly smaller than that expected for a late - type star at the estimated distance .
subsequent phase - resolved spectroscopic observations by @xcite confirmed the unlikeliness of the star being the binary companion , as no significant radial velocity variations consistent with the expected binary orbit were found .
the nature of the massive binary companion remains unclear , and could be either an early - type main - sequence star or a compact remnant . though @xcite reported small variations in the dispersion measure of psrj1740@xmath23052 near periastron , which would be consistent with a stellar companion , they could not conclusively rule out a stellar mass black hole as the binary companion
. in this paper we report on interferometric radio observations to obtain an accurate position of the psrj1740@xmath23052 and adaptive optics corrected near - infrared observations to search for a stellar counterpart to psrj1740@xmath23052 .
section2 describes the interferometric observations , while section3 details the astrometry and photometry of the near - infrared imaging .
we discuss the results in section4 and conclude in section5 .
the nrao very large array ( vla ) was used to determine the position of the pulsar under project code as769 .
the observation was made over a 5h period starting at utc 23:19 on september 3rd , 2003 using the vla in its most extended a - configuration .
this observation included the pie town antenna for even greater east - west resolution .
spectral line mode mode 2ac was used to provide 25mhz of bandwidth centered at 1678.4mhz on each circular polarization . the correlator integration time
was set to 5s .
the observation consisted of a 600s observation of 3c286 followed by 14 loops containing three sources : 90s on pmnj1751@xmath22524 , 90s on pmnj1820@xmath22528 and 930s on psrj1740@xmath23052 .
data reduction was performed within aips in a standard manner .
flux density and bandpass calibration were performed on 3c286 .
calibrator source pmnj1751@xmath22524 was used as the position reference .
its assumed position was taken to be @xmath10 , @xmath11 @xcite . phases and amplitudes determined on this source were applied to pmnj1820@xmath22528 and psrj1740@xmath23052 .
imaging was performed using the robust weighting factor of unity as a compromise between resolution and sensitivity ; the resultant beam size was @xmath12 oriented @xmath13 east of north .
wide - field imaging of the psrj1740@xmath23052 data was performed in order to clean side - lobes from numerous bright sources over the entire @xmath14 field of view .
most notably , the @xmath15 diameter object known as _ the tornado _ ( see @xcite and references within ) , is only @xmath16 southwest of the pulsar . due to the lack of short interferometer spacings ,
the ability to model and clean this source well was limited .
fortunately the pulsar lies outside the worst side - lobes which emanate mostly north and south from the object .
the pulsar was detected less than one beam - width from the position reported by @xcite with a flux density of @xmath17mjy . despite the best effort to clean the side - lobes of confusing sources ,
the pulsar was found on top of residual large scale structure . in order to fit the position of the pulsar ,
linear flux density slices were made along the right ascension and declination axes through the center of the detected pulsar .
each of these two slices were then fit for a 1-d gaussian and an underlying linear gradient .
the resultant best fit position of the pulsar is @xmath18 , @xmath19 .
the @xmath20 uncertainties on this position are @xmath21 in right ascension and @xmath22 in declination .
the effect of an incorrectly modeled underlying gradient was found to result in deflections much smaller than the fit uncertainty so are not considered .
the final uncertainty reported above is entirely due to measurement error .
the 0.8jy vla calibrator source pmnj1820@xmath22528 has nearly the same angular separation ( @xmath23 ) from the calibrator source pmnj1751@xmath22524 as does psrj1740@xmath23052 ( @xmath24 ) and was used as an astrometric check source .
it was imaged using the same parameters as used for psrj1740@xmath23052 .
its fit position ( @xmath25 , @xmath26 ) differed from its cataloged position ( @xmath27 , @xmath28;@xcite ) in each axis by an amount much smaller than the statistical uncertainty of the pulsar position .
adaptive optics corrected near - infrared imaging of the field containing psrj1740@xmath23052 was obtained with naco ( naos - conica ; @xcite ) at the eso very large telescope on paranal , chile in july and august of 2006 .
a log of the observations is given in table[tab : log ] .
images were obtained in the @xmath29 , @xmath30 and @xmath31-bands using the s27 camera , which provides a 1k@xmath321k chip with a @xmath33pix@xmath34 pixel scale yielding a @xmath35 field - of - view .
the late - type candidate counterpart of @xcite was used as a natural guide star for the wavefront sensor which corrects the atmospheric distortions in the incoming wavefront .
the ` n90c10 ` dichroic was used for @xmath31 observations , while the ` k ` dichroic was used for @xmath29 and @xmath30-band imaging .
the dithered naco observations were used to create a sky frame containing the contribution of both the dark current and the inhomogeneities in the sky for each filter and exposure time combination .
after subtraction of this sky frame , the science frames were registered using integer pixel offsets and averaged .
our aim was to obtain an accurate absolute astrometric calibration of the naco images using the usno ccd astrograph catalog ( ucac3 ; @xcite ) to allow comparison with the interferometric pulsar position .
the positional uncertainties of the standards in this catalog are 20mas for stars with @xmath36 and 70mas for @xmath37 , but with 1.3 stars per square arc - minute the star density in the direction of psrj1740@xmath23052 is low and none of the standards coincides with the naco field - of - view . in order to allow calibration against the ucac3 catalog , imaging observations were obtained with two further instruments . narrow - band near - infrared images at 1.19 and 2.25@xmath38 m were taken with isaac @xcite , also at the vlt in june 2006 .
the total exposure times were 936s and 800s respectively .
this instrument has a 1k@xmath321k ccd with a @xmath39pix@xmath34 pixel scale with a @xmath40 field - of - view .
the observations were corrected for dark current and flat - field effects and registered and averaged using standard routines .
finally , @xmath41 , @xmath42 and @xmath43-band images were obtained with the wide field imager ( wfi ; @xcite ) at the 2.2 m eso telescope on la silla , chile in august 2006 .
this instrument is a mosaic of 8 2k@xmath324k ccds each with a pixel size of @xmath44pix@xmath34 and a @xmath45 field - of - view .
we only use the data from the ccd that contains the pulsar location .
bias - subtraction and flat - fielding were preformed using standard routines . the @xmath42 and @xmath43-band images were further corrected for fringe frames which were constructed from the dithered science observations .
all data reduction was done using the munich image data analysis system ( midas ) .
stars on an @xmath46 subsection of a 45s @xmath41-band wfi image were compared to ucac3 standards .
a total of 79 astrometric standards coincided with the image , of which 72 were not saturated and appeared stellar and unblended .
an astrometric solution was determined by fitting for position and a four parameter transformation matrix . by
iteratively removing outliers the solution converged using 67 standards providing rms residuals of @xmath47 in right ascension and @xmath48 in declination .
this astrometric solution was transferred from the wfi image to the naco images by using the 1.19@xmath38 m and 2.25@xmath38 m isaac images to generate tertiary and quarternary astrometric catalogs .
pixel positions on the isaac images were corrected for geometric distortion using the may 2005 transformation provided by eso .
the transfer of the astrometric solution from the @xmath41-band wfi image to the 1.19@xmath38 m isaac image used 56 stars , providing rms residuals of @xmath49 in right ascension and @xmath50 in declination .
the solution was then transferred to the 2.25@xmath38 m image , using 120 stars yielding rms residuals of @xmath51 in right ascension and @xmath52 in declination .
finally , 29 stars were used to transfer the solution to the naco @xmath31-band image obtained on july13 .
the residuals of the transfer were @xmath51 in right ascension and @xmath53 in declination .
the total uncertainty in the astrometry of the naco image is the quadratic sum of the uncertainties in the astrometric solutions , which amount to @xmath54 in right ascension and @xmath55 in declination and is dominated by the uncertainties in the calibration of the wfi image by the ucac3 catalog . even though multiple transfers of the astrometric solution are needed , it is still more accurate than the direct calibration of the 2.25@xmath38 m isaac image using 160 stars from the 2mass catalog @xcite , which yields residuals of @xmath56 in both coordinates .
combined with the uncertainty of the interferometric position of psrj1740@xmath23052 the 95% confidence region on the naco images is an ellipse with a semi - minor axis of @xmath57 in right ascension and a semi - major axis of @xmath58 in declination .
this confidence region is shown in fig.[fig : fig1 ] and is located @xmath59 west and @xmath60 south of the late - type star , which itself is located just inside the 95% confidence region .
inspection of the naco images shows that the july 13th , 2006 @xmath31-band image contains a star located inside the error ellipse , see fig.[fig : fig2 ] .
the position of the counterpart in that image is @xmath61 and @xmath62 , which is offset from the interferometric position of psrj1740@xmath23052 by @xmath63 in right ascension and @xmath64 in declination .
based on the stellar density of the @xmath31-band image , we estimate that the probability of finding a random star this close to the interferometric position of psrj1740@xmath23052 is only 0.05percent .
.naco observation log .
the columns give the date and time of the observations , the integration time @xmath65 , the filter used , the observed fwhm @xmath66 of the stellar profiles and the observed magnitude or limit on the magnitude in this filter used . [ cols="<,<,>,^,^,>",options="header " , ] subsection of the @xmath31-band images obtained with naco on july13th , 2006 .
the ellipse denotes the 95% confidence error region of the interferometric radio position of the pulsar .
the bright star near the center is the late - type star which @xcite showed is unlikely to be the binary companion to psrj1740@xmath23052 .
this star is located on the edge of the error region .
only a fainter star is present in the error ellipse .
the @xmath67 box is the field shown in fig.[fig : fig2].,width=302 ] subsection of fig.[fig : fig1 ] centered on the bright late - type star , showing the original image and with the stellar profile of the late - type star subtracted .
some artifacts from the late - type star are still present due to imperfect definition of the point - spread - function .
the ellipse denotes the 95% confidence error region on the interferometric position of psrj1740@xmath23052 .
the counterpart to psrj1740@xmath23052 is encircled in both panels.,width=302 ] point - spread - function ( psf ) photometry was performed on the averaged naco images using daophotii @xcite running within midas .
an analytical moffat profile @xcite with a fixed exponent of @xmath68 combined with an empirical look - up table was used to represent the psf .
the parameters of the psf were kept constant over the image . due to the low apparent stellar density on the naco images , the majority of the stars on the images were used to determine the psf .
the instrumental @xmath69 psf magnitudes were calibrated against values from the 2mass catalog @xcite . due to the small field - of - view of naco ,
4 stars were used to determine zero - point magnitude offsets .
no color terms were fitted for .
the rms residuals of the fits were 0.09mag in @xmath29-band and 0.06mag in both @xmath30 and @xmath31 .
the counterpart to psrj1740@xmath23052 is detected in the july 13th @xmath31-band image , where it has @xmath3 .
limits on the magnitude of the counterpart in the other images were determined using simulations . for each image a set of copies were generated where an artificial star of a certain magnitude was placed on the position of the counterpart .
each of these copies was then photometered using the same routines as the original .
the @xmath70 detection limit was set when the artificial star was recovered with a signal - to - noise of 3 .
the limiting magnitudes are given in table[tab : log ] .
the most stringent limits set @xmath71 and @xmath72 .
we will now use the @xmath31-band detection and @xmath29 and @xmath30-band non - detections of the counterpart and compare them with values predicted for the companion of psrj1740@xmath23052 .
the radio timing observations by @xcite place several constraints on the properties of the companion . assuming the pulsar has a canonical mass of 1.4m@xmath1 , the measurements of the orbital period and semi - major axis limit the mass of the companion to @xmath73m@xmath1 .
the radius of the companion is constrained by the radius of the roche lobe at the periastron distance of the orbit .
a further upper limit on the radius is set by the lack of eclipses of the pulsar seen in the radio observations .
figure[fig : fig3 ] shows these constraints .
assuming the magnetic field of the pulsar is a dipole , the rotational properties of the pulsar can be used to determine the age of the pulsar . for psrj1740@xmath23052
this characteristic age is 350kyr , and binary evolution dictates that the companion must be older than that , as the prior evolution of the pulsar progenitor must be involved @xcite .
3052 . assuming a canonical pulsar mass of 1.4m@xmath1 , the orbital parameters measured by @xcite limit the companion to a mass of @xmath73m@xmath1 .
the roche lobe radius at periastron is shown with the dashed line , setting an upper limit to the companion radius . a further upper limit , shown with the dotted line , on the companion radius
is set by the absence of eclipses of the radio signal .
stars with radii larger than this limit would eclipse the pulsar when it passes the line - of - sight .
stellar isochrones from the padova models by @xcite are shown with the solid lines , ranging from zero - age main - sequence models to models at ages of 0.5 , 1 , 2 , 5 , 10 and 20myr .
each isochrone shows the location where the star reaches absolute @xmath31-band magnitudes of @xmath74 , @xmath75 and @xmath76 .
the observed range of @xmath77 shows that the companion is on the main - sequence.,width=302 ] at a galactic longitude of @xmath78 and latitude of @xmath79 , psrj1740@xmath23052 is located only @xmath80 from the galactic center , and hence the electron and dust column densities are substantial . at a dispersion measure ( dm ) of 740.9@xmath81pc@xmath34 @xcite ,
psrj1740@xmath23052 is one of the pulsars with the largest dm . for its line - of - sight ,
distance estimates from models of the galactic electron density distribution based on the observed dm range from 8 to 14kpc @xcite and 6.3 to 11kpc @xcite .
the cobe dust maps by @xcite set the maximum reddening for this line - of - sight at @xmath82 . using the @xcite relative extinction coefficients ,
this is equivalent to a maximum @xmath83 and @xmath31-band absorption of @xmath84 and @xmath85 .
the nearest distinct line - of - sight of the galactic extinction model by @xcite is within @xmath86 of psrj1740@xmath23052 and estimates @xmath31-band absorption of @xmath87 at a distance of 5.0kpc , up to @xmath88 at 10.4kpc .
this model determines the extinction distribution by comparing 2mass observations with predictions of a galactic stellar population synthesis model @xcite . for comparison ,
the galactic extinction model by @xcite , where red clump stars in the 2mass observations are used to derive distance dependent extinction , estimates @xmath89 of 1.2 at 5.0kpc , 1.7 at 10kpc and 3.0 at 15kpc .
hence , based on the observed @xmath31-band magnitude of the counterpart and a conservative estimate on the range of distances ( @xmath90 to 15kpc ) and absorptions ( @xmath91 to 3.0 ) , the absolute @xmath31-band magnitude of the counterpart falls in the range of @xmath92 to @xmath93 .
stellar isochrones by @xcite show that main - sequence stars have @xmath94 .
hence the observed limit on @xmath4 limits the @xmath95 reddening to @xmath96 . using the @xcite extinction coefficients
this sets @xmath97 , which is consistent with the estimates from galactic absorption models
. the observed limit on @xmath98 does not provide a better constraint .
figure[fig : fig3 ] shows the predictions for mass and radius of the @xcite isochrones .
they show that zero - age main - sequence stars with masses up to 25m@xmath1 will have absolute @xmath31-band magnitudes that satisfy the observed range of @xmath92 to @xmath93 .
less massive but more evolved main - sequence stars up to ages of 20myr at 11m@xmath1 also satisfy this range .
the constraints on the mass and brightness of the companion of psrj1740@xmath23052 show that the companion must be on the main - sequence .
furthermore , since the isochrones predict that main - sequence stars more massive than 11m@xmath1 have @xmath99 the @xmath31-band distance modulus must be larger than @xmath100 , which suggests psrj1740@xmath23052 is located at the far end of the distance ranges estimated from galactic electron distribution models .
since its discovery , it was unclear if the binary companion of psrj1740@xmath23052 was an early - type main - sequence star , a late - type giant , or a stellar mass black hole , though the data favored a main - sequence type companion . using accurate astrometric radio interferometry and adaptive optics corrected near - infrared imaging observations ,
we show that the late - type star located near the pulsar position is located at the edge of the 95% confidence error ellipse of the interferometric radio position of the pulsar .
this further strengthens the case made by @xcite and @xcite that it is not the binary companion to psrj1740@xmath23052 . in the wings of the late - type star , and near the center of the error ellipse
, we find a counterpart whose observed @xmath31-band magnitude is consistent with the expected mass and age of the binary companion to psrj1740@xmath23052 at the estimated distance and reddening to the system .
we argue that this counterpart is the binary companion to psrj1740@xmath23052 .
our observations show that the companion is on the main - sequence , which is consistent with the observations of both the small variations in the dispersion measure near periastron of the binary orbit , and the periastron advance being due to tidal and spin quadrupoles @xcite .
these observations rule out the need for a black hole or a giant as the binary companion . at @xmath101
the companion is within reach of near - infrared spectrographs .
adaptive optics would be a necessity to resolve the companion in the wings of the late - type star which is almost 6mag brighter in @xmath31 . to extract the spectrum of the companion an algorithm such as that of @xcite would be needed to resolve the blend .
spectra of the companion could be used to further confirm the identification through radial velocity variations and the spectral classification . with a massive main - sequence companion ,
the psrj1740@xmath23052 system likely experienced a previous phase of mass transfer in which the progenitor of the pulsar lost its hydrogen envelope and the companion accreted a significant amount of mass ( for a review , see @xcite ) .
the subsequent ( type ib ) supernova caused the orbit to become eccentric .
in the next few million years the companion will evolve off the main - sequence and tidal forces will circularize the orbit when the star approaches roche lobe overflow . the ensuing common envelope phase will likely lead to a merger and the formation of a thorne - zytkow object @xcite as there is not enough energy in the orbit to unbind the envelope of the subgiant ( e.g.@xcite ) .
the national radio astronomy observatory ( nrao ) is a facility of the national science foundation operated under cooperative agreement by associated universities , inc .
this research is based on observations made with eso telescopes at the la silla or paranal observatories under program i d 077.d-0683 .
pulsar research at ubc is supported by an nserc discovery grant .
r. , hartung m. , brandner w. , et al .
2003 , in m. iye & a. f. m. moorwood ed . ,
society of photo - optical instrumentation engineers ( spie ) conference series vol .
4841 of society of photo - optical instrumentation engineers ( spie ) conference series , naos - conica first on sky results in a variety of observing modes .
pp 944952 g. , lacombe f. , puget p. , et al .
2003 , in p. l. wizinowich & d. bonaccini ed . , society of photo - optical instrumentation engineers ( spie ) conference series vol .
4839 of society of photo - optical instrumentation engineers ( spie ) conference series , naos , the first ao system of the vlt : on - sky performance .
pp 140149 i. h. , manchester r. n. , lyne a. g. , kaspi v. m. , camilo f. , bell j. f. , damico n. , kramer m. , crawford f. , morris d. j. , possenti a. , mckay n. p. f. , lumsden s. l. , tacconi - garman l. e. , cannon r. d. , hambly n. c. , wood p. r. , 2001 , , 325 , 979 m. h. , bassa c. g. , jacoby b. a. , jonker p. g. , 2005 , in rasio f. a. , stairs i. h. , eds , binary radio pulsars vol .
328 of asp conference series , optical studies of companions to millisecond pulsars .
asp , p. 357 | we report on the identification of a near - infrared counterpart to the massive ( @xmath0m@xmath1 ) binary companion of pulsar j1740@xmath23052 .
an accurate celestial position of psrj1740@xmath23052 is determined from interferometric radio observations .
adaptive optics corrected near - infrared imaging observations show a counterpart at the interferometric position of the pulsar .
the counterpart has @xmath3 and @xmath4 .
based on distance and absorption estimates from models of the galactic electron and dust distributions these observed magnitudes are consistent with those of a main - sequence star as the binary companion .
we argue that this counterpart is the binary companion to psrj1740@xmath23052 and thus rule out a stellar mass black hole as the pulsar companion .
[ firstpage ] binaries : general stars : early - type pulsars : general pulsars : individual : psrj1740@xmath23052 infrared : stars |
* experimental setup : * the liquid viscosity was varied in our experiments by using glycerin / water mixtures of various mass fractions , 0 , 40 , 50 , 60 , 65 , 70 , 75 and 80% of glycerin .
the impact velocity was changed by releasing the drops from between 0.16 - 2.16 m height . by varying the circular steel nozzle size
, we used four different drop diameters .
further details of the experiments can be found in thoroddsen et al .
@xcite .
* numerical simulations : * direct comparison between the experimental observations and numerical results with gerris @xcite were done on a set of different impact conditions .
a convergence study of a bumping case showed that a minimum cell size 1,000 times smaller than the drop diameter was necessary to observe the bumping shapes , explaining why previous studies did not observe this evolution .
however , a much finer grid is needed to reproduce the evolution of very thin ejecta sheets .
this was accomplished using up to 16 levels of refinement , which means separating the domain in two smaller cells 16 times in each direction ( the size of the smallest cell is @xmath31 times smaller than the domain ) . in our simulations , this corresponds to 28,800 cells per drop diameter . at this level of refinement
, an equivalent uniform grid would have more than 4 billion cells ( @xmath32 ) , while our simulation have of the order of 20 million cells . at the highest level of refinement ,
the ejecta sheet can become as thin as 500 nm ( only 3 cells ) near the tip , due to the extreme stretching .
this value is consistent with indirect experimental measurements @xcite .
the curvature of the tracer isolines was estimated by fitting a bzier curve of order 3 through 5 points . for smaller curvature cases ,
the number of points was increased to 11 , improving the estimate of the position of maximum curvature . | the splashing of a drop impacting onto a liquid pool produces a range of different sized micro - droplets . at high impact velocities ,
the most significant source of these droplets is a thin liquid jet emerging at the start of the impact from the neck that connects the drop to the pool .
we use ultra - high - speed video imaging in combination with high - resolution numerical simulations to show how the ejecta gives way to _ irregular splashing_. at higher reynolds number , its base becomes unstable , shedding vortex rings into the liquid from the free surface in an axisymmetric von krmn vortex street , thus breaking the ejecta sheet as it forms .
liquid drop splashing is part of our daily lives , from the morning shower to natural rain @xcite .
while it has been studied for more than one hundred years @xcite , it is only recently that advances in high - speed imaging techniques @xcite have revealed its early dynamics @xcite .
splashing refers herein to the breakup of a drop into smaller droplets during impact .
understanding the underlying mechanism which produces the smallest droplets is important for example for the number of microscopic aerosols which remain when those satellite droplets evaporate .
such aerosols affect human health and can act as nucleation sites during cloud formation . for high speed drop impact on a liquid pool , the _ ejecta sheet _
is the first stage leading to splashing .
it was first observed in the inviscid numerical simulations of weiss & yarin @xcite and in the experiments of thoroddsen @xcite .
when the drop impacts at higher velocity , the speed of these ejecta sheets increases and they become thinner . the radial stretching of the sheets reduces their thickness even further and they can remain intact even at thicknesses well under a micron @xcite . when they eventually rupture , they can produce a myriad of very fine spray droplets .
however , this mechanism does not continue for ever , at a critical reynolds number the smooth ejecta gives way to a more random splashing , which counterintuitively may produce fewer small droplets . to understand the mechanisms leading from continuous ejecta sheets to _
irregular splashing _ , a systematic study of the early dynamics was conducted with ultra - high - speed video imaging , over a range of impact velocities @xmath0 , liquid viscosities @xmath1 and droplet diameters @xmath2 @xcite .
figure [ fig : kre ] shows a classification of the results in terms of reynolds number @xmath3 , where @xmath4 is the liquid density , and splashing parameter @xmath5 , which relates to the weber number @xmath6 , where @xmath7 is the surface tension , i.e. @xmath8 .
we are interested here in the higher @xmath5 regime , where splashing occurs @xcite . ) irregular splashing .
( @xmath9 ) bumping .
( ) quartering .
( ) protrusions rising up along the side of the drop . ] the classification in fig .
[ fig : kre ] focuses on the ejecta shapes . in the lower range of @xmath10 ( more viscous liquids )
, a smooth ejecta sheet emerges between the drop and the pool ( ) .
however in the highest range of @xmath10 , isolated droplets emerge from the neck , followed by a disturbed liquid surface , and no coherent ejecta can be identified , i.e. _ irregular splashing _ occurs ( @xmath11 ) . in the intermediate regime ( @xmath12 ) , the ejecta sheets show a large variety of repeatable shapes .
we have grouped them into 3 classes . at lower @xmath5 ( lower impact velocities , ) ,
surface tension prevents the formation of an ejecta sheet .
however , we observe some protrusions travelling up along the side of the drop , without ejection of droplets outwards . at higher @xmath5 ( ) ,
the ejecta sheet is more developed ; however , it stays attached to the drop , stretching the ejecting sheet between the expanding tip of ejecta and the drop entering the pool .
this stretching can lead to the explosive rupturing of the sheet , which generates fast droplets of a large range of sizes through slingshot @xcite . in the upper range of @xmath5 ( @xmath9 ) we observe an intriguing phenomenon where the free - standing sheet interacts strongly with the downwards moving drop surface .
this is shown in the sequence of fig .
[ fig : bumping](a ) , herein referred to as the _ bumping _ of the ejecta .
the ejecta is strongly bent by the drop , and then folds at its apex .
overall snapshot of a bumping ejecta was included in thoroddsen _
et al . _
@xcite [ their fig .
8(c ) ] . .
from top to bottom , observations at time 30 , 80 , 130 , 180 and 230 @xmath13 after contact .
( a ) experimental observation .
the static dark points correspond to dust on the camera sensor .
the video was taken at 200,000 frames per second .
( b ) numerical simulation of the drop impact for the same conditions as in the experiment presented in ( a ) . in the last image , the leading part of the ejecta sheet becomes smaller than the grid size , by stretching between the apex and the tip , and thus breaks into non - physical droplets .
the axisymmetric simulations can not include the three - dimensional effects , such as the breakup of the tip observed in ( a ) .
supplemental videos show the two evolutions @xcite . ]
those experimental results clearly show the effect of the reynolds number on the transition towards _
irregular splashing_. moreover , it shows new dynamics of the ejecta sheet interacting with the drop .
it suggests that those interactions could underlie the irregular splashing .
to test this idea , we have chosen to reproduce the impact by numerical simulations .
it is only recently that numerical simulations managed to identify the ejecta sheet @xcite , due to the extreme range of scales involved , and the challenges of interfacial flow simulations @xcite .
the intricate shapes observed herein were thus not seen numerically until now .
we use the freely available code gerris @xcite , for its high parallelization and dynamic adaptive grid refinement , which allow us for the first time to reach enough precision to fully resolve the dynamics of the ejecta .
this code uses the _ volume - of - fluid _ method to solve the incompressible navier - stokes equations .
axisymmetric simulations faithfully reproduced all of the experimentally observed features , as we demonstrate in fig .
[ fig : bumping ] for the bumping case .
the shape of the drop in the simulation is perfectly spherical , ruling out the hypothesis that small deviations from spherical drop shapes in the experiments could be responsible for the drop interaction with the ejecta sheet . to study this transition to irregular
splashing we increase @xmath10 , while keeping @xmath5 constant , from a smooth ejecta sheet [ fig .
[ fig : kre ] ( ) ] to irregular splashing ( @xmath11 ) .
this was done for two different @xmath5 values , corresponding to the bumping ( @xmath9 ) and quartering ( ) regimes . at @xmath14 .
quantities are nondimentionalized by the drop diameter d , the drop impact velocity u and the drop entry time @xmath15 .
( a ) definition sketch .
the base of the ejecta sheet is defined as the segment between the two points of maximum curvature of the interface ( t on the drop side and b on the side of the pool ) .
the angle of the ejecta sheet @xmath16 is the angle between the horizontal and the normal to the base .
( b ) evolution of the ejecta base radial position , defined as the distance from the axis of symmetry to the middle of tb in ( a ) , vs. the nondimensional time @xmath17 , for @xmath10 from 1,000 to 6,000 .
the solid curve is @xmath18 , where @xmath19 is the radius where an undisturbed drop would meet the pool .
( c ) evolution of @xmath16 ( in degrees ) for @xmath10 from 1,000 to 6,000 .
( d ) same curves as ( c ) , where the angle is scaled by @xmath20 .
( e , f ) evolution of the maximum positive vorticity ( red ) and maximum absolute negative vorticity ( blue ) in the liquid near the ejecta base for @xmath21 ( e ) and @xmath22 ( f ) .
the positive maximum is located near t , and the negative maximum near b. ] the position of the base of the ejecta @xmath23 [ fig .
[ fig : base](a , b ) ] follows very closely the geometric relation predicted by josserand & zaleski @xcite , @xmath24 , independent of @xmath10 , where @xmath25 is the radius where an unperturbed drop would meet the pool surface .
pure kinematics suggests that the ejecta emerges at an angle intermediate between the drop surface and the pool , thus @xmath16 should increase with @xmath25 .
a simple geometric model @xcite suggested that @xmath16 increases as @xmath26 , where @xmath27 is the time nondimensionalized by @xmath15 , whereas the simulations show that @xmath16 grows linearly before bumping [ fig .
[ fig : base](c ) ] .
however , the ejecta rises faster for higher @xmath10 .
the collapse of the curves in fig .
[ fig : base](d ) shows that @xmath16 grows at a rate proportional to @xmath20 .
the angle of the ejection velocity vector at the middle of the base also follows a similar trend , and increases proportionally to @xmath20 . at lower @xmath10
, @xmath16 increases slowly enough for the ejecta to escape the drop .
however , from @xmath28 , the ejecta sheet rises too fast , thus impacting the drop surface .
the resulting bumping sharply decreases @xmath16 .
this interaction of the drop and the ejecta sheet observed experimentally occurs earlier at higher reynolds numbers , eventually breaking the ejecta sheet .
this is consistent with the interpretation that this interaction is responsible for the irregular splashing observed at higher @xmath10 .
vorticity also appears to play an important role in the dynamics of the ejecta sheet [ fig .
[ fig : base](e , f ) ] .
the flow around a curved free surface generates vorticity ( see for instance batchelor @xcite 5.14 ) .
therefore , concentrated vorticity is observed in the simulations near points t and b at the base of the ejecta , as the flow moves around the base to enter the ejecta . at the early stage of the ejecta formation , both sides of the base produce a similar strength of vorticity .
this initial vorticity scales as @xmath20 as observed previously @xcite .
however , the difference in vorticity ( absolute values ) between the two sides is observed to increase initially linearly with time , before decreasing again .
moreover , this difference is higher for larger @xmath10 [ fig .
[ fig : base](f ) ] . by looking closely at the neck region during the impact
, we can identify fundamental changes in the vorticity structure as @xmath10 is increased ( fig .
[ fig : vortstructures ] ) .
note that in fig .
[ fig : vortstructures](a ) most of the liquid in the sheet originates from the pool , in agreement with dye visualizations @xcite . for the lower range of @xmath10 [ fig .
[ fig : vortstructures](a , d ) ] , the vorticity stays concentrated near the free surface at the neck of the ejecta sheet . as there is stronger vorticity generated at the top of the ejecta base
, a vorticity layer of one sign separates the drop and the pool liquids but it remains stable .
@xmath5 affects the shape of the outer part of the ejecta sheet , as we observe by comparing fig .
[ fig : vortstructures](a , d ) , consistently with experimental observations @xcite . for intermediate @xmath10 ,
the interface remains stable in its early evolution . in the bumping case ( b ) , the rising ejecta sheet contacts the downwards moving drop surface .
this creates a shear instability , generating a toroidal vortex structure around the entrapped bubble . in the quartering case ( e )
, the ejecta sheet leaves the neck region to climb up the drop , pulled by higher surface tension .
this also creates a shear instability between the climbing liquid from the pool and the drop liquid moving down , forming a row of vortex rings of the same sign .
these vortices near the free surface leave their signature @xcite by creating waves below the rising sheet , a feature also observed experimentally .
however , all such vortical effects are absent from inviscid theory and simulations @xcite at even higher @xmath10 [ fig .
[ fig : vortstructures](c , f ) ] , vorticity is shed behind the base of the ejecta sheet , in a way reminiscent of the von krmn vortex street , here forming alternating - sign vortex rings . for the first 7 shedding cycles , the local reynolds number based on the radial speed and width of the neck takes value around 70 and the strouhal number @xmath29 is around @xmath30 , in good agreement with related krmn streets .
moreover , the free surface at the base is also affected by the shedding , reinforcing the oscillations @xcite . during the early shedding [ fig . [
fig : vortstructures](g ) ] , surface tension effects are higher because of the sharper surface geometry . as the angle of the neck increases , the amplitude of the oscillations increases .
the ejecta can then climb on the drop at lower @xmath5 ( f ) , or impact alternatively on the drop and the pool ( c ) , in a similar way to the bumping , entrapping a row of bubble rings @xcite .
four bubble rings can be clearly identified in fig .
[ fig : vortstructures](c ) , and a fifth one being created .
only well resolved bubbles and droplets ( larger than 30 cells ) are kept in the numerics , suggesting that smaller bubbles could be entrapped earlier . from systematic experimental observations , reproduced with axisymmetric simulations
, we have detailed a new mechanism explaining the irregular splashing of a water drop .
previously studied mechanisms have described the droplet separation from the rim of the ejecta @xcite , or the destabilization of a liquid sheet @xcite .
our mechanism however explains the breakup of the ejecta sheet by the destabilization of its base , through vortex shedding from the free surface .
the authors are grateful to a. prosperetti and g. agbaglah for comments , b. marchand and a. clo for assistance , t. el sayed for sharing numerical resources , and h. pottmann for advices on curvature calculations .
we thank also the gerris community for their support .
m .- j.t . and
s.t.t .
were partly supported by kaust - berkeley aea grant 7000000028 .
_ fluid dyn .
res . _
* 12 * , 61 ( 1993 ) .
_ annu .
rev .
fluid mech . _ * 38 * , 159 ( 2006 ) .
_ proc .
r. soc .
lond . _ * 34 * , 217 ( 1882 ) .
_ ieee trans .
electron devices _ * 50 * , 144 ( 2003 ) .
_ annu .
rev .
fluid mech .
_ * 40 * , 257 ( 2008 ) .
_ phys .
rev .
lett . _ * 106 * , 034501 ( 2011 ) .
_ j. fluid mech . _ * 690 * , 5 ( 2012 ) .
_ j. fluid mech . _ * 385 * , 229 ( 1999 ) .
_ j. fluid mech . _ * 451 * , 373 ( 2002 ) .
_ proc .
r. soc .
lond . a _ * 373 * , 419 ( 1981 ) . _
int .
j. multiph . flow _
* 21 * , 151 ( 1995 ) .
_ exp
. fluids _ * 22 * , 463 ( 1997 ) .
_ phys .
fluids _ * 23 * , 022105 ( 2011 ) .
_ phys .
fluids _ * 15 * , 1650 ( 2003 ) .
_ direct numerical simulations of gas - liquid multiphase flows _
( cambridge univ . press , 2011 ) . _
j. comput .
phys . _ * 228 * , 5838 ( 2009 ) . _
c. r. mec . _ * 339 * , 194 ( 2011 ) . _ gerris flow solver _ http://gfs.sf.net ( 2011 ) . _ an introduction to fluid dynamics _
( cambridge , 1967 ) .
_ j. fluid mech . _ * 218 * , 547 ( 1990 ) .
_ j. fluid mech . _
* 542 * , 1 ( 2005 ) .
_ annu .
rev .
fluid mech . _ * 36 * , 413 ( 2004 ) .
_ chem .
eng .
sci . _ * 57 * , 3639 ( 2002 ) . _
c. r. acad .
sci .
iib mec .
_ * 326 * , 839 ( 1998 ) .
_ phys .
fluids _ * 22 * , 122105 ( 2010 ) . _ chem .
eng .
sci . _ * 18 * , 203 ( 1963 ) .
_ j. fluid mech . _ * 462 * , 341 ( 2002 ) . |
the rotational evolution of young pulsars ( @xmath5 years old ) is often observed to be interrupted by glitches - sudden increases in spin frequency @xmath6 .
the range of glitch sizes @xmath7 to @xmath8 and modes of post - glitch recovery is observed to be quite diverse across the pulsar population @xcite .
one of the most studied pulsars , vela , quasi - periodically produces large glitches @xmath9 @xcite ; the relatively large number of glitches observed from vela ( 21 ) has made it a test - bed for proposed glitch mechanisms .
much effort has been devoted to examining two - component glitch models , in which one component of the neutron star is for most of the time decoupled from the solid crustal lattice to which the magnetic field lines are anchored and whose rotational evolution we directly observe .
this decoupled component acts as an angular momentum reservoir , and occasionally re - couples to the rest of the star , transferring some of its angular momentum and spinning the star up @xcite .
the most studied class of two - component models posit that the angular momentum reservoir consists of the free superfluid neutrons that permeate the inner crust .
the mechanism by which they are decoupled from the core , and periodically recoupled , is still a matter of considerable uncertainty owing to the difficulty in marrying the complex microphysics with the global rotational dynamics of the fluid components . in the pinning paradigm ,
the neutron superfluid vortices in the crust could interact with the lattice of nuclei to become ` pinned ' to the crust @xcite , therefore preventing them from moving radially outwards from the rotation axis in response to the secular spin - down of the star @xcite .
they thus become decoupled from the rotational evolution of the charged component of the crust and that part of the core that couples strongly to it . as a lag builds up between the angular frequency of the charged component ( that which we observe ) and the frequency of the crustal superfluid
, the magnus force acting radially outwards on the vortices grows , until eventually it overcomes the pinning force , and the vortices unpin _ en masse _ and couple to the charged component , transferring angular momentum . despite recent advances in combining the microphysics and hydrodynamics of this model @xcite certain aspects of the model such as the exact unpinning trigger and a convincing account of the process by which a large number of vortices subsequently moves remain to be addressed @xcite .
in addition , it is possible that pinning of neutron vortices occurs elsewhere in the star @xcite , or that at least some glitches arise by other means such as hydrodynamical instabilities and turbulence @xcite . in this paper
we consider the class of models in which the free inner crustal neutrons are the component which acts as the effective angular momentum reservoir . on
can analyze the tenability of these models in general terms relatively free of the uncertainties outlined above .
the observed glitch activity of the vela pulsar can be used to infer the fractional moment of inertia of the angular momentum reservoir @xmath10 compared to that of the portion of the star it couples to at the time of glitch , @xmath11 ( which we shall referred to as the charged component of the star ) , @xmath12 @xcite .
this is the most stringent constraint from any pulsar , with the next largest glitch activity observed coming from psr j0537 - 6910 @xmath13 @xcite from which we have observed the most glitches of any pulsar .
assuming an upper limit to the moment of inertia of the angular momentum reservoir to be that of the entire crustal neutron superfluid @xmath14 , and that the charged component it couples to is essentially the whole star @xmath15 ( i.e. the entire core couples to the charged component of the crust on glitch - rise timescales ) , many realistic neutron star ( ns ) equations of state ( eoss ) can satisfy the condition @xmath12 , and constraints can be placed on the ns eos @xcite .
however , recent calculations of the strength of the entrainment of superfluid neutrons by the crustal lattice via bragg scattering suggest that only a fraction of the crustal neutrons are effectively free , and consequently the upper limit to the moment of inertia of the angular momentum reservoir is reduced : @xmath16 .
initial analyses suggested that makes @xmath17 too small to explain the observed glitch sizes @xcite , and suggests one must at least involve other components of the star in addition to the crust superfluid as the store of angular momentum .
these studies assume a strong coupling between crust and core so that @xmath18 .
more recently , however , analyses using a wider range of possible equations - of - state ( eoss ) suggest that a crustal angular momentum reservoir might still be consistent with vela glitch observations @xcite . however , estimates of the crust - core coupling timescales due to interactions of neutron vortices with core protons ( which may form a type i or type ii superconductor ) @xcite raise the possibility of only a small fraction of the core neutrons being coupled to the crust on the glitch rise timescale @xmath19s @xcite . therefore it is possible that @xmath20 , allowing the ratio @xmath17 to satisfy the lower bound of 1.6% again with only crustal superfluid neutrons involved for a wider range of eoss . the aim of this paper is to examine the range of predictions for @xmath21 taking into account systematically variations in the most uncertain nuclear matter parameters over their experimentally and theoretically constrained ranges , and crustal entrainment in a way that is consistent with the crust models using the recently calculated values in @xcite .
we will also go beyond similar studies by allowing for a fraction @xmath22 of core neutrons to be coupled to the crust at the time of glitch .
we shall apply systematically and consistently generated sequences of crust and core eoss together with the relevant crust compositions @xcite to calculate the relevant moments of inertia including that of the free crustal neutrons explicitly .
the consistent modeling of crust and core properties when exploring the dependence of neutron star observables has been presented before @xcite , and here extends to modeling the crust thickness , density of superfluid neutrons throughout the crust , core eos and core proton fraction using the same underlying nuclear matter eoss
. much effort has been devoted to constraining the eos of nuclear and neutron star matter , particularly through constraining the density dependence of the symmetry energy at nuclear saturation density @xmath23 , parameterized by @xmath24 where @xmath25 is the pressure of pure neutron matter at saturation density , which is strongly correlated with the pressure in neutron stars at that density .
nuclear experimental probes ( e.g. @xcite ) give a conservative range @xmath26 mev , although some more recent results on the nuclear experimental side @xcite , as well as tentative constraints from neutron star observation @xcite and from _ ab - initio _ pure neutron matter calculations @xcite tend to favor the lower half of that range ( although , for a counter - example , see e.g. @xcite ) . particularly , combining inferred values of @xmath0 from several nuclear experimental probes suggests a range @xmath4 mev @xcite .
the high - density behavior of the eos is even more uncertain both theoretically and experimentally @xcite , even if one restricts the composition to purely nucleonic matter , with some of the only constraints coming from analysis of heavy - ion collisions @xcite and the recent observations of @xmath27 neutron stars @xcite . in this paper
we shall explore the impact of systematically varying the density dependence of the symmetry energy @xmath0 at saturation density on the the predictions for @xmath28 . in section 2
we describe our glitch modeling and set of eoss and our calculation of observable quantities . in section 3
we present and discuss our results and in section 4 we discuss our conclusions .
the observed angular frequency , @xmath29 , of a pulsar is presumed to be that of its ionic crustal lattice in which the magnetic field lines are anchored .
when considering glitch sizes and immediate post - glitch evolution , it is important to define that component of the star strongly coupled to the lattice on timescales comparable with the glitch rise time , which is observationally constrained to be @xmath30 40s @xcite . in our minimal model of the core which contains purely nucleonic matter , this component contains the core protons and some fraction of the core neutrons , and we shall refer to it as the charged component of the star . the moment of inertia of a star of radius @xmath31 in the limit of small angular frequency @xmath29 @xcite is given by @xmath32 where @xmath33 is energy density of matter in the star , @xmath34 is the pressure and @xmath35 is the mass contained in radius @xmath36 .
@xmath37 is a radially - dependent metric function given by @xmath38 and @xmath39 is the frame dragging angular velocity @xmath40 where @xmath41 for @xmath42 .
the charged component of the star includes the crust lattice plus the protons in the core and some fraction of core superfluid neutrons .
neutron vortices can entrain protons and become magnetized , allowing electrons to scatter off them and therefore coupling to the charged component on timescales of @xmath43s for the vela pulsar @xcite . in regions where protons are type ii superconductors
, they form fluxtubes to which neutron vortices can become pinned , coupling them to the charged component on timescales of days or longer @xcite . by comparing the above estimates of coupling timescales with the upper limit of the glitch rise time @xmath44s , one may infer that only some fraction of the core neutron superfluid will contribute to the charged component of the star at the time of glitch .
that fraction is quite uncertain , and enters into the model as a free parameter @xmath1 , but above estimates indicate that it is possible to have @xmath45 @xcite .
we also denote the total neutron fraction of the core at a given radius @xmath36 by @xmath46 , determined by the eos .
then the moment of inertia of the charged component can be expressed @xcite [ fig:2 ] @xmath47 e^{-\nu(r)}\frac{\bar{\omega}(r)}{\omega}\frac{\left ( \mbox{\large{$ \varepsilon $ } } ( r)+p(r ) \right)}{\sqrt{1 - 2gm(r)/r}}\mathrm{d}r.\ ] ] the total moment of inertia of superfluid neutrons in the inner crust of the star is given by @xmath48 where @xmath49 is the energy density of crustal superfluid neutrons , @xmath50 is the pressure of the crustal superfluid neutrons and @xmath51 and @xmath52 are the radius boundaries for the inner crust .
entrainment of superfluid neutrons by the crust s lattice reduces the mobility of the neutrons with respect to that lattice .
it can be shown that this effect is encoded by introducing an effective `` mesoscopic '' neutron mass @xmath53 @xcite ; larger values correspond to stronger coupling between the neutron superfluid and the crust , and a reduction in the fraction of superfluid neutrons able to store angular momentum for the glitch event .
one can include this effect by modifying the integrand eq .
[ eq : new1 ] by multiplying by a factor of @xmath54 where @xmath55 is the effective mass at radius @xmath36 in the crust .
we obtain @xmath55 from the results of chamel @xcite by interpolating between the values calculated at specific densities to find the effective mass at arbitrary locations in the inner crust .
it is yet to be determined how entrainment is modified when one includes hitherto neglected effects such as the spin - orbit interaction in the calculations@xcite . in order to account for possible variations in the strength of entrainment , and to systematically trace the effect of introducing entrainment
, we introduce a parameter @xmath56 which we use to control the strength of the entrainment : 1 + ( m_n^ * m_n - 1)e where @xmath57 corresponds to no entrainment and @xmath58 corresponds to the full strength entrainment as predicted by @xcite .
we will consider also entrainment reduced by 50% @xmath56=0.5 and entrainment enhanced by 50% @xmath56=1.5 .
the analysis of @xcite of the vela pulsar glitches , recently updated to include the most recently observed glitches @xcite identifies the minimum amount of angular momentum stored in the crustal superfluid reservoir @xmath59 relative to that of the charged component of the star @xmath60 with the parameter @xmath61 , defined by g = 0.016 where @xmath62 is the glitch activity parameter of the pulsar , the fraction of frequency spin - down reversed by glitches determined by a linear fit to a plot of the cumulative relative glitch size over time @xcite .
the microphysical ingredients in the glitch model include the total pressure and energy density @xmath63 , @xmath64 and those of the superfluid neutrons @xmath65 , @xmath66 as a function of baryon density throughout the core _ and _ the crust , as well as the crust - core transition baryon density @xmath67 and the mesoscopic effective mass of neutrons in the crust @xmath68 . in order to calculate
the crust and core eoss a model for uniform nuclear matter is required .
nuclear matter models can be characterized by their behavior around nuclear saturation density @xmath69 @xmath70 , the density region from which much of our experimental information is extracted .
we can denote the energy per particle of nuclear matter around saturation density by @xmath71 , where @xmath72 is the baryon density and @xmath73 the isospin asymmetry , where @xmath74 is the proton fraction .
@xmath75 corresponds to symmetric nuclear matter ( snm ) , and @xmath76 to pure neutron matter ( pnm ) . by expanding @xmath77 about @xmath78
we can define the _ symmetry energy _
@xmath79 , [ fig:3 ] [ eq : eos1 ] e(n , ) = e_0 ( ) + s(n)^2 + ... , which encodes the energy cost of increasing the isospin asymmetry of matter . expanding the symmetry energy about @xmath80 where the density parameter @xmath81 , we obtain [ eq : eos3 ] s(n ) =
j + l + k_sym ^2 + ... , where @xmath82 , @xmath0 and @xmath83 are the symmetry energy , its slope and its curvature at saturation density . at present , the energy of snm around saturation density is well constrained by experiment .
much experimental effort has been focused on determining the symmetry energy @xmath82 and its density dependence @xmath0 around nuclear saturation density @xcite , and from these results we take as a conservative range @xmath84 mev in this work .
we will also pay attention , however , to the fact that the congruence of the experimental results @xcite favors a range @xmath85 mev .
we calculate the crust and core eoss and the transition density consistently using the widely used skyrme nuclear matter model . as the baseline skyrme parameterization
, we choose the skiufsu model used in previous work @xcite , which shares the same saturation symmetric nuclear matter ( snm ) properties as the relativistic mean field ( rmf ) iufsu model @xcite , has isovector nuclear matter parameters obtained from a fit to state - of - the - art pnm calculations , and describes well the binding energies and charge radii of doubly magic nuclei @xcite .
two parameters in the skyrme model are purely isovector - they can be systematically adjusted to vary the symmetry energy @xmath82 and its density slope @xmath0 at saturation density while leaving snm properties unchanged @xcite .
the constraint from pure neutron matter at low densities induces correlation between the magnitude and slope of the symmetry energy at saturation density described by @xmath86 mev . in this work
we will create eoss by varying @xmath0 between 20mev and 120mev under this constraint . for softer symmetry energies at high densities ,
the resulting eos is matched onto two successive polytropic equations of state as described in @xcite in order to match the constraint on the maximum mass of @xmath87m@xmath88 @xcite .
the crust eos and crust - core transition densities used in this work are obtained from a simple compressible liquid drop model ( cldm ) for the crust @xcite . this model gives the composition of the crust ( including the free neutron density ) and the extent of the so - called ` pasta ' phases , in which nuclei become deformed into exotic shapes . to illustrate the relevant correlations between nuclear matter properties and neutron star properties , we plot in fig . 1 as a function of @xmath0 the crust - core transition densities @xmath67 and pressures @xmath89 ( left plot ) and the radii @xmath31 and inner crust thicknesses @xmath90 of a @xmath91 .
although @xmath67 decreases with @xmath0 , the relevant quantity that determines the crust thickness for a star of a given radius is @xmath89 , which increases with @xmath0 .
a larger stellar radius @xmath31 also gives a thicker crust , and the right plot shows the well - known positive correlation of @xmath31 with @xmath0 .
these two trends together give the increase of @xmath90 with @xmath0 shown in the right plot . over the range @xmath26 mev , the radius increases by just over 2 km and the inner crust thickness triples from approximately 300 m to 1000 m .
thus a stiffer symmetry energy will give a larger proportion of crust neutrons and we thus expect the capacity of the crust to deliver giant glitches will increase with @xmath0 . [ fig:4 ] fig .
2 shows , as a function of the radial distance from the center of a @xmath92 neutron star , the total enthalpy density @xmath93 and that of the neutrons only . in the inner crust
, we include the free neutrons _ and _ those bound in nuclei , and thus our results represent an upper limit to those that would be obtained using free neutrons alone .
we show the enthalpy profiles for @xmath94mev ( a ) , @xmath95mev ( b ) , and @xmath96mev ( c ) .
the location of the crust core transition density is indicated , as is the location of the density @xmath97@xmath70 above that , because we will later on consider extending the angular moment reservoir into the outer core .
the previously discussed trends of @xmath31 and @xmath90 with @xmath0 are evident once more
. a higher value of @xmath0 gives a higher value of the symmetry energy at super - saturation densities ; therefore the proton fraction in the core @xmath98 increases with @xmath0 , increasing the moment of inertia of the charged stellar component coupled to the crust at the time of glitch .
we begin by calculating the ratio of the moment of inertia of the inner crustal neutrons to that of the charged component of the star , @xmath61=@xmath99 , as a function of mass for a representative range of the model parameters ; selected results are displayed in fig . 3 . to recap , the model parameters are : the fraction of core neutrons coupled to the crust at the time of glitch @xmath1 , the strength of the crustal neutron entrainment @xmath56 and the slope of the symmetry energy at saturation density @xmath0 . as mass @xmath100 increases , so does the compactness of the star @xmath101 , while the relative thickness of the crust decreases , giving a smaller contribution to the moment of inertia of the star .
hence the generic trend of @xmath61 decreasing as mass increases apparent in all plots in fig .
3 . in fig . 3 ,
the inferred lower limit on @xmath61 , 1.6% , from the vela pulsar is shown as the horizontal bold line . in fig . 3 ( a - c ) we show @xmath61 as a function of mass @xmath100 for @xmath1 = 0.1 ( a ) , 0.5 ( b ) and 1.0 ( c ) , and in each plot we show results for @xmath102 20 , 70 and 120 mev ( softest to stiffest symmetry energy ) .
we use entrainment set at the level predicted @xcite , @xmath56=1 .
as @xmath0 increases ( i.e. the symmetry energy at saturation density becomes stiffer ) , @xmath61 becomes larger .
this relation is the result of the folding together of competing trends .
as @xmath0 increases , the radius of the star and the core proton fraction increases and thus so does the moment of inertia of the charged component of the star @xmath60 : if the crust thickness were held constant , this would result in a smaller value of @xmath61 .
in addition , though , as @xmath0 increases , the crust thickness increases ( fig .
1b ) ( as a result of the increasing radius _ and _ the increasing crust - core transition pressure ) thus increasing @xmath59 .
the overall trend only emerges in a consistent calculation .
it is important to note that the emergence of the relation between @xmath61 and @xmath0 is very sensitive to the particular relation between the crust - core transition pressure @xmath89 and @xmath0 , which is dependent on the family of eoss used @xcite .
we choose to vary @xmath0 while maintaining good fits to the results of low - density pure neutron matter calculations .
one can choose instead to vary @xmath0 while fixing the symmetry energy at sub - saturation densities , motivated by fits to nuclear mass models and neutron skin data @xcite , and this leads to a different ( and potentially non - monotonic ) dependence of @xmath89 on @xmath0 , and hence of @xmath61 on @xmath0 .
[ fig:5 ] increasing @xmath1 for a fixed mass star increases the fraction of the core superfluid that couples to the crust at the time of glitch and thus @xmath60 , therefore decreasing @xmath103 .
if the entire core is taken to be coupled to the crust , as it is in the analyses of @xcite , then we see from fig .
3c that we require stiff eoss @xmath104100 mev to obtain @xmath1051.6% for a 1.4@xmath2 star , in line with the analyses of @xcite .
these are stiffer eoss than are currently inferred from most experimental data . with full entrainment
, we have to lower the fraction of core neutrons coupled to the crust to @xmath1060.5 in order to satisfy @xmath1051.6% for a 1.4@xmath2 star for eoss in the range @xmath10770 mev , more in line with experimental constraints . a very low fraction of core neutrons coupled to the crust @xmath1 allows the whole range of eoss to comfortable
satisfy @xmath1051.6% for a 1.4@xmath2 star . in fig . 3 ( d - f ) we consider a range of possible strengths of entrainment relative to that calculated : no entrainment @xmath56=0 , reduced entrainment @xmath56=0.5 , full entrainment @xmath56=1.0 and enhanced entrainment @xmath108 .
results are displayed for @xmath0=70 mev over the same range of @xmath1 .
note that since the moment of inertia of the crustal neutrons which participate in the glitch depends inversely on the strength of entrainment , increasing that strength in equal increments has successively smaller effect on the predicted value of @xmath61 .
increasing entrainment by 50% compared to @xcite gives a relatively small reduction in @xmath61 ; at @xmath1=0.5 , our @xmath0=70 mev eos is still able to just satisfy @xmath1051.6% for a 1.4@xmath2 star .
equally , reducing entrainment by 50% still requires relatively stiff eoss to satisfy @xmath1051.6% ; for a fully coupled core @xmath1=1.0 , @xmath0=70 mev only marginally satisfies the constraint for @xmath56=0.5 . for a given set of model parameters
, we can identify a mass @xmath100 below which the vela constraint on the moment of inertia ratio , @xmath109 , is satisfied .
these upper mass limits are plotted as a function of @xmath0 in figs .
results are shown @xmath1 = 0.1 ( a ) , 0.5 ( b ) and 1.0 ( c ) . in each plot , the lines represent the upper limit on the mass as a function of @xmath0 for entrainment strengths of @xmath1100 , 0.5 , 1.0 and 1.5 .
masses which satisfy @xmath109 lie _ below _ the curve , as indicated by the arrows . when entrainment of crust neutrons is neglected @xmath57 , @xmath109 is satisfied for all values of the slope of the symmetry energy @xmath0 for masses below @xmath111 for @xmath112 , below @xmath113 for @xmath114 and for all masses considered up to @xmath115 for @xmath116 . for low coupled core neutron fractions down to @xmath116 , the vela constraint is satisfied for masses less than @xmath117 for all eoss even with enhanced entrainment @xmath56=1.5 . for a fully coupled core ,
even reduced entrainment restricts significantly the range of eoss that can satisfy the vela constraint to @xmath11870 mev for canonical neutron star masses @xmath119 . for each eos
( that is , for each value of @xmath0 ) one can find an upper limit on @xmath1 for which @xmath109 is satisfied below a given mass of neutron star .
5a - c plots those values of @xmath0 vs @xmath1 ; only values of @xmath1 _ to the left _ of the curves satisfy the observational constraint on @xmath61 as indicated by the arrows .
we show results for reduced entrainment @xmath56=0.5 ( a ) , full entrainment @xmath56=1.0 ( b ) and enhanced entrainment @xmath56=1.5 ( c ) .
each plot shows results for masses of 1.0 , 1.4 and 1.8@xmath2 .
as one can see from fig . 5c , we are able to satisfy the vela constraint for any eos for stars of mass @xmath119 and below if the coupled core neutron fraction is below @xmath120 even with enhanced entrainment . for full entrainment ,
to be consistent with _ any _ eos in the range @xmath121 30 - 60 mev from the concordance of a number of experimental results , then for full entrainment @xmath56=1.0 we require @xmath1220.2 for a @xmath119 star . for a high mass @xmath123 star ,
this is reduced to @xmath1220.1 .
if we only require consistency with an @xmath124mev eos , we require @xmath1220.5 for a @xmath119 star and @xmath1220.2 for a @xmath123 star .
there is a model dependence in the calculation of the crust - core transition density and pressure from both the eos and the method of calculation .
indeed , a different method of calculation and a different set of eoss has been shown to be able to satisfy the vela constraint for values of @xmath0 in line with experimentally inferred values @xcite . in addition , it is possible that core neutrons can also be involved in driving the glitch .
we therefore consider increasing the density range over which we calculate the moment of inertia of neutrons responsible for the glitch by an amount @xmath125 .
taking the canonical model of a fully coupled core @xmath1=1.0 and a @xmath119 star , we calculate the size of the density increase @xmath125 we must apply in order to satisfy the vela constraint for each value of @xmath0 .
the results are displayed in fig .
6 for reduced entrainment @xmath56=0.5 , full entrainment @xmath56=1.0 and enhanced entrainment @xmath56=1.5 .
one important point can be taken from these results . in order to satisfy the vela constraint in the canonical model for @xmath121 30 - 60 mev
, we require an increase in density of @xmath1260.01@xmath70 above our crust - core transition densities _ even with enhanced entrainment _ .
this is within the reasonable uncertainties in the crust - core transition density @xcite , indicating that even with @xmath1=1.0 we can not rule out the purely crustal origin of glitches .
alternatively , only a small amount of outer core neutrons are required to contribute to the angular momentum transfer to the crust ( perhaps through pinning on type ii superconducting flux tubes ) in order to satisfy the vela constraint .
[ fig:6 ] above the crust - core transition to which we must extend the region of neutrons which act as the angular momentum reservoir for the glitch in order to satisfy the vela constraint .
results are shown for a fully coupled core @xmath1=1.0 , a 1.4@xmath2 star and for reduced entrainment @xmath56=0.5 ( dash - dotted line ) , full entrainment @xmath127 ( dashed line ) and enhanced entrainment @xmath56=1.5 ( solid line).,width=302,height=226 ]
using neutron star crust and core eoss and compositions calculated using the same skyrme nuclear matter model we have explored whether crust neutrons provide a sufficient angular moment reservoir to reproduce the observed glitch activity of the vela pulsar allowing for the reduction of free neutrons by crustal entrainment .
we go beyond similar studies by allowing for only a fraction of core neutrons to be coupled to the crust on the timescale of the glitch .
the skyrme model contains two parameters that allow for systematic variation of the symmetry energy magnitude and slope at saturation density without changing symmetric nuclear matter properties , which we exploit to examine predictions spanning a conservative range of uncertainty in the density - slope of the symmetry energy @xmath128 mev , while simultaneously being constrained to fit the results of state - of - the - art pure neutron matter calculations at low densities .
other than identifying the crust neutrons as the angular momentum reservoir , our results are independent of the details of the glitch mechanism .
we compare our results to the average glitch activity of the vela pulsar , which results in the most stringent constraint , @xmath109 .
we explore variations in the following model parameters : the slope of the symmetry energy @xmath0 , the fraction of core neutrons coupled to the crust at the time of glitch @xmath1 , the strength of the crustal neutron entrainment @xmath56 relative to the results of bragg scattering of neutrons off crustal nuclei @xcite .
we take as the baseline model @xmath1 = 1.0 , @xmath56=1 and a 1.4@xmath2 star . for our baseline model , we find that we require stiff eoss @xmath3mev to generate glitch activities at the level observed in the vela pulsar for a 1.4@xmath2 star .
this is in agreement with the results of @xcite , although we have now quantified the required stiffness in terms of @xmath0 .
comparison with the tentative range @xmath12130 - 60 mev from the congruence of several experimental probes @xcite suggests that if the whole core is coupled to the crust , the crust neutrons are not sufficient to drive the glitch .
two other studies have conducted systematic explorations of equation - of - state parameter space in order to determine compatibility with the vela glitch activity @xcite .
both studies find _ some _ models that are compatible even with entrainment in full effect . in both cases ,
such models occupy a small and specific region of parameter space , but it is nevertheless necessary to outline the difference between these models and those used in this study .
@xcite arrived at the conclusion that the crust neutrons can drive large enough glitches for moderate values of @xmath12170 mev .
the core eoss used in that work are derived using families of parameterizations of relativistic - mean field models ; in our work , we use a parameterization of the non - relativistic skyrme model that is fit at first to a particular parameterization of one such rmf model , and therefore reproduces its predicted observables .
both rmf and skyrme models contain two purely isovector parameters - that is , two parameters whose adjustment affects only predictions of observables sensitive to neutron - proton asymmetries .
starting from our baseline parameterization , skiufsu we re - fit both purely isovector parameters in order to match the results of low - density pure neutron matter ( pnm ) calculations @xcite ( which we emphasize are expected to be quite robust . ) .
subseqeuntly , we vary @xmath0 by adjusting one of the isovector parameters , while adjusting the second to maintain our fit to the pnm calculations .
the study of @xcite starts from a number of baseline rmf parameterizations , all the results of fits to slightly different sets of experimental data as well as , in the case of the one model that satisfies the vela glitch activity , a hypothetical neutron skin measurement on @xmath129pb .
subsequently , only one of the two isovector parameters is adjusted ( an adjustment which is tracked by the resulting change in the neutron skin thickness of @xmath129pb ) in order to find the value that gives the highest value of the crust - core transition pressure @xmath130 and hence the best chance of reproducing the vela glitch activity .
the second isovector parameter remains fixed during this .
there is no comparison with the low density pnm eos , but such a procedure will result only in accidental agreement with the results of pnm calculations .
the matching to pnm matter becomes particularly important when we consider the second main difference between the work of @xcite and that presented here .
we use a model of the inner crust composition and eos , and crust - core transition pressure , that is fully consistent with the core eos in that it is derived from the same underlying skyrme model .
the key component of the inner crust eos in these studies is the low - density neutron fluid in the interstices of the crustal lattice ; it is therefore important to use our best knowledge of the low - density pnm eos when constructing such models , and ensure it is free of artifacts ( such as , for example , bound neutron matter ) .
the work of @xcite , while calculating the crust - core transition pressure consistently , uses a polytropic eos for the inner crust , which is matched onto the core eos ; such an eos will necessarily be inconsistent from a nuclear physics point of view with the core eos , and any low - density artifacts that may arise in their eoss as a result of ignoring the low - density eos in pnm matter will remain hidden . without a direct comparison with their models , we can not comment on
what difference this will make on the predictions of the crustal moment of inertia . the second study to conduct a systematic exploration of eos parameter space @xcite does use the results of low - density pnm matter calculations to parameterize the low - density eoss used ; they also use the same crustal eoss as used here .
nevertheless , they do find a small portion of parameter space consistent with vela glitch activity at moderate values of @xmath131mev .
the reason lies in the high density eos , which are polytropes in that work .
the models which are compatible with vela s glitch activity have particularly stiff eoss in the higher density regions of the core , giving rise to much larger radii and hence crustal thicknesses and moments of inertia .
such stiff eoss at high density predict high radii for 1.4@xmath2 neutron stars , in excess of 14 km , and can therefore be tested by future radius measurements . in our work ,
we use the same eos at high densities and low densities , and only append a polytrope at the highest densities when we need to stiffen the eos sufficiently to reproduce a 2@xmath2 neutron star .
therefore , the higher and lower density regions are more strongly correlated , and we lack much freedom to vary the high density eos independently of the low density component .
our high density eos is nevertheless based on a microphysical model , which is lacking from a simple polytropic description .
more work needs to be done on exploring the high - density eos parameter space on a microscopic footing , given is current uncertainty both theoretically and experimentally @xcite , even if one restricts the composition to purely nucleonic matter . even taking into account pure neutron matter calculations , there is sufficient uncertainty in the remaining parameters of the nuclear model underlying the eos to cause uncertainties of @xmath12610% in the crust - core transition density and pressure @xcite . to allow for this uncertainty ,
we examine how much the crust - core transition density would have to be increased in order to match the observations . in the range @xmath12130 - 60
mev we find an increase of @xmath1320.01@xmath70 is required , which is of order the eos model uncertainties .
thus we must conclude that for the baseline model , it is still marginally possible to match the vela glitch activity using crust neutrons with eoss that are also consistent with nuclear experiment _ and _ the results of low density pure neutron matter calculations .
one can also interpret the increase in density @xmath125 as allowing for the participation of core neutrons in the glitch mechanism , in which , given the set of eos parameters used here , we set a lower limit @xmath1330.01@xmath70 on the density above the crust - core transition that this participation must occur ( although this assumes that the participating core neutrons must be contiguous with the crust neutrons , which need not be the case ) .
we explore the effect of reducing and increasing entrainment for 50% . for reduced entrainment ,
we require @xmath13475 mev for @xmath1 = 1.0 , @xmath100=1.4@xmath2 . for enhanced entrainment , we require @xmath134115 mev .
we have noted that increasing entrainment by 50% ( which is equivalent to increasing the neutron effective masses in the crust by 50% ) does not greatly increase the problem of matching the vela results ; an increase in density @xmath1320.01@xmath70 would still provide enough neutrons to allow for such agreement .
we also consider lower and higher mass stars .
if vela is a high mass neutron star @xmath100=1.8@xmath2 , then we can not match the observed glitch activity for any eoss considered for @xmath1350.5 for full strength entrainment @xmath127 .
allowing for only a fraction of the core neutrons to be coupled to the crust at the time of the glitch , we find that for consistency with experiment @xmath13660 mev and for a 1.4@xmath2 star , we require @xmath1370.5 for @xmath56=1.0 , @xmath1370.7 for @xmath138 and @xmath1370.4 for @xmath108 . for a 1.8@xmath2 , we would require @xmath1370.2 for @xmath56=1.0 , @xmath1370.3 for @xmath138 and @xmath1370.1 for @xmath108 .
considering the crust - core coupling to be mediated by core mutual friction , @xmath1 can be related to the strength of mutual friction in the core via the crust - core coupling timescale @xcite and therefore , in principle , calculated consistently with the eos , thereby removing it as a free model parameter .
this would be a possible next step in a microphysically - consistent analysis of glitch models , to be reported in an upcoming paper .
note that the strength of mutual friction and hence @xmath1 will vary throughout the core as the composition and overall density changes . to summarize , for a 1.4@xmath2 star and a fully coupled core
, we require very stiff eoss to match glitch observations at odds with current experiment .
these results are consistent with @xcite within the set of eoss used here constrained by pure neutron matter calculations .
allowing for uncertainties in the crust - core transition pressure of order 10% would be sufficient to remove this problem ; alternatively , only a small amount of core neutrons need to be involved to alleviate the problem . if less than @xmath13940% of the core is coupled to the crust at the time of glitch , crust - originated glitches are possible for eoss consistent with experimental data and pure neutron matter calculations , even if the strength of crustal entrainment is increased to 50% above its predicted values .
we thank brynmor haskell and pierre pizzochero for invaluable discussions .
we thank the referee for help improving the manuscript .
this work is supported in part by the national aeronautics and space administration under grant nnx11ac41 g issued through the science mission directorate and the national science foundation under grants phy-0757839 and phy-1068022 . | in order to assess the ability of purely crust - driven glitch models to match the observed glitch activity in the vela pulsar , we conduct a systematic analysis of the dependence of the fractional moment of inertia of the inner crustal neutrons on the stiffness of the nuclear symmetry energy at saturation density @xmath0 .
we take into account both crustal entrainment and the fact that only a fraction @xmath1 of the core neutrons may couple to the crust on the glitch - rise timescale .
we use a set of consistently - generated crust and core compositions and equations - of - state which are fit to results of low - density pure neutron matter calculations .
when entrainment is included at the level suggested by recent microscopic calculations and the core is fully coupled to the crust , the model is only able to account for the vela glitch activity for a 1.4@xmath2 star if the equation of state is particularly stiff @xmath3 mev .
however , an uncertainty of about 10% in the crust - core transition density and pressure allows for the vela glitch activity to be marginally accounted for in the range @xmath4mev consistent with a range of experimental results .
alternatively , only a small amount of core neutrons need be involved .
if less than 50% of the core neutrons are coupled to the crust during the glitch , we can also account for the vela glitch activity using crustal neutrons alone for eoss consistent with the inferred range of @xmath0 .
we also explore the possibility of vela being a high - mass neutron star , and of crustal entrainment being reduced or enhanced relative to its currently predicted values .
[ firstpage ] dense matter - equation of state - stars : neutron - ( stars : ) pulsars : general - ( stars : ) pulsars : individual : vela |
_ planck _ ( ( * ? ? ?
* tauber et al . 2010 ) , ( * ? ? ?
* planck collaboration i. 2011 ) ) provides 8 full sky surveys at the frequencies of its lfi instrument ( 30 , 44 , and 70 ghz ) , and 5 full surveys at the 6 hfi frequencies in the range 100857 ghz .
the satellite rotates with 1rpm around an axis kept fixed for a pointing period ( pid ) of about 50 minutes , then shifts along the ecliptic by 2. this keeps a point source near the ecliptic in the main beam ( 2 fwhm ) for about @xmath0 pids ( depending on frequency ) for each survey ; for sources near the ecliptic poles the coverage can be much larger . analysing the time information in planck data for a particular sky direction
thus allows to search for variability in the sky signal .
variability mapping is based on four - dimenional healpix ( ( * ? ? ? * grski etal . 2005 ) ) constructs called 4d - maps , which record for every sky pixel @xmath1 all contributions of a given detector at times @xmath2 and beam orientation @xmath3 , where the index @xmath4 refers to planck pids . to construct an average sky signal @xmath5 free from beam orientation effects
, we use the _ artdeco _ beam deconvolution code ( ( * ? ? ?
* keihnen & reinecke 2012 ) ) .
a variability map is then a two - dimensional healpix map of a quantity @xmath6\ ] ] where + @xmath7 @xmath8 is the number of entries for pixel @xmath1 in the 4d signal map @xmath9 , @xmath10 is the beam - reconvolved average 4d - map based on the _ artdeco _ map @xmath5 , @xmath11 is the detector white noise , and @xmath12 combines all instrumental fluctuations which factor on the signal ( e.g. , calibration , inaccuracies of the beam model ) .
the definition of @xmath13 is motivated by the chernoff bound on the cdf @xmath14 , with for @xmath15 , and @xmath16 for @xmath17 .
the distribution of x - values over a sky map is indicateive not only of variability , but also to incorrect estimations of noise and instrumental variations ( see fig .
[ fig1 ] ) . , and detector noise @xmath18 as given in ( * ? ? ?
* planck collaboration ii .
( 2014 ) ) , thin dotted lines show the distribution expected for pure gaussian noise .
the tail to large values of @xmath13 indicates the presence of true sky variability .
a tail to large negative values of @xmath13 would indicate an overestimation of @xmath12 , an offset of the peak from @xmath19 an incorrect estimation of @xmath18.,scaledwidth=80.0% ]
as the analysis of time variations in the sky is essentially background - free , our method provides a way to extract time resolved planck fluxes down to a time resolution of a few hours .
if the position of a variable point source is known , the residual 4d - map , @xmath20 , can be beam - deconvolved to the source position by a simple division and thus provide an estimate of the flux variation , @xmath21 , where @xmath22 is a _ beamfactor map _ expressing the measured flux of a unit emitter at the source position in a beam centered at pixel @xmath1 and time @xmath2 .
this method is used , e.g. , to extract time - resolved planck fluxes for co - eval monitoring of blazars with the f - gamma program ( ( * ? ? ?
* fuhrmann et al . 2007 ) , ( * ? ? ?
* rachen et al . 2015 ) ) .
our analysis tools work on unified data structures , which ensures that all methods for variability analysis and flux extraction are applicable to lfi and hfi data in the same way .
interfaces for intitial 4d - mapping are in place at both the lfi and hfi dpc .
we expect that all planck frequencies will be analysed for variability , and science results prepared for publication well before the planck legacy release . | the sky is full of variable and transient sources on all time scales , from milliseconds to decades . _
planck _
s regular scanning strategy makes it an ideal instrument to search for variable sky signals in the millimetre and submillimetre regime , on time scales from hours to several years .
a precondition is that instrumental noise and systematic effects , caused in particular by non - symmetric beam shapes , are properly removed .
we present a method to perform a full sky blind search for variable and transient objects at all planck frequencies . |
the reduction of background levels is a key point of double beta decay experiments @xcite , not only by the use of radiopure materials but also by pattern recognition techniques .
this latter topic has been widely treated for germanium detectors @xcite , using the pulse shape to discriminate single from multisite events . in xenon
calorimeter approaches , the study made by gothard in @xcite for a small tpc introduced several ideas , like the presence of a big charge deposit at both ends of the track , and proved the effectiveness of using this distinctive double beta topological signature to discriminate signals from background .
the development of these ideas is the aim of @xcite , including the full simulation of the signal and background events in a high pressure xenon tpc ( hpxe tpc ) equipped with a pixelized detector .
this proceeding is a short summary of this paper and it also includes the latest work on discrimination algorithms . in the first section
, we present the complete simulation chain of a hpxe tpc with a pixelized readout .
the topology of the neutrinoless double beta decay ( @xmath2 ) of @xmath0xe and the expected backgrounds is then compared , based on the physical processes they suffer to deposit energy in the range of interest ( roi ) . from this comparison ,
two pattern recognition algorithms have been created and are described in the following section .
their rejection power is then presented and we finish with some conclusions and an outlook of possible improvements .
the simulation presented here replicates the physical processes involved in the generation of signals in a hpxe tpc .
it can be divided into three logical blocks .
the first one consists of a monte carlo simulation , including all the physical processes involved in the passage of gamma - rays or charged particles through matter .
it provides the interaction points and the corresponding energy losses within the gas ( figure [ fig : simulation ] , left ) .
this part of the simulation is based on decay0 @xcite + geant4 simulation packages @xcite and includes a conceptual design of a xenon tpc .
the second block describes the generation of electrons in the gas , the diffusion effects during the drift to the readout plane and the reconstruction of the 3d event by the readout .
we have considered a pixel size of @xmath3 @xmath4 , a gas pressure of 10 bar and two possible gas diffusion coefficients : low ( ldxe , figure [ fig : simulation ] , center ) and high ( hdxe , figure [ fig : simulation ] , right ) .
the final block includes the discrimination algorithms and is described in detail below .
of @xmath0xe in a hpxetpc at 10 bar , directly plotted from geant4 ( left ) and as acquired by a pixelized detector in ldxe ( center ) and hdxe ( right ) .
the event consists in two electrons emitted from the black point named vertex .
both electrons show a big energy deposit at the end of their tracks , marked by black stars . a pixel length of 5 mm has exceptionally been set to show the importance of diffusion.__,title="fig:",width=188 ] of @xmath0xe in a hpxetpc at 10 bar , directly plotted from geant4 ( left ) and as acquired by a pixelized detector in ldxe ( center ) and hdxe ( right ) .
the event consists in two electrons emitted from the black point named vertex .
both electrons show a big energy deposit at the end of their tracks , marked by black stars . a pixel length of 5 mm has exceptionally been set to show the importance of diffusion.__,title="fig:",width=188 ] of @xmath0xe in a hpxetpc at 10 bar , directly plotted from geant4 ( left ) and as acquired by a pixelized detector in ldxe ( center ) and hdxe ( right ) .
the event consists in two electrons emitted from the black point named vertex .
both electrons show a big energy deposit at the end of their tracks , marked by black stars . a pixel length of 5 mm has exceptionally been set to show the importance of diffusion.__,title="fig:",width=188 ]
in a detailed simulation , several contaminations must be considered to generate the expected background spectrum .
however , a first study can reduce the number of contributions to be simulated by the roi . in our case ,
we have fixed a roi between 2400 and 2500 kev , wider than the energy resolution window of current xenon experiments @xcite . 1 . *
two - neutrinos double beta decay mode * : for the half - life of @xcite and our roi , the background level is negligible . 2 .
* neutron activation * : the only worrisome isotope is the @xmath5xe .
it produces @xmath6 emissions with maximum energy of 4173 kev ( 67% ) and maximum energy of 3717.5 kev plus a @xmath7 of 455 kev ( 30% ) .
this contribution is irrelevant as the computed production ( 0.6 nuclei / year / kg ) is equal to the short half - life of the element ( 3.8 min ) .
* outer contaminations * : the contribution by muons ( direct and high energy photons ) is negligible in underground laboratories and the background level of gammas is @xmath8 kev@xmath9 kg@xmath9 yr@xmath9 with a 25 cm - thick lead shielding . 4 .
* inner contaminations * : both alpha and beta emissions are normally absorbed at the limits of the fiducial volume and can be easily rejected . from gamma emitters , there are only two important contributions : @xmath10tl ( @xmath7-line at 2614.5 kev ) and @xmath11bi ( @xmath7-line at 2447 kev and @xmath6 emissions with maximum energies of 3272 and 824 kev ) .
note that @xmath12co also contributes to the roi ( @xmath7-lines at 1173 kev and 1332 kev ) , but these events will be easily identified by their two separated energy deposits .
the topological features of background events in the roi are different depending on the physical processes that suffer .
they are also different from the expected signal , which is a continuous staggered track with a high energy deposit at both ends ( figure [ fig : discrimination ] , left )
. we will show in this section these differences , studying both @xmath10tl and @xmath11bi .
simple physical processes ( photoelectric , compton scattering or pair production ) of the @xmath10tl @xmath7-line can not generate an event in the roi , but by three more complex processes .
the first one is a compton scattering in an external volume followed by a photoelectric absorption or compton scattering in the gas .
the second one is a multi - compton scattering in the gas with a partial energy loss in form of low energy gammas . in these two cases ,
there are two energy deposits spatially separated ( figure [ fig : discrimination ] , center ) , whereas the expected signal will generally have only one track .
this fact does not appear in the third process , which consists in a photoelectric absortion followed by the emission of bremstrahlung photons which escape from the tpc .
however , it will normally show only a big energy deposit at one track s end .
most background events of @xmath11bi consist in the photoelectric absorption of the intense @xmath7-line at 2448 kev .
these events will have a single component but also only a big energy deposit at one track s end ( figure [ fig : discrimination ] , right ) .
other contributions are the compton scattering of less intense higher - energy @xmath7-lines and the @xmath6 emissions from the vessel s inner walls . in those cases
, there would be either multiple tracks or an energy deposit at the fiducial volume limits . in summary ,
two topological features have been identified , which are not present in background events : the absence of energy deposits far from the main component ( or _ trac_k ) and the presence of a big charge deposit ( generally called _
blob _ ) at both track ends . the discrimination algorithms based on these features
are described in the following section .
of @xmath0xe detected in a hpxe tpc .
the event consists in two electrons of 1.11 and 1.35 mev emitted from the black point named vertex .
both electrons show a big energy deposit ( or _ blo_b ) at the track s end , marked by a black star .
center : an event in the roi generated by @xmath10tl .
there are four tracks and the main one shows a big charge deposit at only one end .
right : a single - track event in the roi generated by @xmath11bi .
only one of the track s end show a blob.__,title="fig:",width=188 ] of @xmath0xe detected in a hpxe tpc .
the event consists in two electrons of 1.11 and 1.35 mev emitted from the black point named vertex .
both electrons show a big energy deposit ( or _ blo_b ) at the track s end , marked by a black star .
center : an event in the roi generated by @xmath10tl
. there are four tracks and the main one shows a big charge deposit at only one end .
right : a single - track event in the roi generated by @xmath11bi .
only one of the track s end show a blob.__,title="fig:",width=188 ] of @xmath0xe detected in a hpxe tpc .
the event consists in two electrons of 1.11 and 1.35 mev emitted from the black point named vertex .
both electrons show a big energy deposit ( or _ blo_b ) at the track s end , marked by a black star .
center : an event in the roi generated by @xmath10tl .
there are four tracks and the main one shows a big charge deposit at only one end .
right : a single - track event in the roi generated by @xmath11bi .
only one of the track s end show a blob.__,title="fig:",width=188 ]
each event is firstly transformed to a mathematical graph : pixels are identified with vertex and adjacent pixels are linked by a segment .
we then use a classical method of graph theory to find tracks or components @xcite .
most of background events are multiple - track but also only @xmath1360% of signal events are single - track , due to the emission of xenon x - rays ( energy @xmath13 30 kev ) or low energy bremsstrahlung photons .
these two effects produce a dramatic reduction of signal efficiency if only single - track events are selected . for this reason and
based one the factor of merit , events with one extra short track of less than 100 kev are also accepted .
this algorithm is based on the method to find the longest track - line in a graph @xcite . as a first step ,
several blob candidates are identified setting an energy threshold of 150 kev .
then , two track - lines are generated : one between two blob candidates and another between one blob and a normal pixel .
the algorithm choses the first line except if the second one is 30% longer than the first track - line .
this criterium is a compromise to recognize : 1 .
signals with two blobs or those where one of the electrons has not enough energy to generate a blob .
backgrounds generally show one blob but random energy deposits ( like @xmath14-rays , bremsstrahlung photons or low compton interacting close to the main track ) may create additional blobs at the main track . finally , the charge of both blobs is calculated considering a fixed radius around them .
the blob with less charge is a good parameter to select signal events , as already shown by gothard collaboration @xcite . in our case , we have set a threshold of 440 kev .
however , background events with fake blobs at the main track could not be rejected as the track - line will generally end at those charge deposits . in those cases ,
an important part of the charge will be out of this track - line and the track - to - total - charge ratio ( named _ track coverage _ ) can be used to reject them . fixing a 40% signal efficiency
, track criterium rejects mainly @xmath10tl events ( 4.8 - 7.1% background efficiency , depending on the vertex ) , as events in the roi suffer several physical processes .
the efficiency for @xmath11bi is modest ( 20% ) as most events are generated by a photoelectric absorption .
the topological criterium shows an efficiency of 12% for ldxe , near the value of 8.6% published by gothard @xcite , while the hdxe case is a factor 3 worse due to a modest blob identification .
current work is focused on a better determination of blob positions and the elimination of fine - tuning variables from our algorithms .
preliminary results show a cleaner separation of the charge distribution of the little blob ( figure [ fig : littlecharge ] , left ) , that respectively improves the background rejection efficency to 6.0 and 9.2% in ldxe and hdxe , independently of the vertex position and the type of contamination . as a result
, the background level is reduced more than two orders of magnitude in the roi ( figure [ fig : littlecharge ] , right ) . 0@xmath15
( blue line ) and @xmath10tl ( magenta line ) and @xmath11bi ( green line ) vessel contaminations .
events are in the roi and have been previously selected by the track criterium .
right : energy spectrum generated by @xmath10tl events emitted from the vessel ( with an activity of 10 @xmath16bq / kg ) in a ldxe after the succesive application of the selection criteria : initial ( black line ) , track ( red line ) , two - blobs ( blue line ) and fiducial volume ( magenta line).__,title="fig:",width=283 ] 0@xmath15 ( blue line ) and @xmath10tl ( magenta line ) and @xmath11bi ( green line ) vessel contaminations .
events are in the roi and have been previously selected by the track criterium .
right : energy spectrum generated by @xmath10tl events emitted from the vessel ( with an activity of 10 @xmath16bq / kg ) in a ldxe after the succesive application of the selection criteria : initial ( black line ) , track ( red line ) , two - blobs ( blue line ) and fiducial volume ( magenta line).__,title="fig:",width=283 ]
a detailed simulation of a hpxe tpc with a pixelized readout for a double beta decay experiment of @xmath0xe has been presented .
we have compared the topology of signal and backgrounds to identify two key topological features of signals : a single track and one big charge deposit at both ends .
two pattern recognition algorithms based on graph theory have been created to identify tracks and to detect blobs , with good results .
current work is focused on a better determination of blob position , the elimination of fine - tuning variables and the study of background rejection dependence on other tpc features like pressure and pixel size .
we are grateful to our colleagues of the groups of the university of zaragoza , cea / saclay as well as the next and rd-51 collaborations .
we acknowledge support from the european commission under the european research council t - rex starting grant ref .
erc-2009-stg-240054 of the ideas program of the 7th eu framework program .
we also acknowledge support from the spanish ministry of economics and competitiveness ( mineco ) , under contracts ref .
fpa2008 - 03456 and fpa2011 - 24058 , as well as under the cup project ref .
csd2008 - 00037 and the cpan project ref .
csd2007- 00042 from the consolider - ingenio 2010 program of the micinn .
part of these grants are funded by the european regional development fund ( erdf / feder ) .
f.i . acknowledges support from the eurotalents program .
9 f. t. avignone , s. elliott & j. engel , _ rev .
* 80 * ( 2008 ) 481 .
h. gmez _ et al .
_ , _ astropart .
_ * 28 * ( 2007 ) , 435 .
m. agostini _ et al .
_ , _ jinst _ * 6 * ( 2011 ) p04005 . h. t. wong _ et al .
. instrum .
a _ * 329 * ( 1993 ) 163 . s. cebrian _ et al .
_ , article in preparation .
pokratenko _ et al .
_ , _ phys . at
* 63 * ( 2000 ) 1282 .
s. agostinelli _ et al .
_ , _ nucl .
instrum .
a _ * 506 * ( 2003 ) 250 . j. allison _ et al .
_ , _ ieee trans . nucl .
* 53 * ( 2006 ) 270 .
n. ackerman _ et al .
* 107 * ( 2011 ) 212501 .
a. gando _ et al .
_ , _ phys .
c _ * 85 * ( 2012 ) 045504 .
b. bollobs , modern graph theory , springer - verlag , 1998 . | the observation of the neutrinoless double beta decay may provide essential information on the nature of neutrinos . among the current experimental approaches ,
a high pressure gaseous tpc is an attractive option for the search of double beta decay due to its good energy resolution and the detailed topological information of each event .
we present in this talk a detailed study of the ionization topology of the @xmath0xe double beta decay events in a high pressure xenon tpc , as well as that of the typical competing backgrounds .
we define some observables based on graph theory concepts to develop automated discrimination algorithms .
our criteria are able to reduce the background level by about three orders of magnitude in the region of interest of the @xmath0xe q@xmath1 for a signal acceptance of 40% .
this result provides a quantitative assessment of the benefit of topological information offered by gaseous tpcs for double beta decay search , and proves that it is a promising feature in view of future experiments in the field .
possible ideas for further improvement in the discrimination algorithms and the dependency of these results with the gas diffusion and readout granularity will be also discussed .
+ * email * : [email protected] |
light - front ( lf ) quantization is the natural frame - independent framework for the description of non - perturbative relativistic bound - state structure in quantum field theory . in principle
, one can solve qcd by diagonalizing the lf qcd hamiltonian @xmath11 , by using , for example , the discretized light - cone quantization method @xcite .
the spectrum and lf wavefunctions ( lfwfs ) , which contain hadronic information , are then obtained from the eigenvalues and eigenfunctions of the heisenberg equation @xmath12 .
the result is an infinite set of coupled integral equations for the lf components in a fock expansion @xcite .
unfortunately , solving these equations is a formidable computational task for the case of a non - abelian quantum field theory such as qcd in four - dimensional space - time .
consequently , alternative methods are necessary ; for a recent comprehensive review , see refs .
@xcite for details . in recent years ,
a semiclassical first approximation to strongly coupled qcd
light - front holographic ads / qcd has been developed @xcite .
this color - confining approach predicts the spectroscopy of light - quark hadrons , dynamical observables such as form factors and structure functions , and the behavior of the running coupling in the nonperturbative domain .
only one mass scale in addition to the quark masses appears .
this approach to hadron dynamics in physical four - dimensional spacetime at fixed lf time @xmath13 is holographically dual to the dynamics of a gravitational theory in five - dimensional anti - de sitter ( ads ) space .
the lf eigenvalue equation can be reduced in this theoretical framework to an effective single - variable quantum - mechanical wave equation for @xmath14 which is given by @xcite @xmath15 the function @xmath16 is the effective potential acting on the valence states @xcite ; it is holographically related to a unique dilation profile in ads space . as a result ,
one arrives at a concise form of a color - confining harmonic oscillator in impact space after the holographical mapping , @xmath17 .
the emergence of the mass scale @xmath18 is consistent with the procedure of de alfaro , fubini , and furlan @xcite in which a mass scale can appear in a hamiltonian without affecting the conformal invariance of the action@xcite .
the eigenvalues of the light - front schrdinger equation , eq .
( [ eq : scheq ] ) , are the squares of the meson masses .
the remarkably simple features of the empirical regge trajectories for both meson and baryon families are reproduced by lf holographic qcd with only one parameter , the mass scale @xmath18 @xcite .
the predictions are in good agreement with the observed spectroscopy .
the eigensolutions of eq .
( [ eq : scheq ] ) provide the @xmath19 light - front wavefunctions which control the dynamics of the mesons .
after factoring out the longitudinal and orbital dependence , the lfwf can be written as @xmath20 where @xmath21 is the poincare invariant radial variable of lf hamiltonian , and @xmath22 is the invariant transverse impact variable .
the hadronic lfwf @xmath14 in the soft - wall holographic model encodes the dynamical properties of the mesons .
if one also includes the light quark masses , it is given by @xcite @xmath23 in impact space .
note that the lf kinetic energy @xmath24 is also the invariant mass squared @xmath25 of the hadronic constituents .
the holographic lfwf given by eq .
( [ eq : lfwfb ] ) has been successfully used to describe diffractive @xmath26 meson electroproduction at hera @xcite as well as the spectroscopy and distribution amplitudes of light and heavy mesons @xcite . after introducing the lf spinor structure of the wavefunctions for light vector mesons in analogy with that of the photon ,
the authors of refs @xcite have predicted the light - front distribution amplitudes ( lfdas ) of the @xmath26 and @xmath27 vector mesons and have used them to evaluate the branching fractions of @xmath28 and @xmath29 decays .
in addition , the @xmath30 form factors are computed and applied to rare @xmath31 and @xmath32 decays @xcite .
the helicity dependence of the lfwfs for the vector mesons is introduced in these analyses @xcite in order to predict specific helicity dependent observables .
the helicity dependence of the holographic lfwf is assumed to decouple from the dynamics , which in turn leads to simple factorizable formulae for physical quantities , such as the decay constants . in the past few years
, several qcd - inspired approaches , such as qcd factorization ( qcdf ) @xcite , perturbative qcd ( pqcd ) @xcite and soft - collinear effective theory ( scet ) @xcite , have been developed in order to evaluate the hadronic matrix elements of local operators which control two - body nonleptonic @xmath33 decays .
however , the convolution integrals of the hard kernels with the asymptotic form of distribution amplitudes of light final states suffer from an end - point divergence , such as @xmath34 and @xmath35 .
this divergence limits the prediction power of the theoretical approaches and prevents reliable results . at present
, the dynamical origin of the end - point divergence is still unclear , even though there are conjectures about the reliability of the collinear approximation and concerns about our limited understanding of the qcd dynamics of hadrons . in this paper
, we will explore helicity - improved lfwfs for light pseudoscalar mesons , and then test their predictions for hadronic observables including the decay constants of @xmath0 and @xmath1 mesons , their @xmath2-moments and the pion - to - photon transition form factor .
we will also explore new applications to two - body nonleptonic @xmath33 decays , focusing especially on the measured pure annihilation @xmath36 and @xmath37 decay channels .
we will also show that the problem of end - point divergences in the annihilation amplitudes is mitigated by the improved behavior of the lfdas near their end - points .
several schemes for regulating the end - point divergences have been previously proposed . in the scet approach , a zero - bin subtraction @xcite is assumed .
the annihilation diagrams are found to be factorizable and bring no any strong phase in the leading terms of order @xmath38 @xcite . in the qcdf approach ,
the end - point divergent integrals are treated as signals of infrared - sensitive contributions which can be regularized by introducing a complex quantity @xmath39 @xcite .
alternatively , one can introduce an infrared - finite dynamical gluon propagator which moves the end - point singularity into an integral over the time - like gluon momentum ; the divergence then vanishes , and a large strong phase is predicted @xcite . in the pqcd approach , the end - point singularity is avoided by introducing parton transverse momentum @xmath40 , but at the expense of having to model the additional @xmath40 dependence of the meson distributions ; this again predicts a large complex annihilation correction @xcite .
in contrast , in the framework of the lf holographic qcd , the end - point contribution is naturally suppressed by the exponential factor in lfwf due to non - vanishing effective quark masses , @xmath41 and @xmath42 . in this paper
we will test if the effective quark mass regulation of the end - point divergences in the annihilation amplitudes obtained from lf holographic qcd can provide viable predictions for pure annihilation heavy hadron decays .
our paper is organized as follows . in section 2 ,
the connections between holographic lfwfs and lfdas for light pseudoscalar mesons are explored within the framework of lf quantization , based on two different assumptions for the helicity - dependence of the hadronic lfwfs .
sections 3 and 4 are devoted to numerical results and discussions in which the decay constants , the @xmath2-moments and the pion - to - photon transition form factor are evaluated using the helicity - improved lfwfs and lfdas . in section 5 ,
the annihilation amplitudes and the pure annihilation @xmath3 and @xmath4 decays are studied in detail using the lfdas predicted by lf holographic qcd .
finally , we give our summary in section 6 .
our starting point is the definition of the distribution amplitudes ( das ) of light pseudoscalar meson @xcite .
the das parameterize the operator product expansion of meson - to - vacuum matrix elements @xcite , @xmath43 where @xmath44 , @xmath45 is the decay constant of a pseudoscalar meson ( @xmath46 ) , @xmath47 and @xmath48 are twist-2 and twist-3 das , respectively . in the following derivation
, we will adopt lepage - brodsky ( lb ) convention @xcite and assume light - front gauge , @xmath49 . at equal lf time
, the das can be expressed using eqs .
( [ eq : twi2 ] ) and ( [ eq : twi3a ] ) as @xmath50 by performing the fourier transformation with respect to @xmath51 .
the main remaining task is to deal with the hadronic matrix elements in eqs .
( [ eq : twi2p2 ] ) and ( [ eq : twi3ap2 ] ) . in the framework of lf quantization @xcite
, a hadronic eigenstate @xmath52 can be expanded on a complete fock - state basis of noninteracting 2-particle states as @xmath53 in which , @xmath54 is the lfwf of the pseudoscalar meson with helicity - dependence included ; @xmath55 and @xmath56 are the helicities of quark and anti - quark , respectively ; and the one - particle state is defined , for instance , by @xmath57 .
the dirac ( quark ) field is expanded in terms of particle creation and annihilation operators as @xmath58 \ ,
, \label{eq : qfexp}\ ] ] assuming lf helicity spinors @xmath59 and @xmath60 the equal lf - time anti - commutation relations are @xmath61 equipped with the above formulae , the hadronic matrix element in eqs .
( [ eq : twi2p2 ] ) and ( [ eq : twi3ap2 ] ) can be expressed as @xmath62 in which @xmath63 and @xmath64 , and the scale @xmath65 is introduced as an ultraviolet cut - off on transverse momenta .
using eq .
( [ eq : meder1 ] ) and integrating over @xmath66 and @xmath67 , we can further obtain a general expression for the rhs of eqs .
( [ eq : twi2p2 ] ) and ( [ eq : twi3ap2 ] ) , @xmath68 to proceed with the derivation , we will need the explicit form of the holographic lfwf , @xmath54 .
as mentioned in the introduction , the helicity - dependence of the holographic lfwf has been assumed in previous works to decouple from the dynamics , and therefore @xmath69 , where @xmath70 is given by the fourier transformation of eq .
( [ eq : lfwfb ] ) .
this assumption leads to a universal formula for predicting physical quantities for different kinds of mesons ; however , it is obviously disfavored by experiment . in order to restore the proper helicity dependence
, the holographic lfwf in the @xmath71 space needs to be modified as @xmath72 where @xmath73 is the helicity - dependent wavefunction , @xmath74 is the normalization factor determined by the normalization condition @xmath75 and @xmath70 is the radial wavefunction obtained by performing the fourier transformation of eq .
( [ eq : lfwfb ] ) , @xmath76 in the case of a vector meson , one can work in analogy with the lowest - order helicity structure of the photon lfwf in qed ; the following structure of @xmath77 is thus assumed @xcite @xmath78 this form has been successfully used to study the production of the @xmath26 and @xmath27 vector mesons and the decays related to @xmath79 transitions @xcite . in the case of a pseudoscalar meson , following such a strategy , @xmath80 in eq .
( [ eq : lfwfsv2 ] ) would be replaced simply by @xmath64 @xcite .
very recently , this spin structure has been used to evaluate the holographic da of @xmath0 meson in ref .
the helicity - dependent wavefunction is written explicitly as @xmath81 where the factor `` @xmath82 '' is now added to be consistent with the convention for the definition in eqs .
( [ eq : twi2 ] ) and ( [ eq : twi3a ] ) , and the abbreviation @xmath83 is used for convenience .
an additional multiplying factor `` @xmath84 '' , which is added in ref .
@xcite , has been absorbed into the normalization constant .
it should be noted , however , that this spin structure corresponds to light quark and antiquark of the pseudoscalar meson , such as the pion , to have parallel spin projections , and thus @xmath85 .
this state has twist@xmath86 , and it is thus not the meson eigenstate of the ads / qcd theory .
instead of @xmath64 , the dirac structure like @xmath87 is also allowed .
we therefore consider an alternative form of @xmath88 : @xmath89 in which , the structure @xmath90 implies that the light quark and antiquark have only opposite helicities .
this is the helicity assignment that couples the pion to the axial - vector current and thus pion decay constant @xmath91 in : @xmath92 it is thus the leading - twist lfwf , and is the solution from ads / qcd for light quarks . since @xmath93 is the invariant mass of @xmath94 pair in the @xmath46 meson , the dimensions of the two terms in @xmath88 , eq . ( [ eq : lfwfsp3 ] ) , are consistent . in the following , for convenience of discussion , the two helicity - dependent wavefunctions defined by eqs . and ( [ eq : lfwfsp3 ] )
will be referred to as scenario 1 ( s1 ) and scenario 2 ( s2 ) , respectively .
they are related by the gell - mann - oakes - renner ( gmor ) relation and are thus not independent @xcite . using lb convention @xcite , the two helicity - dependent wavefunctions @xmath88 are given explicitly as @xmath95\,,\qquad \qquad~{\rm scenario~1 } \\ \frac{i}{\sqrt{z\bar{z}}}\left [ -|\mathbf{k}_{\bot}|e^{\mp i \theta_k}\delta_{h\pm,\bar{h}\pm}\pm\left(zm_{\bar{q}}+\bar{z}m_{q}+z\bar{z}\widetilde{m}_p\right ) \delta_{h\pm,\bar{h}\mp}\right]\,,\quad { \rm scenario~2 } \end{aligned}\right.\end{aligned}\ ] ] and the spinor currents in eq .
can be written as @xmath96\,,\end{aligned}\ ] ] in which , @xmath97 is specified . finally , in the @xmath71 space ,
using the building blocks given above , the holographic das of @xmath46 meson can be written as @xmath98 & = & \frac{\sqrt{n_c } } { \pi f_p } \int^{|\mathbf{k}| < \mu } \frac{\d^2\mathbf{k}_{\bot}}{(2\pi)^2}\ , \frac{n_1}{(z\bar{z})^{1/2}}(\bar{z}m_q+zm_{\bar q})\,\psi(z,\mathbf{k}_{\bot})\,,\\ \label{eq : lfdak3p } \phi(z , \u)[{\rm s1 } ] & = & \frac { \sqrt{n_c } } { 2\pi \u_{p}f_p } \int^{|\mathbf{k}| < \mu } \frac{\d^2\mathbf{k}_{\bot}}{(2\pi)^2}\frac{n_1}{(z\bar{z})^{3/2}}\big\{\mathbf{k}_{\bot}^2 + ( zm_{\bar{q}}+\bar{z}m_q)^2 \big\ } \,\psi(z,\mathbf{k}_{\bot})\,,\end{aligned}\ ] ] in the case of s1 meson is also evaluated with a @xmath88 similar to s1 .
our result is more general in comparison with that of ref .
@xcite . ] , and @xmath99 & = \frac{\sqrt{n_c } } { \pi f_p } \int^{|\mathbf{k}| < \mu } \frac{\d^2\mathbf{k}_{\bot}}{(2\pi)^2}\ , \frac{n_2}{(z\bar{z})^{1/2}}(\bar{z}m_q+zm_{\bar q}+z\bar{z}\widetilde{m}_p)\,\psi(z,\mathbf{k}_{\bot})\,,\\ \phi(z , \u)[{\rm s2 } ] & = \frac { \sqrt{n_c } } { 2\pi \u_{p}f_p } \int^{|\mathbf{k}| < \mu } \frac{\d^2\mathbf{k}_{\bot}}{(2\pi)^2}\frac{n_2}{(z\bar{z})^{3/2}}\big\{\mathbf{k}_{\bot}^2 + ( zm_{\bar{q}}+\bar{z}m_q)(zm_{\bar{q}}+\bar{z}m_q+z\bar{z}\widetilde{m}_p ) \big\ } \,\psi(z,\mathbf{k}_{\bot})\ , , \label{eq : lfdak3p2}\end{aligned}\ ] ] in the case of s2 , where @xmath100 and @xmath101 are the corresponding normalization factors determined by eq .
( [ eq : norc ] ) . the expression in the impact space can be obtained through fourier transformation .
these formulae , which exhibit the connections between holographic lfdas and lfwfs , are one of the main theoretical results in this paper . using the theoretical framework given above , we will present numerical results and applications of these holographic lfdas and lfwfs in the following sections .
before presenting our numerical results , we now clarify the values of input parameters used in our evaluation .
one of the most important inputs is the mass scale parameter @xmath102 is used . ] , which could be extracted from many observables .
for example , to fit the light - quark mass spectrum , the values @xmath103 and @xmath104 are suggested in ref .
@xcite for light pseudoscalar and vector mesons , respectively .
a mean value , @xmath105 , is obtained in ref .
@xcite by fitting all of the slopes of the different regge trajectories for mesons and baryons including all excitations .
this result is also favored by the recent high accuracy computation of the perturbative qcd scale parameter @xmath106 @xcite .
the fit to the bjorken sum - rule data at low @xmath107 yields @xmath108 @xcite . in ref .
@xcite , the value @xmath109 is used for determining the freezing value of @xmath110 and the interface between perturbative and nonperturbative qcd .
in addition , in order to describe the hera data on diffractive @xmath26 and @xmath111 electroproduction , the values @xmath112 and @xmath113 are suggested @xcite . besides @xmath102 ,
the light - quark masses appearing in the holographic lfwfs are the other important inputs , which will be specified below . in this paper , for s1
, we follow entirely the inputs suggested by the recent study of holographic da of @xmath0 meson with a similar lfwf of s1 @xcite .
explicitly , the following input values are used @xcite : @xmath114 where the constituent quark masses are adopted , and are also used for studying the @xmath26 and @xmath27 mesons @xcite .
it should be noted that , as pointed out in ref .
@xcite , the light - quark masses introduced in the holographic lfwf are not the traditional constituent masses in the non - relativistic theories , but are the effective quark masses from the reduction of higher fock states as functionals of the valence states
. such effective quark masses , in principle , should be universal in a specific theoretical framework of holographic qcd . for s2 , on the other hand ,
we take @xmath115 which will be further explained in detail in the next subsection .
it should be noted that such input values are very similar to the results @xcite , @xmath116 obtained by fitting the regge trajectories of pseudoscalar mesons in the framework of lf holographic qcd @xcite .
= .[tab : dc ] numerical results of decay constants of @xmath117 and @xmath118 mesons in unit of @xmath119 . [ cols="<,^,^,^,^,^,^,^,^,^ " , ] in the numerical evaluations , we will use the values of ckm parameters fitted by the ckmfitter group @xcite , @xmath120 the averaged values of the @xmath33-meson decay constants @xcite , @xmath121 and the central values of the other inputs , such as the well - determined masses and lifetimes of @xmath33 mesons , and the fermi constant _ etc .
, _ given by pdg @xcite . using these inputs ,
our numerical results for the cp - averaged branching ratios of @xmath122 and @xmath37 decays are listed in table [ pikbr ] , in which the experimental data and the previous theoretical results based on the qcdf with parameterization scheme @xcite and the pqcd approach @xcite are also given for comparison .
our results are evaluated at the renormalization scale @xmath123 with an assigned uncertainty @xmath124 . for the case of s2 , the theoretical errors caused by the ckm parameters and @xmath33-meson decay constants , the holographic inputs given by eq . , and the renormalization scale @xmath65 are obtained by evaluating separately the uncertainties induced by each input parameter and then adding them in quadrature . from table
[ pikbr ] , we find that the results in s1 are similar to the central values obtained in qcdf @xcite with traditional parameterization scheme , but are about one order of magnitude smaller than the data , which is mainly due to the fact that the holographic das in s1 are relatively narrow as shown in fig .
[ fig : das ] , and the contributions with @xmath125 and @xmath126 are strongly suppressed . in contrast
, our prediction for @xmath127 in s2 is in good agreement with the data ; within the experimental and theoretical uncertainties our prediction for @xmath128 in s2 also agrees with the data .
this implies that , compared to s1 , s2 is much more favored by the data of @xmath127 and @xmath128 . in the following discussions , we will focus only on the results using s2 .
comparing with the previous evaluations in qcdf with parameterization scheme for the end - point divergence , we find that the theoretical predictions are remarkably improved by using the holographic das .
moreover , the predicting power is retained . comparing our predictions with the ones in pqcd
, we find good agreement for @xmath127 ; however , our result for @xmath128 is smaller than that obtained in pqcd .
the significant difference between @xmath127 and @xmath128 in our evaluation is easily understandable due to the following facts : 1 . for the @xmath122 decay , because @xmath129 , its decay amplitude , eq .
, can be simplified as @xmath130 in which @xmath131 because the @xmath132- and @xmath133-quark difference is not distinguished in this paper . for the @xmath37 decay , on the other hand , its amplitude can be simplified as @xmath134\ ,
. \end{aligned}\ ] ] comparing with eq .
, one can easily find that the first and second terms in eq .
are relatively suppressed by additional cabibbo factors @xmath135 and @xmath136 , respectively .
thus , a large ratio @xmath137 is generally expected .
moreover , for the @xmath138 meson , as shown by fig .
[ fig : das ] , the holographic das near the end - point where the ( anti-)strange quark carries small momentum fraction is suppressed due to @xmath139 . as a result ,
both twist-2 and twist-3 contributions are relatively suppressed for the @xmath37 decay compared to the @xmath122 decay .
in addition , since @xmath140 and the phase space of @xmath141 decay is larger than that of @xmath142 decay , the ratio @xmath143 is further enhanced .
it should be noted that our evaluations are performed at leading order and the theoretical uncertainties , especially the one induced by the renormalization scale , are still quite large .
moreover , the refined measurements , especially for the @xmath142 decay , are required for a definite conclusion . from the phenomenological point of view , an annihilation amplitude with a large strong phase is generally welcome in order to fit experimental data and to explain some puzzles observed in @xmath33-meson decays @xcite . as a result
, a complex parameter @xmath39 has been introduced in the traditional parameterization scheme within the framework of qcdf @xcite . by using the dynamical gluon mass @xmath144 in qcdf approach @xcite or by introducing transverse momentum @xmath40 degree in pqcd approach @xcite , a large imaginary part in the annihilation amplitudes
is also obtained because the singularities exist in the integral over momentum fractions .
in contrast to the above regulation schemes , the leading - order annihilation contributions are real by using the holographic das . this result is understandable due to the fact that , although the leading - order annihilation corrections are evaluated at the order @xmath145 , they are in fact tree " contributions and there is no independent internal momentum ; while , the strong phases are generally induced by the loop integration , such as in the vertex and penguin diagrams . in the scet approach , real annihilation contributions for the leading terms of order @xmath38
have also been predicted @xcite .
in addition , it should be noted that complex annihilation contributions are of course possible if , for instance , final - state interactions or higher - order corrections are taken into account .
motivated by the rapid development of the lf holographic qcd , the lfwfs for light pseudoscalar mesons and their applications are studied in this paper . in order to restore the dynamical spin effects of quarks and to improve the predictability of lfwfs for different pseudoscalar mesons ,
the traditional lfwfs are modified according to two assumptions for the helicity - dependent wavefunctions , corresponding to the structures @xmath146 ( named as s1 ) and @xmath147 ( named as s2 ) , respectively .
the lf holographic das of pseudoscalar mesons are then extracted using the helicity - improved lfwfs .
the decay constants , the @xmath2-moments , the pion - to - photon transition form factor and the @xmath3 and @xmath4 decays are then evaluated and compared with experiment .
our main findings are summarized as follows : * in contrast to the lfwf for s1 , we find that the lfwf for s2 can provide sufficient flavor - asymmetry resources for predicting the decay constants of @xmath0 and @xmath1 mesons .
moreover , the results based on s2 for all of the observables considered in this paper are in a much better agreement with experiment than the ones based on s1 . * taking the decay constants of @xmath0 and @xmath1 mesons as constraints , we perform a @xmath8-fit for the holographic parameters , @xmath9 and effective quark masses @xmath148 and @xmath149 .
interestingly , our fitted results are remarkably consistent with the ones obtained by fitting the regge trajectory of light - quark pseudoscalar mesons . *
a new scheme with lf holographic das for regulating the end - point divergence in the annihilation amplitudes of @xmath10 decays is presented . in this scheme ,
the leading - order annihilation contributions are real .
numerically , our predictions for the branching fractions @xmath127 and @xmath128 using the lf holographic das in s2 agree well with current data and result in a relatively large flavor - symmetry breaking effect .
these predictions will be further tested by future refined measurements .
we thank junfeng sun at hnnu and xingbo yuan at kiast for helpful discussions .
this work is supported by the national natural science foundation of china ( grant nos .
11475055 , 11675061 and 11435003 ) .
q. chang is also supported by the foundation for the author of national excellent doctoral dissertation of p. r. china ( grant no .
201317 ) , the program for science and technology innovation talents in universities of henan province ( grant no .
14hastit036 ) , the excellent youth foundation of hnnu and the csc ( grant no .
201508410213 ) .
x. li is also supported in part by the srf for rocs , sem , and by the self - determined research funds of ccnu from the colleges basic research and operation of moe ( ccnu15a02037 ) .
is supported by the department of energy , contract de - ac02 - 76sf00515 .
slac - pub-16849 .
g. f. de teramond and s. j. brodsky , aip conf .
* 1296 * ( 2010 ) 128
. v. de alfaro , s. fubini and g. furlan , nuovo cim .
a * 34 * , 569 ( 1976 ) .
doi:10.1007/bf02785666 g. f. de teramond , h. g. dosch and s. j. brodsky , phys .
d * 91 * ( 2015 ) no.4 , 045040 .
m. beneke , g. buchalla , m. neubert and c. t. sachrajda , phys .
* 83 * ( 1999 ) 1914 .
m. beneke , g. buchalla , m. neubert and c. t. sachrajda , nucl .
b * 591 * ( 2000 ) 313 .
m. beneke , g. buchalla , m. neubert and c. t. sachrajda , nucl .
b * 606 * ( 2001 ) 245 . | in this paper the dynamical spin effects of the light - front holographic wavefunctions for light pseudoscalar mesons are studied using two different assumptions .
these improved wavefunctions are then confronted with a number of sensitive hadronic observables : the decay constants of @xmath0 and @xmath1 mesons , their @xmath2-moments , the pion - to - photon transition form factor , and the pure annihilation @xmath3 and @xmath4 decays .
taking @xmath5 , @xmath6 and their ratio @xmath7 as constraint conditions , the @xmath8 analyses for holographic parameters , including the mass scale parameter @xmath9 and effective quark masses , are all consistent with the mass scale which controls the slopes of the light - quark hadronic regge trajectories .
in addition , we also show how the improved light - front holographic distribution amplitudes regulate the end - point divergences which appear in the annihilation amplitudes of @xmath10 decays . *
pacs numbers : * 12.38.aw , 11.25.tq , 14.40.nd , 13.25.hw |
there is a strong consensus that spiral galaxies are embedded in massive , non - luminous halos , the structure and scaling properties of which have been intensively investigated ( e.g. , @xcite ) .
studies of galaxy kinematics and of the density and temperature distributions of hot intergalactic gas in galaxy groups ( @xcite ) and clusters ( e.g. , @xcite ) have demonstrated the dominance of dark matter on larger scales . in the context of existing theories of the formation of galaxies and large scale structure , it is therefore natural to expect non - luminous material within elliptical galaxies as well .
however , until very recently , the evidence for dark matter in ellipticals has been less forthcoming and more controversial
. dark matter in ellipticals has been elusive largely due to the lack of internal dynamical indicators at optical and radio wavelengths that are as ubiquitous , spatially extensive , and unambiguous as hi rotation curves in spirals .
although mergers have endowed some early - type galaxies ( many classified as s0 ) with neutral ( e.g. , @xcite ) or ionized ( @xcite ) gas disks , these are rare and may not be representative of the population of ellipticals as a whole .
studies of stellar kinematics have , until recently , been inadequate in quality and extent to significantly constrain the large scale mass distribution in ellipticals in light of the complicating effects of projection and anisotropic velocity distributions .
most , if not all , giant elliptical galaxies contain extended distributions of hot x - ray emitting gas that provide a means for probing the large scale dark matter distribution a means potentially as powerful as rotation curves in spiral galaxies . since these hot gaseous halos should be close to hydrostatic equilibrium ( @xcite ) , measurements of the density and temperature distributions from x - ray imaging and spectroscopy yield constraints on the mass distribution on the scale and with the accuracy and spatial resolution of x - ray temperature profiles .
observations of elliptical galaxies by the _ einstein observatory _ ipc were presented as strong evidence for the presence of extensive dark halos in early - type galaxies ( @xcite ; @xcite ; @xcite ) ; however , the temperature profiles were of sufficiently poor quality to foster a continuing skepticism ( @xcite , @xcite ) .
progress on a number of fronts has been rapid during recent years . in the optical , improvements in detector capabilities , data analysis , and modeling
have now provided robust stellar dynamical constraints on the mass distribution out beyond the optical half - light radius @xmath5 for several ellipticals ( @xcite , @xcite , @xcite , @xcite ) , while gravitational lensing observations are providing the first evidence for dark matter in intermediate redshift galaxies ( @xcite ; @xcite ; @xcite ) .
x - ray observations using the _ rosat _ and _ asca _ satellites have improved the accuracy , spatial resolution , and extent of derived hot gas density and temperature profiles . of particular interest
is the recent study of _ asca _ observations of @xmath15 early - type galaxies by matsushita ( 1997 ) that shows that gas temperature profiles are relatively flat out as far as can be reliably measured ( 5@xmath16 , depending on the intrinsic luminosity and galaxy distance ) . the existence of isothermal 0.51 kev gaseous halos extending , in some cases , to well beyond 10@xmath5 is strong evidence for the presence of extended dark matter in elliptical galaxies .
the case for dark matter halos in a number of elliptical galaxies is now overwhelming , and the following pattern seems to be emerging from the latest results on individual galaxies : dark matter comprises a significant but not dominant fraction of the total mass within @xmath5 ( @xcite ; @xcite ; @xcite ; @xcite ; but , see @xcite for an exceptional case : ngc 4636 ) ; dark matter becomes increasingly dominant at large radii : @xmath17m@xmath18 kpc@xmath19 to within a factor of two for @xmath20 kpc ( @xcite , @xcite , @xcite ) , corresponding to mass - to - light ratios @xmath21m@xmath13l@xmath22 measured at ( or extrapolated to ) @xmath23 kpc ( @xcite ; @xcite ) .
while there is now little doubt that some elliptical galaxies do contain dark matter , a number of broader issues remain to be addressed . among the most important with respect to our understanding of galaxy formation is whether dark matter in ellipticals is ubiquitous , and how the relative scaling of dark and luminous matter varies both within galaxies as a function of radius , and between galaxies with different optical properties . in this paper
we use the recently derived collection of x - ray temperatures in a complete , optically selected sample from davis & white ( 1996 ; hereafter dw ) to constrain the _ average _ properties of dark halos in the _ population _ of @xmath2 elliptical galaxies .
dw analyzed x - ray spectra from 42 of the 43 optically brightest elliptical galaxies , using _
pspc data if available and _ einstein _ ipc data otherwise ( _ i.e. _ for four galaxies ) .
see dw and white & davis ( 1998 ) for details about the sample and data analysis . for the purposes of this work , we use only the mean relationship between x - ray temperatures , @xmath24 , and central projected optical velocity dispersions , @xmath25 ( hereafter , the `` @xmath0@xmath1 relation '' ) @xmath26 or equivalent x - ray / optical correlations .
more precisely , we consider the variation of the observable @xmath9 with optical luminosity @xmath27 ( figure 1 ) , where @xmath28 and @xmath29 is the mean mass per particle . if dark matter dominates the gravitational potential on large scales , then @xmath24 is essentially a measure of the dark matter content within the extraction radii used by dw ( @xmath8 , where @xmath5 is the optical effective radius ) .
since the observed scaling relations ( _ i.e. _ the fundamental plane ) for elliptical galaxies provide a link between @xmath25 and the global luminosity , @xmath9 is an excellent diagnostic of the dark - to - luminous matter ratio within the optical radius .
temperature profiles can provide additional dark matter constraints , but are measurable for only the x - ray brightest galaxies in the sample .
likewise , extended stellar velocity dispersion profiles are useful , but data of sufficient detail to be unambiguously interpreted are not generally available . on the other hand ,
central velocity dispersions have been accurately measured for the entire sample ; central velocity dispersions are known to correlate with other optical properties and are less likely to suffer from the uncertainties caused by anisotropic stellar orbits .
our goal in this paper is to make general inferences about the presence of dark matter halos and their mean systematic variation with optical luminosity . toward that end ,
we utilize a set of physically and observationally well - motivated mass models and scaling relations in order to reproduce the @xmath0@xmath1 relation .
although we address the scatter in the @xmath0@xmath1 relation , a detailed examination of individual galaxies is beyond the scope of this paper and is deferred to future investigation .
our results can be considered in three stages .
first , we demonstrate how the observed range of @xmath9 _ necessarily _ requires the existence of dark halos in elliptical galaxies .
second , we derive conservative lower limits on the dark matter scale length and dark - to - luminous mass fraction , and discuss how the dark matter content must scale with optical luminosity so as to reproduce the observed @xmath0@xmath1 relation .
finally , we discuss possible implications of our results for cosmology and galaxy formation , by embedding our models within particular scenarios for large scale structure formation .
as stated above , we use the observable @xmath9 , defined in equation ( 1 ) , to constrain the dark matter distribution in elliptical galaxies .
this parameter is the ratio of suitable averages ( to be defined below ) of the square of the stellar velocity dispersion and the gas temperature .
therefore , we must derive the expected distributions of stellar velocity dispersions @xmath1 and gas temperatures @xmath0 , given specified total gravitational mass distributions .
recently published _ hubble space telescope _
( _ hst _ ) observations of the centers of elliptical galaxies show that their surface brightness profiles can generally be characterized by the function @xmath30^{(c - b)/a},\ ] ] where @xmath31 is the projected radius and @xmath32 is a break radius such that @xmath33 for @xmath34 and @xmath35 for @xmath36 , while @xmath37 characterizes the sharpness of the break ( @xcite ) . in order to accurately calculate the central stellar velocity dispersion ,
this observed departure from distributions with `` analytic cores '' must be accounted for . in the appendix
we demonstrate that , for @xmath38 and @xmath39 , equation ( 2 ) provides a good approximation to the projection of a space density profile @xmath40 ( @xmath41 ) for @xmath42 and @xmath43 for @xmath44 . to derive the luminosity density distribution @xmath45 , we numerically deproject the surface brightness profile : @xmath46 where @xmath47 , and the change of variables @xmath48 has been made .
the dw sample is magnitude - limited , and we are primarily interested in the most luminous ellipticals , since they have the most accurate x - ray temperatures ( see 3.3 ) .
we adopt the typical values for @xmath49 galaxies ( adopting @xmath50 km s@xmath19 mpc@xmath19 ) of @xmath38 , @xmath51 , and @xmath52 ( @xcite ) .
for the nine galaxies in common with dw ( excluding the highly flattened ngc 4697 ) , the mean surface brightness parameters and variances from faber et al .
( 1997 ) are as follows : @xmath53 , @xmath54 , @xmath55 .
the resulting stellar density profile diverges slightly faster than @xmath56 for @xmath42 .
the asymptotic surface brightness slope @xmath57 is flatter than in conventional surface brightness parameterizations and the integrated luminosity @xmath58 does not converge .
therefore , we truncate the optically luminous distribution at radius @xmath59 , determined so that @xmath60 consequently @xmath61 . as an alternative stellar model
, we consider the density profile from hernquist ( 1990 ) , @xmath62 where @xmath63 is the total luminosity and @xmath64 ( as adopted here ) provides the best match to a de vaucouleurs ( dev ) model with effective radius @xmath5 ( @xcite ) .
an advantage of this model is that the density and mass profiles are analytic .
we shall subsequently refer to the stellar distribution of equation ( 3 ) as the `` hst '' model , and that of equation ( 5 ) as the `` hernquist '' model .
a comparison of the luminosity density ( in units of the average density , @xmath65 ) as a function of normalized radius ( @xmath66 ) in these models with the corresponding dev model is shown in figure 2 , where @xmath67 .
all three models produce similar slopes at very small @xmath68 ; however , dev and hernquist models fail to reproduce the flattening at @xmath69 observed by _
hst_. on the other hand , while @xmath70% of the total luminosity is contained within @xmath71 in all three models , the hst model concentrates the entire remaining @xmath72% of the mass between @xmath73 and @xmath8 as opposed to spreading it out from @xmath73 to @xmath74 .
in order to accurately characterize the observed central stellar properties as required by the definition of @xmath9 without introducing additional parameters ( _ i.e. _ an additional break radius and a distinct asymptotic form for the density profile ) , such a redistribution of the stellar light is necessary .
if the light profile does indeed steepen outside of the _ hst _ field - of - view , then the only effect on our results for the hst model is that , by overestimating the stellar mass at large radii , we would underestimate the dark halo mass required to reproduce any observed value of @xmath9 .
the stellar velocity dispersion is derived by solving the jeans equation in the form @xmath75 where @xmath76 is the stellar density distribution obtained from multiplying the luminosity density ( see equations 3 and 5 , above ) by a constant stellar mass - to - light ratio ; @xmath77 is the total mass within radius @xmath68 ( the dark matter component of which is described in the next section ) ; @xmath78 is the velocity dispersion anisotropy parameter ( @xmath79 ) ; and @xmath80 ( @xmath81 ) is the radial ( tangential ) component of the velocity dispersion . we make the reasonable assumption that the velocity dispersion becomes isotropic at small @xmath68 but radial at large @xmath68 ( @xcite ) , and adopt the functional form @xmath82 where the transitional scale radius @xmath83 is a free parameter of our models . the dark matter mass distribution is assumed to follow the `` universal '' density profile ( @xcite ) , @xmath84 that , when integrated , yields @xmath85.\ ] ] it is not clear how appropriate equations ( 8) and ( 9 ) are for elliptical galaxy dark halos . on one hand
, there is evidence for cores in spiral galaxy dark matter distributions ( e.g. , @xcite ) ; on the other , dissipational collapse of the baryonic component in ellipticals can increase the dark matter concentration .
moreover , models with cores can not generally be consistently coupled with stellar mass distributions as cuspy as those often observed in elliptical galaxies ( @xcite ) .
we have verified the existence of self - consistent two - component mass models with dark matter following equation ( 8) and hernquist stellar models ( equation 7 ) , in the parameter range of interest , according to the criterion of ciotti and pelligrini ( 1992 ) . from the virial theorem it follows that the integrated mass can be robustly estimated from the mean temperature .
constraints on the dark matter concentration can also be placed from consideration of the @xmath0@xmath1 relation ; however , any inferences about dark matter at small or large radii are contingent on the corresponding assumed asymptotic dark matter density slopes .
our choice of equation ( 8) for the dark matter distribution does allow us to connect our results with predictions of dark halo structure from numerical structure formation simulations .
however , the true shape of the dark matter spatial distribution can only be constrained from temperature and stellar velocity dispersion _
profiles _ of sufficient extent and resolution .
the hot , x - ray emitting gas is assumed to be in hydrostatic equilibrium so that @xmath86 the mass in hot gas is not significant compared to the stellar mass , so its contribution to the gravitational potential is neglected . since we are interested in the average properties of the dw sample , and because spatial analysis is not available for most of the sample , we assume that @xmath87 ; this has been found to be a good approximation for x - ray bright galaxies , and is expected if heating is balanced by cooling locally ( @xcite ) .
both assumptions , hydrostatic equilibrium and @xmath87 , are more likely to break down for galaxies with low x - ray - to - optical luminosity ratios , where a partial or transient outflow might obtain ; however , the @xmath0@xmath1 relation for these relatively x - ray faint systems does not differ , within the statistics of the sample , from that of the sample as a whole .
a significant unifying simplification for the hst stellar models follows from the observations that the ratios of core mass to total stellar mass ( @xmath88 , where @xmath89 is @xmath90 [ equation 4 ] multiplied by the stellar mass - to - light ratio ) , and break radius @xmath32 to @xmath5 ( the `` true '' @xmath5 as observed , not the half - light radius in our models ) are approximately constant : @xmath91 and @xmath92 ( @xcite ) .
our adoption of these scaling relations , along with the empirical relation @xmath93 , insures that our model galaxies are consistent with the fundamental plane ( fp ) as long as the core is baryon - dominated , since it follows that @xmath94 where @xmath95 is the central velocity dispersion .. ] in order to provide us with a one - parameter family of models ( the quality and quantity of the x - ray data is insufficient to attempt to derive a `` @xmath0@xmath1@xmath5 '' or equivalent fp - like relation ) , we take a `` faber - jackson cut '' through the fp such that @xmath96 , normalized so that @xmath97 kpc ( and @xmath98 km s@xmath19 ) at @xmath99 for @xmath50 km s@xmath19 mpc@xmath19 .
the root - mean - square deviation of galaxies in the dw sample from this projection of the fp is only 34 km s@xmath19 ( 13.5% of the mean ) .
the hernquist models ( equation 5 ) likewise adhere to the fp , since we universally adopt @xmath64 .
equations ( 6 ) and ( 10 ) are solved in dimensionless form .
there are two fundamental ( luminosity - dependent ) dimensional quantities that we denote as @xmath100 ( chosen to be @xmath32 for the hst model or @xmath101 for the hernquist model ) and @xmath102 ( @xmath89 for the hst model or total stellar mass @xmath103 for the hernquist model ) .
thus , adopting the above scaling relations , the length- , mass- , density- , velocity- , and temperature - scaling relations are as follows : @xmath104 @xmath105 @xmath106 @xmath107 and @xmath108 here @xmath109 is the hubble constant in units of 80 km s@xmath19 mpc@xmath19 , @xmath110 with @xmath111l@xmath14 , and @xmath112 is defined such that the stellar mass - to - light ratio @xmath113m@xmath13l@xmath14 for @xmath114 and @xmath115 .
@xmath116 corresponds to @xmath99 , _
i.e. _ @xmath117 .
the dark matter mass distribution is scaled accordingly , _
i.e. _ @xmath118 where @xmath119 is the ratio of dark to stellar mass at the hst stellar model cutoff radius , i.e. @xmath120 and @xmath121 is the dimensionless dark matter scale length . in the above , @xmath122 , @xmath123 , and the function @xmath124
is defined in equation ( 9 ) . to determine the average values of @xmath0 and @xmath1 , the hydrostatic and jeans equations
are integrated inward from a zero - pressure boundary at @xmath125 ( see @xcite ) .
although the stellar mass has a cutoff ( @xmath59 ) for the hst model , the gas and dark matter halos are likely to extend beyond this and are not truncated in the models .
this results in a pressure contribution to the gas virial temperature within @xmath59 that leads to lower values of @xmath9 for a given @xmath119 , and is therefore conservative in the sense of minimizing the required amount of dark matter .
the average quantities that constitute the observable @xmath9 are calculated as follows : we define @xmath25 as the average projected stellar velocity dispersion within a circular aperture of size @xmath32 , and @xmath24 as the emission - weighted projected gas temperature within @xmath59 ( the same aperture over which the temperature is measured in dw , corresponding to the radius enclosing @xmath126% of the optical light ) .
the value of @xmath9 is robust to the precise choice of these radii since @xmath127 .
the aperture - averaged , projected velocity dispersion for all central radii @xmath128 ( @xmath129 ) is within 10% of its value at @xmath32 for the hst ( hernquist ) model , assuming isotropic orbits and that the central potential is dominated by the stellar component . by averaging over constant multiples of the fundamental stellar scale length ( @xmath100 )
, we assure that our computed values of @xmath9 depend only on the _ relative _ ( to stellar ) dimensionless dark matter and velocity dispersion anisotropy parameters .
these parameters are the following : the ratio of dark to luminous matter scale length , @xmath121 ; the mass ratio @xmath130 within @xmath59 , @xmath119 ; and the dimensionless anisotropy scale length , @xmath131 .
the only additional model parameters are the stellar distribution slopes .
for a given triplet ( @xmath131 , @xmath132_{r_{\rm max}}$ ] , @xmath133 ) , a general solution for the dimensionless gas temperature and stellar velocity dispersion radial distributions follows that can be scaled to physical units for any @xmath109 and @xmath27 using equations ( 1216 ) ; @xmath112 is determined by normalizing to @xmath134 km s@xmath19 for @xmath99 and @xmath115 ; note that @xmath135 .
in this section we consider the dependences of @xmath9 on @xmath131 , @xmath133 , and @xmath132_{r_{\rm max}}$ ] for both hst and hernquist models .
while the latter provides a more accurate global representation of elliptical galaxy stellar distributions , the former is more accurate around and inside the break radius @xmath32 where @xmath25 is calculated . for models that are dark matter dominated beyond several @xmath5 ,
the redistribution of the stellar mass at large radii in the hst model is irrelevant . in the absence of dark matter ,
the only free parameter in our models is the anisotropy length scale @xmath83 ; recall that this is defined such that orbits become isotropic for @xmath136 and radial for @xmath137 . therefore ,
if stars dominate the gravitational potential within @xmath59 , and if the shape of the velocity ellipsoid does not vary systematically with @xmath27 , then @xmath9 should be independent of @xmath27 and @xmath25 ( _ i.e. _ , @xmath138 ) , contrary to the observed @xmath0@xmath1 relation ( dw ) .
moreover , as we now show , the typical observed value of @xmath139 can not be reproduced in the absence of dark matter .
the solid and dotted curves in figure 3 show the variation of @xmath9 with the dimensionless anisotropy scale length @xmath140 for , respectively , the hst and hernquist stellar models without dark matter .
evidently , @xmath9 falls with increasing @xmath83 , but is essentially constant for @xmath141@xmath142 ( @xmath143@xmath144 ) , since @xmath25 is an average over a much smaller radius .
the asymptotic values of @xmath9 are @xmath145 and @xmath146 for the the hst and hernquist stellar models , respectively .
the lower value for the hernquist model is the result of a temperature maximum at several @xmath32 that substantially contributes to the emission - averaged temperature because of the high stellar ( and , by assumption , gas ) density there . as dark matter
is introduced , the temperature profile becomes more nearly isothermal , and differences in @xmath9 between hst and hernquist models lessen ( see next paragraph , and figure 3 ) .
varying the slope parameters of the hst models ( equation 2 ) does not result in any values of @xmath9 less than 0.75 .
the values of @xmath9 derived without dark matter lie well above what is observed in a typical elliptical galaxy ( in fact , no galaxy in the dw sample has a best - fit value of @xmath9 greater than 1 ; see figure 1 ) .
one of our major conclusions is already apparent .
while a typical bright elliptical galaxy has @xmath139 , _ even under the most extreme assumptions , @xmath9 is never less than @xmath147 without the presence of an extended dark halo_. while 10 of the 30 galaxies in the dw sample have upper limits to @xmath9 greater than 0.75 , only 3 have upper limits greater than 1.2 ( see figure 1 ) . of the 16 galaxies where @xmath9 is determined to better than 20%
, only 1 has an upper limit greater than 0.75 .
the dash - dotted and dashed curves show the corresponding models containing dark matter halos with dark matter scale length @xmath148 set equal to @xmath5 ( _ i.e. _ @xmath149 ) , and with @xmath150_{r_{\rm max}}$ ] chosen so that @xmath139 for @xmath151 .
the non - luminous fraction of the total mass within @xmath5 is @xmath152 and @xmath153 for the hst and hernquist models , respectively . from here on
we assume isotropic stellar orbits .
as shown in figure 3 , for the practical purposes of calculating @xmath9 , this is equivalent to assuming that the orbits are mostly isotropic inside @xmath154 .
since @xmath9 increases with decreasing @xmath140 , more dark matter is required to obtain a given value of @xmath9 if this assumption is discarded once again we adopt the most conservative assumption as regards the required amount of dark matter .
figure 4a shows @xmath9 as a function of the ratio of dark matter scale length to stellar break radius , @xmath155 , for the hst ( solid curve ) and hernquist ( dotted curve ) models with @xmath150_{r_{\rm max}}$ ] .
( for the hernquist model , the scaling relation @xmath156 has been utilized . ) as @xmath148 decreases below @xmath157 , dark matter plays an increasingly important role in determining @xmath25 , so @xmath9 rises . for fixed @xmath132_{r_{\rm max}}$ ] , @xmath9 is nearly independent of @xmath155 provided that @xmath158 .
it is clear that measurements of @xmath9 can only place _ lower _ limits on @xmath148 .
there is a fair amount of freedom in how the dark matter can be distributed and still produce @xmath139 .
this is illustrated in figure 4b , which displays the variation in baryon fraction at @xmath5 ( solid curve for the hst stellar model , dotted curve for the hernquist model ) and at @xmath32 ( dashed curve for the hst stellar model , dot - dashed curve for the hernquist model ) .
@xmath9 changes very slowly with @xmath155 for @xmath159 , while the baryon fraction inside @xmath32 varies from @xmath160 and @xmath161 for the hst and hernquist stellar models , respectively , to @xmath162 for both models as @xmath148 becomes very large
. meanwhile , the baryon fraction inside @xmath163 increases from @xmath164 ( @xmath165 ) if @xmath166 , to @xmath147 ( @xmath167 ) for the hst ( hernquist ) model as @xmath148 becomes large compared to the stellar cutoff radius @xmath59 .
figure 5 shows @xmath9 as a function of @xmath132_{r_{\rm max}}$ ] for hst and hernquist stellar models with @xmath148 set equal to 0.5 , 1.0 , and @xmath168 .
as discussed above , @xmath169 ( 0.75 ) as @xmath170 for the hst ( hernquist ) model .
the curves pass through @xmath139 at @xmath171_{r_{\rm max}}$ ] if @xmath172 or @xmath173 , while @xmath174_{r_{\rm max}}$ ] is required if @xmath175 for hst ( hernquist ) models . as @xmath132_{r_{\rm max}}$
] becomes sufficiently large for fixed @xmath148 , dark matter comes to dominate the mass even at small radii ( _ i.e _ within @xmath32 where @xmath25 is calculated ) and @xmath9 becomes independent of @xmath119 .
that is , in order to maintain consistency with the fp and the faber - jackson relation , the physical mass must remain constant and the increase in dark - to - luminous mass ratio @xmath119 is compensated for by decreasing the stellar mass - to - light ratio ( _ i.e. _ , @xmath112 in equations 1316 ) accordingly .
we have shown ( section 4.1 ) that the lower limits on @xmath9 for the hst and hernquist stellar models without dark matter are 0.75 and 1.2 , respectively .
to explain observed values of @xmath9 that are less than this requires the addition of dark matter halos with sufficient mass to raise the value of @xmath24 , but with a large enough scale length to leave @xmath25 ( which is averaged over a much smaller scale ) relatively unaltered . by calculating solutions to equations ( 6 ) and ( 10 ) for the entire ( @xmath132_{r_{\rm max } } ,
r_{\rm dm}/r_*$ ] ) parameter space , we have derived limits on the amount and scale length of any dark halo able to reproduce a given observed value of @xmath9 . as discussed in section 4.3 and shown in figure 5 , for a fixed value of @xmath155
there is a minimum value of @xmath9 that is obtained when @xmath119 becomes sufficiently large that both @xmath25 and @xmath24 are determined by the non - luminous mass component .
since this minimum @xmath9 increases as @xmath155 decreases , there is a minimum value of @xmath155 required to explain any observed @xmath9 .
figure 6a shows these minimum values of @xmath155 for the hst ( solid curve ) and hernquist ( dotted curve ) models .
the minimum value of @xmath148 for a @xmath176 galaxy with @xmath139 is @xmath177 kpc ( @xmath143 ) ; the dimensional value scales as @xmath178 .
these extreme models are dark matter dominated , even inside the break radius @xmath32 , and have very low stellar mass - to - light ratios . requiring models to have @xmath179m@xmath13l@xmath14 for @xmath180 and @xmath115 ( or , equivalently , that @xmath460% of the mass within @xmath32 be stellar ) roughly doubles the minimum values of @xmath155 for @xmath181 .
this reasonable requirement then implies that @xmath3 kpc .
the maximum baryon fractions within @xmath32 , @xmath5 , and @xmath59 for any given @xmath9 . are shown in figure 6b .
the lower limits on the dark matter fractions in models required to explain @xmath182 are less than 5% within @xmath32 , and less than 20% within @xmath5 non - luminous halos could be quite inconspicuous unless data well beyond @xmath5 are considered .
the maximum baryon fractions at @xmath59 for the hst ( hernquist ) model are 0.43 ( 0.15 ) , 0.53 ( 0.27 ) , and 0.61 ( 0.45 ) to produce @xmath183 , 0.5 , and 0.6 , respectively .
these models have very large dark matter scale lengths ( @xmath184 ) and are analogous to `` maximum disk '' models of spiral galaxies in that they maximize the contribution of stars to the mass - to - light ratio . in order to investigate the scaling characteristics of dark halos with optical luminosity
, we consider the equivalent of the @xmath0@xmath1 relation in the @xmath9@xmath27 plane .
correlations of @xmath9 and log@xmath185 with log(@xmath27 ) are derived using the dw data set , blue magnitudes and @xmath186 colors from the rc3 catalog ( @xcite ) , and velocity distances from faber et al .
( the @xmath9log(@xmath27 ) data and best - fit correlation are re - plotted in figure 7b . )
galaxies in the centers of cluster cooling flows and those dominated by nuclear point source emission are excluded .
the log@xmath185log(@xmath27 ) correlation , equivalent to @xmath187 where @xmath188 is the visual luminosity in units of @xmath111l@xmath14 , is essentially identical to the convolution of the @xmath0@xmath1 correlation and faber - jackson relation ( see section 3.4 ) . nevertheless , requiring that models conform to the observations in this form
makes use of the tilt of the fp for scaling of the the stellar component , thus circumventing concerns about the thickness of the projection of the fp onto the @xmath27@xmath1 plane . note that the galaxies considered span @xmath189 , although most of the sample lies within the range @xmath190 .
given our homologous elliptical galaxy models , which are based on the scaling properties of the fp , the observed variation in @xmath9 implies that there is a systematic variation of relative dark matter properties which breaks the self - similarity for galaxies of different luminosity .
the sense of the observed @xmath0@xmath1 relation could also be explained by a fine - tuned systematic variation of the velocity dispersion anisotropy ( having more radial orbits for more luminous galaxies ) , or if more luminous galaxies have cuspier stellar distributions ( larger values of @xmath191 in equation 2 ) .
however , the latter is contrary to observations ; and as we have shown , the observed values of @xmath9 are too low to be explained without the introduction of dark matter .
it is possible to explain the @xmath9@xmath27 trend , assuming the dark - to - luminous mass ratio within @xmath59 is constant , if the ratio of dark matter to stellar break ( or , equivalently , effective ) radius , @xmath155 decreases with increasing galaxy luminosity @xmath27 .
however , since @xmath9 is flat for @xmath192 and steeply rises for @xmath193 ( figure 4a ) this requires fine - tuning .
moreover , as we discuss shortly , @xmath193 is not expected .
more naturally , the increase in @xmath9 with @xmath27 could be explained if @xmath132_{r_{\rm max}}$ ] _ decreases _ with _ increasing _ @xmath27 ( that is , if more luminous galaxies are less dark matter dominated ) and @xmath155 is relatively large ( see figure 5 ) .
we now discuss the plausibility of such a correlation within the context of the hierarchical clustering formalism of structure formation , simulations of which produced the dark matter distribution expressed by equations ( 8) and ( 9 ) .
this formalism provides specific predictions for @xmath148 as a function of virial mass ( defined here as the mass within a radius where the mean dark matter density is 200 times the average density in the universe ) for a given cosmology and power spectrum of initial fluctuations . in order to extend our models to the scale of the virial radius and mass
, we transform the dark matter parameters @xmath194 and @xmath132_{r_{\rm max}}$ ] that are directly constrained by observations into the following pair of global parameters : the dark matter concentration parameter @xmath195 and the global dark - to - luminous mass ratio @xmath196_{r_{200}},\ ] ] where @xmath197 refers to the radius within which the overdensity of the dark matter ( only ) is 200 .
( @xmath198 and @xmath197 refer to the original mass of the dark halo , but this may have been subsequently reduced in rich environments through tidal stripping . ) note that for @xmath199 and @xmath200 , @xmath201 kpc , @xmath202 , and @xmath132_{r_{\rm max}}=1.6 $ ] for @xmath203 ( @xmath204 )
. in bottom - up hierarchical clustering theories of structure formation , lower mass halos are more concentrated a reflection of the higher density in the universe at the earlier epoch of their formation ( @xcite ) .
we parameterize this dependence with the function @xmath205 for the concentration , where @xmath206 and @xmath207m@xmath18 is on the order of the present - day characteristic mass . the second term in the denominator of equation ( 21 )
is introduced to prevent the concentration from exceeding a maximum value @xmath208 , that we set equal to 100 , and driving the stellar mass - to - light ratio to very small values .
note that @xmath209^{-(1+\gamma)/\gamma}$ ] becomes insignificant for @xmath210 ( e.g , in standard cdm models ) . utilizing the `` characteristic density '' program kindly provided by j. navarro , a standard cold dark matter ( cdm ) fluctuation spectrum results in @xmath211 and @xmath212 . for these parameters
, @xmath213 ( @xmath214 for @xmath215 and @xmath216 ) and @xmath148 is indeed much greater than @xmath217 in the luminosity range of interest .
the dot - dashed curve in figure 7a shows @xmath132_{r_{\rm max}}$ ] as a function of @xmath188 for @xmath218 ( _ i.e. _ baryon fraction 0.06 ) cdm models ( and for two alternative models to be discussed in the next section ) .
clearly @xmath132_{r_{\rm max}}$ ] _ increases _ with optical luminosity , implying a _ decrease _ of @xmath9 with increasing @xmath27 as shown by the dot - dashed curve in figure 7b .
a comparison with the observed @xmath9@xmath188 correlation ( solid curve in figure 7b ) , demonstrates that _ a constant baryon fraction cdm model for elliptical galaxy dark matter halos clearly fails to match the observed correlation _ , even over the limited range of @xmath188 where the correlation is well - established .
a successful alternative model must produce dark halos where @xmath132_{r_{\rm max}}$ ] _ decreases _ with increasing optical luminosity in this range .
we will discuss the form such possible alternatives must take in the next section .
a natural explanation for why @xmath9 , the ratio of the central stellar velocity dispersion to the globally - averaged gas temperature , _ increases _ with @xmath25 and therefore @xmath27 is that the dark - to - luminous mass ratio within the luminous parts of elliptical galaxies ( @xmath219 ) _ decreases _ with @xmath27 ( see figures 5 , 7a and 7b ) .
in fact , the observed trend requires that the total mass - to - light ratio within @xmath59 be nearly constant .
this can be understood from a simple argument based on the virial theorem and fp , as follows .
the virial temperature within @xmath59 , @xmath220 since @xmath221 ( all masses are evaluated at @xmath59 ) . since , from the fp @xmath222
, it follows that @xmath223 , or @xmath224 if we associate @xmath225 with the integrated x - ray temperature , which should be a fair approximation .
the implication from the observed @xmath9@xmath27 trend that the total mass - to - light ratio within @xmath59 be nearly constant ( see below ) is contrary to expectations from cdm models of large scale structure if all elliptical galaxies have the same global baryon fraction .
the dark - to - luminous mass ratio is predicted to _ increase _ with @xmath27 , in such models , since the ratio of the dark matter scale length @xmath148 to the optical half - light radius is large and weakly varying , thus producing a predicted correlation between @xmath9 and @xmath27 in the opposite sense to what is observed ( figures 7a and 7b ) .
we consider two simple variations on the above model to try and recover the observed @xmath9@xmath27 trend .
first we assume that the dark matter concentration depends only weakly on halo mass , as cdm predicts , but relax the constant baryon fraction assumption .
then we assume that the concentration varies much more steeply with virial mass than the cdm prediction .
if the weak dependence of concentration on halo mass predicted in cdm simulations ( equation 21 with @xmath226 and @xmath227 ) is maintained , then the assumption of a constant baryon fraction within @xmath197 must be relaxed : the required increase in dark matter fraction within @xmath59 for less luminous galaxies must be a reflection of a larger _ total _
( _ i.e. _ within @xmath197 ) dark - to - luminous mass ratio .
if all galaxies start off with identical baryon fractions , this then implies that smaller galaxies lose an increasingly large fraction of their original baryonic mass .
( similar conclusions for spiral galaxies have been reached by @xcite . )
we parameterize this variation in the dark / stellar mass ratio as @xmath228,\ ] ] where , as previously defined , @xmath229_{r_{200}}$ ] , @xmath188 is the visual luminosity in units of @xmath111l@xmath14 , and @xmath230 is a fitting parameter . equation ( 22 ) has the properties @xmath231 as @xmath232 ( complete baryonic mass loss ) , and @xmath233 as @xmath234 ( presumably , @xmath235 is the average baryon fraction in the universe ) .
the observed @xmath9@xmath27 correlation is satisfactorally reproduced in models with @xmath236 , @xmath237 , and @xmath2382.5 .
the dotted curves in figures 7a and 7b represent hst stellar models with @xmath239 ; the observed trend is very well matched ( figures 7b ) .
( hernquist stellar models require even larger values of @xmath240 ; however , this is an artifact of the underestimate of @xmath9 for low @xmath241 ; see section 4.1 ) .
alternatively , if the dark matter halo concentration increases much more steeply with virial mass than predicted for a cdm fluctuation spectrum , the observed @xmath9@xmath27 correlation can be reproduced assuming a constant total baryon fraction ( _ i.e. _ , @xmath242 ) . in this case , the necessary decrease in @xmath132_{r_{\rm max}}$ ] for more luminous galaxies results not from a smaller relative global amount of dark matter , but from a larger fraction of the dark matter lying outside of the luminous portion of the galaxy ( _ i.e. _ @xmath59 ) .
we have found that equation ( 21 ) with @xmath243 , @xmath244 , and @xmath245 provides models that adequately reproduce the data ( smaller values of @xmath246 require larger values of @xmath242 ) . @xmath132_{r_{\rm
max}}$ ] and @xmath9 are shown as functions of @xmath27 , for @xmath218 , by the dashed curves in figure 7a and 7b .
such a steep dependence of the concentration on total mass would seem to indicate a flatter - than - cdm primordial fluctuation spectrum ( @xcite ) .
these alternative scenarios were designed to reproduce the @xmath9@xmath27 correlation without consideration of the scatter or that of galaxies about the mean fp .
we address these concerns by computing @xmath247 for individual pairs @xmath248 , @xmath249 corresponding to each of the 25 non - cluster - cooling - flow , non - point - source - dominated galaxies in the dw sample , for the three scenarios of figures 7a b .
a comparison with the observed values of @xmath9 , as shown in figures 8a c , confirms that the constant baryon fraction cdm scenario provides the poorest match to the observations , on average , and further illustrates the magnitude of variations about the mean correlation .
further investigation into these variations will be pursued in a future work that examines the individual galaxies in detail . both of these alternative scenarios have drawbacks .
mass loss induced by supernovae - driven galactic winds during the star - forming epoch of elliptical galaxies has been inferred from both the color - magnitude diagram and the enrichment of the intracluster medium ( e.g. , @xcite , @xcite ) .
however the magnitude of this effect must be extreme ( _ i.e _ @xmath240 , in equation 22 , on the order of 1 or greater ) if it is to explain the increasing dominance of dark matter in less luminous galaxies required to explain the observed @xmath0@xmath1 relation . the mass loss implied by equation ( 22 ) with the required set of parameters , when integrated over the elliptical galaxy luminosity function ( @xcite ; @xcite ) , corresponds to @xmath250% of the original baryonic mass of @xmath2 galaxies .
moreover , in such a scenario , all elliptical galaxies with @xmath251l@xmath14 would span a range in total mass of less than a factor of three ; the large range in mass loss could not , therefore , be due to less luminous galaxies lying within shallower potential wells . instead , one must presume that variations in the relative timing of star formation and the merging of pre - galactic fragments is responsible for the large range in mass loss , with galaxies with smaller present - day luminosities having lost the largest fraction of their original baryonic mass .
light would be a very poor tracer of mass for elliptical galaxies , and the luminosity function would be a reflection of the rarity of elliptical galaxies sufficiently relaxed at the star formation epoch to prevent copious mass loss from galactic winds .
it might therefore seem more appealing to resort to a concentration - mass relationship for dark halos ( equation 21 ) that is steeper than predicted in cdm simulations in such a way that optically less luminous galaxies are more dark matter dominated within @xmath59 , despite having the same total ( within @xmath197 ) baryon fraction .
the basic requirement is that the dark matter scale length @xmath148 increase from @xmath252 at the low - luminosity end of the observed range , to @xmath253 at @xmath254l@xmath14 , to @xmath255 ( @xmath256 ) at the high luminosity range .
these criteria are met for only a rather narrow volume in ( @xmath257 ) parameter space , and thus this scenario seems to suffer from a fine - tuning problem ; nonetheless , the range that successfully reproduces the observed @xmath0@xmath1 relation is not out of line for cosmogenic models with flat primordial fluctuation spectra ( @xcite ) .
figures 9a , 9b , and 9c show the luminosity dependences of mass , mass - to - light ratio , and baryon fraction at various radii for the two models discussed above that reproduce the observed @xmath0@xmath1 ( @xmath9@xmath27 ) relation : dotted lines show the model with cdm - predicted halo concentration as a function of mass and increasing mass loss with decreasing luminosity ( @xmath226 and @xmath227 in equation 21 , @xmath239 , @xmath236 , and @xmath258 in equation 22 ) ; dashed curves show the model with steeper - than - cdm concentration as a function of mass and constant baryon fraction ( @xmath244 and @xmath243 in equation 21 , @xmath259 and @xmath260 in equation 22 ) .
we refer to these as `` mass loss '' and `` non - cdm '' scenarios , respectively .
the stellar model is assumed to follow equation ( 3 ) , and the luminosity range in the plots roughly corresponds to that of the dw sample ( @xmath152@xmath261 ) .
the radii of interest are @xmath262 kpc , @xmath263 , and @xmath264 ; @xmath265 increases from 270 to 400 kpc for the mass loss scenario and @xmath266 kpc for the @xmath218 , non - cdm scenario , where @xmath110 with @xmath111l@xmath267 .
there are a number of robust properties that hold for any model ( including hernquist models ) to successfully reproduce the @xmath0@xmath1 relation .
since the models are constrained by the observed properties of the stars and by the average temperature of hot gas in hydrostatic equilibrium within @xmath264 , the inferred quantities at @xmath59 are nearly model - independent : @xmath268m@xmath18/l@xmath14 , @xmath269 .
note also that @xmath270m@xmath18 for any model where the central potential is dominated by the stellar component .
these values agree strikingly well with the constraints from statistical studies of gravitational lensing ( @xcite ; @xcite ; @xcite ) .
the solid lines in figure 9a shows the mass as a function of luminosity , evaluated at @xmath5 and @xmath59 for the softened isothermal sphere ( sis ) model that provides a best - fit to the observed weak gravitational shear from @xcite .
it is remarkable that the mass - to - light scaling derived from the shear that is dominated by @xmath271 galaxies at intermediate redshifts is in such excellent agreement with our results using a completely independent method on local galaxies primarily with @xmath2 . for the constant baryon fraction cdm model ,
the total mass within @xmath59 is a significantly steeper function of luminosity than is demonstrated by the curves in figure 9a . on scales
both larger and smaller than @xmath59 , the two scenarios depicted in figures 9a - c diverge . in the mass loss scenario
the baryon fraction within @xmath5 is generally greater than 0.7 and @xmath272m@xmath13l@xmath14 , while in the non - cdm scenario the mass - to - light ratio is nearly constant , @xmath273m@xmath13l@xmath14 , so that the baryon fraction within @xmath5 becomes less than 0.5 below @xmath274 . extrapolated out to the dark matter virial radius , @xmath197 , the total mass scalings differ ( by assumption ) . for the mass loss scenario , @xmath275 h_{80}{\rm m}_{\odot}/{\rm l}_{v_\odot},\ ] ] where acceptable models have @xmath2765 and @xmath2778 ( @xmath278 , @xmath279 for the model shown by dotted curves in figures 7a - b and 9a - c ) . for the non - cdm ( or any other constant baryon fraction )
scenario , @xmath280 the ( constant ) total baryon fraction @xmath281 for the dashed curves shown in figures 7a - b and 9a - c .
we found that @xmath282 in non - cdm models is required to reproduce the observed @xmath9@xmath27 correlation .
non - cdm models with relatively small global @xmath283 have larger baryon fractions within @xmath5 for @xmath284 ( because of larger dark matter scale lengths ) and agree somewhat better with the lensing constraints at this radius .
integrating over the early - type galaxy luminosity function ( @xcite ; @xcite ) yields an average total ( _ i.e. _ within the virial radius ) mass - to - light ratio @xmath285m@xmath13l@xmath14 for the non - cdm ( constant baryon fraction ) scenario and @xmath286 , which can be compared to an average total mass - to - light ratio of @xmath287m@xmath13l@xmath14 for spirals ( @xcite ) , and @xmath2880.06 . in the mass loss scenario ,
the average total mass - to - light ratio is sensitive to the lower luminosity cutoff , @xmath289 , for the luminosity function and can be as high as @xmath290m@xmath13l@xmath14 ( @xmath291 ) for @xmath292l@xmath14 since even the lowest luminosity elliptical galaxy has a total mass of @xmath293m@xmath18 .
the average density in the universe of virialized mass on galactic scales is @xmath294 of the critical density .
we have calculated the velocity dispersion distributions for both the dark matter and stellar density distributions , assuming isotropic orbits .
these are compared in figure 10 for an @xmath180 galaxy for both the mass loss and non - cdm scenarios .
both velocity dispersion profiles have maxima since the total gravitational potential is not isothermal .
the ratio ( dark - matter - to - stars ) of the squares of these maxima is greater than 1.4 over the luminosity range in figures 9a - c for the two models under consideration , and is @xmath295 over the range @xmath296 . in fact , the minimum value of this ratio for any hst or hernquist model that produces @xmath297 is greater than one . in this sense ,
the dark matter is hotter than the stars , as predicted by davis & white ( 1996 ) .
in this work , we have addressed two essential features of the x - ray temperatures derived by davis & white ( 1996 ) for an optically complete sample of elliptical galaxies : ( 1 ) the x - ray emitting gas is always hotter than the stars and , typically , twice as hot ( @xmath298 ) ; ( 2 ) the gas / stellar temperature ratio tends to be higher for galaxies with lower velocity dispersions ( the `` @xmath0@xmath1 relation '' ) .
we have constructed physically plausible models of the mass distribution in bright elliptical galaxies in an effort to constrain their average dark matter properties by matching these observations .
the stellar models ( described in 3 ) are fully consistent with the fundamental plane scaling relations , and are designed to either conform to the latest published _ hst _
results on the structure of the centers of elliptical galaxies or to follow the hernquist ( 1990 ) approximation to a de vaucouleurs profile ( see figure 2 ) .
we have made other plausible , but conservative , assumptions , such as ( 1 ) maximizing the non - gravitational ( _ i.e _ pressure ) contribution to the gas temperature by allowing the gas and dark matter distributions to extend to infinity and ( 2 ) assuming that the stellar orbits are isotropic for @xmath299 .
a basic and general result of our calculations is that , in the absence of dark matter , @xmath300 ( see figure 3 ) . since @xmath139 is observed , a convincing case is made that dark matter is an extremely common , if not ubiquitous , constituent of elliptical galaxies . although x - ray ( and other ) observations have been used to infer the presence of dark matter in individual cases , we have shown that dark halos are generic to luminous , nearby elliptical galaxies .
furthermore , the dark matter / stellar temperature ratios ( derived at the peak of their velocity dispersion distributions , assuming isotropic orbits ) are greater than one for models with @xmath297 .
thus , the observation that the temperature of the extended hot gas exceeds the central stellar temperature is a reflection of the fact that the dark matter is dynamically hotter " than the stars , as suggested by davis & white ( 1996 ) . in 4 we described how @xmath9 varies as functions of the two relative parameters of the `` universal '' dark matter distribution ( equation 8) the ratio of dark matter to stellar scale lengths , @xmath301 , and the ratio of dark - to - luminous matter within @xmath264 , @xmath302_{r_{\rm max}}$ ] ( see figures 4a - b and 5 ) .
because we attempt to match only the single global observable @xmath9 , there is an allowed range in the details of the dark matter spatial distribution ; however , we have derived absolute limits on the dark matter parameters required to obtain any particular value of @xmath9 ( see figures 6a - b ) .
the observations do not require that dark matter dominate the inner luminous regions of elliptical galaxies : more than half of the mass within @xmath5 is baryonic for models with @xmath303 if @xmath158 .
the most natural explanation of the tendency for galaxies with lower stellar temperatures to have larger gas - to - stellar temperature ratios is that @xmath132_{r_{\rm max}}$ ] decreases with @xmath27 in such a way that , on average , the total mass - to - light ratio inside @xmath59 is nearly independent of optical luminosity .
this ratio , @xmath304m@xmath13l@xmath14 , is exactly what is predicted for mass models of elliptical galaxies designed to explain the gravitational shear of background field galaxies .
if one specifies a scaling relation for the dark halo concentration , one can extend the dark matter distribution out to the virial radius and calculate the total baryon fraction and mass - to - light ratio .
when we attempt to embed our models within the cdm theory of hierarchical halo formation , the implied dark matter scaling badly fails to reproduce the observed @xmath0@xmath1 relation unless smaller galaxies lose an increasingly larger fraction of their initial baryonic content ( see figure 7b ) , such that the average @xmath2 galaxy has lost most of its initial baryonic mass .
alternatively , the global dark - to - luminous mass ratio could be constant if the dark halo concentration declines much more steeply with virial mass than cdm models predict , so that the decrease in @xmath132_{r_{\rm max}}$ ] for larger systems is a result of a more diffuse dark matter halo rather than a less massive ( relative to the stars ) halo . in this latter scenario ,
dark matter may become increasingly important inside @xmath5 as @xmath27 decreases , becoming dominant for @xmath305 .
this deviation from cdm predictions of dark halo scaling could conceivably be due to a relatively flat primordial fluctuation spectrum on mass scales @xmath306m@xmath18 or to the effects on the dark matter density profile of the evolution of the baryonic component ; to date , large scale structure numerical simulations that can resolve halos on galactic scales _ and _ include a dissipational component have not been attempted .
in this paper we have shown that the observed relationship between optical velocity dispersions and x - ray temperatures in giant elliptical galaxies implies that they have dark matter halos with @xmath12m@xmath13l@xmath14 within @xmath8 . in the future
, we plan to fully utilize the available x - ray and optical imaging and spectroscopic data for our sample on a case - by - case basis in order derive the mass distributions in individual galaxies in the highest possible detail and to investigate the scatter in the @xmath0@xmath1 relationship . for galaxies with x - ray temperature profiles
, we will be able to constrain the detailed form of the dark matter distribution as well as its integrated properties .
we thank lars hernquist for reminding us of the consistency considerations of ciotti & pelligrini ( 1992 ) , and richard mushotzky for feedback on the original manuscript .
comments from an anonymous referee led to significant improvements in the quality of this paper . r.e.w .
acknowledges partial support from nasa grants nag 5 - 1718 and nag 5 - 1973 .
suppose the stellar density profile is given by a broken power - law of the form @xmath307 the projected surface mass , @xmath308 can be recast , utilizing the change of variable @xmath309 , into the form @xmath310,\ ] ] where @xmath311 @xmath312 , and @xmath313 .
note that the bracketed term in equation ( a3 ) has the asymptotic slopes at small and large @xmath31 that one would expect from dimensional analysis ( _ i.e. _ @xmath314 ) .
the definite integral of equation ( a4 ) can be evaluated as follows ( @xcite ) : @xmath315 where @xmath316 is one - half the asymptotic density slope as @xmath317 , @xmath318 is the beta function , and @xmath319 is the hypergeometric function ( @xcite ) .
we have empirically found that we can introduce a parameter @xmath320 , recast equation ( a3 ) as @xmath321 where @xmath322 and determine a value of @xmath320 ( for any @xmath323 ) that results in a very nearly constant @xmath324 as a function of @xmath31 ( with no significant structure near @xmath325 ) . equation ( a6 )
has the expected slope as @xmath326 ; and , we find @xmath327 as @xmath328 ( the `` @xmath329 model '' ) , and @xmath330 as @xmath331 .
we conclude that the projection of a density profile of the form expressed by equation ( a1 ) is well - described by equation ( 2 ) ( at least for @xmath38 ) , thus perhaps lending some physical justification to its use .
bahcall , n. a. , lubin .
l. m. , dorman , v. 1995 , , 447 , l81 bertola , f. , pizzella , a. , persic , m. , & salucci , p. 1993
, , 416 , l45 brainerd , t. g. , blandford , r. d. , & smail , i. 1996 , , 466 , 623 buote , d. a. , & canizares , c. r. 1997 , , 474 , 650 carollo , c. m. , de zeeuw , t. , van der marel , r. p. , & danziger , i. j. 1995 , , 441 , l25 ciotti , l. , & pelligrini , s. 1992 , , 255 , 561 davis , d. s. , & white , r. e. iii 1996 , , 470 , l35 ( dw ) de paolis , f. , ingrosso , g. , & strafella , f. 1995 , , 438 , 83 de vaucouleurs , g. , et al .
1991 , third reference catalogue of bright galaxies ( new york : springer ) de zeeuw , t. , & franx , m. 1991 , , 29 , 239 fabbiano , g. 1989 , , 27 , 87 faber , s. m. , et al .
1997 , , 114 , 177 faber , s. m. , wegner , g. , burstein , d. , davies , r. l. , dressler , a. , lynden - bell , d. , & terlevich , r. j. 1989 , , 69 , 763 fabian , a. c. , thomas , p. a. , white , r. e. iii , & fall , s. m. 1986 , , 221 , 1049 forman , w. , jones , c. , & tucker , w. 1985 , , 293 , 102 franx , m. , van gorkom , j. , & de zeeuw , t. 1994 , , 436 , 642 gradshteyn , i. s. , and ryzhik , i. m. 1980 , table of integrals , series , and products , new york : academic press , 287 and 413 gerhard , o. , jeske , g. , saglia , r. p. , & bender , r. 1998 , , 295 , 197 gibson , b. k. 1997 , , 290 , 471 griffiths , r. e. , casertano , s. , i m , m. , & ratnatunga , k. u. 1996 , , 282 , 1159 hernquist , l. 1990 , , 356 , 359 irwin , j. a. , & sarazin , c. l. 1996 , , 471 , 683 kochanek , c. 1995 , , 445 , 559 lin , h. , kirshner , r. p. , shectman , s. a. , landy , s. d. , oemler , a. , tucker , d. l. , & schechter , p. l. 1996 , , 464 , 60 loewenstein , m. 1994 , , 431 , 91 loewenstein , m. , & mathews , w. g. 1987 , , 319 , 614 loewenstein , m. , & mushotzky , r. f. 1996 , , 466 , 695 marzke , r. o. , huchra , j. p. & geller , m. j. 1994 , , 428 , 43 matsushita , k. 1997 , ph.d .
thesis , university of tokyo mulchaey j. s. , davis , d. s. , mushotzky , r. f. , & burstein , d. 1996 , , 456 , 80 mushotzky , r. f. 1998 , proc .
usa , 95 , 72 mushotzky , r. f. , loewenstein , m. , awaki , h. , makishima , k. , matsushita , k. , & matsumoto , h. 1994 , , 436 , l79 navarro , j. f. , frenk , c. s. , & white , s. d. 1997 , , 490 493 persic , m. , salucci , p. , & stel , f. 1996 , , 281 , 27 rix , h .- w .
, de zeeuw , t. , c. m. , cretton , n. , van der marel , r. p. & carollo , c. m. 1997 , , 488 , 702 saglia , r. p. et al . 1993 , , 403 , 567 white , r. e. iii , & davis , d. s. 1998 , in preparation zepf , s. e. , & silk , j. 1996 , , 466 , 114 | given the recently deduced relationship between x - ray temperatures and stellar velocity dispersions ( the `` @xmath0@xmath1 relation '' ) in an optically complete sample of elliptical galaxies ( @xcite ) , we demonstrate that @xmath2 ellipticals contain substantial amounts of dark matter _ in general_. we present constraints on the dark matter scale length and on the dark - to - luminous mass ratio within the optical half - light radius and within the entire galaxy .
for example , we find that minimum values of dark matter core radii scale as @xmath3 kpc and that the minimum dark matter mass fraction is @xmath420% within one optical effective radius @xmath5 and is @xmath6@xmath7% within @xmath8 , depending on the stellar density profile and observed value of @xmath9 .
we also confirm the prediction of davis & white ( 1996 ) that the dark matter is characterized by velocity dispersions that are greater than those of the luminous stars : @xmath10@xmath11 .
the @xmath0@xmath1 relation implies a nearly constant mass - to - light ratio within six half - light radii : @xmath12m@xmath13l@xmath14 .
this conflicts with the simplest extension of cdm theories of large scale structure formation to galactic scales ; we consider a couple of modifications which can better account for the observed @xmath0@xmath1 relation . |
there have been major breakthroughs in understanding high - density qcd evolution recently .
several important observations have been made : ( i ) gluon number fluctuations have a big effect on the evolution towards gluon saturation @xcite , ( ii ) a connection between high - energy qcd evolution and statistical physics models of reaction - diffusion type was found @xcite which has clarified the interpretation of fluctuations in an event - by - event picture , and ( iii ) it was realized that the relevant fluctuations are missed @xcite by the existing balitsky - jimwlk equations balitsky:1995ub+x , jalilian - marian:1997jx+x .
these observations have led to the creation of new qcd evolution equations which include gluon number fluctuations or pomeron loops mueller:2005ut , iancu:2005nj , kovner:2005nq , hatta:2005rn .
the new equations describe the usual bfkl evolution , pomeron splittings and pomeron mergings and , thus , through iterations , pomeron loops .
a lot is known about the structure of the new equations kovner:2005en+x , blaizot:2005vf+x , marquet:2005hu , levin:2005au , enberg:2005cb , however , because of their complexity , little about their solutions .
so far , analytical results for the energy dependence of the saturation momentum and for the scaling behavior of the @xmath1-matrix at asymptotic energies mueller:2004se , iancu:2004iy and some numerical simulations of approximations to the exact equations @xcite are available . solutions to the new equations in the range of collider energies are desired . in this paper
, we consider a toy model which has a structure similar to that of the new qcd evolution equations , but which has no transverse dimensions .
this simplification allows an exact solution of the evolution equations .
the particle - particle scattering amplitude is calculated .
the effects of saturation and unitarity , or pomeronloops which are the characteristics of the new evolution equations , are investigated .
a decade ago , mueller @xcite studied a similar toy model without transverse dimensions which was motivated by the qcd dipole model @xcite
. he has been able to solve the toy model exactly and calculate the @xmath2-matrix .
observations made by studying this toy model ( e.g. the one that fluctuations in particle numbers lead to a divergent multiple scattering series ) have been shown numerically salam:1995uy , salam:1995zd to be valid also in the four - dimensional qcd .
the hope that some of the results obtained in the toy model discussed in this paper may also apply in four - dimensional qcd is one of the motivations for this work . in the toy model proposed by mueller @xcite ,
particle - particle scattering in the center - of - mass frame at total rapidity @xmath3 was considered . in this model
, each particle evolves to a dense system through particle branching ( splitting ) .
unitarity effects ( at leading order accuracy ) are correctly described due to the multiple scatterings between the dense systems .
there are no particle mergings in the wavefunctions of the particles in this model , therefore , the model fails to describe saturation effects in the wavefunctions of the particles which become important at very high rapidities @xmath4 . in the language of loops , only the loops stretching over the rapidity interval greater than @xmath5 ( large loops ) shown e.g. in fig .
[ reggeon]b are included ( at leading - order accuracy and for @xmath6 ) .
large loops are created by matching the evolution ( pomeron splittings ) of both particle s wavefunctions at @xmath5 . however , there are no loops inside the particle s wavefunctions ( small loops ) where the pomeron splits and merges , as shown e.g. in fig .
[ reggeon]c(see , e.g. ref.@xcite ) , within the rapidity interval @xmath5 .
the above limitations of the toy model are , of course , properties of the qcd dipole model @xcite as well and are well - known .
mueller s toy model breaks down at @xmath7 . in a later work mueller and salam
@xcite did include saturation effects by imposing boost invariance on the small-@xmath8 evolution in a toy model , getting an expression for the @xmath2-matrix which is valid also for @xmath9 .
the effect of saturation was an extension of the validity region of the result obtained for @xmath10 in @xcite also to the region where @xmath9 .
our toy model which is inspired by the new qcd evolution equations mueller:2005ut , iancu:2005nj , kovner:2005nq , hatta:2005rn allows us to study unitarity effects as well as saturation effects in a more refined and systematic way .
we calculate the leading order ( lo ) and the next - to - leading order ( nlo ) loop contributions to the particle - particle scattering amplitude .
the lo result , which is the lo contribution of the large loops , agrees with the results from @xcite .
our nlo result consisting of the nlo contribution of the large loops and the lo contribution of the small loops is new .
the nlo correction with respect to the lo result is of order @xmath11 .
this is the main result of the paper .
it tells us that the nlo contribution is negligible as compared with the lo contribution up to rapidities @xmath12 .
our lo result agrees with the result derived within mueller s toy model up to @xmath7 . beyond
@xmath13 mueller s toy model fails to include saturation effects .
the result in @xcite which involves saturation effects is the same as our lo result until @xmath14 which is parametrically much larger than @xmath13 . however , beyond @xmath14 the nlo corrections become important which are not taken into account in mueller and salam s toy model @xcite .
presumably our toy model reveals the properties of the four - dimensional qcd evolution to some extent .
in fact , its behavior is a heuristic sign that the nlo contribution can be neglected in the qcd evolution as well as long as @xmath15 .
such a nlo correction seems to appear also in real qcd .
considering the one - loop diagram for example , the lo contribution in qcd is @xmath16 which comes from varying the size of the loop from @xmath17 to @xmath18 , where @xmath19 .
the nlo contribution @xmath20 comes from changing the location of a fixed zero - size loop from @xmath21 to @xmath18 .
the suppression of the nlo contribution by the factor @xmath22 directly leads to the general correction @xmath23 ( the same happens in our toy model ) .
the paper is organized as follows : in sec .
[ sec_the_toy_model ] , we show the toy model .
two methods are developed in order to solve the toy model and some discussion is provided on the consequences of the nlo corrections .
they are explained in sec .
[ solution_tm ] . in the appendices we show the lorentz - invariance of the toy model and give the solution to another commonly used zero - dimensional model .
finally , we discuss the results and give the conclusions .
in this section , we consider a toy model without transverse dimensions which has a structure similar to that of the recently found four - dimensional qcd evolution equation mueller:2005ut , iancu:2005nj , kovner:2005nq , hatta:2005rn .
the toy model describes the rapidity evolution of the particle number @xmath24 of a system .
the dynamics in zero - transverse dimensions is given by the following langevin equation in appendix .
[ other_langevin_equation ] ] @xmath25 where @xmath26 is a gaussian white noise : @xmath27 and @xmath28 .
( [ langevin ] ) is known as the zero transverse dimensional reggeon field theory equation and it contains the projectile - target duality ( see appendix [ duality ] ) . the structure of eq .
( [ langevin ] ) is obvious : the first term on the r.h.s .
represents the growing of the particles , the second term describes particle recombinations which limit the growth at a maximum occupancy , and the third term describes the particle number fluctuations .
these are essentially the main ingredients of the real qcd equations . with ito s calculus
, one can write eq .
( [ langevin ] ) into an infinite hierarchy of coupled evolution equations , @xmath29 where @xmath30 is the expectation value of @xmath31-particles during the evolution and it should be understood as a normal ordered number operator ( i.e. the factorial moment @xmath32 ) according to its physical interpretation . here , @xmath33 is the growth term , @xmath34 the fluctuation term , and @xmath35 the recombination term . eqs .
( langevin , [ h1 ] ) are the zero - transverse - dimensional analog of the real qcd equations ( see @xcite ) .
a simple derivation and a discussion in the context of qcd evolution of eqs .
( [ langevin ] , [ h1 ] ) can be found in @xcite .
the solution to another widely discussed langevin equation is shown in appendix other_langevin_equation .
( [ h1 ] ) is quite complicated since @xmath36 is coupled with @xmath37 and @xmath38 in the evolution , i.e. , eq .
( h1 ) is just a particular equation within an infinity hierarchy of coupled equations .
however , the following observation helps to solve eq .
( [ h1 ] ) : one can neglect the recombination term if one starts with a dilute object at the beginning of rapidity evolution , on the other hand , one can see that @xmath39 is an approximate fixed point of eq .
( h1 ) in extremely high rapidity and @xmath40 limit .
this will be more clear if one thinks about this in terms of the evolution equation of @xmath41(one can define @xmath42 . ) : @xmath43 \ , $ ] in which it is easy to find @xmath44 . therefore ,
if one starts with a dilute object at @xmath45 ( @xmath46 ) , then one can consider the evolution equation without the recombination term at the early stages of the evolution , and turn on the recombination term as a perturbation at the end of the evolution which forms the pomeron loops .
following this philosophy , we have found two different methods to solve eq .
( [ h1 ] ) which we will present in the following subsections . before entering the calculations
, we should state that we always work in a parametrical limit in which @xmath47 and @xmath48 , but @xmath49 is fixed and finite .
( [ h1 ] ) without the recombination term , @xmath50 can be solved exactly if given some physically chosen initial conditions . in this paper
, we consider the evolution of a single particle for which the initial conditions are @xmath51 and @xmath52 .
the solution to eq .
( [ h2 ] ) with the above initial conditions reads @xmath53and corresponds to the @xmath31-particle density obtained via splitting from a single particle after the evolution over the rapidity @xmath54 . to turn on the perturbation ,
let us start with the first equation in the infinite hierarchy ( [ h1 ] ) , @xmath55 which has the solution @xmath56 it is straight forward to recognize that the first term in eq . ( [ s ] ) is the usual bfkl term , and the second term corresponds to the one - loop diagram with the right intrinsic minus sign .
one can also see that @xmath57 is the nlo contribution in the parametrical limit we mentioned above . to get the lo two - loop diagram shown in fig .
[ reggeon ] b , we start with the second equation in eq .
( [ h1 ] ) , @xmath58 whose solution is @xmath59 then turn on the perturbation , @xmath60 , in order to take into account the recombination from @xmath61 to @xmath62 ( in addition to the recombination from @xmath63 to @xmath61 given by eq .
( [ s0 ] ) ) , which leads to @xmath64 the third term in eq .
( [ ss2 ] ) represents the lo two - loop diagram shown in fig .
[ reggeon ] b. in order to calculate the nlo two - loop diagram shown in fig .
[ reggeon ] c , one has to enforce a recombination to happen after a splitting .
thus , starting with @xmath65 , one splitting given by @xmath66 , followed by one merging @xmath67 , yields the result for the one - loop diagram .
a further splitting and a subsequent merging gives the result for the nlo two - loop diagram , @xmath68 in fact , in order to calculate any loop diagram , one just needs to know the merging and splitting vertices which according to eq .
( [ h1 ] ) read , respectively , @xmath69 for particle - particle scattering , the @xmath70-th order diagram contains @xmath31-splittings and @xmath31-mergings : the leading - order diagram corresponds to the diagram in which all @xmath31-splittings occur before all the @xmath31-mergings , the next - to - leading order diagram is always the diagram in which exactly one merging takes place before one splitting .
the sum over all lo and nlo loop diagrams gives the lo and nlo contributions to the @xmath1-matrix , respectively . for particle - particle scattering ,
the sum over the lo loop diagrams ( leading and next - to - leading contribution ) reads @xmath71 while the sum over the nlo loop diagrams ( leading contribution ) is @xmath72 the resulting @xmath2-matrix , @xmath73 ( with @xmath74 ) , which includes lo and nlo loop diagrams , can be written in the form @xmath75 where we have put @xmath76 in close analogy with the four - dimensional qcd and @xmath77 .
@xmath78 and @xmath79 are functions of @xmath31 , which are too complicated to be calculated in this method . the series in eq .
( [ smatrix ] ) is a divergent series , however , it is borel - summable .
after the borel - summation , the expression for the @xmath2-matrix becomes rather complicated , @xmath80 -\left [ \frac{7}{\alpha_{s}^{2}e^{\alpha y } } + \frac{3}{(\alpha_{s}^{2}e^{\alpha y})^2}\right ] \right\ } \nonumber\\&+ & ( \alpha_s^2 ) \left\{\tilde{f_1}(\alpha_{s}^{2}e^{\alpha y})\right\ } + ( \alpha_{s}^{2}\alpha y)^2 \left\{\tilde{f_2}(\alpha_{s}^{2}e^{\alpha y})\right\ } + ... \ , \label{s_exact}\end{aligned}\ ] ] however , it gets considerably simplified when @xmath81 becomes large , @xmath82 \ , \label{approx}\end{aligned}\ ] ] where @xmath83 is the incomplete gamma function .
the nlo correction term @xmath84 inside the brackets with respect to the leading contribution is new . in this section ,
we develop another more powerful method , the @xmath86-representation , to solve the toy model .
this method exhibits the pole - structure of the toy model which bears some resemblance to the pole - structure of the bfkl equation in qcd .
the @xmath87-representation is equivalent to the integral representation method in sec .
integral_representation , however , it enables us to go one step further by resumming all the @xmath23 terms in the @xmath2-matrix .
therefore , this method allows us to get an expression for the @xmath2-matrix which is assumed to be valid up to the rapidity @xmath88 where the next - to - next - to - leading order @xmath89 becomes of order one .
let us show how to solve eq .
( [ h1 ] ) with the @xmath86-representation method .
one starts with the laplace transform of eq .
( [ h1 ] ) @xmath90 where we have used the definitions @xmath91 and @xmath92 . with the initial condition @xmath93 and @xmath94 , eq .
( [ lp ] ) can be written into a matrix form : @xmath95 it is fairly complicated to solve this equation directly because of the infinite dimension it has . nevertheless , there are two ways to attack this problem .
first , we can follow the idea that the @xmath96 is small , which reduces the importance of recombination at the beginning of the evolution .
then we can solve the above linear algebra equation perturbatively which will prove that this approach is completely equivalent to the integral representation method we developed in the previous section .
second , noticing that eq .
( [ matrix ] ) has an exact solution at finite dimensions , we can truncate the eq .
( [ matrix ] ) to a @xmath97 linear algebra equation and then take the @xmath98 limit to get the solution of our infinite dimension matrix . first , let us turn off the perturbation and only solve the unperturbed matrix to get the corresponding @xmath99 .
one can easily find the solution to be @xmath100 .
after the inverse laplace transformation , @xmath101the result is exactly the same as what we have obtained in the previous section .
turning on the perturbation , we find that the first - order correction @xmath102 to @xmath99 satisfies the equation : @xmath103 the general solution of this equation is too complicated to be written down . however , the strategy of how to get it is straightforward .
for example , we can easily obtain @xmath104 which corresponds to the one - loop diagram .
it yields the same result as given in eq .
( [ s ] ) after the inverse laplace transformation .
this example shows that the result for each diagram calculated in the previous section becomes pretty much simplified in the @xmath86-representation . to see this more clearly , one can do the laplace transform of eq .
( [ merging ] ) and eq .
( [ splitting ] ) which yields @xmath105 according to these equations , one can directly write down the contribution for any diagram in this @xmath86-representation since the merging of @xmath106-particles to @xmath31-particles is nothing but adding a @xmath107 pole with a prefactor @xmath108 in @xmath86-complex plane and the splitting of @xmath109-particles to @xmath31-particles is nothing but adding a @xmath110 pole with a prefactor @xmath111 . although the problem has been simplified a lot in the @xmath86-representation , the exact solution to eq .
( [ matrix ] ) is nevertheless unavailable this way .
we get the solution to eq .
( [ matrix ] ) as follows : we first truncate the infinite dimension matrix ( [ matrix ] ) into a finite @xmath97 matrix , then exactly solve the amputated equation .
let us show the solutions for the first four @xmath31-values : * @xmath112 case : one obtains @xmath113 which transforms to @xmath114 .
* @xmath115 case : one obtains @xmath116 which transforms to @xmath117 * @xmath118 case : one obtains @xmath119 } { % \left ( \omega -1\right ) \left ( \omega -2\right ) \left ( \omega -3\right ) + 2\alpha _ { s}^{2}\left ( 7\omega -9\right ) } $ ] .
then one can set the determinant @xmath120 which yields three solutions .
each of those three solutions stands for three poles
@xmath121 , @xmath122 and @xmath123 , respectively .
the inverse laplace transform changes the integer exponents to non - integer renormalized exponents , @xmath124 * @xmath125 case : one obtains @xmath126 } { \left ( \omega -1\right ) \left ( \omega -2\right ) \left ( \omega -3\right ) \left ( \omega -4\right ) + 2\alpha _ { s}^{2}\left ( 72 - 91\omega + 25\omega ^{2}\right ) + 72\alpha _ { s}^{4 } } $ ] .
the denominator has 4 poles which are @xmath127 , @xmath128 , @xmath129 and @xmath130 .
then , the inverse laplace transform gives @xmath131 from the above calculation for @xmath132 , one can already see that for an arbitrary @xmath31 , the first @xmath109 poles are stabilized to the corresponding solutions @xmath133 @xmath134 with a fixed coefficient @xmath135 and the last pole is always @xmath136 with a wrong coefficient since it misses the recombination from @xmath137 particle mode as a result of truncation at the @xmath31 particle mode .
in addition , it is easy to see that the @xmath138 term is corresponding to the sum of the nlo contributions @xmath139 of the leading loop diagrams ( e.g. fig .
[ reggeon ] b ) and the leading contributions ( @xmath140 ) of the nlo loop diagrams ( fig . [ reggeon ] c ) . on the other hand
, the @xmath141 term in the last pole only corresponds to the leading order contributions of the nlo loop diagrams ( e.g. fig .
reggeon c ) as a consequence of the truncation at dimension @xmath31 which rules out the leading loop diagram but allows the nlo loop diagrams at the @xmath31 particle mode .
the pattern of the exact solution is transparent from the above discussion and the generalization to infinite dimensions follows by simply taking the limit @xmath98 .
the exact solution for the @xmath2-matrix in infinite dimensions reads @xmath142 this result agrees with what we have found in the previous section , see eq .
( [ smatrix ] ) , however , it goes a step further since the @xmath143 term is now in the exponent which can be understood as a resummation of all @xmath144 terms via exponentiation in eq .
( [ smatrix ] ) .
we notice that the @xmath2-matrix is not borel - summable since @xmath145 is beyond the borel - summability according to nevanlinna s theorem @xcite .
however , one can generalize the borel - summation technique and show that the above series still has a finite sum by using analytic continuation .
generally speaking , one can see that the @xmath2-matrix is the analytic function of two independent variables , @xmath146 and @xmath147 , when @xmath96 is negative in the @xmath148 complex plane .
since @xmath149 is not a singularity , one can use analytic continuation to define the other half plane when @xmath148 becomes positive . in the following ,
we provide a generalized borel - summation method which yields a finite result .
we use a technique together with a subsequent borel - summation in order to maneuver the divergent series eq . ( [ s_exact_2 ] ) into a definite and well - defined sum .
let us define @xmath150 and @xmath151 , then we obtain @xmath152 to check the convergence of this integral , one can expand @xmath153 at @xmath154 which gives @xmath155 , showing that the above integral is definitely finite and well - defined .
the @xmath8-integration can not be performed analytically .
however , one can use the saddle point approximation to evaluate the above @xmath8-integral , in which one finds the saddle point @xmath156 , and @xmath157 a good agreement of this result with the numerical evaluation of eq .
( num ) is shown in fig .
( [ tab_2],[tmatrix ] ) .
-matrix as a function of @xmath158 for @xmath159 in logarithmic scale : the solid line represents the solution of the zero - transverse dimensional kovchegov equation .
the dashed line corresponds to the leading order result , the dotted line stands for the saddle point approximation ( [ exact2 ] ) and the dot - dashed line is extracted from the numerical evaluation of eq .
( [ num ] ) . ] -matrix as a function of @xmath158 for @xmath159 : the solid line represents the solution of the zero - transverse dimensional kovchegov equation .
the dashed line corresponds to the leading order result , the dotted line stands for the saddle point approximation ( [ exact2 ] ) and the dot - dashed line is extracted from the numerical evaluation of eq .
( [ num ] ) . ] eq .
( [ exact2 ] ) is a fairly exact result in the high - energy limit since it presumably resums all the sub - leading diagrams up to @xmath160 level , only neglects @xmath161 ( in the prefactor ) and @xmath162 ( in the exponent ) corrections . if we compare this with eq .
( [ approx ] ) of the previous section , we can see that eq .
( [ approx ] ) is an approximate solution which results from eq .
( [ exact2 ] ) when @xmath163 .
the @xmath86-representation is equivalent to the integral representation method in sec .
[ integral_representation ] , however , it enables us to go one step further by resumming all the @xmath11 terms in the @xmath2-matrix .
thus , the @xmath2-matrix in eq .
( exact2 ) is assumed to be valid up to the rapidity @xmath88 where the next - to - next - to - leading order @xmath164 becomes of order one . in mueller s toy model @xcite , which was inspired by the qcd dipole model ( dm ) @xcite ,
the @xmath2-matrix for particle - particle scattering in the center - of - mass frame is given by @xmath165 where @xmath166 is the probability density for having @xmath24 particles in the wavefunction of an initial single particle after the evolution up to @xmath18 .
for large rapidities , @xmath166 reads @xcite @xmath167 with @xmath168 . using @xmath169 and @xmath170 and changing the summations in eq .
( [ borels ] ) into integrals ( this change is rigorous since @xmath171=@xmath172=@xmath173 is infinitesimal ) for large rapidities , one obtains @xmath174 \ \label{dist_ns_int}\end{aligned}\ ] ] which after the integration over @xmath175 and @xmath176 leads to @xmath177 \ . \label{dm_result}\end{aligned}\ ] ] this result reduces for @xmath178 to @xmath179 \label{y_sy0}\end{aligned}\ ] ] while for @xmath180 it gets @xmath181 \ . \label{y_ly0}\end{aligned}\ ] ] it is easy to see that the result in eq .
( [ dm_result ] ) agrees with our lo contribution given in eq .
( [ approx ] ) only when @xmath182 . for @xmath183
the result from mueller s toy model given in eq .
( [ y_ly0 ] ) can not be trusted since it does not include saturation effects within the interacting particle wave functions which start becoming important at @xmath184 .
mueller s toy model fails to reproduce the lo contribution of the lo diagrams ( e.g. fig [ reggeon ] b ) at @xmath185 , it does not include the nlo contribution of the lo diagrams ( terms including @xmath186 in eq .
( [ lo_loop ] ) ) and it also misses the nlo loop diagrams ( e.g. fig [ reggeon ] c ) .
mueller and salam @xcite have been able to calculate the probability distribution @xmath166 which includes saturation effects by imposing boost invariance on the small - x evolution .
they calculated the boost invariant s matrix in the rest frame of the target , and obtained the same lo result in eq .
( [ approx ] ) .
we also find that it is interesting to calculate the s matrix in the center of mass frame .
inserting their result , @xmath187 , \label{dist_ws}\end{aligned}\ ] ] in eq .
( [ borels ] ) , converting the summations into integrals , changing the integration variables from @xmath188 to @xmath189 and from @xmath190 to @xmath24 , setting @xmath191 , and using the integration variables @xmath192 and @xmath193 , one ends up with the boost - invariant result for the @xmath2-matrix in the center of mass frame , @xmath194 \label{limit_0 } \\ & = & \frac{1}{\alpha_s^2 e^{\alpha\,y}}\ e^{\frac{1}{\alpha_s^2 e^{\alpha\,y}}% } \ \gamma(0,\frac{1}{\alpha_s^2 e^{\alpha\,y } } ) \label{res_ws}\\ & \simeq & \frac{1}{\alpha _ { s}^{2}e^{\alpha y}}\ln \left [ \alpha _ { s}^{2}e^{\alpha y}\right ] .
\label{res_ws1}\end{aligned}\ ] ] this result agrees with our lo contribution in eq .
( [ approx ] ) .
note , however , that the mueller - salam toy model does not incorporate the nlo correction given in eq .
( [ approx ] ) . our toy model , eq .
( [ approx ] ) , tells us that the nlo correction is negligible for @xmath195 .
this means that the result in eq .
( [ res_ws ] ) is valid over a much larger rapidity range , up to @xmath196 , as compared to the result in ( [ dm_result ] ) which is inspired from the dipole model and which breaks down at @xmath197 .
therefore , the effect of saturation is to replace the result in eq .
( [ y_ly0 ] ) by the result in eq .
( [ y_sy0 ] ) even when @xmath198 .
it is interesting to note that the saturation effects are mathematically included by setting the lower limits in eq .
( [ dist_ns_int ] ) to @xmath17 as can be seen by comparing eq .
( dist_ns_int ) with eq .
( [ limit_0 ] ) without modifying the probability distribution.(in addition , one can easily prove that the @xmath2 matrix of dipole approach is boost invariant if and only if the lower limits in eq .
( [ dist_ns_int ] ) are set to @xmath17 . )
we have investigated a stochastic toy model without transverse dimensions ( equivalently , a infinite dimension hierarchy of evolution equation ) which naturally generates pomeron loops .
we have computed the pomeron loop diagrams to the nlo using two different methods .
we find that the lo calculation agrees with the exisiting results in the literature mueller:1994gb , mueller:1996te .
the study of nlo graphs in this toy model generates the @xmath199 corrections .
observations made by investigating mueller s toy model @xcite have been shown numerically by salam @xcite to be valid also in the four - dimensional qcd .
presumably our toy model reveals properties of the four - dimensional qcd evolution to some extent .
we believe that the nlo correction in the four - dimensional qcd would be of order @xmath23 as well ( here we define @xmath200 to be the bfkl pomeron intercept ) .
the behavior of our toy model is heuristically indicating that nlo contributions can be neglected in the four - dimensional qcd evolution as long as @xmath201 as well .
this believe is strengthened even more by noticing that the nlo correction @xmath202 seem to naturally appear in qcd : there are two types of nlo contributions .
* the first type of nlo contribution comes from nlo contribution of lo graphs(see example fig .
[ reggeon ] b ) . for the one - loop diagram , for instance
, the lo contribution in qcd is @xmath203 which comes from varying the size of the loop from @xmath17 to @xmath18 .
the nlo contribution @xmath204 comes from changing the location of a fixed zero - size loop from @xmath21 to @xmath18 .
the nlo contribution is suppressed by the factor @xmath22 with respect to the lo contribution , as in the toy model ( see eq .
( [ lo_loop ] ) ) , leading to the general form @xmath11 . *
the second type of nlo contribution comes from lo contribution of nlo graphs(see example fig .
[ reggeon ] c ) .
for the two - loop diagram fig .
[ reggeon ] c , for instance , the lo contribution @xmath205 comes from shifting the location of the fixed zero - size connecting pomeron between two loops from @xmath21 to @xmath18 while the sizes of these two loops are changed in the same time . since the total amplitude of this graph does not depend on the location of the connecting pomeron
, one can obtain an extra factor of @xmath206 by integrating from @xmath21 to @xmath18 . the nlo contribution ( see eq .
( [ nloloop ] ) ) is suppressed by the factor @xmath207 with respect to the lo contribution of fig .
[ reggeon ] b , as in the toy model ( see eq .
( [ lo_loop ] ) ) , leading to the general form @xmath23 .
therefore , we can see that the nlo correction generally takes the form of @xmath23 . a qcd model in which all pomeron splittings occur before all pomeron mergings in a boost - invariant manner ( as in this toy model ) would lead to the right lo result which would be valid up to the limit @xmath208 .
we would like to thank alfred mueller for numerous stimulating and insightful discussions .
we are also grateful to stephane munier for many useful remarks and comments .
a. sh . acknowledges financial support by the deutsche forschungsgemeinschaft under contract sh 92/2 - 1 .
in this part , we show that eq .
( [ h1 ] ) gives lorentz invariant results . to show this ,
let us start with a single particle , then split it to three particles which yields @xmath209 according to eq .
( [ split ] ) .
following kovchegov @xcite , the scattering amplitude is @xmath210 in which @xmath211 @xmath212 and the @xmath213 can be understood as the fact that those @xmath31-particles in the target are the same .
therefore , it is very easy to see that @xmath214 which corresponds to the diagram shown in fig .
splitmerge a. on the other hand , we can put the evolution into the target , let them merge into one object before scattering with the projectile ( fig .
[ splitmerge ] b ) .
this means that one starts with an initial condition @xmath215 and @xmath216 in solving the eq .
( [ h1 ] ) . after merging twice from @xmath217 ,
one obtains @xmath218 . finally , it is straight forward to see that @xmath219 after setting @xmath220 . presumably , @xmath221 should be same as @xmath222 since they actually describe the same process , and they are indeed the same once we set @xmath223 . moreover , it is tempting to replace the target by a large nucleus with @xmath224 nucleons inside it
. then , we need to define @xmath211 @xmath225 , and indeed one can see that @xmath226 as long as we set @xmath227 .
we notice that there has been a lot of theoretical discussion and numerical simulation on a slightly different stochastic differential equation which has a noise term @xmath228 , namely the stochastic fisher - kolmogorov - petrovsky - piscounov(sfkpp ) equation . for completeness
, we provide the solution to this equation by using the @xmath229representation we developed above .
the dynamics of this equation is given by two random processes : by an increase @xmath230 in rapidity , a particle can split into two particles with some rate @xmath231 ( @xmath232 ) or two particles can merge into one with a rate @xmath233 ( @xmath234 ) . from the equation , @xmath235one can get@xmath236similarly , one can write eq .
( [ hierarchy2 ] ) into a matrix in the @xmath237-representation , then get the similar @xmath2-matrix of the scattering between two dipoles , where @xmath239 and @xmath240
. the difference of between this result and the one in eq .
( [ exact2 ] ) is at the nlo correction , however , one can see that eq .
( [ langevin ] ) has a more transparent physical meaning in qcd
since it contains the projectile - target duality and one can easily relate it to the zero transverse dimension reggeon theory . 99 a. h. mueller , nucl .
b * 437 * ( 1995 ) 107 .
a. h. mueller and g. p. salam , nucl .
b * 475 * , 293 ( 1996 ) .
a. h. mueller and a. i. shoshi , nucl .
b * 692 * ( 2004 ) 175 .
e. iancu , a. h. mueller and s. munier , phys .
b * 606 * ( 2005 ) 342 .
s. munier and r. peschanski , phys .
* 91 * ( 2003 ) 232001 ; phys .
d * 69 * ( 2004 ) 034008 .
e. iancu and d. n. triantafyllopoulos , nucl .
a * 756 * ( 2005 ) 419 .
i. balitsky , nucl .
b * 463 * ( 1996 ) 99 ; phys .
* 81 * ( 1998 ) 2024 ; phys .
b * 518 * ( 2001 ) 235 .
j. jalilian - marian , a. kovner , a. leonidov and h. weigert , nucl .
b * 504 * ( 1997 ) 415 ; phys .
d * 59 * ( 1999 ) 014014:e .
iancu , a. leonidov and l. d. mclerran , phys .
b * 510 * ( 2001 ) 133 ; nucl .
a * 692 * ( 2001 ) 583:h .
weigert , nucl .
a * 703 * ( 2002 ) 823 .
a. h. mueller , a. i. shoshi and s. m. h. wong , nucl .
b * 715 * ( 2005 ) 440 .
e. iancu and d. n. triantafyllopoulos , phys .
b * 610 * ( 2005 ) 253 .
a. kovner and m. lublinsky , phys .
d * 71 * ( 2005 ) 085004 .
y. hatta , e. iancu , l. mclerran , a. stasto and d. n. triantafyllopoulos , arxiv : hep - ph/0504182 .
a. kovner and m. lublinsky , phys .
* 94 * ( 2005 ) 181603 ; jhep * 0503 * ( 2005 ) 001 ; phys .
d * 72 * ( 2005 ) 074023 ; `` more remarks on high energy evolution , '' arxiv : hep - ph/0510047 .
j. p. blaizot , e. iancu , k. itakura and d. n. triantafyllopoulos , phys .
b * 615 * ( 2005 ) 221.y .
hatta , e. iancu , l. mclerran and a. stasto , nucl .
phys . a * 762 * ( 2005 ) 272.e .
iancu , g. soyez and d. n. triantafyllopoulos , on the probabilistic interpretation of the evolution equations with pomeron arxiv : hep - ph/0510094 . c. marquet , a. h. mueller , a. i. shoshi and s. m. h. wong , nucl .
a * 762 * ( 2005 ) 252 .
e. levin and m. lublinsky , `` towards a symmetric approach to high energy evolution : generating functional with pomeron loops , '' arxiv : hep - ph/0501173 .
e. levin , nucl .
a * 763 * , 140 ( 2005 ) .
r. enberg , k. golec - biernat and s. munier , phys .
d * 72 * ( 2005 ) 074021 .
g. soyez , phys .
d * 72 * ( 2005 ) 016007 .
a. h. mueller , nucl .
b * 415 * ( 1994 ) 373 ; nucl .
b * 425 * ( 1994 ) 471 .
g. p. salam , nucl .
b * 461 * ( 1996 ) 512 ; g. p. salam , nucl .
b * 449 * ( 1995 ) 589 .
y. v. kovchegov , phys .
d * 72 * ( 2005 ) 094009 .
hardy , divergent series(oxford university , oxford , 1963 ) . | we analyze a toy model which has a structure similar to that of the recently found qcd evolution equations , but without transverse dimensions .
we develop two different but equivalent methods in order to compute the leading - order and next - to - leading order pomeron loop diagrams .
in addition to the leading - order result which has been derived within other toy models mueller:1994gb , mueller:1996te , we can also calculate the next - to - leading order contribution which provides the @xmath0 correction .
we interpret this result and discuss its possible implications for the four - dimensional qcd evolution . |
simulations of nano scale liquid systems can provide insight into many naturally - occurring phenomena , such as the action of proteins that mediate water transport across biological cell membranes @xcite .
they may also facilitate the design of future nano devices and materials ( e.g. high - throughput , highly selective filters or lab - on - a - chip components ) .
the dynamics of these very small systems are dominated by surface interactions , due to their large surface area to volume ratios
. however , these surface effects are often too complex and material - dependent to be treated by simple phenomenological parameters @xcite or by adding ` equivalent ' fluxes at the boundary @xcite .
direct simulation of the fluid using molecular dynamics ( md ) presents an opportunity to model these phenomena with minimal simplifying assumptions .
some md fluid dynamics simulations have been reported @xcite , but md is prohibitively computationally costly for simulations of systems beyond a few tens of nanometers in size .
fortunately , the molecular detail of the full flow - field that md simulations provide is often unnecessary ; in liquids , beyond 510 molecular diameters ( @xmath0 for water ) from a solid surface the continuum - fluid approximation is valid and the navier - stokes equations with bulk fluid properties may be used @xcite .
hybrid simulations have been proposed @xcite to simultaneously take advantage of the accuracy and detail provided by md in the regions that require it , and the computational speed of continuum mechanics in the regions where it is applicable .
an example application of this technique is shown schematically in figure [ figure_postsandnanochannel ] , where a complex molecule is being electrokinetically transported into a nanochannel for separation and identification@xcite . only the complex molecule ,
its immediately surrounding solvent molecules , and selected near - wall regions require an md treatment ; the remainder of the fluid ( comprising the vast majority of the volume ) may be simulated by continuum mechanics .
a hybrid simulation would allow the effect of different complex molecules , solvent electrolyte composition , channel geometry , surface coatings and electric field strengths to be analysed , at a realistic computational cost . in order to produce a useful , general simulation tool for hybrid simulations ,
the md component must be able to model complex geometrical domains .
this capability does not exist in currently available md codes : domains are simple shapes , usually with periodic boundaries .
the most important , computationally demanding , and difficult aspect of any md simulation is the calculation of intermolecular forces .
this paper describes an algorithm that is capable of calculating pair force interactions in arbitrary , unstructured mesh geometries that have been parallelised for distributed computing by spatial decomposition .
the conventional method of md force evaluation in distributed parallel computation is to use the cells algorithm to build ` neighbour lists ' for interacting pairs with the ` replicated molecule ' method of providing interactions across periodic boundaries and interprocessor boundaries , where the system has been spatially partitioned @xcite . the spatial location of molecules in md is dynamic , and hence not deducible from the data structure that contains them .
a neighbour list defines which pairs of molecules are within a certain distance of each other , and as such need to interact via intermolecular forces .
_ neighbour lists are unsuitable _ when considering systems of arbitrary geometry , that may have been divided into irregular and complex mesh segments using standard mesh partitioning techniques ( see , for example , figure [ figure_testmolconfig_complex_bw ] ) for two reasons : 1 .
* interprocessor molecule transfers : * a molecule may cross an interprocessor boundary at any point in time ( even part of the way through a timestep ) , at which point it should be deleted from the processor it was on and an equivalent molecule created on the processor on the other side of the boundary . given that neighbour lists are constructed as lists of array indices , references or pointers to the molecule s location in a data structure , deleting a molecule would invalidate this location and require searching to remove all mentions of it .
likewise , creating a molecule would require the appropriate new pair interactions to be identified .
clearly this is not practical .
it is conventional to allow molecules to ` stray ' outside of the domain controlled by a processor and carry out interprocessor transfers ( deletions and creations ) during the next neighbour list rebuild .
this is only possible when the spatial region associated with a processor can be simply defined by a function relating a position in space to a particular cell on a particular processor ( i.e. a uniform , structured mesh , representing a simple domain ) . in a geometry where the space in question is defined by a collection of individual cells of arbitrary shape ,
this is not possible .
for example the location the molecule has ` strayed ' to may be on the other side of a solid wall on the neighbour processor , or across another interprocessor boundary . for the molecule to end up unambiguously in the correct place
, interprocessor transfers must happen as molecules cross a face .
* spatially resolved flow properties : * md simulations used for flow studies must be able to spatially resolve fluid mechanic and thermodynamic fields .
this is achieved by accumulating and averaging measurements of the properties of molecules in individual cells of the same mesh that defines the geometry .
if a molecule is allowed to stray outside of the domain controlled by a processor , as above , then it would not be unambiguous and automatic which cell s measurement the molecule should contribute to .
while both of these problems could be mitigated by , for example , working out which cell a molecule outside the domain should be in on another processor and sending its information , doing so would result in an inelegant and inflexible arrangement with each additional simulation feature ( i.e. a different measured variable or class of intermolecular potential ) requiring special treatment .
we have , therefore , developed a new algorithm which is of comparable computational cost to neighbour lists , but designed to be powerful and generic for simulation in arbitrary geometries .
as will be shown , neighbour lists also have some unfavourable features that could be improved upon .
when parallelising an md calculation , the spatial domain of the simulation is decomposed and each processor is given responsibility for a single region @xcite .
molecules are , of course , able to cross the boundaries between these regions and therefore need to be communicated from one processor to the next
. processors also communicate when carrying out intermolecular force calculations , in which molecules close to processor boundaries need to be replicated on their neighbours to provide interactions .
this process is illustrated in figure [ figure_referredmolecules ] .
periodic boundaries also require information about molecules that are not adjacent physically in the domain ( see figure [ figure_referredmoleculesperiodic ] ) , these required interactions can also be constructed by creating copies of molecules outside the boundary .
it is possible to handle processor and periodic boundaries in exactly the same way , since they have the same underlying objective : molecules near to the edge of a region need to be copied either between processors or to other locations on the same processor at every timestep to provide interactions .
this is a useful feature because decomposing a mesh for parallelisation will often turn a periodic boundary into a processor boundary .
the issue is how to efficiently identify which molecules need to be copied , and to which location , because this set continually changes as the molecules move .
the new arbitrary interacting cells algorithm ( aica ) we propose here is an extension of the conventional cells algorithm ( cca ) @xcite . in the cca ,
a simple ( usually cuboid ) simulation domain is subdivided into equally sized cells . for computational and theoretical reasons
, intermolecular potentials do not extend to infinity ; they are assigned a cut - off radius , @xmath1 , beyond which they are set to zero .
the minimum dimension of the cca cells must be greater than @xmath1 , so that all molecules in a particular cell interact with all other molecules in their own cell and with those in their nearest neighbour cells ( i.e. those they share a face , edge or vertex with 26 in 3d ) .
aica uses the same type of mesh as would be used in computational fluid dynamics ( cfd ) to define the geometry of a region .
_ there are no restrictions on cell size , shape or connectivity .
_ each cell has a unique list of other cells that it is to interact with , this list is known as the direct interaction list ( dil ) for the cell in question ( ciq ) .
it is constructed by searching the mesh to create a set of cells that have at least one vertex within a distance of @xmath1 from any of the vertices of the ciq , see figure [ figure_interactingcellidentification ] .
the dils are established prior to the start of simulation and are valid throughout because the spatial relationship of the cells is fixed , whereas the set of molecules they contain is dynamic . in a similar way to the cca , at every timestep
a molecule in a particular cell calculates its interactions with the other molecules in that cell and consults the cell s dil to find which other cells contain molecules it should interact with .
information is required to be maintained stating which cell a molecule is in this is straightforward and computationally cheap .
the construction of the dils and accessing molecules is done in such a way as to not double - count interactions , similarly to the cca .
for example , if cells a and b need to interact , cell b will be on cell a s dil , but cell a will not be on cell b s dil .
when a molecule in cell a interacts with one in cell b the molecule in cell b will receive the inverse of the force vector calculated to be added to the molecule in cell a , because the pair forces are reciprocal , i.e. @xmath2 .
it is possible in rare cases for small errors to be caused by this algorithm : small slices of volume in cells may not be identified for interaction ( see figure [ figure_errorsintwistedmeshes ] ) .
the errors introduced will be slight because the relative volume not accounted for will be very small : in figure [ figure_errorsintwistedmeshes ] , only when molecules are in both the indicated regions of each cell will either miss out on an interaction , and the intermolecular potential will be small at this distance . a small guard distance could be added to @xmath1 when constructing the dil to reduce this error . this is less of a problem in good quality meshes , e.g. hexahedral rather than tetrahedral cells . replicated molecule parallelisation and periodic boundaries are handled in the same way using _ referred cells _ ( see figure [ figure_interactingcellidentification ] ) and _ referred molecules_. referred molecule : : : a copy of a real molecule that has been placed in a region outside a periodic or processor boundary in order to provide the correct intermolecular interaction with molecules inside the domain .
a referred molecule holds only its own position and i d ( i.e. identification of which type of molecule it is for heterogenous simulations ) .
referred molecules are created and discarded at each timestep , and do not report any information back to their source molecules . therefore if molecule _
j _ on processor 1 needs to interact with molecule _
k _ on processor 2 , a separate referred molecule will be created on each processor .
referred cell : : : referred cells define a region of space and hold a collection of referred molecules .
each referred cell knows + * which real cell in the mesh ( on which processor ) is its source ; * the required transformation to refer the positions and orientations of the real molecules in the source cell to the referred location ( see below for the details of how cells and molecules are referred ) ; * the positions of all of its own vertices .
these are the positions of the vertices of the source cell which have been transformed by the same referring transform as the referred molecules it contains ; * which real cells are in range of this particular referred cell and hence require intermolecular interactions to be calculated .
this is constructed once at the start of the simulation in the same way as the dil for real - real cell interactions the vertices of the real cells in the portion of mesh on the same processor as the referred cell are searched , those with at least one vertex in range of any vertex of the referred cell are found .
a spatial transformation is required to refer a cell across a periodic or processor boundary .
figure [ figure_cellreferringtransform ] shows the most general case of a cell being referred across a separated , non - parallel boundary , where , [ cols= " > , < " , ] all cell combinations have been created once , i.e. when @xmath3 , @xmath4 would produce the 21 combination , which is already identified as 12 .
this non double counting procedure means that some cells will have few or no other cells in their dil , however , as long as the correct cell pair is created in one of the cell s dil , this is not an issue .
it is also designed so that a cell does not end up on its own dil , the interactions between molecules in the same cell are handled separately . _ coupled patches _ are the basis of periodic and interprocessor communication for creating referred cells .
patches , in general , are a collection of cell faces representing a mesh boundary of some description they may provide solid surfaces , inlets , outlets , symmetry planes , periodic planes , or interprocessor connections . coupled patches provide two surfaces ; whatever exits one enters the other and vice versa .
two types are used in aica : * _ periodic patches _ on a single processor are arranged into two halves , each half representing one of the coupled periodic surfaces . when a molecule crosses a face on one surface , it is wrapped around to the corresponding position on the corresponding face on the other surface .
* _ processor patches _ provide links between portions of the mesh on different processors , one half of the patch is on each processor . when a molecule crosses a face on one surface , it is moved to the corresponding position on the corresponding face on the other surface , on the other processor . decomposing a mesh for parallelisation will often require a periodic patch to be changed to a processor patch
the surfaces of coupled patches may have any relative orientation , and may be spatially separated as long as the face pairs on each surface correspond to each other .
this allows them to also be used as a ` recycling ' surface in a flow simulation whatever exits from an outlet is reintroduced at an inlet , with the option of scaling the molecule s properties , e.g. temperature and pressure . in the decomposed portion of the mesh on each processor
, each processor and periodic patch should be split into segments , such that : * faces on a processor that were internal to the mesh prior to decomposition end up on a segment ; * faces on a processor patch that were on separate periodic patches on the undecomposed mesh end up on different segments .
these segments are further split such that faces that were on different halves of the periodic patch on the undecomposed mesh end up on different segments ; * faces on different halves of a periodic patch end up on different segments .
each segment must produce a single vector / tensor transformation pair ( see section [ text_cellpositionreferringtransformations ] ) which will be applied to all cells referred across it . for each patch segment
: 1 . find all real and existing referred cells in range ( with at least one vertex within a distance of @xmath5 ) of at least one vertex of the faces comprising the patch segment .
2 . refer or re - refer this set of cells across the boundary defined by the patch segment using its transformation , see section [ text_cellpositionreferringtransformations ] . in the case of a segment of periodic boundary , this creates new referred cells on the same processor . for a segment of a processor patch ,
these cells are communicated to , and created on the neighbouring processor . before creating any new referred cell a check
is carried out to ensure that it is not * a duplicate referred cell , one that has been created already by being referred across a different segment ; * a referred cell trying to be duplicated on top of a real cell , i.e. a cell being referred back on top of itself . + if the proposed cell for referral would create a duplicate of an existing referred cell , or end up on top of a real cell , then it is not created . to be a duplicate , the source processor , source cell and the vector part of the transformation ( see section [ text_cellpositionreferringtransformations ] )
must be the same for the two cells ( note , the vector part of transformation for a real cell is zero by definition ) .
a single run of these steps will usually not produce all of the required referred cells .
they are repeated until no processor adds a referred cell in a complete evaluation of all segments , meaning all possible interactions are accounted for . in iterations after the first , in step 1 it is enough to only search for referred cells in range of the faces on the patch segment , because the real cells will not have changed , and would all be duplicates .
the final configuration of referred cells does not depend on the order of patch segment evaluation .
appendix [ text_exampleconstructionofreferredcells ] contains an example of the results of the cell referring process .
figures [ figure_parallelgridpolyhedra1step_bw ] and [ figure_parallelgridpolyhedra2step_bw ] show the start and end points of the cell referring operation on a mesh that has been decomposed to run in parallel .
the example is in 2d for clarity but the process is exactly the same in 3d . in this example
the number of referred cells created far exceeds the number of real cells , which would lead to much costly interprocessor communication .
this is because the mesh has been made deliberately small relative to @xmath5 to demonstrate as many features of the algorithm as possible and to be practical to understand . in realistic systems
, the mesh portions would be significantly bigger than @xmath5 and the referred cells would form a relatively thin ` skin ' around each portion .
decomposition of the mesh should preferably be carried out to minimise the number of cells that need to be referred , and to ensure that the vast majority of the intermolecular interactions happen between real - real molecule pairs ; in this way the communication cost is minimised . .
some of those molecules may be on a periodic image of the system , and as such , in implementation terms , will reside on the other side of the domain .
serial calculations in simple geometries typically use the minimum image convention @xcite , but this is not suitable for parallelisation.,width=272 ] that interact with the ciq ( dark ) are shaded in grey .
the required referred cells are hatched in alternate directions according to which boundary they have been referred across .
realistic systems would be significantly larger compared to @xmath1 than shown here.,width=264 ] drawn from the indicated vertex on cell a intersects a face of cell b , therefore , molecules near this vertex should interact with molecules in the shaded region .
similarly , an arc of radius @xmath1 drawn from the indicated face on cell b encompasses the small shaded region of cell a. the cell searching algorithm will not , however , identify cells a and b as needing to interact because none of their vertices lie within @xmath1 of each other.,width=264 ] on each processor . a circle of radius @xmath5 drawn from any vertex of a real cell
would be fully encompassed by other real or referred cells , thereby providing all molecules in that cell with the appropriate intermolecular interactions , either across a periodic boundary or from another processor .
note that many cells are referred several times ; for example , the source cell marked with @xmath6 on processor 2 is referred to processors 0 and 1 twice on each.,width=529 ] | a new algorithm for calculating intermolecular pair forces in molecular dynamics ( md ) simulations on a distributed parallel computer is presented . the arbitrary interacting cells algorithm ( aica )
is designed to operate on geometrical domains defined by an unstructured , arbitrary polyhedral mesh , which has been spatially decomposed into irregular portions for parallelisation .
it is intended for nano scale fluid mechanics simulation by md in complex geometries , and to provide the md component of a hybrid md / continuum simulation .
aica has been implemented in the open - source computational toolbox openfoam , and verified against a published md code .
and molecular dynamics , nano fluidics , hybrid simulation , intermolecular force calculation , parallel computing 31.15.qg , 47.11.mn |
while in the soft band ( 0.5 - 2 kev ) _ rosat _ ( @xcite ) and especially _ chandra _ ( @xcite ) has resolved almost all the xrb , in the hard band ( 2 - 10 kev ) the xrb has been resolved at a 25%-30% level with _ bepposax _ and _ asca _ surveys ( @xcite ; @xcite ) and recently at a 90% with _ chandra _ ( @xcite ) . moreover , in the very hard band ( 5 - 10 kev ) the fraction resolved by _ bepposax _ is around 30% ( @xcite ) and recently in the _ xmm - newton _ lockman hole deep pointing a 60% is reached ( @xcite ) . + the optical counterparts of the objects making the xrb are predominantly active galactic nuclei ( agn ) . in the soft band
the predominant fraction is made by unabsorbed agn , with a small fraction of absorbed agn ( @xcite ) .
the fraction of absorbed type-2 agn rises if we consider the spectroscopic identifications of hard x - ray sources in _ bepposax _ , _ asca _ and _ chandra _ surveys ( @xcite ; @xcite ; @xcite ) . + the x - ray and optical observations are consistent with current xrb synthesis models ( @xcite ; @xcite ) , which explain the hard xrb spectrum with an appropriate mixture of absorbed and unabsorbed agn , by introducing the corresponding luminosity function and cosmological evolution .
however , these models require the presence of a significant population of type-2 qsos ( @xcite ) , not yet detected in sufficient quantities .
type-2 qsos are rare ( so far , only a few are known ) , luminous and hard ( heavily absorbed in the soft band ) .
a good way of finding them is to perform surveys in the hard x - ray bands , covering large solid angles .
the large throughput and effective area , particularly in the harder bands , make _ xmm - newton _ currently the best satellite to perform hard x - ray surveys .
+ in this poster contribution we present results from the hellas2xmm survey ( @xcite ) , one of its main goals is to constrain the contribution of absorbed agn to the xrb .
we use the xmm - sas analysis software tasks @xmath4 and @xmath5 to linearize the event files . before processing
, datasets are corrected for the attitude of the satellite in order to have absolute positions in the sky .
the attitude history file(ahf ) coordinates are given to the sas task @xmath6 which performs the correction .
+ the event files produced by @xmath4 and @xmath5 are cleaned from : * hot pixels ; with a procedure ( developed at ifc / cnr - milan by a. de luca ) which uses cosmic ray iraf tasks to localize the pixels to be rejected in each ccd and xmm - sas task @xmath7 for removing them from the event files ; * soft proton flares ; analysing the light curves at energies greater than 10kev and setting a threshold for good time intervals of 0.15 cts / s for mos units and of 0.35 cts / s for pn unit . a complete set of mos1 , mos2 and pn 600x600 pixel images ( 1 pixel = 4.35 arcsec ) is generated using xmm - sas task @xmath7 in the 0.5 - 2 , 2 - 10 , 2 - 4.5 , 4.5 - 10 and 0.5 - 10 kev bands .
mos and pn images are merged together in order to increase the signal - to - noise ratio of the sources and go deeper .
+ a corresponding set of exposure maps is generated to account for spatial quantum efficiency , mirror vignetting and field of view of each instrument , running xmm - sas task @xmath8 .
the so - created exposure maps are not completely satisfactory , since the evaluation of quantum efficiency , filter transmission , and vignetting is performed assuming an event energy which corresponds to the mean of the energy boundaries . in the 2 - 10 kev band , this may lead to inaccuracies in the estimate of these key quantities , thus we create the 2 - 10 kev band exposure map as a weighted mean of the 2 - 4.5 kev and the 4.5 - 10 kev exposure maps ( assuming an underlying power - law spectral model with @xmath9 ) .
+ xmm - sas task @xmath10 creates a background map by removing all the sources above a fixed maximum likelihood threshold and fitting the remaining with a cubic spline . even using the maximum number of spline nodes ( 20 ) ,
the fit is not enough flexible to take into account the local variations of the background .
we correct the background map produced by xmm - sas , pixel by pixel , comparing the mean value in it with that of the so - called cheesed image , within a radius of @xmath11 ( @xmath12 = radius corresponding to an encircled energy fraction ( eef ) of 0.68 ) .
a comparison between xmm - sas standard background map and our corrected map is shown in figure [ abaldi - f_fig : bkgmap ] .
a preliminary detection run , using xmm - sas @xmath13 , is performed in each energy band , in order to create a list of candidate sources .
each source is characterized using a radius corresponding to an eef of @xmath14=0.68 .
+ the source counts @xmath2 and their error @xmath15 are determined as @xmath16 @xmath17 the countrate is @xmath18 where @xmath19 is the sum of mos1 , mos2 and pn exposure times .
the corresponding flux is @xmath20 where @xmath21 is calculated from the @xmath22 moreover we compute @xmath23 , the poissonian probability that counts originate from a background fluctuation , from the @xmath24 and choose a threshold of @xmath25 to decide whether to accept or not a detected source .
the hellas2xmm survey ( @xcite ) currently uses the 15 _ xmm - newton _ calibration and performance verification phase fields shown in table [ abaldi - f_tab:15fields ] .
all the fields are at high galactic latitude ( @xmath26 @xmath27 27@xmath28 ) , have low galactic @xmath29 ( a few @xmath30 @xmath31 ) and at least 15 ksec of good observing time .
lrr + target&n@xmath32 ( @xmath31)&@xmath33(@xmath28 ) + + pks0537 - 286&@xmath34&-27.3 + pks0312 - 770&@xmath35&-37.6 + ms0737.9 + 7441&@xmath36&29.6 + lockman hole&@xmath37&53.1 + mkn 205&@xmath38&41.7 + bpm 16274&@xmath39&-65.0 + ms1229.2 + 6430&@xmath40&52.8 + pks0558 - 504&@xmath41&-28.6 + mkn 421&@xmath42&65.0 + abell 2690&@xmath43&-78.4 + g158 - 100&@xmath44&-74.5 + gd153&@xmath45&84.7 + iras13349 + 2438&@xmath46&60.6 + abell 1835&@xmath47&60.6 + mkn 509&@xmath48&-29.9 + + the sky coverage of the sample has been computed using the exposure maps of each instrument , the background map of the merged image and a model for the psf .
we adopt the off - axis angle dependent psf model implemented in _ xmm - sas _ @xmath13 task .
+ at each image pixel @xmath49 we evaluate , within a radius @xmath12 , the total background counts ( from the background map ) . from these
we calculate the minimum total counts ( source + background ) necessary for a source to be detected at a probability @xmath25 ( defined in section [ abaldi - f_sec : srcdet ] ) . the mean exposure times for _ mos1 _ , _ mos2 _ and _ pn _ , evaluated from the exposure maps within @xmath12 ,
are used to compute the count rate @xmath50 . from the count rate - to - flux conversion factor @xmath21 ( computed as in section [ abaldi - f_sec : srcdet ] ) we build a flux limit map and straightforwardly calculate the sky coverage of a single field .
+ summing the contribution from all fields we obtain the total sky coverage of the survey , which is plotted in figure [ abaldi - f_fig : skycov ] , in three different energy bands . +
the @xmath0 @xmath1 - @xmath0 @xmath2 distributions are plotted in figure [ abaldi - f_fig : lognlogs ] and contain 1022 , 495 and 100 sources , for the 0.5 - 2 kev , 2 - 10 kev and 5 - 10 kev band , respectively .
+ in the 0.5 - 2 kev band the distribution shows a flattening around @xmath3 erg @xmath31 s@xmath51 , similarly to _ rosat
_ data ( @xcite ) although with a flatter differential slope index at faint fluxes .
we are also in good agreement with _
chandra _ deep field south data ( @xcite ) .
in the 2 - 10 kev band we find that the distribution is significantly sub - euclidean , in contrast to _ bepposax _ and _ asca _ findings ( @xcite ; @xcite ; @xcite ) , indicating that the @xmath0 @xmath1 - @xmath0 @xmath2 flattens at fainter fluxes .
anyhow it represents a link between _ chandra _
( @xcite ) and _ bepposax - asca _ observation , sampling an intermediate flux range .
the 5 - 10 kev @xmath0 @xmath1 - @xmath0 @xmath2 is consistent with an euclidean slope and samples an intermediate flux range between _ xmm - newton _ deeper observations ( @xcite ) and _ bepposax _ shallower hellas survey ( @xcite ) .
we divided the sample of sources detected both in 0.5 - 2 kev band and in 2 - 4.5 kev band in two subsamples containing the brighter ( @xmath52 erg @xmath31 s@xmath51 ) and the fainter ( @xmath53 erg @xmath31 s@xmath51 ) sources , respectively and we compute for them the hardness ratio : @xmath54 lcc + & n(@xmath55)&n(@xmath56 ) + + faint sample&654&84 + bright sample&107&0 + + as shown in figure [ abaldi - f_fig : hr1 ] and in table [ abaldi - f_tab : hr1distrib ] , the faint sample shows a tail of hard sources which is not present in the bright sample .
the probability of having 84 sources with @xmath56 in the faint sample and no sources with @xmath56 in the bright sample is @xmath57 , so the progressive hardening of the sources towards fainter fluxes seems to be highly significant .
+ a further analysis of the hardness ratio has been carried out on the 4.5 - 10 kev sample . in figure [ abaldi - f_fig : hr2 ]
the relation between the 2 - 10 kev flux and the hardness ratio @xmath58 for sources detected in the 4.5 - 10 kev band is plotted .
the diagram shows a progressive hardening of the sources when going towards fainter fluxes .
while the lack of sources in the bottom left of the figure is due to a selection effect ( because we are more sensitive in the 2 - 4.5 kev than in the 4.5 - 10 kev band ) , there is no apparent reason for having no hard sources at fluxes greater than a few @xmath59 erg @xmath31 s@xmath51 .
+ the population of harder sources we detect , probably consists of agn having substantial absorbing column densities ( @xmath60 @xmath31 ) .
we are carrying out a serendipitous _ xmm - newton _ survey .
we currently cover nearly three square degrees in 15 fields observed during satellite calibration and performance verification phase .
this is , to date , the _ xmm - newton _ survey with the largest solid angle . + the main results can be summarized as follows : * the @xmath0 @xmath1-@xmath0 @xmath2 relations in the 0.5 - 2 kev , 2 - 10 kev and 5 - 10 kev band are in agreement with previous determinations ; * in the hard bands we sample an intermediate flux range : deeper than _ asca _ and _ bepposax _ and shallower than _ chandra _ and _ xmm - newton _ deep surveys * we find an evidence for hard sources emerging below 0.5 - 2 kev fluxes of @xmath61 erg @xmath31 s@xmath51 and 2 - 10 kev fluxes of 10@xmath62 erg @xmath31 s@xmath51 .
we thank a. de luca for developing the hot pixel cleaning algorithm .
we are also grateful to g. zamorani , g. c. perola and all members of the hellas2xmm team for useful discussions .
ab and sm acknowledge partial financial support by asi i / r/190/00 contract . | we present the first results from an _ xmm - newton _ serendipitous medium - deep survey , which covers nearly three square degrees .
we show the @xmath0 @xmath1 - @xmath0 @xmath2 distributions for the 0.5 - 2 , 2 - 10 and 5 - 10 kev bands , which are found to be in good agreement with previous determinations and with the predictions of xrb models . in the soft band
we detect a break at fluxes around @xmath3 cgs . in the harder bands , we fill in the gap at intermediate fluxes between deeper _ chandra _ and _ xmm - newton _ observations and shallower _ bepposax _ and _ asca _ surveys .
moreover , we present an analysis of the broad - band properties of the sources . |
there has been much recent interest in theories of gravity arising from scenarios with extra spatial dimensions .
many examples of these are based on the dvali - gabadadze - porrati ( dgp ) model @xcite a @xmath3 dimensional theory with action consisting simply of separate einstein - hilbert terms in the bulk and on a codimension-1 brane , to which standard model particles are also confined .
the model results in a 4d gravitational force law at sufficiently small scales , which transitions to a 5d gravitational force law at a crossover length scale @xmath4 , determined by the @xmath5d and @xmath6d gravitational couplings @xmath7 and @xmath8 respectively . to yield interesting cosmological dynamics ,
this crossover scale is usually chosen to be of order the horizon size .
much of the phenomenology of the dgp model is captured by its decoupling limit @xmath8 , @xmath9 with the strong - coupling scale @xmath10 kept fixed @xcite . in this limit
, the difference between dgp gravity and general relativity is encoded in the behavior of a scalar degree of freedom , @xmath11 .
the dynamics of this scalar are invariant under internal galilean transformations @xmath12 , with @xmath13 a constant and @xmath14 a constant vector .
this symmetry proves to be extremely restrictive , with a leading order self - interaction term which is a higher - derivative coupling cubic in @xmath11 , and yet which yields second order equations of motion .
higher order couplings with these properties were derived independently of the dgp model @xcite and dubbed galileons . "
see @xcite for cosmological studies of galileon theories .
it is natural to explore induced gravity models in co - dimension greater than one @xcite , and recently multi - galileon actions arising in the relevant @xmath6-dimensional decoupling limit have been derived @xcite .
the theories studied in @xcite are invariant under individual galilean transformations of the @xmath11 fields , and also under an internal @xmath2 symmetry rotating the fields into one another , thus forbidding the existence of terms containing an odd number of @xmath11 fields , in contrast to the co - dimension one dgp case . in this paper
we explore the nature of spherically symmetric solutions in theories with an @xmath2 internal symmetry among the galileon fields , and couplings to matter that respect this symmetry .
spherical solutions for a more general bi - galileon action were discussed in @xcite , for the specific case of a linear coupling @xmath15 to matter , where @xmath16 is the trace of the matter energy momentum tensor .
this form of coupling arises from decoupling limits of dgp - like theories , because @xmath11 arises through a conformal mixing with the graviton .
however , while this coupling is therefore the natural form to consider in the case of a single galileon field , it breaks the new internal symmetry satisfied by multiple galileons ( and breaks the galilean symmetry if the matter is dynamical ) .
we instead study general non - derivative couplings to matter fields which respect the @xmath2 internal symmetry . at the background level
, our solution can always be rotated to lie along a single field direction , say @xmath17 , while the other field variables remain trivial , thus exhibiting spontaneous symmetry breaking .
the solution exhibits vainshtein screening @xcite , characteristic of galileon theories : we find @xmath18 sufficiently close to the source , whereas @xmath19 far away , with the crossover scale determined by a combination of the galileon self - interaction scale and the coupling to the source .
however , when we turn to the stability of spherically symmetric solutions under small perturbations , we find that , sufficiently close to the source , perturbations in @xmath17 suffer from gradient instabilities along the angular directions .
morever , they propagate superluminally both along the radial and angular directions ( in the regime that angular perturbations are stable ) . perturbations in the remaining @xmath20 galileon fields are stable but propagate superluminally in the radial direction .
the gradient instability and superluminal propagation found here for the @xmath17 field are multi - field generalizations of single galileon instabilities @xcite .
our findings thus present significant hurdles for @xmath2 galileon models with non - derivative matter coupling .
one of the main lessons to be drawn is that more general matter couplings , including derivative interactions , are necessary for the phenomenological viability of @xmath2 multi - galileon theories .
for instance , the coupling @xmath21 naturally arises from brane - world constructions @xcite and maintains both the galilean and the internal rotation symmetries .
in co - dimension @xmath22 , the @xmath6-dimensional effective theory contains @xmath22 fields @xmath0 , @xmath23 , representing the @xmath22 brane - bending modes of the full @xmath24-dimensional theory . the extended symmetry of the vacuum lagrangian is ^i=^i _ x^+^i+^i _ j^j , [ multisymmetry ] where @xmath25 , @xmath26 and @xmath27 are constant transformation parameters .
( see @xcite for the geometric setup and origin of this symmetry ) .
this transformation consists of a galilean invariance acting on each of the @xmath0 fields , and an @xmath2 rotation symmetry under which @xmath28 transforms as a vector .
the unique four dimensional lagrangian density respecting this is @xcite _ = & & - 12_^i^_i + & & - , + where @xmath29 is a coupling with dimension [ mass]@xmath30 , containing the strong interaction mass scale .
the @xmath31 indices are raised and lowered with @xmath32 .
it remains to couple this theory to matter .
the natural coupling we might consider , the lowest dimension coupling that preserves the galilean and internal rotation symmetries , is @xmath21 .
this is the coupling that naturally arises from brane matter in the construction of @xcite .
however , for static non - relativistic sources @xmath33 , and since @xmath34 for static solutions there are no nontrivial spherically symmetric solutions with this coupling .
linear couplings @xmath35 arise naturally from dgp - like setups , since the @xmath11 s conformally mix with the graviton .
these lead to spherical solutions @xcite , but break the @xmath2 internal symmetry .
we therefore do not consider these couplings further , and instead concentrate on the most general non - derivative coupling that preserves the @xmath2 symmetry .
_ coupling = t2 p(^2 ) , [ couplings ] where @xmath36 is an arbitrary function of the invariant @xmath37 .
our focus is on the existence and viability of spherically symmetric solutions sourced by a delta function mass distribution t = -m ^3(r ) .
the equations of motion , including the coupling ( [ couplings ] ) , are ^i & - & + & & = mp(^2(0))^i(0)^3(r ) , where @xmath38 . restricting to spherically symmetric configurations
@xmath39 , this reduces to = mp(^2(0))^i(0)^3(r ) , where y^i , and @xmath40 .
note that , due to the shift symmetry of the lagrangian , the equations of motion of galileon fields always take the form of a total derivative .
thus we can integrate once to obtain the equations of motion [ finalequations ] y^i+2y^i y^2=m4r^3p(^2(0))^i(0 ) . dividing these equations by each other
, we obtain the relations = ^i(0)^j(0 ) , which , when integrated from the origin , gives = ^i(0)^j(0 ) .the various components of the solution are therefore always proportional to each other .
thus , by a global @xmath2 rotation , we can rotate the solution into one direction in field space , say the @xmath41 direction , so that the solution takes the form @xmath42 and @xmath43 for @xmath44 .
this model therefore exhibits a kind of spontaneous symmetry breaking of the internal @xmath2 symmetry , since any non - trivial solution must pick a direction in field space .
equation ( [ finalequations ] ) now takes the form y+2y^3=m4r^3p(^2(0))(0 ) .
[ otheryeqn ] as @xmath45 ranges from zero to infinity , the left hand side is monotonic , and is positive or negative depending on the sign of @xmath46 .
for there to be a continuous solution for @xmath47 as a function of @xmath45 , the left hand side must be invertible when it is positive ( negative ) . for a solution to exist ,
this requires ( for non - trivial @xmath29 ) > 0 .
thus @xmath47 is also positive ( negative ) , is monotonic with @xmath45 , and ranges from zero to ( negative ) infinity as @xmath45 ranges from infinity to zero .
this in turn implies that @xmath48 does not cross zero , and hence @xmath11 is monotonic .
equation ( [ otheryeqn ] ) yields a solution for @xmath47 , and hence @xmath48 , as a function of @xmath45 and the parameters of the theory .
integrated from @xmath49 to infinity , this will give a relation between @xmath50 and the asymptotic value of the field @xmath51 .
the asymptotic field value is essentially a modulus of the theory it will be set by whatever cosmological expectation value is present .
it is a physically meaningful parameter as it affects the coupling to the source by determining @xmath50 .
near the source , where the non - linear term dominates , the solution is linear in @xmath45 , _
rr_(r ) ~(0)+^1/3r , whereas far from the source , where the linear term dominates , the solution goes like @xmath52 , _
rr_(r ) ~()-m4p(^2(0))(0)1r , where the transition between these regimes occurs at the radius r_~(m^2 ^2)^1/6 . [ vainshteinradius ]
note that this crossover radius , and hence the distance at which non - linearities become important , depends on the modulus @xmath50 .
the equation of motion for @xmath53 is readily solved numerically , and the solution obtained is plotted schematically in fig .
[ figure ] .
while the existence of static , spherically - symmetric configurations is encouraging , there are , of course , other important checks that our solution must pass to be physically viable .
specifically , following @xcite , we must study the stability of these spherically symmetric solutions and to determine the speed at which fluctuations propagate , since superluminal propagation can be an obstacle to finding an ultraviolet completion of the effective theory @xcite .
we expand the field in perturbations around the background solution @xmath54 , ^i=^i_0+^i .away from the source , the linearized equations of motion for the perturbations are of the form -k^t_i(r ) _ t^2 ^i + 1r^2 _ r
( r^2 k^r_i(r ) _
r ^i ) + k^_i(r ) _ ^2 ^i = 0 , where the coefficients @xmath55 , @xmath56 and @xmath57 depend on @xmath45 through the background field @xmath39 .
we find k^t_1&=&13r^2dr ; + k^r_1&=&1 + 6y^2 ; + k^_1&=&12rdr ; + k^t_i1&=&13r^2dr ; + k^r_i1&=&1 + 2y^2 ; + k^_i1&=&12rdr .
[ k1 ] applying the implicit function theorem to the function @xmath58 , we have = -_r f_y f=-3ry+2y^31 + 6y^2 .this allows us to eliminate @xmath59 from ( [ k1 ] ) : k^t_1&=&(1 - 6y^2)^21 + 6y^2 ; + k^r_1&=&1 + 6y^2 ; + k^_1&=&1 - 6y^21 + 6y^2 ; + k^t_i1&=&1 + 12 ^ 2 y^41 + 6y^2 ; + k^r_i1&=&1 + 2y^2 ; + k^_i1&=&1 + 2y^21 + 6y^2 .
[ k2 ] stability of the spherically symmetric background solutions against small perturbations requires @xmath60 for all @xmath61 s .
the @xmath62 directions in field space are stable , but the @xmath63 direction exhibits a gradient instability sufficiently close to the source along the angular directions . in other words , @xmath64 near the source .
therefore , localized perturbations can be found near the source that lower the energy of the solution through their gradients .
this instability plagues very short - wavelength fluctuations , right down to the uv cutoff , so decay rates are dominated by the shortest distances in the theory and can not be reliably computed within the effective theory .
equations ( [ k2 ] ) also allow us to compute the speeds of propagation of our small perturbations , in both the radial and angular directions .
these are given by ( c^2)^r_1 & = & k^r_1k^t_1= ( 1 + 6y^2)^2 ; + ( c^2)^_1 & = & k^_1k^t_1=11 - 6y^2 ; + ( c^2)^r_i1 & = & k^r_i1k^t_i1= ( 1 + 2y^2)(1 + 6y^2)1 + 12 ^ 2 y^4 ; + ( c^2)^_i1 & = & k^_i1k^t_i1 = 1 + 2y^21 + 12 ^ 2 y^4 . note that @xmath65 , and hence these perturbations always propagate superluminally .
the same is true of @xmath66 , in regions where these perturbations are stable .
the speed @xmath67 is always superluminal , and @xmath68 is always subluminal .
whether superluminal propagation of signals is problematic for a low - energy effective theory is still an arguable issue , but it seems that at the least it may preclude the possibility of embedding the theory in a local , lorentz - invariant uv completion @xcite .
it is interesting to note in passing that if a mechanism exists to tame the instabilities we have identified , then precision tests of gravity within the solar system already place useful constraints on multi - galileon theories .
the galileon is screened at radii below the vainshtein radius @xmath69 , given by equation ( [ vainshteinradius ] ) , restoring the behavior of general relativity .
requiring the solar system to be screened to @xmath70 thus yields a constraint on @xmath29 and @xmath50 . however , lunar laser ranging data constrain the departure from the gravitational potential predicted by gr to satisfy @xmath71 ( at radius @xmath72 cm ) , and we may translate this into a constraint on a different combination of @xmath29 and @xmath50 for example , consider the choice of @xmath73 , giving a linear coupling between the radial @xmath11 field and matter . in the interesting case when the constraints are saturated , and detection of an effect is therefore imminent ,
the relevant constraint simply becomes 10 ^ -9 ev .
note that this is an extremely low cutoff for the effective theory , as is also found in the dgp model .
we have derived spherically - symmetric solutions in an @xmath2 multi - galileon theory with general , non - derivative couplings to matter .
these solutions exhibit a vainshtein screening effect , characteristic of galileon models .
however , a study of the behavior of fluctuations around these solutions shows that one of the fields has imaginary sound speed along the angular directions , signaling an instability to anisotropic modes of arbitrarily short wavelength .
moreover fluctuations inevitably propagate superluminally .
these results raise serious concerns about the phenomenological viability of @xmath2 multi - galileon theories .
( of course , this does not preclude their effectiveness in early universe physics @xcite , for instance during inflation , as long as they become massive or decouple before the present epoch . ) a key input in our analysis is the restriction to non - derivative coupling to matter .
the main lesson to be drawn is that more general , derivative couplings are necessary . for instance , the lowest - dimensional coupling invariant under the galilean and internal rotation symmetries is @xmath21 .
this coupling in fact naturally arises in the higher - codimension brane picture @xcite . as mentioned earlier ,
the galileon fields are oblivious to static , spherically - symmetric sources in this case ; thus exhibiting a screening mechanism .
however , they will be excited by orbital motion , and we leave a study of the phenomenological implication of this coupling to future work .
our analysis also highlights a distinct advantage to explicitly breaking the symmetry ( [ multisymmetry ] ) , for example through the introduction of a sequence of regulating branes of different co - dimensions , as in the cascading gravity case @xcite .
the explicit breaking of @xmath2 symmetry allows for more general terms in the action , which can lead to a healthier phenomenology @xcite .
this work is supported in part by nasa atp grant nnx08ah27 g , nsf grant phy-0930521 , and by department of energy grant de - fg05 - 95er40893-a020 .
is also supported by the fay r. and eugene l. langberg chair . the work of k.h . and j.k .
is supported in part by funds from the university of pennsylvania . | the 4-dimensional effective theory arising from an induced gravity action for a co - dimension greater than one brane consists of multiple galileon fields @xmath0 , @xmath1 , invariant under separate galilean transformations for each scalar , and under an internal @xmath2 symmetry .
we study the viability of such models by examining spherically symmetric solutions .
we find that for general , non - derivative couplings to matter invariant under the internal symmetry , such solutions exist and exhibit a vainshtein screening effect . by studying perturbations about such solutions , we find both an inevitable gradient instability and fluctuations propagating at superluminal speeds .
these findings suggest that more general , derivative couplings to matter are required for the viability of @xmath2 galileon theories . |
the extraordinary ( optical ) transmission ( eot ) of electromagnetic waves through periodic arrays of subwavelength holes @xcite or slits @xcite made in opaque mostly metallic screens has been exhaustively studied along the last decade .
comprehensive review papers reporting on the basic physics behind the phenomenon are available nowadays @xcite . the research of periodic structures has also stimulated the study of the transmission properties through single holes or slits .
sometimes the physics of eot through periodic structures is closely related with the physics of enhanced transmission through single apertures , as it is the case of slits or holes around which the conducting surface is periodically structured @xcite .
the periodically perforated screen or the periodically structured surface ( a metal surface with corrugations , for instance ) support surface waves ( the so - called _ spoof plasmons _
@xcite ) which are strongly excited around certain critical frequencies .
however , any other mechanism provoking field enhancement at the aperture level would automatically induce enhanced transmission .
thus , it seems almost obvious that high quality factor resonators placed close to or inside the holes will produce resonant enhancement of the fields and , consequently , enhanced transmission @xcite .
other methods have been proposed to enhance transmission of electromagnetic waves through electrically small slits or holes @xcite .
a common feature to all the methods mentioned in the previous paragraph is their intrinsic narrow band behavior .
narrow bandwidth is associated with the resonant nature of the underlying cause of the phenomenon .
a completely different approach is proposed in this contribution . the new approach to obtain extraordinary transmission
is based on light squeezing controlled by non - uniform and non - resonant dielectric permittivity distributions around the aperture .
coordinate transformation is applied to design the permittivity distributions and appropriate approximations are introduced to realize a practical all - dielectric device .
this design is tested by in - house fdtd based simulations and a great enhancement is verified when the novel devices are applied to the sub - wavelength aperture .
although some frequency dependence of transmissivity is observed , the bandwidth of the relevant enhanced transmission frequency region is much larger than any previously reported .
the method of coordinate transformation offers us an easy way to control the travelling route of the electromagnetic waves @xcite . in a cartesian coordinates , an incident electromagnetic wave travels straightly , while in a distorted coordinates it turns according to the mesh . on the other hand ,
maxwell s equations are form - invariant under coordinate transformations @xcite , and , therefore , one can manipulate the electromagnetic wave in a distorted space by engineering the background with the necessary material parameters . the main scheme to enhance the transmission through a sub - wavelength aperture is to build a distorted space which can smoothly induce much more energy from a larger aperture into a smaller one .
[ label:2spaces](a ) shows a space in the cartesian coordinates .
when a plane wave travels into the space from the top , the wave vector @xmath0 is in the direction towards the bottom along the edge of a grid , as illustrated with black arrows , and so is the poynting vector @xmath1 , which represents the flowing of energy . for a perfect electric conductor ( pec ) with a @xmath2 ( @xmath3 the wavelength of free space ) wide slit ,
most incident energy is reflected and little energy could pass through the slit . when the aperture of the slit is enlarged to , for instance , @xmath4 , the transmitted energy increases significantly . a distorted space is designed in fig . [ label:2spaces](b )
, where the wave vector @xmath0 is bending towards the slit along the curvature of the grids and the energy is guided from a virtual electrically - large aperture to an electrically - small slit . in this way
, the transmission through a sub - wavelength slit could be enhanced dramatically in the distorted space .
a discrete coordinate transformation is used to design the background material for the distorted space . instead of using the general coordinate transformation in the whole space , the transformation is operated between every pair of local coordinates @xcite ( a pair of local coordinates is defined as the example in fig .
[ label:2spaces ] ) . to simplify the problem
, here we only discuss the two dimensional ( 2d ) case with @xmath5 , @xmath6 and @xmath7 components .
the jacobian transformation matrix @xmath8 is used to map one space to the other .
the relative permittivity and permeability values of each block in the distorted space are calculated as @xcite @xmath9 notice that @xmath8 is defined for every pair of local coordinates and it has a simpler form in this 2-d case as @xmath10 a practical way to realize the distorted space is to use an all - dielectric background material .
it has been proved that if all the local coordinates in the distorted space are quasi - conformal , the background permeability could be considered as unity and quasi - isotropic for the e - polarization @xcite , then , only the permittivity map is required in the z direction . in this scenario , it is difficult to generate a strictly - orthogonal grid to map the distorted space linking the sub - wavelength slit with an electrically large aperture . instead
, very fine meshes are carefully generated in the distorted space to ensure most cells nearly orthogonal , especially in the central region above the slit .
although such an engineering approximation may not represent a mathematically strict transformation , and , hence , is sensitive to wave incident angles for the proposed device , the performance enhancement in terms of extraordinary wave transmission is still evident over a broadband of frequencies , as demonstrated in the following sections .
the relative permittivity ( @xmath11 ) map is given in fig .
[ label : eps ] .
relative permittivity is very close to unity in most parts of the distorted space , which means most of the background material could be considered as air and dielectrics with varying @xmath11 are mainly required in the central area surrounding the slit . the dashed black line in fig
. [ label : eps ] outlines the profile of such an enhancement device and the detailed permittivity map .
the device has a size of @xmath12 at 8 ghz .
fdtd simulations @xcite are used to test the performance of the designed device without loss of generality .
the operating frequency is set to 8 ghz , the wavelength in free space ( @xmath3 ) being 37.5 mm accordingly .
the sub - wavelength slit on the pec plate is @xmath2 ( 7.5 mm ) wide and @xmath13 ( 2.5 mm ) thick .
the simulation domain is shown in fig .
[ label : field](a ) .
periodic boundary conditions ( pbc ) are applied to realize an incident plane wave with e - polarization impinging from the top .
perfectly matched layers ( pml ) are added at the left and right sides to eliminate the interference from the neighbours @xcite .
the amplitude of the electric field is plotted to quantify the distribution of energy . in fig .
[ label : field](a ) , the incident plane wave irradiates on the pec plate with the sub - wavelength slit .
extremely low field intensity is observed on the other side of the plate .
[ label : field](b ) shows the field distribution when the slit is enlarged to be @xmath4 ( 60 mm ) wide .
much energy passes through the slit and propagates further , as expected . when the proposed enhancement device , designed in fig .
[ label : eps ] , is located over the @xmath2 wide slit ( inside the region defined by the black lines in fig . [
label : field](c ) ) , a significant enhancement of the field is observed on the other side of the pec plate .
ideally , from the point of view of coordinate transformation , the transmitted energy should be the same in ( b ) and ( c ) .
however , the device includes only the central part of the distorted space and the orthogonality in the distorted space is not guaranteed everywhere .
the anisotropic property needs to be included for even enhancing the transmission .
nevertheless , the proposed device still produces strong field enhancement , when comparing figures [ label : field](a ) and [ label : field](c ) . a conventional focusing lens is also applied in fig .
[ label : field](d ) to compare the performances .
the lens has an aperture of @xmath4 and its focal point is set to be at the @xmath2 wide slit to focus more energy directly inside the slit .
the transmission through the @xmath2 wide slit increases only slightly .
thus the proposed device outperforms the conventional dielectric lens for field enhancement .
furthermore , the device can be mirrored to the other side of the pec plate as shown in fig .
[ label : field](f ) , and the transmitted electromagnetic wave will travel through a second distorted space and leads to much more enhanced field distribution . fig .
[ label : field](e ) shows that behind the pec plate the transmitted wave is constrained in a bundle and high directivity is achieved .
wide slit .
( b ) the incident plane wave illuminates a pec plate with a @xmath4 wide slit .
( c ) the incident plane wave illuminates the above mentioned sub - wavelength slit when the proposed enhancement device is applied in the black lines .
( d ) the incident plane wave illuminates the sub - wavelength slit after crossing a focusing lens with its focal point located on the slit .
( e ) the incident plane wave illuminates a pec plate with a pair of enhancement devices at both sides of the sub - wavelength slit .
( f ) the permittivity map around the slit in case ( e ) . the map is symmetric.,width=302 ] radiation patterns are depicted to further compare the field enhancement performances . in fig .
[ label : scattering](a ) , the amplitude of @xmath14 is recorded along the semicircle of radius @xmath15 centered in the slit . the @xmath16 axis is the value of @xmath17 defined in fig .
[ label : field](a ) .
the amplitude of the transmitted @xmath14 field is low when there is only a @xmath2 wide slit on the pec plate .
once the lens is added , the field increases slightly .
however , a significant field enhancement is observed when the proposed device is put above the slit .
the amplification at @xmath18 is about 3.25 for amplitude and 10.56 for power .
when the mirrored device is added on the other side , the amplification increases remarkably . around the angle of @xmath19 ,
the e field is not much weaker than that one gets when there is a @xmath4 wide slit on the pec plate .
note that when the slit is @xmath4 wide , the two side lobes come from the refraction of the pec plate , which is already observed in fig .
[ label : field](b ) .
[ label : scattering](b ) shows the radiation pattern at the semicircle @xmath20 away from the center of the slit ( far field ) .
the patterns in ( b ) are similar to the patterns in ( a ) .
the amplifications are observed as before while the amplitudes decrease uniformly when the waves travel further .
away from the center of the slit .
( b ) the radiation pattern recorded at the semicircle @xmath20 away from the center of the slit.,width=264 ] fig .
[ label : propagation ] also shows the energy distribution along the propagation path .
the energy is recorded at the direction of @xmath19 , from the center of the slit towards the bottom .
when the transmitted waves travel to the far field , the enhancement of energy is preserved once the designed device is applied .
a pair of devices produce much more transmitted energy than a single one all along the propagation path .
note that the blue curve , which represents the energy when a pair of devices are used , is not smooth from @xmath21 to @xmath22 .
this area is inside the mirrored device and the reflections between different dielectric blocks are considerable .
to quantify the transmitted energy , a broadband gaussian pulse is applied to excite the slit .
its behaviour in the frequency domain is depicted as the sub - figure in fig .
[ label : fre1](a ) by using fourier transformation .
the poynting vector over the slit represents the energy passing through . by integrating the poynting vector at the lower surface of the pec plate , the total energy transmitted over time is available . notice that at the lower surface of the pec plate , only perpendicular electric field and parallel magnetic field exist .
so , the transmitted energy could be calculated as shown in fig .
[ label : poynting ] .
table 1 gives the values of the transmitted energy in different cases .
it is obvious that extremely poor transmission is obtained for the 7.5 mm slit if compared with the energy crossing the 60 mm slit . when a focusing lens is applied , the enhancement is very limited .
but once the enhancement device studied in this paper is located above the 7.5 mm wide slit , the transmission increases dramatically ( it is about 13 times the energy transmitted without the device ) . when applying the second device the transmission
is further increased ( the enhancement factor is about 16 ) . with the implementation of the devices ,
the transmitted energy increases to about @xmath23 of the value when the slit has an aperture of 60 mm .
the discrepancy between case 1 and case 4 is due to the fact that the device is mismatched to the free space and the anisotropic material properties is ignored in the design .
c c l ccase no .
& & the pec plate with : & transmitted energy ( arb . units ) + 1 & & a 60 mm wide slit & 965 + 2 & & a 7.5 mm wide slit & 19 + 3 & & a 7.5 mm wide slit and a lens & 28 + 4 & & a 7.5 mm wide slit and a device & 248 + 5 & & a 7.5 mm wide slit and a pair of devices & 300 +
to investigate the broadband performance of the devices , an incident pulse spaning from 4 ghz to 8 ghz is applied and the average amplitude of transmitted ez field at the lines @xmath24 and @xmath15 ( at 6 ghz ) away behind the pec plate is recorded during every time step .
the frequency domain responses are calculated and plotted in fig .
[ label : fre1 ] .
slight enhancement is observed at some frequencies , when the conventional lens is used .
once the proposed devices are applied , the transmitted electric field is considerably higher than the values without any devices over the whole frequency domain . when the transmitted waves travel from the very near field of @xmath24 [ fig .
[ label : fre1](a ) ] to the far field of @xmath15 [ fig .
[ label : fre1](b ) ] , the significant enhancement is still held .
note that in fig .
[ label : fre1](a ) , since the observation line is inside the mirrored device , the case with a couple of devices is not plotted .
( at 6 ghz ) away behind the pec plate .
( b ) the average amplitude of ez field recorded at the line @xmath15 ( at 6 ghz ) away behind the pec plate.,width=238 ] the transmission enhancements produced by the devices and the lens are also calculated and plotted in fig .
[ label : fre2 ] .
the amplitude of electric field is normalized at each frequency point by the value of the green curve , which represents the transmitted field through the sub - wavelength slit without any devices .
the results give the detailed information of the amplification and prove that the designed device has significant enhancement performance over a wide frequency range .
( at 6 ghz ) away behind the pec plate over the frequency spectrum .
( b ) amplifications at the line @xmath15 ( at 6 ghz ) away behind the pec plate over the frequency spectrum .
the amplitude of electric field is normalized at each frequency point by the value of the green curve ( which represents the transmitted field through the sub - wavelength slit ) to represent the amplification factor.,width=238 ]
a device made by dielectric blocks is proposed using the concept of discrete coordinate transformation .
extraodinary transmission in a sub - wavelength aperture is achieved over a broad band from 4 ghz to 8 ghz when the device is applied , thanks to the nonresonant nature of the device . the implementation of a pair of mirrored devices results in even more transmission and higher directivity behind the aperture .
fdtd based simulation results have verified the expected performances .
the device can be modified to operate for all angles of incidence by rotating the proposed structures .
overall , the device demonstrates superior performance over other existing approaches for broadband extraordinary transmission in a single slit .
more importantly , the design based on transformation electromagnetics can be applied to optical frequencies as the required material properties can be easily found from the nature .
the authors would like to thank the support of the `` spanish ministry of science and innovation '' ( mobility program , grant number p2009 - 0405 ) .
they d also thank mr . c. argyropoulos for his help during the preparation of the manuscript . | coordinate transformation is applied to design an all - dielectric device for extraordinary transmission ( et ) in a single sub - wavelength slit .
the proposed device has a broadband feature and can be applied from microwave to visible frequency bands
. finite - difference time - domain ( fdtd ) simulations are used to verify the device s performance .
the results show that significantly increased transmission is achieved through the sub - wavelength aperture from 4 ghz to 8 ghz when the device is applied .
in contrast with previously reported systems , the frequency sensitivity of the new device is very low .
99 t. w. ebbesen , h. j. lezec , h. f. ghaemi , t. thio and p. a. wolff , `` extraordinary optical transmission through sub - wavelength hole arrays , '' nature * 391 * , 667 - 669 ( 1998 ) .
j. a. porto , f. j. garca - vidal and j. b. pendry , `` transmission resonances on metallic gratings with very narrow slits , '' phys .
rev .
lett . * 83 ( 14 ) * , 2845 - 2848 ( 1999 ) .
c. genet and t. w. ebbesen , `` light in tiny holes , '' nature * 445 * , 39 - 46 ( 2007 ) .
f. j. garca - de - abajo , `` colloquium : light scattering by particle and hole arrays , '' rev .
mod .
phys . * 79 * , 1267 - 1290 ( 2007 ) .
f. j. garca - vidal , l. martn - moreno , t. w. ebbesen and l. kuipers , `` light passing through subwavelength apertures , '' rev .
mod .
phys .
* 82 * , 729 - 787 ( 2010 ) .
h. j. lezec , a. degiron , e. devaux , r. a. linke , l. martn - moreno , f. j. garca - vidal and t. w. ebbesen , `` beaming light from a subwavelength aperture , '' science * 297 ( 5582 ) * , 820 - 822 ( 2002 ) .
f. j. garca - vidal , h. j. lezec , t. w. ebbesen and l. martn - moreno , `` multiple paths to enhance optical transmission through a single subwavelength slit , '' phys .
rev .
lett . * 90 * , 213901(1 - 4 ) ( 2003 ) .
l. a. dunbar , m. guillaume , f. de len - prez , c. santschi , e. grenet , r. eckert , f. lpez - tejeira , f. j. garca - vidal , l. martin - moreno and r. p. stanley , `` enhanced transmission from a single subwavelength slit aperture surrounded by grooves on a standard detector , '' app .
. lett . * 95 * , 011113(1 - 3 ) ( 2009 ) .
j. b. pendry , l. martn - moreno and f. j. garca - vidal , `` mimicking surface plasmons with structured surfaces , '' science * 305 * , 847 - 848 ( 2004 ) .
k. aydin , a. o. cakmak , l. sahin , f. bilotti , l. vegni and e. ozbay , `` split - ring - resonator - coupled enhanced transmission through a single subwavelength aperture , '' phys .
rev .
lett . * 102 * , 013904(1 - 4 ) ( 2009 ) .
a. o. cakmak , k. aydin , e. colak , z. li , f. bilotti , l. vegni and e. ozbay , `` enhanced transmission through a subwavelength aperture using metamaterials , '' app .
. lett . * 95 * , 052103(1 - 3 ) ( 2009 ) .
f. bilotti , l. scorrano , e. ozbay and l. vegni , `` enhanced transmission through a sub - wavelength aperture : resonant approaches employing metamaterials , '' j. opt .
a : pure appl . opt . *
11 * , 114029 ( 2009 ) .
f. j. valdivia - valero and m. nieto - vesperinas , `` resonance excitation and light concentration in sets of dielectric nanocylinders in front of a subwavelength aperture .
effects on extraordinary transmission , '' opt .
express * 18 ( 7 ) * , 6740 - 6754 ( 2010 ) .
j. b. pendry , d. schurig and d. r. smith , `` controlling electromagnetic fields , '' science * 312 * , 1780 - 1782 ( 2006 ) .
u. leonhardt , `` optical conformal mapping , '' science * 312 * , 1777 - 1780 ( 2006 ) .
d. schurig , j. b. pendry and d. r. smith , `` calculation of material properties and ray tracing in transformation media , '' opt .
express * 14 ( 21 ) * , 9794 - 9804 ( 2006 ) .
j. b. pendry and j. li , `` hiding under the carpet : a new strategy for cloaking , '' phys .
rev .
lett .
* 101 ( 20 ) * , 203901(1 - 4 ) ( 2008 ) .
j. b. pendry , a. j. holden , w. j. stewart and i. youngs , `` extremely low frequency plasmons in metallic mesostructures , '' phys .
rev .
lett . * 76 ( 25 ) * , 4773 - 4776 ( 1996 ) .
j. b. pendry , a. j. holden , d. j. robbins and w. j. stewart , `` magnetism from conductors and enhanced nonlinear phenomena , '' ieee trans . microwave theory tech .
* 47 ( 11 ) * , 2075 - 2084 ( 1999 ) .
y. zhao , c. argyropoulos and y. hao , `` full - wave finite - difference time - domain simulation of electromagnetic cloaking structures , '' opt .
express * 16 ( 9 ) * , 6717 - 6730 ( 2008 ) . c. argyropoulos ,
y. zhao and y. hao , `` a radially - dependent dispersive finite - difference time - domain method for the evaluation of electromagnetic cloaks , '' ieee trans .
antennas propag .
* 57 ( 5 ) * , 1432 - 1441 ( 2009 ) .
a. taflove and s. c. hagness , _ computational electrodynamics : the finite - difference time - domain method ( 3rd edition ) _ ( artech house , boston , 2005 ) . |
this paper is a continuing work of @xcite , focusing on neutrinoproduction of photons and pions from nuclei with neutrino energy @xmath0 . in refs .
@xcite , we introduced the @xmath1 resonance as a manifest degrees of freedom to the effective field theory ( eft ) , known as _ quantum hadrodynamics _ or qhd @xcite .
( the motivation for this eft and some calculated results are discussed in refs .
@xcite . ) to calibrate the reaction mechanism on the nucleon level , we studied the productions from nucleons @xcite .
the calculations are motivated by the fact that the neutrinoproductions of @xmath6 and photons from nuclei ( and nucleons ) are potential backgrounds in neutrino - oscillation experiments ( e.g. , miniboone @xcite ) .
currently , it is still a question whether the neutral current ( nc ) photon production might explain the excess events seen at low _ reconstructed _ neutrino energies , which the microboone experiment plans to answer @xcite .
moreover , the authors of refs .
@xcite point out the possible role of anomalous interaction vertices involving @xmath7 , @xmath8 , and the photon in nc photon production .
so it is necessary to calculate the cross sections for these processes . here by using the qhd eft , we study incoherent production ,
in which the nucleus is excited .
coherent production with the nucleus being intact is a topic of future work .
we will discuss the power - counting .
see refs.@xcite for detailed discussions about power - counting in qhd eft .
] of the calculations through which we will show that the contributions of the anomalous interactions are small in the incoherent nc production of photons ( where they contribute at next - to - next - to - leading - order ) . to benchmark the approximation scheme , we study electron scattering in both quasi - elastic and pion production channels .
there have been several experiments measuring the weak response of nuclei across the quasi - elastic region to the @xmath1 excitation peak . in most experiments @xcite , which have @xmath9 and @xmath10 as the primary target nuclei ,
the mean energy of the beam is around @xmath11 .
as emphasized in @xcite , we expect our theory to work up to @xmath12 gev , so we do not rely on these experiments to constrain the theory at this stage . on the theoretical side ,
much work has been done ( e.g. , in @xcite ) .
most of these papers are based on the global or local fermi gas approximation and include contributions from one - body currents , with improved treatment for final - state interaction ( fsi ) and @xmath1 dynamics in the medium .
the same approach has also been applied in electron scattering ( e.g. , in @xcite ) . in @xcite , scaling approaches are used to address quasi - elastic scattering . moreover ,
the contribution from two - body currents was studied nonrelativistically , for example , in @xcite . in most of these calculations ,
the @xmath1 dynamics in nuclei is based on the work of @xcite , in which the @xmath1 self - energy has been studied using a nonrelativistic model .
parallel to the nonrelativistic studies , some work has been initiated in the relativistic framework , qhd eft , using the local fermi gas ( lfg ) approximation and including one - body currents @xcite .
the two - body current was investigated relativistically in @xcite .
these works mainly focus on electron scattering .
but the handling of the @xmath1 resonance in these papers is somewhat phenomenological .
moreover , in both nonrelativistic and relativistic studies , photon production is rarely investigated . in this paper
, we also apply the lfg approximation @xcite to study the one - body current contribution .
as shown in @xcite , we make use of meson dominance to generate form factors for various currents . because of the built in symmetries in the lagrangian , conservation of vector current and the partial conservation of axial current are satisfied .
these properties are well preserved in the lfg approximation . especially for photon production ,
vector current conservation is crucial .
the @xmath1 dynamics , as a key component in this work , is explored to some extent .
we introduce interactions between @xmath1 and non - goldstone meson fields to generate the spin - orbit ( s - l ) coupling that has been introduced in phenomenological models @xcite . on the other hand , phenomenological knowledge about s - l coupling
puts constraints on these couplings .
moreover , the @xmath1 decay width increases in the nucleus , because more decay channels are opened up and this effect overcomes the reduction of pion decay phase space . here
we follow the phenomenological studies and separate the width to the pion decay width and anything else parametrized by the imaginary part of the @xmath1 spreading potential . as a result of opening new decay channels ,
the flux having excited a @xmath1 resonance can be transferred to channels that do not involve pion or photon production .
moreover , pauli blocking can reduce the pion and photon production cross section further , because of the reduction of the final particle s phase space . in this paper , we explore how both @xmath1 and nonresonant contributions are reduced compared to those in free nucleon scattering .
however , we do not include fsi effects for pions and knocked out nucleons .
the simple treatment can be found in @xcite , while the complete treatment is implemented in various event generators of experiments ( e.g. , nuance @xcite ) , and the gibuu model @xcite .
hence we only compare our predictions with the output of nuance without fsi . dominates .
the axial mass @xmath13 gev is used which is the same as that used by the miniboone experiment for their _ baseline _ calculations @xcite .
however , the actual backgrounds used in their final analyses were scaled to data in a separate exercise . ]
the paper is organized as follows . in sec .
[ sec : quasi ] , we first discuss the lfg approximation and then apply it to electron quasi - elastic scattering , which serves as a benchmark . in sec .
[ sec : pionprod ] , the calculation scheme for pion ( photon ) production is briefly introduced .
then the @xmath1 dynamics is studied with emphasis on the connection between @xmath1-meson interactions and s - l coupling .
the modification of the @xmath1 width is also discussed .
after that , electron scattering at the @xmath1 peak is studied , and results are compared with data with explanation of the missing strength .
the cross sections of neutrinoproduction of pions are also shown and compared to nuance s output .
[ sec : photonprod ] is dedicated to nc photon production .
finally sec .
[ sec : sum ] contains a short summary . in the appendices , we show detailed kinematic analyses for both quasi - elastic scattering and pion production
this section serves as an illustration of the lfg approximation used for quasi - elastic scattering and for photon and pion production .
( see ref .
@xcite for discussion on the free nucleon interaction amplitude in all these processes . ) here we make use of the mean - field approximation to calculate the nuclear ground state . the relevant leading order lagrangian is @xmath14n \ ] ] ( where the full lagrangian can be found in @xcite for example ) . the mean - field approximation is presented simply as follows .
inside nuclear matter , vector @xmath15 and @xmath16 , and scalar @xmath17 fields develop nonzero expectation values . in the laboratory frame of the matter ,
only two fields ( @xmath17 and @xmath18 ) have nonzero values ( but in the isospin asymmetric case , @xmath19 can also develop a nonzero value ) . as a result ,
the nucleon s mass is modified : @xmath20 . at the lowest order ,
the spectrum of nucleons is @xmath21 . inside a finite nucleus , due to
different boundary conditions , the mean - field expectation value is space dependent and can be calculated numerically . by using this approximation
, we can calculate the bulk properties of the nucleus , the details of which can be found in ref .
@xcite for example .
following @xcite , we calculate the local density @xmath22 and field expectation value in @xmath9 ( the major nucleus in the miniboone s detector ) .
[ fig : c12density ] and [ fig : c12fields ] show the results based on @xmath23 and @xmath24 parameter sets in @xcite .
we will explore the difference due to the two sets in electron quasi - elastic scattering . to calculate the electroweak response of nuclei
, we use the lfg approximation .
this approach has been applied in @xcite to study electron quasi - elastic scattering .
first , by assuming the impulse approximation ( ia ) , the interaction happens every time between probe and each individual nucleon this only holds when the transferred momentum is high enough that the interference between different nucleons is reduced due to the big recoil .
second , the response of the nucleus is the incoherent sum of the response of the fermion gas in different regions .
this works when the probe s wave length is small enough compared to a characteristic length scale of the nucleus density profile .
the discussion can be summarized in the following equation : @xmath25 in this equation , @xmath26 and @xmath27 are the incoming and outgoing lepton momenta , respectively , @xmath28 is the momentum transfer , @xmath29 @xmath30 and @xmath31 @xmath32 are the scattered nucleon s initial and final momenta and spin projection , and @xmath33 is the one - body interaction amplitude . the kinematic configuration is shown in fig .
[ fig : nucleilabframeconfiguration ] .
the integration over initial and final nucleon momenta depends on the space dependent fermi momentum .
a detailed discussion about this equation can be found in appendix [ app : quasikinematics ] . the interaction amplitude @xmath33 in eq .
( [ eqn : quasitcrosssection ] ) can be expressed in terms of various current matrix elements ( where @xmath34 , @xmath35 , and @xmath36 are , respectively , the isovector vector current , the isovector axial - vector current and the baryon current , @xmath37 @xcite ) . for electron quasi - elastic scattering , @xmath38 \langle j_{em}^{(had ) \mu } \rangle & \equiv & { \langlen , b\vert } v^{0\mu}+{\ensuremath{\frac{1}{2}}}j_{b}^{\mu}{\vertn , a\rangle } \ , \label{eqn : matrixelementemcurrent } \end{aligned}\ ] ] where @xmath39 and @xmath40 in the state are nucleon isospin . for charged current ( cc ) quasi - elastic scattering , @xmath41 \langle j_{l}^{(had ) i \mu } \rangle & \equiv & { \langlen , b\vert } { \ensuremath{\frac{1}{2}}}(v^{i \mu}+a^{i \mu}){\vertn , a\rangle } \ , \label{eqn : matrixelementcccurrent}\end{aligned}\ ] ] where @xmath42 , @xmath43 is the fermi constant and @xmath44 is the @xmath45 and @xmath46 quark mixing element in the cabibbo - kobayashi - maskawa ( ckm ) matrix .
for nc quasi - elastic scattering , @xmath47 \langle j_{nc}^{(had ) \mu } \rangle & \equiv & { \langlen , b\vert } j_{l}^{0\mu } - \sin^{2}\theta_{w}j_{em}^{\mu } { \vertn , a\rangle } \ , \label{eqn : matrixelementnccurrent}\end{aligned}\ ] ] where @xmath48 is the weak mixing angle .
the electroweak currents of leptons are well known , and @xmath49 can be found in @xcite .
but we need to include the nucleon spectrum modification to the results of @xcite , which is straightforward to complete in the lfg approximation .
a short discussion on fsi is in order here .
the picture is the following : the interaction channels are opened in the initial interacting vertex , and then these channels would couple to each other when particles are traveling through the nucleus . the flux among all the initial channels are redistributed due to fsi .
this picture is adopted in the gibuu model calculations for example for cc and nc processes @xcite
@xcite . from conservation of probability , assuming the picture mentioned above is valid , we should expect the sum of these channels in the initial vertex to match the inclusive data .
moreover , coulomb distortion of the electron is not included in this calculation . in the upper panels of figs .
[ fig : emscattering_0.62_60 ] , [ fig : emscattering_0.68_36 ] , and [ fig : emscattering_0.73_37.1 ] , we present differential cross sections @xmath50 for electron scattering off @xmath9 at given electron energies and scattering angles . in this section
we only focus on the so - called quasi - elastic peak at the lower energy region , which is believed to be dominated by one nucleon knock out .
the higher energy peak will be discussed in sec .
[ sec : pionprod ] . in fig .
[ fig : emscattering_0.62_60 ] , the electron energy is @xmath51 , and the scattering angle is @xmath52 .
the plots `` g1 '' and `` g2 '' are the calculations done with g1 and g2 parameter sets @xcite .
the difference between the two is small .
the data are from ref .
the validity of the form factors realized by meson dominance needs to be discussed here . in this figure , @xmath53 and @xmath54 at the peak . below the peak , @xmath55 is slightly higher than @xmath56 , and above the peak , @xmath57 . as discussed in @xcite ,
meson dominance works when @xmath57 , and hence it can be applied here .
this is also true for figs .
[ fig : emscattering_0.68_36 ] and [ fig : emscattering_0.73_37.1 ] . in fig .
[ fig : emscattering_0.68_36 ] , the electron energy is @xmath58 , and the scattering angle is @xmath59 . in fig .
[ fig : emscattering_0.73_37.1 ] , the electron energy is @xmath60 , and the scattering angle is @xmath61 .
the data are from @xcite and @xcite .
again , we see only small differences between the `` g1 '' and `` g2 '' parameter sets .
by using the lfg approximation detailed before , the formula for the cross section can be written as @xmath62 & & \times ( 2\pi)^{4 } \delta^{4 } ( q+{p_{ni}^{\ast}}-{p_{nf}^{\ast}}-k_{\pi } ) \underset{s_{f},s_{i}}{\sum } \vert m_{fi } \vert^{2 } \ .
\label{eqn : xsecpion}\end{aligned}\ ] ] the details can be found in appendix [ app : pionprodkinematics ] .
the notations for various momenta are explained in fig .
[ fig : nucleipionprodlabframe ] .
all the integrations except the volume integration depend on the space coordinate @xmath63 through the space dependent fermi momentum . the amplitude @xmath33 in eq .
( [ eqn : xsecpion ] ) is similar to those in eqs .
( [ eqn : matrixelementemcurrent ] ) , ( [ eqn : matrixelementcccurrent ] ) , and ( [ eqn : matrixelementnccurrent ] ) , except that the hadronic currents should be changed to those relevant to pion production : @xmath64 several feynman diagrams contribute here , as shown in fig .
[ fig : feynmanpionproduction ] , including diagrams with the @xmath1 [ ( a ) and ( b ) ] and all the rest which we define as nonresonant diagrams . see ref .
@xcite for details about them . among the medium - modifications of the matrix elements , the behavior of the @xmath1 needs to be singled out .
first , let us focus on the real part of the @xmath1 self - energy .
we start from the following lagrangian ( and a similar lagrangian can be found in @xcite ) : @xmath65 & & { } -\frac{\widetilde{f}_{\rho}h_{\rho}}{4 m } { \overline{\delta}}_{\lambda } \rho_{{\mu\nu } } { \sigma^{{\mu\nu}}}\delta^{\lambda } -\frac{\widetilde{f}_{v}h_{v}}{4 m } { \overline{\delta}}_{\lambda } v_{{\mu\nu } } { \sigma^{{\mu\nu } } } \delta^{\lambda } \ .
\label{eqn : deltabg } \end{aligned}\ ] ] here the @xmath1 field is given by the rarita - schwinger representation , and @xmath66 are @xmath1 isospin indices @xcite .
at the normal nuclear density , the @xmath1 is not stable in the nuclear medium .
so the expectation values of meson fields are not changed in normal nuclei at the mean - field level .
similar to the nucleon case , the @xmath1 spectrum in nuclear matter ( without the @xmath1-pion interaction ) is given by @xmath67 & \equiv & h_{v } \langle v^{0 } \rangle + p_{\delta}^{\ast 0 } \notag \\[5pt ] & = & h_{v } \langle v^{0 } \rangle + \sqrt{{\mstar}^{2}+\vec{p}_{\delta}^{\ast 2 } } \ , \notag \\[5pt ] \mstar & \equiv & m - h_{s } \langle \phi \rangle \ .
\notag \end{aligned}\ ] ] the effect of introducing @xmath68 and @xmath69 couplings on the equation of state ( eos ) was analyzed in @xcite .
some constraints on the couplings , @xmath70 and @xmath71 , were calculated in @xcite . here
we resort to the scattering problem to find other constraints . in pion - nucleus scattering studies
@xcite , s - l coupling of the @xmath1 inside the nucleus was introduced by hand , although its origin is not clear in the nonrelativistic model . in this model , a mechanism similar to the generation of the nucleon s s - l coupling is used to generate @xmath1 s . following discussions in @xcite and using the lagrangian in eq .
( [ eqn : deltabg ] ) , we can estimate the s - l coupling of the @xmath1 as @xmath72 \vec{s } \cdot \vec{l } \notag \\[5pt ] & \equiv & \alpha(r ) \vec{s } \cdot \vec{l } \ .
\label{eqn : lscoupling } \end{aligned}\ ] ] here , @xmath73 . in fig .
[ fig : ls_c12 ] , we compare our estimates of @xmath74 defined in eq .
( [ eqn : lscoupling ] ) with two different phenomenological fits for @xmath9 .
we can see that our estimates based on three different parameter sets , @xmath75 , are consistent with the `` nakamura '' result in @xcite , while the `` horikawa '' result in @xcite is significantly larger than the `` nakamura '' result when @xmath76 .
meanwhile , all the couplings are consistent with the `` naturalness '' assumption , which also motivates our choice of @xmath77 .
we do not show the consequence of @xmath78 , since there is no s - l coupling generated in this case .
second , we turn to the imaginary part of the self - energy .
it is known that pauli blocking effects decrease the width due to reduction of the pion - decay phase space , while the collision channels , @xmath79 for example , increase the width .
the two competing processes have been investigated in nonrelativistic models . at normal nuclear density ,
the net result is to increase the width @xcite . in phenomenological fits
@xcite @xcite , this increase is taken into account by introducing a density - dependent complex spreading potential for the @xmath1 . here
we follow this approach . above the pion threshold ,
@xmath80 @xmath81 is the @xmath1 pion - decay width @xcite .
calculation , only pauli - blocking is considered .
modifications of the real part of the nucleon and @xmath1 self - energies are not included . ]
@xmath82 ( @xmath83 ) is the width in other channels , and it has been fitted in @xcite . below the pion threshold ( which is useful in photon production ) , @xmath84 in the cross channel of the @xmath1 diagram , we set the width to zero .
moreover , in the literature @xcite , the simple increase of the @xmath1 width by @xmath85 has been used for pion production : @xmath86 this procedure turns out to work qualitatively , as will be shown later . furthermore , in @xcite , the @xmath1 self - energy calculated in @xcite is used for inclusive electron scattering off nuclei . in sec .
[ subsection : eletroproduction ] , we will compare our results using eqs .
( [ eqn : phenomenology_delta_width1 ] ) and ( [ eqn : phenomenology_delta_width2 ] ) with those using eq .
( [ eqn : phenomenology_delta_width3 ] ) and the width in @xcite . here we focus on the region beyond quasi - elastic scattering in the upper panels of figs .
[ fig : emscattering_0.62_60 ] , [ fig : emscattering_0.68_36 ] , and [ fig : emscattering_0.73_37.1 ] .
it is believed that the second peak mainly comes from the @xmath1 excitation inside the nucleus . in the upper panels of these figures ,
we provide our pion - production results ( without fsi ) due to six different calculations .
we include the full set of feynman diagrams in the first five calculations , and diagrams with the @xmath1 in @xmath87 and @xmath45 channels in the sixth .
the difference among the first three calculations is the choice of @xmath88 parameter sets : @xmath89 , @xmath90 , and @xmath91 . in these three ,
the @xmath1 width shown in eqs .
( [ eqn : phenomenology_delta_width1 ] ) and ( [ eqn : phenomenology_delta_width2 ] ) is applied . in the fourth calculation , we set @xmath90 and apply the constant shift of the @xmath1 width as shown in eq .
( [ eqn : phenomenology_delta_width3 ] ) .
the fifth calculation is done by using the @xmath1 self - energy as calculated in @xcite , which is essentially repeating the calculations in @xcite .
the sixth calculation has @xmath90 and uses the same @xmath1 width as used in the first three .
first , let us discuss the location of the @xmath1-peak along the @xmath92 axis .
the different choice of @xmath93 and @xmath94 indicates different binding potentials for @xmath1 . for @xmath95 ,
the real part of the self - energy is the same as in vacuum without any binding . for @xmath96
, @xmath1 has the same attractive potential as the nucleon .
the vector part tuned by @xmath94 provides a repulsive potential .
so we can see that @xmath97 has a deeper binding potential than @xmath98 . hence , @xmath95 is less bound than @xmath99 and @xmath99 is less bound than @xmath97 . in figs .
[ fig : emscattering_0.62_60 ] , [ fig : emscattering_0.68_36 ] , and [ fig : emscattering_0.73_37.1 ] , the location of the @xmath1 peak in first three calculations indeed follows this argument .
( we can estimate the location of the @xmath1 peak in a global fermi gas model . , we assume a global fermi gas , with the nucleon effective mass @xmath100 , @xmath101 , @xmath102 ( in the laboratory frame ) , and @xmath103 . meanwhile in the @xmath87 channel ,
the @xmath1 momentum is @xmath104 .
it is easy to calculate the @xmath92-location of the peak by setting @xmath105 . for @xmath106 and @xmath107 ( see fig . [
fig : emscattering_0.62_60 ] ) , @xmath108 if @xmath109 and @xmath110 if @xmath111 . ] ) the fourth calculation with @xmath99 and constant width does not give the correct peak position in the three figures : it underestimates ( overestimates ) @xmath1 s contribution on the left ( right ) side of the peak , because the constant width assumption overestimates ( underestimates ) @xmath1 s width on the left ( right ) side .
the fifth calculation by using @xmath1 modification calculated in @xcite gives the correct location of the peak .
comparing the second with the sixth calculation , we can see the significance of nonresonant contributions ( they use the same set of parameters and the @xmath1 width ) .
however , the pion production channel could not explain the full strength of the @xmath1 peak .
meanwhile in the `` dip '' region between the quasi - elastic scattering and the @xmath1 peak , the calculations also miss strength .
this indicates we miss other channels from dip region to the @xmath1 peak .
missing strength at the _ peak _ position can be qualitatively explained by considering the fourth calculation in the upper panels , whose simple treatment of the @xmath1 width makes analysis transparent . according to eqs .
( [ eqn : phenomenology_delta_width1 ] ) and ( [ eqn : phenomenology_delta_width3 ] ) , we estimate @xmath112 in the sense of averaging over @xmath9 , and hence @xmath113 .
the comparable width of other decay channels shows their importance to the inclusive data
. moreover , there are contributions from two - body currents without @xmath1 as an intermediate state . in the lower panels of figs .
[ fig : emscattering_0.62_60 ] , [ fig : emscattering_0.68_36 ] , and [ fig : emscattering_0.73_37.1 ] , we add up three different channels : quasi - elastic , pion production , and two - body current contributions [ labeled as meson - exchange - current ( mec ) in the plots ] .
the label `` @xmath114 '' for pion production applies under the assumption that @xmath90 and that no new channel takes away the flux from the @xmath1 pion production .
the mec - contributions are from @xcite @xcite . here
the total strength matches well with the inclusive data .
however a detailed study of different channels in the qhd eft framework is needed to address this issue conclusively .
the difference between our calculations and those in @xcite where the qhd model is also applied should be mentioned here .
ours are strictly based on the field theory , while in @xcite the @xmath1 is introduced by hand ( where they were convoluted with the cross section based on a `` stable '' @xmath1 theory with a lorentzian weight function ) .
moreover , we take into account the contribution from other diagrams , which are not considered in @xcite .
the two results are different somewhat , but our choice @xmath115 is consistent with the analysis in @xcite .
moreover , we can see that the differences in cross sections obtained using @xmath115 are not significant , which indicates that the total cross sections of neutrinoproduction processes are not sensitive to them either .
this will be confirmed by the results in sec .
[ subsec : ccncpion ] .
[ fig : p_ppion+_c12 ] shows the total cross section averaged over proton or neutron number for cc pion production in ( anti)neutrino@xmath9 scattering .
we also compare our result with nuance s output without fsi . in each figure ,
our calculations including different diagrams and using different @xmath93 and @xmath94 are shown .
the `` only @xmath1 '' calculation only takes into account @xmath1 diagrams . in the others , all the diagrams up to @xmath116 are included . systematically in all the channels , our `` only @xmath1 '' calculation is close to the nuance output
but other diagrams contained in `` @xmath116 '' calculations are not negligible in all the channels around the resonance region , especially when the @xmath87-channel @xmath1 contribution is suppressed by the small clebsch - gordan coefficients ( for example , @xmath117 ) . in the very low energy region away from the resonance , nonresonant diagrams dominate .
see @xcite for the power counting of diagrams .
moreover , we check that the contributions of higher order ( @xmath118 ) diagrams are tiny .
we also can see that below 0.5 gev , the @xmath119 results are bigger than the @xmath99 results and the @xmath99 results are bigger than the @xmath95 results . following the discussions in sec .
[ subsection : eletroproduction ] about the location of the @xmath1 peak in pion electroproduction , we expect that , at a given energy , @xmath1 excitation occupies more phase space in @xmath97 than in @xmath99 and more in @xmath99 than in @xmath95 .
so the pattern among the three different calculations is consistent with the qualitative analysis . here
@xmath95 is presented simply for the purpose of comparison , and its conflict with the @xmath1 s - l coupling is presented in sec .
[ subsec : deltamodification ] .
one question needs to be raised : do we have @xmath1-dominance in the _ nuclear _ scattering around @xmath12 gev ? if we compare the `` @xmath116 '' calculations with `` only @xmath1 '' calculations in every channel , the answer is no .
it turns out that the @xmath1 contribution is strongly reduced due to the broadening of its width , compared to its contribution in free nucleon scattering .
meanwhile nonresonant contributions are reduced by pauli blocking . to see this qualitatively , compare our results here with the cross sections shown in @xcite for production from free nucleons . in @xcite ,
two different calculations can be found , including `` only @xmath1 '' and `` @xmath116 '' .
we just show the total cross sections at @xmath120 gev in tab .
[ tab : freeboundcc ] for neutrino scattering .
for example , `` @xmath121 '' indicates the channel @xmath122 . ``
( f ) '' and `` ( b ) '' correspond to scattering from free nucleons and from bound nucleons in @xmath9 respectively . in both `` only @xmath1 ( b ) '' and `` @xmath116 ( b ) , '' @xmath123 ( and
note that calculations with only @xmath1 and @xmath123 are not shown in the figures ) .
`` nonresonant ( b ) '' is the difference between the two , and can be viewed qualitatively as the contributions of the nonresonant diagrams .
and other diagrams . at @xmath120 gev
, we can assume in most of phase space that the @xmath1 is `` on shell '' while contributions from other diagrams are real , and hence the interferences are small . ]
the labeling for free nucleon scattering is the same .
we can see that the @xmath1 contribution in nuclear scattering has been reduced systematically by around 50% in all channels , compared to its contribution in nucleon scattering ; the nonresonant contributions are also strongly reduced .
clearly , the nonresonant contributions are not negligible in both nucleon and nuclei scattering .
the same situation occurs in the antineutrino scattering channels and hence are not shown explicitly .
this underscores the importance of including nonresonant contributions in cc pion production .
.total cross sections averaged over number of protons or nucleons for cc pion production in neutrino@xmath9 scattering at @xmath120 gev .
see the text for detailed explanations . in the nuclear scattering , @xmath123 .
[ cols="^,^,^,^,^,^,^",options="header " , ]
neutrinoproduction of photons and pions from nuclei provides an important background in neutrino - oscillation experiment and must be understood quantitatively .
especially , we are interested in the possible role of nc photon production in the excess events seen in the miniboone experiment at low _ reconstructed _ neutrino energy . in ref .
@xcite , we have calibrated our theory
qhd eft with @xmath1 introduced by calculating photon and pion production from free nucleons up to @xmath124 . in this work
, the theory is applied to study the production from nuclei . here
we make use of the lfg approximation and impulse approximation , and include only one - body current contributions . in the mean - field approximation of the nuclear ground state ,
the change of the baryon spectrum is represented by introducing an effective mass for baryons , which leads to the change of one - body currents in this calculation .
the calculation for electron quasi - elastic scattering and electroproduction of pion serves as a benchmark for our approximation schemes .
we then proceed to calculate the neutrinoproduction of pion and photon from @xmath9 , and show the plots for total cross section in every channel .
first , we present calculations for pion production up to next - to - leading - order with different @xmath93 and @xmath94 parameters as constrained by the phenomenological study .
it turns out that total cross sections are not very sensitive to changes of these parameters .
then in nc production of photon , although we show the result up to @xmath125 order , there are no @xmath116 contributions from contact terms , and as we have checked already the @xmath125 contributions due to @xmath126 and @xmath127 , related to the so - called anomalous interactions , are tiny ( the same has been shown for nucleon scattering in @xcite ) . again , the total cross section of photon production is not sensitive to choice of different @xmath93 and @xmath94 . in all the plots ,
the @xmath1 contributions are singled out and compared with the full calculations .
moreover , we also compare our results with the output from nuance , and we find that the nuance output is close to our `` only @xmath1 '' calculation with @xmath89 for both pion and photon production , which should be expected since the @xmath1 dominates in nuance . in the calculation ,
the @xmath1 dynamics in nuclei is a key component .
the dynamics has been investigated in a nonrelativistic framework and also initiated in the qhd model .
parallel to the modification of the nucleon s spectrum , the @xmath1-meson couplings ( related to @xmath93 and @xmath94 ) introduced in our theory dictates the real part of the @xmath1 self - energy .
the couplings are used to explain the s - l coupling of @xmath1 .
meanwhile the phenomenological result about s - l coupling based on nonrelativsic isobar - hole models puts an interesting constraint on the @xmath1-meson coupling strengths , which is complementary to the constraints based on an eos consideration .
the @xmath1 width is treated in a simplified way , as we take advantage of the existing result that shows an increase of the width due to the opening of other decay channels . in pion electroproduction
, the pion - production ( without fsi ) result gives a correct prediction for the location of the @xmath1-peak .
we argue that this deficit is due to the absence of other channels . by adding contributions from two - body currents ( from other relativistic studies ) to our quasi - elastic and pion production ( and turning off @xmath1 broadening )
, we can explain the inclusive electron scattering strength .
the investigation on @xmath1 dynamics and two - body currents , which plays an important role in nuclear response and other problems , certainly needs to be pursued further in qhd eft .
moreover , because of the broadening of the @xmath1 width , we expect that in both pion and photon productions , the @xmath1 contribution is much less in nuclear scattering than in nucleon scattering .
but the reduction of nonresonant contributions would be less at higher energies ( beyond 0.5 gev ) , because the pauli blocking effect should be less important . in tabs .
[ tab : freeboundcc ] , [ tab : freeboundnc ] , and [ tab : freeboundphoton ] , we have shown explicitly the cross sections at @xmath120 gev due to @xmath1 and nonresonant contributions in both neutrino nucleon and neutrino nucleus scattering . although we see the reduction of nonresonant contributions for pion production in tabs .
[ tab : freeboundcc ] and [ tab : freeboundnc ] , we see a smaller reduction for photon production in tab . [
tab : freeboundphoton ] .
this is consistent with the picture that the nonresonant contribution is reduced because of pauli blocking .
the same situation occurs in antineutrino scattering .
this conclusion is important for future investigations of higher energy neutrino scattering , which may be relevant to miniboone s excess event problem .
since our calculation is based on a qhd eft lagrangian with all the relevant symmetries built in , conservation of vector current is manifest .
this is crucial for photon production .
also partial conservation of the axial current is a necessary constraint in the problem . by using the mean - field approximation and the lfg model ,
these constraints are satisfied in a transparent way .
we are currently working on coherent pion and photon production from nuclei by applying this qhd eft , which may also be relevant to the miniboone low energy excess event problem .
xz would like to thank t. william donnelly , gerald t. garvey , joe grange , charles j. horowitz , teppei katori , j. timothy londergan , william c. louis , rex tayloe , and geralyn zeller for their valuable information , useful discussions , and important comments on the manuscript .
this work was supported in part by the department of energy under contract no .
de fg0287er40365 .
the analysis of the kinematics is for scattering from nuclear matter , and can be easily generalized in the lfg model .
the kinematic variables are shown in fig .
[ fig : nucleilabframeconfiguration ] , and discussed following eq .
( [ eqn : quasitcrosssection ] ) . from the mean - field theory in qhd eft
, we know that the leading order hamiltonian gives rise to the nucleon spectrum in nuclear matter as @xmath128 , @xmath129 . then we can define @xmath130 .
this can be generalized from the laboratory frame to an arbitrary frame . in the lfg model
, we consider each neighborhood inside the nucleus as a homogeneous system ; the field expectations , @xmath131 and @xmath132 , are space - time dependent ( and in the laboratory frame , they only depend on the space coordinate ) . in the following , we always work in the nuclear laboratory frame .
the covariance of our calculation is more transparent with the @xmath133 variables than with @xmath134 .
for example energy momentum conservation is @xmath135 .
next we derive the formula for the total cross section .
suppose @xmath33 is the covariant interaction amplitude between the probe and each individual nucleon with specific initial and final states .
we have @xmath136 pauli blocking leads to constraints on the integration of @xmath137 and @xmath138 , i.e. @xmath139 and @xmath140 . here
@xmath141 is the fermi momentum related with the local density .
the two constraints can be expressed by using factors @xmath142 $ ] and @xmath143 $ ] . in the following
, we will not include them explicitly .
we know that @xmath144 by using this , we have the total cross section as @xmath145 meanwhile to make our phase space analysis simple , we can integrate over @xmath146 and @xmath147 : @xmath148 now , we need to calculate the boundary of the phase space in eq .
( [ eqn : xsectionquasi3 ] ) . from the lepton kinematics
, we can determine the boundary of @xmath149 ( @xmath150 ) : @xmath151 { { \vert \vec{q}_{}\vert}}_{min } & = & { \vert \vec{p}_{li}\vert } - \sqrt{e_{li}^{2}-m_{lf}^{2 } } \ .
\label{eqn : quasiphasespaceq2 } \end{aligned}\ ] ] for a given @xmath149 , we have the following constraints based on the lepton kinematics : @xmath152 q^{0 } & \geqslant & e_{li}-(e_{lf})_{max } = 0 \ .
\label{eqn : boundsonq02}\end{aligned}\ ] ] however , there are further constraints on @xmath92 for a given @xmath149 due to the hadron kinematics . for a given set of @xmath92 , @xmath149 , @xmath153 has to be physical .
this requires @xmath154 \longleftarrow & & \vert \vec{p}_{ni}^{\ast}\vert \geqslant \left| \frac{{\vert \vec{q}_{}\vert}}{2 } -\frac{q^{0}}{2}\sqrt{1-\frac{4{\mstar}^{2}}{q^{2 } } } \right| \equiv p^{- } \ .
\label{eqn : quasiphasespacepni1 } \end{aligned}\ ] ] eq .
( [ eqn : quasiphasespacepni1 ] ) gives a lower bound of @xmath155 which is also required to be below the fermi surface : @xmath156 .
combining @xmath157 and the constraints in eqs .
( [ eqn : boundsonq01 ] ) and ( [ eqn : boundsonq02 ] ) , we find @xmath158 \
, \label{eqn : quasiphasespaceq01 } \\[5pt ] q^{0}_{max } & = & \min \left[\sqrt{({\vert \vec{q}_{}\vert}+p_{f})^{2}+{\mstar}^{2 } } -e_{f } , e_{li}-\sqrt{({\vert \vec{p}_{li}\vert}-{\vert \vec{q}_{}\vert})^{2 } + m_{lf}^{2 } } \right ] \ .
\label{eqn : quasiphasespaceq02 } \end{aligned}\ ] ] moreover , the constraint @xmath159 is not present in the former discussion , but is taken care of in the numerical calculation .
the kinematic variables are defined in fig . [ fig : nucleipionprodlabframe ] in the laboratory frame . except for the @xmath160 momentum @xmath161 ,
all the others are defined in appendix [ app : quasikinematics ] .
the variables defined in other frames will be mentioned explicitly .
first we have @xmath62 & & \times ( 2\pi)^{4 } \delta^{4 } ( q+{p_{ni}^{\ast}}-{p_{nf}^{\ast}}-k_{\pi } ) \underset{s_{f},s_{i}}{\sum } \vert m_{fi } \vert^{2 } \ . \notag\end{aligned}\ ] ] the constraints on @xmath162 and @xmath163 , i. e. @xmath139 and @xmath140 , are always implicit in the formula . one way to think about the phase space as follows : given specific values for @xmath164 and @xmath165 , the final pion and nucleon invariant mass @xmath166 are fixed , and then the degrees of freedom in the isobaric frame ( final pion and nucleon s center - of - mass frame ) is the angle of @xmath167 , i.e. @xmath168 .
so we have @xmath169 & = & \int dm_{n\pi } d\omega_{\vec{k}_{i\pi } } \frac{1}{(2\pi)^{2 } } \frac{{\vert \vec{k}_{i\pi}\vert}}{2 } \delta[(q+p_{ni}^{\ast})^{2}-m_{\pi n}^{2 } ] \ .
\notag \end{aligned}\ ] ] in the above , we have made use of the following identities : @xmath170 e_{i\pi } & = & \frac{m_{\pi n}^{2}-{\mstar}^{2}+m_{\pi}^{2 } } { 2m_{\pi n } } \qquad e_{inf } = m_{\pi n}-e_{i \pi } \ , \notag \\[5pt ] \frac{d e_{i \pi}}{dm_{\pi n } } & = & \frac{e_{inf}}{m_{\pi n } } \ . \notag \end{aligned}\ ] ] then analogous to the analysis in the quasi - elastic scattering case , we have @xmath171 \notag \\[5pt ] & = & \int \frac{d p_{ni}^{\ast 0 } d \phi_{\vec{p}_{ni}^{\ast } } } { 4 { \vert \vec{q}_{}\vert}(2\pi)^{3 } } \vert _ { \cos(\angle\hat{q}\hat{p}_{ni}^{\ast } ) = ( 2q^{0}p_{ni}^{\ast 0 } + q^{2}+{\mstar}^{2 } - m_{\pi n}^{2})/2{\vert \vec{q}_{}\vert } \vert \vec{p}_{ni}^{\ast } \vert } \ .
\notag \end{aligned}\ ] ] so finally : @xmath172 & = & \int dv dq^{0 } d{\vert \vec{q}_{}\vert } dm_{\pi n } dp_{ni}^{\ast 0 } d\phi_{\vec{p}_{ni}^{\ast } } d\omega_{\vec{k}_{i\pi } } \frac{1}{(2\pi)^{7 } } \frac{{\vert \vec{k}_{i\pi}\vert}}{32 ( p_{li}^{0})^{2 } } \underset{s_{f},s_{i}}{\sum } \vert m_{fi } \vert^{2 } \ . \notag\end{aligned}\ ] ] next , we need to determine the boundary of phase space in terms of these variables .
first , it is clear that no constraint needs to be applied to @xmath168 and @xmath173 .
second , for a given set of @xmath174 , and @xmath166 , to make sure @xmath175 , there is a constraint on @xmath176 besides @xmath177 : @xmath178 & \longleftrightarrow & \begin{cases } ( \vert \vec{p}_{ni}^{\ast } \vert + \frac{\lambda+1}{2 } { \vert \vec{q}_{}\vert } ) ^{2 } + \frac{{\mstar}^{2 } ( q^{0})^{2}}{q^{2 } } -\frac{(\lambda+1)^{2}}{4 } ( q^{0})^{2 } \geqslant 0 \ , \qquad \lambda \equiv \frac{m_{\pi n}^{2}-{\mstar}^{2 } } { -q^{2 } } \ ; \notag \\[5pt ] q^{0 } \sqrt { \vert \vec{p}_{ni}^{\ast}\vert^{2 } + { \mstar}^{2 } } + { \vert \vec{q}_{}\vert } \vert \vec{p}_{ni}^{\ast } \vert \geqslant -\frac{q^{2}}{2 } + \frac{m_{\pi n}^{2}-{\mstar}^{2}}{2 } \ , \notag \end{cases } \notag \\[5pt ] & \longleftrightarrow & \vert \vec{p}_{ni}^{\ast } \vert \geqslant p^{- } \equiv \left| \frac{\lambda+1}{2 } { \vert \vec{q}_{}\vert } -\frac{q^{0}}{2 } \sqrt{(\lambda+1)^{2 } -\frac{4{\mstar}^{2}}{q^{2 } } } \right| \ .
\label{eqn : boundsoneni}\end{aligned}\ ] ] third , for a given set of @xmath174 , there is a constraint on @xmath166 , such that @xmath179 . from @xmath180 ,
we have @xmath181 and to open the pion production threshold , we need @xmath182 so , we have @xmath183 \leqslant m_{\pi n}^{2 } \leqslant q^{2 } & + & { \mstar}^{2 } + 2q^{0 } e_{f } + 2 { \vert \vec{q}_{}\vert } p_{f } \ .
\label{eqn : boundsonmpin}\end{aligned}\ ] ] fourth , for a given @xmath149 , there is a constraint on @xmath92 such that @xmath184 . we have @xmath185 \longleftrightarrow & & q^{0 } \geqslant \max ( \sqrt { ( m_{\pi}+\mstar)^{2 } + ( { \vert \vec{q}_{}\vert}-p_{f})^{2 } } -e_{f } , 0 ) \ .
\label{eqn : pionprodboundsonq01}\end{aligned}\ ] ] however , there are further constraints on @xmath92 due to the lepton kinematics , which has been shown in eqs .
( [ eqn : boundsonq01 ] ) and ( [ eqn : boundsonq02 ] ) . together with eq .
( [ eqn : pionprodboundsonq01 ] ) , we have the boundary of @xmath92 : @xmath186 q^{0}_{min } & = & \max ( \sqrt { ( m_{\pi}+\mstar)^{2 } + ( { \vert \vec{q}_{}\vert}-p_{f})^{2 } } -e_{f } , 0 ) \ . \label{eqn : pionprodboundsonq03}\end{aligned}\ ] ] eqs .
( [ eqn : quasiphasespaceq1 ] ) and ( [ eqn : quasiphasespaceq2 ] ) give constraints on @xmath149 . and
there are further constraints due to hadron kinematics .
we have to make sure that @xmath187 .
but it is complicated to obtain an analytic expression for @xmath149 based on this constraint . in the numerical calculations , we made use of another boundary by assuming a static nucleon in vacuum
. then , we can say the allowed region of @xmath149 is always inside the previous region .
solving @xmath187 with @xmath188 and @xmath189 gives us : @xmath190 { \vert \vec{q}_{}\vert } & \geqslant & \frac { \beta_{a } \frac { m_{a}^{2 } + ( m_{\pi}+m)^{2 } -m_{lf}^{2 } } { e_{li}+m } -\sqrt{\delta } } { 2(1-\beta_{a}^{2 } ) } \ .
\label{eqn : boundsonq2}\end{aligned}\ ] ] in the above , @xmath191 m_{a } & = & \sqrt{(e_{li}+m)^{2}-e_{li}^{2 } } \ , \notag \\[5pt ] \delta & = & \frac{[m_{a}^{2}+(m_{\pi}+m)^{2}-m_{lf}^{2}]^{2 } } { ( e_{li}+m)^{2 } } -4(1-\beta_{a}^{2})(m_{\pi}+m)^{2 } \ . \notag\end{aligned}\ ] ] so , eqs .
( [ eqn : quasiphasespaceq1 ] ) , ( [ eqn : quasiphasespaceq2 ] ) , ( [ eqn : boundsonq1 ] ) and ( [ eqn : boundsonq2 ] ) are the bounds used in the numerical calculations . and to map out the physical region , we simply try and check .
it is also complicated to determine an analytic expression for the threshold value of @xmath192 in the lfg model .
however , it is simpler to work out the value for pion production off a static nucleon , which is @xmath193 so , the difference between the true threshold and the value calculated above is essentially the binding energy .
b. d. serot and x. zhang , in _ advances in quantum field theory _ , edited by sergey ketov ( intech , rijeka , croatia , 2012 ) , chap .
b. d. serot and x. zhang , phys .
c * 86 * 015501 ( 2012)[arxiv:1206.3812 ] . b. d. serot and j. d. walecka , adv .
. phys .
* 16 * , ( 1986 ) 1 .
b. d. serot and j. d. walecka , int .
j. mod .
e * 6 * , ( 1997 ) 515 .
r. j. furnstahl , b. d. serot , and h .- b .
tang , nucl .
a * 615 * , ( 1997 ) 441 ; * 640 * , ( 1998 ) 505 ( e ) .
r. j. furnstahl and b. d. serot , nucl .
a * 671 * , ( 2000 ) 447 .
r. j. furnstahl and b. d. serot , nucl .
a * 673 * , ( 2000 ) 298 .
r. j. furnstahl and b. d. serot , comments mod .
* 2 * , ( 2000 ) a23 .
b. d. serot , lecture notes in physics * 641 * , g. a. lalazissis , p. ring , and d. vretenar , eds .
( springer , berlin heidelberg , 2004 ) , p. 31 .
b. d. serot , ann .
( ny ) * 322 * , ( 2007 ) 2811 .
m. a. huertas , phys .
c * 66 * , 024318 ( 2002 ) ; * 67 * , ( 2003 ) 019901 ( e ) .
m. a. huertas , acta phys .
b * 34 * , ( 2003 ) 4269 .
m. a. huertas , acta phys .
polon . b * 35 * , ( 2004 ) 837 .
j. mcintire , acta phys .
b * 35 * , ( 2004 ) 2261 .
j. mcintire , arxiv : nucl - th/0507006 .
j. d. walecka , in _ theoretical nuclear and subnuclear physics _ , 2nd ed .
( world scientific , singapore , 2004 ) , chap . 24 .
j. mcintire , y. hu , and b. d. serot , nucl .
a * 794 * , ( 2007 ) 166 .
y. hu , j. mcintire , and b. d. serot , nucl .
a * 794 * , ( 2007 ) 187 .
j. mcintire , ann .
( ny ) * 323 * , ( 2008 ) 1460
. b. d. serot , phys .
c * 81 * , ( 2010 ) 034305 .
a. a. aguilar - arevalo _ et al .
_ , ( miniboone collaboration ) , phys . rev
. lett .
* 98 * , ( 2007 ) 231801 .
a. a. aguilar - arevalo _ et al .
_ , ( miniboone collaboration ) , phys . rev
. lett . * 102 * , ( 2009 ) 101802 .
a. a. aguilar - arevalo _ et al .
_ , ( miniboone collaboration ) , phys . rev
. lett . *
105 * , ( 2010 ) 181801 .
t. katori ( for the microboone collaboration ) , aip conf .
proc . * 1405 * , ( 2011 ) 250 .
j. a. harvey , c. t. hill , and r. j. hill , phys .
* 99 * , ( 2007 ) 261601 .
j. a. harvey , c. t. hill , and r. j. hill , phys .
d * 77 * , ( 2008 ) 085017 .
richard j. hill , phys .
d * 81 * , ( 2010 ) 013008 .
s. s. gershtein , yu . ya .
komachenko , and m. yu . khlopov , sov .
j. nucl .
phys . * 33 * , ( 1981 ) 860 .
m. martini , m. ericson , g. chanfray and j. marteau , phys .
d * 80 * , ( 2009 ) 065501 .
s. nakayama _
et al . _ ,
( k2k collaboration ) , phys .
b * 619 * , ( 2005 ) 255 .
m. hasegawa _
_ , ( k2k collaboration ) , phys . rev
. lett . * 95 * , ( 2005 ) 252301 .
et al . _ ,
( k2k collaboration ) , phys .
d * 74 * , ( 2006 ) 052002 .
a. rodriguez _
et al . _ ,
( k2k collaboration ) , phys .
d * 78 * , ( 2008 ) 032003 .
k. hiraide _
_ , ( sciboone collaboration ) , phys . rev .
d * 78 * , ( 2008 ) 112004 .
a. a. aguilar - arevalo _ et al .
_ , ( miniboone collaboration ) , phys . rev .
100 * , ( 2008 ) 032301 .
a. a. aguilar - arevalo _ et al .
_ , ( miniboone collaboration ) , phys .
b * 664 * , ( 2008 ) 41 .
a. a. aguilar - arevalo _ et al .
_ , ( miniboone collaboration ) , phys . rev
. lett .
* 103 * , ( 2009 ) 081801 .
a. v. butkevich , phys .
c * 78 * , ( 2008 ) 015501 .
a. v. butkevich , phys .
c * 80 * , ( 2009 ) 014610 .
s. k. singh and e. oset , nucl .
a * 542 * , ( 1992 ) 587 .
s. k. singh and e. oset , phys .
c * 48 * , ( 1993 ) 1246 .
a. meucci , c. giusti , and f. d. pacati , nucl .
a * 739 * , ( 2004 ) 277 .
a. meucci , c. giusti , and f. d. pacati , nucl .
a * 744 * , ( 2004 ) 307 .
j. nieves , j. e. amaro , and m. valverde , phys .
c * 70 * , ( 2004 ) 055503 ; * 72 * , ( 2005 ) 019902 ( e ) .
m. c. martinez _ et al .
_ , phys . rev .
c * 73 * , ( 2006 ) 024607
. t. leitner , l. alvarez - ruso , and u. mosel , phys .
c * 73 * , ( 2006 ) 065502
. t. leitner , l. alvarez - ruso , and u. mosel , phys .
c * 74 * , ( 2006 ) 065502 .
t. leitner , o. buss , l. alvarez - ruso , and u. mosel , phys .
c * 79 * , ( 2009 ) 034601 .
a. kartavtsev , e. a. paschos , and g. j. gounaris , phys .
d * 74 * , ( 2006 ) 054007 . c. praet , o. lalakulich , n. jachowicz , and j. ryckebusch , phys .
c * 79 * , ( 2009 ) 044603 .
t. leitner , u. mosel and s. winkelmann , phys .
c * 79 * , ( 2009 ) 057601
. s. k. singh , m. j. vicente - vacas , and e. oset , phys .
b * 416 * , ( 1998 ) 23 ; * 423 * , ( 1998 ) 428 ( e ) .
t. sato , d. uno , and t. s. h. lee , phys .
c * 67 * , ( 2003 ) 065201 .
b. szczerbinska , t. sato , k. kubodera , and t. s. lee , phys .
b * 649 * , ( 2007 ) 132 .
j. e. amaro , m. b. barbaro , j. a. caballero , t. w. donnelly , a. molinari , and i. sick , phys .
c * 71 * , ( 2005 ) 015501
. j. a. caballero , j. e. amaro , m. b. barbaro , t. w. donnelly , c. maieron , and j. m. udias , phys . rev
. lett . * 95 * , ( 2005 ) 252502 .
j. e. amaro , m. b. barbaro , j. a. caballero , and t. w. donnelly , phys .
lett . * 98 * , ( 2007 ) 242501 . m. martini , g. co , m. anguiano , and a. m. lallena , phys . rev .
c * 75 * , ( 2007 ) 034604
. j. e. amaro , m. b. barbaro , j. a. caballero , t. w. donnelly , and j. m. udias , phys . rev .
c * 75 * , ( 2007 ) 034613
. m. v. ivanov , m. b. barbaro , j. a. caballero , a. n. antonov , e. moya de guerra , and m. k. gaidarov , phys .
c * 77 * , ( 2008 ) 034612 .
a. gil , j. nieves and e. oset , nucl .
a * 627 * , ( 1997 ) 543 .
w. m. alberico , m. ericson , and a. molinari , ann .
( ny ) * 154 * , ( 1984 ) 356 .
e. oset , l.l .
salcedo , nucl .
a * 468 * , ( 1987 ) 631 .
r. rosenfelder , ann .
( ny ) * 128 * , ( 1980 ) 188 .
k. wehrgerger , c. bedau , and f. beck , nucl .
a * 504 * , ( 1989 ) 797 .
k. wehrgerger , and r. wittman , nucl .
a * 513 * , ( 1990 ) 603 .
t. herbert , k. wehrgerger , and f. beck , nucl .
a * 541 * , ( 1992 ) 699 .
k. wehrgerger , phys .
rept . * 225 * , ( 1993 ) 273 . m. j. dekker , p. j. brussaard , and j. a. tjon , phys .
c * 49 * , ( 1994 ) 2650 .
a. de pace , m. nardi , w. m. alberico , t. w. donnelly , and a. molinari , nucl .
a * 726 * , ( 2003 ) 303 .
s. x. nakamura , t. sato , t .- s .
h. lee , b. szczerbinska , and k. kubodera , phys .
c * 81 * , ( 2010 ) 035502 . y. horikawa , m. thies , and f. lenz , nucl . phys . a * 345 * , ( 1980 ) 386
. s. l. adler , s. nussinov and e. a. paschos , phys .
d * 9 * , ( 1974 ) 2125 .
paschos , l. pasquali and j.y.yu , nucl .
b * 588 * , ( 2000 ) 263 .
d. casper , nucl .
suppl . *
112 * , ( 2002 ) 161 .
a. a. aguilar - arevalo _ et al .
_ , ( miniboone collaboration ) , phys . rev .
d * 81 * , ( 2010 ) 092005 .
_ et al_. , nucl .
a * 402 * , ( 1983 ) 515 .
j. s. oconnell _ et al_. , phys .
c * 35 * , ( 1987 ) 1063 .
m. hirata , j. h. koch , f. lenz and e. j. moniz , ann .
( ny ) * 99 * , ( 1976 ) 374 .
j. boguta , phys .
b , * 109 * , ( 1982 ) 251 .
d. s. kosov , c. fuchs , b. v. martemyanov , amand faessler , phys .
b * 421 * , ( 1998 ) 37 .
r. j. furnstahl , john j. rusnaka and b. d. serot , nucl .
a * 632 * , ( 1998 ) 607 .
i. j. d. macgregor _
et al . _ ,
lett . * 80 * , ( 1998 ) 245 .
j. nieves , e. oset , and c. garcia - recio , nucl .
a * 554 * , ( 1993 ) 554 . | we study the incoherent neutrinoproduction of photons and pions with neutrino energy @xmath0 .
these processes are relevant to the background analysis in neutrino - oscillation experiments [ for example , miniboone ; a. a. aquilar - arevalo _ et al . _
( miniboone collaboration ) , phys .
rev .
lett . * 100 * , 032301 ( 2008 ) ] .
the calculations are carried out using a lorentz - covariant effective field theory ( eft ) , which contains nucleons , pions , the delta ( 1232 ) ( @xmath1 ) , isoscalar scalar ( @xmath2 ) and vector ( @xmath3 ) fields , and isovector vector ( @xmath4 ) fields , and has @xmath5 chiral symmetry realized nonlinearly .
the contributions of one - body currents are studied in the local fermi gas approximation .
the current form factors are generated by meson dominance in the eft lagrangian .
the conservation of the vector current and the partial conservation of the axial current are satisfied automatically , which is crucial for photon production .
the @xmath1 dynamics in nuclei , as a key component in the study , is explored .
introduced @xmath1-meson couplings explain the @xmath1 spin - orbit coupling in nuclei , and this leads to interesting constraints on the theory .
meanwhile a phenomenological approach is applied to parametrize the @xmath1 width .
to benchmark our approximations , we calculate the differential cross sections for quasi - elastic scattering and incoherent electroproduction of pions without a final state interaction ( fsi ) .
the fsi can be ignored for photon production . |
long - term numerical integrations play an important role in our understanding of the dynamical evolution of many astrophysical systems ( see , e.g. , duncan , these proceedings , for a review of solar - system integrations ) .
an essential tool for long - term integrations is a fast and accurate integration algorithm .
symplectic integration algorithms ( sias ) have become popular in recent years because the newtonian gravitational @xmath0-body problem is a hamiltonian problem and sias enforce certain conservation laws that are intrinsic to hamiltonian systems ( see sanz - serna & calvo 1994 for a general introduction to sias ) . for an autonomous hamiltonian system ,
the equations of motion are @xmath1 where @xmath2 is the explicitly time - independent hamiltonian , @xmath3 are the @xmath4 canonical phase - space coordinates , @xmath5 is the poisson bracket , and @xmath6(@xmath7 ) is the number of degrees of freedom .
the formal solution of eq.([heq ] ) is @xmath8 if the hamiltonian @xmath9 has the form @xmath10 , where @xmath11 and @xmath12 are separately integrable , we can devise a sia of constant timestep @xmath13 by approximating @xmath14 as a composition of terms like @xmath15 and @xmath16 .
for example , a second - order sia is @xmath17 for the gravitational @xmath0-body problem , we can write @xmath18 , where @xmath19 and @xmath20 are the kinetic and potential energies . then the second - order sia eq.([sia2 ] ) becomes @xmath21 where @xmath22 and @xmath23 ; this is the familiar leapfrog integrator . for solar - system type integrations ,
a central body ( the sun ) is much more massive than the other bodies in the system , and it is better to write @xmath24 , where @xmath25 is the part of the hamiltonian that describes the keplerian motion of the bodies around the central body and @xmath26 is the part that describes the perturbation of the bodies on one another .
symplectic integrators using this decomposition of the hamiltonian were introduced by wisdom & holman ( 1991 ) , and they are commonly called mixed variable symplectic ( mvs ) integrators .
the constant timestep sias have the following desirable properties : \(1 ) as their names imply , sias are symplectic , i.e. , they preserve @xmath27 .
\(2 ) for sufficiently small @xmath13 , sias solve almost exactly a nearby `` surrogate '' autonomous hamiltonian problem with @xmath28 .
for example , the second - order sia in eq.([sia2 ] ) has @xmath29 consequently , we expect that the energy error is bounded and the position ( or phase ) error grows linearly .
\(3 ) many sias ( e.g. , eq.[[sia2 ] ] ) are time reversible .
note , however , that there are algorithms that are symplectic but not reversible .
1 shows the energy error @xmath30 of an integration of the @xmath31 kepler orbit using the constant timestep leapfrog integrator eq.([lf ] ) .
( in this and all subsequent figures , the orbits are initially at the apocenter @xmath32 , where @xmath33 and @xmath34 are the semi - major axis and eccentricity . )
1 illustrates that there is no secular drift in @xmath30 .
for problems with large variations in timescale ( due to close encounters or high eccentricities ) , it is desirable to use a variable timestep . a common practice is to set the timestep using the phase - space coordinates @xmath35 at the beginning of the timestep : @xmath36 . if the sias discussed in 1
are implemented with this simple variable timestep scheme , they are still symplectic if we assume that the sequence of timesteps determined for a particular initial condition are also used to integrate neighboring initial conditions ( see skeel & gear 1992 for another point of view ) .
however , tests have shown that this and similar simple variable timestep schemes [ and also @xmath37 destroy the desirable properties of the integrators ( e.g. , gladman , duncan , & candy 1991 ; calvo & sanz - serna 1993 ) . in fig .
2_a _ we show the energy error of an integration with @xmath31 and @xmath38 .
although the error is initially smaller than that in fig . 1 ( the integrations shown in figs . 1 and 2 use nearly the same number of timesteps per orbit ) , it shows a linear drift . in fig.2_a
_ we also show the errors for two neighboring initial conditions ( @xmath39 ) .
they were integrated using the timesteps determined for the @xmath31 orbit .
note that these errors grow even faster .
the degradation in performance is due to the fact that the properties ( 2 ) and ( 3 ) listed in 1 are no longer true .
the algorithm is not time reversible because the timestep depends only on the coordinates at the beginning of the timestep . since @xmath40 depends explicitly on the timestep @xmath13 ( see , e.g. , eq.[[herr ] ] ) , a variable timestep changes the surrogate hamiltonian problem that the integrator is solving from step to step .
thus the solution from @xmath41 to @xmath42 after @xmath43 steps is not in general the solution of a nearby autonomous hamiltonian problem .
hut , makino , & mcmillan ( 1995 ; see also funato et al .
1996 ; hut , these proceedings ) pointed out that the performance of a variable timestep integrator can be improved if time reversibility is restored by choosing the timestep in a time symmetric manner : @xmath44 with @xmath45 .
for example , @xmath46/2 $ ] .
we have tested this idea in detail by ( i ) integrating a series of problems ( the pendulum , kepler orbits , and the restricted @xmath47-body problem ) using a symmetrized variable timestep leapfrog integrator and ( ii ) integrating the restricted @xmath47-body problem using a symmetrized mvs integrator .
we have found that time symmetrization usually reduces the drift in energy ( or the jacobi constant ) to negligible levels .
2_b _ shows an integration of the @xmath31 kepler orbit using the symmetrized leapfrog integrator , and it should be compared to fig .
2_a_. in fig
. 2_b _ we also show the errors for two neighboring initial conditions that were integrated using the timesteps determined for the @xmath31 orbit .
there is no improvement in the error growth for the neighboring initial conditions .
integrating an orbit using the timesteps determined for another orbit may seem to be somewhat artificial , but similar situations do occur in realistic integrations .
for example , if we integrate an @xmath0-body system using a shared timestep scheme , the timestep used by all of the particles is determined by the few having the strongest close encounter . in practice
, since the symmetrized timestep criterion depends on @xmath48 , a symmetrized integrator also has the disadvantage that it requires iteration and can be significantly slower than the original integrator ( unless the original integrator is also implicit ) .
in this section we describe a `` variable timestep '' sia that has all the desirable properties of the constant timestep sias . the algorithm is based on ideas proposed by skeel & biesiadecki ( 1994 ; see also macevoy & scovel 1994 ) .
it is also similar to the individual timestep scheme of saha & tremaine ( 1994 ) . for simplicity ,
let us consider in particular the kepler problem with @xmath49 and @xmath50 .
we choose a set of cutoff radii @xmath51 and decompose the potential @xmath52 into @xmath53 , or equivalently the force @xmath54 into @xmath55 , such that ( i ) @xmath56 , ( ii ) @xmath57 ( except @xmath58 ) is zero at @xmath59 , and ( iii ) @xmath57 is `` softer '' than @xmath60 .
the force @xmath57 is to be applied with a timestep @xmath61 .
if we assume that @xmath62 is an integer , we can apply the second - order sia in eq.([sia2 ] ) recursively to obtain the following second - order algorithm : @xmath63\ , \exp({\tau_0 \over 2 } k_0 ) \nonumber \\ & \approx&\!\ ! \exp({\tau_0 \over 2 } k_0)\ , \biggl [ \exp({\tau_1 \over 2 } k_1)\ , \exp[\tau_1 ( d + k_2 + k_3 + \cdots ) ] \nonumber \\ & & \quad \times \exp({\tau_1 \over 2 } k_1)\biggr]^{m_1}\ , \exp({\tau_0 \over 2 } k_0 ) \label{vsia2 } \\ & \vdots & \nonumber\end{aligned}\ ] ] where @xmath64 and @xmath65 .
hereafter , we adopt @xmath66 . the multiple timescale algorithm eq.([vsia2 ] )
has an overall timestep @xmath67 , but it is effectively a variable timestep scheme because the recursion terminates at level @xmath68 if @xmath69 during a substep of length @xmath61 .
for example , if the particle is in the region @xmath70 during an overall step , eq.([vsia2 ] ) reduces to @xmath71 since eq.([vsia2 ] ) is based on the recursive application of eq.([sia2 ] ) , it has all the desirable properties of a constant timestep sia .
it is obviously symplectic and time reversible
. we can also derive the surrogate autonomous hamiltonian solved by this integrator .
for example , if we decompose @xmath52 into two levels @xmath72 and @xmath73 only , the error hamiltonian is ( cf .
eq.[[herr ] ] ) @xmath74 after some experiments , we found two force decompositions that work well .
if we write @xmath75 and @xmath76 for @xmath77 , one of the decompositions is @xmath78 \vecr/4 r_i^3 & \mbox{if $ r < r_i$. } \end{array}\right.\ ] ] an alternative decomposition uses @xmath79 where @xmath80 . unlike the forces suggested by skeel & biesiadecki ( 1994 ) , these forces have continuous first derivatives and decrease rapidly ( or exactly ) to zero at @xmath81 ( see fig .
we shall not provide the details here , but we can understand why these properties are important from an analysis of @xmath40 . in fig .
4 we show the energy error @xmath30 of two integrations using the multiple timescale symplectic integrator with @xmath82 . as expected , there is no secular drift in @xmath30 .
note also that , with the chosen integration parameters ( @xmath83 ) , the maximum error is almost independent of the pericentric distance ( which changes by @xmath84 in the two cases shown ) .
this again agrees with the expectation from an analysis of @xmath40 .
one of the goals of this study is to develop a variable timestep integrator for solar system integrations .
we have developed a second - order multiple timescale mvs integrator based on the algorithm described in this section ( see levison & duncan 1994 for another approach ) .
we are currently testing this integrator in detail .
initial results indicate that the integrator is fast and accurate and has all the desirable properties of the constant timestep symplectic integrators .
calvo , m. p. , & sanz - serna , j. m. 1993 , siam j. sci .
, 14 , 936 funato , y. , hut , p. , mcmillan , s. , & makino , j. 1996 , , 112 , 1697 gladman , b. , duncan , m. , & candy , j. 1991 , celest .
astr . , 52 , 221 hut , p. , makino , j. , & mcmillan , s. 1995 , , 443 , l93 levison , h. f. , & duncan , m. j. 1994 , icarus , 108 , 18 macevoy , w. , & scovel , j. c. 1994 , preprint saha , p. , & tremaine , s. 1994 , , 108 , 1962 sanz - serna , j. m. , & calvo , m. p. 1994
, numerical hamiltonian problems ( london : chapman & hall ) skeel , r. d. , & biesiadecki , j. j. 1994 , ann . numer . math . , 1 , 191 skeel , r. d. , & gear , c. w. 1992 , physica d , 60 , 311 wisdom , j. , & holman , m. 1991 , , 102 , 1528 | symplectic integration algorithms have become popular in recent years in long - term orbital integrations because these algorithms enforce certain conservation laws that are intrinsic to hamiltonian systems . for problems with large variations in timescale ,
it is desirable to use a variable timestep .
however , naively varying the timestep destroys the desirable properties of symplectic integrators .
we discuss briefly the idea that choosing the timestep in a time symmetric manner can improve the performance of variable timestep integrators
. then we present a symplectic integrator which is based on decomposing the force into components and applying the component forces with different timesteps .
this multiple timescale symplectic integrator has all the desirable properties of the constant timestep symplectic integrators . |
the last decade has seen the coming of age of extremely high multiplex fibre spectroscopy , as implemented by the 2-degree field ( 2df ) instrument @xcite and the sloan digital sky survey ( sdss ; york et al .
these new facilities have allowed order of magnitude increases in sample sizes over the previous generation of surveys . the 2df qso redshift survey ( 2qz ; croom et al . 2001a ; 2004 ) and the sdss qso survey ( schneider et al . 2007 ) have allowed precise measurement of the evolution of qsos ( e.g. boyle et al . 2000 ; croom et al .
2004 ; richards et al . 2006 ) , qso clustering ( e.g. croom et al . 2001b ; croom et al . 2005 ; shen et al .
2007 ) , spectral properties ( e.g. croom et al . 2002 ; corbett et al . 2003 ; vanden berk et al . 2001
; richards et al .
2002a ) and a range of other significant results .
the published sample sizes ( @xmath10 qsos in 2qz ; @xmath11 qsos in sdss ) are large enough that in many cases measurements are now limited by systematic uncertainties rather than random errors .
however , one of the important limitations of the 2qz and sdss surveys are their relative depths . the sdss qso survey is limited to @xmath12 , or @xmath13 for the high redshift sample @xcite , which does not reach the break in the qso luminosity function .
2qz is somewhat deeper , limited in the bluer @xmath14-band to @xmath15 .
the 2qz clearly shows the break in the qso luminosity function ( lf ) , typically reaching @xmath16 mag fainter than the break at @xmath17 .
the observed break in the lf is a gradual flattening towards faint magnitudes ; as a result the constraints from the 2qz on the actual slope of the faint end are fairly uncertain , as evidenced by the difference between the results from the first release ( boyle et al . 2000 ) and the final release ( croom et al .
2004 ; c04 ) . in comparison ,
x - ray surveys , in particular using _ chandra _
( e.g. giacconi et al .
2002 ; alexander et al .
2003 ) and _ xmm - newton _
( e.g. hasinger et al . 2001 ; worsley et al . 2004
; barcons et al . 2007 ) , reach to fainter depths , but over a much smaller area .
the largest samples contain @xmath18 objects over a few square degrees .
these surveys have demonstrated that the pure luminosity evolution that appears to model the evolution of the most extensive optical samples ( e.g. 2qz , sdss ) fails to trace the evolution of the faint agn populations at @xmath19 .
it now appears that the activity in faint agn peaks at a lower redshift than that of more luminous agn ( e.g. hasinger et al 2005 ) ; this process has been described as agn downsizing ( e.g. barger et al . 2005 ) . whether the downsizing is due to lower mass black holes being more active at low redshift ( e.g. heckman et al . 2004 ) or massive black holes at lower rates of accretion ( e.g. babic et al . , 2007 )
remains unclear .
both effects are likely to play a role .
substantial advances have been made in the theoretical understanding of agn formation and the connection to galaxy formation ( e.g. hopkins et al .
this work has been largely driven by the observational evidence that most massive galaxies with bulges contain super - massive bhs ( smbhs ) ( e.g. tremaine et al . 2002 ) .
smbh accretion is thought to be triggered ( at least for moderate to high luminosity agn ) by the merger of gas rich galaxies ; while the timescale for the merger may be as long as @xmath16gyr , during the majority of this time the accretion is obscured from view .
it is only when the agn finally expels the surrounding gas and dust that it shines like a quasar for a brief period ( @xmath20myr ) , before exhausting its fuel supply ( e.g. di matteo et al . 2005 ) .
this feedback of the agn into the host also heats ( and possibly expels ) the gas in the galaxy , which suppresses star formation leading to `` red and dead '' ellipticals or bulges .
these models match a number of previous observations and predict that the faint end of the qso luminosity function is largely comprised of higher mass bhs at lower accretion rates ( i.e. below their peak luminosity ) @xcite .
the 2df sdss lrg and qso ( 2slaq ) survey was designed to survey optically faint agn / qsos within a sufficiently large volume to obtain robust measurements of both the luminosity function and qso clustering . throughout this paper
we will use the term qso to refer to any broad line ( type 1 ) agn , irrespective of luminosity .
the qso portion of the survey shared fibres with a related program to target luminous red galaxies ( lrgs ) at @xmath21 ( cannon et al .
both the lrgs and qsos were selected from single epoch sdss imaging data , and then observed spectroscopically with the 2-degree field ( 2df ) instrument at the anglo - australian telescope ( aat ) .
the 2slaq qso sample has already produced a preliminary qso luminosity function ( richards et al .
2005 ; r05 ) , measured the clustering of qsos as a function of luminosity ( da angela et al .
2008 ) and studied the distribution of qso broad line widths ( fine et al .
2008 ) . in this paper
we present the final spectroscopic qso catalogue of the 2slaq sample .
we then carry out a detailed analysis of the survey completeness .
the analysis of the qso luminosity function from the final 2slaq sample is presented in a companion paper ( croom et al . in prep ) . in section [ sec : imaging ] we discuss the selection of qso candidates from the sdss imaging data .
this has largely been described by r05 , but is summarized here for completeness . in section [ sec : spec ]
we present the spectroscopic observations , and in section [ sec : cat ] we describe the composition and quality of the resulting catalogue .
section [ sec : comp ] contains our detailed completeness analysis .
we summarize our results in section [ sec : sum ] . throughout this paper
we will assume a cosmology with @xmath22 , @xmath23 and @xmath24 .
the photometric measurements used as the basis for our catalogue are drawn from the data release 1 ( dr1 ) processing ( stoughton et al . 2002 ; abazajian et al .
2003 ) of the sdss imaging data .
the astrometric precision at the faint limit of the survey is @xmath25 arcsec @xcite .
the sdss data are taken in five photometric pass - bands ( @xmath6 ; fukugita et al .
1996 ) using a large format ccd camera @xcite on a special - purpose 2.5-m telescope ( gunn et al .
the regions covered by the 2slaq survey were complete in dr1 , so no further updates to more recent data releases are required . except where otherwise stated , all sdss magnitudes discussed herein are `` asinh '' point - spread - function ( psf ) magnitudes @xcite on the sdss pseudo ab magnitude system @xcite that have been dereddened for galactic extinction according to the model of schlegel , finkbeiner & davis ( 1998 ) .
the sdss quasar survey @xcite extends to @xmath12 for @xmath26 and @xmath13 for @xmath27 ; our work herein explores the @xmath26 regime to @xmath28 ( @xmath29 , based on the median colours of sdss qsos , e.g. richards et al .
2001 ) . at the faint limit of the 2slaq sample ( @xmath30 ) ,
the photometric errors are typically @xmath31 , @xmath32 , @xmath33 , @xmath34 , @xmath35 .
our quasar candidate sample was drawn from 10 sdss imaging runs .
due to the slightly poorer image quality in the 6th row of ccds in the sdss camera we did not include these data .
thus the 2slaq survey regions are @xmath36 wide rather than the usual @xmath37 for an sdss imaging run .
we rejected any objects that met the `` fatal '' or `` non - fatal '' error definitions of the sdss quasar target selection @xcite .
although our survey covers the southern equatorial stripe 82 region which has been scanned multiple times @xcite , the co - added data ( annis et al .
2006 ) were not available at the time of our spectroscopic observations and so single scan data were used .
we apply a limit to the ( extinction corrected ) @xmath38-band psf magnitude of @xmath39 and @xmath40 .
we also placed restrictions on the errors in each of the other four bands : @xmath41 , @xmath42 , @xmath43 and @xmath44 .
note that this selection of error constraints effectively limits the redshift to less than 3 , as the ly@xmath45 forest suppresses the @xmath46 flux at higher redshifts .
based on spectroscopic identifications from sdss and 2qz of this initial set of objects , we implement additional colour cuts that are designed to select faint uv - excess qsos with high efficiency and completeness at redshifts @xmath1 .
an analysis of the completeness of the selection algorithm is given as a function of redshift and magnitude in section [ sec : photcomp ] .
we reject hot white dwarfs using the following cuts , independent of magnitude .
specifically , we rejected objects that satisfy the condition : a and ( ( b and c and d ) or e ) , where the letters refer to the cuts : @xmath47 these constraints are similar to the white dwarf cut applied by richards et al .
( 2002a ; their eq .
2 ) except for the added cut with respect to the @xmath7 colour . to efficiently target both bright and faint targets we use different colour cuts as a function of @xmath8-band magnitude .
the bright sample is restricted to @xmath48 and is designed to allow for overlap with previous sdss and 2df spectroscopic observations .
the faint sample has @xmath49 and probes roughly one magnitude deeper than 2qz .
these cuts are made in @xmath8 , rather than the @xmath38-band that the sdss quasar survey uses , since we are concentrating on uv - excess quasars and would like to facilitate comparison with the results from the @xmath14-based 2qz .
at this depth an @xmath38-band limited sample selected from single epoch sdss data would also contain substantial stellar contamination .
the combination of the @xmath50 and @xmath39 cuts will exclude objects bluer than @xmath51 ; however , objects this blue are exceedingly rare ( @xmath52 deviations from the observed qso spectral slope distribution ; richards et al 2006 ) . as pointed out by r05 ,
the @xmath14 and @xmath8 bands are almost equivalent , with a mean @xmath53 found for qsos in common between the sdss and 2qz .
we note that the relative transmission curves plotted in r05 ( their fig .
6 ) were in error ( the energy to photon conversion was reversed ) .
we plot the correct comparison in fig .
[ fig : bjg ] . in general , we would prefer to avoid a morphology - based cut since we do not want to exclude low-@xmath54 quasars and because our selection extends beyond the magnitude limits at which the sdss s star / galaxy separation breaks down .
however , scranton et al . ( 2002 ) have developed a bayesian star - galaxy classifier that is robust to @xmath55 . as a result ,
in addition to straight colour - cuts , we also apply some colour restrictions on objects with high @xmath56-band galaxy probability ( referred to below as `` galprob '' ) according to scranton et al .
( 2002 ) in an attempt to remove contamination from narrow emission line galaxies ( nelgs ; i.e. blue star - forming galaxies ) from our target list .
bright sample objects are those with @xmath48 and that meet the following conditions @xmath57 in the combination a and ( not b ) and ( not c ) and ( not d ) and ( not e ) , where cut a selects uvx objects , cuts b and c eliminate faint f - stars whose metallicity and errors push them blueward into the quasar regime , and cuts d and e remove nelgs that appear extended in the @xmath56 band . among the bright sample objects , those with @xmath58 were given priority in terms of fibre assignment ( see section [ sec : targpri ] ) .
faint sample objects are those with @xmath49 and that meet the following conditions @xmath59 in the combination a and ( not b ) and ( not c ) and ( not d ) and ( not e ) , where cut a selects uvx objects , cuts b , c and d eliminate faint f - stars whose metallicity and errors push them blueward into the quasar regime , and cut e removes nelgs .
these faint cuts are more restrictive than the bright cuts to avoid significant contamination from main sequence stars that will enter the sample as a result of larger errors at fainter magnitudes .
the low redshift colour cuts ( @xmath60 and @xmath7 ) are shown in fig .
[ fig : ugi_gz ] ( see also fig . 1 of r05 ) .
in addition to the main low redshift ( @xmath1 ) sample described above , we also target a sample of higher redshift qso candidates , analogous to the high redshift sample selected in the main sdss qso survey which selected qsos up to @xmath61 at @xmath62 ( richards et al .
2002a ) . the
2slaq high redshift sample was limited to @xmath63 and an additional constraint that @xmath64 was applied to the @xmath54-band photometry .
we then selected candidates in three redshift intervals .
qso candidates at redshift @xmath65 satisfied the following cuts : @xmath66 for the redshift range @xmath67 this selection becomes @xmath68 in the combination a and b and c and d and e. for the redshifts above @xmath69 we use @xmath70 these samples have a high degree of contamination from the stellar locus due to photometric errors .
these candidates were therefore targeted at a lower priority than the main low redshift sample , and we do not present a detailed analysis of completeness for the high redshift sample . | we present the final spectroscopic qso catalogue from the 2df - sdss lrg and qso ( 2slaq ) survey .
this is a deep , @xmath0 ( extinction corrected ) , sample aimed at probing in detail the faint end of the broad line agn luminosity distribution at @xmath1 .
the candidate qsos were selected from sdss photometry and observed spectroscopically with the 2df spectrograph on the anglo - australian telescope .
this sample covers an area of 191.9 deg@xmath2 and contains new spectra of 16326 objects , of which 8764 are qsos , and 7623 are newly discovered ( the remainder were previously identified by the 2qz and sdss surveys ) . the full qso sample ( including objects previously observed in the sdss and 2qz surveys ) contains 12702 qsos . the new 2slaq spectroscopic data set also contains 2343 galactic stars , including 362 white dwarfs , and 2924 narrow emission line galaxies with a median redshift of @xmath3 .
we present detailed completeness estimates for the survey , based on modelling of qso colours , including host galaxy contributions .
this calculation shows that at @xmath4 qso colours are significantly affected by the presence of a host galaxy up to redshift @xmath5 in the sdss @xmath6 bands .
in particular we see a significant reddening of the objects in @xmath7 towards fainter @xmath8-band magnitudes .
this reddening is consistent with the qso host galaxies being dominated by a stellar population of age at least @xmath9 gyr .
the full catalogue , including completeness estimates , is available on - line at http://www.2slaq.info/. quasars : general galaxies : active galaxies : seyfert stars : white dwarfs catalogues surveys |
in recent years , n - body numerical simulations have given us a good understanding of dark matter structure for standard cosmological scenarios , while large scale observational surveys have done the same for the distribution of galaxies . in this way
, we are now developing a good picture of how mass and luminosity are distributed in the universe .
however , it is still not well known how to connect the two pictures .
while it is well established that dark matter haloes are the hosts of the observed galaxies , it is still poorly understood how the former are related to the latter .
further , the picture is complicated by the fact that what is usually taken as a halo in simulations would often host multiple galaxies , especially for higher masses . to analyse the issue fully
, it is necessary to look at the halo substructure , since each subhalo can host a galaxy .
establishing such a link between halo mass and galaxy luminosity would be important because , first of all , it would allow us to have a direct connection between theory and observation , dark matter haloes and galaxies .
further , it could also shed some light into the theory of galaxy formation .
there are several ways in which this problem can be studied .
the more direct ones involve either numerical simulations including gas dynamics @xcite , or semi - analytical models of galaxy formation @xcite , but , while they explicitly give the properties of galaxies located in a given halo , they have the added difficulty that many of the mechanisms involved in galaxy formation are poorly understood , and difficult to compute .
their complexity could also mask any fundamental relations that might be present between halo and galaxy properties .
more indirect approaches have also been studied .
the halo occupation distribution ( hod ) model @xcite is based on the probability @xmath4 that a halo of mass m is host to n galaxies . by specifying the @xmath4 function , along with some form for the distribution of dark matter and galaxies within each halo ,
it is then possible to relate different statistical indicators of the dark matter and galaxy distributions , such as correlation functions , to each other . this fully specifies the bias between the galaxy and the underlying matter distributions . a recent paper by @xcite ,
has done a detailed study of results from simulations and related them to the hod model , and has concluded that the form of @xmath5 for the subhaloes is approximately universal , where @xmath6 is the subhalo mass scaled to an appropriate minimum mass .
this paper also give results for the relation between galaxy absolute magnitude and halo circular velocity .
other work @xcite has taken this approach one step further by studying not only the number of galaxies associated with each halo , but also their luminosity , by building the conditional luminosity function , @xmath7 .
this gives the number of galaxies with luminosities in the range @xmath8 contained in a halo of mass @xmath1 .
while this work directly relates the halo mass to the galaxy luminosity , it does so only to the average values , lacking the full statistical treatment which is analysed in the hod models .
other authors have used a slightly different method . instead of trying to specify the number of galaxies in each halo
, they treat the halo as a whole and identify it with a galaxy group .
then , by comparing the group luminosity function with the halo mass function , they obtain the luminosity associated with each halo @xcite , and also develop ways to estimate the number of galaxies hosted in a halo , thus coming back partly to the @xmath4 estimate of the hod models . in the present paper
, we follow a new and conceptually clear approach based on one simplified and testable hypothesis : there is a one to one , monotonic correspondence between halo / subhalo mass and resident galaxy luminosity . we might call this the empirical ( rather than the semi - analytical ) approach because there is no attempt to physically model the galaxy formation process .
instead we take from observations the galaxy luminosity distribution and match it with the theoretical halo / subhalo distribution .
this has the additional advantage of naturally giving a lower mass threshold for haloes that host luminous galaxies , as the luminosity decreases sharply with mass for less massive haloes .
it also implicitly gives rise to galaxy systems , if one identifies a system of a massive halo and its subhaloes with the central galaxy and its satellites in groups and clusters .
we find that the single assumption is very powerful and allows us to compute , and compare to observations many quantities from bias to the void distribution function to the spatial correlation function .
this paper is organized as follows : in section 2 we present our model for the subhalo mass distribution in a parent halo , and build the global subhalo mass function . in section 3 ,
we derive the relation between mass and luminosity , as well as some other functions such as the luminosity function of cluster galaxies , the group luminosity function and the multiplicity function . in section 4
we study how to apply the relation we obtain to get the light density and the number density of galaxies as a function of mass density , and also obtain results for the distribution function of light density and the void probability function .
finally , we conclude in section 5 . throughout
we have used a concordance cosmological model , with @xmath9 , @xmath10 , @xmath11 and @xmath12 @xcite .
the first step in building our model is to specify the mass distribution of subhaloes for a given parent halo .
we adopt the following function : @xmath13 which gives the number of subhaloes with masses in the range @xmath14 to @xmath15 , for a parent halo of mass @xmath1 .
the normalisation @xmath16 is such that the total mass in these subhaloes , @xmath17 , is a fraction of the parent halo mass , @xmath18 ( where the factor @xmath19 accounts for the added mass of the original , unstripped , subhaloes ) . with this definition ,
we can write @xmath16 as @xmath20 this expression is motivated by recent analysis of high resolution dark matter simulations of @xcite , and we use for the parameters the values @xmath21 , @xmath22 and @xmath23 .
its results are similar to those obtained by the simulations of @xcite , who find a power law fit to the subhalo mass function with a slope close to -2 ( that is , in terms of equation [ nsubhalo ] , @xmath24 ) .
however , they also find that if they only include the lowest mass bins where statistical errors are smallest , this slope is reduced to values around -1.9 , very similar to what we use here .
it further agrees well with the cumulative mass function derived analytically by @xcite ( see section below on the occupation numbers for further discussion ) , and it is also similar to the power law form for the subhalo number given in @xcite , only instead of having a sharp cutoff at the parent halo mass , we introduce a smooth exponential cutoff from a mass @xmath25 . since it is a schechter function , it is also similar to the halo mass function , and the slope @xmath26 is close to the expected value of the slope of the halo mass function .
it gives a total mass in subhaloes of 18% , close to but slightly higher than the values of @xmath27 obtained in different studies ( see , for example , @xcite ) .
@xcite derive a lower value of around 6% at radius @xmath28 ( where the average overdensity is 200 times the critical ) , though they also find that this fraction can be as high as 10 - 15% in some cases .
however , @xmath28 is smaller than the virial radii typically measured from the simulation results by @xcite , which helps to explain the difference .
it is worth noting that , like previous results from simulations and analytical modelling ( e.g. , @xcite ) , the shape of the subhalo mass function given by equation [ nsubhalo ] is independent of parent halo mass .
this does not mean , however , that one should automatically expect the same to be true of satellite galaxy distribution in galaxies like the milky way and clusters .
in fact , as can be see from our results further below , the mass luminosity relation has very different behaviour depending on the mass of the host halo / subhalo ; this leads to parent halos of very different masses ( such as a cluster and a galaxy ) having very different satellite luminosity distributions , even though the scaled mass functions of their subhaloes are identical . a very important point when analysing this expression is to note exactly what mass is being accounted by @xmath14 .
in fact , the distribution is valid for the present mass of the subhalo satellites , obtained when @xmath29 , which is measured after the tidal stripping of the outer parts of the subhalo in the halo potential well .
however , to use this function to build the host distribution and to subsequently compare it to the galaxy luminosity function , it is necessary to use the original , unstripped mass of the subhalo , since only then can we establish a monotonic correspondence between host halo mass and galaxy luminosity .
the factor @xmath19 put into the expression takes this mass loss into account , since we can then treat the mass @xmath14 in equation ( [ nsubhalo ] ) as the original mass , with the stripped mass being then @xmath30 with @xmath31 . obviously ,
the actual factor for a given subhalo will be highly variable , but for reasons of simplicity and also because it is not well known ( see e.g. @xcite for a treatment of the profile of stripped subhaloes ) , we will be treating this as an average applicable to all subhaloes , and using @xmath32 in the present work .
it should be noted that , since we are taking the total subhalo mass to be 18% of the parent halo mass , this factor can not be more than 5 , assuming that the majority of the mass in the parent halo was built up by stripping of the subhaloes .
a recent paper by @xcite includes a comparison between maximum and present mass of subhaloes in the dwarf mass range .
their results ( see their figure 4 ) seem to support that the average change in mass does not depend on the mass of the subhalo , and an average mass stripping factor of 3 as we adopt here seems a fair agreement .
all plots and further expressions ( where we will drop the factor @xmath19 ) presented refer to this original , unstripped mass ; to invert our approximation is simply a case of dividing the subhalo mass by a factor of 3 to represent the actual stripped mass which would be measured in a simulation .
ideally , we should use instead a distribution for the maximum circular velocity of the subhaloes . even though this quantity also shows a large scatter between the maximum value and that at present after stripping @xcite , its relative change should be smaller than what would be expected for a given change in mass .
this should be the case especially for subhaloes massive enough to host galaxies , which would have only their outer parts stripped off , and where therefore the inner parts which determine the peak circular velocity are left relatively undisturbed .
this would mean that a calculation based on the cumulative number function like we use here would incurr a smaller error , even with a large scatter , and in later work we will transform to that variable . using the expression for the subhalo number ( [ nsubhalo ] ) together with the halo mass function , @xmath33 , it is then possible to build the global subhalo mass function , that is , the number density of subhaloes in a given mass range , by @xmath34 if we assume that the halo mass function has a schechter form , @xmath35 it is possible to write an analytical expression for it : @xmath36 } \big(\frac{m}{\beta m_*}\big)^{-\frac{a+\alpha}{2 } } k_0\big(2 \sqrt{\frac{m}{\beta m_*}}\big ) \
, , \ ] ] where @xmath37 stands for the modified bessel function . with more general mass functions ,
the calculation has to be done numerically . in the present work ,
we use the more accurate sheth tormen mass function @xcite , @xmath38 with @xmath39 , @xmath40 , @xmath41 and @xmath42 ; as usual , @xmath43 is the variance on the mass scale @xmath1 , @xmath44 is the growth factor , and @xmath45 is the linear threshold for spherical collapse , which in the case of a flat universe is @xmath46 .
the result for the global subhalo mass function is shown in figure [ massffig ] , where the sheth tormen mass function is also shown . ) , while the subhalo one is the one obtained using our model for the subhalo number distribution with parent halo mass.,width=317 ] with the expression for the subhalo number , equation ( [ nsubhalo ] ) , it is also possible to calculate the halo occupation number , that is , the number of subhaloes in a halo of mass @xmath1 : @xmath47 because the integral diverges , it is necessary to set a minimum threshold for the subhaloes
. this can be set as the minimum mass for a halo to host a galaxy , in which case this occupation number corresponds to the number of galaxies contained in a given halo .
this is one of the basic ingredients of hod models , where it is taken as the average number of galaxies in a halo .
figure [ haloocupfig ] shows the occupation number as a function of the parent halo mass , in units of the minimum mass considered .
the most prominent feature is the cutoff for @xmath48 ; this is inherent to our model , since we consider the total mass in subhaloes to be only 18% of the halo mass , and we also have a cutoff in the number of subhaloes , given by the parameter @xmath49 . at the high mass end ,
the function is well fit by a power law , with @xmath50 for @xmath51 , which is similar to what is predicted by most hod models ( see e.g. , @xcite ) .
this behaviour is also analogous to that found in the analytically derived halo mass function of @xcite , who also find a slope of roughly 0.9 at high @xmath52 , but a steeper value close to 1 in the lower range , similar to our results .
the values they obtain are also similar to ours , with one caveat : they correspond to those we would get if we took the values for the stripped mass of the subhaloes , instead of what is plotted in figure [ haloocupfig ] ; this would correspond roughly to diving the numbers we get by a factor of 3 . our result is also similar to what has been observed for the case of galaxies in a cluster , with @xcite obtaining a relation for the number of galaxies with luminosity greater than @xmath53 , @xmath54 , where all quantities are normalised to the usual mass overdensity of 200 . if we assume that their @xmath53 galaxy would have a mass of around @xmath55 ( see below for the results we obtain ) , we also obtain approximately the same number of galaxies in a @xmath56 parent halo . the overall shape of our occupation number is also similar to the results obtained by @xcite in their simulations , although in their case the cutoff is less pronounced , and in fact is only noticeable for @xmath57 . as referred , however , our cutoff around @xmath58 is in fact a feature of our model , and something similar should be present as long as a cutoff in subhalo mass is introduced and there is an upper limit to the total mass in subhaloes .
the normalisation of the curves is also similar , with a value of @xmath59 for @xmath60 .
it should however be cautioned that the way in which the subhalo mass is accounted may not be exactly the same in both cases ( see discussion above ) .
, solid line).the halo mass is scaled to the minimum mass considered for the subhaloes .
the solid line shows the results for the model used in this paper ( @xmath22 ) .
the two dotted lines show the result of changing the subhalo cutoff ( respectively for @xmath61 and @xmath62).,width=317 ]
once we have the total mass function for haloes and their subhaloes , it is then possible to compare it to the galaxy luminosity function to obtain a mass luminosity relation .
as noted , a few prior assumptions go into this : first , that the galaxies are hosted individually in the haloes or subhaloes , and that each contains a single one ( or , in the case of the parent haloes , they have one in their centre , plus the ones in their subhaloes ) ; and second , that the luminosity is a monotonic function of the halo mass .
the first assumption is supported by previous studies of the group mass function .
for example , @xcite obtain a group mass function from the 2df catalogue which matches well with theoretical mass functions like the sheth tormen one , equation ( [ stmf ] ) , indicating that each of their identified groups corresponds to a halo .
as for the second , @xcite have shown that subhaloes identified in a set of simulations have a correlation function and power spectrum that matches the galaxies in the pscz survey , which shows it is possible to understand the spatial distribution of galaxies by identifying galaxies brighter than a given luminosity with haloes larger than a certain mass .
this is further justified by studies which show that the only halo property dependent on the large scale environment is the mass distribution ( e.g. , @xcite ) , which coupled with the current understanding of galaxy formation theories should imply that we have captured most of the environmental dependence of the galaxy luminosity .
we take the galaxy luminosity function to have the usual schechter form : @xmath63 the values of the different parameters are taken from the @xmath64 band 2df galaxy luminosity function @xcite , with @xmath65 , @xmath66 and @xmath67 .
although this fit was determined for the magnitude range @xmath68 , here we will be extrapolating its result to higher or lower luminosities as necessary .
the mass luminosity relation is then calculated by setting the luminosity @xmath69 of a galaxy hosted in a halo of mass @xmath1 to be such that the number of galaxies with luminosity greater than @xmath69 equals the number of haloes plus subhaloes with mass greater than @xmath1 : @xmath70 dm \ , .\ ] ] it should be noted that , because the values for the subhalo mass function are generally lower than those of the halo one , the subhaloes make only a small contribution to this expression , only being important at lower mass values .
this guarantees that a possible second generation of subhaloes ( subhaloes of the subhaloes considered here ) would not influence this result much , since their mass function would have values that much lower than the original halo mass function , and then only at the lowest masses considered .
it is also possible to obtain the group luminosity associated with each halo . in order to do this ,
we first assume that a galaxy system ( which would be either an isolated galaxy , a group or a cluster , depending on the mass of the halo containing it ) can be represented by a halo and its associated subhaloes .
the group luminosity for a halo of mass @xmath1 is then simply @xmath71 refers to the parent halo mass only .
also shown as a dashed line is the result when the mass is taken to be the actual stripped mass of the subhaloes ( which is obtained by dividing the original mass by a factor of 3 ) .
the different points are estimates from different observational studies ( see discussion in text for references).,width=317 ] there is an important point that should be considered when analysing this expression , and that is the possibility of putting a lower limit in the integral , since there may be a minimum mass for a halo to be host to a luminous galaxy .
however , as can be seen in figure ( [ masslightfig ] ) , which shows the results for the luminosity and group luminosity associated with a halo , the calculated luminosity decreases sharply with mass at the low end , becoming negligible for haloes with masses lower than approximately @xmath72 .
therefore , instead of putting in a cutoff , the value of which is not clear from what is known of galaxy formation , we simply use the natural drop off in the derived relation .
this lower limit seems interestingly to be roughly the mass which is usually considered for the host haloes of dwarf galaxies ( for example , @xcite give a lower limit of @xmath73 to the maximum circular velocity of halo capable of hosting a dwarf galaxy ; see also @xcite for their simulation results on the halo hosts of dwarf galaxies , where they give similar values for the less massive of them ) , but it is quite low when compared with the values used in some studies ( for example , in the hod model in @xcite ) , although this may be related to the fact that most dark matter simulations used in these studies do not reach such low masses .
another important aspect that can be seen in the figure is that the group luminosity only starts to depart from the halo luminosity for masses above @xmath74 .
this is to be expected given the estimated mass of the host haloes of rich groups and clusters .
this difference is the contribution of the subhaloes , and it is clearly seen that , although their numbers may not be very significant to the total mass function , they are very important for the luminosity of high mass haloes , and therefore to the luminosity function . due to their small numbers when compared with subhaloes , it is to be expected that subsubhaloes make only a small contribution to this group luminosity , and that this would only be noticeable for the most massive parent haloes , where the subsubhaloes can be massive enough to be luminous .
noticeably , the group luminosity has a middle region of intermediate mass with a shallower slope , and then an upturn for high mass , which was also present in similar studies which used the group luminosity function ( e.g. , @xcite ) , while the single halo luminosity has a much shallower high end slope .
we find that the mass luminosity relation we obtain can be fairly well fit by a double power law of the type : @xmath75 for the luminosity of individual galaxies ( the solid line in figure [ masslightfig ] ) , the parameters are @xmath76 , @xmath77 , @xmath78 , @xmath79 , @xmath80 and @xmath81 .
therefore , at the low mass end , we have @xmath0 .
in fact , if we go to very low luminosities , the slope will actually be slightly steeper ; nonetheless , the region of interest essentially begins for luminosities above a few times @xmath82 ( which corresponds to the fainter dwarf galaxies ; however , note that such low luminosities are outside the range of the schechter fit to the 2df luminosity function , so this is based on an extrapolation of the 2df results to this luminosity range ) , and for these the slope is approximately 4 .
such a slope is actually what you would expect from a straightforward comparison between the halo mass function ( assuming a slope of -1.8 at the low mass end ) and the galaxy luminosity function . since at the low mass end the subhaloes
give an important contribution to the total number of hosts ( see figure [ massffig ] ) , this is most likely a coincidence , arising from the fact that the total host distribution ( haloes+subhaloes ) has a slope similar to the halo mass function .
at the high mass end , we essentially obtain the relation between the halo mass and the luminosity of the brightest cluster galaxy , which has a much shallower slope ( @xmath83 ) .
this is most likely due to the fact that , by construction , the mass term refers to the mass of the entire halo hosting the cluster , and not just to the mass in the region of the galaxy itself .
the halo mass / group luminosity relation is not really well fit by a double power law , since it has a third region for middle values of mass with a shallower slope than either of the asymptotic values .
however , we find that it is possible to describe it as two different double power law branches , which provide a fair fit to the results .
thus , for @xmath84 , the group luminosity is essentially the same as the luminosity of the parent halo ( as these haloes do not have subhaloes massive enough to be luminous ) , so we fit it also with equation [ fit ] , and use the same parameters as before . for parent
halo masses higher than @xmath85 , we find that a simple double power law of the form @xmath86 , with @xmath87 is a good fit .
such a behaviour , of a curve with a relatively flat slope in the intermediate mass range and steeper slopes at both the low and high mass ends is similar to what was observed by @xcite , who compared the ags group luminosity function with the halo mass function , although the actual shape of the curve is somewhat different from what we find here . at the high mass end ,
the cluster luminosity is almost directly proportional to halo mass ( in fact , @xmath3 ) .
this means that the resulting mass to light ratio will be almost constant , rising only very slowly with halo mass , which matches well with previous results for the mass to light ratio of clusters ( e.g. , @xcite ) .
however , the values we obtain for the group luminosity seem to be smaller than the observational results .
further , the value derived by @xcite for the cluster blue mass to light ratio is @xmath88 , which is roughly consistent with the value we obtain around @xmath89 , but smaller than what we get at @xmath56 . since the mass luminosity relation for a single halo is dominated at the high end by the parent haloes , it seems unlikely that this result could be much modified in the scope of our model , since both the halo mass function and the luminosity function are well known .
this then means that the problem would lie in the subhalo distribution for these massive parent haloes ; it would be necessary either to have more of them , or else for them to be more luminous .
the two are in fact related , since an increase in the number of subhaloes causes an increase in the total number of hosts , and thus a decrease in the luminosity of a halo ( especially in the low mass range , where subhalo number is more significant ) .
however , our results for the occupation number and the luminosity of lower mass haloes seem to be in good agreement with observations and prior theoretical models , which leads us to believe that the most likely cause of this result is the simplistic way in which we treated subhalo mass stripping . we can also compare our results with those in @xcite , who fit the mass to light ratio in different models they study to a double power law .
in general , their results are in fair agreement with what we obtain . at the low mass end , the results are quite similar , with these authors obtaining a minimum of the mass to light ratio at a slightly lower mass . at the high mass end , they obtain a steeper mass to light ratio as a function of mass , although this is due to the flattening off we find in our results , since in the intermediate range @xmath90 we obtain higher mass to light ratios and a steeper relation .
this discrepancy is most likely due to the factor that these authors are fitting the mass to light ratio to a double power law , which as was discussed above does not provide a good fit to our results .
in fact , when they adopt a different model with a fixed mass to light ratio at high mass , their results agree with ours slightly better in this intermediate region .
overall , our results seem a fairly good match to estimates of mass taken from a range of observational results across the entire mass range . shown
are points for the three most luminous of the dwarf spheroidals in the local group , where the luminosity was taken from the review by @xcite , and the mass was estimated from the results of @xcite ( see the results in their figure 13 for bounds to the peak circular velocity function of unstripped nfw haloes estimated to be possible hosts of the local group dwarf spheroidals ) , by assuming that the relation between mass and luminosity is monotonic .
the dashed box represents the relation obtained from the weak lensing study of @xcite , for galaxies around @xmath91 ( where we have taken the values applicable to a nfw halo ) .
the results for poor groups are taken from @xcite , while those for clusters are from @xcite , where the bounds come from the two different methods the authors use for estimating fore / background corrections .
also included is a point for the milky way @xcite .
the major difficulty in obtaining a luminosity function for galaxies in clusters lies in distinguishing what haloes should be treated as rich clusters in our model , and which are simply groups .
we adopt the abell definition for rich cluster , namely that it must have upwards of 30 objects brighter than @xmath92 , where @xmath93 is the magnitude of the third brightest galaxy in the cluster .
using our derived mass luminosity relation , it is possible to transform this magnitude threshold into a mass one . in our model , this would correspond to the mass of the second most massive subhalo ( since we consider the brightest galaxy to be in the parent halo ) . in order to calculate this
, we assume that its mass is given by a distribution which is the product of the poisson probability that on average there exists a single subhalo more massive than it , by the probability of finding a subhalo at that mass .
that is , its average value for a parent halo of mass @xmath94 is given by @xmath95 where @xmath96 is the probability of the second most massive subhalo having mass @xmath1 , @xmath97 where @xmath98 is the mass distribution function of the subhaloes , equation [ nsubhalo ] , and @xmath99 is the average number of subhaloes more massive than @xmath1 in a parent halo of mass @xmath94 .
since this mass threshold depends on the mass of the parent halo , in the end we obtain a condition on halo mass for it to be treated as a host to a rich cluster . using our results
, we find this to be @xmath100 .
if we now combine the distribution of all haloes more massive than this , together with their subhaloes , with the mass luminosity relation , we can obtain the luminosity function of galaxies in clusters .
in fact , it is possible to make a rough prediction for its shape at the high end by comparing the luminosity of the brightest galaxy in the cluster ( given by the solid line in figure [ masslightfig ] , which represents the luminosity of the galaxy associated with the parent halo itself ) with the total cluster luminosity .
the average luminosity of the brightest galaxy can be estimated as @xmath101 , where @xmath102 is the cluster galaxy luminosity function , and @xmath103 is such that on average there is only one galaxy more luminous than @xmath103 , i.e. @xmath104 ( see @xcite ) . with @xmath102 normalized to the total cluster luminosity , @xmath105 , it is then possible to extract a relation between @xmath106 and @xmath105 , which will depend on the shape of the luminosity function .
taking our results from figure [ masslightfig ] that @xmath107 and @xmath108 , it is possible to work out that , at the high end , @xmath102 is given by a power law with slope -4.21 . for the mass of a halo associated with a rich cluster .
the dotted line shows the 2df global luminosity function , the short dashed one the luminosity function of cluster galaxies also from 2df , and the long dashed line is a double power with the same parameters as the previous function , but with @xmath109 at high @xmath69.,width=317 ] we show the results we obtain in figure [ clusterlffig ] , where we have excluded the first brightest galaxies ( in effect , we are just accounting for the galaxies in subhaloes ) .
for comparison , we include the 2df global luminosity function of @xcite , as a dotted line .
we also include the luminosity function for galaxies in clusters in the 2df survey , as derived by @xcite , as the short dashed line .
this is also given by a schechter function , with parameters @xmath110 and @xmath111 .
since the normalisation is not given , we adjusted this to have values in the same range as the ones we obtain . also shown , as the long dashed line , is a double power law with the same low end slope , normalisation , and @xmath53 as the ones used to plot the previous curve , but where the exponential cutoff has been replaced with a power law with the same slope as calculated above , namely @xmath112 .
our results show a good agreement to those obtained from 2df , which is even better if a double power law is considered instead of a schechter function .
another way to check our results is to build the group luminosity function .
this is the equivalent of the galaxy luminosity function , only applied to groups , and it is usually obtained from galaxy catalogues by building groups of gravitationally bound galaxies . in our case , we start with the same assumptions for groups described above , and use the halo mass function and the relation between halo mass and group luminosity given by equation ( [ grouplum ] ) so that @xmath113 ) ; the two other lines are observational group luminosity functions extracted from galaxy catalogues , respectively the ags @xcite ( dotted ) and the vslf @xcite ( dashed).,width=317 ] we show our result in figure [ grouplumfig ] , along with the ags for the cfa survey @xcite and the vslf @xcite group luminosity functions .
while at higher luminosities our model would seem to underpredict the abundance of groups , it reproduces the slope of the ags luminosity function quite well .
this is undoubtedly related to the fact that our calculated group luminosity is somewhat lower than the observed values ( see figure [ masslightfig ] ) ; an increase in group luminosity would cause a shift to the right of our curve , providing a better match for the observational results .
it should also be noted that the inclusion of subsubhaloes would probably tend to slightly increase the values at the high end , relatively to the low end , since most of these subsubhaloes would be dark for low mass parent haloes , but be more massive and therefore contribute a small luminosity to the group total in the case of massive parent haloes .
it is also curious to note that , since the fits shown in the figure are a double power law in the case of the ags and a schechter function in the case of the vslf , our result would probably not be well fit by a schechter type function .
it is also possible to derive the multiplicity function , the number of groups / clusters as a function of their richness , by a process similar to the one used for the group luminosity function . only in this case
, we use the occupation number ( plus one to account for the central galaxy , which is hosted by the parent halo itself ) instead of the group luminosity .
then , using an expression similar to ( [ grouplf ] ) with the subhalo number @xmath114 put in place of the group luminosity @xmath115 , we obtain the result shown in figure [ multfig ] . the main point to take into account in this calculation
is , as was referred in our above discussion of the occupation number , the need to introduce a minimum mass for the subhaloes . in the case of the figure shown ,
this was taken to be the mass equivalent to @xmath116 , but in general , and to compare to results from observational studies , this should be set to equal the minimum luminosity considered for objects in the observations .
there is also a sharp upturn at @xmath117 due to all the haloes that are massive enough to be considered to host a galaxy , but not enough so that they can have subhaloes with galaxies in them .
figure [ multfig ] shows our results in comparison with some observational data .
the points were taken from the analysis of @xcite , while the two lines were constructed from the two group luminosity functions shown in figure [ grouplumfig ] , by using the luminosity functions of the two surveys to relate total group luminosity to the number of galaxies above the magnitude threshold .
the galaxy numbers for both the points and the results derived from the group luminosity functions were then multiplied by a factor of 0.66 to take into account the difference in radius between friends of friends estimates and the usual definition of virial radius ( see @xcite ) . even though our results lie in the range between those estimated from the vslf and the ags group luminosity functions , they seem to be a bit lower than the observational points .
since the group abundance is directly related to the halo mass function , which is well known , this discrepancy must be caused by the occupation number .
a slightly higher occupation number would shift our curve further to the right in the figure , bringing it into better agreement with the observational data .
incidently , a higher occupation number would also bring our results for the group luminosity and the related group lumininosity function into better agreement with observational values , so we believe that this is where the problem lies .
our calculated occupation numbers seem to be in fair agreement with expectations ; however , they depend sensitively on our prescription for subhalo stripping , where we adopted a simplistic approach . a more detailed and correct model for this effect would most likely produce better results . .
the different points are taken from the results derived by @xcite from the cfa survey , while the two additional lines are derived from the group luminosity functions shown in figure [ grouplumfig ] , where the magnitude limit for all of these is also @xmath116.,width=317 ]
to derive a relation between the mass and light densities , we first need to obtain the mass function of haloes for regions of different density . in order to do this ,
we follow the method outlined in @xcite . in essence
, we treat the evolution of the matter distribution in a region as if it derived from a universe with the same cosmological parameters as that region , namely the same average mass density as is found in the local region .
we start by labelling each region by its average mass density , @xmath118 , to which corresponds a given value of the parameter @xmath119 , with @xmath120 the critical density .
overdense regions will have @xmath121 , where the barred quantities refer to the background universe , while voids will have the opposite .
the growth of perturbations will be affected by the different matter content , and this is reflected by changing the normalisation of the power spectrum according to the differential of the growth factor , @xmath44 , relative to the background : @xmath122 where @xmath123 is some initial redshift for which the fluctuations were equal in the background and in this region ; in the present case , we use @xmath124 .
we also use the usual normalisation for the growth factor , @xmath125 . to obtain the mass function for a region with a particular average density , @xmath126
, we then apply this new normalisation to the sheth tormen mass function given in equation ( [ stmf ] ) , changing the @xmath127 term as appropriate , and also the value for @xmath45 , which has a small dependence on @xmath128 ( see for example @xcite ) .
an important point is that this prescription does not utilise an explicit smoothing radius for the region considered .
instead , these are labelled by their average density .
however , when regions of limited size are considered , there exists a maximum mass for objects in them , given by the total mass they contain . as the above prescription
does not take this into account , we introduce an additional term to compensate for this effect . therefore ,
after we obtain the mass function , we put in a further cutoff of the form @xmath129 $ ] , where @xmath14 is the mass of the halo and @xmath130 the total mass in the region of radius @xmath131 and average density @xmath132 .
the parameter @xmath133 allows for some tuning of the actual cutoff , and in the present case we use @xmath134 .
once this cutoff is introduced , it becomes necessary to renormalise the mass function , so that it still gives the appropriate density when the mass of all haloes is calculated .
two examples of the resultant mass functions are shown in figure [ voidmffig ] , with curves for regions with the average density , @xmath135 of the average and @xmath136 times the average , for two different radii . , the lower for @xmath137 . in both , the solid curves are the background sheth tormen mass function , the dotted ones are for densities of(top to bottom ) : @xmath138 , @xmath139 , @xmath140 . , title="fig:",width=317 ] , the lower for @xmath137 . in both , the solid curves are the background sheth tormen mass function , the dotted ones are for densities of(top to bottom ) : @xmath138 , @xmath139 , @xmath140 . , title="fig:",width=317 ]
once the mass function for an over or under dense region has been determined , it is then simple to find the equivalent light density , @xmath141 , by using the group luminosity correspondent to each halo , given by equation ( [ grouplum ] ) : @xmath142 our result is shown in figure [ mldensityfig ] .
the results for @xmath143 and @xmath144 are quite similar , with the curve for @xmath145 being different .
this is due to the large suppression of high mass haloes , even at high overdensities , which can be seen in figure [ voidmffig ] , which alters the proportion of luminous to non - luminous haloes in favour of the latter .
the curve we obtain for @xmath146 is similar to the one in @xcite , where the authors have used a mass function for different densities derived from simulation results , to which they then apply the conditional luminosity function .
the agreement is quite good with our result , even though we used a theoretical model for the mass function instead of taking it from simulations .
, as a function of mass density , @xmath147 .
both axis are scaled to the average background values , so regions with @xmath148 are overdense and regions with @xmath149 are underdense .
the different lines are for different smoothing lengths : @xmath145 ( solid ) , @xmath150 ( dotted ) , @xmath146 ( short dashed ) .
the long dashed line marks @xmath151 .
the dashed dotted line is the fit to the same relation mentioned in @xcite.,width=317 ] the most striking feature in these results is the sharp decline in the light density in underdense regions .
this would mean that the majority of void regions would be quite dark .
another important feature is that for the most part @xmath152 , and the slope of the curve at high density is lower than 1 .
the overall results ( e.g. , darkness of the voids ) are similar to those from hydrodynamical simulations by @xcite .
but the simulation results show a sharper cutoff at low mass density and light being more overdense than mass for overdense regions and a slope to the relation greater than one .
this difference is the more remarkable since the results of @xcite are similar to ours while using a mass function derived from simulations , which otherwise might be considered the most likely origin of the discrepancy .
however , previous results shown in @xcite , also taken from simulations , seem to be more in line with what we obtain in the present work .
these authors in fact give an explanation for the apparent increasing antibias at higher mass densities : at low redshift these regions usually consist of rich clusters and superclusters , whose stellar population tends to be old .
therefore , young blue stars tend to be rare and consequently the total luminosity in the blue band is lower than what could be expect from their high mass , giving rise to the slight antibias .
it is therefore likely that this difference is coming from the way in which light is being counted in the two methods .
in fact , one problem with the approach presented here is that it breaks down for haloes whose size is comparable to the size of the region being considered , since it is then possible for a sphere to encompass an outer region of the halo where a subhalo and a galaxy lie , and which is therefore luminous but is considered dark in our prescription . since @xcite study the light distribution directly , they would account for the light present in such a situation .
this effect would be the more noticeable for high density regions and small smoothing lengths , and in fact the results presented in @xcite are for @xmath145 , while the major discrepancy with our results occurs for high density regions .
the differences in underdense regions can most likely be explained by limited resolution effects .
we also derive results for the bias between galaxy numbers and the dark matter mass density .
the procedure is similar to the adopted for the luminosity , replacing the @xmath153 term in ( [ rholight ] ) with the number of galaxies in a halo of mass @xmath1 , i.e. @xmath154 , corresponding to the central galaxy hosted in the halo and the ones present in the subhaloes .
the integral should also be taken with a lower limit at @xmath155 .
the minimum mass @xmath155 corresponds to the intended cutoff in mass ( or luminosity ) for the galaxies to be counted .
our result is shown in figure [ mndensityfig ] .
the general behaviour is similar to what was observed for the luminosity .
according to these results , the galaxy number distribution would be a fairly unbiased tracer of the underlying dark matter mass in moderately overdense regions , as would be expected from previous studies ( e.g. , @xcite ) , but the galaxy numbers severely underestimate the mass density in underdense regions and consequently assuming a direct proportionality can lead to a significant underestimate of @xmath128 @xcite .
( curves have same label ) , but for @xmath156 , the overdensity in the number of galaxies .
the minimum threshold was set at halo mass corresponding to a luminosity @xmath157 .
the average number of galaxies brighter than this limit in the background , taken from the luminosity function , is @xmath158.,width=317 ] using the relations obtained in the previous section , it is also possible to derive a distribution function for the light density ( i.e. , the probability of a region of a given size having a certain average light density , the light pdf ) . to make this calculation , besides the relation between the mass and light densities
it is also necessary to have the mass distribution function . in general , this can be obtained from dark matter simulations . for the present work , we will use the approximation of treating this as a lognormal , with an appropriate dispersion dependent on the radius being considered and on the cosmological model .
it has been known for some time that such a function is a fair approximation to the real mass distribution function , at least in regions of moderate over / under density ( @xcite ; see also @xcite and @xcite for some recent analysis on this subject ) .
the form of the lognormal is given by : @xmath159 ^ 2}{2\omega^2}\big ) \
, , \ ] ] where @xmath160 and the parameter of the distribution @xmath161 is related to the smoothing length and the variance @xmath162 of the mass density field by @xmath163 here , the variance @xmath162 refers to the full , nonlinear spectrum . to relate this to the linear variance , and hence to the cosmological parameters we use , we follow the prescription outlined in @xcite to convert the linear power spectrum of mass density fluctuations to the nonlinear one . to obtain the light distribution function ,
we then follow a similar procedure to the one used for the group luminosity function : @xmath164 where we use the notation @xmath165 and @xmath166 is the light distribution function .
such a transformation assumes a monotonic dependence of the light density with the mass density , which is true of our results ( see figure [ mldensityfig ] ) .
another important point to note is that this function is not normalised to one , since some of the underdense areas ( in terms of mass ) are non luminous ( see discussion in @xcite ) .
our result is shown in figure [ distfuncfig ] .
the general behaviour is what would be expected : the distribution functions are similar in the overdense regions , but in general differ substantially in the underdense regions , with the light distribution function in general not well fit by a lognormal . , with peak height increasing with radius .
the x - axis represents either mass or light overdensity , depending on the curve .
the distribution functions are shown scaled appropriately for logarithmic binning.,width=317 ] the void probability function is defined as the probability of having no galaxies in a sphere of radius @xmath131 .
although voids are a particularly striking feature in the galaxy distribution , their study is impaired by their large size and low number of galaxies present in the observational case , and by resolution difficulties when the study is conducted through numerical simulations including galaxy formation .
due to these factors , in the past studies of the voids in the galaxy distribution have been relatively few ( see , for example , @xcite for observational results , and @xcite and @xcite for comparisons with theoretical results ; there are also two recent papers with observational results for the 2df survey , @xcite ) . despite this
, the void probability is a powerful tool in the analysis of a particular model of galaxy formation , since it probes a highly non - linear regime and is not derivable from the low order correlation functions ( in fact , it depends on all of the n - point correlation functions ) . in the case of our model
, we can determine the void probability function by combining the dark matter distribution function with the expected galaxy number in regions of a given size .
thus the number density of galaxies in a region of radius @xmath131 and average density @xmath118 , given a minimum mass cutoff of @xmath167 will be : @xmath168 where @xmath169 is the mass function in a region of size @xmath131 and average density @xmath118 ( see discussion in section 4.1 , also figure [ voidmffig ] ) , and @xmath170 is the number of subhaloes in a halo of mass @xmath1 .
the number is then obtained by multiplying this density by the region volume , @xmath171 . we treat this as the average number and assume a poissonian distribution around this average ( see @xcite , where the authors have shown that the full hod is consistent with a poisson distribution for large host masses ) , so that the probability of having 0 galaxies will in fact be @xmath172 .
finally , the void probability function is obtained by integrating this number times the probability of a region having this given density , @xmath173 where @xmath174 is the probability of a region of radius r having density @xmath118 , which we take to be given by the lognormal distribution , equation ( [ lognormal ] ) .
our results are shown in figure [ voidpffig ] , where we also show observational results for the 2df survey taken from @xcite , for both the ngp and the sgp . at low radius ,
we obtain a good agreement , but our model would seem to overestimate the probability at large radii .
our results are also in better agreement with the values observed for the sgp than for the ngp , since @xcite have found the former to be somewhat emptier than the latter .
there is an additional effect that may go some way towards explaining this discrepancy , and that is corrections for the peculiar velocity distortions .
these would lead , especially for the larger radii , to the inclusion in voids of galaxies that are not actually there , due to the smearing effect caused by peculiar velocities to the positions of galaxies in redshift space .
this effect is not symmetric , since the voids are areas of low mass density and therefore the galaxies in them will have low velocity dispersions . the final result of this would be to overpopulate the voids and consequently to underestimate the measured void probability function , which would lead to a better agreement with our results . , and the corresponding mass was taken as the minimum mass limit in the calculations ( equation [ avgno]).,width=317 ]
in this paper we present a new model for relating halo mass to hosted galaxy luminosity , based on the dark matter substructure .
we feel this is a potentially powerful way to approach this problem , since it is based on two main inputs , the halo / subhalo distribution and the galaxy luminosity function , which can be well tested and adjusted to results from simulations and surveys , respectively .
additionally , the model requires only one further assumption : that there is a one to one and monotonic relation between imbedded halo / subhalo mass and galaxy luminosity .
this is more explicit but in general agreement with the general assumptions made in past studies and general assumptions made in similar work of the same subject .
it is a much less restrictive assumption than the still sometimes utilized ansatz of a linear `` bias '' between galaxy numbers and dark matter density .
we have shown how , starting with a prescription we describe for the subhalo mass distribution in a parent halo , it is possible to obtain a relation for the luminosity of a hosted galaxy , as well as the group luminosity when the system of a halo and its subhaloes is identified with groups or clusters of galaxies .
the subsequent results appear to match well with general assumptions of the behaviour of such a relation , as well as results for the mass and luminosity at scales of dwarf galaxies , @xmath91 galaxies and massive clusters . from this model ,
it is then possible to derive many quantities that can be directly compared with further simulation or observational results .
we have shown four examples of this , namely the occupation number , the luminosity function of cluster galaxies , the group luminosity function and the multiplicity function . using further assumptions on how to calculate the mass function for regions of different average densities , and on the shape of the dark matter distribution function ,
we have also obtained a relation between mass and light and number densities for different smoothing lengths close to what is expected from previous bias studies .
we also obtain the distribution function of light density , and the void probability function .
the latter is a powerful additional test , since it probes a highly non - linear regime , and our results seem to match well with previous observations
. the major difficulty with the model as it is presented here has to do with the identification of the mass of the subhaloes . to be able to apply the monotonic correspondence to the galaxy luminosity
, the original mass in the subhaloes has to be taken into account .
however , the mass distribution we use is measured for the stripped mass of the subhaloes in their parent halo . in order to build the model
, we took the approximation of taking an average for the stripping factor .
this is , however , a not wholly satisfactory approach , since in general the stripping history will be highly variable from subhalo to subhalo , and also in different parent haloes .
ideally , we would like to use the maximum circular velocity instead of the mass to identify the subhaloes , since this quantity should less sensitive to stripping for the relatively massive subhaloes which can host galaxies .
unfortunately , we do not have at present a good distribution for subhalo abundance as a function of the maximum circular velocity , although we wish to study this further in the future . nonetheless , although somewhat crude , the approximation we took is reasonable , and since this work deals with statistical averages for most of the quantities , it does fit somewhat well into the whole structure .
more importantly , the results we obtain seem for the most part to match well with observational data . in future work which takes into account the statistical variation , however
, this factor will certainly be of great importance .
we wish to thank michael strauss and jochen weller for useful discussions .
we are also very grateful to the anonymous referee for a most helpful report .
av acknowledges financial support from fundao para a cincia e tecnologia ( portugal ) , under grant sfrh / bd/2989/2000 .
99 bahcall n. a. , ostriker j. p. , perlmutter s. , steinhardt p. j. , 1999 , sci , 284 , 1481 bahcall n. a. , cen r. , dav ' e r. , ostriker j. p. , yu q. , 2000 , apj , 541 , 1 benson a. j. , 2001 , mnras , 325 , 1039 benson a. j. , baugh c. m. , cole s. , frenk c. s. , lacey c. g. , 2000a , mnras , 316 , 107 benson a. j. , cole s. , frenk c. s. , baugh c. m. , lacey c. g. , 2000b , mnras , 311 , 793 benson a. j. , frenk c. s. , baugh c. m. , cole s. , lacey c. g. , 2001 , mnras , 327 , 1041 benson a. j. , frenk c. s. , baugh c. m. , cole s. , lacey c. g. , 2003 , mnras , 343 , 679 benson a. j. , hoyle f. , torres f. , vogeley m. s. , 2003 , mnras , 340 , 160 berlind a. a. , weinberg d. h. , 2002 , apj , 575,587 berlind a. a. et al .
, 2003 , apj , 593 , 1 bullock j. s. , wechsler r. h. , somerville r. s. , 2002 , mnras , 329 , 246 coles p. , jones b. , 1991 , mnras , 248 , 1 cox a. n. , ed .
, 2000 , allen s astrophysical quantities 4th edn . , aip press , new york croton d. j. et al .
, 2004 , astroph/0401406 , submitted to mnras de lucia , g. , kauffmann , g. , springel , v. , white , s. d. m. , lanzoni , b. , stoehr , f. , tormen , g. , & yoshida , n. 2004 , mnras , 348 , 333 de propris r. et al . , 2003 , mnras , 342 , 725 fukugita m. , hogan c. j. , peebles p. j. e. , 19 98 , apj , 503 , 518 ghigna s. , moore b. , governato f. , lake g. , quinn t. , stadel j. , 2000 , apj , 544 , 616 girardi m. , borgani s. , giuricin g. , mardirossian f. , mezzetti m. , 2000 , apj , 530 , 62 gottl " ober s. , okas e. l. , klypin a. , hoffman y. , 2003 , mnras , 344 , 715 governato f. , baugh c. m. , frenk c. s. , cole s. , lacey c. g. , quinn t. , stadel j. , 1998 , natur , 392 , 359 hayashi e. , navarro j. f. , taylor j. e. , stadel j. , quinn t. , 2003 , apj , 584 , 541 hoekstra h. , yee h. k. c. , gladders m. d. , 2004 , apj , 606 , 67 hoyle f. & vogeley m. s. , 2003 , astro - ph/0312533 , submitted to apj kauffmann g. , nusser a. , steinmetz m. , 1997 , mnras , 286 , 795 kauffmann g. , colberg j. m. , diaferio a. , white s. d. m. , 1999a , mnras , 303 , 188 kauffmann g. , colberg j. m. , diaferio a. , white s. d. m. , 1999b , mnras , 307 , 529 kayo i. , taruya a. , suto y. , 2001 , apj , 561 , 22 kochanek c. s. , white m. , huchra j. , macri l. , jarrett t. h. , schneider s. e. , mader j. , 2003 , apj , 585 , 161 kravtsov a. v. , berlind a. a. , wechsler r. h. , klypin a. a. , gottlber a. , allgood b. , primack j. r. , 2003 , astro - ph/0308519 , submitted to apj kravtsov a. v. , gnedin o. y. , klypin a. a. , 2004 , astro - ph/0401088 , submitted to apj lemson g. , kauffmann g. , 1999 , mnras , 302 , 111 magliocchetti m. , porciani c. , 2003 , mnras , 346 , 186 marinoni c. , hudson m. j. , 2002 , apj , 569 , 101 marinoni c. , hudson m. j. , giuricin g. , 2002 , apj , 569 , 91 martnez h. j. , zandivarez a. , merch ' an m. e. , domnguez m. j. l. , 2002 , mnras , 337 , 1441 mateo m. l. , 1998 , ara&a , 36 , 435 mathis h. , white s. d. m. , 2002 , mnras , 337 , 1193 mo h. j. , yang x. , van den bosch f. , jing y. p. , astro - ph/0310147 , submitted to mnras moore b. , frenk c. s. , white s. d. m. , 1993 , mnras , 261 , 827 moore b. , ghigna s. , governato f. , lake g. , quinn t. , stadel j. , tozzi p. , 1999 , apj , 524 , l19 nagamine k. , fukugita m. , cen r. , ostriker j. p. , 2001
, apj , 558 497 navarro j. f. , frenk c. s. , white s. d. m. , 1997 , apj , 490 , 493 neyrinck m. c. , hamilton a. j. s. , gnedin n. y. , 2004 , mnras , 348 , 1 norberg p. et al . , 2002 ,
mnras , 336 , 907 oguri m. & lee j. , 2004 , astro - ph/0401628 , submitted to mnras ostriker j. p. , nagamine k. , cen r. , fukugita m. , 2003 , apj , 597 , 1 peacock j. a. , dodds s. j. , 1996 , mnras , 280 , l19 peacock j. a. , smith r. e. , 2000 , mnras , 318 , 1144 pearce f. r. , jenkins a. , frenk c. s. , white s. d. m. , thomas p. a. , couchman h. m. p. , peacock j. a. , efstathiou g. , 2001 , mnras , 326 , 649 ramella m. , geller m. j. , pisani a. , da costa l. n. , 2002 , aj , 123 , 2976 seljak u. , 2000 , mnras , 318 , 203 sheth r. k. , diaferio a. , 2001 , mnras , 322 , 901 sheth r. k. , jain b. , 2003 , mnras , 345 , 529 sheth r. k. , tormen g. , mnras , 308 , 119 somerville r. s. , lemson g. , sigad y. , dekel a. , kauffmann g. , white s. d. m. , 2001 , mnras , 320 , 289 spergel d. n. et al . , 2003 , apjs , 148 , 175 stoehr f. , white s. d. m. , tormen g. , springel v. , 2002 , mnras , 335 , l84 thoul a. a. , weinberg d. h. , 1996 , apj , 465 , 608 tormen g. , diaferio a. , syer d. , 1998 , mnras , 299 , 728 van den bosch f. c. , yang x.
h. , mo h.j .
, 2003 , mnras , 340 , 771 vogeley m. s. , geller m. j. , park c. , huchra j. p. , 1994
, aj , 108 , 745 wechsler r. h. , somerville r. s. , bullock j. s. , kolatt t. s. , primack j. r. , blumenthal g. r. , dekel a. , 2001 , apj , 554 , 85 weller j. , ostriker j. p. , bode p. ,
astro - ph/0405445 , submitted to mnras white m. , hernquist l. , springel v. , 2001 , apj , 550 , l129 yang x. h. , mo h. j. , van den bosch f. c. , 2003 , mnras , 339 , 1057 yoshikawa k. , taruya a. , jing y. p. , suto y. , 2001 , apj , 558 , 520 zheng z. , tinker j. l. , weinberg d. h. , berlind a. a. , 2002 , apj , 575 , 617 | in this paper we present a new , essentially empirical , model for the relation between the mass of a dark matter halo / subhalo and the luminosity of a galaxy hosted in it . to estimate this , we replace the assumption of linearity between light and mass fluctuations with the assumption of monotonicity between galaxy light and halo or subhalo mass .
we are enabled to proceed with this less restrictive ansatz by the availability of new , very high resolution dark matter simulations and more detailed and comprehensive global galactic luminosity functions .
we find that the relation between halo / subhalo mass and hosted galaxy luminosity , is fairly well fit by a double power law .
that between halo mass and group luminosity has a shallower slope for an intermediate mass region , and is fairly well fit by a two branch function , with both branches double power laws .
both relations asymptote to @xmath0 at low @xmath1 , while at high mass the former follows @xmath2 and the latter @xmath3 .
in addition to the mass - luminosity relation , we also derive results for the occupation number , luminosity function of cluster galaxies , group luminosity function and multiplicity function .
then , using a prescription for the mass function of haloes in under / overdense regions and some further assumptions on the form of the mass density distribution function , we further derive results for biasing between mass and light and mass and galaxy number , light distribution function and the void probability distribution .
our results for the most part seem to match well with observations and previous expectations .
we feel this is a potentially powerful way of modelling the relation between halo mass and galaxy luminosity , since the main inputs are readily testable against dark matter simulation results and galaxy surveys , and the outputs are free from the uncertainties of physically modelling galaxy formation .
galaxies : haloes cosmology : theory dark matter
large - scale structure of the universe |
it is a famous result that a real brownian motion conditioned in doob s sense to remain positive , is distributed as a bessel @xmath0 process , i.e. as the radial process of a @xmath0-dimensional brownian motion .
more generally , if one considers the adjoint action of a compact lie group on its lie algebra , the radial part process of a brownian motion in the lie algebra is distributed as a brownian motion in a cartan subalgebra conditioned to remain in a weyl chamber . in this paper , we consider the coadjoint action of a loop group of a compact group on the dual of the corresponding centrally extended loop algebra and prove that a brownian motion in a cartan subalgebra conditioned to remain in an affine weyl chamber - which can be seen as a space time conditioned brownian motion - is distributed as the radial part process of a brownian sheet on the underlying lie algebra .
let us be more precise .
let @xmath1 be a connected compact lie group , @xmath2 its lie algebra , and @xmath3 a cartan subalgebra .
one considers a weyl chamber in @xmath3 .
then the orbits of @xmath2 under the adjoint action of @xmath1 are parametrized by the weyl chamber .
actually for any @xmath4 , it exists a unique vector in the weyl chamber which is in the same orbit as @xmath5 .
this vector is called the radial part of @xmath5 .
the lie algebra @xmath2 is equipped with an @xmath6-invariant scalar product , whose restriction to the cartan subalgebra is invariant for the action of the weyl group .
if we consider the radial part process of a standard brownian motion on @xmath2 , it is a classical fact that this process is distributed as its projection on the cartan subalgebra @xmath3 - which is a brownian motion on @xmath3 - conditioned in doob s sense to remain forever in the weyl chamber .
the kirillov s character formula is at the heart of the connection between the two processes .
an affine lie algebra can be realized as a central extension of a loop algebra @xmath7 .
considering the coadjoint action of the loop group @xmath8 on the dual of the centrally extended loop algebra , one defines the radial part of an element of this dual algebra ( see pressley and segal @xcite ) .
frenkel established in @xcite a kirillov character type formula in the framework of affine lie algebras considering a gaussian measure on the dual of @xmath7 - basically a brownian motion on @xmath2 . in his approach , the conditional law of a brownian motion on @xmath2 given
its radial part provides a natural measure on the corresponding orbit under the action of @xmath8 .
it highly suggested that one could construct a process on a loop algebra , whose the radial part process would be distributed as a projection on a cartan subalgebra , conditioned to remain in an affine weyl chamber .
the affine weyl chamber is a fundamental domain for the action of the weyl group on the tits cone .
the first difficulty is that there is no euclidean structure on the cartan subalgebra of an affine lie algebra , which would be invariant for the action the weyl group .
so there is no natural brownian motion to consider on it .
the kirillov s orbit method , on which we based our intuition , suggests a connection between coadjoint orbits for the action of a loop group on the dual of an affine lie algebra and irreducible representations of the affine lie algebra .
tensor product of irreducible representations of an affine lie algebra makes appear a drift in the direction of the fundamental weight @xmath9 .
the idea is to consider a process with such a drift in the direction of @xmath9 - which can be seen as a time component - living in an affine lie algebra . in this paper
, we construct such a process , considering a brownian sheet on the lie algebra @xmath2 and prove that the corresponding radial part process is distributed as a projection on a cartan subalgebra , conditioned to remain forever in an affine weyl chamber .
the paper is organized as follows . in section [ action orbits ]
we describe the orbits for the coadjoint action of a loop group of a compact group on the dual of the associated centrally extended loop algebra .
in particular we define a notion of radial part for this action , which is suitable for our context . in section [ affine ]
we briefly recall the necessary background on representation theory of affine lie algebras .
sections [ cond - law - end - point ] and [ kirillov - frenkel ] are basically a reformulation of the main results of @xcite . in section [ cond - law - end - point ]
we compute the conditional law of a brownian motion indexed by @xmath10 $ ] given the end point of its stochastic exponentiel . in section [ kirillov - frenkel ]
we recall how frenkel proves that this conditional law leads to a kirillov character formula for affine lie algebras . in section [ cond - affine ]
we introduce a brownian motion on a cartan subalgebra of an affine lie algebra conditioned - in doob s sense - to remain forever in an affine weyl chamber .
we prove in section [ cond - rad ] that this conditioned doob process has the same law as the radial part process of a brownian sheet on @xmath2 .
[ [ loop - group - and - its - action ] ] * loop group and its action * + + + + + + + + + + + + + + + + + + + + + + + + + + + the following presentation is largely inspired by the one given in @xcite .
let @xmath1 be a connected , simply connected , compact lie group and @xmath11 its lie algebra . by compactness , without loss of generality
, we suppose that @xmath1 is a matrix lie group .
the adjoint action of @xmath1 on itself , which is denoted by @xmath12 , is defined by @xmath13 , @xmath14 .
the induced adjoint action of @xmath1 on its lie algebra @xmath11 is still denoted by @xmath12 and is defined by @xmath15 , @xmath16 , @xmath17 .
the lie bracket on @xmath11 is denoted by @xmath18_\mathfrak{k}$ ] .
we equipp @xmath2 with an @xmath19invariant inner product @xmath20 , i.e. the negative of the killing form .
we denote by @xmath21 the identity of @xmath1 .
we consider the group @xmath22,k)$ ] of borel measurable maps from @xmath10 $ ] to @xmath1 , the group law being pointwise composition , and the subgroup @xmath8 of smooth loops from @xmath10 $ ] to @xmath1 , @xmath23\to k,\ , f\textrm { is smooth } , \ ,
f(0)=f(1)\}.\ ] ] the lie algebra @xmath24 over @xmath25 of @xmath8 is the loop algebra of smooth loops from to @xmath10 $ ] to @xmath2 .
we define a centrally extended lie algebra @xmath26 equipped with a lie bracket given by @xmath27=[\xi,\eta]_\mathfrak{k}+\omega(\xi,\eta)c,\ ] ] for @xmath28 , @xmath29 , where @xmath30 is the cocycle defined by @xmath31 and @xmath32_\mathfrak{k}$ ] is defined pointwise .
a cartan subalgebra of the extended loop algebra is @xmath33 , where @xmath3 is identified with the set of @xmath3-valued constant loops .
the lie bracket actually defines a lie algebra action of @xmath7 on @xmath34 given by @xmath35+\omega(\xi,\eta)c,\ ] ] for any @xmath36 .
this action comes from the adjoint action of @xmath8 on @xmath34 defined by @xmath37 for any @xmath38 .
the corresponding coadjoint action of @xmath8 on the algebraic dual @xmath39 , where @xmath40 and @xmath41 , is defined by @xmath42+\lambda\lambda_0,\end{aligned}\ ] ] for any @xmath43 , where
@xmath44 stands for the linear form defined by @xmath45 notice that the coadjoint action of the loop group does nt affect the level , i.e. the coordinate in @xmath9 .
let us equipp @xmath7 with the @xmath46-norm , and consider its completion @xmath47,\mathfrak k)$ ] with respect with the @xmath46-norm .
we define @xmath48,k)&=\{\gamma:[0,1]\to k , \gamma\textrm { is absolutely continuous , } \ , \gamma^{-1}\gamma ' \in l_2([0,1],\mathfrak k),\ , \gamma(0)=\gamma(1 ) \}\\
k)&=\{f:([0,1]\to\mathfrak k , f\textrm { is absolutely continuous } , f'\in l_2([0,1],\mathfrak k ) , \ , f(0)=0\},\\ h^1([0,1],k)&=\{h:[0,1]\to k , h\textrm { is absolutely continuous , } \ , h^{-1}h ' \in l_2([0,1],\mathfrak k),\ , h(0)=e \}.\end{aligned}\ ] ] if we denote by @xmath49,\mathfrak k))'$ ] the topological dual of @xmath47,\mathfrak k)$ ] then ( [ coadjoint ] ) defines an action of @xmath50,k)$ ] on @xmath49,\mathfrak k))'\oplus \r\lambda_0 $ ] , given by @xmath51 for any @xmath52,k)$ ] , where @xmath53 is a linear form defined on @xmath47,\mathfrak k)$ ] by @xmath54 for any @xmath55,\mathfrak k)$ ] , and @xmath56,\mathfrak k)$ ] .
this action gives rise to an action @xmath57 of @xmath50,k)$ ] on @xmath58,\mathfrak{k})\oplus \r\lambda_0 $ ] defined by @xmath59 for any @xmath52,k)$ ] , @xmath56,\mathfrak{k})$ ] and @xmath60 , which satisfies @xmath61 there is another way to make this action appear naturally .
one defines a group action of the loop group @xmath50,k)$ ] on @xmath58,k)$ ] by letting for any @xmath52,k)$ ] , @xmath62 for @xmath63,k)$ ] and @xmath64 $ ] .
the set @xmath58,k)$ ] is in one to one correspondence with the set @xmath58,\mathfrak k)$ ] .
actually for @xmath65 , this is a classical result of differential geometry that for any path @xmath56,\mathfrak{k})$ ] it exists a unique @xmath66,k)$ ] such that @xmath67 . for any @xmath56,\mathfrak{k}),$ ] and @xmath65 , one defines @xmath68 as the unique map in @xmath58,k)$ ] which satisfies this differential equation . as @xmath69 the action of @xmath50,k)$ ] on @xmath58,k)$ ] , and
the action of @xmath50,k)$ ] on @xmath58,\mathfrak{k})\oplus \r\lambda_0 $ ] , satisfy @xmath70 when @xmath71 .
[ [ roots - and - weights ] ] * roots and weights * + + + + + + + + + + + + + + + + + + + we choose a maximal torus @xmath72 of @xmath1 and denote by @xmath73 its lie algebra .
we denote by @xmath74 the complexification of @xmath11 , i.e. @xmath75 .
we consider the set of real roots @xmath76=2i\pi\alpha ( h)x\}.\ ] ] we choose a set @xmath77 of simple roots of @xmath78 and denote by @xmath79 the set of positive roots .
the half sum of positive roots is denoted by @xmath80 . letting for @xmath81 , @xmath82=2i\pi\alpha(h)x\},\ ] ] the coroot @xmath83 of @xmath84
is defined to be the only vector of @xmath73 in @xmath85 $ ] such that @xmath86 .
the dual coxeter number denoted by @xmath87 is equal to @xmath88 , where @xmath89 is the highest coroot .
we denote respectively by @xmath90 and @xmath91 the root and the coroot lattice .
the weight lattice @xmath92 is denoted by @xmath93 and the set @xmath94 of dominant weights is denoted by @xmath95 .
[ [ orbits - and - radial - part - of - a - path - from - h101mathfrak - k . ] ] * orbits and radial part of a path from @xmath58,\mathfrak k)$ ] . * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let @xmath96,\mathfrak{k})$ ] and @xmath97 . 1 . for any @xmath98,k)$ ] , one has @xmath99 if and only if @xmath100 2 .
it exists @xmath98,k ) , $ ] such that @xmath101 if and only if it exists @xmath102 such that @xmath103 the first point comes from identity ( [ correspondence ] ) . for the second
we write that if @xmath104 , and @xmath105 then @xmath52,k)$ ] and @xmath106 . for @xmath84 the fundamental reflection @xmath107
is defined on @xmath73 by @xmath108 we consider the extended affine weyl group generated by the reflections @xmath109 and the translations by @xmath83 , @xmath110 for @xmath84 .
the fundamental domain for its action on @xmath73 is @xmath111 ( see for instance section 4.8 of @xcite ) . for @xmath112 ,
one defines @xmath113 as the adjoint orbit through @xmath5 , i.e. @xmath114 we consider the exponential map @xmath115 . as @xmath1
is simply connected , the set of conjugacy classes @xmath116 is in one - to - one correspondence with the fundamental domain @xmath117 , i.e. for all @xmath102 , it exists a unique @xmath118 such that @xmath119 ( see @xcite for instance ) .
thus given @xmath120 , every paths in @xmath58,\mathfrak k)$ ] is conjugated to a straight path @xmath121\mapsto sy,\ ] ] for some @xmath122 , and one can give the following definition for the radial part of @xmath56,\mathfrak k)$ ] , given a positive level @xmath97 .
for @xmath123,\mathfrak k)\times \r_+^*$ ] , one defines the radial part of @xmath124 as the unique element @xmath125 in @xmath117 such that @xmath126 .
it is denoted by @xmath127 .
[ [ radial - part - of - a - continuous - semi - martingale ] ] * radial part of a continuous semi - martingale * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the aim of this part is to define the radial part of a @xmath2-valued brownian path .
such a path is not in @xmath58,\mathfrak k)$ ] but one can define a stochastic exponential of a brownian path , which allows us to define its radial part , by analogy with what we ve done above .
let @xmath128},\p)$ ] be a filtered probability space .
the following results have been proved for instance in @xcite or @xcite .
if @xmath129}$ ] is an @xmath2-valued continuous semi - martingale and @xmath130 , then the stochastic differential equation @xmath131 where @xmath132 stands for the stratonovitch integral , has a unique solution starting from @xmath21 .
such a solution is a @xmath1-valued process , that we still denote by @xmath133}$ ] .
this is the stratonovitch stochastic exponential of @xmath134 .
note that given @xmath97 , @xmath5 can be recovered from @xmath68 as @xmath135 . for @xmath97 , and @xmath136}$ ] a continuous @xmath2-valued semi - martingale
, one defines the radial part of @xmath137 as the unique element @xmath125 in @xmath117 such that @xmath126 .
it is denoted by @xmath127 .
we denote by @xmath138 the kernel of the restriction @xmath139 and by @xmath140 the set of integral weights @xmath141 , which is included in @xmath93 since @xmath142 ( see @xcite for instance ) .
thus , we define an application @xmath143 on @xmath72 , when @xmath144 , by letting @xmath145 , for @xmath146 .
the irreducible representations of @xmath1 are parametrized by the set @xmath147 , where @xmath148 . here
, @xmath1 is supposed to be simply connected , so that @xmath149 and @xmath150 .
we denote by @xmath151 the character of the irreducible representation with highest weight @xmath152 .
we recall that @xmath3 is equipped with a @xmath153-invariant scalar product @xmath20 .
we identify and @xmath3 and @xmath154 via @xmath20 and still denote by @xmath20 the scalar product on @xmath155 we write sometimes @xmath156 instead of @xmath157 , for @xmath4 .
[ [ brownian - motion - on - k ] ] * brownian motion on @xmath1 * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let @xmath158},\mathbb p)$ ] be a probability space , where @xmath159}$ ] is the natural filtration of a @xmath11-valued standard continuous brownian motion @xmath160}$ ] defined on @xmath161 , with variance @xmath162 . in this section we consider the stochastic exponential @xmath163 only for @xmath164 . as for any @xmath97 , @xmath165
, there is no loss of generality . in the sequel we let @xmath166 .
the stochastic exponential @xmath167}$ ] is a left levy process on @xmath1 starting from @xmath21 , with transition probability @xmath168}$ ] with respect to the haar measure on @xmath1 , defined on @xmath1 by @xmath169 @xmath170 $ ] , @xmath171 . in the sequel
we write @xmath172 instead of @xmath173 .
this process is a brownian motion on @xmath1 .
the following proposition states a girsanov formula for a brownian motion on a compact lie group .
it is proved for instance in @xcite , @xcite or @xcite . for a @xmath2-valued @xmath46 function @xmath174 , and a @xmath2-valued continuous semi - martingale @xmath129}$ ] ,
@xmath175 is defined as the stochastic integral of @xmath174 with respect to @xmath5 , for any @xmath64 $ ] .
[ girsanov ] let @xmath176}$ ] be a brownian motion on @xmath2 , with variance @xmath177 , and @xmath63,k)$ ] . if @xmath178 is the law of @xmath167}$ ] , then @xmath179 where @xmath180 is the law of @xmath181 . for @xmath182 , we write @xmath183 , for the probability defined on @xmath184 as the conditional probability @xmath185 .
one has for any @xmath186 , @xmath187 note that under @xmath183 , @xmath188 remains a continuous semi - martingale ( see for instance @xcite , theorem 14 ) , so that the stochastic integral @xmath189 is well defined under @xmath190 , for any @xmath2-valued @xmath46 function @xmath174 and @xmath191 .
theorem [ girsanov ] implies the following proposition , which appears in proposition ( 5.2.12 ) of @xcite .
[ condendpoint ] let @xmath192}$ ] be a standard brownian motion on @xmath2 with variance @xmath162 and @xmath174 be a @xmath2-valued @xmath46 function . if @xmath63,k)$ ] satisfies @xmath193 then for any @xmath191 , @xmath194 for @xmath191 , @xmath195 let @xmath196 such that @xmath193 .
theorem [ girsanov ] implies that @xmath197 which gives the proposition .
the following lemma has been proved in @xcite .
[ endorbit ] for @xmath198 , @xmath199 for any @xmath200 , the characters @xmath151 satisfies @xmath201 for every @xmath198 .
thus @xmath202 previous lemma and proposition [ condendpoint ] imply the following one .
[ condorbitendpoint ] let @xmath203}$ ] be a standard brownian motion on @xmath2 starting from @xmath204 , with variance @xmath162 . for @xmath174 a @xmath2-valued @xmath46 function , @xmath205 , and @xmath206 , @xmath207 where @xmath63,k)$ ] satisfies @xmath193 .
@xmath208 which proves the proposition thanks to lemma [ endorbit ] .
in this part , we consider an affine lie algebra whose cartan matrix is an extended cartan matrix of the simple finite dimensional complex lie algebra @xmath74 . such an algebra is a nontwisted affine lie algebra . for our purpose , we only need to consider a realization of its cartan subalgebra . if @xmath209 and @xmath210 are respectively the sets of simple real roots and coroots of @xmath1 previously considered , we let @xmath211 and @xmath212 where @xmath213 the killing form on @xmath2 is from now on normalized such that @xmath214 .
we consider its restriction to @xmath3 and extend it to @xmath215 by @xmath216linearity , and by letting @xmath217 the following definitions mainly come from chapters @xmath218 and @xmath219 of @xcite .
the linear isomorphism @xmath220 identifies @xmath221 and @xmath222 .
we still denote @xmath20 the induced bilinear form on @xmath222 .
we record that @xmath223 notice that here @xmath224 and @xmath225 .
one defines the weyl group @xmath226 , as the subgroup of @xmath227 generated by fundamental reflections @xmath228 , @xmath229 , defined by @xmath230 under the identification of @xmath215 and @xmath231 , the action of the affine weyl group on @xmath221 is defined by @xmath232 , @xmath233 , @xmath234 .
the form @xmath235 is @xmath236-invariant .
the affine weyl group @xmath236 is the semi - direct product @xmath237 ( proposition 6.5 chapter 6 of @xcite ) where @xmath238 is the group of transformations @xmath239 , @xmath240 , defined by @xmath241 thus , if @xmath242 , @xmath243 , and @xmath244 , @xmath245 notice that if one identifies the quotient space @xmath246 with @xmath3 , then @xmath236 is identified with the extended affine weyl group and the fundamental domain of @xmath247 for the action of @xmath236 is identified with the fundamental domain of @xmath3 for the action of the extended affine weyl group .
actually one has @xmath248 where @xmath249 is the fundamental chamber defined by @xmath250 the affine weyl chamber @xmath249 for the @xmath251 type is drawn in figure @xmath218 below : it is the area delimited by yellow and orange half - planes .
the alcove @xmath117 is highlighted in green .
the following definitions and properties mainly come from chapters @xmath252 and @xmath253 of @xcite .
we denote @xmath254 ( resp .
@xmath255 ) the set of integral ( resp .
dominant ) weights defined by @xmath256 @xmath257 where @xmath258 is the pairing between @xmath215 and its dual @xmath231 . for @xmath259 we denote @xmath260 the irreducible module with highest weight @xmath261 .
the weyl - kac character s formula ( theorem 10.4 , chapter 10 of @xcite ) states that @xmath262 where @xmath263 is chosen such that @xmath264 , for all @xmath265 and @xmath266 is the dimension of the root space associated to the root @xmath267 .
in particular @xmath268 in the sequel we choose @xmath269 .
writing @xmath270 , the weyl character s formula becomes by identity ( [ actionofwaff ] ) @xmath271 where the sums run over @xmath243 and @xmath242 . letting @xmath272 , @xmath273 ,
the formal character @xmath274 can be seen as a function defined on its region of convergence .
actually the series @xmath275 converges absolutely for every @xmath273 such that @xmath276 .
we denote by @xmath277 its limit ( see chapter @xmath278 of @xcite ) .
the limiting function @xmath279 is analytic on the set @xmath280 ( see chapter @xmath281 of @xcite ) . for @xmath282 ,
we let @xmath283 .
[ [ the - kirillovs - character - formula - for - compact - groups ] ] * the kirillov s character formula for compact groups * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + for @xmath284 , we denote by @xmath285 the coadjoint orbit in @xmath286 through @xmath261 and by @xmath287 the distribution of @xmath288 where @xmath289 is a random variable distributed according the haar measure on @xmath1 . when the restriction of @xmath261 to @xmath3 ( also denoted by @xmath261 ) is a dominant weight , the kirillov s character formula @xcite states that for @xmath4 , one has @xmath290 where @xmath291 for @xmath292 and @xmath293 , one defines @xmath294 the weyl s character formula for compact lie groups states that @xmath295 for any @xmath296 .
thus the kirillov s character formula becomes , by approximation and analytical continuation , @xmath297 for any @xmath292 and any @xmath293 .
this formula gives the fourier transform of the pushforward measure of the measure @xmath298 under the application @xmath299 which is known as the duistermaat - heckman measure associated to @xmath261 .
[ [ a - kirillov - frenkel - character - formula - for - affine - lie - algebras ] ] * a kirillov - frenkel character formula for affine lie algebras * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let @xmath300 be a brownian motion on @xmath2 , with variance @xmath162 , and its stochastic exponential @xmath301 . a kirillov character type formula for affine lie algebras has been proved by frenkel in @xcite .
it can be formulated as in the following theorem .
[ kiraff ] for @xmath302 , and @xmath259 such that @xmath303 , one has @xmath304 where @xmath305 , @xmath306 , and @xmath307 for @xmath296 , @xmath308 , @xmath309 , we let @xmath310 as @xmath236 is the semi - direct product @xmath237 one has also @xmath311 the previous definition seems to be not valid for @xmath174 such that @xmath312 .
actually , in proposition [ phi - char ] , we give another expression for @xmath313 , which shows that it is actually well defined for every @xmath308 .
the poisson formula implies the following lemma . for @xmath296 , and @xmath314 ,
one has @xmath315 where @xmath316 is a positive constant independent of @xmath5 and @xmath317 . as @xmath318 is a weight in @xmath93 if and only if @xmath319 the poisson summation formula gives for every bounded continuous function @xmath320 defined on @xmath154 , @xmath321 where @xmath322 , and @xmath316 is a positive constant independent of @xmath320 .
the lemma follows taking @xmath320 defined by @xmath323 , @xmath324 .
[ phi - char ] for @xmath325 , and @xmath177 , one has @xmath326 where @xmath316 is a positive constant independent of @xmath5 , @xmath174 and @xmath327 . one has to prove that @xmath328 is equal to @xmath329 where @xmath316 is a positive constant independent of @xmath5 , @xmath174 and @xmath327 . on the one hand , on has @xmath330 on the other hand , @xmath331 by analytical continuation , the identity remains true for @xmath308 , @xmath332 , @xmath333 .
propositions [ phi - char ] and [ condorbitendpoint ] imply the following theorem , which implies the kirillov character type formula of theorem [ kiraff ] .
[ endpoint ] for @xmath55,\mathfrak k)$ ] , and @xmath334 , one has for @xmath206 , @xmath335 where @xmath193 and @xmath336 , with @xmath337 .
in particular for @xmath338 , one has @xmath339 let us make some non rigorous remarks about these formulae .
previously we have considered the identification of @xmath47,\mathfrak k)'\oplus \r\lambda_0 $ ] with @xmath340,\mathfrak k)\oplus \r\lambda_0 $ ] letting @xmath341 for @xmath342,\mathfrak k)$ ] . in the first formula
the stochastic integral @xmath343 can be seen as a random linear functional .
its conditional law given @xmath344 has to be thought as a measure on a coadjoint orbit through @xmath345 where @xmath346 is the straight path @xmath347 , @xmath170 $ ] . for the second formula
, we notice that the restriction of @xmath348 to @xmath3 is equal to @xmath349 .
thus the law of @xmath350 given @xmath344 has to be thought as a duistermaat - heckman distribution associated to @xmath345 .
[ cond - affine ] let us consider a standard brownian motion @xmath352 on @xmath353 .
we consider a random process @xmath354 on @xmath355 as follows
. for @xmath356 , denote @xmath357 a probability under which @xmath358 for all @xmath359 , and @xmath352 is a standard brownian motion starting from @xmath360 .
we let @xmath361 , for @xmath362 .
the process @xmath363 is a space - time brownian motion , @xmath364 being the time component .
let us consider the stopping time @xmath365 [ mart ] let @xmath366 and @xmath352 be a standard brownian motion on @xmath3 . for @xmath367
, @xmath368 is a true martingale under @xmath369 .
let us suppose that @xmath370 .
choose an orthonormal basis @xmath371 of @xmath73 and consider for @xmath372 a function @xmath373 defined on @xmath374 by @xmath375 where @xmath376 .
letting @xmath377 , the function @xmath373 satisfies @xmath378 which implies that @xmath379 is a local martingale .
its quadratic variation is easily shown to be integrable , so that , it is a true martingale .
the sum @xmath380 converges for the @xmath381 norm , which implies that @xmath382 is also a true martingale .
continuity in @xmath174 ensures that lemma is true for any @xmath367 .
let @xmath383 . if @xmath384 , i.e. @xmath385 , then @xmath386 with equality if and only if @xmath387 is on the boundary of @xmath388 , i.e. @xmath5 is on the boundary of @xmath389 .
proposition [ phi - char ] implies that for any @xmath390 , @xmath391 where @xmath316 is a positive constant , which ensures in particular that @xmath392 for @xmath393 , with equality if and only if @xmath394 .
let @xmath395 be the natural filtration of @xmath363 . as @xmath396 is a positive true martingale under @xmath397 such that @xmath398 , one defines a measure @xmath399 on @xmath400 as below .
let @xmath366 and @xmath401 such that @xmath402 is in the interior of @xmath388 , i.e. @xmath403 in the interior of @xmath117 .
one defines a probability @xmath404 letting @xmath405 actually this conditioned doob process can be obtained as a limit of a space time brownian motion starting in the affine weyl chamber , with drift within the affine weyl chamber , conditioned to remain forever in it , when the drift goes to zero ( see @xcite and @xcite for more details ) .
lemma [ mart ] and the fact that for @xmath308 , @xmath406 imply immediately the following proposition .
[ phiq ] for @xmath407 , @xmath5 in the interior of @xmath408 , and @xmath308 , one has @xmath409 and @xmath410
in this last section we prove that the conditioned doob process in an affine weyl chamber previously introduced , has the same law as the radial part process of a brownian sheet on @xmath2 .
it is stated in theorem [ maintheo ] .
let @xmath411,t\in \r_+}$ ] be a standard brownian sheet on @xmath2 , i.e. for any @xmath412 , @xmath413}$ ] is a @xmath2-valued random process independent of @xmath414,r\le t)$ ] , having the same law as @xmath415}$ ] , which is a standard brownian motion on @xmath2 , with variance @xmath416 . in the sequel
we choose a continuous version of it .
proposition [ pointconvergence ] and corollary [ entrancelaw ] prove the existence of an entrance law for the conditioned process in the affine weyl chamber introduced in section [ h - process ] , and the entrance point @xmath204 .
[ conv1 ] let us fix @xmath383 .
if @xmath421 is the law of @xmath422 under @xmath423 then for any @xmath308 , @xmath424 converges towards @xmath425 when @xmath426 goes to @xmath204 within the affine weyl chamber . if @xmath421 is the law of @xmath422 under @xmath427 then @xmath428 proposition [ phi - char ] and the fact that @xmath429 imply that @xmath430 converges to @xmath431 , uniformly in @xmath432 , when @xmath433 goes to @xmath204 .
thus lemma follows from lemma [ convergence ] [ pointconvergence ] let @xmath434,t\in \r_+}$ ] be a standard brownian sheet on @xmath2 and @xmath383 . under @xmath435 , @xmath422 converges in law towards the radial part of @xmath436}$ ] , when @xmath426 goes to @xmath204 within the affine weyl chamber .
let @xmath383 . by analytical continuation ,
the second identity of theorem [ endpoint ] remains valid for any @xmath308 . as @xmath436}$ ] is a standard brownian motion on @xmath2 with variance @xmath437 , theorem [ endpoint ] gives for @xmath432 , @xmath438 , @xmath439 in particular the law @xmath318 of @xmath440 satisfies , for any @xmath438 , @xmath441proposition [ phi - char ] and the fact that @xmath429 imply that @xmath442 where @xmath443 , is smooth on @xmath117 .
thus the peter - weyl theorem ensures that for any smooth function @xmath433 defined on @xmath117 , and @xmath432 , @xmath444 where @xmath445 , and the convergence stands uniformly and absolutely .
actually @xmath446 for all @xmath447 . as @xmath448 remains positive on @xmath117 , @xmath449 as @xmath450 and @xmath451
is bounded on @xmath117 , the uniform and absolute convergence implies that for any probability measure @xmath452 on @xmath117 , @xmath453 expression ( [ seconddef ] ) for the definition of @xmath454 gives , @xmath455 for any @xmath200 .
thus one has @xmath456 and identity ( [ idnu ] ) becomes for @xmath457 , @xmath458 identity ( [ idnu ] ) becomes for @xmath459 @xmath460 as @xmath461 lemma [ conv1 ] implies that @xmath462 as @xmath426 goes to @xmath204 within the affine weyl chamber , which proves the proposition .
the weyl integration formula implies that the density of the radial part of @xmath464 is equal to @xmath465 , up to a multiplicative constant .
thus corollary follows from propositions [ pointconvergence ] and [ phi - char ] . under @xmath466 , @xmath471 is equal in law to @xmath472 for any @xmath383 . thus for any @xmath383 , @xmath471 is equal in law to @xmath473 .
as @xmath474 converges towards the haar measure on @xmath1 when @xmath317 goes to @xmath204 , one obtains by the weyl integration formula that @xmath475 which can be also deduced from ( [ q0 + ] ) . [ intertwining ] let @xmath411,t\in \r_+}$ ] be a standard brownian sheet on @xmath2 .
let @xmath138 be a kernel on @xmath477,{\mathfrak k } ) ) $ ] , such that for any @xmath478 , @xmath479 is the law of a brownian motion @xmath480}$ ] on @xmath2 , given @xmath481 .
let @xmath482 be the transition probability of the markov process @xmath483 under @xmath484 , and @xmath485 the transition probability of @xmath486 .
then for any @xmath383 and @xmath487 , @xmath488 it is sufficient to prove that for every measurable fonction @xmath55,\mathfrak k)$ ] , @xmath487 , and @xmath489 , @xmath490 let @xmath55,\mathfrak k)$ ] , @xmath487 , @xmath489 , @xmath63,k)$ ] such that @xmath193 , and @xmath337 such that @xmath491 . on the one hand
, one has @xmath492 on the other hand , @xmath493 for the following theorem we consider a continuous version of the doubly indexed process @xmath494,t\in\r_+^*}$ ] ( see for instance @xcite and references therein for the existence of a continuous version ) . applying the criterion described by rogers and pitman in @xcite ,
the theorem is a consequence of the proposition [ pointconvergence ] and the intertwining relation established in proposition [ intertwining ] . | we consider the coadjoint action of a loop group of a compact group on the dual of the corresponding centrally extended loop algebra and prove that a brownian motion in a cartan subalgebra conditioned to remain in an affine weyl chamber - which can be seen as a space time conditioned brownian motion - is distributed as the radial part process of a brownian sheet on the underlying lie algebra .
this paper is dedicated to professor philippe bougerol to thank him for many discussions which have been a great source of joy and inspiration to me , and a valuable aid especially in the work presented here . |
time - reversal - invariance non - conservation has for the first time been unequivocally demonstrated in a direct measurement at cplear.[1 ] the experiment measured directly the difference in the transition probabilities @xmath3 and @xmath4 .
a non - zero value of this difference gives model independent evidence of time - reversal - invariance non - conservation .
the deduction does not depend on the validity of the @xmath5 rule .
the result obtained for a@xmath6 with @xmath7 is in good agreement with the measure of cp violation in neutral kaon decay .
starting with cpt conservation and the well established non - conservation of cp in kaon decays , time - reversal - invariance should also be broken .
the cplear measurement is the first direct confirmation of that .
the question that one then can ask is : what about time - reversal - invariance non - conservation in systems other than the kaon system ?
tests of time - reversal - invariance can be distinguished as belonging to two classes : the first one deals with time - reversal - invariance - odd ( t - odd)/parity violating ( p - odd ) interactions , while the second one deals with t - odd / p - even interactions ( assuming cpt invariance this implies c - conjugation non - conservation ) .
however , it should be noted that constraints on these two classes of interactions are not independent since the effects due to t - odd / p - odd interactions may also be produced by t - odd / p - even interactions in conjunction with standard model parity violating radiative corrections .
the latter can occur at the @xmath8 level and consequently this presents a limit on the constraint of t - odd / p - even interactions , which can be derived from experiments .
limits on a t - odd / p - odd interaction follow from measurements of the electric dipole moment of the neutron ( which currently stands at @xmath9 e.cm [ 95% c.l . ] ) .
it provides a limit on a t - odd / p - odd pion - nucleon coupling constant which is less than @xmath1 times the weak interaction strength .
measurements of atomic electric dipole moments of @xmath10xe and @xmath11hg ( @xmath12 e.cm [ 95% c.l . ] ) give similar constraints .
[ see ref .
2 ] experimental limits on a t - odd / p - even interaction are much less stringent . following the standard approach of describing the nucleon - nucleon interaction in terms of meson exchanges ,
it can be shown that only charged rho - meson exchange and a@xmath2-meson exchange can lead to a t - odd / p - even interaction.[3 ] the better constraints stem from measurements of the electric dipole moment of the neutron and from measurements of charge symmetry breaking in neutron - proton elastic scattering .
all other experiments , like gamma decay experiments [ 4 ] , detailed balance experiments [ 5 ] , polarization - analyzing power difference measurements , and five - fold correlation experiments with polarized incident nucleons and aligned nuclear targets , have been shown to be at least an order of magnitude less sensitive .
haxton , hoering , and musolf [ 2 ] have deduced constraints on a t - odd / p - even interaction from nucleon , nuclear , and atomic electric dipole moments with the better constraint coming from the electric dipole moment of the neutron . in terms of a ratio to the strong rho - meson nucleon coupling constant , they deduced for the t - odd / p - even rho - meson nucleon coupling : @xmath13 . here
one should note that the ratio of the theoretical to the measured value of @xmath14 may be as large as 15 ! [ 6 ] in the standard model a t - odd / p - even nucleon - nucleon interaction can hardly be accommodated .
it requires c - conjugation non - conservation , which can not be introduced at the first generation quark level .
it can neither be introduced into the gluon self - interaction .
consequently one needs to consider c - conjugation non - conservation between quarks of different generations and/or between interacting fields.[7 ]
the nucleon - nucleon scattering matrix , assuming conservation of angular momentum , parity , time - reversal - invariance , and isospin , consists of five complex amplitudes , a , b , c , d , and e , which are functions of centre - of - mass energy e and scattering angle @xmath15 .
if isospin is broken ( which leads to charge symmetry breaking in the neutron - proton system ) , the neutron - proton scattering matrix contains a sixth complex amplitude , f. if in addition one no longer assumes time - reversal - invariance , the neutron - proton scattering matrix has two additional amplitudes , g and h ; the second one of these , h , is simultaneously time - reversal - invariance violating and charge symmetry breaking .
[ 8 ] @xmath16 where @xmath17 , @xmath18 , and @xmath19 are unit vectors given as @xmath17 = @xmath20 + @xmath21 + @xmath22 ; @xmath18 = @xmath23 - @xmath24 - @xmath25 ; @xmath19 = @xmath20 @xmath26 @xmath21 @xmath26 @xmath22 and @xmath27 and @xmath28 are the initial and final state centre - of - mass nucleon momenta . @xmath29 and @xmath30 are pauli spin matrices for the two nucleons .
the proton - proton system may only contain one additional , sixth time - reversal - invariance non - conserving amplitude , g. in a partial wave decomposition the four lowest , parity conserving , transition amplitudes in which time - reversal - invariance violation may occur are @xmath31 @xmath32 @xmath33 , @xmath34 , @xmath35 , @xmath36 . of these only the latter one is allowed for identical nucleons because of the pauli exclusion principle . thus time - reversal - invariance violating effects ,
if these exist , are strongly suppressed in the proton - proton and neutron - neutron systems . as shown by arash , moravcsik , and goldstein [ 9 ] null tests of time - reversal - invariance
do not exist in a two - particle in and two - particle out reaction .
observables are bilinear product combinations of scattering amplitudes
. tests of time - reversal - invariance can be accomplished only through a comparison of two distinct observables . as prime example
, polarization - analyzing power difference determinations , @xmath37 , are based on two independent measurements with their own systematic errors in polarimeter analyzing power and beam polarization calibrations .
the best calibration standards todate carry uncertainties of a few parts in @xmath38 . writing the polarization - analyzing power difference as @xmath39 , with @xmath40 the normalized spin - flip differential cross sections difference and with d the depolarization parameter
, one can show that most such difference determinations for the neutron - proton system were made in an angular region where d is close to one [ 10 ] , causing a measure of insensitivity to time - reversal - invariance non - conservation .
similarly , time - reversal - invariance tests performed in the proton - proton system are rather inconclusive.[8 ] as noted by conzett [ 11 ] transmission experiments are not included in the nonexistence proof of arash , moravcsik , and goldstein .
indeed , since the total cross section can be expressed in terms of the imaginary part of the forward scattering amplitude , null tests of time - reversal - invariance can be devised for transmission experiments ( a selected spin - dependent total cross section linearly dependent on the time - reversal - non - invariant amplitude ) .
transmission experiments that test a t - odd / p - even interaction require spin 1/2 particles incident on target particles with spin @xmath41 or vice versa .
charge symmetry breaking ( csb ) in neutron - proton elastic scattering manifests itself as a non - zero difference of the neutron @xmath42 and proton @xmath43 analyzing powers , @xmath44/\sigma_0 $ ] .
the three precision experiments performed ( at triumf at 477 mev [ 12 ] and at 369 mev [ 13 ] , and at iucf at 183 mev [ 14 ] ) have unquestionably shown that charge symmetry is broken and that the results for @xmath45 at the zero - crossing angle of the average analyzing power , are very well reproduced by meson exchange model calculations ( see fig.1 ) .
as shown above a t - odd / p - even interaction corresponds to a term in the scattering amplitude which is simultaneously charge symmetry breaking .
thus , simonius [ 15 ] deduced an upper limit on a t - odd / p - even csb interaction from a comparison of the experimental results with the theoretical predictions for the above mentioned three csb experiments .
the upper limit so derived is @xmath46 [ 95% c.l . ] .
this is therefore comparable to the upper limit deduced from the electric dipole moment of the neutron , taking the present experimental limit of @xmath14 , and is considerably lower than the limits inferred from direct tests of a t - odd / p - even interaction .
for instance the detailed balance experiments give a limit on @xmath47 . [
16 ] as remarked above , it is inconceivable in the standard model to account for a t - odd / p - even interaction .
nevertheless , there is a need to clarify the experimental constraint on a t - odd / p - even interaction by providing a better experimental limit .
such a better experimental constraint may be provided by an improved upper limit on the electric dipole moment of the neutron .
in fact a new measurement with a sensitivity of @xmath48 e.cm has been proposed at the los alamos neutron science center.[17 ] but performing an improved n - p elastic scattering csb experiment appears to be a very attractive alternative .
one can calculate with a great deal of confidence the contributions to csb due to one - photon exchange ( the neutron magnetic moment interacting with the current of the proton ) and due to the n - p mass difference affecting charged one - pion and rho - meson exchange .
furthermore , one can select an energy where the @xmath49 meson mixing contribution changes sign at the same angle where the average of the analyzing powers a@xmath50 and a@xmath51 changes sign and therefore does not contribute .
this occurs at an incident neutron energy of 320 mev and is caused by the particular interplay of the n - p phase shifts and the form of the spin / isospin operator for the @xmath49 mixing term .
but also the contribution due to one - photon excange changes sign at about the same angle at 320 mev . the contribution due to two - pion exchange with an intermediate @xmath52
is expected to be no more than one tenth of the overall csb effect , essentially presenting an upper limit on the theoretical uncertainty ( see fig .
2).[18 ] it has been shown that simultaneous @xmath53 exchanges can only contribute through second order processes and can therefore be neglected.[19 ] also the effects of inelasticity are neglegibly small at 320 mev .
it appears therefore well within reach to reduce the theoretical uncertainty in the comparison between experiment and theory .
subtracting the calculated difference in the neutron and proton analyzing powers from the measured difference in these permits establishing an upper limit on a p - even / t - odd / csb interaction . the second triumf experiment measuring csb in n - p elastic scattering at 347 mev obtained the result @xmath52a@xmath54 at the zero - crossing angle of the average of @xmath55 and @xmath56 . in the experiment
polarized neutrons were scattered from unpolarized protons and vice versa .
the polarized ( or unpolarized ) neutron beam was obtained using the ( p , n ) reaction with a 369 mev polarized ( or unpolarized ) proton beam incident on a 0.20 m long ld@xmath57 target . at the triumf energies one makes use of the large sideways - to - sideways polarization transfer coefficient @xmath58 at 9 degrees in the lab .
the only difference in obtaining the unpolarized and polarized neutron beams was the turning off of the pumping laser light in the optically pumped polarized ion source ( oppis ) .
the polarized proton target was of the frozen spin type with butanol beads as target material . the same target after depolarization was used as the unpolarized proton target .
great care was taken that the two interleaved phases of the experiment were performed with identical beam and target parameters except for the polarization states .
scattered neutrons and recoiling protons were detected in coincidence in the c.m .
angular range 53.4 to 86.9 degrees in two left - right symmetric detector systems .
rather than measuring @xmath55 and @xmath56 directly ( which would be troubled by not having polarization calibration standards of the required precision ) , the zero - crossings of @xmath55 and @xmath56 were determined by fitting the partial angular distributions with polynomials , deduced from n - p phase shift analyses .
the difference @xmath59 followed by multiplying the difference in the zero - crossing angles by the average slope of the analyzing powers ( the experiment measured the slope of @xmath56 at the zero - crossing angle , which is a very good approximation for the average slope at the zero - crossing angle and introduces a negligible error ) .
the principle of the measurement is shown schematically in fig .
3 . the execution of the experiment depended on a great deal of simultaneous monitoring and control measurements ( see fig . 4 for a schematic view of the experiment ) .
both the statistical and systematic errors obtained in the experiment can be considerably improved upon . with the oppis developments that have taken place in the intervening years and with the addition of a biased na - ionizer cell
, it will be possible to obtain up to 50 @xmath60a of beam with a polarization of 80% incident on the neutron production ld@xmath57 target ( a factor of 50 increase in neutron beam intensity over the previous csb experiment).[20 ] a 342 mev proton beam incident on a 0.20 m long ld@xmath57 target would present a heat load of 500 w. lh@xmath57 targets allowing such heat loads have been developed for electron scattering experiments .
however , reducing the ld@xmath57 target thickness to 0.05 m would give a better defined neutron energy spectrum reducing the uncertainty in the difference in apparent energies of the a@xmath50 and @xmath56 measurements .
correcting for the apparent energy difference contributes significantly to the systematic error budget .
the median of the intensity distribution of the proton beam incident on the ld@xmath57 target will again be kept fixed to within 0.05 mm at two places through feedback loops to two sets of steering magnets placed upstream in the beam line .
this freezes position and direction of the proton beam at the ld@xmath57 target , necessary to meet the secondary neutron beam energy and direction stability requirement .
the polarized proton target should again be of the frozen spin type ; choosing a target material with improved ratio of free protons to bound protons and a lower magnetic holding field ( @xmath61 0.2 t ) would greatly improve on the systematic error by reducing the uncertainty in the background subtraction .
it appears entirely feasible to reduce the statistical and systematic errors by a factor three to four .
such an experiment would constitute a measurement of csb in n - p elastic scattering of unprecedented precision , of great value on its own , and would be simultaneously provide the best upper limit on a t - odd / p - even interaction .
as remarked above true null tests of time - reversal - invariance do not exist for spin 1/2 particles scattered from spin 1/2 particles .
however , null tests exist as transmission measurements for spin 1/2 particles interacting with aligned nuclear targets.[11 ] in such tests one measures the total cross section asymmetry a@xmath62 of vector polarized spin 1/2 particles interacting with an aligned nuclear target ( for instance a tensor polarized spin 1 deuteron target ) .
huffman et al [ 16 ] have extracted the five - fold correlation parameter a@xmath62 by observing polarized neutron transmission through nuclear spin - aligned @xmath63ho .
this resulted in an upper limit for @xmath64 of @xmath65 even though the measured value of a@xmath62 was @xmath66 .
it is to be noted that only the valence proton in @xmath63ho contributes to a@xmath62 .
storage rings provide a completely different environment , with the advantages outweighing the disadvantages possibly , for high precision tests of fundamental symmetries ( like testing time - reversal - invariance in the gev range ) , as discussed for instance by s.e .
vigdor.[21 ] at cosy a proton - deuteron transmission experiment has been proposed to measure the t - odd / p - even observable a@xmath62 using a polarized internal proton beam ( polarization p@xmath67 ) and an internal polarized deuterium target p@xmath68 .
tensor polarized deuterium atoms are produced in an atomic beam source based on stern - gerlach separation in permanent sextupole magnets and adiabatic high frequency transitions .
adequate luminosities can be obtained using a window - less storage cell placed on the axis of the proton beam [ 22 ] .
the polarized proton beam is obtained from an atomic crossed beam polarized ion source . for this test of a t - odd / p
- even interaction the cosy ring will serve as accelerator , forward spectrometer , and detector.[23 ] figure 5 presents a schematic view of the cosy facility with the edda internal target and detector system for this t - odd / p - even test .
crucial to the experiment is a current monitor which can register with the required precision the decrease in circulating beam intensity with time as function of the circulating proton beam spin state . also of great importance
is the precise alignment of the proton beam vector and deuteron beam tensor polarizations in order to suppress unwanted spin correlation coefficients , which could produce a false result .
an accuracy of @xmath69 in a@xmath62 has been planned for the experiment , which would give a predicted sensitivity to @xmath70 of @xmath71 and to @xmath72 of @xmath73 for center of mass momenta in the range 200 to 400 mev / c .
the predicted sensitivity is decreasing for higher momenta .
there is a further measure of uncertainty due to the relatively large inelasticity ( pion production ) in the momentum range 2 - 3 gev / c , for which the cosy experiment is planned.[24 ] the dilution effect of coulomb multiple scattering decreases significantly with increasing proton beam energy .
it will be a tour - de - force for the experiment to reach a sensitivity comparable to the present n - p csb experiments .
other searches for a t - odd / p - even interaction are made in particle decays , e.g. , in the decay @xmath74 .
a non - zero value of the muon polarization transverse to the decay plane would be an indication of time - reversal - invariance non - conservation .
several experiments have been performed using both neutral and charged kaons .
there is a unique feature to the transverse muon polarization in that it does not have contributions from the standard model at tree level and that higher order effects are of order @xmath69 . with
only one charged particle in the final state , a final state interaction , which can mimic a time - reversal - invariance breaking effect , is greatly reduced and is estimated to occur at the same level of @xmath69 .
the more recent effort of measuring the time - reversal - invariance non - conserving transverse muon polarization is at kek using a stopped @xmath75 beam .
the experiment reports a result for p@xmath6 = -0.0042 @xmath76 0.0049(stat ) @xmath76 0.0009(syst ) , based on the data taken in 1996 and 1997 , which translates into a value of im@xmath77 = -0.0013 @xmath76 0.0169(stat ) @xmath76 0.0009(syst).[25 ] the quantity @xmath77 is defined as the ratio of the two form factors , @xmath78 ) and @xmath79 , in the k@xmath80 decay matrix element ; im@xmath77 must be equal to zero for time - reversal - invariance to hold.[26 ] with the data allready in hand and with the approved data taking time , it is anticipated to arrive at a statistical error of @xmath81 in im@xmath77 .
the best previous experimental limits were obtained with both neutral and charged kaons at the bnl - ags.[27 ] a combination of both experimental results provided a limit on the imaginary part of the hadron form factors , im@xmath77 = -0.01 @xmath76 0.019 .
a new search for the time - reversal - invariance non - conserving transverse muon polarization with in - flight decays @xmath82 was proposed at the bnl - ags.[28 ] it was intended to obtain a sensitivity to the tranverse muon polarization of @xmath83 , corresponding to a sensitivity to im@xmath77 of @xmath84 .
the possibility of similar searches for the time - reversal - invariance non - conserving transverse @xmath85 polarization in b semileptonic decays , @xmath86 , has been discussed recently by y. kuno.[29 ] it is estimated that the polarization of the @xmath85 leptons could reach as high as as 30% in @xmath87 decays
. a non - zero value of the transverse muon polarization in @xmath88 decay and of the @xmath85 lepton in semi - leptonic b decays would constitute clear evidence for new physics .
in summary , searches for a t - odd / p - even interaction have sofar resulted in only very modest limits on such an interaction .
most promising are the continuing efforts to measure the electric dipole moment of the neutron , to measure charge - symmetry breaking in neutron - proton elastic scattering at around 320 mev , and to measure the five - fold correlation parameter a@xmath62 in a proton - deuteron transmission experiment , as well as searches of transverse lepton polarizations in k and b decays . for a recent review
see : f.boehm in symmetries and fundamental interactions in nuclei , ed . w.c .
haxton and e.m .
henley ( world scientific , singapore , 1997 , p. 67 . ) .
e. blanke et al .
51 , 355 ( 1983 ) .
eversheim et al . , in polarized gas targets and polarized beams ( 1997 ) , ed .
holt and m.a .
miller , aip conf .
421 , p. 419 .
f. hinterberger , lanl archives / nucl - ex/9810003 , unpublished ; cosy proposal , spokesperson p.d . | time - reversal - invariance non - conservation has for the first time been unequivocally demonstrated in a direct measurement at cplear .
one then can ask the question : what about tests of time - reversal - invariance in systems other than the kaon system ?
tests of time - reversal - invariance can be distinguished as belonging to two classes : the first one deals with parity violating ( p - odd)/time - reversal - invariance - odd ( t - odd ) interactions , while the second one deals with p - even / t - odd interactions ( assuming cpt conservation this implies c - conjugation non - conservation ) .
limits on a p - odd / t - odd interaction follow from measurements of the electric dipole moment of the neutron ( with a present upper limit of @xmath0 e.cm [ 95% c.l . ] ) .
it provides a limit on a p - odd / t - odd pion - nucleon coupling constant which is less than @xmath1 times the weak interaction strength .
experimental limits on a p - even / t - odd interaction are much less stringent . following the standard approach of describing the nucleon - nucleon interaction in terms of meson exchanges , it can be shown that only charged rho - meson exchange and a@xmath2-meson exchange can lead to a p - even / t - odd interaction .
the better constraints stem from measurements of the electric dipole moment of the neutron and from measurements of charge - symmetry breaking in neutron - proton elastic scattering .
the latter experiments were executed at triumf ( 497 and 347 mev ) and at iucf ( 183 mev ) .
weak decay experiments may provide limits which will possibly be comparable .
all other experiments , like gamma decay experiments , detailed balance experiments , polarization - analyzing power difference determinations , and five - fold correlation experiments with polarized incident nucleons and aligned nuclear targets , have been shown to be at least an order of magnitude less sensitive .
the question then emerges : is there room for further experimentation ? |
studying near - threshold @xmath9 photoproduction presents an interesting opportunity to search for new baryon resonances .
measurements made by previous experiments have produced relatively high - precision cross sections at most production angles ; however , precise spin density matrix elements have only been measured at very forward angles and at higher energies @xcite . in the near - threshold region ,
the only previously published spin density matrix results , which come from the saphir collaboration , constitute a total of 8 data points in the energy range from @xmath9 photoproduction threshold up to a center - of - mass energy , @xmath10 , of about @xmath11 gev @xcite .
a number of theoretical studies have been undertaken with the goal of extracting resonance contributions to @xmath9 photoproduction from these data .
all of the authors agree on the importance of contributions from @xmath3 exchange in the @xmath2-channel ; however , discrepancies exist on the importance of various resonance contributions . in the calculations of oh _ et al_. @xcite
, the dominant resonance contributions are found to be from a `` missing '' @xmath12 state ( _ i.e. _ a state predicted by the constituent quark model but not observed experimentally ) and from a @xmath13 state .
in contrast to this , titov and lee @xcite find the most significant resonance contributions to be from the @xmath14 and @xmath4 states .
the quark model calculations made by zhao @xcite find that the two most important resonance contributions to @xmath9 photoproduction come from the @xmath15 and @xmath4 states .
the @xmath16 and @xmath17 states were found to be the dominant resonance contributions in the coupled - channel analysis of penner and mosel @xcite .
all of the models mentioned above were fit solely to differential cross sections .
a more recent analysis @xcite also included the spin density matrix elements published by the saphir collaboration @xcite .
this work found the largest resonant contributions to @xmath9 photoproduction to be from the sub - threshold @xmath18 and @xmath4 states .
the authors noted the importance of the strong additional constraints placed on their model by the polarization information and concluded that : _ there is urgent need for precise measurements of the spin density matrix in more narrow energy bins to pin down the reaction picture . _
recently published results from the cebaf large acceptance spectrometer ( clas ) have provided such measurements @xcite . in the center - of - mass ( c.m . )
energy range from threshold up to @xmath19 gev , differential cross section results were reported at @xmath20 points in @xmath10 and @xmath21 .
the experiment did not use a polarized beam or a polarized target ; thus , only the @xmath22 , @xmath23 and @xmath24 elements of the spin density matrix could be determined ( the definitions of which can be found in @xcite ) .
these results were reported at @xmath25 points in @xmath10 and @xmath21 .
the increase in precision for @xmath26 , in the energy range overlapping the saphir results , is approximately a factor of @xmath27 . in this paper , we present an _ event - based mass - independent _ partial wave analysis ( pwa ) of these data , _
i.e. _ the data are only divided into narrow c.m .
energy bins . in each of these narrow bins ,
the spin - independent part of any resonance propagator a complex function of @xmath10 is approximated as a constant complex number .
this allows us to reduce model dependence in our treatment of resonances ( see section [ section : amps : resonant ] for further discussion on this topic ) .
the data used in our analysis were obtained using the clas housed in hall b at the thomas jefferson national accelerator facility .
real photons were produced via bremsstrahlung from a @xmath28 gev electron beam .
the momenta of the recoiling electrons were then analyzed in order to obtain the energy of the photons with an uncertainty of @xmath29% @xcite .
the physics target was filled with liquid hydrogen .
the @xmath30 decay was used to select the reaction of interest .
the momenta of the charged particles ( @xmath31 ) were determined using the clas detector with an uncertainty of approximately @xmath32% .
the neutral @xmath3 was reconstructed using kinematic fitting .
more details concerning the analysis techniques can be found in @xcite .
a detailed description of the clas can be found in @xcite . in total , the dataset consists of over @xmath33 million signal events divided into @xmath34 @xmath33-mev wide c.m .
energy bins .
our primary interest is in extracting possible nucleon resonance contributions ; thus , we have restricted our pwa to include only bins with @xmath35 gev . in total ,
@xmath36 c.m .
energy bins were used in the pwa ( the @xmath37 gev bin was excluded due to issues with the normalization calculation @xcite ) .
this work represents the first event - based pwa results on baryons from photoproduction data .
as stated above , to limit theoretical model dependence we divided our data into 10-mev wide @xmath10 bins .
thus , all formulas written below are intended to describe data from a narrow c.m . energy range . in all of the work that follows
, @xmath38 and @xmath39 will be used for the initial proton , final proton , @xmath9 and photon 4-momenta .
the @xmath40-axis in the overall c.m .
frame , defined by @xmath41 , is used as the angular momentum quantization axis .
the mandelstam variables are defined as [ eq : mandlestam ] @xmath42 the mass of the proton and @xmath9 are denoted as @xmath43 and @xmath44 , respectively . the lorentz invariant transition amplitude , @xmath45 , of the process @xmath46 ,
can be written as @xmath47 where @xmath48 are the initial proton , incident photon and final proton spin projections , @xmath49 are the partial wave amplitudes ( the form of which is discussed in section [ section : amps ] ) , @xmath50 denotes the complete set of kinematic variables describing the reaction and @xmath51 are the unknown parameters to be determined by the fit .
we denote the detector acceptance by @xmath52 and the phase - space volume as @xmath53 .
a more detailed description of the work presented in this section is given in @xcite .
all of our fits are event - based ; thus , the data were only binned in @xmath10 . to obtain estimators for the unknown parameters , @xmath54
, we employ the extended unbinned maximum likelihood method . the work detailed in this section
is based on that of chung @xcite ; however , the normalizations we have developed differ from his work .
the likelihood function is defined as @xmath55 where the term in parentheses is the poisson probability of obtaining @xmath56 events given the expected number @xmath57 ( the calculation of which is discussed below ) , @xmath58 represents the complete set of kinematic variables of the @xmath59 event and @xmath60 is the probability density function given by @xmath61 > from left to right , eq .
( [ eq : pdf ] ) accounts for the relative strength of the transition amplitude , the detection probability and the available phase space for the @xmath59 event .
calculation of the denominator , which normalizes the probability density function , is discussed below .
the estimators @xmath54 are then found by maximizing @xmath62 .
the expected number of signal events for a given set of parameters is given by @xmath63 which includes the average over initial spin states .
@xmath64 is the `` target factor '' obtained from the target density , @xmath65 , length , @xmath66 , and atomic number , @xmath67 ; along with avogadro s number , @xmath68 ; the branching fraction of @xmath69 , @xmath70 ; and the integrated photon flux in each @xmath10 bin , @xmath71 .
the integral in eq .
( [ eq : nbar - def ] ) must be done numerically due to the lack of an analytic expression for the detector acceptance .
monte carlo events were generated in each @xmath10 bin according to @xmath72 ( the @xmath9 mass was generated according to a breit - wigner distribution ) , @xmath73 phase space and then run through a geant - based detector simulation package ( discussed in detail in @xcite ) .
this procedure simulates the acceptance of the detector by rejecting events that would not have survived the data analysis , _
i.e. _ for each generated event , the acceptance factor @xmath74 or @xmath75 .
the integral can then be approximated by @xmath76 where @xmath77 is the number of generated ( accepted ) monte carlo events and @xmath78^{1/2 } } { 4(2\pi)^5 s}\ ] ] is the volume of the 2-body @xmath79 phase space ( the @xmath80 phase - space volume is factored into the normalization of the @xmath9 decay amplitude ) . using eqs .
( [ eq : nbar - integral ] ) and ( [ eq : phase - space ] ) , eq .
( [ eq : nbar - def ] ) can be rewritten as @xmath81^{1/2 } } { 64 \pi s ( s - w_p^2 ) } \nonumber \\
\times \frac{\mathcal{t}(s)}{n_{gen } } \sum\limits_i^{n_{acc } } |\mathcal{m}(\vec{\alpha},\vec{x}_i)|^2.\end{aligned}\ ] ] this normalization allows us to use physical coupling constants in our event - based fits , _
i.e. _ it allows us to put our parameters on an absolute scale .
thus , our normalization scheme permits direct theoretical input . due to the monotonically increasing nature of the natural logarithm , the likelihood , @xmath62 , defined in eq .
( [ eq : likelihood ] ) can be maximized by minimizing @xmath82 which , using eqs .
( [ eq : pdf ] ) and ( [ eq : nbar - def ] ) , can be rewritten as @xmath83 neglecting terms that do not depend on the parameters , we can then rewrite eq .
( [ eq : log - l - def - v2 ] ) as follows : @xmath84 we note here that for any set of estimators that minimize eq .
( [ eq : log - l ] ) , the expected number of events is @xmath85 . to accurately extract partial wave contributions to @xmath9 photoproduction , background events , _
i.e. _ non-@xmath9 events , must be separated from the signal in a way that preserves all kinematic correlations .
the method we applied to our data , described in detail in @xcite , assigned each event a signal quality factor , or @xmath86-factor .
this background was assumed to be non - interfering .
following our previous work @xcite , we can rewrite eq .
( [ eq : log - l ] ) using these @xmath86-factors as @xmath87 where @xmath88 is the @xmath86-factor for the @xmath59 event .
thus , the @xmath86-factors are used to weight each event s contribution to the likelihood .
we also note that in the literature , the @xmath2- and @xmath89-channel contributions are often referred to as _
background_. this _ theoretical _ background is not to be confused with the experimental background discussed here . in this analysis ,
the theoretical backgrounds were allowed to interfere with the @xmath90-channel amplitudes in our pwa .
the choice of which amplitudes to include to describe the data is partially motivated by experimental measurements .
the @xmath9 photoproduction cross section is known to have a strong forward peak , even at near - threshold energies @xcite . at higher energies , the cross section develops a rather pronounced backwards peak as well @xcite .
these features are typically associated with meson and nucleon exchanges in the @xmath2- and @xmath89-channel , respectively .
the recent clas data also possess a number of features in the cross sections and spin density matrix elements suggestive of resonance contributions @xcite .
thus , it would seem that @xmath90- , @xmath2- and @xmath89-channel amplitudes may be required to fully describe the data ( see fig .
[ fig : feynman - diagrams ] ) .
the formalism used to construct these amplitudes is described in detail elsewhere @xcite , below we simply give an overview of the different types of amplitudes used in our analysis .
all of these amplitudes were computed using the qft++ package @xcite , which performs numerical computations of quantum field theory expressions .
+ + the @xmath92 amplitude can be written in terms of the isovectors , @xmath93 , and the 4-momenta , @xmath94 , of the pions , along with the @xmath9 4-momentum ( @xmath95 ) , polarization ( @xmath96 ) , and spin projection , @xmath97 as @xmath98 which is fully symmetric under interchange of the three pions . for this reaction , where all final states contain @xmath69
, the isovector triple product simply contributes a factor to the global phase of all amplitudes . in the @xmath9 rest frame , eq .
( [ eq : omega - decay - amp ] ) simplifies to @xmath99 which is the standard non - relativistic result @xcite .
previous studies of forward @xmath9 photoproduction data have shown that the reaction is dominated at low energies by pion exchange and at higher energies by diffractive processes , _
i.e. _ pomeron exchange ( see , _
e.g. _ , @xcite ) .
we have chosen to use the non - resonant terms included in the model of oh , titov and lee ( otl ) @xcite in our partial wave analysis .
the otl pomeron exchange amplitude follows the work of donnachie and landshoff @xcite with the unknown parameters fixed by fitting to high energy vector meson cross section data .
the otl model also includes pseudoscalar meson exchange amplitudes obtained from the following lagrangians : @xmath100 where @xmath101 , @xmath102 and @xmath103 denote the pseudoscalar , photon and proton fields , respectively .
the vertices in these amplitudes were dressed using form factors of the type @xmath104 where @xmath105 is the cutoff parameter for the interaction and @xmath106 is the mass of the exchanged particle .
the larger mass and weaker coupling constants of the @xmath107 suppress its contribution relative to that of the pion .
nucleon pole terms were obtained from the lagrangians @xmath108 the form factor @xmath109 where @xmath110 , was included to dress the corresponding vertices .
the amplitudes were also modified to preserve gauge invariance .
the details concerning these modifications , along with the values of the parameters used in the model can be found in @xcite . in the near - threshold region ,
the high precision spin density matrix results published by clas confirm the dominance of @xmath2-channel @xmath3 exchange in the forward direction ; however , at higher energies the existing theoretical models fail to reproduce the clas data @xcite .
the pomeron amplitudes are able to describe the energy dependence of the forward cross section , but fail to adequately describe the spin density matrix elements .
the unknown parameters present in the nucleon exchange amplitudes can be modified to describe the backward - angle data at higher energies if some assumptions , the reliability of which are unknown , are made @xcite .
our analysis is restricted to the energy range from threshold up to 2.4 gev . for c.m .
energies below 2 gev , the @xmath3 exchange amplitude dominates the @xmath2-channel contributions . in the higher energy range used in our analysis ,
the @xmath3 contribution is still substantially larger than that of the pomeron .
thus , the deficiencies in the pomeron amplitudes ( discussed above ) should not greatly affect our pwa results .
for this reason , we have chosen to use the otl @xmath2-channel terms with the parameter values obtained in that analysis @xcite . due to the unreliability of the assumptions under which
the nucleon exchange parameters were determined , we have decided not to include these amplitudes in our analysis , _
i.e. _ we do not include any @xmath89-channel terms in our fits .
the effect of this choice on our conclusions was found to be negligible ( see section [ section : sys ] ) .
the formalism used to construct our resonant amplitudes is described fully in @xcite .
it involves the use of relativistic tensor operators and is similar to the framework employed by anisovich _
et al . _
@xcite .
as discussed above , we do not impose resonant - like shapes on our @xmath90-channel waves .
instead , we divide our data into narrow c.m . energy bins . in each of these bins ,
the mass - dependence of an @xmath90-channel wave with spin @xmath111 and parity @xmath112 , which we will denote @xmath113 , is approximated by a constant complex number : @xmath114 where the sum is over the 10-mev wide c.m .
energy bins , @xmath115 are the strength and phase of the mass dependence in each bin , respectively , and @xmath116 mev is the maximum distance in any bin from the centroid @xmath117 .
the resonant waves then enter into our fits according to @xmath118 where @xmath119 are the angular momentum quantum numbers of the initial ( final ) state , @xmath120 are the unknown coupling constants to these states and @xmath121 is the covariant amplitude obtained using the formalism described in @xcite .
the values extracted for each @xmath90-channel wave s @xmath122 and @xmath123 parameters can be used to search for evidence of nucleon resonance contributions in that wave . for
every fit iteration run in each @xmath10 bin ( multiple iterations are run to alleviate problems caused by local minima ) , these parameters are started at random values that include the entire physically allowed range of the parameter .
for example , @xmath123 is started randomly in the range @xmath124 .
estimators for the parameters are then found by maximizing the likelihoods independently in each @xmath10 bin . in this way
, the mass dependence of the waves is extracted unbiasedly .
if resonant - like features are found in the cross sections and phase motion of the @xmath90-channel waves , then this is very strong evidence that resonances do contribute to the scattering amplitude .
if the strength observed in an @xmath90-channel wave is due to a single resonant state , then @xmath113 should be ( at least qualitatively ) described by a constant - width breit - wigner line shape of the form : @xmath125 where @xmath126 and @xmath127 denote the mass and width of the state , respectively .
if , however , the strength in the wave is caused by multiple resonant states or from a non - resonant process , then the use of a breit - wigner line shape is not valid .
the line shape given in eq .
( [ eq : bw ] ) neglects the kinematics and dynamics of the mass dependence of the resonance .
this can be an issue , _
e.g. _ , near threshold .
this will be addressed below .
before examining our results , it is important to reiterate the goals of our analysis .
we are attempting to extract strong resonance contributions to @xmath9 photoproduction in a least model - dependent way .
we do not enforce resonance shapes on the mass dependence of our partial waves .
instead , we first extract the strengths and phase motion of our partial waves independently in each @xmath10 bin . this stage will be referred to as a partial wave extraction ( pwe ) .
the second stage involves comparing the results of the pwe to what is expected from resonances in a mass - dependent fit ( mdf ) .
the pwe s are performed using unbinned extended maximum likelihood fits to the data in each @xmath10 bin .
the mdf s are simply @xmath128 fits to the phase differences obtained between partial waves in a given @xmath10 region .
we are not looking to build a complete model of @xmath9 photoproduction , _
i.e. _ we do not claim that the fits discussed below contain all of the amplitudes that contribute to this reaction .
because of this , we do not expect the physical observables extracted from our fits to provide perfect descriptions of our measurements ; however , the descriptions in many kinematic regions are very good .
finally , we are not attempting to extract all resonance contributions to @xmath9 photoproduction , only the most significant .
it is important to have a systematic method for selecting wave sets . as discussed in section [ section : amps : non - resonant ] , all of the wave sets used in our analysis contain the otl @xmath2-channel terms ( which contain no free parameters ) and no @xmath89-channel terms .
systematic studies show that the effects on extracted resonance parameters due to the choice of the non - resonant model are small ( see section [ section : sys ] ) . the mass - independent nature of our procedure , _
i.e. _ the bin - to - bin freedom of the resonance parameters , makes the use of smaller wave sets advantageous . for this reason
, we began our wave selection process by scanning the entire energy range of interest , @xmath129 gev@xmath130 gev , using the otl @xmath2-channel terms along with waves from a single spin - parity , @xmath131 .
the goal of this scan was to identify ( possible ) energy ranges where waves of a given @xmath131 perform significantly better than waves of any other spin - parity .
this information alone does not constitute evidence of resonance production ; however , it can serve as a guide as to which waves are more likely to contribute strongly to @xmath9 photoproduction .
given a pair of fits run with different wave sets , the difference in the log likelihoods obtained from the fits , @xmath132 , can be used to quantitatively determine which fit best describes the data . if @xmath133 , then wave set @xmath134 provides a better description of the data than wave set @xmath70 , while @xmath135 implies the converse is true .
figure [ fig : dlogl - example ] shows two examples comparing the likelihood differences between fits with different @xmath90-channel waves .
the @xmath136 quantities are shown for two separate fits , one with @xmath137 and one with @xmath138 , each of which is compared to a fit with @xmath139 .
> from threshold up to @xmath140 gev , the fit with @xmath138 is clearly the best , while in the energy range 1.85 gev@xmath1412 gev the preferred wave is @xmath137 .
it is also clear in fig .
[ fig : dlogl - example ] that both the @xmath138 and @xmath137 waves provide better descriptions of our data than the @xmath139 wave in this energy range .
similar fits were run using any single @xmath90-channel wave with @xmath142 of both parities . in the region from threshold up to @xmath140 gev , the @xmath138 wave was found to provide a better description of our data than any other wave .
similarly , in the energy range @xmath143 gev@xmath1412 gev , the @xmath137 wave was found to provide the best description .
scans were also performed using two @xmath90-channel waves and the otl @xmath2-channel terms .
the quantity @xmath136 can also be used in these fits to determine which wave sets best describe our data . in the @xmath144 gev region ,
the best fit was obtained using the @xmath90-channel waves with @xmath145 .
this wave set had the best likelihood in every bin in this energy range ( typically by a large amount ) .
given the results of the single wave @xmath90-channel scans discussed above , this is not a surprising result . above 2 gev , the preferred wave set consisted of the @xmath146 waves , along with the @xmath2-channel terms .
as in the lower energy range , this wave set had the best likelihood in every energy bin for @xmath147 gev .
the results presented for the waves below were not affected by our choice of wave set ; however , as the number of waves was increased so did the noise .
for this reason , we have chosen to present the results from fits with at most three @xmath90-channel waves .
see section [ section : sys ] for more discussion on fits with a larger number of waves .
vs @xmath10(mev ) : example likelihood differences from @xmath90-channel scans .
each fit contained the locked otl @xmath2-channel terms , along with a single @xmath90-channel wave . shown are @xmath148 ( open squares ) and @xmath149 ( closed triangles ) .
see text for details and discussion .
, scaledwidth=49.0% ] our preliminary @xmath90-channel scans showed that the best fit using two @xmath90-channel waves , along with the otl @xmath2-channel terms , in the energy range @xmath129 gev @xmath141 @xmath8 gev is obtained using @xmath150 . to extract any possible resonance contributions in these waves ,
event - based pwe fits were run using the locked otl @xmath2-channel terms along with @xmath145 @xmath90-channel waves parametrized according to eq .
( [ eq : resonance ] ) . in each energy bin
, multiple fit iterations were run using random starting values for the parameters ; the results presented below are always from the fit with the best likelihood .
the strength and phase of each @xmath90-channel wave were completely free to vary in each energy bin , _
i.e. _ they were fit independently .
the cross sections extracted for the @xmath90-channel waves are consistent with either near or sub - threshold resonance states ( see fig . [
fig : thresh : sigma ] ) .
the pdg lists two states in these waves consistent with this hypothesis : ( 1 ) the 4-star @xmath4 , which has a well known very large coupling to @xmath151 ; ( 2 ) the 3-star @xmath5 , which is currently rated as having only a 2-star coupling to @xmath151 .
we note here that the masses of the states can not be ( precisely ) estimated by simply examining the cross sections due to threshold suppression effects .
figure [ fig : thresh : dphase ] shows the phase motion between the two @xmath90-channel waves obtained from the pwe fits .
the phase differences were then fit in a mdf using the constant width breit - wigner line shapes of the form given in eq .
( [ eq : bw ] ) .
we chose not to use mass - dependent widths in the breit - wigner line shapes ( despite the proximity to @xmath79 threshold ) since the @xmath4 ( and perhaps the @xmath5 as well ) is below threshold , which introduces model dependence in a single - channel analysis .
for this reason , extracting precise values for the resonance parameters may not be possible ; however , the use of eq .
( [ eq : bw ] ) is sufficient to provide evidence for the presence of known pdg states in our data . in principle
, our mdf fits could have been made to the partial - wave intensities as well as the phase difference . in order to do this , form factors would need to be introduced at both the production and decay vertices . while these form factors do not strongly influence the phase difference , they are very important for describing the shapes of the cross sections .
the intensities of the partial waves ( in this fit , and in the following sections ) are qualitatively consistent with the expected resonance shapes .
obtaining good quantitative agreement requires the extra degrees of freedom introduced by the form factors ; however , including these factors also introduces additional model dependence .
thus , we have decided to only fit the phase difference which , as noted , is nearly independent of the form factors .
our results are , qualitatively , in good agreement with those expected from the pdg states mentioned above .
the dashed - line in fig . [ fig : thresh : dphase ] was fit requiring all parameters to be within the limits quoted by the pdg for the @xmath4 and @xmath5 .
there is a minor discrepancy in the near - threshold bins .
the parameters of the @xmath5 are not as well known as those of the @xmath4 ; thus , we also performed a mdf allowing the @xmath152 parameters to vary freely .
this fit resulted in a mass of @xmath153 mev and a width of @xmath154 mev for the @xmath5 , which are very close to the pdg limits .
uncertainties of @xmath155 mev for the mass and @xmath156 mev for the width were estimated by examining the variation in the @xmath152 parameters while using various parameter values ( all within the pdg limits ) for the @xmath157 , along with fitting different sub - ranges in @xmath10 of the phase motion .
vs @xmath10(mev ) : total cross sections from all of the waves included in the fit ( filled squares ) , only @xmath2-channel waves ( open squares ) , only @xmath137 waves ( circles ) and only @xmath138 waves ( triangles ) .
the cross sections extracted for both @xmath90-channel waves are consistent with near / sub - threshold resonances .
the error bars are purely statistical . , scaledwidth=50.0% ] ( radians ) vs @xmath10(mev ) : the dashed line was fit using constant width breit - wigner distributions requiring the parameters to be within the limits quoted by the pdg for the 4-star @xmath4 and 3-star @xmath5 .
the solid line was fit allowing the @xmath152 parameters to vary freely , the results are listed in the text .
the error bars are purely statistical .
, scaledwidth=50.0% ] a single channel analysis is not the best environment for extracting precise resonance parameters ; however , the qualitative agreement of the phase motion obtained from the pwe s to that of the two pdg states is very suggestive of their presence in our data .
recall that the phase parameters were each started pseudo - randomly in the range @xmath124 in each @xmath10 bin .
yet the results are in good qualitative agreement with the phase motion expected using simple constant width breit - wigner distributions for the pdg @xmath4 and @xmath5 states .
we can also compare the production helicity couplings extracted from our fits to those quoted by the pdg .
the ratio of the helicity amplitudes is obtained using the production couplings , @xmath158 in eq .
( [ eq : resonance ] ) , extracted by the fits and the @xmath90-channel production amplitudes . due to the nature of the covariant formalism
, these amplitudes are energy - dependent ; _ i.e. _ the ratio of the helicity amplitudes is a function of @xmath10 .
the pdg quotes these values at the resonance masses . for the @xmath5 , the pdg reports the ratio of the helicity amplitudes as @xcite @xmath159 our fits extract this value to be in the range @xmath160
$ ] , depending on the mass of the @xmath138 state .
this is in a very good agreement with the pdg value .
the ratio of the helicity amplitudes for the @xmath4 extracted by our fits is consistent with the pdg @xcite value .
however , projecting this ratio from the @xmath79 threshold down to the required mass makes obtaining a precise quantitative value difficult .
fit i consists of three production mechanisms : ( 1 ) otl @xmath2-channel terms ( with no free parameters ) , which are dominated by @xmath3 exchange in this energy range ; ( 2 ) @xmath138 @xmath90-channel waves , whose extracted parameters are consistent with the pdg @xmath5 state ; ( 3 ) @xmath161 @xmath90-channel waves , whose extracted parameters are consistent with the pdg @xmath4 state ( at least , near threshold ) .
this is almost certainly not all of the physics contributing to @xmath9 photoproduction in this energy range .
thus , we would not expect fit i to provide a perfect description of our data . before we examine the quality of fit i
, we note that the otl @xmath2-channel terms do provide a good description of our forward data in the energy regime of fit i. figures [ fig : fit1:compare - dsigma ] and [ fig : fit1:compare - rho ] show comparisons of the differential cross sections and spin density matrix elements extracted from the pwa fits compared to our measurements @xcite .
recall that we do not fit to the experimental observables directly .
we perform event - based fits to the data used to obtain the measured results .
the forward cross section and polarization observables are very well described in this energy range , confirming that @xmath3 exchange in the @xmath2-channel does dominate the amplitude in this region .
there is a discrepancy in the description of the cross section at backwards angles that increases with energy .
this could be due to the lack of @xmath89-channel terms .
it could also be a signature of unaccounted - for @xmath90-channel amplitudes .
though we did not start off by including known pdg resonance states , the fit has extracted evidence for them from our data .
we also note here that the large - angle cross section at @xmath162 gev is virtually flat . without polarization information ,
the production mechanism could have easily been mistaken for a @xmath163 wave .
this demonstrates the importance of the spin density matrix elements .
the quality of the description of the observables decreases slightly with increasing energy .
this signifies that there is another production mechanism that is not accounted for in the fit .
this is expected due to the limited number of waves included in the pwa .
adding additional waves improves the description of the data but has virtually no effect on the strengths and phase motion of the two @xmath90-channel waves presented in this section .
thus , the conclusions drawn about resonance contributions are robust , and do not change when additions are made to the wave set . see section [ section : results : additional - states : fitiii ] for results obtained by adding an additional @xmath90-channel wave to this fit .
our preliminary @xmath90-channel scans showed that the best fit using two @xmath90-channel waves and the otl @xmath2-channel terms in the energy range @xmath8 gev@xmath164 gev is obtained using waves with @xmath146 . to extract any possible resonance contributions in these waves ,
the same procedure used for fit i was employed . as in fit
i , the strength and phase of each @xmath90-channel wave were completely free to vary in each energy bin in the pwe s .
the cross sections extracted for the @xmath90-channel waves , shown in fig . [ fig : fit2:sigma ] , are consistent with the tail of a @xmath161 state with a mass below 2 gev ( as seen in fit i ) and a @xmath165 state with a mass near 2.2 gev .
the pdg lists a state consistent with this hypothesis : the 4-star @xmath6 , which currently is only listed as having 1-star coupling to @xmath151 .
figure [ fig : fit2:dphase ] shows the phase motion between the two @xmath90-channel waves extracted from the pwe s . our results do not agree with those expected from the pdg @xmath4 and @xmath6 states , assuming their mass dependencies are well described by the constant width breit - wigner line shape described in eq .
( [ eq : bw ] ) . allowing the @xmath137 parameters to vary freely
gives us better agreement and yields a mass around 1.95 gev .
the presence of a second state in the @xmath137 wave near this mass would have virtually no effect on the results obtained in fit i ; however , it would mean that the use of a breit - wigner distribution in the energy range where both @xmath157 states are contributing significantly , _
e.g. _ the energy region examined in fit ii , is invalid . instead
, we can employ a two - pole single channel k - matrix for the @xmath157 states of the form @xcite @xmath166 where @xmath167 and @xmath168 are the k - matrix resonance masses and coupling constants to the @xmath79 final state and @xmath169 are the centrifugal barrier factors ( see , _ e.g. _ , @xcite ) .
the mass dependence of the amplitude is then written in terms of the production vector , @xmath112 , and 2-body phase - space factor , @xmath170 , as @xmath171 where @xmath172 with production coupling constants
@xmath173 and @xmath174 for this mdf , we required the @xmath175 parameters to be within the limits quoted by the pdg for the @xmath6 .
one of the @xmath137 k - matrix poles was required to be consistent with the @xmath4 .
the exact location of this pole depends on how one treats the opening of the @xmath79 threshold ( this is a single channel analysis ) .
the parameters of the second @xmath137 k - matrix pole were obtained from the mdf yielding @xmath176 mev for the mass and @xmath177 mev for the width . even with these constraints ,
the results provide a very good description of the phase motion obtained from the pwe . in principle
, poles in the t - matrix and poles in the k - matrix can be quite different .
the relationship between the two can also depend on the specific k - matrix model employed .
for these reasons , care must be taken when interpreting the parameters obtained for the `` second '' @xmath157 resonance . while the k - matrix parameters may not coincide exactly with the physical t - matrix values , the observed strength in this wave and its phase motion relative to the @xmath6 support it having a mass around @xmath178@xmath8 gev and a width of approximately @xmath179@xmath180 mev .
these values are in good agreement with the missing @xmath181 state predicted by @xcite .
a check using the ( unitarity violating ) two breit - wigner prescription also resulted in @xmath157 resonance parameters in this range . to extract precise resonance parameters for this state ( and to confirm its existence ) , a coupled - channel analysis should be employed .
vs @xmath10(mev ) : total cross sections from all of the waves in the fit ( filled squares ) , only @xmath2-channel waves ( open squares ) , only @xmath137 waves ( circles ) and only @xmath175 waves ( triangles ) .
the cross section extracted for @xmath137 is consistent with the tail of a lower mass state ( as seen in fit i ) .
the @xmath175 cross section is indicative of a state near @xmath182 gev .
the errors are purely statistical . ,
scaledwidth=50.0% ] ( radians ) vs @xmath10(mev ) : the dot - dashed line is the phase motion expected using constant width breit - wigner distributions and the parameters quoted by the pdg for the @xmath4 and @xmath6 .
the dashed line required the @xmath183 parameters to be within the pdg limits for the @xmath6 , while allowing the @xmath7 parameters to vary freely .
the solid line used a constant width breit - wigner distribution for the @xmath6 , but a 2-pole single channel k - matrix for the @xmath7 wave .
the parameters obtained from these fits are listed in the text .
the error bars are purely statistical .
, scaledwidth=50.0% ] the ratio of the helicity amplitudes for the @xmath4 was discussed in section [ section : results : threshold : helicity ] . without employing a model we can not separate out the possible missing @xmath181 production amplitudes .
the ratio of the helicity amplitudes for the @xmath6 is extracted to be @xmath184 due to its 1-star coupling to @xmath151 , the pdg does not quote a value for this ratio . as in fit
i , we do not expect the limited number of waves used in fit ii to include all of the physics at these energies .
thus , we again do not expect to provide a perfect description of the observables in this energy regime . before we examine fit ii
, we note that the otl @xmath2-channel terms provide a good description of our forward cross sections ; however , there are some noticeable discrepancies with the spin density matrix elements at these energies @xcite . figures [ fig : fit2:compare - dsigma ] and [ fig : fit2:compare - rho ] show the differential cross sections and spin density matrix elements extracted from the pwa fits compared to our measurements @xcite . as the energy increases , the quality of the descriptions of the spin density matrix elements decreases .
this is partly due to the issues with the non - resonant model employed ( as discussed in section [ section : amps : non - resonant ] ) . at these energies ,
the otl @xmath2-channel terms begin to fail to adequately describe the polarization observables
. the discrepancies in the backwards direction could be due to the lack of inclusion of any @xmath89-channel terms .
the effects on our results , _
i.e. _ the effects on conclusions drawn about resonance contributions , due to possible issues with the non - resonant terms are discussed in section [ section : sys ] .
the lack of perfect description of the data signifies that there are other production mechanisms that are not accounted for in the fit .
this is , again , expected due to the limited number of waves included in the pwa .
adding additional waves improves the description of the data .
the strengths and phase motion of the two @xmath90-channel waves presented in this section become noisier in the presence of these additional waves ; however , the conclusions drawn about resonance contributions are unaffected by additions to the wave set .
we are unable to determine from our pwe s which additional waves may coincide with unaccounted - for physical processes ; thus , we do not present them here .
perhaps future measurement of additional polarization observables in this energy range might help determine the nature of these waves .
one of the prime motivating factors in undertaking this study was to search for missing resonances . the strongest evidence for resonance contributions to @xmath9 photoproduction found in fits i and ii was for well - known pdg states .
suggestive evidence was also found for a missing @xmath181 state .
below we examine possible additional resonance contributions .
quark model calculations predict three missing resonances with @xmath186 in the energy range of fit i which couple to @xmath79 @xcite .
figure [ fig : fit3:results ] shows the cross sections and phase motion obtained if we add a @xmath186 wave to the pwe in fit i. below @xmath187 mev , the range of production angles over which the clas has acceptance is limited .
this makes it difficult to cleanly separate contributions from three waves ; thus , this energy range has been excluded from these fits .
the strengths and phases of the @xmath145 waves are virtually unaltered by the addition of the extra @xmath90-channel wave .
the cross section of the @xmath186 wave does show some fairly smooth structure ; however , its phase motion , relative to the other two resonant states , is not consistent with a single constant width breit - wigner hypothesis .
if we instead perform a mdf using the k - matrix formalism described in section [ section : results : higher : cs_and_dphase ] for the @xmath186 wave , the phase motion between the @xmath188 waves is well described .
the poles in the k - matrix are at 1850 mev and 1950 mev .
we could apply the same procedure for the @xmath189 phase motion ; however , both waves would have k - matrices and the number of free parameters would leave the fit under - constrained .
figures [ fig : fit3:compare - dsigma ] and [ fig : fit3:compare - rho ] show the comparisons of the pwa results with our measurements @xcite .
the additional @xmath185 wave improves the description of the spin density matrix elements obtained in fit i ( see fig . [
fig : fit1:compare - rho ] ) .
it is difficult to make firm conclusions about the @xmath190 wave .
the observed strength suggests there is significant overlap of the scattering amplitude with this partial wave ; however , this is not sufficient evidence to claim resonance contributions .
the phase motion of the @xmath185 wave relative to the @xmath152 and @xmath191 waves is not consistent with a single resonant state . this does not , however , rule out the existence of the multiple states predicted by the quark model .
numerous other fits that we have performed yielded inconclusive evidence for states of various spin - parities .
these fits are very similar to fit iii .
generally , smooth structures are found in the extracted cross sections ; however , the phase motion is inconsistent with a single resonant state .
it is possible that a number of resonant states exist which couple relatively strongly to @xmath79 .
it is also possible that the smooth cross sections are simply the result of overlap of various partial waves with unaccounted - for non - resonant terms .
it would appear that we have reached the limits of what our technique can extract from our data .
more polarization information may be required to cleanly extract additional resonances . improved theoretical input for the non - resonant ( non @xmath90-channel ) terms may also be necessary . while our studies have shown that the strong @xmath90-channel signals extracted by this analysis are not affected by the way that the non - resonant terms are modeled ( see sec .
[ section : sys ] ) , this is almost certainly not the case for weaker signals .
this is particularly true at higher energies , where the current theoretical models do a poor job of describing the new clas data .
the amplitudes that are currently being generated by several groups from a coupled - channel approach ( see , _ e.g. _ , @xcite ) may well allow for the extraction of much weaker resonance signals from these data .
in section [ section : results : scans ] , we found that the single @xmath90-channel waves with the best likelihoods were @xmath192 for @xmath193 gev and @xmath137 for 1.85 gev@xmath194 gev .
we proceeded to add single @xmath90-channel waves to these fits to determine which had the best likelihoods ; these were the basis for the wave sets used in fits i and ii . we can also examine the `` discarded '' wave sets and examine the @xmath90-channel contributions as a systematic check on our results . in the @xmath195 gev energy range , the @xmath90-channel waves with @xmath145 , used in fit i , had the best likelihood of all two @xmath90-channel wave combinations ( when combined with the otl @xmath2-channel waves ) .
in this fit , the contributions extracted for the two @xmath90-channel waves were approximately equal in size for @xmath196 gev . for all other @xmath197 $ ] combinations ,
the extracted contribution for the @xmath138 wave was the bigger of the two @xmath90-channel terms in this energy range .
in the 1.85 gev@xmath1982 gev energy range , the @xmath157 wave had the larger of the two @xmath90-channel contributions for every @xmath131 used for the other @xmath90-channel wave .
fits were also run using all two @xmath90-channel wave combinations ( with the otl @xmath2-channel terms ) over the entire energy range .
the contributions of the waves presented in this paper were consistent , regardless of which other waves they were fit with .
the robustness of the results presented in this paper was also tested by performing the pwe s with larger wave sets .
the presence of any additional @xmath199 wave does not effect the conclusions drawn about resonance contributions to fit i or fit ii .
fits run with very large wave sets that included all @xmath90-channel waves with @xmath142 also confirm the large contributions from the @xmath138 and @xmath157 below 2 gev and from @xmath200 waves around 2.2 gev ; however , with this many waves it was not possible to unambiguously determine the spin - parity of the large @xmath111 contribution .
another possible cause of systematic effects is our lack of inclusion of any @xmath89-channel terms . for @xmath201
gev , any @xmath89-channel contribution must be small due to the lack of any visible peak in the backwards cross section . thus
, the conclusions drawn from fit i are independent of whether or not @xmath89-channel terms are included .
the same can not be said for the energy regime of fit ii . in @xcite
, we were able to modify the @xmath89-channel parameters of the oh , titov and lee model to better describe our highest energy data ; however , these amplitudes were not included in our pwa fits due to a lack of confidence in the assumptions used to obtain the parameters .
we can examine what effect adding these terms would have on the resonance parameters extracted in fit ii .
figure [ fig : effects - of - u ] shows the phase motion obtained from fit ii with and without @xmath89-channel amplitudes .
the agreement is very good in the region where both of the @xmath90-channel waves have strong contributions to the cross section .
it is only in the regions where the cross section of one of the @xmath90-channel terms is very small that including @xmath89-channel terms leads to a discrepancy in the extracted phase motion .
therefore , the conclusions drawn from fit ii regarding the resonance states are unaffected by how the @xmath89-channel terms are modeled .
we also note here that the likelihoods of the fits containing the @xmath89-channel terms were worse in all bins .
perhaps this is not surprising since the modified otl terms were obtained assuming the entire backward production amplitude is due to @xmath89-channel mechanisms at our highest energies . to obtain a better @xmath89-channel model
, the otl parameters should be fit including @xmath90-channel waves ; however , to simply estimate the effects of neglecting @xmath89-channel terms in fit ii , these parameters are sufficient .
-channel : @xmath202(radians ) vs @xmath10(mev ) : phase motion obtained with and without @xmath89-channel terms in fit ii .
including @xmath89-channel terms only creates discrepancies in the phase motion where one of the @xmath90-channel waves has a small contribution to the cross section .
the error bars on both phase motion plots are purely statistical .
, scaledwidth=45.0% ]
an event - based mass - independent partial wave analysis has been performed on data obtained using the clas at jefferson lab .
evidence has been found for contributions from the @xmath4 and @xmath5 nucleon resonance states .
these states are found to be dominant near threshold .
the data also strongly support the presence of the @xmath6 state .
suggestive evidence for an additional @xmath203 state with a mass around @xmath178@xmath8 gev has also been found .
the data shows definite strength in this partial wave over a very large energy range .
the phase motion between this wave and the @xmath6 supports the presence of a second @xmath157 state near @xmath204 gev . some evidence for other states exists , although the interpretations are more difficult .
the strength seen in the @xmath205 wave around @xmath206 gev , for example , is not consistent with a single resonant state ; however , we can not rule out the possibility that multiple @xmath207 resonances could be contributing to our data at these energies . to extract additional resonance signals from our data ,
improved theoretical input for the non - resonant terms may be required . in particular ,
including the amplitudes currently being generated by coupled - channel analyses would be highly desirable .
we thank the staff of the accelerator and the physics divisions at thomas jefferson national accelerator facility who made this experiment possible .
this work was supported in part by the u.s .
department of energy ( under grant no .
de - fg02 - 87er40315 ) , the national science foundation , the italian istituto nazionale di fisica nucleare , the french centre national de la recherche scientifique , the french commissariat lenergie atomique , the science and technology facilities council ( stfc ) , and the korean science and engineering foundation .
the southeastern universities research association ( sura ) operated jefferson lab under united states doe contract de - ac05 - 84er40150 during this work .
99 j. ballam _
et al . _ ,
d * 7 * , 3150 ( 1973 ) .
r. w. clift _
et al . _ ,
b * 72 * , 144 ( 1977 ) .
d. p. barber _ et al . _
( lamp2 group collaboration ) , z. phys .
c * 26 * , 343 ( 1984 ) .
m. battaglieri _
( clas collaboration ) , phys .
lett . * 90 * , 022002 ( 2003 ) .
j. barth _ et al .
_ ( saphir collaboration ) , eur .
j. a * 18 * , 117 ( 2003 ) .
y. oh , a. i. titov , and t. s. h. lee , phys .
c * 63 * , 025201 ( 2001 ) .
a. i. titov and t. s. h. lee , phys .
c * 66 * , 015204 ( 2002 ) .
q. zhao , phys .
c * 63 * , 025203 ( 2001 ) .
g. penner and u. mosel , phys .
c * 66 * , 055212 ( 2002 ) .
v. shklyar , h. lenske , u. mosel , and g. penner , phys .
c * 71 * , 055206 ( 2005 ) .
m. williams _ et al . _
( clas collaboration ) , submitted to phys .
c , ( 2009 ) , arxiv:0908:2910 [ nucl - ex ] .
k. schilling , p. seyboth and g. wolf , nucl . phys . * b15 * , 397 ( 1970 ) .
d. i. sober _ et al . _ ,
methods * a440 * , 263 ( 2000 )
. m. williams , ph.d .
thesis , carnegie mellon university , 2007,`www.jlab.org / hall - b / general / clas_thesis.html ` .
b. a. mecking _
et al . _ ,
methods * a503 * , 513 ( 2003 ) .
chung , bnl preprint bnl - qgs-93 - 05 ( 1999 ) .
m. williams , m. bellis and c. a. meyer , jinst * 4 * , p10003 , ( 2009 ) .
m. williams , comp .
. comm . * 180 * , 1847 , ( 2009 ) .
d. binosi and l. theussl , comp .
. comm . * 161 * , 76 ( 2004 ) . c. zemach , phys
. rev . * 140 * , b109 ( 1965 ) .
a. donnachie and p. v. landshoff , phys .
b * 296 * , 227 ( 1992 ) .
m. williams _ et al .
_ , _ the impact on the phenomenology of @xmath0 of the first high - precision measurements of @xmath9 polarization observables _ , to be submitted to phys .
a. v. anisovich _
et al . _ ,
j a * 24 * , 111 ( 2005 ) . w .-
_ , j. phys .
g * 33 * , 1 ( 2006 ) . c. amsler
_ et al . _ , ( crystal barrel collaboration ) phys
b355 * , 425 ( 1995 ) .
s. capstick and w. roberts , phys .
d * 49 * , 4570 ( 1993 ) .
mark paris , phys .
c**79 * * , 025208 ( 2009 ) . | an event - based partial wave analysis ( pwa ) of the reaction @xmath0 has been performed on a high - statistics dataset obtained using the clas at jefferson lab for center - of - mass energies from threshold up to 2.4 gev .
this analysis benefits from access to the world s first high - precision spin density matrix element measurements , available to the event - based pwa through the decay distribution of @xmath1 .
the data confirm the dominance of the @xmath2-channel @xmath3 exchange amplitude in the forward direction .
the dominant resonance contributions are consistent with the previously identified states @xmath4 and @xmath5 near threshold , as well as the @xmath6 at higher energies .
suggestive evidence for the presence of a @xmath7 state around @xmath8 gev , a `` missing '' state , has also been found .
evidence for other states is inconclusive . |
pulsars are highly magnetized , rotating neutron stars which emit beams of radio emission at a low height above the star s magnetic poles .
the emission is confined to the open field lines and the beam only sweeps out a small fraction of the celestial sphere .
for this reason we see emission from only one pole in the vast majority of pulsars . in fortunate geometrical circumstances ,
however , the second pole should also be observable as a radio emitting site , if the spin axis is almost orthogonally orientated to the line - of - sight to earth .
such pulsars can be identified by the detection of an `` interpulse '' separated by @xmath3 in pulse longitude from the main pulse .
less than 30 pulsars have been identified to show interpulses @xcite from the known population of @xmath21800 pulsars .
not all interpulses may arise from the second magnetic pole , but some of them can also be interpreted within a geometry of an aligned rotator where the observer s line - of - sight never , or hardly , leaves the emission cone during the pulse period ( e.g. manchester & lyne 1977 ) .
the geometry of a pulsar may be determined from observations of their polarization which has been a major industry since shortly after their discovery . in the more than 40 years since ,
a few key points are well established ( see e.g. lorimer & kramer 2005 ) .
pulsars are essentially rotating dipoles and the super - strong magnetic field naturally implies a high degree of linear polarization with the position angle ( pa ) of the radiation parallel ( or perpendicular ) to the magnetic field direction .
therefore , as the dipole field sweeps past the observer once per rotation of the pulsar , the observer sees a characteristic s shaped sweep of the position angle .
known as the rotating vector model ( rvm ) , it provides a simple mapping between the relevant geometrical angles and the position angle variations @xcite .
the pa as a function of pulse longitude , @xmath4 , can be expressed as @xmath5 here , @xmath6 is the angle between the rotation axis and the magnetic axis , and @xmath7 , with @xmath1 being the angle at closest approach of the line of sight to the magnetic axis .
@xmath8 is the corresponding pulse longitude at which the pa is then pa@xmath9 .
note that we follow the observer s sign convention for the pa , which differs from that introduced by damour & taylor ( 1992 , see also everett & weisberg 2001 ) .
the rvm has been used extensively to derive geometrical angles in a number of pulsars @xcite .
unfortunately there are many factors which mean that @xmath6 and @xmath1 are often subject to large uncertainties .
these include a covariance between @xmath6 and @xmath1 which can not be broken over the usually small longitude range over which pulsars emit ( e.g. von hoensbroech & xilouris 1997 , everett & weisberg 2001 ) , but also orthogonal modes and propagation effects in the pulsar magnetosphere @xcite . for these reasons
multi - frequency studies of a given pulsar rarely give similar values for @xmath6 and @xmath1 as the uncertainties are large ( e.g. mitra & li 2004 ) .
some information about the geometry can also be obtained from the evolution of pulse profiles with frequency .
observational data show that low frequency pulse profiles tend to be wider in extent than high frequency profiles @xcite and the explanation for this is that radio emission originates at different heights in diverging dipolar magnetic field lines . in this model ,
high frequency radio emission originates from closer to the stellar surface than low frequency emission with emission heights ranging from several tens to several hundreds of km @xcite .
recent evidence tends to suggest that , at a given frequency , the emission height varies across the polar cap with lower emission heights over the magnetic axis @xcite . also , observations indicate that low - frequency profiles are often dominated by strong central components whereas high - frequency profiles show strong outrider components in many cases .
this has been interpreted as indicating a different spectral index for core and cone emission @xcite , while geometrical factors may be at least partly responsible @xcite . when considering emission heights , it is intriguing to also consider the observational consequences of relativistic effects in the magnetosphere .
blaskiewicz , cordes & wasserman ( 1991 ) pointed out that , to first order , the main effect is that the pa swing is delayed with respect to the total intensity profile and the magnitude of this effect is directly propotional to the emission height .
this effect should therefore be frequency dependent if different frequencies originate at different heights .
the studies by blaskiewicz et al .
( 1991 ) , von hoensbroech & xilouris ( 1997 ) and mitra & li ( 2004 ) showed evidence for this effect in a sample of pulsars . in this paper , we present observations and their analysis of a 100-ms pulsar discovered by damico et al . ( 1988 ) which allows us , as it turns out , to study our described understanding of pulsar emission .
initial observations showed that the profile consisted of two widely separated components .
polarization observations at 0.66 and 1.6 ghz were carried out by qiao et al .
( 1995 ) and wu et al .
( 1993 ) respectively .
they showed that the profile is virtually 100 per cent linearly polarized with an indication of a small fraction of circular polarization .
the time resolution of the observations was low and each component was largely featureless .
( 1993 ) attempted an rvm fit to the pa swing and determined that @xmath10 and @xmath11 .
this fit indicated that the emission arises from a single pole and supported the idea that young pulsars had wide profiles @xcite . in a survey looking for pulsar wind nebulae , gaensler et al .
( 1998 ) discovered an unusual looking nebula around .
the head of the nebula is clearly resolved and a bow - shock indicates that the direction of motion of the pulsar is at an angle of @xmath2315(measured as north through east ) .
polarization properties of the pulse profile can give an indication of the position angle of the pulsar s rotation axis and we will investigate if the correlation between the velocity and rotation axis seen by johnston et al .
( 2005 ) is also evident in this pulsar , and whether it confirms the direction of motion suggested by gaensler et al .
( 1998 ) .
the organisation of the paper is as follows . in section 2
we briefly discuss multi - frequency observations of as part of a bigger campaign to measure polarization in southern pulsars generally . in section 3
we show how the profile of evolves with frequency and present our rvm fits , and re - examine in particular the question of whether this pulsar shows emission from a single , wide pole or from two poles .
we discuss the implications of the rvm fits in section 4 and derive a model for the beam of .
all observations were carried out using the parkes 64-m radio telescope located in central new south wales , australia . was observed on a number of occasions at a range of frequencies over the period 2004 december to 2007 july .
the observing bands were centred on frequencies of 1.37 , 3.10 and 8.6 ghz .
the bandwidths used were 256 , 1024 and 512 mhz respectively . in all cases , the backend used was the wideband correlator specifically designed for high - time - resolution polarimetric observations of pulsars . at all frequencies we sampled the data with 1024 phase bins across the pulse period .
the correlator performs on - line folding and de - dispersion at the topocentric period and dispersion measure . before each observation of the pulsar ,
a pulsed calibration signal was observed to allow correction of the gain and phase between the two feed probes . at each frequency observations
were made of the flux calibrator hydra a to provide accurate flux densities .
data reduction was carried out using the psrchive software package @xcite .
the final product was a fully polarimetric- and flux - calibrated pulse profile at each frequency .
we are able to obtain absolute position angles for the linear polarization using techniques described in johnston et al .
( 2005 ) . from multi - frequency observations made in 2005 july ,
the rotation measure of was determined to be [email protected] rad m@xmath13 ( cf [email protected] rad m@xmath13 in qiao et al .
figure [ fig : profiles ] shows the polarization profiles of obtained at 1.4 , 3.1 and 8.6 ghz .
zero longitude coincides with the fiducial point derived from beam considerations discussed in sections [ sec : abb ] and [ sec : beam ]
. values for @xmath8 quoted in table [ tab : paresults ] were measured relative to this longitude .
the higher time and frequency resolution of our data compared to that of wu et al .
( 1993 ) allows us to distinguish features in the profile not previously seen .
the main pulse ( mp ) consists of a small leading component blended with a stronger , narrower trailing component .
in contrast , the interpulse ( ip ) is a separated double profile with the trailing component somewhat stronger than the leading component .
johnston & weisberg ( 2006 ) have pointed out that a double - pulse structure with the trailing component dominating appears typical of young pulsars generally and both the mp and ip show this behaviour .
the mp - ip separation remains essentially constant at all frequencies and the intensity ratio of mp and ip remains unchanged , indicating similar radio spectra .
the ip shows hardly any frequency evolution , although the central part between the prominent components becomes noticeably weaker , leading to sharper inner edges of the components .
this behaviour is typical for many pulsars @xcite .
in contrast , the mp undergoes clear changes with frequencies .
the dominant component remains the strongest at all frequencies , while the outer components ( in particular the leading one ) become weaker at higher frequencies .
typical of young pulsars , the profile of ( both mp and ip ) is virtually 100 per cent linearly polarized . only at 8.4 ghz
does the polarization appear to decrease with both mp and ip having a fractional polarization near 70 per cent . at 1.4
ghz the mp shows positive circular polarization in the first two components with a hint of negative circular in the trailing component .
the value of @xmath14 is 4.1 per cent . in the ip ,
there is no circular polarization in the leading component but about 9 per cent of negative circular polarization in the trailing component . at 3.1
ghz the polarization has changed sign in the dominant component of the mp - there is little or no circular polarization at the leading component but about 9 per cent of negative circular polarization in the trailing component the peak of which is offset from the total intensity peak . in the ip
the circular polarization remains negative in the trailing component with a value of about 7 per cent . at 8.6 ghz the circular polarization has increased in the main component of the mp .
there is now 15 per cent of negative circular polarization and the peak of the circular is aligned with the total intensity peak . in the ip ,
the circular polarization is rather low with an upper limit of about 6 per cent but appears to be positive in sign . in other words
, it seems that the mp and ip show a mirror - symmetry in changing handedness of circular polarisation .
we note that in the low frequency profiles of qiao et al .
( 1995 ) , there appears to be little circular polarization in the interpulse but significant positive circular in the main pulse .
the picture where circular polarization is stronger in the trailing component in young pulsars was pointed out by johnston & weisberg ( 2006 ) but this is one of the few pulsars in which the sign of @xmath15 changes as a function of frequency . in the compilation of han et al .
( 1998 ) only psr b1240@xmath064 shows unambiguous evidence for a sign change from low to high frequencies . [ cols="<,^,^,^,^,^,^",options="header " , ] figure [ fig : pafit ] shows the measured values of the pas of the linear polarisation , @xmath16 , and fits of the rvm to those .
in contrast to the large majority of pulsars where the small duty cycle prevents a reliable determination of the rvm parameters @xcite , the highly polarized ip of allows a very precise measurement of the magnetic inclination angle @xmath6 , the impact parameter @xmath1 ( or alternatively , the line - of - sight angle @xmath7 ) , and the offset angles @xmath17 and @xmath8 , determining the symmetry positions of the fitted curve in pa and longitude , respectively .
table [ tab : paresults ] lists the results obtained for each frequency .
we obtain excellent fits with a reduced @xmath18 of close to unity at all three frequencies , clearly indicating an orthogonal geometry .
this is in contrast to the results of wu et al .
( 1993 ) , who obtained @xmath6=58 . inspecting their fig .
4 , we note that our pa values for the ip are offset from theirs by 180 ( an inherent ambiguity to measured pas ) .
performing a fit to our pa values in such an arrangement is possible but results in much worse fits ( @xmath19 ) , so that we believe that our solution is the correct one .
indeed , a free fit to all parameters results in consistent values at all frequencies , in particular , in @xmath6 close to 96 . given the problems often encountered in obtaining reliable rvm fits , this is an extraordinary result .
the results are so precise that they reveal a surprising , unexpected decrease in the magnitude of the impact angle @xmath1 with increasing frequency .
we confirm this observation by keeping @xmath6 fixed in subsequent fits ( see table [ tab : paresults ] ) and a visiual inspection of figure [ fig : pafit ] which indeed shows a steeping of the pa curve at the inflexion point @xmath8 .
we see the same trend , albeit with much larger uncertainties , by fitting the mp and ip separately .
such fits can also be used to determine the @xmath8 independently for mp and ip . doing this by keeping @xmath6 fixed to @xmath20 ( see below ) to increase the accuracy , we measure an mp - ip separation of @xmath8(mp)@xmath21(ip)@xmath22 . for reason discussed in detail in section [ scattering ]
, we consider the solution obtained at the highest frequency as the correct intrinsic geometry .
we therefore come to the conclusion that @xmath23=96 , @xmath24=@xmath06 , @xmath25=84 , @xmath26=6 .
this implies that we are seeing the mp and ip at the same impact angle on opposite sides of the pole .
this is unusual in itself , compare for example the situation in psr b1702@xmath019 , where the impact angles for the mp and ip are quite different @xcite .
finally , the inflexion point of the rvm occurs _
later _ than either the profile midpoint or the profile peaks . as outlined in the introduction , gaensler et al .
( 1998 ) detected a bow - shock nebula around , with the inference that the direction of motion was @xmath2315 .
the rvm fits described above show that the rotation axis points towards 48or 228 .
there is therefore an offset of 87 between the proper motion direction and the rotation axis .
we have shown elsewhere @xcite that the probable implication of this is that the proper motion vector and rotation axis are likely to be aligned , with the pulsar emitting linear polarisaion in a mode with the plane of polarisation orthogonal to the magnetic field direction .
this is consistent with our earlier results obtained from different samples of pulsars , and also confirms the interpretation of the bow - shock nebula by gaensler et al .
the geometrical interpretation of the rvm implies that the determined values for @xmath6 and @xmath1 should be independent of frequency .
therefore , our results of a changing @xmath1 and @xmath8 reported in figure [ fig : pafit ] and table [ tab : paresults ] are somewhat surprising .
usually , the geometry of a pulsar can not be determined with such high precision as is possible here , so that this effect may not have been noticed for other sources before .
however , rather than questioning the geometrical interpretation of the rvm , other explanations may apply . indeed , as we show in the following , the apparent change can be explained by an undetected scattering of the pulse profile at lower frequencies .
we investigated the impact of interstellar scattering on the observed pa swing in a similar fashion to that presented in the research note of li & han ( 2003 ) .
we confirm their results that large scattering leads to flat pa curves over the longitudes of the scattering tail .
such obvious effects of scattering are , however , not visible in the data for presented here .
hence , we studied the impact of much smaller scattering times . for this we used a noise - free template constructed from the determined gaussian decomposition of the observed profile .
we assigned a constant degree of linear polarisation and computed stokes @xmath27 and @xmath28 from the rvm fits to the 8.4 ghz data as shown in table [ tab : paresults ] .
we then convolved the stokes @xmath29 , @xmath27 and @xmath28 with a scattering tail as caused by a thin screen in the interstellar medium located half - way between the pulsar and the observer ( e.g. williamson 1972 ) .
the resulting pa swing was computed from the scattered stokes @xmath27 and @xmath28 intensities and subjected to the same rvm fitting process as the real data .
this process was repeated for scattering times in the range from zero to five phase bins , yielding the results shown in figure [ fig : sim ] .
the general agreement between the simulations and the observations is remarkable . while the predicted change in @xmath8 is somewhat too large , the behaviour of the measured @xmath6 and @xmath1 is well reproduced .
this confirms that , based on the polarization profile observed at 8.4 ghz , the pa swings ( and the derived values for @xmath6 and @xmath1 ) at 3.0 and 1.4 ghz can be well explained by a scattering of the profile with scattering times of only about 200@xmath30s and 500@xmath30s , respectively .
these scattering times are so small , that the corresponding total power profiles are hardly distinguishable from the unscattered profile in our simulations .
future high - resolution low - frequency studies should enable us to trace the frequency evolution of these scattering effects . while the geometry derived from the rvm
is therefore indeed independent of radio frequency , we note that the simulations also produce a shift in absolute pa that does depend on observing frequency . even though this shift is only 4 over a frequency range from 1.4 to 8.4 ghz , we point out that this effect might affect rotation measures determined from rvm fits over widely separated frequencies .
however , the precision to which the geometry can be determined for is unrivalled , so that in practice this effect may be small compared to other sources of uncertainties .
the precision obtained in fitting the rvm model to the observed data , promises to allow us to study aberration effects in pulsar radio emission . a first - order relativistic treatment of
the rvm leads to the expectation that the inflexion point of the pa swing , @xmath8 , should be delayed with respect to the profile midpoint by an amount @xmath31 where @xmath32 is the emission height @xcite .
the main difficulty in showing whether aberration is occuring or not is the determination of the location of the pulsar s fiducial point . for pulsars with symmetrical pulse profiles ,
one generally locates the fiducial point in the profile centre but for more complex pulse profiles the situation is much less clear although one can also use the profile s frequency evolution as a help . in the case of
, the mp is rather asymmetric ( and becomes more so at higher frequencies ) . for the ip ,
the profile is much more symmetrical ( and remains so at high frequencies ) , suggesting that the ip midpoint should be separated by 180 from the fiducial point of the mp .
this is indeed how we chose the zero - degree reference point for the pulse longitude in fig [ fig : profiles ] , even though it is clearly not centred on the mp profile .
we discuss this issue further in the next section .
we can draw all the strands of the above arguments together to come up with a plausible beam model for for the following reasons .
we believe the best location for the fiducial point of the mp and ip is at 0 and 180as shown in fig [ fig : profiles ] .
first , this is the midpoint of the symmetrical ip profile .
secondly , there is a hint of a central component at this location in the ip at 1.4 ghz and this has a steeper spectral index than the outer components in accordance with the general rule for central components .
this implies that the leading component in the mp is more centrally located ( and also has a steeper spectral index ) and we see emission at the trailing edge but not at the leading edge of the emission cone .
the offset between our purported profile fiducial point and the inflexion point as determined by the rvm is therefore some 11 .
this value can be used to determine an emission height of @xmath2230 km and the emission from both the mp and ip must originate from approximately the same height . in turn , the emission height can be used to determine the half - opening angle of the open fieldline region , @xmath33 , which must then be @xmath218 .
this is about a factor of two larger than the measured half - width of the profile which is about 9 for both the mp and ip ( although we do not see the leading emission in the mp ) .
interestingly , the derived value of @xmath33 for is very close to that given by the relationship @xmath34 with the pulse period @xmath35 measured in seconds ( e.g. kramer et al .
1994 ) , determined empirically from a large sample of pulsars . in figure
[ fig : geo ] we have drawn a sketch of a possible configuration of the beam for both the mp and ip .
the emission appears to arise from a cone of thickness @xmath23 which is located about half way out to the last open field lines .
the emission cone is not fully illuminated but rather consists of active patches .
both the mp and the ip sight lines appear to cross two such patches .
what is remarkable is that the active regions in the magnetosphere near both poles happen to lie on the same cone .
high time resolution observations over a range of frequencies for have enabled us to determine the geometry and beam shape of the pulsar .
uniquely amongst pulsars , even amongst those with interpulses , we have been able to determine highly constrained rvm fits which show conclusively that is an orthogonal rotator .
we have shown that the value of @xmath1 appears to be frequency dependent ; however careful modelling shows that this appears to be entirely due to the effect of interstellar scattering of the profiles .
while other factors usually dominate the uncertainties in the geometry derived from rvm fits , it should be considered in particular for low - frequency observations or observations of short - period pulsars .
it will also be relevant for future high - sensitivity detections and studies of interpulse sources with the square - kilometre - array @xcite .
interestingly , the relativlely high ratio of spin - down luminosity @xmath36 to distance @xmath37 as measured by @xmath38 erg s@xmath39 kpc@xmath13 makes this pulsar potentially detectable with the gamma - ray large area space telescope ( glast ) .
this would be an exciting opportunity to test competing @xmath40-ray emission models , given the very well determined geometry of this pulsar .
for instance , outer gap models ( e.g. chiang & romani 1994 ) predict that orthogonal rotators should be @xmath40-ray bright as we look directly into the null - charge surface , i.e. the bottom part of the outer gap ( yadigaroglu & romani 1995 ) . we have shown that the total intensity structure of the mp and the ip in have similarities . while the degree of linear polarisation is very large in both mp and ip , the circular polarisation changes handedness with increasing frequency with an intriguing mirror - symmetry between the mp and ip .
the emission comes from a height of @xmath2230 km in both and is consistent with an origin in a cone with a radius of only one half of the beam opening angle .
the cone is patchy as we do not detect the leading emission from the mp . in summary
, this pulsar represents a show - case for many features of pulsar radio emission , combined in a single , remarkable source .
but most of all , this source provides conclusive evidence for the geometrical interpretation of the pa swing within the rvm .
many pulsars , however , in particular recycled pulsars , show large deviations from a simple rvm swing in which cases the geometrical interpretation has to be questioned . in this case
, however , the evidence is incontrovertible .
the australia telescope is funded by the commonwealth of australia for operation as a national facility managed by the csiro .
we thank andrew lyne and the referee jarek dyks for useful and stimulating comments on the manuscript . | high time resolution observations of ( or psr j0908@xmath04913 ) over a wide range of frequencies have enabled us to determine the geometry and beam shape of the pulsar .
we have used the position angle traverse to determine highly - constrained solutions to the rotating vector model which show conclusively that is an orthogonal rotator . the accuracy obtained in measuring
the geometry is unprecedented .
this may allow tests of high - energy emission models , should the pulsar be detected with glast . although the impact parameter , @xmath1 , appears to be frequency dependent
, we have shown that this is due to the effect of interstellar scattering . as a result
, this pulsar provides some of the strongest evidence yet that the position angle swing is indeed related to a geometrical origin , at least for non - recycled pulsars .
we show that the beam structures of the main pulse and interpulse in are remarkably similar .
the emission comes from a height of @xmath2230 km and is consistent with originating in a patchy cone located about half way to the last open field lines .
the rotation axis and direction of motion of the pulsar appear to be aligned .
[ firstpage ] pulsars : general pulsars : individual : psr j0908@xmath04913 stars : neutron |
type ia supernovae ( sne ia ) have been successfully utilized over the past decade to measure the cosmic expansion history @xcite and explore the nature of dark energy ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the foundation for the utility of sne ia as a cosmological tool is that some distance - independent observables , such as the light - curve shape parameters @xcite , the color parameters @xcite , or both @xcite , have been found to correlate with their peak luminosity . these empirical correlations can be used to calibrate the luminosities of sne ia and measure their distances with a precision of @xmath19% .
a recent result suggests that the luminosity standardization of sne ia can be improved to a level of @xmath16% by separating the sne into two groups based on a spectroscopic criterion [ for details , see @xcite , hereafter w09 ] .
the expansion velocity ( @xmath0 ) of the sn ejecta is inferred from the blueshift of the absorption minimum of siii @xmath96355 , and the sne are divided into one group with normal @xmath0 ( hereafter normal " sne ia ) and the other group with high @xmath0 ( hereafter hv " sne ia ) .
w09 found that the two groups have either different extinction laws or color evolution .
the cause for such a dichotomy might be related to the properties of their progenitors .
the hv sne ia are characterized by stronger absorption features of intermediate - mass elements ( imes , such as si , s , and ca ) at higher velocities as well as a red @xmath10 color around maximum brightness .
sn 2002bo and sn 2006x are two of the best - studied examples of this class @xcite .
in particular , the interstellar na i d lines were found to show significant variations in the spectra of sn 2006x , likely pointing to the presence of csm produced by the progenitor system @xcite .
the possible detection of csm around sn 2006x is also supported by a flat evolution of the late - time light curve @xcite and a detection of light echoes @xcite .
similar variability of the na i d lines was also observed in sne 1999cl and 2007le @xcite , two other members of the hv sn ia class .
contrasting with this , multi - epoch , high - resolution spectral observations of sn 2007af , a normal sn ia , do not reveal any significant signature of csm absorption @xcite .
this raises the possibility that csm might be preferentially present for the sne ia in the hv class . in this paper
, we present optical observations of another member of the hv sn ia class , sn 2007gi .
our goal is to increase the sample and understand the properties of well - observed hv sne ia .
observation and data reductions are described in 2 , while 3 presents the @xmath11 light curves , color curves , reddening estimate , and an analysis of the rise time .
section 4 presents the spectral evolution .
our discussions and conclusions are given in 5 .
sn 2007gi was discovered by k. itagaki ( iauc 8864 ) on july 23.49 ( ut dates are used throughout this paper ) at @xmath12 = 12@xmath1301@[email protected] , @xmath16 = + 61@xmath1753@[email protected] ( j2000.0 ) in the lenticular galaxy ngc 4036 .
an optical spectrum taken on 2007 aug .
4.90 revealed that sn 2007gi was a very young sn ia at a phase @xmath110 days before maximum brightness @xcite , with an expansion velocity measured from the absorption minimum of si ii @xmath96355 as @xmath120,500 km s@xmath2 .
this velocity is much higher than what is observed for a normal sn ia at a similar phase , but is typical for members of the hv sn ia class , such as sne 1983 g , 1997bp , and 2006x @xcite .
photometry of sn 2007gi was obtained primarily with the 0.8-m tsinghua - naoc telescope ( tnt ) located at naoc xinglong observatory .
this telescope is equipped with a @xmath20 pixel back - illuminated ccd , with a field of view ( fov ) of @xmath21 ( scale @xmath22 pixel@xmath2 ) .
observations of sn 2007gi began on 2007 aug .
8 , @xmath16 days before @xmath3-band maximum , and continued for the next 6 months .
sn 2007gi was also observed during its early rising phase ( from @xmath23 to @xmath247 days ) with the ultraviolet / optical telescope [ uvot ; @xcite ] onboard the _ swift _ observatory . as shown in figure 1 ,
sn 2007gi is located just outside the central bulge of ngc 4036 . to remove the light contamination from the host galaxy , we applied an image - subtraction technique before performing the photometry .
template images of ngc 4036 were obtained on 2008 may 24 , roughly 300 days after maximum brightness . to perform the image subtraction ,
the image containing sn 2007gi is first geometrically registered to the corresponding template image ; the fluxes of the foreground stars in the images are then scaled to the same level .
the point - spread functions ( psfs ) of these two images are convolved to match , and the template
is then subtracted from the sn images .
we finally performed standard psf - fitting photometry to obtain the instrumental magnitudes for the sn and the local standard stars with the iraf daophot package @xcite .
transformations from the instrumental magnitudes to the standard johnson @xmath25 @xcite and kron - cousins @xmath26 @xcite systems were established by observing on photometric nights a series of @xcite standard stars covering a wide range of air masses and colors .
a total of 4 photometric nights were used to calibrate 8 local standard stars in the field of sn 2007gi .
table 1 lists their final calibrated @xmath11 magnitudes @xcite .
these local stars are then used to transform the instrumental magnitudes of sn 2007gi to the standard @xmath11 system , and the final results of the photometry are listed in table 2 . the error bars ( in parentheses ) include both the uncertainty in the calibration of the local standard stars and the uncertainty of the instrumental magnitudes due to photon noise and image subtraction . the @xmath3- and @xmath27-band magnitude obtained by uvot on three earlier epochs
are also listed in table 2 .
the swift / uvot photometry was reduced in a manner similar to that described in @xcite and using the photometric zeropoints of @xcite . in the absence of template images of the host galaxy
, regions around the galaxy at the same distance from the nucleus as the sn were used to estimate the underlying count rate and its uncertainty .
given the sparse spectroscopic observations of sn 2007gi ( 4 ) , no @xmath28- or @xmath29-corrections were applied to the photometry presented in this paper .
optical spectra of sn 2007gi were obtained primarily with the 3-m shane telescope at lick observatory using the kast double spectrograph @xcite .
a spectrum was also obtained on 2007 nov .
12 with the low resolution imaging spectrometer ( lris ; @xcite ] mounted on the 10 m keck i telescope on mauna kea , hawaii .
the journal of spectroscopic observations is given in table 3 .
all of the spectra were reduced by standard iraf routines . in order to avoid contamination from the host galaxy ,
each spectrum of the sn was extracted carefully , and the flux calibration was done with spectra of standard stars observed on the same night at similar air masses .
the spectra were corrected for continuum atmospheric extinction using mean extinction curves for lick observatory , and telluric lines were removed from the data . during the observations , the slit was always aligned near the parallactic angle to avoid chromatic losses due to atmospheric dispersion @xcite .
the @xmath11 light curves of sn 2007gi are presented in figure 2 . a polynomial fit to the @xmath3-band light curve around
maximum brightness yields @xmath30 = [email protected] mag on jd 2,454,[email protected] ( 2007 august 14.5 ) .
this indicates that our observations started from about @xmath2411 days and extended to + 188 days with respect to the @xmath3-band maximum .
the @xmath27-band light curve reached a peak magnitude of [email protected] mag on jd 2,454,[email protected] , about 1.5 days after @xmath31(b ) . using a template - fitting method [ similar to that used in hamuy et al .
( 1996 ) but expanded to include more templates ] to fit the @xmath3- and @xmath27-band light curves , we derived @xmath32 = [email protected] mag and @xmath33 = [email protected] mag .
the @xmath10 color around the maximum is found to be slightly redder than the unreddened loci @xcite , suggesting some reddening toward sn 2007gi .
the salt2 method @xcite yields a best - fit stretch factor of @xmath34 = [email protected] , which corresponds to @xmath35(b ) = [email protected] from the transforming relation between @xmath34 and @xmath35(b ) by @xcite .
the best - fit templates from mlcs2k2 @xcite are also overplotted in figure 2 .
one can see that the light curves of sn 2007gi can be well fit by both salt2 and mlcs2k2 . in figure 3 , we compare the light curves of sn 2007gi with those having similar @xmath32 values : sne 1994d @xcite , 1996x @xcite , 2002bo @xcite , 2002dj @xcite , 2002er @xcite , 2004eo @xcite , and 2006x @xcite . among these objects ,
sne 2002bo , 2002dj , 2006x , and 2007gi belong to the hv group , while sne 1994d , 1996x , and 2004eo belong to the normal group .
sn 2002er may be a transitional object linking the normal and hv groups ( w09 ) .
compared to the normal sne ia in the @xmath3 band , sn 2007gi appears relatively brighter and declines slowly when entering the nebular phase . for convenience of comparison
, we introduce a quantity @xmath36 , which is the flux ratio measured at @xmath37 and @xmath6 days with respect to the @xmath3 maximum .
the ratio @xmath36 is found to be [email protected] for sn 2007gi , which is similar to that of sn 2006x ( [email protected] ) but noticeably smaller than that of sn 1994d ( [email protected] ) , sn 1996x ( [email protected] ) , and sn 2004eo ( [email protected] ) .
thus , the hv objects seem to have systematically smaller peak - to - tail contrast than the normal objects . close inspection of figure 3a also reveals that there may be a break in the light - curve evolution for sn 2007gi at @xmath38 days . the decay rate @xmath39 is found to be [email protected] mag ( 100 days)@xmath2 during the period from @xmath40 days to @xmath41 days , similar to that measured for sn 1984a and sn 2002bo at the same phase . the decay becomes steeper after @xmath41 days , with @xmath39 = [email protected] mag ( 100 days)@xmath2 .
such a two - stage evolution also exists in sn 2006x , but it is not unambiguously present in the cases of sne 1984a and 2002bo due to poor photometric coverage in the nebular phase .
future observations will be needed to determine whether this trend is universal for all hv sne ia . in contrary to the @xmath3-band light curve
, the @xmath42-band light curves of sn 2007gi do not exhibit a two - stage evolution in the nebular phase ( see figure 3b d ) . neither the flux contrast between the tail and peak nor the decay rate of the tail show significant differences among the objects in the comparison .
a noticeable difference is that the secondary shoulder / maximum features at around + 25 days in the @xmath42 bands appear more pronounced in the hv objects than those in the normal ones ( see figures 3c and 3d ) .
part of this effect could be caused by significant extinction which shifts the effective passband to redder wavelengths [ e.g. , @xcite ] .
as the extinction suffered by sn 2007gi is relatively low ( 3.2 ) , its prominent secondary shoulder is likely to be mostly intrinsic .
according to @xcite , the variation of the secondary maximum of the near - infrared ( nir ) light curves is caused by the abundance stratification in sne ia and/or the progenitor metallicity .
one feature in figures 3(a)(d ) , though not easily discernible due to the crowdness of the data points , is that the hv sne 2007gi , 2006x , and 2002bo seem to rise to the maximum at a faster rate than the normal objects .
such a trend was also mentioned by @xcite in the qualitative comparison of sn 2002dj .
a quantitative analysis of the rise time is given in 3.3 .
figure 4 presents the optical color curves of sn 2007gi ( @xmath43 , @xmath44 , @xmath45 ) .
overplotted are the color curves of the other sne in the light - curve comparison in figure 3 .
all of the color curves are shifted to match the observed values of sn 2007gi at @xmath3 maximum . for the @xmath10 color
, sn 2007gi exhibited a similar color evolution as the other sne in the early phase ( @xmath46 days ) . at @xmath47 days , it has the reddest color of all the sne .
after the red peak at @xmath48 25 days , sn 2007gi seems to become progressively bluer than the other sne in the comparison , so by @xmath40 days it has the bluest color .
unfortunately , there is a gap in the range @xmath4960 days when sn 2007gi was behind the sun , so the detailed evolution during this phase is not clear .
the slope measured during @xmath690 days is @xmath50 mag day@xmath2 , steeper than the lira - phillips relation [ e.g. , @xmath51 mag day@xmath2 ; @xcite ] .
this was also noticed by @xcite in the comparison study of sn 2006x .
in fact , a steeper @xmath10 color - change slope than the lira - phillips relation in the nebular phase is found to be the case for most of the hv events ( wang , x. , et al .
2009c , in preparation ) , suggesting that the reddening derived for them from their tail colors might be potentially biased toward a lower value .
despite the relatively large differences seen in the @xmath10 color evolution , the two classes of sne ia ( hv and normal ) share similar @xmath52 and @xmath53 color evolution ( figures 4b c ) .
the galactic extinction toward ngc 4036/sn 2007gi is @xmath54 mag @xcite , corresponding to a color excess of @xmath55 mag [ adopting the standard reddening law of @xcite ] .
the peak @xmath10 color , corrected for the galactic reddening , is [email protected] mag .
this corresponds to a host - galaxy reddening of @xmath56 mag under the assumption that the hv sne ia have intrinsic @xmath10 colors similar to those of the normal objects near maximum brightness ( w09 ) .
assuming @xmath57 , as suggested by w09 for the hv sne ia , the extinction caused by the dust within the host galaxy of sn 2007gi is @xmath58 mag
. the late - time evolution of sne can provide valuable constraints on the sn environments as well as their explosion models [ e.g. , @xcite ] .
the field of sn 2007gi was imaged with the wide field planetary camera 2 ( wfpc2 ) onboard the _ hubble space telescope ( hst ) _ on 2009 march 26 , @xmath1600 days after discovery , as part of _ hst _ program go-10877 ( pi : weidong li ) .
a cosmic - ray - split pair of observations was observed with each of the f435w , f555w , f675w , and f814w filters . to locate sn 2007gi on these wfpc2 images
, we follow the technique detailed by @xcite .
we use an @xmath8-band image of sn 2007gi taken with the lick observatory 1-m nickel telescope on 2007 dec .
15 , and identify four stars that are present in the wfpc2 images .
an astrometric solution using these four stars yields a relatively poor precision of @xmath59 due to the faintness of the stars and the resolution of the nickel image .
nevertheless , within the 1@xmath60 error radius of the astrometric solution , we identify a single star which we presume to be sn 2007gi .
figure 5 demonstrates this process . to measure the magnitudes of sn 2007gi in the wfpc2 images
, we have used the
hstphot " package developed by @xcite . after performing the preprocessing suggested by the hstphot manual , including removal of cosmic rays , defects , bad pixels , and hot pixels
, we measure the magnitudes in the wfpc2 system .
these magnitudes are then converted to the ground - based broad - band @xmath11 magnitudes using the procedure described by @xcite .
we measure the following magnitudes for sn 2007gi : @xmath3 = [email protected] , @xmath27 = [email protected] , @xmath8 = [email protected] , and @xmath61 = [email protected] .
sn 2007gi was found to decline by @xmath111.1 mag in the @xmath27 band by 590 days after the @xmath3 maximum .
this flux drop is consistent with the corresponding values observed in sn 1996x and sn 2003hv at comparable phases @xcite .
sn 1991 t and 1998bu , which showed evidence for the presence of light echoes @xcite , declined by @xmath62 mag within 590 days since the maximum .
in addition , the @xmath10 color of sn 2007gi in the nebular phase is found to be inconsistent with that predicted by the light - echo scenario . thus .
we conclude that an interstellar echo may not be present in sn 2007gi . the time spent by a sn from explosion to maximum brightness
is defined as the rise time @xmath63 , an important parameter that links the explosion physics to the progenitors of sne ia @xcite .
the rise time of a sn ia is determined primarily by the rate at which the interior energy is released and subsequently diffuses to the surface of the supernova .
assuming that the luminosity of a sn ia evolves as an expanding fireball at very early phases @xcite , the rise time @xmath63 can be derived from the relation @xcite @xmath64 where @xmath12 is the `` speed '' of the rise of the luminosity and @xmath65 is the time elapsed since maximum brightness . in this subsection , we attempt to examine the difference in @xmath66 between hv and normal sne ia .
instead of adopting the @xmath3-band light curve @xcite , our analysis uses the @xmath8-band light curve , as our earliest photometric data points are usually obtained with a clear filter ( used in our sn search ) which is very close to broad - band @xmath8 @xcite . only those data obtained earlier than one week before the @xmath8-band maximum are used in the fitting .
the computed rise times for sn 2007gi and several other well - observed sne ia are plotted against the decline rate @xmath5 in figure 6 .
the sne are divided into the hv ( filled circles ) and normal ( open squares ) groups according to the classification scheme defined by w09 . in each group
, @xmath63 correlates with @xmath32 : slower decliners have longer rise times .
the hv sne ia appear to have a faster ( i.e. , shorter ) rise time than the normal ones at the same @xmath32 with the exception of sn 1994d , which lacks published photometry during the very early phases .
a similar trend was also noted by @xcite .
assuming comparable amounts of nickel are synthesized during the explosion , the rise time of a sn ia is affected by the efficiency of the photon diffusion , which depends on the opacity and the location of the photosphere .
a shorter @xmath63 could indicate that compared with a normal sn ia , the photosphere of a hv sn ia is less effectively heated by the @xmath67-ray photons inside the ejecta , or becomes optically thin at a faster pace due to the more rapid expansion ( see the discussion in 4.3 ) .
we have 5 optical spectra of sn 2007gi obtained with the 3.0-m shane telescope at lick observatory and the keck i 10-m telescope at the w. m. keck observatory , spanning from @xmath68 to @xmath69 days since @xmath3-band maximum brightness .
as shown in figure 7 , the spectral evolution around maximum brightness generally follows that of a normal sn ia but is characterized by very broad and highly blueshifted absorption features at 3600 (ca ii h&k ) , 6000 ( si ii @xmath96355 ) , and 8000 ( ca ii nir triplet ) . in figure 8 , we compare the spectra of sn 2007gi with those of sne ia having similar @xmath35 values at four selected epochs ( @xmath70 , @xmath71 , + 6 , and + 3 months since the @xmath3 maximum ) .
all spectra have been corrected for the reddening and the redshifts of the host galaxies .
the reddening of each comparison sn ia is taken from table 1 in w09 , and the @xmath72 value is assumed to be 1.57 for the hv sne ia and 2.36 for the normal objects according to the analysis w09 . at @xmath73 days ( figure 8a ) , the spectrum of sn 2007gi is characterized by lines of singly ionized imes ( si , s , mg , and ca ) . the absorption features of si ii @xmath96355 and ca ii h&k in sn 2007gi are deep and broad , comparable to those in the hv sn ia 2002bo but significantly stronger than those in the normal sne 1994d , 2002er , and 2004eo .
around maximum brightness ( figure 8b ) , we measured a line - strength ratio of si ii @xmath95972 to si ii @xmath96355 , known as @xmath74(si ii ) @xcite , to be @xmath75 for sn 2007gi .
this is a relatively small @xmath74(si ii ) value and usually indicates an overluminous event @xcite .
similar low values of @xmath74(si ii ) are observed for the other two hv objects , sne 2006x and 2002bo . by one week
after maximum ( figure 8c ) , the strong absorptions of si ii and ca ii lines are still present in the hv objects ( sne 2007gi , 2002bo , 2002dj , and 2006x ) , but the w - shape s ii lines have almost vanished . at @xmath76 months ( figure 8d ) ,
the spectra are dominated by iron lines whose strength is similar in all of the objects , and a strong ca ii nir absorption trough .
sn 2006x was found to show a prominent near - uv excess starting from @xmath77 days , perhaps due to a local light echo or an interaction of the sn ejecta with circumstellar material @xcite .
close inspection of the @xmath78 day spectrum of sn 2007gi reveals a flux excess at the blue end when compared to the normal sne ia , consistent with its bluer @xmath79 color at this phase ( 3.2 ) .
the blueshift of the absorption minima of some spectral features may approximately trace the location of the photosphere in the early phases of a sn . in this subsection
, we examine some strong lines such as si ii @xmath96355 and ca ii h & k , as well as a weak line s ii @xmath95640 .
the derived evolution of the expansion velocity ( @xmath0 ) of the sne is shown in figure 9 .
all of the velocities have been corrected for the redshifts of the host galaxies .
sn 2007gi is one of the objects with the highest expansion velocities in the comparison , with @xmath0 = 15,500 @xmath4 300 km s@xmath2 at @xmath80 day . by comparison ,
the typical value of @xmath0 for a normal sn ia is 10,600 @xmath4 400 km s@xmath2 at this epoch ( w09 ) .
the velocities inferred from the ca ii h @xmath81 k and s ii @xmath95640 lines are about 21,500 km s@xmath2 and 12,800 km s@xmath2 near maximum , respectively , which are also significantly higher than those of the normal sne ia .
these high @xmath0 values of the imes ( si , s , and ca ) can often be interpreted as being caused by a density / abundance enhancement in the outer ejecta of the sne @xcite .
several scenarios have been proposed to account for such an enhancement .
the metallicity effect was examined by @xcite , who found that increasing the metallicity in the c+o layers by a factor of 10 is still far from explaining the high @xmath0 seen in sn 2002bo .
the delayed detonation models were proposed to explain the hv features in sn 1984a and sn 2002bo @xcite , as intense nucleosynthesis in the outer layers results in an enhancement of the density of the imes at higher velocities , which subsequently increases the opacity in the outer regions and results in a photosphere that is moving at a high velocity .
an interaction between the sn ejecta and csm could produce the high - velocity ca ii lines @xcite and may also affect the si ii lines @xcite .
detailed analysis of the reasons for the hv features is beyond the scope of this paper .
as discussed by @xcite , @xcite , and @xcite , the flux ratio @xmath74(si ii ) between si ii @xmath95972 and si ii @xmath96355 is an indicator of the photospheric temperature and the luminosity of a sn ia : a value of @xmath74(si ii ) means a more luminous sn ia with a higher photospheric temperature .
the unusually low @xmath74(si ii ) value of @xmath10.07 for sn 2007gi at @xmath80 day thus suggests a rather high photospheric temperature . on the other hand , the strength of si iii @xmath94560 (
actually si iii @xmath824553 , 4568 ) has been proposed as another photospheric temperature tracer @xcite , and is stronger for a sn ia having a higher temperature .
the si iii @xmath94560 line is weak in sn 2007gi , suggesting a rather low photospheric temperature .
the two temperature indicators thus provide conflicting results for sn 2007gi . in an attempt to further constrain the photospheric temperature in sn 2007gi
, we used the parameterized resonance scattering synthetic - spectrum code synow @xcite to fit the near - maximum spectra of sn 2007gi .
the basic input parameters are taken from table 1 in @xcite , with the blackbody temperature @xmath83 and the velocity at the photosphere @xmath84 as free parameters .
figure 10 shows our best synow fit for the @xmath80 day spectrum , with @xmath84 = 15,500 km s@xmath2 and @xmath85 k. the high @xmath84 is consistent with the high @xmath0 measured directly from the spectrum ( 4.2 ) .
the blackbody temperature @xmath83 , on the other hand , is rather typical of other sne ia at this phase .
for example , @xcite estimated @xmath86 k for the normal sn ia 1994d at @xmath87 day , while our best synow fit to a spectrum of the normal sn ia 2004eo at @xmath88 day yields @xmath89 k. our synow analysis thus suggests that @xmath74(si ii ) is not a good photospheric temperature indicator for the hv sne ia .
the exact cause for this is not clear .
one possibility is contamination by fe ii and fe iii lines .
@xcite suggested that the fe lines might weaken the si ii @xmath95972 absorption line even at low @xmath83 .
in this paper , we present optical photometry and spectroscopy of sn 2007gi , a sn ia with an extremely high expansion velocity measured from imes ( si , ca , and s ) . we also conduct a comparison study for a sample of sne ia with high and normal expansion velocities ( hv and normal , respectively ) to investigate differences in their photometric and spectroscopic behaviors .
the @xmath3-band light curve of sn 2007gi shows a two - stage evolution : a decay rate of @xmath90 mag ( 100 days)@xmath2 during @xmath9190 days and @xmath92 ( 100 days)@xmath2 thereafter .
the latter decay rate is similar to those observed in normal sne ia .
this two - stage evolution is also present in the hv sn ia 2006x .
more late - time observations of the hv objects are necessary to establish whether this is a universal property .
the @xmath93 color of sn 2007gi is found to evolve at a faster slope than that of normal sne ia during the nebular phase , a trend that is also observed for the hv sn ia 2006x @xcite and most of the other hv events ( wang , x. , et al .
2009c , in preparation ) .
sn 2007gi was detected at @xmath94 days after @xmath3 maximum in _
hst_/wfpc2 images . the magnitudes and colors at this very late - time epoch suggest that sn 2007gi may not be contaminated by a light echo on interstellar scales .
using a dozen well - observed sne ia , we confirm the previous claim that hv sne ia tend to have a faster rise time than normal sne ia @xcite : at the same value of @xmath35 , the rise time of the hv objects is shorter than that of the normal objects by 12 days .
the spectral evolution of sn 2007gi is characterized by high expansion velocities measured from the absorption lines .
the value of @xmath0 is measured to be 15,500 km s@xmath2 at @xmath80 day , @xmath150% higher than what is observed for normal sne ia . a small flux ratio @xmath74(si ii ) = [email protected]
is observed , which traditionally means a hot photospheric temperature in the sn ejecta .
our synthetic spectral analysis using the synow code , on the other hand , suggests that sn 2007gi has a typical photospheric temperature compared with other normal sne ia .
the two - stage evolution in the @xmath3-band light curve of sn 2007gi may suggest that an additional energy source besides radioactive decay plays a role in the nebular phase .
dust scattering ( a light echo ) seems unlikely because the tail luminosity would have remained nearly constant for a long time , which is not the case for sn 2007gi .
nevertheless , the possibility of a local light echo on a small scale ( due to csm dust ) can not be completely ruled out .
another potential energy source is interaction with csm ; detection of variable sodium lines in the spectra of some hv events provides some evidence for this scenario . assuming that the csm lies at a distance of @xmath95 cm from the supernova , it is expected that the outermost ejecta will begin to interact with the csm at @xmath130 days after the @xmath3 maximum .
this naturally accounts for a flatter evolution seen in the @xmath3-band light curve of the hv sne ia after @xmath96 days .
the difficulty with this scenario is the lack of convincing observational evidence for stripped material , such as low - velocity h@xmath12 emission in the nebular spectra [ e.g. , @xcite , but see @xcite for alternative explanations regarding the observed lack of hydrogen ] , and the lack of spectral evidence in support of ongoing csm interaction .
a possible explanation for the hv features observed in the hv sne ia is that there are density enhancements of the imes in the outer layers of the ejecta of the hv objects ( due to a metallicity effect , delayed detonations , or csm interaction ) ; the @xmath67-ray heating is less effective , and the photosphere is formed at an outer layer with a higher expansion velocity compared to the normal objects .
we thank peter nugent for help with the synow fitting .
this study is supported by the chinese national natural science foundation through grants 10673007 , 10673012 , and 10603006 , and by the china-973 program 2009cb824800 .
a.v.f.s group at u.c .
berkeley is grateful for nsf grants ast-0607485 and ast-0908886 , the tabasgo foundation , and us department of energy grants de - fc02 - 06er41453 ( scidac ) and de - fg02 - 08er41563 .
financial support for this work was also provided by nasa through grant go-10877 from the space telescope science institute , which is operated by associated universities for research in astronomy , inc .
, under nasa contract nas 5 - 26555 . the work of l.w .
is supported by nsf grant ast-0708873 .
j.d . and y.q .
are supported by the nsfc grant no .
10673014 and the china-973 program 2009cb824800 .
some of the data presented herein were obtained at the w. m. keck observatory , which is operated as a scientific partnership among the california institute of technology , the university of california , and nasa ; the observatory was made possible by the generous financial support of the w. m. keck foundation .
we are grateful to the staffs at the lick and keck observatories for their assistance with the observations . , @xmath44 , and @xmath45 color curves of sn 2007gi compared with those of sne 1994d , 1996x , 2002bo , 2002er , 2002dj , 2004eo , and 2006x .
all of the comparison sne have been dereddened .
the solid line in the @xmath10 panel shows the unreddened lira - phillips locus .
the data sources are cited in the text.,width=566 ] region of an @xmath8-band image taken with the lick observatory 1-m nickel telescope , and the _
hst_/wfpc2 f450w image , respectively , after the images have been astrometrically registered .
the position of sn 2007gi in the nickel image is marked as a circle in the middle panel .
the right panel shows a @xmath97 region around the site of sn 2007gi in the wfpc2/f450w image .
the location of sn 2007gi is marked with a circle having a radius that is three times the astrometric precision .
there is a single object near the center of the circle which we identify as sn 2007gi.,width=566 ] -band rise time of sn 2007gi and other sne ia .
the sample is divided into two groups : hv ( solid circles ) [ sne 2002bo @xcite , 2003w @xcite , 2006le @xcite , 2006x @xcite , 2007gi ( this paper ) ] and normal ( open squares ) [ sne 1990n @xcite , 1994d @xcite , 2001el @xcite , 2002er @xcite , 2003cg @xcite , 2004eo @xcite , 2005cf @xcite , 2007af @xcite and sne 1999cp , 2001en , 2001ep , 2002dj , 2002eb , 2003fa , 2006gr ( ganeshalingam et al .
2009 , in preparation ; wang x. et al .
2009c , in preparation)].,width=566 ] , @xmath71 , + 6 , and + 3 months from @xmath3-band maximum , from top to bottom ) , compared to those of sne 1994d @xcite , 1996x @xcite , 2002bo @xcite , 2002dj @xcite , 2002er @xcite , 2004eo @xcite , and 2006x @xcite at the same phase .
all of the spectra have been corrected for reddening and host - galaxy redshift , and shifted in the vertical direction for clarity.,width=566 ] lcccccc star & @xmath27 & @xmath43 & @xmath98 & @xmath44 & @xmath45 + 1 & 13.860(002 ) & 0.823(003 ) & 2.392(016 ) & 0.475(002 ) & 0.726(002 ) + 2 & 14.233(002 ) & 0.860(003 ) & 2.702(020 ) & 0.482(003 ) & 0.697(002 ) + 3 & 14.833(002 ) & 0.742(004 ) & 2.399(022 ) & 0.395(004 ) & 0.585(004 ) + 4 & 15.375(004 ) & 1.413(008 ) & 3.367(067 ) & 0.929(005 ) & 1.363(004 ) + 5 & 15.298(003 ) & 0.800(006 ) & 2.532(032 ) & 0.443(005 ) & 0.648(005 ) + 6 & 15.408(004 ) & 0.762(007 ) & 2.538(034 ) & 0.451(005 ) & 0.654(006 ) + 7 & 15.869(005 ) & 0.878(009 ) & 2.655(045 ) & 0.522(007 ) & 0.783(007 ) + 8 & 16.175(006 ) & 0.781(010 ) & 2.474(047 ) & 0.410(009 ) & 0.611(010 ) + lcrcccc ut date & jd @xmath24 2,450,000 & phase & @xmath3 ( mag ) & @xmath27 ( mag ) & @xmath8 ( mag ) & @xmath61 ( mag ) + 2007 aug 3 & 4316.17 & -10.85 & 15.170(0.048 ) & 14.940(0.055 ) & & + 2007 aug 4 & 4317.70 & -9.32 & 14.510(0.050 ) & 14.420(0.045 ) & & + 2007 aug 7 & 4320.18 & -6.84 & 13.870(0.052 ) & 13.840(0.053 ) & & + 2007 aug 8 & 4321.02 & -6.00 & 13.678(0.016 ) & 13.615(0.029 ) & 13.488(0.021 ) & 13.553(0.035 ) + 2007 aug 9 & 4322.03 & -4.99 & 13.505(0.015 ) & 13.450(0.028 ) & 13.391(0.034 ) & 13.446(0.029 ) + 2007 aug 11 & 4324.03 & -2.99 & 13.323(0.021 ) & 13.269(0.048 ) & 13.190(0.018 ) & + 2007 aug 12 & 4325.02 & -2.00 & 13.321(0.023 ) & 13.189(0.028 ) & 13.197(0.017 ) & 13.320(0.014 ) + 2007 aug 14 & 4327.01 & -0.01 & 13.259(0.037 ) & 13.106(0.052 ) & 13.153(0.023 ) & 13.383(0.023 ) + 2007 aug 15 & 4328.01 & + 0.99 & 13.237(0.028 ) & 13.077(0.029 ) & 13.130(0.033 ) & 13.374(0.032 ) + 2007 aug 21 & 4334.00 & + 6.98 & 13.571(0.023 ) & 13.238(0.018 ) & 13.347(0.067 ) & 13.647(0.095 ) + 2007 sep 3 & 4347.02 & + 20.00 & 15.236(0.024 ) & 13.899(0.015 ) & 13.769(0.028 ) & 13.642(0.018 ) + 2007 sep 4 & 4347.99 & + 20.97 & 15.330(0.029 ) & 13.944(0.018 ) & 13.753(0.045 ) & 13.612(0.033 ) + 2007 sep 5 & 4348.99 & + 21.97 & 15.414(0.028 ) & 13.997(0.025 ) & 13.756(0.036 ) & + 2007 sep 6 & 4349.99 & + 22.97 & 15.533(0.047 ) & 14.089(0.041 ) & 13.804(0.050 ) & + 2007 oct 13 & 4387.30 & + 60.28 & 16.514(0.020 ) & 15.753(0.012 ) & 15.572(0.017 ) & 15.671(0.026 ) + 2007 oct 16 & 4390.36 & + 63.34 & 16.532(0.035 ) & 15.807(0.014 ) & 15.695(0.027 ) & 15.791(0.036 ) + 2007 oct 21 & 4395.39 & + 68.37 & 16.586(0.020 ) & 15.972(0.013 ) & 15.846(0.019 ) & 15.973(0.035 ) + 2007 oct 22 & 4396.34 & + 69.32 & 16.626(0.021 ) & 15.986(0.013 ) & 15.893(0.026 ) & 16.015(0.031 ) + 2007 oct 24 & 4398.34 & + 71.32 & 16.613(0.031 ) & 16.014(0.012 ) & 15.980(0.038 ) & 16.128(0.047 ) + 2007 oct 25 & 4399.36 & + 72.34 & 16.622(0.033 ) & 16.034(0.019 ) & 16.000(0.049 ) & 16.154(0.050 ) + 2007 oct 29 & 4403.40 & + 76.38 & 16.736(0.059 ) & 16.179(0.081 ) & 16.228(0.081 ) & + 2007 nov 12 & 4417.38 & + 90.36 & 16.857(0.016 ) & 16.551(0.020 ) & 16.612(0.017 ) & 16.643(0.056 ) + 2007 nov 20 & 4425.36 & + 98.34 & 17.028(0.014 ) & 16.735(0.010 ) & 16.889(0.014 ) & 16.979(0.027 ) + 2007 nov 28 & 4433.36 & + 106.34 & 17.174(0.033 ) & 16.961(0.022 ) & 17.120(0.020 ) & 17.145(0.035 ) + 2007 dec 22 & 4457.36 & + 130.34 & 17.606(0.049 ) & 17.425(0.053 ) & 17.771(0.041 ) & 17.729(0.057 ) + 2008 jan 6 & 4472.36 & + 145.34 & 17.845(0.023 ) & 17.861(0.032 ) & 18.301(0.044 ) & 18.218(0.086 ) + 2008 jan 9 & 4475.26 & + 148.24 & 17.852(0.057 ) & 17.907(0.058 ) & 18.315(0.065 ) & 18.274(0.140 ) + 2008 feb 18 & 4515.29 & + 188.27 & 18.480(0.030 ) & 18.520(0.126 ) & & + cccccc stretch & @xmath35 & @xmath99 & explosion velocity & @xmath100 & temperature + & mag & mag & km s@xmath2 & mag & k + 0.88 @xmath4 0.02 & 1.31 @xmath4 0.09 & -19.26 @xmath4 0.10 & 15 , 500 @xmath4 300 & 0.17 @xmath4 0.04 & 11 , 500 + lcrlc ut date & jd @xmath24 2,450,000 & phase & range( ) & resolution( ) + 2007 aug 7 & 4319.50 & -7.52 & 330010,000 & 5 - 12 + 2007 aug 14 & 4326.50 & -0.52 & 330010,000 & 5 - 12 + 2007 aug 21 & 4333.50 & + 6.48 & 330010,000 & 5 - 12 + 2007 nov 12 & 4416.50 & + 89.48 & 31009200 & 6 + 2008 jan 15 & 4480.50 & + 153.48&330010,000 & 5 - 12 + altavilla , g. , et al .
2004 , , 349 , 1344 astier , p. , et al .
2006 , , 447 , 31 benetti , s. , et al .
2004 , , 348 , 261 benetti , s. , et al .
2005 , , 623 , 1011 blondin , s. , et al .
2009 , , 693 , 207 bongard , s. , baron , e. , smadja , g. , branch , d. , & hauschildt , p. h. 2006 , , 647 , 513 bongard , s. , baron , e. , smadja , g. , branch , d. , & hauschildt , p. h. 2008 , , 687 , 456 branch , d. , et al .
2005 , , 117 , 545 branch , d. , et al .
2006 , , 118 , 560 brown , p. et al .
2009 , , 137 , 4517 cappellaro , e. , et al .
2001 , , 549 , l215 cardelli , j. a. , clayton , g. c. , & mathis , j. s. 1989 , , 345 , 245 chugai , n. , n. , 2008 , astronomy letters , 34 , 389 cousins , a. w. j. 1981 , s. african astron .
6 , 4 crotts , a. , & yourdon , d. 2008 , , 689 , 1186 dolphin , a. e. 2000a , , 112 , 1383 dolphin , a. e. 2000b , , 112 , 1397 elias - rosa , n. , et al .
2006 , , 369 , 1880 filippenko , a. v. 1982 , , 94 , 715 fisher , a. , et al .
1999 , , 304 , 67 gerardy , c. , et al .
2004 , , 607 , 391 goldhaber , g. , et al .
1998 , , 113 , 1325 guy , j. , et al .
2005 , , 443 , 781 guy , j. , et al .
2007 , , 466 , 11 hachinger , s. , mazzali , p. a. , & benetti , s. 2006 , , 370 , 299 harutyunyan , a. , benetti , s. , & cappellaro , e. 2007 , cbet , 1021 , 1 hicken , m. , et al .
2009 , arxiv:0901.4787v3 jha , s. , riess , a. g. , & kirshner , r. p. 2007
, , 695 , 122 johnson , h. l. , iriarte , b. , mitchell , r. i. , & wisniewskj , w. z. 1966 , comm .
lunar planet . lab . , 4 , 99z .
r. wang ( cambridge : cambridge univ . press ) , 41 kasen , d. 2006 , , 649 , 939 kotak , r. , et al . 2005 , , 436 , 1021 krisciunas , k. , et al .
2003 , , 125 , 166 landolt , a. u. 1992 , , 104 , 340 leibundgut , b. , & pinto , p. a. 1992 , , 401 , 49 leloudas , g. , et al .
2009 , arxiv:0908.0537 lentz , j. e. , baron , e. , branch , d. , & hauschildt , p. h. 2001 , , 547 , 402 leonard , d. c. , et al .
2008 , , 120 , 1259 li , j. 2008 , phd thesis , tsinghua university li , w. , et al .
2002 , , 114 , 403 li , w. , et al .
2003 , 586 l9 li , w. , et al .
2007 , , 661 , 1013 lira , p. 1995 , master thesis , univ .
chile lira , p. , et al .
1998 , , 115 , 234 mattila , s. , et al .
2005 , , 443 , 649 miller , j. s. , & stone , r. p. s. 1993 , lick obs .
tech . rep .
66 milne , p. a. , the , l. s. , & leising , m. d. 2001 , , 559 , 1019 nugent , p. , et al .
1995 , 455 , l147 oke , j. b. , et al .
1995 , , 107 , 375 pastorello , a. , et al .
2007 , , 376 , 1301 patat , f. , et al .
1996 , , 278 , 111 patat , f. , et al .
2007 , science , 317 , 924 perlmutter , s. , et al .
1997 , , 483 , 565 perlmutter , s. , et al .
1999 , , 517 , 565 phillips , m. m. 1993 , , 413 , l105 phillips , m. m. , et al .
1999 , , 118 , 1766 pignata , g. , et al .
2004 , , 355 , 178 pignata , g. , et al .
2008 , , 388 , 971 poole , t. s. , et al .
2008 , , 383 , 627 roming , p. w. , et al .
2005 , space science reviews , 120 , 95 richmond , m. w. , et al .
1995 , , 109 , 2121 riess , a. g. , press , w. h. , & kirshner , r. p. 1995
, , 438 , l17 riess , a. g. , et al .
1998 , , 116 , 1009 riess , a. g. , et al .
1999 , , 118 , 2675 riess , a. g. , et al .
2004 , , 607 , 665 riess , a. g. , et al .
2007 , , 659 , 98 salvo , m. e. , et al .
2001 , , 321 , 254 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 schmidt , b. p. , et al .
1994 , , 434 , l19 simon , j. d. , et al .
2007 , , 671 , l25 simon , j. d. , et al .
2009 , , 702 , 1157 sollerman , j. , et al .
2004 , , 428 , 555 stetson , p. b. 1987 , , 99,191 strovink m. 2007 , , 671 , 1084 tanaka , m. , mazzali , p. a. , maeda , k. , & nomoto , k. 2006 . , 645 , 470 tully , r. b. 1988 , nearby galaxy catalog , cambridge university press tripp , r. 1998 , , 331 , 815 wang , l. , et al .
2003 , , 590 , 944 wang , x. , et al . 2005 , , 620 , l87 wang , x. , et al . 2008a , , 675 , 626 wang , x. , et al .
2008b , , 677 , 1060 wang , x. , et al .
2009a , , 699 , l139 ( w09 ) wang , x. , et al . 2009b , , 697 , 380 wood - vasey , w. m. , et al .
2007 , , 666 , 694 yamanaka , m. , et al .
2009 , , in press ( arxiv:0904.2763 ) | we present optical photometry and spectra for the type ia supernova ( sn ia ) 2007gi in the nearby galaxy ngc 4036 .
sn 2007gi is characterized by extremely high - velocity ( hv ) features of the intermediate - mass elements ( si , ca , and s ) , with expansion velocities ( @xmath0 ) approaching @xmath115,500 km s@xmath2 near maximum brightness ( compared to @xmath110,600 km s@xmath2 for sne ia with normal @xmath0 ) .
sn 2007gi reached a @xmath3-band peak magnitude of [email protected] mag with a decline rate of @xmath5(true ) = [email protected] mag .
the @xmath3-band light curve of sn 2007gi demonstrated an interesting two - stage evolution during the nebular phase , with a decay rate of [email protected] mag ( 100 days)@xmath2 during @xmath690 days and 1.61@xmath7 mag ( 100 days)@xmath2 thereafter .
such a behavior was also observed in the hv sn ia 2006x , and might be caused by the interaction between supernova ejecta and circumstellar material ( csm ) around hv sne ia .
based on a sample of a dozen well - observed @xmath8-band ( or unfiltered ) light curves of sne ia , we confirm that the hv events may have a faster rise time to maximum than the ones with normal @xmath0 . |
the one - dimensional diffusion in an external potential @xmath0 is described by the fokker - planck equation @xcite @xmath2,\ ] ] where @xmath3 is related to the potential @xmath0 by @xmath4 with @xmath5 , equation ( [ 1eq1.1 ] ) reduces to the time - independent form @xmath6=k^2\phi(x).\ ] ] in addition to being a fundamental equation for nonequilibrium phenomena , the fokker - planck equation is of particular importance in its relation to the schrdinger equation .
setting @xmath7 , we can transform ( [ 1eq1.3 ] ) into a steady - state schrdinger equation @xmath8 where @xmath9 , and @xmath10 historically , methods developed for the schrdinger equation have been applied to the study of the fokker - planck equation . but this should be the other way around as well .
theoretically , the fokker - planck equation is often more convenient to deal with than the schrdinger equation . with the fokker - planck equation
we can carry out a more systematic analysis , and the results obtained for the fokker - planck equation can be applied to quantum - mechanical problems described by the schrdinger equation . in this paper , we study the high - energy ( large-@xmath11 ) behavior of the green function for ( [ 1eq1.3 ] ) or , equivalently , ( [ 1eq1.4 ] ) .
although we shall mainly work with the fokker - planck equation , the results of this paper are directly applicable to the schrdinger equation , too .
one - dimensional quantum scattering has a long history .
in particular , high - energy asymptotic behavior of the green function and related functions have been studied over the decades by both physicists and mathematicians .
this classical topic has recently attracted attention in connection with inverse problems and the theory of integrable systems , and this area of research remains active even to the present day @xcite@xcite .
an essential matter in the study of an asymptotic expansion is the estimation of the remainder term . in conventional methods , which are mostly based on an integral equation ,
it is necessary to impose strong conditions on the potential in order to control the remainder term .
for example , the integrability of @xmath12 is often required . for non - integrable @xmath12
, different specific methods need to be used .
there has not yet been a formalism in which the asymptotic analysis of the green function can be carried out systematically and in a unified way for a variety of potentials that may not necessarily be integrable and that may not even be finite as @xmath13 . in this paper , we take a totally new approach to this problem , and present a new method for the high - energy analysis of the green function which is applicable to a larger class of potentials . our method is based on the analysis of reflection coefficients .
reflection coefficients are fundamental quantities in scattering theory , and they serve as building blocks for constructing the green function .
the green function can be expressed solely in terms of reflection coefficients for semi - infinite intervals @xcite .
high - energy behavior of the reflection coefficients was studied in a previous paper @xcite , and formulas for their asymptotic expansion were derived there . in the present paper , we apply these results to the green function , and derive the expansion in powers of the inverse wave number .
the coefficients of the expansion are expressed in a simple form in terms of a linear operator , and the remainder term is expressed in terms of transmission and reflection coefficients for finite intervals .
the validity of the asymptotic expansion can be studied by using this expression for the remainder term .
the short - time expansion of the green function can also be obtained by this method .
we assume that the ( fokker - planck ) potential @xmath0 is a real function which is finite for finite @xmath14 , and which either converges to a finite limit or diverges to infinity ( @xmath15 or @xmath16 ) as @xmath17 .
the quantity @xmath18 is taken to be a complex number with @xmath19 . in sections 2 and 3 , we review the relevant results of previous papers . we derive the asymptotic expansion of the green function in sections 46 , and discuss the conditions for its validity in sections 7 and 8 . in section 9 ,
the short - time expansion is studied . in section 10 , we explain how our method can be applied to the schrdinger equation .
various examples are given in section 11 .
we define the time - dependent green function @xmath20 as the solution of @xmath21g_{\rm f}(x , y;t)=\delta(x - y)\delta(t)\ ] ] with the condition that @xmath22 for @xmath23 . physically , @xmath20 is the probability density of finding the diffusing particle at position @xmath14 at time @xmath24 , under the condition that it was initially at position @xmath25 .
we also define its fourier transform @xmath26 this @xmath27 is the green function for equation ( [ 1eq1.3 ] ) with @xmath28 . in the same way , we define the retarded green function for the schrdinger equation as the solution of @xmath29g_{\rm s}(x , y;t)=\delta(x - y)\delta(t)\ ] ] satisfying the condition @xmath30 for @xmath23 .
its fourier transform defined by @xmath31 is the green function for the steady - state schrdinger equation ( [ 1eq1.4 ] ) , satisfying @xmath32g_{\rm s}(x , y;e ) = \delta(x - y).\ ] ] it is easy to see that @xmath33 and @xmath34 are related by @xmath35/2}g_{\rm s}(x , y;e = i\omega).\ ] ] for convenience , we define @xmath36 as a function of complex @xmath18 , and deal with this @xmath37 instead of @xmath33 or @xmath34 .
( we shall refer to @xmath37 as the green function , too . ) since @xmath38 , without loss of generality we assume that @xmath39 . to introduce our basic expression for the green function ,
let us first define the transmission and reflection coefficients for finite intervals .
we consider an interval @xmath40 , and define @xmath41 namely , @xmath42 is identical to @xmath0 within @xmath40 and constant outside this interval .
we consider equation ( [ 1eq1.3 ] ) with @xmath3 replaced by @xmath43 : @xmath44=k^2\phi(x).\ ] ] ( in general , the left - hand side of this equation contains delta functions at @xmath45 and @xmath46 coming from the derivative of @xmath47 . ) since @xmath48 for @xmath49 and @xmath50 , equation ( [ 1eq2.9 ] ) has two solutions of the form [ 1eq2.10 ] @xmath51/2 } \tau(x_2,x_1;k ) e^{-ik(x - x_1 ) } & $ x < x_1 $ , \\
e^{-ik(x - x_2 ) } + r_r(x_2,x_1;k ) e^{ik(x - x_2 ) } & $ x > x_2 $ , \\ } \ ] ] @xmath52/2 } \tau(x_2,x_1;k ) e^{ik(x - x_2 ) } & $ x > x_2$. \\ } \ ] ] ( the factor @xmath53/2}$ ] in front of @xmath54 is necessary since @xmath55 and @xmath56 are solutions of the fokker - planck equation , not the schrdinger equation . ) the transmission coefficient @xmath54 , the right reflection coefficient @xmath57 , and the left reflection coefficient @xmath58 for the interval @xmath40 are defied by ( [ 1eq2.10 ] ) .
we can let @xmath59 in ( [ 1eq2.10]a ) or @xmath60 in ( [ 1eq2.10]b ) to define the reflection coefficients for semi - infinte intervals : @xmath61 let us define @xmath62 and @xmath63 it was shown in @xcite that the green function can be expressed in terms of this @xmath64 as @xmath65[1-s(y , k ) ] } } \exp\left [ ik(x - y)-ik\int_y^x s(z , k)\,dz \right].\ ] ] the logarithm of ( [ 1eq2.14 ] ) reads @xmath66-\frac{1}{2}\log[1-s(y , k ) ] .
\nonumber \\\end{aligned}\ ] ] we shall carry out the analysis of the green function on the basis of this expression .
in this section , we review the formulas derived in @xcite for the high - energy expansion of the reflection coefficients . here
we deal only with @xmath57 .
( corresponding formulas for @xmath58 can be obtained in the same way . ) first , we define the generalized transmission and reflection coefficients , with an additional argument @xmath67 , as [ 1eq3.1 ] @xmath68 @xmath69 @xmath70 we also define the operator @xmath71 which acts on functions of @xmath14 and @xmath67 as @xmath72 + \xi^{-1}\int_0^\xi d\xi \,\frac{\partial}{\partial x}g(x,\xi),\ ] ] where @xmath73 is an arbitrary function , and @xmath3 is the function in ( [ 1eq1.1])([1eq1.3 ] ) .
it was shown in @xcite that , for an arbitrary nonnegative integer @xmath74 , @xmath75 where @xmath76 @xmath77 @xmath78 in ( [ 1eq3.5 ] ) , @xmath79 is the quantity obtained from @xmath80 by the substitution @xmath81 .
it is convenient to define @xmath82 so that ( [ 1eq3.4 ] ) and ( [ 1eq3.6 ] ) read @xmath83 @xmath84 it is easy to calculate @xmath85 from ( [ 1eq3.8 ] ) by using definition ( [ 1eq3.2 ] ) .
we have @xmath86 obviously @xmath85 is an @xmath87th degree polynomial in @xmath67 , whose coefficients consist of the powers of @xmath88 and its derivatives . the @xmath89 s
are obtained from ( [ 1eq3.9 ] ) and ( [ 1eq3.10 ] ) as @xmath90 equations ( [ 1eq3.3])([1eq3.6 ] ) hold for any @xmath3 as long as @xmath91 and @xmath92 make sense .
however , ( [ 1eq3.3 ] ) is meaningful as a high - energy expression only if @xmath93 using ( [ 1eq3.5 ] ) with ( [ 1eq3.9 ] ) , we can study the conditions for ( [ 1eq3.12 ] ) to hold . for simplicity ,
let us assume that @xmath3 and all its derivatives are monotone for sufficiently large @xmath94 .
( this condition is unnecessarily strong and can be relaxed , but we make this assumption in order to simplify the presentation . )
then it can be shown , but this is sufficient .
( see the comment below equation ( 5.16 ) of @xcite . )
if ( [ 1eq3.12 ] ) holds for @xmath95 then it holds for any @xmath67 .
this can be easily shown by substituting the expansion of @xmath57 ( equation ( [ 1eq3.3 ] ) with @xmath95 ) into the right - hand side of ( [ 1eq3.1]a ) . ]
that : 1 . when @xmath96 with fixed @xmath97 in the range @xmath98 , equation ( [ 1eq3.12 ] ) holds if ( @xmath88 is @xmath99 times differentiable and ) @xmath100 is continuous and piecewise smooth .
2 . when @xmath96 with fixed @xmath101 , equation ( [ 1eq3.12 ] ) holds if @xmath100 is continuous and piecewise smooth , and if @xmath102 for any positive number @xmath103 .
3 . when @xmath96 with @xmath104 , equation ( [ 1eq3.12 ] ) holds if @xmath100 is continuous and piecewise smooth , and if @xmath105 is finite .
this means that if the conditions in ( i ) , ( ii ) or ( iii ) hold , then ( [ 1eq3.12 ] ) holds in the region ( i ) @xmath106 , ( ii ) @xmath107 , ( iii ) @xmath19 , respectively , where @xmath108 is an arbitrary positive number .
( we shall always let @xmath108 stand for a positive quantity . )
we wish to derive the @xmath110-expansion of @xmath111 from ( [ 1eq2.15 ] ) by using the formulas given in the previous section .
to do so , we need to study the expansion of @xmath64 and @xmath112 . from ( [ 1eq2.12 ] ) and ( [ 1eq3.1]a )
, we find @xmath113.\ ] ] therefore , the expansion of @xmath114 is obtained by substituting ( [ 1eq3.3 ] ) into ( [ 1eq4.1 ] ) as @xmath115 where , using definition ( [ 1eq3.7 ] ) , @xmath116 @xmath117 setting @xmath118 in ( [ 1eq3.10 ] ) , we have @xmath119 substituting ( [ 1eq3.5 ] ) with ( [ 1eq3.1 ] ) into ( [ 1eq4.4 ] ) gives the expression for @xmath120 , @xmath121 ^ 2 k_n(z,\eta_l)\,dz,\ ] ] where we have defined @xmath122 the expressions for @xmath123 corresponding to ( [ 1eq4.2 ] ) , ( [ 1eq4.6 ] ) and ( [ 1eq4.7 ] ) can be derived in the same way , using the analogues of the formulas of section 3 for @xmath58 . the result is @xmath124 @xmath125 ^ 2 k_n(z,\eta_r)\,dz,\ ] ] @xmath126 the coefficients @xmath127 in ( [ 1eq4.8 ] ) are the same ones as in ( [ 1eq4.2 ] ) .
the expansion of @xmath128 ( equation ( [ 1eq2.13 ] ) ) is obtained by adding ( [ 1eq4.2 ] ) and ( [ 1eq4.8 ] ) : @xmath129+\sigma_n(x , k),\ ] ] @xmath130 where @xmath131 note that @xmath132 for any even number @xmath74 .
obviously , a sufficient condition for @xmath133 is that the following two equations hold : [ 1eq4.15 ] @xmath134 from ( [ 1eq4.4 ] ) we can see that the first equation of ( [ 1eq4.15 ] ) holds if ( [ 1eq3.12 ] ) holds .
( the limit @xmath135 does not interfere with the limit @xmath96 .
we can also show this directly by substituting the expansion of @xmath57 into the first equation of ( [ 1eq2.12 ] ) . )
therefore , the first equation of ( [ 1eq4.15 ] ) holds under the conditions given in section 3 .
the conditions for the second equation are obvious analogues .
hence it is apparent that ( [ 1eq4.14 ] ) holds in the sector @xmath106 if @xmath100 is continuous and piecewise continuously differentiable .
if , in addition , both @xmath136 and @xmath105 are finite , then ( [ 1eq4.14 ] ) holds for @xmath137 .
from ( [ 1eq4.11 ] ) we can derive the expansion of @xmath139 as @xmath140= \frac{1}{(2ik)^2}\alpha_2(x)+\frac{1}{(2ik)^4}\alpha_4(x)+\cdots + \frac{1}{(2ik)^{n'}}\alpha_{n'}(x)+\rho_n(x , k ) , \nonumber \\ \fl \end{aligned}\ ] ] where the coefficients @xmath141 are expressed in terms of @xmath142 as @xmath143 ( the second sum in ( [ 1eq5.2 ] ) is over @xmath144 for each @xmath145 @xmath146 with the constraint @xmath147 . ) the remainder term @xmath148 of ( [ 1eq5.1 ] ) can be written as @xmath149,\ ] ] where @xmath150 , so that @xmath151 ( see ( [ 1eq4.11 ] ) ) .
the quantity in the brackets on the right - hand side of ( [ 1eq5.3 ] ) is @xmath152 as @xmath96 . hence @xmath153+o(1/\vert k
\vert^{n+1 } ) \qquad ( \vert k \vert \to \infty)\ ] ] provided that @xmath154 vanishes as @xmath155 .
note also that @xmath156 the coefficients @xmath157 of ( [ 1eq5.1 ] ) , which are given by ( [ 1eq5.2 ] ) , can be expressed in a more compact form . from equations ( [ 1eqa.3 ] ) of appendix
a , we obtain @xmath158 = f(x ) + ik[s_r(x , k)-s_l(x , k)].\ ] ] substituting ( [ 1eq4.2 ] ) and ( [ 1eq4.8 ] ) into the right - hand side of ( [ 1eq5.5 ] ) yields @xmath159 = \frac{1}{(2ik)^2}s_3(x)+\frac{1}{(2ik)^4 } s_5(x)+\frac{1}{(2ik)^6 } s_7(x)+\cdots,\ ] ] where we have used @xmath160 . comparing ( [ 1eq5.1 ] ) with ( [ 1eq5.6 ] )
we find that @xmath161 for any even number @xmath145 .
integrating both sides of this equation gives @xmath162 the integral on the right - hand side is uniquely determined in such a way that @xmath163 includes no additional integration constant when expressed in terms of @xmath88 and its derivatives .
for example , @xmath164 , @xmath165 , as can be seen from ( [ 1eq4.5 ] ) . in this way
, @xmath166 is a total derivative for odd @xmath167 .
hence we have @xmath168 if @xmath169 or @xmath170 , we can write the right - hand side of ( [ 1eq5.8 ] ) as @xmath171 or @xmath172 .
but ( [ 1eq5.8 ] ) holds even if neither @xmath169 nor @xmath170 . equations ( [ 1eq5.1 ] ) and ( [ 1eq5.5 ] ) also give the relation between the remainder terms @xmath173.\ ] ]
substituting ( [ 1eq4.11 ] ) , ( [ 1eq5.1 ] ) and ( [ 1eq5.8 ] ) into ( [ 1eq2.15 ] ) , we obtain , for any integer @xmath176 , @xmath177 where [ 1eq6.2 ] @xmath178 @xmath179 we can also write the right - hand side of ( [ 1eq6.2]b ) as @xmath180 with @xmath163 given by ( [ 1eq5.2 ] ) . for ( [ 1eq6.1 ] ) to be meaningful as a high - energy expansion , @xmath181 must satisfy @xmath182 let us study the conditions for ( [ 1eq6.5 ] ) using expression ( [ 1eq6.3 ] ) . we first remark
that ( [ 2eq5.4 ] ) and ( [ 2eq5.5 ] ) imply @xmath183+o(1/\vert k \vert^{n+1 } ) \qquad ( \vert k \vert \to \infty).\ ] ] now suppose that @xmath184 .
then , @xmath185 and so ( [ 1eq6.5 ] ) holds .
( since @xmath186 is uniformly bounded in the interval @xmath187 , the limit @xmath96 and the integral can be interchanged .
the second equation follows from ( [ 2eq6.6 ] ) . )
we know that @xmath184 is satisfied under the conditions stated at the end of section 4 , with @xmath74 replaced by @xmath188 .
hence it follows that ( [ 1eq6.5 ] ) holds for @xmath189 if @xmath190 is continuous and piecewise continuously differentiable .
it holds for @xmath137 if , in addition , both @xmath105 and @xmath136 are finite .
the conditions for ( [ 1eq6.5 ] ) mentioned at the end of the last section are sufficient conditions , not necessary ones . here
we show that @xmath190 need not be continuous for ( [ 1eq6.5 ] ) to hold .
suppose that @xmath192 is @xmath193 times continuously differentiable , @xmath194 is continuously differentiable except at @xmath195 , and that @xmath196 has a finite jump at @xmath195 : @xmath197 here @xmath198 is the heaviside step function , and @xmath199 is a constant .
then @xmath200 contains a delta function as @xmath201 .
it is easy to show that @xmath89 ( equation ( [ 1eq3.9 ] ) ) has the form @xmath202 , where the remaining terms on the right - hand side do not contain derivatives of order @xmath145 or higher .
( see ( [ 1eq3.11 ] ) . ) therefore , @xmath203 substituting this into ( [ 1eq4.6 ] ) and ( [ 1eq4.9 ] ) with @xmath204 , and then into ( [ 1eq4.12 ] ) , we have @xmath205 ^ 2 \delta(w)\,dw \nonumber \\ \fl & \qquad - \frac{c}{(-2ik)^{m+1}}\int_z^\infty \left[\frac{\tau(w , z)}{1+r_l(w , z)}\right]^2 \delta(w)\,dw + \cdots \nonumber \\ \fl & = \frac{c}{(2ik)^{m+1 } } \left\ { \left[\frac{\tau(z,0)}{1+r_r(z,0)}\right]^2 \theta(z ) + ( -1)^m\left[\frac{\tau(0,z)}{1+r_l(0,z)}\right]^2 \theta(-z ) \right\}+\cdots . \nonumber \\
\fl\end{aligned}\ ] ] the quantity in the curly brackets in the last line of ( [ 1eq7.4 ] ) does not vanish as @xmath96 when @xmath18 is real .
( see ( [ 1eqa.2 ] ) . ) but its integral does vanish in this limit even when @xmath104 . indeed , using ( [ 1eqa.5 ] ) , ( [ 1eqa.1 ] ) and ( [ 1eqa.2 ] ) of appendix a we find @xmath206 ^ 2 \theta(z ) + ( -1)^m\left[\frac{\tau(0,z)}{1+r_l(0,z)}\right]^2 \theta(-z ) \right\ } dz \nonumber \\ \fl & \qquad=\frac{1}{2ik } \biggl [ \left(e^{2ikx}-e^{2ik\max(y,0)}\right)\theta(x ) + ( -1)^m\left(e^{-2iky}-e^{-2ik\min(x,0)}\right)\theta(-y ) \biggr ] \nonumber \\ \fl & \qquad \qquad + o(1/\vert k \vert)\qquad
( \vert k \vert \to \infty , \ 0 \leq \arg k \leq \pi ) .
\end{aligned}\ ] ] assuming that the part represented by `` @xmath207 " in ( [ 1eq7.4 ] ) is @xmath208 , we have @xmath209 and , from ( [ 2eq6.6 ] ) , @xmath210 from ( [ 1eq6.3 ] ) , ( [ 1eq7.5 ] ) and ( [ 1eq7.6 ] ) it follows that @xmath211 , so ( [ 1eq6.5 ] ) holds for @xmath212 in spite of the discontinuity of @xmath196 .
the same argument holds when @xmath196 has two or more finite jumps . in summary , the following conclusion can be drawn : 1 .
if @xmath88 is @xmath99 times continuously differentiable and @xmath100 is piecewise continuously differentiable , and if @xmath190 is piecewise continuously differentiable , then ( [ 1eq6.5 ] ) holds in the angular region @xmath106 with arbitrary @xmath213 .
( here @xmath190 may possibly have a finite number of finite jumps . )
2 . if the conditions of ( i ) are satisfied , and if @xmath214 for any @xmath215 , then ( [ 1eq6.5 ] ) holds in the half plane @xmath216 with arbitrary @xmath213 .
3 . if the conditions of ( i ) are satisfied , and if both @xmath136 and @xmath105 are finite , then ( [ 1eq6.5 ] ) holds in the upper half plane @xmath19 including the real axis .
( as noted below equation ( [ 1eq3.12 ] ) , we are assuming that @xmath192 and its derivatives are monotone for sufficiently large @xmath217 , at both @xmath218 and @xmath219 . )
the conditions of ( i ) are still not necessary conditions .
actually , ( [ 1eq6.5 ] ) may hold even when @xmath88 is not @xmath99 times differentiable , as will be shown in the next section .
here we show that ( [ 1eq6.5 ] ) may hold even for @xmath221 when @xmath196 has a discontinuity as in ( [ 1eq7.1 ] ) . to this end , let us study @xmath222 for the case of ( [ 1eq7.1 ] ) . since @xmath223 , we have @xmath224 where @xmath225 is the derivative of the delta function . as before ,
we substitute ( [ 1eq8.1 ] ) into ( [ 1eq4.6 ] ) and ( [ 1eq4.9 ] ) , and then into ( [ 1eq4.12 ] ) . after integrating by parts
, we obtain @xmath226 ^ 2 \right\vert_{w=0}\theta(z ) \right\ } \nonumber \\ \fl & \qquad + \frac{c}{(-2ik)^{m+2 } } \left\ { \delta ( z ) + \left.\frac{\partial}{\partial w}\left[\frac{\tau(w , z)}{1+r_l(w , z)}\right]^2 \right\vert_{w=0}\theta(-z ) \right\ } + \cdots.\end{aligned}\ ] ] using ( [ 1eqa.4 ] ) of appendix a , we can write ( [ 2eq8.2 ] ) as @xmath227 ^ 2 \theta(z ) + ( -1)^m\left[\frac{\tau(0,z)}{1+r_l(0,z)}\right]^2 \theta(-z ) \right\ } \nonumber \\ \fl & \qquad + \frac{2\,c}{(2ik)^{m+2 } }
a_m\delta(z ) + \cdots,\end{aligned}\ ] ] where @xmath228 in ( 8.3 ) , we have disregarded the terms of order @xmath229 which have the form @xmath230 with @xmath231 such that @xmath232 vanishes as @xmath96 .
such terms are irrelevant to the following discussion .
in addition to @xmath233 , the right - hand side of ( 8.1 ) contains another singular term proportional to @xmath234 , but the contribution from this term can be disregarded for the same reason .
we assume that the contribution to ( 8.3 ) from non - singular terms of @xmath235 is @xmath236 as @xmath96 .
now we consider the two cases , @xmath237 and @xmath238 .
\(a ) @xmath237 the integral of ( [ 1eq8.2 ] ) can be calculated by using ( [ 2eq7.5 ] ) . for the case
@xmath237 we have @xmath239 + o(1/\vert k\vert^{m+2}).\ ] ] ( note that the contribution from the delta function in ( [ 1eq8.2 ] ) is canceled . ) from ( [ 2eq5.4 ] ) , ( [ 1eq7.4 ] ) and ( [ 1eqa.2 ] ) ( or from ( [ 2eq6.6 ] ) , ( [ 1eq8.2 ] ) and ( [ 1eqa.2 ] ) ) we obtain [ 1eq8.4 ] @xmath240 @xmath241 on substituting ( [ 1eq8.3 ] ) , ( [ 1eq8.4]a ) and ( [ 1eq8.4]b ) into ( [ 1eq6.3 ] ) with @xmath204 , the terms of order @xmath242 cancel out , and we have @xmath243 namely , ( [ 1eq6.5 ] ) holds for @xmath204 . the effect of the discontinuity of @xmath196 on @xmath244 etc can be studied in the same way .
singular terms involving derivatives of the delta function such as @xmath245 appear in @xmath246 , respectively , but they cause no problems .
singular contributions coming from these terms cancel out in the expressions for @xmath244 etc , as we have seen above for @xmath247 .
( the explanation is omitted here , but the cancellation between @xmath248 and @xmath148 is guaranteed by relation ( [ 1eq5.11 ] ) . ) as a result , the derivatives of the delta function do not produce a term of order @xmath249 ( or lower ) in @xmath181 . in ( [ 1eq6.2 ] ) and ( [ 1eq6.3 ] ) ,
the delta function and its derivatives appear only within an integral , and these expressions are well defined for @xmath250 . in ( [ 1eq6.2 ] ) and ( [ 1eq6.3 ] ) ,
the function multiplying @xmath251 is always @xmath252 times continuously differentiable if @xmath253 ( for ( [ 1eq6.2 ] ) ) or @xmath254 ( for ( [ 1eq6.3 ] ) ) .
an expression like @xmath255 may appear in @xmath256 , where @xmath257 .
] as long as @xmath258 .
however , ( [ 1eq6.3 ] ) is not well defined for @xmath259 since the square of the delta function appears in it .
we can easily see that @xmath260 contains a term proportional to @xmath261 .
in addition , terms proportional to @xmath262 ( @xmath263 ) are contained in @xmath260 .
these terms are ill defined , too , since they produce the square of the delta function by integration by parts . to study @xmath264
, we need to go back one step , as we can not directly use ( [ 1eq6.3 ] ) . by induction
, it can be shown that @xmath265 the terms explicitly shown in ( [ 2eq8.9 ] ) are the ones that give rise to the ill - defined terms in @xmath260 .
we substitute ( [ 2eq8.9 ] ) into the expression for @xmath256
. then we can study @xmath264 by using this expression together with the relation @xmath266 .
after some calculation ) is relevant .
anomalous contribution comes from where @xmath267 appears in formal calculation whereas it should really be @xmath268 or @xmath269 .
for example , @xmath270 @xmath271 , where the first term on the right - hand side is not @xmath272 but @xmath273 . in deriving ( [ 2eq8.10 ] )
, we also use the @xmath110-expansion of @xmath57 , @xmath58 and @xmath54 for finite intervals .
( see the comments below ( [ 1eqa.2 ] ) in appendix a. ) in the expansion , we need only to keep track of the terms which are linear in @xmath88 and its derivatives . ]
we find that @xmath264 contains a term of order @xmath274 as @xmath275 so ( [ 1eq6.5 ] ) does not hold for @xmath259 .
we can see that ( [ 2eq8.10 ] ) gives correction to the coefficient of order @xmath276 as @xmath277+\cdots.\end{aligned}\ ] ] thus , the correct coefficient of order @xmath276 is not @xmath278 but @xmath279 . in summary , when @xmath196 has a finite jump of the form of ( [ 1eq7.1 ] ) , and when the jump is located between @xmath25 and @xmath14 , equation ( [ 1eq6.5 ] ) holds for @xmath280 in the sector @xmath281
provided that the part represented by `` @xmath207 " in ( [ 1eq7.1 ] ) is sufficiently differentiable .
if , in addition , @xmath105 and @xmath136 are both finite , then ( [ 1eq6.5 ] ) holds for @xmath280 even for real @xmath18 . in any case ,
( [ 1eq6.5 ] ) does not hold for @xmath282 .
although @xmath111 can be asymptotically expanded in powers of @xmath110 even beyond the term of order @xmath274 , the coefficients @xmath283 given by ( [ 1eq6.2 ] ) are not correct for @xmath284 .
in particular , @xmath278 is shifted by @xmath285 as in ( [ 2eq8.11 ] ) .
( see examples 58 of section 11 . )
\(b ) @xmath238
similarly , for @xmath238 we have , by using ( [ 2eq7.5 ] ) , @xmath286 @xmath287 while the expression for @xmath288 is the same as ( [ 1eq8.4]a ) . substituting ( [ 1eq8.6 ] ) , ( [ 1eq8.4]a ) and ( [ 1eq8.7 ] ) into ( [ 1eq6.3 ] ) , we obtain @xmath289 if the limit @xmath96 is taken with fixed @xmath97 in the sector @xmath106 , then @xmath290 since @xmath291 falls off exponentially .
the same can be said for @xmath292 , @xmath293 , and so on ; the contribution to @xmath181 coming from the discontinuity of @xmath196 is exponentially small at large @xmath11 as long as @xmath106 . when @xmath18 is real , however , @xmath294 does not vanish as @xmath96 .
so , unlike case ( a ) , equation ( [ 1eq6.5 ] ) does not hold for @xmath221 when the limit is taken along the real axis .
( see example 6 of section 11 . )
the case @xmath295 can be treated in exactly the same way as ( b ) .
it is straightforward to extend the arguments of ( a ) and ( b ) above to the cases where there are two or more such discontinuities .
the expansion of @xmath37 is obtained by exponentiating ( [ 1eq6.1 ] ) as @xmath296,\ ] ] where @xmath297 @xmath298 -\sum_{m=0}^n \frac{b_m}{(2ik)^m}.\ ] ] ( here the prime does not denote a derivative . ) it is obvious that @xmath299 as @xmath96 if @xmath300 .
the time dependent green function for the fokker - planck equation is obtained from @xmath301 by the inverse fourier transformation as @xmath302/2 } \int_{-\infty}^\infty \frac{1}{2i\kappa}g(x , y;\kappa)e^{-i \omega t}d\omega,\ ] ] where @xmath303 is defined by @xmath304 ( when necessary , the integral in ( [ 1eq9.4 ] ) is to be understood as @xmath305 with positive infinitesimal @xmath108 . ) by substituting ( [ 1eq9.1 ] ) into ( [ 1eq9.4 ] ) and carrying out the integration term by term , we can derive an expansion of @xmath33 in powers of @xmath24 .
the result is @xmath306/2}}{\sqrt{4 \pi t } } \exp\left[-\frac{(x - y)^2}{4t}\right ] \left(1+g_1t+g_2 t^2 + g_3 t^3+\cdots+g_n t^n+\tilde\delta_n\right ) , \nonumber \\\end{aligned}\ ] ] where @xmath307 ( see appendix b for the derivation .
the expression for @xmath308 is shown there . )
when @xmath309 , the remainder term @xmath308 satisfies @xmath310 if ( [ 1eq6.5 ] ) is satisfied for @xmath106 .
( see appendix b. ) for ( [ 1eq9.8 ] ) to hold , it is not necessary that ( [ 1eq6.5 ] ) hold for @xmath104 .
so this short - time expansion is valid even when @xmath311 are infinite .
( here we are assuming that @xmath24 is real . )
note that the right - hand side of ( [ 1eq9.7 ] ) is not infinite at @xmath312 in spite of the appearance of the @xmath313 .
for @xmath312 we have @xmath314 as can be directly calculated . the right - hand side of ( [ 1eq9.7 ] ) approaches ( [ 1eq9.11 ] ) as @xmath315 .
from ( [ 1eq4.5 ] ) , we may note that @xmath127 for @xmath316 can be expressed in terms of the schrdinger potential @xmath12 ( equation ( [ 1eq1.5 ] ) ) and its derivatives @xmath317 this can be confirmed as follows . from ( [ 1eqa.3]c ) of appendix
a and ( [ 1eq2.12 ] ) , we can show that @xmath114 satisfies the differential equation @xmath318+f(x)[1 - 2s_r(x , k)].\ ] ] substituting ( [ 1eq4.2 ] ) into ( [ 1eq10.2 ] ) , we obtain @xmath319 hence it follows that the @xmath127 s satisfy the recursion relation @xmath320 where we have used @xmath160 .
we can obtain @xmath321 from this recursion relation , starting with @xmath322 .
therefore , any @xmath127 ( @xmath323 ) can indeed be expressed in terms of @xmath12 and its derivatives . substituting ( [ 1eq10.1 ] ) into ( [ 1eq6.2 ] ) , we have @xmath324dz , \nonumber \\ \fl a_4=2v_{\rm
s}^2(x)-v_{\rm s}''(x)+2v_{\rm s}^2(y)-v_{\rm s}''(y ) , \quad { \rm etc}.\end{aligned}\ ] ] substituting ( [ 2eq10.5 ] ) into ( [ 1eq6.1 ] ) , and returning to definition ( [ 1eq2.7 ] ) , we can write @xmath325 -\frac{1}{(2ik)^3}\int_y^x\left[v_{\rm s}^2(z)-v_{\rm s}''(z)\right]dz+\cdots.\end{aligned}\ ] ] the recursion relation ( [ 1eq10.4 ] ) is familiar in soliton theory .
the quantities @xmath127 obtained from it are , apart from the sign , identical to the conserved charge densities for the kdv equation @xcite .
it is known that these quantities also appear in the asymptotic expansion for jost solutions @xcite .
our results for the expansion of the green function are valid even when jost solutions do not exist .
so far , we have been assuming that the schrdinger equation was derived from the fokker - planck equation ( [ 1eq1.3 ] ) .
let us now go in the opposite direction . to transform a given schrdinger equation into a fokker - planck equation , it is , in general , necessary to shift the energy level .
let @xmath326 be a solution of ( [ 1eq1.4 ] ) with @xmath327 .
then the schrdinger equation ( [ 1eq1.4 ] ) is equivalent to the fokker - planck equation @xmath328=p^2\phi(x),\ ] ] where @xmath329 the relation between @xmath12 and @xmath88 is now @xmath330 so ( [ 1eq1.5 ] ) is a special case of ( [ 1eq10.8 ] ) with @xmath331 . if we want @xmath3 to be real and finite for any finite @xmath14 , we need to choose @xmath332 such that @xmath333 for any finite @xmath14 . we may take @xmath332 to be the ground state energy if @xmath12 has a bound state .
equation ( [ 1eq10.5 ] ) was derived by using ( [ 1eq1.5 ] ) , which corresponds to @xmath331 . however , since the schrdinger equation ( [ 1eq1.4 ] ) is invariant under the replacements @xmath334 and @xmath335 , expansion ( [ 1eq10.5 ] ) is valid even when @xmath336 .
we can check this by an explicit calculation .
suppose that @xmath337 .
then it is obvious that a correct expression for @xmath338 is obtained by making the replacements @xmath334 and @xmath335 in ( [ 1eq10.5 ] ) as @xmath339 dz \nonumber \\ \fl & \qquad -\frac{1}{(2ip)^2}\left[v_{\rm s}(x)+v_{\rm s}(y)-2e_0\right ] + \cdots.\end{aligned}\ ] ] if we rearrange ( [ 1eq10.9 ] ) into an expansion in powers of @xmath110 by substituting @xmath340 and so on , then it becomes ( [ 1eq10.5 ] ) .
thus , ( [ 1eq10.5 ] ) holds for @xmath337 as well .
hence we know that ( [ 1eq6.1 ] ) , too , is valid for @xmath337 .
( see example 8 of the next section .
note that @xmath37 is not @xmath341 but @xmath342 . ) unlike the coefficients @xmath283 , the remainder term @xmath181 of ( [ 1eq6.1 ] ) is not expressed solely in terms of @xmath12 .
but the conditions for the validity of ( [ 1eq6.5 ] ) remain unchanged when @xmath337 .
we can understand this by writing ( [ 1eq10.9 ] ) with the remainder term , and then turning it into an expansion in powers of @xmath110 as above . is is not difficult to interpret the results of sections 7 and 8 in the language of the schrdinger equation .
we see from the second equation of ( [ 1eq10.7 ] ) that the conditions on @xmath88 can be interpreted as the conditions on @xmath343 . for example
, @xmath344 takes a nonzero finite value when the ground state wave function decays like @xmath345 as @xmath346 .
the differentiability conditions on @xmath88 can be directly related to the differentiability of @xmath12 .
namely , @xmath88 is @xmath145 times differentiable if @xmath12 is @xmath87 times differentiable .
let us consider some simple potentials for which the exact green function can be obtained , and compare the exact expression with our results for the expansion @xmath347 the coefficients @xmath283 are obtained from ( [ 1eq6.2 ] ) , ( [ 1eq4.3 ] ) and ( [ 1eq3.8 ] ) .
explicit forms of the quantities @xmath127 ( equation ( [ 1eq4.3 ] ) ) are shown in ( [ 1eq4.5 ] ) for @xmath348 .
( alternatively , we may use expressions ( [ 2eq10.5 ] ) in terms of @xmath12 . )
we omit the derivation of the exact expressions .
* example 1.*@xmath349 as the simplest example , let us first consider a linear potential . substituting @xmath350 into ( [ 1eq4.5 ] ) , and then into ( [ 1eq6.2 ] ) , we obtain the coefficients @xmath283 as @xmath351 the exact green function for this potential is @xmath352.\ ] ] it is easy to see that ( [ 1eq11.1 ] ) with ( [ 1eq11.2 ] ) is the correct expansion of the logarithm of ( [ 1eq11.3 ] ) for @xmath19 . since the exact @xmath353 does not have any singularities in @xmath354 , the right - hand side of ( [ 1eq11.1 ] ) is a convergent infinite series for @xmath354 .
the short - time expansion ( [ 1eq9.6 ] ) with @xmath355 calculated from ( [ 1eq9.7 ] ) and ( [ 1eq9.2 ] ) reads @xmath356 \left ( 1 - t + \frac{t^2}{2 } -\frac{t^3}{6 } + \cdots\right).\ ] ] the series in parentheses on the right - hand side is @xmath357 , so this expansion is convergent for any @xmath24 .
* example 2.*@xmath358 next , we consider a parabolic potential .
this example satisfies conditions ( i ) and ( ii ) of section 7 for any @xmath74 but not ( iii ) . from ( [ 1eq6.2 ] ) and ( [ 1eq4.5 ] ) ,
we obtain @xmath359 the exact green function for this potential can be expressed as @xmath360 where @xmath361.\end{aligned}\ ] ] here @xmath362 is the gamma function , and @xmath363 is the confluent hypergeometric function defined by @xmath364 . we can check that ( [ 1eq11.1 ] ) with ( [ 1eq11.6 ] ) is the correct asymptotic expansion of the logarithm of ( [ 1eq11.7 ] ) when @xmath97 is fixed in @xmath98 ( figure 1(a ) ) . now
( [ 1eq11.1 ] ) is divergent as an infinite series
. for the potential @xmath365 ( example 2 ) , with @xmath366 and @xmath367 , ( a ) plotted as functions of @xmath11 , with @xmath97 fixed at @xmath368 ; ( b ) plotted as functions of @xmath369 , with @xmath370 fixed at @xmath371 .
solid lines : the exact values . dashed lines : the expansion ( [ 1eq11.1 ] ) to order @xmath249 ( @xmath372 and @xmath373 ) . ]
this asymptotic expansion is also valid when @xmath101 is kept fixed as @xmath96 ( figure 1(b ) ) .
when @xmath104 , however , ( [ 1eq11.1 ] ) is not valid since @xmath374 oscillates and does not vanish as @xmath96 .
the short - time expansion ( [ 1eq9.6 ] ) now reads @xmath375 \left(1+g_1t+g_2 t^2 + g_3 t^3 + \cdots \right),\ ] ] where the @xmath355 s are obtained from ( [ 1eq9.7 ] ) , ( [ 1eq9.2 ] ) and ( [ 1eq11.6 ] ) as @xmath376 the exact time - dependent green function has the well - known form @xmath377}\right)^{1/2 } \exp \left ( -\frac{[x - y\exp(-2t)]^2}{1-\exp(-4 t ) } \right).\ ] ] the expansion of ( [ 1eq11.11 ] ) indeed has the form of ( [ 1eq11.9 ] ) with ( [ 1eq11.10 ] ) .
this @xmath378 has singularities in the complex @xmath24 plane where @xmath379 .
the nearest singularities to the origin are at @xmath380 .
so , the infinite series ( [ 1eq11.9 ] ) is convergent for @xmath381 .
this is a typical case where the short - time expansion is convergent for small @xmath24 although the high - energy expansion , from which it was derived , is divergent .
* example 3.*@xmath382 this exponential potential satisfies conditions ( i ) of section 7 for any @xmath74 but not ( ii ) or ( iii ) . for this potential
we have @xmath383 the exact green function has the form @xmath384 \chi_-(y , k),\ ] ] where @xmath385 are defined in terms of the bessel functions as @xmath386 , \qquad \nu_\pm \equiv \frac{1}{2}\pm ik.\ ] ] equation ( [ 1eq11.1 ] ) with ( [ 1eq11.12 ] ) gives the correct asymptotic expansion when @xmath97 is fixed in @xmath98 ( figure 2 ) . for the potential @xmath387 ( example 3 ) , ( a ) plotted as functions of @xmath11 , with @xmath97 fixed at @xmath368 ; @xmath388 , @xmath367 ; ( b )
plotted as functions of @xmath14 , with @xmath389 , @xmath367 .
solid lines : the exact values . dashed lines : the expansion ( [ 1eq11.1 ] ) to order @xmath249 ( @xmath372 and @xmath373 ) . ]
however , this asymptotic expansion is not correct when @xmath96 with fixed @xmath370 , irrespective of whether @xmath101 or @xmath104 .
the short - time expansion ( [ 1eq9.6 ] ) with @xmath355 calculated from ( [ 1eq11.12 ] ) is shown in figure 3 .
for the potential @xmath387 ( example 3 ) , plotted as a function of @xmath24 with various @xmath74 .
( a ) @xmath390 , @xmath367 ; ( b ) @xmath391 , @xmath367 . ]
* example 4.*@xmath392 this @xmath0 tends to @xmath393 linearly as @xmath394 . both @xmath136 and @xmath105
are finite , and conditions ( iii ) of section 7 are satisfied for any @xmath74 . in this case
, we have @xmath395 , \quad { \rm etc}.\end{aligned}\ ] ] the exact @xmath37 for this potential is @xmath396 } \frac{\eta_+(x ) \eta_-(y)}{\cosh x \cosh y},\end{aligned}\ ] ] where @xmath397 with @xmath398 and @xmath399 . here
@xmath363 is the hypergeometric function defined by @xmath400 . since @xmath311 are finite , expansion ( [ 1eq11.1 ] ) with ( [ 1eq11.15 ] ) is valid even when @xmath104 ( figure 4 )
. for the potential @xmath401 ( example 4 ) , ( a ) plotted as functions of real @xmath18 , with @xmath402 , @xmath367 ; ( b ) plotted as functions of @xmath14 , with @xmath403 , @xmath367 .
solid lines : the exact values . dashed lines : the expansion ( [ 1eq11.1 ] ) to order @xmath249 .
( since @xmath18 is real , @xmath404 are the same as @xmath405 , respectively , for the imaginary part . ) ] as in example 1 , the infinite series ( [ 1eq11.1 ] ) is convergent for @xmath354 .
* example 5.*@xmath406 this is an example where @xmath192 has a jump at @xmath195 .
this belongs to the case of ( [ 1eq7.1 ] ) with @xmath407 and @xmath408 . for @xmath237 ,
the exact green function is @xmath409,\ ] ] and hence we have , as @xmath96 ( @xmath137 ) , @xmath410 on the other hand , @xmath411 and @xmath412 are obtained from ( [ 2eq10.5 ] ) as [ 1eq11.25 ] @xmath413dz = x - y-2 , \qquad a_2=-1 + 2\delta(x)-1 + 2\delta(y)=-2.\ ] ] comparing ( [ 1eq11.25 ] ) with ( [ 1eq11.24 ] ) , we can see that @xmath411 is the correct coefficient of the expansion but @xmath412 is not . as shown in section 8 , the coefficient @xmath412 needs to be corrected by @xmath414 .
since @xmath415 , we can see that ( [ 2eq8.11 ] ) indeed agrees with ( [ 1eq11.24 ] ) .
* example 6 .
* @xmath416 this example belongs to the case of ( [ 1eq7.1 ] ) with @xmath417 and @xmath418 .
now @xmath192 is continuous and @xmath419 has a jump at @xmath195 . for @xmath237 ,
the exact @xmath37 has the form @xmath420 with @xmath421 , where @xmath422 and @xmath423 are defied by ( [ 1eq11.14 ] ) . as shown in figure 5(a ) , we can check that ( [ 1eq6.5 ] ) holds for @xmath424 but not for @xmath425 . for examples 6 and 7 , as functions of real @xmath18 .
( @xmath426 and @xmath427 . )
( a ) example 6 ; @xmath402 , @xmath428 .
( b ) example 7 ; @xmath402 , @xmath428 . ]
it can be seen that the limit of @xmath429 as @xmath96 is @xmath430 as predicted by ( [ 2eq8.10 ] ) .
the coefficients of the expansion for @xmath237 are obtained from ( [ 1eq6.2 ] ) as @xmath431 since both @xmath105 and @xmath136 are finite , expansion ( [ 1eq11.1 ] ) with ( [ 1eq11.28 ] ) is correct to order @xmath432 for @xmath19 . the correct expansion to order @xmath433 is obtained by adding @xmath434 to @xmath435 as in ( [ 2eq8.11 ] ) .
( see figure 6(a ) . ) for the potential of example 6 , plotted as functions of real @xmath18 ; ( a ) @xmath402 , @xmath436 ; ( b ) @xmath437 , @xmath438 . in ( a )
, the dashed curve labeled `` @xmath439 " shows the expansion to order @xmath433 with the correction term @xmath440 added to @xmath435 as in equation ( [ 2eq8.11 ] ) . in ( b ) , the envelope of the oscillation of @xmath441 falls off like @xmath442 . ] for @xmath238 , the exact green function is @xmath443 where @xmath444 the @xmath283 obtained from ( [ 1eq6.2 ] ) for @xmath238 are @xmath445 we can show that the quantity @xmath446 is @xmath447 as @xmath96 ( @xmath137 ) .
therefore , from ( [ 1eq11.29 ] ) we can see that ( [ 1eq11.1 ] ) with ( [ 1eq11.31 ] ) is correct only up to order @xmath110 when @xmath104 ( figure 6(b ) ) . for @xmath448 ,
this asymptotic expansion is correct to any order since @xmath449 in ( [ 1eq11.29 ] ) vanishes faster than any power of @xmath110 .
* example 7.*@xmath450 , \quad f(z)=-\frac{1}{4}\left(1+\vert z \vert\right)^{-1/2}$ ] , @xmath451 this is another case where @xmath419 has a finite jump at @xmath195 .
the exact green function for @xmath237 can be expressed in terms of confluent hypergeometric functions as @xmath452 where @xmath453 \nonumber \\ \fl & \
\pm \gamma\left(\textstyle{q+\frac{1}{2}}\right)\left [ f\left(\textstyle{q+\frac{1}{2},\frac{1}{2};-2ik(1\pm x)}\right ) \mp \frac{\sqrt{1\pm x}}{2 } f\left(\textstyle{q+\frac{1}{2},\frac{3}{2};-2ik(1\pm x)}\right ) \right ] \biggr\ } \nonumber \\ \fl\end{aligned}\ ] ] with @xmath454 , and @xmath455 . for @xmath237 ,
the coefficients of the expansion calculated by our method are @xmath456 -\frac{1}{4}\left[(1+x)^{-1/2}-(1-y)^{-1/2}\right ] , \nonumber \\
\fl a_2&=-\frac{1}{16}\left[(1+x)^{-1}+(1-y)^{-1}\right ] -\frac{1}{8}\left[(1+x)^{-3/2}-(1-y)^{-3/2}\right ] , \quad { \rm etc}.\end{aligned}\ ] ] this , too , is a case of ( [ 1eq7.1 ] ) with @xmath417 , and so ( [ 1eq6.5 ] ) holds for @xmath424 but not for @xmath425 .
expansion ( [ 1eq11.1 ] ) is now correct to order @xmath432 for @xmath19 ( figure 7 ) .
we can see from figure 5(b ) that @xmath429 indeed approaches the predicted value @xmath457 as @xmath96 .
( in this case , @xmath458 . ) for the potential of example 7 , plotted as functions of real @xmath18 ; @xmath402 , @xmath428 . ]
* example 8.*@xmath459 .
here we consider the case where @xmath12 is given , and where @xmath337 ( see section 10 ) .
for @xmath237 , the coefficients @xmath283 are obtained from ( [ 2eq10.5 ] ) ( or ( [ 1eq10.4 ] ) and ( [ 1eq6.2 ] ) ) as @xmath460 the exact green function for @xmath237 is @xmath461 where @xmath462 is the airy function , and @xmath463 is its derivative . from the second equation of ( [ 1eq10.7 ] ) we have @xmath464 , where @xmath332 is the smallest number satisfying @xmath465 .
( numerically , @xmath466 . )
it is easy to see that @xmath467 , so that this is a case of ( [ 1eq7.1 ] ) with @xmath468 , @xmath469 .
expansion ( [ 1eq11.1 ] ) with ( [ 2eq11.30 ] ) is correct to order @xmath470 for @xmath107 .
( conditions ( ii ) of section 7 are satisfied since @xmath192 behaves like @xmath471 as @xmath472 .
this expansion is not valid for @xmath104 . )
we can also check that the correct coefficient of order @xmath473 is not @xmath474 but @xmath475 .
in this paper , we studied the high - energy asymptotic behavior of the green function .
the expansion of @xmath111 in powers of @xmath110 ( with @xmath37 defined by ( [ 1eq2.7 ] ) and ( [ 1eq2.5 ] ) ) is given by ( [ 1eq6.1 ] ) .
the coefficients of the expansion ( equations ( [ 1eq6.2 ] ) and ( [ 1eq4.3 ] ) ) are expressed in terms of the coefficients @xmath85 for the expansion of the generalized reflection coefficient , which are calculated by using the formula ( [ 1eq3.8 ] ) with ( [ 1eq3.2 ] ) .
the remainder term @xmath181 is also expressed in terms of @xmath85 ( see ( [ 1eq6.3 ] ) , ( [ 1eq5.3 ] ) , ( [ 1eq4.12 ] ) , ( [ 1eq4.9 ] ) , ( [ 1eq4.6 ] ) and ( [ 1eq3.9 ] ) ) .
sufficient conditions for the validity of the expansion to order @xmath74 ( equation ( [ 1eq6.5 ] ) ) are given by ( i ) , ( ii ) , ( iii ) of section 7 .
these are not necessary conditions .
equation ( [ 1eq6.5 ] ) holds under broader conditions as shown in section 8 .
we assumed that the potential @xmath0 is monotone for sufficiently large @xmath476 , but this is not an essential restriction for our formalism .
the formulas for the coefficients of the expansion and the remainder term are valid for any potential , as long as these quantities make sense .
the particular shape of the potential is relevant only to the conditions for the validity of ( [ 1eq6.5 ] ) .
the above mentioned assumption on the potential is used only in deriving the conditions for ( [ 1eq3.12 ] ) quoted at the end of section 3 . even for other kinds of potentials
, we can use the same method to derive the criterion for ( [ 1eq6.5 ] ) .
for example , in this paper we excluded the cases where @xmath0 oscillates indefinitely as @xmath13 , but our method can be applied to these cases as well , if only we study the validity of ( [ 1eq3.12 ] ) ( and its analogue for @xmath58 ) for such potentials in a similar way as in @xcite .
here we summarize some properties of the transmission and reflection coefficients for finite intervals .
( for details , see @xcite and references therein . )
first , it is obvious that @xmath477 let us assume that @xmath88 is piecewise smooth .
we have , for finite @xmath14 and @xmath25 , [ 1eqa.2 ] @xmath478,\ ] ] @xmath479 as @xmath96 ( @xmath137 ) .
the first equation of ( [ 1eqa.2]b ) is a special case of ( [ 1eq3.3 ] ) with @xmath480 . for finite @xmath14 and @xmath25 , equations ( [ 1eqa.2 ] )
can be derived more directly from integral representations of the scattering coefficients .
( see equations ( 1.14 ) of @xcite .
alternatively , we can use equations ( 3.8 ) of @xcite and the asymptotic forms of the functions @xmath481 and @xmath482 to derive ( [ 1eqa.2 ] ) . ) from equations ( 3.5 ) and ( 3.8 ) of @xcite , we can derive the differential equations [ 1eqa.3 ] @xmath483 @xmath484 @xmath485,\ ] ] @xmath486 @xmath487 @xmath488.\ ] ] from ( [ 1eqa.2 ] ) and ( [ 1eqa.3 ] ) it follows that , as @xmath96 ( @xmath137 ) , [ 1eqa.4 ] @xmath489 ^ 2 = -2ik\left[\frac{\tau(x , z)}{1+r_r(x , z)}\right]^2 + 2f(z)e^{4ik(x - z ) } + o(1),\ ] ] @xmath490 ^ 2 = 2ik\left[\frac{\tau(z , x)}{1+r_l(z , x)}\right]^2 + 2f(z)e^{4ik(z - x ) } + o(1),\ ] ] [ 1eqa.5 ] @xmath491 ^ 2dz = \frac{1}{2ik}\left\ { \left[\frac{\tau(b , c)}{1+r_r(b , c)}\right]^2-\left[\frac{\tau(a , c)}{1+r_r(a , c)}\right]^2 \right\ } + o(1/\vert k \vert),\ ] ] @xmath492 ^ 2dz = \frac{1}{2ik}\left\ { -\left[\frac{\tau(c , b)}{1+r_l(c , b)}\right]^2+\left[\frac{\tau(c , a)}{1+r_l(c , a)}\right]^2 \right\ } + o(1/\vert k \vert).\ ] ]
let @xmath493 .
when ( [ 1eq9.1 ] ) is substituted into ( [ 1eq9.4 ] ) , the integral of the @xmath252th - order term is proportional to @xmath494 where we have changed the variable of integration from @xmath495 to @xmath496 $ ] , and deformed the contour of integration .
the right - hand side of ( [ 1eqb.1 ] ) can be expanded in powers of @xmath24 by using the formula of taylor expansion @xmath497 where @xmath498 is an arbitrary positive integer , and @xmath499 . applying ( [ 1eqb.2 ] ) to ( [ 1eqb.1 ] ) , and carrying out the integration of each term
, we have @xmath500^{2m+m+2 } } \biggr\}. \nonumber \\ \fl\end{aligned}\ ] ] from ( [ 1eq9.4 ] ) , ( [ 1eq9.1 ] ) and ( [ 1eqb.3 ] ) , we obtain the expansion ( [ 1eq9.6 ] ) with ( [ 1eq9.7 ] ) .
the remainder term @xmath308 can be written as @xmath501^{2n - m+2 } } , \nonumber \\
\fl\end{aligned}\ ] ] where @xmath502 . the second term of ( [ 1eqb.4 ] ) is obviously @xmath503 as @xmath504 . assuming that @xmath505 , the first term is @xmath506 as @xmath504 if @xmath507 as @xmath96 in the region @xmath106 .
10 risken h 1984 _ the fokker - planck equation _ ( berlin : springer ) newton r g 1966 _ scattering theory of waves and particles _
( new york : mcgraw - hill ) deift p and trubowitz e 1979 _ commun .
pure appl .
_ * 32 * 121 chadan k and sabatier p c 1989 _ inverse problems in quantum scattering theory _
( new york : springer ) verde m 1955 _ nuovo cimento _ * 2 * 1001 buslaev v and faddeev l 1960 _ sov .
* 1 * 451 faddeev l d and zakharov v e 1971 _ funct .
appl . _ * 5 * 280 calogero f and degasperis a 1968 _ j. math .
phys . _ * 9 * 90 corbella o d 1970 _ j. math .
phys . _ * 11 * 1695 harris b j 1986 _ proc .
edinburgh _ * 102a * 243 danielyan a a and levitan b m 1988 _ moscow univ .
bull . _ * 43 * 9 hinton d b , klaus m and shaw j k 1989 _ inverse problems _ * 5 * 1049 hinton d b , klaus m and shaw j k 1989 _ differential integral equations _ * 2 * 419 gesztesy f , holden h , simon b and zhao z 1995 _ rev . math .
phys . _ * 7 * 893 gesztesy f and simon b 2000 _ ann
_ * 152 * 593 rybkin a 2002 _ bull .
london math .
_ * 34 * 61 rybkin a 2002 _ proc .
* 130 * 59 miyazawa t 2006 _ j. phys . a : math .
_ 39 _ 10871 miyazawa t 2006 _ j. phys . a : math .
_ 39 _ 7015 miyazawa t 2006 _ j. phys . a : math .
_ 39 _ 15059 ( corrigendum ) novikov s , manakov s v , pitaevskii l p and zakharov v e 1984 _ theory of solitons _ ( new york : consultants bureau ) newell a c 1985 _ solitons in mathematics and physics _ ( philadelphia : siam ) miyazawa t 1998 _ j. math .
* 39 * 2035 | a new formalism is presented for high - energy analysis of the green function for fokker - planck and schrdinger equations in one dimension .
formulas for the asymptotic expansion in powers of the inverse wave number are derived , and conditions for the validity of the expansion are studied through the analysis of the remainder term .
this method is applicable to a large class of potentials , including the cases where the potential @xmath0 is infinite as @xmath1 .
the short - time expansion of the green function is also discussed . |
various phenomenological models developed over the years have raised the possibility that nuggets of strange quark matter ( sqm ) , called strangelets , are present in cosmic rays .
such strangelets will have nearly equal numbers of up , down and strange quarks and so will have an anomalous @xmath0 ratio ( @xmath1 ) @xcite , where @xmath2 is the charge and @xmath3 is the baryon number , compared to ordinary nuclei .
one particular model for strangelet propagation through the earth s atmosphere has strongly hinted at the possible presence of low energy ( few mev
/ n ) strangelets at high mountain altitudes @xcite , although with a very low flux .
also , over the years , various experimental groups have reported the presence of particles with anomalous @xmath0 ratios in cosmic rays @xcite . but none of those groups have been able to make any definitive claim because of lack of statistics .
so the search for strangelets in cosmic rays remains an active area of research @xcite .
one of the best ways to look for low energy strangelets with very low fluxes at high mountain altitudes is through the deployment of very large area arrays of nuclear track detectors ( ntds ) .
such passive detector arrays offer several advantages over many other detector types .
they are relatively inexpensive to deploy , easy to maintain and do not require any power for their operation , a fact which offers particular advantages when it comes to the deployment of large arrays at very remote high altitude locations . also because of their high intrinsic thresholds of registration , some ntds offer a natural way to suppress the huge low-@xmath2 background ( neutron recoil tracks , atmospheric radon alpha tracks etc . )
expected in any such experiment .
nuclear track detectors ( ntds ) like cr-39 , makrofol etc . have been used for charged particle detection for many years @xcite .
we plan to use a particular brand of commercially available polymer , identified as polyethylene terephthalate ( pet ) @xcite , as ntd in the search for exotic particles in cosmic rays , through the deployment of large area arrays of pet films at high mountain altitudes .
pet was found @xcite to have a much higher detection threshold ( @xmath4 , where @xmath2 is the charge and @xmath5 the measure of the velocity of the impinging particle ) compared to other commercially available ntds like cr-39 , makrofol etc .
( @xmath6 lying in the range 6 - 60 ) .
this makes pet particularly suitable for low energy rare event search in cosmic rays .
before any new material can be employed as a detector , it needs to be properly characterized and calibrated . with that aim ,
systematic studies were carried out on pet to determine its ideal etching condition and also to ascertain its charge response to various ions using accelerators as well as natural radioactive sources . a calibration curve for pet ( @xmath7 vs. @xmath8 , where @xmath7 is the specific energy loss , @xmath9 and @xmath10 are the track and the bulk etch rates respectively
while their ratio @xmath8 is the reduced etch rate or charge response ) utilizing @xmath11o , @xmath12s , @xmath13fe , @xmath14u ions was obtained @xcite .
it was then updated @xcite with additional data points corresponding to @xmath15xe , @xmath16kr and @xmath17ti ions obtained from the rex - isolde facility at cern .
also , studies were carried out to determine the charge and energy resolution that could be achieved with pet @xcite .
all these studies have firmly established pet as a very efficient detector of heavily ionizing particles with a detection threshold much higher than the other commercially available detector material cr-39 which is in widespread use today . in addition to such calibration experiments ,
pilot studies were carried out at different high altitude locations where pet , as well as standard cr-39 detectors , were given open air exposures for durations ranging from a few months to two years .
the goal was to study how the detector behaviour changes with exposure to harsh environmental conditions and also to survey the local radiation background .
the results of such studies are presented in this paper .
the sites chosen for these studies are darjeeling in eastern himalayas , hanle in northern himalayas and ooty in the nilgiri hills .
table [ table : parameters ] lists the altitudes ( above mean sea level ) and some other parameters associated with those sites .
one reason for the choice of these sites is the existence of scientific research facilities there , which is going to make the eventual deployment and maintenance of any such large area array easier .
also these facilities have records of climatic conditions at those sites going back many years and they continue to collect and maintain such records .
stacks containing three pet films of a4 size ( @xmath18 ) and thickness @xmath19 as well as cr-39 films of size @xmath20 and thickness @xmath21 were mounted on perspex stands and given open air exposures for durations ranging from three months to two years at those sites .
fig [ detectors ] shows pet films as well as smaller pieces of cr-39 mounted on perspex stands .
the stands in turn are fitted inside a box for convenience of transport . a4 size pet films as well as smaller pieces of cr-39 mounted on perspex stands . ]
.[table : parameters]some parameters of the sites where ntds were given open air exposures [ cols="^,^,^,^,^ " , ] fig
. [ angle ] , fig .
[ diameter ] , fig .
[ tracklength ] , fig .
[ vtvb ] give distribution of the angle of incidence , minor axis diameter , track length and @xmath8 values for the tracks recorded on cr-39 at darjeeling , ooty and hanle .
distribution of the angle of incidence of tracks recorded on cr-39 at darjeeling , ooty and hanle . ]
distribution of the minor axis diameter of tracks recorded on cr-39 at darjeeling , ooty and hanle ( samples etched for 4 h ) . ] distribution of the length of tracks recorded on cr-39 at darjeeling , ooty and hanle ( samples etched for 4 h ) . ]
distribution of the of the @xmath8 values of tracks recorded on cr-39 at darjeeling , ooty and hanle . ]
from the studies conducted so far , high threshold pet seem to be a very good choice as a detector material for the planned rare event search .
also hanle appears to be the place most suitable as a site for the setup of the large area array .
the site is basically a cold desert with robust existing scientific research facilities for astronomical research
. we will continue to conduct further studies along these lines as well as begin the process of large scale search for strangelets using large area ntd arrays .
the authors sincerely thank staff members of iia at hanle and of tifr at ooty for their help in setting up the detectors at those places .
the authors also thank mr .
sujit k. basu and mr .
deokumar rai for technical assistance .
the work is funded by irhpa ( intensification of research in high priority areas ) project ( ir / s2/pf-01/2011 dated 26.06.2012 ) of the science and engineering research council ( serc ) , dst , government of india , new delhi .
99 j. madsen , phys .
d 71 ( 2005 ) 014026 .
s. biswas , j.n . de , p. s. joarder , s. raha , d. syam , phys
b 715 ( 2012 ) 30 .
sayan biswas , j.n . de , partha s. joarder , sibaji raha , debapriyo syam , proc.indian natl.sci.acad . 81 ( 2015 ) 1 , 277 .
s. banerjee , s.k .
ghosh , s. raha , d. syam , phys .
85 ( 2000 ) 1384 .
t. saito et .
al . , phys .
lett . 85 ( 2000 ) 1384 .
basudhara basu , sibaji raha , swapan k. saha , sukumar biswas , sandhya dey , atanu maulik , amal mazumdar , satyajit saha , debapriyo syam , astropart.phys .
61 ( 2015 ) 88 . o. adriani et .
al . , phys .
( 2015 ) 111101 .
fleischer , p.b .
price , r.m .
walker , nuclear tracks in solids , university of california press , 1975 .
durrani , r.k .
bull , solid state nuclear track detection principles , methods and applications , pergamon press , 1987 .
b. basu , s. dey , b. fischer , a. maulik , a. mazumdar , s. raha , s. saha , s. k. saha , d. syam , radiat .
meas . 43 ( 2008 ) s95 .
s. banerjee , s.k .
ghosh , s. raha , d. syam , phys .
85 ( 2000 ) 1384 .
b. basu , a. mazumdar , s. raha , s. saha , s. k. saha , d. syam , ind . j. phys . 79 ( 2005 ) 279 .
r. bhattacharyya , s. dey , sanjay k. ghosh .
a. maulik , sibaji raha , d. syam , nucl .
instr . and meth .
b 370 ( 2016 ) 63 .
d. bhowmik , s. dey , a. maulik , s. raha , s. saha , s. k. saha , d. syam , nucl .
instr . and meth .
b 269 ( 2011 ) 197 .
s. dey , d. gupta , a. maulik , s. raha , s. k. saha , d. syam , j. pakarinen , d. voulot , f. wenander , astropart .
34 ( 2011 ) 805 .
s. dey , a. maulik , sibaji raha , swapan k. saha , d. syam , nucl .
instr . and meth .
b 336 ( 2014 ) 163 . j.f .
ziegler , j.p .
biersack , the stopping and range of ions in matter ( srim computer code ) , version : 2003.26 , 2003 . | various phenomenological models presented over the years have hinted at the possible presence of strangelets , which are nuggets of strange quark matter ( sqm ) , in cosmic rays .
one way to search for such rare events is through the deployment of large area nuclear track detector ( ntd ) arrays at high mountain altitudes . before the deployment of any such array can begin , a detailed study of the radiation background is essential .
also a proper understanding of the response of detectors exposed to extreme weather conditions is necessary . with that aim ,
pilot studies were carried out at various high altitude locations in india such as darjeeling ( 2200 m a.m.s.l ) , ooty ( 2200 m a.m.s.l ) and hanle ( 4500 m a.m.s.l ) .
small arrays of cr-39 as well as high threshold polyethylene terephthalate ( pet ) detectors were given open air exposures for periods ranging from three months to two years .
the findings of such studies are reported in this paper . |
the dynamics of suspended particles whose size is larger than the smallest active scales of the carrier turbulent flow is important for the understanding of many natural and industrial applications . let us mention two concrete problems borrowed from the study of planet formation and of rain processes . in the first situation ,
the role of turbulence is still unsettled in the dynamical mechanisms that take place during the origination of planets in a proto - planetary disk ( see , e.g. , * ? ? ?
a key issue is to understand whether or not the disk turbulence can generate an increase of the rate at which a large body can accrete dust .
in other terms , it requires identifying the turbulent fluctuation mechanisms that lead to a snowball effect where the larger bodies have a growth speed all the more fast as their size increases .
the same kind of question arises in the precipitation process when a large raindrop falls under the effect of gravity and collides with much smaller droplets that are transported by the turbulent airflow in the cloud ( see * ? ? ?
* ) . for this problem
, another question is to know the efficiency of wet deposition ( or scavenging ) of particle pollutants that are present in the atmosphere ( see , e.g. , * ? ? ?
these various phenomena are still poorly understood and are subject to empirical ad - hoc modeling .
such fundamental open questions , which are taken from natural sciences but also arise in a number of industrial settings , reflect an inadequate current level of modeling for large particles in turbulent flows .
current approaches for the dynamics of finite - size particles are commonly assuming that the particles are infinitely small ( point particles ) and that the reynolds number associated to their interaction with the flow vanishes , so that their dynamics can be obtained by matching the solution to the stokes equation in their vicinity to an unperturbed fluid velocity .
recent experimental work showed that the dynamics of finite - size particles is not well described by such models ( see * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) .
significant size effects have been for instance measured in the acceleration statistics of such particles . despite the fact that the flatness of the acceleration distribution depends only weakly on the diameter of the particle , its variance strongly decreases with diameter and temporal correlations of the acceleration vary with the particle size in a manner that can not be predicted by standard point - particle models .
improved models accounting for the so - called `` faxn corrections '' ( spatial averages of fluid velocities and accelerations over the volume of the particle ) are in much better agreement with such qualitative trends but hardly reproduce quantitatively the mentioned experimental data ( see * ? ? ?
one of the most challenging requirement for improving such models is to get a precise understanding on how the carrier flow is modified in the vicinity of the particle in situations where such effects occurs on scales which are comparable to the active scales of the surrounding turbulence .
for instance , it is clear that much work is still needed in order to get a better handling on how the kinetic turbulent energy , the dissipation rate , as well as other standard statistical properties of turbulence , varies in the turbulent boundary layer enclosing the particle .
direct numerical simulations appear to be valuable candidates as they give a unique global and instantaneous access to such quantities .
however , the numerical problem is challenging for several reasons : _
( i ) _ the numerical approach has to resolve all active scales of a fully developed turbulent flow , which span several decades ; _ ( ii ) _ at small scales the boundary layer around the particle is of great importance and has to be well resolved ; _ ( iii ) _ if one considers a moving particle the algorithm has to follow the displacement of the particle accurately and efficiently ; _ ( iv ) _ for a sufficient separation of the different scales involved ( forcing scales , inertial range of scales , dissipative scales ) the algorithm has to parallelize efficiently on massively parallel computers . two different strategies have been recently proposed in order to solve the incompressible navier
stokes equations together with no - slip boundary conditions at the surface of a spherical particle . both make use of a homogeneous grid covering of the complete physical domain .
the first,``physalis '' , which was developed by @xcite and extended and used by @xcite , makes use of known analytic solutions to the stokes equation for the boundary layer in the vicinity of the particle , in order to describe the velocity at the nearest surface points of the grid .
iteratively , the analytic solution and the solution on the homogeneous grid are matched .
an advantage of this strategy is that the no - slip condition is imposed in an exact manner .
however , improvement are required in order to achieve very - large reynolds number carrier flows ( the value @xmath1 of the taylor - microscale reynolds number was chosen in * ? ? ?
the other recently developed approach consists in using and matching two different grids .
a spherical one encloses the particle surface and is designed to resolve the boundary layer and a second homogeneous grid covers the entire domain .
this `` overset grid '' technique proposed by @xcite uses a third - order interpolation to transfer values from one grid to the other in order to couple the two solutions .
this method was shown to reproduce very accurately experimental measurements of the drag coefficient at a particle reynolds number of @xmath2 .
in turbulent settings , this method was used with a reynolds number @xmath3 .
both of these methods reside on finite - difference schemes and are at the moment limited to rather small carrier - flow reynolds numbers .
we propose here a new method that allows one to reach much higher reynolds numbers .
the idea is to combine a standard pseudo - fourier - spectral with a penalty method .
the former is well adapted to incompressible homogeneous and isotropic turbulence and has been extensively used in such settings during the past .
its advantages comprise its degree of accuracy and its performance on massive parallel supercomputers such as the bluegene / p system or large intel / amd clusters . in combination with an immersed boundary and penalization strategy , high reynolds numbers turbulent flow with an extended inertial range of scales can be reached .
the proposed method is a modification of that used by @xcite for investigating the dynamics of neutrally buoyant particles suspended in turbulent flows .
the former version of the method was well - adapted to particles displaying a small slip velocity with respect to the fluid but happened to show some limitations when the slip is too large .
we present here a modification of this method that unravels such shortcomings .
this paper is organized as follows : in section [ sec : method ] we present in details the proposed method , together with some considerations concerning its convergence .
section [ sec : benchmarks ] presents measurements of drag and lift coefficients in various settings and compare them to other numerical and experimental works .
we consider the cases of a stationary sphere in a flow at a constant speed , of a freely rotating sphere and a translating sphere in a flow maintained at rest .
section [ sec : wake ] deals with the characterization of the turbulence decay in the particle wake .
section [ sec : conclusion ] encompasses concluding remarks and open questions .
the dynamics of an incompressible flow in which is embedded a particle is given by the navier - stokes equations @xmath4 together with the no - slip boundary condition on the surface of the particle @xmath5 @xmath6 denotes here external forces that are exerted on the flow , @xmath7 is the full physical domain containing both the fluid and the particle , @xmath8 is the boundary of the physical location @xmath9 of the particle at time @xmath10 , and @xmath11 denotes the velocity of the particle surface .
note that in the case of a solid spherical particle with diameter @xmath12 , one has @xmath13 and @xmath14 \times \bm\omega_p(t)$ ] , where @xmath15 and @xmath16 are the translational and angular velocities of the particle , respectively .
the particle motion is then obtained by solving newton equations where the force exerted by the fluid is obtained by integrating on the surface of the particle the contribution from pressure and viscous stress to the fluid stress tensor . the numerical method presented here
consists in integrating the navier stokes equation ( [ eq : navier - stokes ] ) by using a strongly - stable , low - storage runge - kutta scheme of third order that was introduced by @xcite .
it reads @xmath17 @xmath18 denotes here the time step and @xmath19 and @xmath20 are the velocity fields at time steps @xmath21 and @xmath22 and the operator @xmath23 denotes the right - hand side ( rhs ) of ( [ eq : navier - stokes ] ) .
the later is integrated on a homogeneous rectangular periodic spatial grid with a fourier pseudo - spectral solver using the standard 2/3 rule to avoid aliasing due to its quadratic nonlinearity ( see , e.g. , * ? ? ?
the pressure term is obtained by solving the poisson equation @xmath24 directly in fourier space .
the viscous term is treated implicitly via an exponential factor . to impose the no - slip boundary conditions ( [ eq : bc ] ) at the surface of the particle , we make use of an immersed boundary technique .
it consists in introducing in the right - hand side of the navier - stokes equation ( [ eq : navier - stokes ] ) a penalty force @xmath25 , which acts as a lagrange multiplier associated to the constraint defined by the boundary condition ( [ eq : bc ] ) .
the full problem ( [ eq : navier - stokes])-([eq : bc ] ) can then be rewritten as @xmath26 this system then has to be associated with the problem @xmath27 one can easily see that , when the solution is approximated by the galerkin truncation associated with the fourier spectral method , these two conditions can not be satisfied in an exact manner : an non - zero function that vanishes on a full sub - domain is necessarily non - analytic and can not be represented by a finite number of fourier modes .
the idea of the method which is used here is to relax the problem of finding the penalty force @xmath28 to @xmath29 all the difficulty of integrating this system is now in finding an efficient manner of solving the optimization problem above , which involves local condition in the physical space , while spectral methods try to be as local as possible in fourier space .
as this has to be done at every sub - time steps , the use of standard minimization techniques would introduce a very large number of operations and result in a cost - inefficient algorithm .
efficient strategies consist in guessing acceptable values of the field @xmath28 .
we here make use of the `` direct - forcing '' method proposed by @xcite . to explain this method ,
let us consider a simple euler - discretization in time of ( [ eq : ib ] ) @xmath30.\ ] ] the idea of the direct - forcing method is to use the fact that we want to impose @xmath31 in @xmath32 to write from ( [ eq : euler ] ) @xmath33,\ ] ] where @xmath34 denotes the characteristic function of the particle : @xmath35 if @xmath36 and 0 otherwise .
other similar approaches consist in imposing the boundary conditions by means of a darcy term , i.e. @xmath37 ( see * ? ? ?
the particle volume is then seen as a porous media where a large value of the parameter @xmath38 ensures the no - slip boundary conditions but results in choosing a sufficiently small value of the time step .
the advantage of direct forcing is that it consists in combining the value of the parameter @xmath38 that is optimal for a given choice of the time step @xmath39 , together with a compensation of the overall evolution of the velocity field inside the particle due to @xmath23
. the expression ( [ eq : directforcing ] ) for @xmath40 would be correct if the position of the grid points coincided with the immersed boundary .
this is not true in the case of a spherical particle embedded in a cubic grid .
an interpolation procedure is then needed in order to prescribe the boundary conditions at the grid - points next to the surface . for this
we use a volume - fraction scheme proposed by @xcite and @xcite .
the characteristic function @xmath41 describing the geometry of the particle is averaged over two grid cell : @xmath42 this representation has two advantages .
first , as a typical velocity field around a particle is continuous but not differentiable at the surface , gibbs oscillations are created in the vicinity of the boundary ; replacing the characteristic function by @xmath43 reduces such oscillations .
second , this volume averaged @xmath43 allows for a continuous displacement of the immersed boundary which is important in the case of a moving particle .
another issue arises when combining a spectral with a penalty method .
although applying the force @xmath40 to the velocity field imposes the boundary conditions precisely , the divergence - freeness of @xmath44 requires a subsequent projection of the velocity field onto its solenoidal part which violates the formally correct boundary conditions .
a solution to this conflict is to solve the poisson equation ( [ eq : pressure - poisson ] ) for pressure with modified boundary conditions as for example proposed by @xcite in combination with a finite - difference scheme .
however , this is not possible with a fourier - spectral method as in this case ( [ eq : pressure - poisson ] ) is solved with generic periodic boundary conditions .
we attenuate this intrinsic problem of a fourier - spectral approach by using a predictor for the pressure gradient in ( [ eq : ib ] ) , as considered by @xcite .
we will now explain our procedure .
the usual strategy in spectral codes without immersed boundaries consists in advancing first the velocity without taking into account the pressure gradient in ( [ eq : navier - stokes ] ) @xmath45.\ ] ] the intermediate velocity field @xmath46 is compressible and is projected in a second step onto its solenoidal part @xmath47 this projection is equivalent to subtracting the gradient of the pressure that is given by the poisson equation ( [ eq : pressure - poisson ] ) .
this splitting of the integration is called fractional - step method .
if the domain contains immersed boundaries with no - slip conditions , the penalty force @xmath28 has to be additionally applied .
this can be seen as an intermediate step @xmath48 leading to a compressible @xmath49 , which is again projected onto its solenoidal part as in ( [ eq : discproj ] ) .
a problem arises since the two operators @xmath50 and @xmath51 do not commute .
the application of @xmath50 spoils the result of @xmath51 .
our approach consists in attenuating this conflict by adding a predictor for the pressure gradient to ( [ eq : discnav1 ] ) .
we make of @xmath52 , i.e. of the value of the pressure gradient at the last time step .
this predictor already includes the major part of the pressure needed to keep the velocity field very close to being divergence - free .
the predictor is updated during the projection step according to @xmath53 it is important to stress that we apply the operators p and b and also the update of the pressure predictor during each sub - time step of the third - order temporal scheme ( [ eq : rk3 ] ) . the importance of the predictor for a fourier - spectral scheme is evident from fig .
[ fig : predictor_comp ] where the streamwise velocity profile on a line passing through the particle is shown .
the flow conditions correspond to a numerical wind tunnel experiment where a fixed spherical particle ( @xmath54 ) is embedded in a homogeneous flow .
the velocity fluctuations introduced by the wake of the particle are eliminated in a zone at the right boundary by an additional application of the penalty method .
the accuracy of the upstream boundary condition at the particle surface is strongly affected by the pressure predictor . with the predictor
we achieve a no - slip condition at the surface while the fluid significantly enters the particle when we neglect the pressure predictor .
also the smoothing of the velocity fluctuation at the exit of the wind tunnel is much more efficient with the pressure predictor .
( left ) profile of the velocity ( normalized to the inflow speed ) in the direction of the mean flow on the axis of symmetry .
the solid line corresponds to a computation with predictor @xmath55 while the points correspond to a computation without the predictor .
( right ) @xmath0 norm of the velocity inside the particle as a function of the number of grid points on its diameter , title="fig:",scaledwidth=51.0% ] ( left ) profile of the velocity ( normalized to the inflow speed ) in the direction of the mean flow on the axis of symmetry .
the solid line corresponds to a computation with predictor @xmath55 while the points correspond to a computation without the predictor .
( right ) @xmath0 norm of the velocity inside the particle as a function of the number of grid points on its diameter , title="fig:",scaledwidth=51.0% ] a measure for the accuracy of the boundary conditions is given by the @xmath0-norm @xmath56 of the velocity inside the fixed spherical particle .
this norm is represented in fig .
[ fig : predictor_comp ] ( right ) for a flow with a particle reynolds number @xmath2 and as a function of the number of grid - points @xmath57 along the diameter of the sphere .
this number quantifies the resolution of the geometry of the particle by the underlying homogeneous grid .
the norm @xmath58 shows a power - law decrease with an exponent @xmath59 .
this value can be understood by remembering that a function @xmath60 in a fourier scheme is represented by a @xmath61 , a truncated series of @xmath62 modes . using a function @xmath63 which resembles qualitatively the flow structure in the vicinity of a particle at rest
one can show that @xmath64 .
the force that is exerted by the fluid on the particle is given by the integral over the surface of the particle of the fluid stress tensor @xmath65 this total force @xmath66 can easily be computed from the penalty force @xmath28 because @xmath67 in order to distinguish the pressure and viscous terms appearing in the stress tensor ( [ eq : forceparticle ] ) , our strategy consists in computing the pressure contribution by constructing a homogeneous grid of points at the sphere surface .
then , we integrate the directed pressure @xmath68 , @xmath69 being the normal vector field at the surface of the particle , over this grid . the viscous part is obtained by subtracting the pressure contribution from the total force .
now we are going to present several types of benchmarks of the proposed scheme . for this
we turn back to the numerical wind tunnel experiment already mentioned before .
the principle flow configuration is depicted in fig .
[ fig : re400_9000 ] above the critical reynolds number where a non - stationary wake develops .
turbulent wake at re=400 ( volume rendering of high - vorticity isosurfaces).,scaledwidth=70.0% ] an important quantity of the dynamics of a solid sphere exposed to a homogeneous inflow is the drag - coefficient @xmath70 where @xmath71 denotes the total force exerted on the particle in the stream - wise direction , @xmath72 the inflow velocity , @xmath12 is the sphere diameter , and @xmath73 the fluid mass density ( set to one in our numerical settings ) .
the measured coefficients are given in fig .
[ fig : re_drag_coeff ] in comparison to the empirical formula @xmath74 proposed by @xcite which fits well experimentally measured coefficients . for all reynolds numbers considered ( from @xmath75 to @xmath76 )
the dns coefficients are in good agreement with the fitting formula .
( left ) drag - coefficients @xmath77 as a function of the particle reynolds number ; in parenthesis is given the number of internal grid points .
inset : ratio between the viscous contribution @xmath78 to the drag coefficient and the contribution @xmath79 from pressure .
( right ) convergence rate as a function of the number of grid points @xmath57 inside the sphere.,title="fig:",scaledwidth=51.0% ] ( left ) drag - coefficients @xmath77 as a function of the particle reynolds number ; in parenthesis is given the number of internal grid points .
inset : ratio between the viscous contribution @xmath78 to the drag coefficient and the contribution @xmath79 from pressure .
( right ) convergence rate as a function of the number of grid points @xmath57 inside the sphere.,title="fig:",scaledwidth=51.0% ] the numbers in parenthesis in fig .
[ fig : re_drag_coeff ] ( left ) denote the number of grid - points @xmath57 on the diameter of the sphere . the higher this number the better is the resolution of the boundary layer around the sphere .
the thickness of this layer decreases as @xmath80 thus the higher the reynolds number the more points @xmath57 are required in order to resolve the boundary layer . in this experiment
the sphere is fixed and the inflow velocity is one .
the reynolds number is varied by changing the viscosity of the fluid . in order to check to performance of the scheme in case where the particle is moving we changed the frame of reference of this experiment to the particle frame and thus move the particle in a flow at rest .
also the zone where we remove the velocity fluctuations is moved according to the particle position .
the measured drag - coefficients and the velocity profile around the sphere are nearly indistinguishable from the former wind tunnel experiments .
we observe fluctuating deviations from the values given in fig .
[ fig : re_drag_coeff ] of less than one percent for reynolds number up to 50 and less than five percent up to 400 . the inset of fig .
[ fig : re_drag_coeff ] ( left ) shows the ratio of the viscous @xmath78 to the pressure @xmath79 contribution to the complete drag @xmath81 .
the measured ratios are in reasonable agreement with results of @xcite . here
we stress that the pressure contribution is not needed for the integration of the particle dynamics and might only be of interest for post - processing purposes . the convergence of the drag - coefficient can be quantified by looking at the relative error @xmath82 as a function of the internal points @xmath57 . from fig . [
fig : re_drag_coeff ] ( right ) we observe that the coefficients converge with an exponent of -3/2 for the two reynolds numbers considered .
this is the same rate of convergence as for the @xmath0 norm @xmath58 of the velocity inside the particle ( see previous section ) which suggests that the convergence - rate of the proposed scheme is determined by the convergence - rate of the truncated representation of the @xmath83-velocity field . if the inflow velocity of the wind tunnel is not homogeneous but possesses a constant gradient @xmath84 this linear shear flow creates a lift force @xmath85 on the particle perpendicular to the mean flow .
the rate of shear is usually given in terms of a non - dimensional shear rate @xmath86 where @xmath87 is the dimensional shear in ( [ eq : velprofil ] ) .
recently , the corresponding lift - coefficients @xmath88 where @xmath89 represents the amplitude of the lift force , have been measured by three different groups ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
( left ) lift - coefficients @xmath90 as a function of the particle reynolds number for @xmath91 .
( right ) stationary rotation rate ( normalized to the ambient rotation rate @xmath92 ) as a function of the particle reynolds number defined with the mean fluid velocity gradient @xmath93 .
inset : logarithmic representation . for the exponential fit we used @xmath94 and @xmath95.,title="fig:",scaledwidth=51.0% ] (
left ) lift - coefficients @xmath90 as a function of the particle reynolds number for @xmath91 .
( right ) stationary rotation rate ( normalized to the ambient rotation rate @xmath92 ) as a function of the particle reynolds number defined with the mean fluid velocity gradient @xmath93 .
inset : logarithmic representation . for the exponential fit we used @xmath94 and @xmath95.,title="fig:",scaledwidth=51.0% ] from fig [ fig : lift_coeff_sphere ] ( left )
one observes a quite large spread of the lift coefficients measured by the three groups .
one reason for these differing values is the sensitivity of the lift force on the stresses in the flow and thus on the size of the computational domain .
it has been claimed by @xcite that the box size in the work of @xcite was too small , leading to an overestimation of the lift force .
we have computed the lift force for two different box sizes and observe a decreasing force with increasing box - size .
our measurements are in good agreement with the two recent works and reproduce in particular the transition to negative values of the lift force for @xmath96 . as a last benchmark we present in fig .
[ fig : lift_coeff_sphere ] ( right ) the torque - free rotation rate in a linear shear flow described in the precedent section .
the sphere is fixed but allowed to rotate freely .
after a relaxation period we observe a rotation rate which is exponentially ( see inset of fig . [
fig : lift_coeff_sphere ] ( right ) ) decreasing with the reynolds number .
the measured values agree well with those of @xcite .
the small deviations might again be a result of different box - sizes used .
this section is dedicated to the study of turbulent velocity statistics in the wake of a spherical obstacle .
the development of a fourier - spectral code that integrates the fluid motion and prescribes no - slip boundary conditions on the body surface allows one to achieve precise direct numerical simulations of turbulent wakes at a reasonable computational time .
most models for turbulent flows past an obstacle rely on studying solutions to the reynolds - averaged equations ( see , e.g. , * ? ? ?
this approach consists in writing the velocity field @xmath97 as a sum of its stationary average @xmath98 and turbulent fluctuations @xmath99 .
the mean velocity is then a solution to @xmath100 where @xmath101 designates the reynolds stress tensor @xmath102 . in order to obtain the properties of the mean velocity profile @xmath103 , eq .
( [ eq : rans ] ) is generally closed by means of the eddy viscosity assumption .
this hypotheses consists in relating linearly the anisotropic part of the reynolds stress to the mean rate , that is @xmath104 , \label{eq : eddyviscosity}\ ] ] where @xmath105 is the turbulent kinetic energy and the positive quantity @xmath106 is the eddy viscosity .
when the eddy viscosity @xmath107 is assumed to be constant in space , and in particular independent of the streamwise distance @xmath108 from the particle center , the reynolds - averaged equation ( [ eq : rans ] ) can be integrated to obtain the following self - similar estimate for the streamwise mean velocity @xmath109 ( see , e.g. , * ? ? ?
* ) @xmath110,\ ] ] where the wake half - width @xmath111 is given by @xmath112 such formula , which are very commonly used in industrial applications , have two important consequences .
first , the mean velocity deficit @xmath113 should behave at large distances in the turbulent wake as @xmath114 .
secondly , the normalized velocity deficit profile @xmath115/[u_c - u(x,0,0)]$ ] should collapse to the similar profile @xmath116 where @xmath117 and @xmath118 .
an analysis of the turbulent wake from a direct simulation performed at the moderate value of the particle reynolds number @xmath119 shows that this similar eddy - viscosity theory is able to rightly predict the decay of the mean velocity deficit .
figure [ fig : u_decay ] ( left ) represents the behavior of @xmath120 as a function of @xmath121 .
its displays a power - law decay with a exponent @xmath122 over nearly half a decade . the value @xmath122 is undoubtedly obtained as seen from the inset of fig .
[ fig : u_decay ] , which shows the value of the local slope ( logarithmic derivative ) of the velocity deficit as a function of the distance in the wake .
the oscillations around the value @xmath122 are a signature of the gibbs phenomenon induced by the strip forcing at @xmath123 .
( left ) mean velocity deficit @xmath124 in the axis of symmetry @xmath125 and @xmath126 of the wake as a function of the normalized streamwise distance @xmath127 .
as seen from the inset , which represents its logarithmic derivative @xmath128/[\mathrm{d}\ln ( x / d-1)]$ ] , it displays a power - law behavior @xmath129 with @xmath130 .
( right ) mean velocity deficit profile @xmath131 as a function of the reduced spanwise distance @xmath132 for various values of @xmath127 as labeled ; the bold dashed line represents the gaussian functional form of @xmath131 obtained from a constant - eddy - viscosity similar profile.,title="fig:",scaledwidth=52.0% ] ( left ) mean velocity deficit @xmath124 in the axis of symmetry @xmath125 and @xmath126 of the wake as a function of the normalized streamwise distance @xmath127 .
as seen from the inset , which represents its logarithmic derivative @xmath128/[\mathrm{d}\ln ( x / d-1)]$ ] , it displays a power - law behavior @xmath129 with @xmath130 .
( right ) mean velocity deficit profile @xmath131 as a function of the reduced spanwise distance @xmath132 for various values of @xmath127 as labeled ; the bold dashed line represents the gaussian functional form of @xmath131 obtained from a constant - eddy - viscosity similar profile.,title="fig:",scaledwidth=52.0% ] however , as seen from the right - hand panel of fig .
[ fig : u_decay ] , the mean velocity deficit profile does not seem to display any similarity .
the half width @xmath111 was estimated here as the value of the distance to the axisymmetric axis of the wake such that the velocity deficit is exactly equal to @xmath133 . by definition
, the various curves should go through @xmath134 at @xmath135 and @xmath136 at @xmath137 . despite these two constraints
, it is seen that they do not collapse on the top of each other for values of @xmath138 close to @xmath133 and @xmath139 .
the discrepancy is even stronger for @xmath140 , where @xmath131 tends to negative values that depend on @xmath108 .
these deviation at large @xmath138 s might be due to the periodicity of the spatial domain in the @xmath141 and @xmath142 directions that results in a mean streamwise velocity of the fluid that is larger than @xmath72 at @xmath143 and @xmath144 .
however , the deviation from the similar profile at moderate values of @xmath138 are comparable to those obtained experimentally by @xcite and @xcite . in particular ,
a striking observation is that in all cases with moderately large reynolds number , the mean velocity defect profile seems to vanish for @xmath145 .
such a feature is clearly reproduced in our simulation , as seen from fig .
[ fig : u_decay ] ( right ) where the various curves are pinched together before reaching negative values .
we have seen above that the self - similar wake approach reproduces well the decay @xmath146 of the mean velocity deficit . however , this approach can not be fully true as no similarity is observed in the wake . in order to get a better handling , we relate in this section this observed behavior to the decay of turbulent kinetic energy in the wake of the spherical particle .
it is known since the seminal work of @xcite that the law of decay of turbulent kinetic energy is dictated by the `` principle of permanence of large eddies '' .
we follow here the phenomenological presentation of this law made by @xcite .
if we assume that the initial turbulent velocity correlations follow a infrared asymptotic self - similarity of the form @xmath147 at large scales @xmath148 , the decay behavior of all turbulent quantities is dictated by the value of the exponent @xmath149 .
indeed , by continuity of velocity statistics , the integral - scale velocity should then itself behave as @xmath150 , where @xmath151 is the turbulent kinetic energy and @xmath152 is the integral scale .
however , we know that the energy decays with a rate that can be related to velocity by @xmath153 . replacing @xmath151 by @xmath154 in this equation , we obtain that @xmath155 , so that @xmath156 .
this evolution of the integral scale implies that the kinetic turbulent energy behave as @xmath157 . when considering the decay of wake turbulence ,
all the turbulent quantities have to be of course evaluated in planes perpendicular to the axisymmetric axis and time has to be replaced by @xmath158 by invoking taylor hypothesis .
the above considerations hence imply that if the velocity correlations @xmath159 behave as @xmath160 in a plane @xmath161 located in the developing wake , then at any @xmath162 , the large - scale correlations should remember this initial form and the turbulent kinetic energy is expected to decay as @xmath163 .
figure [ fig : correl_ur ] represents the correlation of the radial component of the fluctuating velocity at various values of @xmath108 .
correlation of the velocity radial fluctuation as a function of the distance @xmath148 to the axisymmetric axis for various values of the streamwise distance @xmath108 as labeled .
inset : same in semi - logarithmic coordinates showing that the various correlation fall on the top of each other and follow an exponential decay.,scaledwidth=60.0% ] as seen from this figure , velocity correlations collapse at large scales on the top of each other onto an exponential behavior @xmath164 that is independent on @xmath108 .
such a fast decay of the correlations is associated in the above phenomenological formalism to a value of the exponent @xmath165 ( an exponential is faster than any power law ) , leading to a decay of the turbulent kinetic energy @xmath166 .
this behavior is confirmed numerically , as can be seen from fig .
[ fig : wake_kin_en ] .
profile of the mean turbulent kinetic energy @xmath151 in the wake of a spherical particle with reynolds number @xmath119 .
the top figure represents the turbulent kinetic energy @xmath151 in the @xmath167 plane ( high values are shown in red and small values in blue ) .
the bottom figure shows the behavior of @xmath151 along the @xmath168 axis , clearly displaying a behavior @xmath169,scaledwidth=60.0% ] finally , to relate such a behavior of the turbulence decay to the way the velocity deficit decreases as a function of the distance downstream of the particle , one has to turn back to the reynolds - averaged equation ( [ eq : rans ] ) .
the mean velocity deficit @xmath170 approximatively obeys the equation @xmath171 @xmath101 denotes here the reynolds tensor introduced in previous subsections .
this shows that the scaling of the velocity deficit is consistent with that of the decay of turbulent kinetic energy .
the work presented in this article aims at validating a new method for integrating the incompressible navier stokes equation with no - slip boundary conditions on the surface of a moving particle .
simulations that were performed with this method in the case of a spherical particle embedded in a homogeneous flow , show that such a strategy reproduces with a high accuracy known results on the variation of the drag and lift forces as a function of the particle reynolds number and give a precise handling on the turbulent fluctuations that arise in the wake downstream the sphere .
such benchmarking will now allow us confidently attacking the physical problems that motivated the development of such a code .
the first question that is subject to ongoing work concerns the dynamics of neutrally buoyant finite - size particles that are transported by a developed turbulent flow in the case when the diameter of such particles falls inside the turbulent inertial range of lengthscales .
the drag , lift , slip velocity associated to such settings are still not well defined because of the fluctuating nature of the fluid surrounding environment .
this situation still requires an important modeling effort and a first step consists in understanding how the fluid velocity statistical properties are modified by the presence of the particle .
preliminary results confirm that a turbulent boundary layer develops around the particle and that its thickness is comparable to the particle size .
an important open question that is worth mentioning here relates to possible improvements of the order of the method that is presented here .
as we have seen , the error of the proposed method is of the order of @xmath172 . to go beyond this first order ,
a natural idea is to make use of the variational formulation that is described in sec .
[ sec : method ] and which consists in expressing the penalty force as a solution to a minimization problem . to avoid gibbs oscillations that , as we showed , are responsible for the observed error , it could be envisaged to relax such minimality to the surface instead of the full particle volume , allowing for instance the fluid flow to develop inside the particle , without affecting the no - slip boundary condition .
such ideas lead to new difficulties related to the non - locality of pressure and which still need to be cautiously worked out .
this work benefited from fruitful discussions with j. dreher .
part of this research was supported by the french agence nationale de la recherche under grant no .
blan07 - 1_192604 .
the research leading to these results has received funding from the european research council under the european community s seventh framework program ( fp7/2007 - 2013 , grant agreement no .
access to the ibm bluegene / p computer jugene at the fz jlich was made available through project hbo22 .
part of the computations were performed on the `` msocentre de calcul sigamm '' and using hpc resources from genci - idris ( grant 2009-i2010026174 ) .
calzavarini , e. , volk , r. , bourgoin , m. , leveque , e. , pinton , j. , f.toschi , 2009 .
acceleration statistics of finite - sized particles in turbulent flow : the role of faxen forces . j. fluid mech . 630 , 179189 .
fadlun , e. a. , verzicco , r. , orlandi , p. , mohd , j. , 2000 . combined immersed - boundary finite - difference methods for three - dimensional complex flow simulations .
journal of computational physics 60 , 3560 .
qureshi , n. m. , bourgoin , m. , baudet , c. , cartellier , a. , gagne , y. , 2007 .
turbulent transport of material particles : an experimental study of finite size effects . physical review letters 99 , 184502184506 . | a novel scheme is presented for the direct numerical simulation of two - way coupling between a particle and an incompressible fluid flow maintained in a high - reynolds - number turbulent regime .
the main idea consists in combining a fourier pseudo - spectral method for the fluid with an immersed - boundary technique to impose the no - slip boundary condition on the surface of the particle .
this scheme is shown to converge as the power 3/2 of the spatial resolution .
this behavior is explained by the @xmath0 convergence of the fourier representation of a velocity field displaying discontinuities of its derivative .
benchmarking of the code is performed by measuring the drag and lift coefficients and the torque - free rotation rate of a spherical particle in various configurations of the carrier flow .
such studies show a good agreement with experimental and numerical measurements from other groups , and validate the code .
a study of the turbulent wake downstream the sphere is also reported .
the mean velocity deficit is shown to behave as the inverse of the distance from the particle , as predicted from classical similarity analysis .
this law is reinterpreted in terms of the principle of `` permanence of large eddies '' that relates infrared asymptotic self - similarity to the law of decay of energy in homogeneous turbulence .
fourier - spectral methods , penalty methods , flow - particle interaction |
in 1958 , while working as an assistant to professor n. obreshkov , blagovest sendov raised the following question . _
assume that all zeros of a polynomial @xmath0 with complex coefficients lie in the closed unit disk .
is it true that there exists a critical point in every disk of radius one centered at a zero of @xmath0 ? _ the conjecture was included in the 1967 book of research problems collected and edited by w. hayman @xcite .
the conjecture was wrongly attributed to ilieff , who in fact only had communicated the problem to the world of mathematicians .
sendov s conjecture naturally fits into the line of classical works on the geometry of polynomials and at the same time it has close connections with potential theory and operator theory , allowing non - trivial matrix theory reformulations . in spite of many ingenious ideas originating in attempts to solve sendov s conjecture ,
only partial results are known today , see for instance the respective chapters in the monographs @xcite .
it was the young julius borcea who pushed the positive solution to this problem up to degree seven @xcite , and it was the mature julius borcea who outlined a series of extensions and modifications of sendov s conjecture in the context of statistics of point charges in the plane .
the authors of the present note had the privilege to work closely with borcea on this topic , individually , and on two occasions in full formation ( generously supported by the mathematics institute at the banff center in canada , and by the american institute of mathematics in palo alto ) . without aiming at completeness ,
the following pages offer a glimpse at borcea s beautiful new ideas . to understand the present status of sendov s conjecture we outline first the main contributions scattered through its half century history .
in the late 1960s soon after the publication of hayman s book , a series of papers by several authors proved sendov s conjecture for polynomials of degree three , four and five in quick succession .
sendov s conjecture for polynomials of degree six turned out to be significantly more difficult ; its solution in julius borcea s first research paper appeared more than a quarter century later in 1996 @xcite . in his next paper published in the same year
, he proved that the conjecture is true for polynomials of degree seven @xcite .
it is now known that sendov s conjecture holds for polynomials with at most eight distinct zeros @xcite . in the twenty - five years between the proofs for polynomials of degree five and those of degree six , some other special cases of sendov s conjecture of
were solved ; these tended to be classes of polynomials defined by conditions on the geometry of their zeros rather than on their degree .
for instance , sendov s conjecture is true in the following cases : - all zeros are real @xcite , - for all trinomials @xcite , - all zeros belong to the unit circle @xcite , - all zeros are contained in the closed unit disk , and at least one lies at the center of the disk @xcite , - if the convex hull of the zeros form a triangle @xcite .
+ the contents are the following .
we introduce borcea s family of conjectures in the next section .
we verify the @xmath1 version of borcea s conjecture for polynomials with at most three distinct roots in section three and the even stronger @xmath2 version of borcea s conjecture for cubic polynomials in section four .
the connection between these family of conjectures and the theory of univalent functions is explored in section five .
the next three sections detail various operator theoretical approaches to these conjectures and other results in the analytic theory of polynomials . using the logarithmic derivative , it is easy to see that the zeros of the derivative of a polynomial can be viewed as the zeros of the cauchy transform of a sum of point measures at the zeros of the polynomial . in section nine ,
we generalize borcea s conjectures to the zeros of cauchy transform of weighted point measures .
our concluding section is somewhat more speculative ; we look at some very interesting ideas in the mathematical literature which could prove useful for tackling these questions .
since we are interested in the geometry of zeros of a polynomial , it is natural to introduce some quantitative estimates of their relative position in the plane .
we denote by @xmath3 the zero set of a polynomial @xmath4 $ ] .
let @xmath5 be a polynomial and let @xmath6 .
define the @xmath7-variance of the zero set of @xmath0 by @xmath8 .
certain values of @xmath9 will be of particular importance to us : @xmath10 is the _ variance _ and @xmath11 is called the _
mean deviation_. we define @xmath12 as : @xmath13 and we identify it as the _ chebyshev radius _ of the set @xmath14 ; the optimal value of @xmath15 is known as the _ chebyshev center _ of the set @xmath14 , see for instance @xcite or ample references in the banach space literature @xcite . for every polynomial @xmath0 and @xmath16
we have @xmath17 let @xmath15 be an optimal center for @xmath18 : @xmath19 in virtue of hlder s inequality : @xmath20^{p / q } n^{1-p / q},\ ] ] hence @xmath21^{1/p } \leq [ \frac{1}{n } \sum_k |z_k - c|^q]^{1/q } = \sigma_q(f)\ ] ] and the first inequality follows .
the second inequality is obtained similarly , starting with an optimal center @xmath15 with respect to @xmath22 .
recall that in the case of variance @xmath23 the optimal center is the _ barycenter _ of the zeros of @xmath0 @xmath24 indeed , the reader can verify with basic inner space techniques that @xmath25 note also that in the above definitions we allow multiple roots , so that @xmath26 really depends on the polynomial @xmath0 rather than its zero set @xmath14 , although in the literature one also finds the notation @xmath27 .
only @xmath28 is intrinsic to @xmath14 .
we define @xmath29 to be the unsymmetrized _ hausdorff distance _ between the zero sets @xmath14 and @xmath30 : @xmath31 while @xmath32 is the symmetrized hausdorff distance between the the two sets .
alternatively , we will use the more suggestive notation @xmath33 and all related variations . a series of conjectures , derived by borcea from the ilieff - sendov original problem ,
are stated below .
[ main ] let @xmath0 be a polynomial of degree at least two and let @xmath34
. then @xmath35 .
when @xmath36 , this is sendov s original conjecture . in view of the above lemma
, the validity of borcea s conjecture for a specific value of @xmath7 implies that the inequality holds for all @xmath37 , including @xmath38 .
we will see shortly that the conjecture is false for @xmath39 .
the problem regarding the extremal configurations for sendov s conjecture was raised by phelps and rodrigues @xcite as follows .
[ sec ] if @xmath40 , then equality in the above conjecture occurs if and only if @xmath0 is of the form @xmath41 for some complex numbers @xmath42 whenever @xmath43 .
* example 1 .
* let @xmath44 and @xmath45 .
then @xmath46 , so that @xmath47 and @xmath48 we derive from here that @xmath49 and @xmath50^{1/p } = \frac{1}{n^{1/p } } \min_{c \in [ 0,1 ] } [ c^p + ( n-1)(1-c)^p]^{1/p}.\ ] ] if @xmath2 , then @xmath51 } [ n-1 -c(n-2 ) ] = \frac{1}{n}$ ] , therefore conjecture [ sec ] does not hold for @xmath2 , while @xmath52 for all @xmath53 .
assume that @xmath39 .
the only critical point in the interval @xmath54 $ ] of the function @xmath55 is @xmath56 , with @xmath57 increasing on the interval @xmath58 $ ] and decreasing on @xmath59 $ ] .
thus @xmath60 showing that conjecture [ main ] can not be extended to the case @xmath39 .
+ another question which may be of interest is whether asymptotic versions of conjecture [ main ] can be proven . for instance : [ asymp ] for all @xmath61 , there exists @xmath62 with @xmath63 such that @xmath64 provided @xmath65 .
the case @xmath36 was validated in @xcite by the value @xmath66 . to a polynomial @xmath67 with derivative @xmath68 we associate the _ discriminant _ : @xmath69 and the _ resultant _ : @xmath70 the fundamental relation between the two is : @xmath71 the proof of this identity follows from the observation that @xmath72 when all points are in the closed unit disk , the maximum value of the discriminant is attained for @xmath73 , for which we get @xmath74 ( equivalently , equally spaced points are fekete points for the disk ) .
thus we immediately obtain that if all the zeros @xmath75 of the polynomial @xmath0 are in the unit disk , then the zeros @xmath76 of @xmath77 must satisfy @xmath78 thus we infer that is further than @xmath12 from @xmath14 can not be in the convex hull of @xmath14 .
see also proposition 4.1 . ]
that sendov s conjecture is true for at least one zero of @xmath0 and a little more : there exists a zero of @xmath0 with the property that the product of all distances to the critical points of @xmath0 is less than or equal to @xmath79 .
the rest of the present article focuses on conjecture [ main ] in the two most important cases @xmath2 and @xmath1 .
in order to prove the variance conjecture for polynomials with at most three distinct zeros and for polynomials with all zeros real we need the following definition .
let @xmath80 where all the @xmath81 are distinct .
let @xmath82 be the zeros of @xmath83 that are not zeros of @xmath0 , counted with multiplicities .
we define the gauss - lucas matrix to be the @xmath84 matrix @xmath85 with entries @xmath86 we note that the gauss - lucas matrix is a stochastic matrix which maps the vector of zeros of the original polynomial into the vector of zeros of its derivative : specifically , @xmath85 has only non - negative entries , the sum of all elements of any row of @xmath85 is equal to @xmath79 , and @xmath87 ( where @xmath88 and @xmath89 are column vectors ) .
indeed , one starts with the identity @xmath90 which yields @xmath91 in particular we notice that every @xmath92 is a convex combination of the @xmath93 .
apparently it was cesro who first wrote the above system of equations relating the zeros of a polynomial to its critical points @xcite .
thus we have proved the following classical result .
the critical points of a polynomial lie in the convex hull of its zeros .
let @xmath80 where all @xmath81 are distinct .
we define the augmented gauss - lucas matrix @xmath94 of @xmath0 to be the @xmath95 matrix whose first @xmath96 rows consist of the gauss - lucas matrix @xmath85 and whose last row is defined as follows @xmath97 .
since every row or column vector of a unitary matrix has euclidean norm equal to one , we have the following well known result , see for instance @xcite .
let @xmath98 $ ] be any @xmath99 matrix and define @xmath100 to be the @xmath99 non - negative matrix whose @xmath101-th entry is @xmath102 .
if @xmath103 is a unitary matrix , then @xmath104 is doubly stochastic , i.e. all entries of @xmath104 are non - negative and that the sums along every column or row equal @xmath79 .
we are now able to prove the key technical result in the case of three distinct zeros .
the case of two distinct zeros can be reduced via a rotation to the case of real zeros , to be discussed in full below .
[ main2 ] any polynomial @xmath0 with at most three distinct zeros has an augmented gauss - lucas matrix which is doubly stochastic .
let @xmath105 , @xmath106 and @xmath107 be the zeros of the polynomial @xmath0 with multiplicites @xmath108,@xmath109 and @xmath110 respectively .
let @xmath111 ( note that we take the transpose for @xmath112 ; but we take the conjugate transpose for @xmath113 ) and @xmath114 .
these three vectors are mutually orthogonal and therefore @xmath115 is an orthonormal basis of @xmath116 where @xmath117 .
let @xmath103 be the unitary matrix whose rows are @xmath115 .
then the augmented gauss - lucas matrix is @xmath118 which is doubly stochastic .
we note that the fourth degree polynomial @xmath119 has an augmented gauss - lucas matrix which is not doubly stochastic so the above lemma does not generalize to higher degree polynomials .
however , we are more fortunate in the case of purely real zeros .
any polynomial @xmath0 all of whose zeros are real has an augmented gauss - lucas matrix which is doubly stochastic .
let @xmath120 be the zeros of the polynomial @xmath0 , @xmath121 be the multiplicity of @xmath122 and let @xmath123 be the zeros of @xmath83 which are nt also zeros of @xmath0 .
let @xmath124 when @xmath125 and let @xmath126 .
then @xmath127 is an orthogonal set .
now , the same argument as in the proof of lemma 3.2 yields the result .
remark that the vectors @xmath128 have the following operator theoretic interpretation . for simplicity
we assume no multiple zeros .
let @xmath129 .
let @xmath130 be the projection on to the orthogonal complement of the vector @xmath131 .
let @xmath132 . according to @xcite , the eigenvalues of @xmath133 are + @xmath134 .
it can be seen that @xmath135 is an eigenvector of @xmath133 corresponding to the eigenvalue @xmath76 and @xmath136 is an eigenvector of @xmath133 corresponding to the eigenvalue @xmath137 . if @xmath94 is hermitian , then so is @xmath133 , therefore the vectors @xmath138 are orthogonal . if @xmath94 is not a linear function of a hermitian matrix , then @xmath133 is not normal by a theorem of fan and pall @xcite ( this is also proved directly in @xcite ) .
we will return to this point of view in a subsequent section .
[ 3main ] any polynomial @xmath0 with at most three distinct zeros satisfies borcea s 2-variance conjecture .
let @xmath122 be a zero of @xmath0 .
we prove that there is a zero of @xmath83 which lies in the disk centered at @xmath139 of radius @xmath10 .
if @xmath122 is a multiple zero then we are done , so let us assume that @xmath122 has multiplicity one
. then by lemma [ main2 ] , @xmath140 , hence there is at least one @xmath141 such that @xmath142 .
let @xmath143 .
then @xmath144 thus , by cauchy - schwarz inequality and in view of our choice of the index @xmath145 , we obtain that : @xmath146 retaining the extreme terms in the inequality we obtain @xmath147 . remark that equality would require that no entry in the @xmath148th column of the gauss - lucas matrix be strictly larger than @xmath149 which would force @xmath150 to be equal to three and all of the @xmath105 , @xmath106 and @xmath107 to be equidistant from @xmath92 . by examining the coefficient of the polynomial @xmath151
one can show that @xmath152 for some complex numbers @xmath153 .
furthermore , as we saw , the above lemma implies that borcea s variance conjecture holds for polynomials with real zeros .
we note that we do not need the augmented gauss - lucas matrix to be doubly stochastic for the above proof to work ; any @xmath150th degree polynomial which has a gauss - lucas matrix in which every column has an element greater than or equal to @xmath149 would satisfy borcea s 2-variance conjecture . unfortunately
, even this weaker conjecture is not true in general as it fails for the 19th degree polynomial found in @xcite .
this polynomial ( after translation ) is @xmath154 .
the two columns of its gauss - lucas matrix corresponding to the conjugate pair @xmath155 of zeros of @xmath0 have all entries strictly less than @xmath156 .
it would be interesting to understand what other conditions besides stochasticity gauss - lucas matrix must satisfy .
for instance , what can be said about its column sums ?
what about the size of the maximal elements in each column ?
we begin with a brief review of _ apolarity _ , a key concept in the geometry of polynomials , see for instance @xcite .
let @xmath157 and @xmath158 .
then @xmath159 and @xmath160 are said to be apolar provided that @xmath161 . if @xmath0 and @xmath85 are apolar polynomials then any closed circular domain which contains all zeros of @xmath85 contains at least one zero of @xmath0 .
there are many equivalent forms of grace s theorem @xcite proved by among others szeg and walsh ( see section 3.4 of @xcite for details ) . at the time of writing ,
the most recent new proof of grace s theorem is by borcea and brndn and can be found in @xcite .
grace s theorem and apolarity are used to prove many results in the analytic theory of polynomials , including many of the known special cases of sendov s conjecture .
a striking example of an application of apolarity to this conjecture is borcea s necessary and sufficient conditions for sendov s conjecture in @xcite we first note that we can prove a mean deviation result for the reverse direction of the unsymmetrized hausdorff distance .
let @xmath0 be a polynomial of degree at least two
. then @xmath162 .
let @xmath163 , where the @xmath75 are not required to be distinct .
let @xmath164 be a zero of @xmath165 that is not also a zero of @xmath0 .
now pick @xmath15 such that @xmath166 , since @xmath167 , we have @xmath168 hence we can show for every @xmath169 that @xmath170 would follow from @xmath171 . the mean deviation conjecture ( @xmath172 has been verified for cubic polynomials by j. borcea . in the remainder of this section
we discuss his proof @xcite .
first we define the _ circular deviation_. let @xmath173 for @xmath5 , set @xmath174 the sets @xmath175 for @xmath176 can be described in simple geometric terms . specifically , denote @xmath177 and @xmath178 ; we say that @xmath179 and @xmath180 are equivalent if there exists @xmath181 with @xmath182 and a permutation @xmath183 such that @xmath184 .
then @xmath185 is the set of all vectors equivalent to @xmath186 ; @xmath187 is the set of all vectors equivalent to @xmath188 where @xmath189 is a primitive cube root of unity ; @xmath190 is the set of all vectors equivalent to @xmath191 where @xmath15 is any complex number of modulus one .
it can easily be shown that @xmath192 .
the inequality can be however strict .
for instance , in the case of @xmath193 we have @xmath194 and @xmath195 .
note that if @xmath0 is a cubic polynomial and @xmath137 is one of its zeros , one may construct a @xmath196 circulant matrix @xmath197 whose characteristic polynomial is @xmath0 .
for the theory of circulant matrices we refer to the monograph @xcite .
then any two by two principal submatrix of @xmath198 has characteristic polynomial @xmath199 . by remarking that the circulant matrix can be diagonalized by the fourier matrix @xmath200 one
proves that @xmath201 .
then one verifies that @xmath202 is apolar to @xmath203 .
since the disk centred at zero of radius @xmath204 contains both zeros of @xmath85 , it must contain at least one of the zeros of @xmath83 .
a classical counterpart to grace theorem was discovered in 1915 by j. w. alexander @xcite and refined in 1917 by kakeya @xcite ( see also @xcite ) , as follows .
[ alexander - kakeya ] if a polynomial @xmath0 of degree @xmath150 has no critical points in a closed disk of radius @xmath205 , then it is univalent in the concentric disk of radius @xmath206 .
the extremal polynomials for the alexander - kakeya theorem ( the ones for which @xmath206 can not be improved ) are the same ones as in conjecture [ sec ] : @xmath207 . we are naturally led now to formulate the following conjecture . if @xmath0 is an @xmath150th degree polynomial with @xmath208 , then @xmath0 is not univalent in any closed disk of radius larger than @xmath209 centered at zero .
of course , if the conjecture is true for @xmath2 , then it follows for all @xmath53 . by the alexander - kakeya theorem
, the truth of this conjecture implies the truth of conjecture [ main ] for the same value of @xmath7 .
it can be easily seen that the strongest case @xmath2 of this conjecture is true for @xmath210 . in this case @xmath211 and @xmath212 .
since @xmath0 has a critical point at @xmath213 it is clearly not univalent in any disk larger than @xmath214 and so our conjecture holds .
_ non__univalence conjecture opens the possibility to use the huge body of known facts about the univalence of polynomials and analytic functions .
traditionally one normalizes the functions by the conditions @xmath208 and @xmath215 . in our polynomial case
we simply consider @xmath216 .
if such a polynomial is univalent in the open unit disk , then it is well - known , for example , that @xmath217 . for a survey of results on univalent polynomials ,
see @xcite .
we present below a different approach to borcea s 2-variance conjecture derived from some recent matrix theory advances of pereira @xcite and malamud @xcite .
we return to a monic polynomial @xmath218 and the variance of its zero set @xmath219 where we repeat the enumeration of multiple zeros .
@xmath220 where @xmath221 is the barycenter : @xmath222 if now @xmath223 and @xmath224 is the sequence of critical points @xmath225 , then recall that the 2-variance conjecture [ main ] states that @xmath226 where @xmath227 is the symmetrized hausdorff distance .
the main idea is to associate with the polynomial ( [ v-1 ] ) a normal operator @xmath94 acting in @xmath228 so that @xmath94 is diagonal with entries @xmath229 along the diagonal . in this case
@xmath230 , where @xmath231 is the characteristic polynomial of @xmath94 . according to the terminology introduced in @xcite , a unit vector @xmath232
is called a _ differentiator _ if @xmath233 where @xmath234 and @xmath130 is the orthogonal projection onto the orthogonal complement @xmath235 in @xmath228 ( so that @xmath133 is acting in @xmath235 ) .
a simple computation shows that any vector @xmath236 satisfying @xmath237 is a differentiator , see for the proof the above section devoted to gauss - lucas matrices .
moreover , there is an orthonormal basis of differentiators @xmath238 , where @xmath239 and @xmath240 we shall return to this basis later . in the above notations ,
if @xmath241 is a differentiator and @xmath242 , then @xmath243 is a rank one operator @xmath244 with the norm @xmath245 which gives a natural operator - theoretical interpretation of the variance .
the conjectured inequality @xmath246 reduces to two independent assertions : + * statement ( i ) * : _ for any @xmath247 , the disk @xmath248 ( centered at @xmath249 and having radius @xmath250 ) contains at least one point from @xmath224_. + * statement ( ii ) * : _ for any @xmath251 , the disk @xmath252 contains at least one point from @xmath14 . _ + we discuss next a matrix theory proof of statement ( ii ) , already settled in the previous section by a different method .
the basic observation goes back to a paper by ptak @xcite . indeed , if we assume that @xmath253 is in @xmath224 and @xmath254 is such that @xmath255 , then @xmath256 which implies @xmath257 and hence @xmath94 must have at least one eigenvalue @xmath139 with @xmath258 since @xmath94 is a normal operator .
the main difficulty in proving the 2-variance conjecture consists in proving statement ( i ) .
a standard reduction to the case where @xmath259 transforms this question to the problem of proving the invertibility of a normal matrix under certain spectral conditions on its principal submatrices .
namely , let @xmath238 be an orthonormal basis of differentiators and @xmath260 denote the orthogonal projections to the hyperplanes @xmath261 , and @xmath262 .
the proof of statement ( i ) then reduces to the following statement .
[ matrix ] let @xmath94 be a normal matrix and let @xmath238 be an orthonormal basis
. denote by @xmath263 the orthogonal projection onto @xmath264 .
if all eigenvalues of the compressions @xmath265 lie outside the unit disk and @xmath266 for all @xmath267 , then @xmath94 is invertible .
this question is similar to known invertibility criteria for diagonally dominant matrices ; this time we deal instead with codimention one principal submatrices rather than diagonal entries .
the principal difficulty of this problem is that the matrices @xmath268 are not normal and the challenge is to transform their known spectral properties into appropriate metric properties . in two special cases
this can be done .
namely , we will give an alternate proof of statement ( i ) ( and hence the variance conjecture ) in the case @xmath269 .
another special case is that of polynomials with real zeros . in this case ,
a much more stronger estimate than statement ( i ) is valid ( see proposition [ v - p-1 ] below ) . a critical step towards the proof is contained in the next statement .
[ v - l1 ] if @xmath133 is an operator on @xmath270 such that all eigenvalues @xmath164 of @xmath271 satisfy @xmath272 , then for any vector @xmath273 @xmath274 in an appropriate orthonormal basis @xmath133 has the triangular form @xmath275 where @xmath276 and @xmath277 are eigenvalues of @xmath133 .
the lemma follows now from the coordinate - wise computation of the left - hand side of the desired inequality .
a similar inequality is no longer true in dimensions @xmath278 , even with a constant in front of @xmath279 on the right hand side .
it is enough to consider @xmath280 and @xmath281 with a sufficiently large real @xmath282 . in view of the discussion
prior to conjecture [ matrix ] , the 2-variance conjecture in the case @xmath269 is a consequence of the next result .
[ v - l2 ] let @xmath94 be a normal operator in @xmath283 and let @xmath284 be an orthonormal basis .
for @xmath285 , let @xmath263 be an orthogonal projection onto @xmath286 and @xmath287 .
assume that @xmath288 and all operators @xmath265 acting in @xmath286 have their spectra outside the unit disk .
then @xmath94 is invertible .
let @xmath289 be small enough , so that @xmath290 let @xmath291 denote @xmath283 and let @xmath292 denote @xmath286 for @xmath285 .
the inequality @xmath293 @xmath294 shows that the row operator @xmath295 considered as acting from @xmath296 into @xmath291 satisfies the estimate @xmath297 .
the property that @xmath94 is a normal operator implies @xmath298 hence another row operator @xmath299 ( again acting from @xmath296 into @xmath291 ) satisfies the estimate @xmath300 this shows that @xmath301 is bounded away from zero and hence @xmath302 is surjective .
therefore , any @xmath303 has the form @xmath304 with appropriate @xmath305 which implies that @xmath94 is invertible since @xmath94 is a normal operator . in the case of polynomials whose all zeros are real
, a much stronger result , recently proved by borcea , is valid .
@xcite [ v - p-1 ] if all zeros @xmath229 of polynomial ( [ v-1 ] ) are real , then for any @xmath306 the disk @xmath307 contains at least one zero of the derivative @xmath77 . as before
, we may assume that @xmath308 and that @xmath309 .
we keep also notations from preceding proof . assuming that all operators @xmath310 do not have eigenvalues in the disk @xmath311
, we shall see that @xmath94 is invertible which gives a contradiction .
since all @xmath229 are real , @xmath94 is self - adjoint and hence all @xmath310 are also self - adjoint .
therefore , @xmath312 for any @xmath313 which implies that @xmath314 hence the row operator @xmath315 considered as acting from @xmath316 into @xmath291 satisfies the estimate @xmath317 for any @xmath303 . on the other hand ,
the row operator @xmath318 satisfies the estimate @xmath319 .
hence @xmath320 is bounded away from zero which shows that @xmath321 is surjective and hence @xmath94 is invertible .
the constant @xmath10 in statement ( ii ) can not be improved for polynomials with real zeros .
a counterexample is @xmath322 and @xmath253 .
statement ( i ) ( and hence the 2-variance conjecture ) is equivalent to the following clear matrix theory conjecture , originally identified by borcea : [ toep ] assume that @xmath278 , @xmath323 and @xmath324 .
then the @xmath325 toeplitz matrix @xmath326 has at least one eigenvalue @xmath327 satisfying @xmath328 indeed , the operator @xmath94 in the orthonormal basis of differentiators @xmath329 given by ( [ v-3 ] ) has the circulant matrix @xmath330 where @xmath331 so that the sequence @xmath332 is the discrete fourier transform of the sequence @xmath333 .
assuming that @xmath334 we find the dependence @xmath324 . since @xmath335 , we see that @xmath336 finally the matrix @xmath133 from conjecture [ toep ] is exactly the matrix of the operator @xmath337 so that its eigenvalues are zeros of @xmath338 .
we recall the definition of the numerical range of a square matrix .
chapter 1 of @xcite is a standard reference for this subject .
let @xmath94 be an @xmath339 matrix .
the numerical range of @xmath94 is the set @xmath340 .
it can easily be seen that @xmath341 is a compact subset of the complex plane which contains the spectrum of @xmath94 and that if @xmath133 is a compression of @xmath94 then @xmath342 . toeplitz and hausdorff independently have shown that @xmath341 is always convex . if @xmath94 is normal , then @xmath341 is the convex hull of the spectrum of @xmath94 .
if @xmath94 is a @xmath343 by @xmath343 matrix then @xmath341 is a ( possibly degenerate ) ellipse whose foci are the eigenvalues of @xmath94 . to demonstrate the usefulness of numerical ranges in the analytic theory of polynomials
, we will use them to give a short proof of marden s theorem .
we recall that the steiner inellipse of a triangle is the inscribed ellipse which touches each side of the triangle at its midpoint .
let @xmath159 be a polynomial with three distinct zeros @xmath344 .
then the critical points of @xmath159 are the foci of the steiner inellipse of the triangle @xmath345 .
let @xmath346 be a three by three diagonal matrix with diagonal entries @xmath344 ; then @xmath347 .
let @xmath348 and let @xmath349 . since @xmath350 is the numerical range of a two by two compression of @xmath346 , @xmath350 is an ellipse contained in @xmath345 .
let @xmath351 . since @xmath352 , @xmath353 .
similarly , @xmath354 and @xmath350 is the steiner inellipse of @xmath345 . finally , as the projection onto @xmath355 is a differentiator of @xmath346 , the foci of @xmath350 are the critical points of @xmath159
. the numerical range may also prove useful in attacking the following strengthening of sendov s conjecture due to schmeisser @xcite .
let @xmath159 be an @xmath150th degree polynomial with @xmath356 and let @xmath179 be any complex number which is in the convex hull of the zeros of @xmath159 .
then the closed disk centred at @xmath179 with radius @xmath12 contains a critical point of @xmath0 .
we note that @xmath12 can not be replaced by @xmath357 for any finite @xmath7 in the above conjecture .
( if @xmath150 is sufficiently large so that @xmath358 , @xmath45 will be a counterexample ) .
borcea has given a matrix theoretical generalization of schmeisser s conjecture . before stating it , we introduce the following notation . for an @xmath339 matrix @xmath94 ,
we denote by @xmath359 its spectrum and by @xmath360 the @xmath361 matrix obtained by removing the @xmath362th row and @xmath362th column of @xmath94 .
let @xmath94 be an @xmath339 normal matrix .
then @xmath363 .
recall that for any @xmath339 normal matrix @xmath94 there is a set of differentiators of @xmath94 which form an orthonormal basis of @xmath364 .
if we only consider normal matrices for which every element of the standard basis is a differentiator then we recover schmeisser s conjecture .
so far we were interested in the shortest distance from a zero of a polynomial to its critical points . in the other direction
quite a few results are known about exclusion of critical points from regions around the zeros of a polynomial .
an ingenious combination of algebraic and geometric observations led j. von sz .
nagy to the following result .
let @xmath105 be a zero of multiplicity @xmath108 of the polynomial @xmath159 of degree @xmath150 .
assume that the equation @xmath365 admits at most @xmath366 zeros on every side of a line passing through @xmath105 and let @xmath367 be a closed disk passing through @xmath105 which does not contain any other zero of @xmath0 .
then @xmath368 on the closed disk internally tangent to @xmath367 , of diameter @xmath369 . without reproducing here the complete proof ( contained in @xcite ) ,
we discuss only a beautiful geometric observation appearing in sz .
nagy paper .
specifically , assume that @xmath370 is a critical point of @xmath371 where the zeros @xmath372 are not necessarily distinct .
if @xmath370 is not a zero of @xmath0 , then @xmath373 choose a real line @xmath374 passing through @xmath370 of slope @xmath375 with respect to the positive real semi - axis , and write the polar decompositions of the denominators above with respect to the origin @xmath370 and @xmath374 as zero - th direction : @xmath376 by taking imaginary parts in the above identity we find @xmath377 the diameter of the circle passing through @xmath75 and which is tangent to @xmath374 at @xmath370 is @xmath378 not excluding the value @xmath379 if the point @xmath75 lies on @xmath374 ( that is @xmath380 ) .
let s rearrange the zeros @xmath75 so that @xmath381 belong to one side of @xmath374 and @xmath382 belong to the other side , so that @xmath383 then we obtain the identity @xmath384 from here sz .
nagy obtains a series of non - trivial quantitative relations , of the form @xmath385 where @xmath121 is the multiplicity of the zero @xmath122 .
a consequence of the above proposition ( and the preceding geometric reasoning ) is the following result , independently obtained by alexander @xcite and walsh @xcite .
[ aw]let @xmath105 be a zero of multiplicity @xmath108 of the polynomial @xmath159 of degree @xmath150 and let @xmath386 denote the shortest distance from @xmath105 to another zero of @xmath0 .
then @xmath387 for @xmath139 belonging to the open disk centered at @xmath105 of radius @xmath388
. we offer below an alternate matrix theory proof . for any @xmath389 with @xmath390
, we define @xmath391 to be the set of all @xmath392-tuples of integers @xmath393 satisfying @xmath394 . if @xmath94 is a @xmath395 matrix and @xmath396 , then @xmath397\vert $ ] denotes the determinant of the @xmath398 principal submatrix of @xmath94 whose @xmath101th entry is the @xmath399th entry of @xmath94 .
we need the following well - known result : @xcite let @xmath94 be a @xmath395 matrix and let @xmath400 be the characteristic polynomial of @xmath94
. then @xmath401 and @xmath402\vert$ ] for @xmath403 .
let @xmath404 denote the @xmath405 identity matrix and let @xmath406 denote the @xmath405 matrix all of whose entries are one . using the previous lemma and some straightforward calculations one can prove the following : let @xmath346 be a @xmath395 diagonal matrix with characteristic polynomial @xmath85 and let @xmath407 .
then the characteristic polynomial of @xmath408 is @xmath409 .
we now prove theorem [ aw ] .
let @xmath410 .
then @xmath160 is an @xmath411th degree polynomial with no zero in the open disk centered at the origin of radius @xmath386 . since @xmath412
, we will be done if we show that @xmath413 has no zeros in the open disk centered at the origin of radius @xmath388 .
now let @xmath346 be an @xmath414 diagonal matrix whose characteristic polynomial is @xmath160 .
it is clear that @xmath415 .
let @xmath416 , then @xmath417 is a positive definite matrix with one eigenvalue equal to @xmath418 and all other eigenvalues equal to one .
hence @xmath419 .
the zeros of @xmath413 are the eigenvalues of @xmath420
. hence if @xmath139 is a zero of @xmath413 , we have @xmath421 and therefore @xmath422 . @xmath423
the localization of zeros of the cauchy transforms of positive measures is closely related to the similar problem for derivatives of polynomials .
indeed , given a polynomial ( [ v-1 ] ) , we have @xmath424 and hence the zeros of @xmath77 can be interpreted as zeros of the cauchy transform of the measure @xmath425 . in the case
when the support of the measure is finite , the problem reduces to the question of localization of zeros of the _ cauchy transform _ : @xmath426 where @xmath427 with @xmath428 and @xmath429 . for simplicity
, we may assume that @xmath430 so that @xmath431 is a probability measure .
as before , we may associate with the measure @xmath431 its baricenter @xmath432 and the variances @xmath433 at this point there are several possibilities to bring operator theory into the play .
the first one is to consider the same normal operator @xmath94 as before , i.e. the diagonal operator with @xmath229 on the diagonal .
if the vector @xmath232 is defined as @xmath434 and @xmath130 is the orthogonal projection to @xmath235 , then one can easily check that the eigenvalues of the operator @xmath234 acting in @xmath235 are exactly the zeros of the cauchy transform ( [ w-1 ] ) .
indeed , @xmath435 moreover , one has @xmath436 and @xmath437 with respect to the orthogonal decomposition @xmath438 , the operator @xmath94 takes the form @xmath439 where @xmath440 and @xmath441 . by analogy with the invertibility criterion for diagonally dominant matrices we are led to the following statement : + [ weighted - conj ] [ wei ] if @xmath442 and all eigenvalues @xmath164 of @xmath133 satisfy @xmath443 , then @xmath94 is invertible .
if the above statement were true , then the following statement about the zero location of cauchy transforms ( [ w-1 ] ) would also be true : [ cauchy - transform ] [ cau ] let @xmath444 and let @xmath445 denote the set which is the union of zeros of the cauchy transform ( [ w-1 ] ) and the point @xmath446 .
then @xmath447 we will see shortly that the above claim is false for @xmath44 . a natural attempt to save
it would be to restrict both sets appearing in the inequality to some relevant subsets .
first we adopt some general notation and restrict @xmath448 . for a compact subset @xmath449 , let @xmath450 denote the chebyshev radius of @xmath367 , i.e. the radius of the smallest closed disk containing @xmath367
. if @xmath431 is a probability measure , we denote by @xmath451 the closed support of @xmath431 and we put @xmath452 , and @xmath453 [ varconj ] @xmath454 a three - point mass example shows that claim ( [ varconj ] ) fails and hence so do claims ( [ wei ] ) and ( [ cau ] ) .
namely , the polynomial @xmath455 has the derivative @xmath456 $ ] and the associated measure @xmath457 thus , @xmath458 , @xmath459 and @xmath460 if @xmath461 , all points of @xmath445 are real and therefore the distance from @xmath462 to @xmath445 is strictly greater than @xmath463 .
this example also disproves a conjecture raised in @xcite .
finally , we restrict @xmath451 by defining for a finite point mass measure @xmath431 : @xmath464 we are led to formulate another statement which still implies sendov s conjecture . this time @xmath465 [ minvar ] for a finite point mass probability measure @xmath431 , @xmath466 nattapong bosuwan of vanderbilt university has shown this conjecture holds in the case when the set @xmath467 lies on the circle @xmath468 , where @xmath15 is the chebyshev center of @xmath451 as well as in the cases when @xmath469 consists either of three point masses lying anywhere on the complex plane or an arbitrary number of collinear point masses .
his proof of the first case is a generalization of the proof from @xcite that sendov s conjecture holds for zeros on the unit circle .
we present bosuwan s proof for the case where @xmath469 consists of three point masses .
@xcite if @xmath470 then conjecture [ minvar ] is true _ _ ( _ _ in fact with @xmath471 replaced by @xmath472__)__. without loss of generality we assume that @xmath473 and @xmath474 and proceed as in the proof of theorem [ 3main ] . define @xmath475 , @xmath476 and @xmath477 where @xmath478 are the elements of @xmath479 .
@xmath480 are mutually orthogonal and @xmath481 is an orthonomal basis of @xmath482 where @xmath483 .
then @xmath484 is an unitary matrix . by lemma 3.1
, @xmath485 is doubly stochastic .
let @xmath486 .
then @xmath487 , @xmath488 and @xmath489 where @xmath490 . since g is doubly stochastic , @xmath491 .
this implies that there exists @xmath492 such that @xmath493 .
let @xmath494 .
then from equation [ gl ] with @xmath150 replaced by @xmath374 and @xmath145 replaced by @xmath495 we get @xmath496 .
the chebyshev radius @xmath497 admits an operator - theoretical interpretation .
namely , it was proved by bjrck and thome @xcite that for any bounded normal operator @xmath94 in a hilbert space @xmath291 @xmath498 if @xmath94 is the operator of multiplication by an independent variable in @xmath499 for a compactly supported probability measure @xmath431 we have @xmath500 since we are dealing with matrix interpretations , an alternative possibility is to define the zeros of the cauchy transform ( [ w-1 ] ) as a spectrum is to use rank one perturbations of the above operator @xmath501 given @xmath502 , we introduce the rank - one operator @xmath503 as @xmath504 one can check that the spectrum of the operator @xmath505 is the set @xmath506 , where @xmath507 are zeros of ( [ w-1 ] ) . in particular , if @xmath508 , then @xmath509 . moreover , in this special case we have in addition the properties @xmath510 and @xmath511 .
sadly , in the labyrinth of observations we have stated and partially disproved in this section , we do not yet know how to state a weighted variance conjecture for probability measures supported by a continuum rather than finitely many points .
as was noticed above , the main difficulty in proving conjecture [ matrix ] consists in transforming known spectral properties of operators @xmath310 to metric properties . as an alternative , one can search for those metric properties of @xmath310 which would imply desired spectral properties .
we describe one possible scenario . as usual , we make first the reduction to the case where @xmath308 and @xmath512 .
further , it is known from matrix inertia theory ( see , e.g. @xcite ) that the following condition is sufficient for the existence of an eigenvalue of an operator @xmath133 ( acting in some hilbert space @xmath291 ) inside the unit disk : there exists an indefinite inner product @xmath513 in @xmath291 such that in connection with the discussion of the weighted variance conjecture in section 9 , it seems natural to raise here a question of plausible extensions to higher dimensions .
note that the rational function ( [ w-1 ] ) , or more generally , the cauchy transform @xmath518 where @xmath431 is any compactly supported probability measure in @xmath519 , can be interpreted as the force field ( i.e. , the complex gradient @xmath520 ) of the logarithmic potential @xmath521 in this context , the gauss - lucas theorem ( for probability measures @xmath431 ) is obvious .
the points where the gradient of the potential @xmath522 vanishes , i.e. , the zeros of the cauchy transform @xmath523 of the measure @xmath431 , all lie in the convex hull of the support of @xmath431 .
( at any point @xmath524 outside of the convex hull there would be a non - trivial component of the force field @xmath523 `` pushing '' a probational positive charge placed at @xmath524 off to infinity . ) in this form , the result extends word for word to higher dimensions if we replace the logarithmic potential by the coulomb ( or , in the context of gravitational force , newtonian ) potential in @xmath525 thus replacing @xmath522 by @xmath526 for @xmath527
. some further generalizations of the gauss - lucas theorem to higher dimensions can be found in @xcite , also cf .
however , it turns out that for atomic measures @xmath431 even the estimate for the total number of critical points where the gradient of the potential vanishes , i.e. , no force is present , is not known .
a conjecture going back to maxwell @xcite asserts that if the measure @xmath431 consists of @xmath528 point charges ( not necessarily all positive ) , the total number of _ isolated _ critical points of the potential does not exceed @xmath529 .
of course , for @xmath530 , which is a rational function in @xmath519 of degree @xmath528 with a zero of order @xmath528 at @xmath531 , the total number of finite critical points equals @xmath532 as follows at once from the fundamental theorem of algebra .
note that in @xmath533 , the critical points of the potential can form lines or curves as well , e.g. , if the @xmath534 charges alternate at the vertices of a square , the line through the center of the square which is perpendicular to the plane of the square will consist entirely of critical points of the potential .
examples of bounded curves , like circles and ellipses , of critical points in the electrostatic fields generated by three charges can be found in @xcite , also see @xcite which is more easily accessible for an english speaking reader .
maxwell simply asserted the claim without giving any justification for its validity . only recently , in @xcite , in a technically brilliant paper , the authors were able to prove that the number of isolated critical points of the potential is finite . yet , even for @xmath535 ( ! ) charges their separate proof for the possible number of isolated critical points yields the estimate @xmath536 instead of @xmath537 asserted by maxwell . a slight improvement of the estimate in @xcite under an additional assumption that all @xmath150 charges ( in @xmath538 ) lie in the same plane can be found in @xcite . also , we refer the reader to @xcite for the history of the problem and a rather complete , although surprisingly short , list of references that exist up to this date .
let @xmath541 , @xmath542 $ ] and let @xmath543 denote the set of all monic @xmath150th degree polynomials @xmath0 for which @xmath544 . since @xmath29 is continuous as a function of @xmath0 on the compact set @xmath543 , @xmath29 has an absolute maximum on @xmath543 .
conjecture 1 is equivalent to the statement that the polynomials @xmath545 where @xmath546 are absolute maxima for @xmath29 on @xmath543 .
conjecture 2 is equivalent to the statement that for @xmath40 , these polynomials are the only absolute maxima for @xmath29 on @xmath543 . in @xcite , miller looked at the local maxima of @xmath29 on @xmath547 .
it was known that all polynomials of the form @xmath545 where @xmath546 are local maxima of @xmath29 on @xmath547 @xcite .
miller found eight properties which together formed a sufficent condition for a polynomial to be a local maximum of @xmath29 on @xmath547 .
he then found polynomials which possess all eight properties ; these polynomials were local maxima but were not of the form @xmath548 @xcite .
these unexpected local maxima are particularly useful as test cases and possible counterexamples for conjectures .
the polynomial which disproved the conjecture in section 3 that every column of a gauss - lucas matrix of @xmath150th degree polynomial has an element greater than or equal to @xmath149 was @xmath154 which is a local maximum of @xmath29 on @xmath549 found by miller in @xcite
. it would be very instructive and useful to study the local maxima of @xmath29 on @xmath543 for finite @xmath7 of which little is known .
d. b. daz , d. b. shaffer , _ a generalization , to higher dimensions , of a theorem of lucas concerning the zeros of the derivative of a polynomial of one complex variable _ , applicable anal .
* 6 * ( 1976/77 ) , 109117 . | closely following recent ideas of j. borcea , we discuss various modifications and relaxations of sendov s conjecture about the location of critical points of a polynomial with complex coefficients .
the resulting open problems are formulated in terms of matrix theory , mathematical statistics or potential theory .
quite a few links between classical works in the geometry of polynomials and recent advances in the location of spectra of small rank perturbations of structured matrices are established .
a couple of simple examples provide natural and sometimes sharp bounds for the proposed conjectures . |
belle reported the first observation of @xmath1 production in @xmath2 meson decays with an inclusive measurement @xcite .
the @xmath20 @xcite momentum distributions in the @xmath18 rest frame ( cm frame ) indicate that most of the @xmath1 mesons come from non - two - body @xmath2 decays @xcite .
still , there have been only a few searches for exclusive @xmath2 decays with a @xmath1 in the final state , @xmath21 @xcite and @xmath22 @xcite .
the @xmath23 decays are found to be highly suppressed with respect to the similar @xmath0 processes @xcite .
the suppression can be explained in the framework of the factorization in two - body @xmath2 decays @xcite , where @xmath1 production is allowed only when one takes into account final state interactions .
due to angular momentum conservation , @xmath24 , @xmath25 and @xmath26 are favored while @xmath27 , @xmath28 , @xmath29 and so on are suppressed .
a study of the multi - body @xmath2 decay modes with @xmath0 and @xmath1 in the final state is important to understand the detailed dynamics of @xmath2 meson decays .
further , one can search for charmonium / charmonium - like exotic states in one of the intermediate final states such as @xmath30 and @xmath31 .
for example , looking at the @xmath32 invariant mass spectrum in @xmath33 decays , one can search for @xmath14 and/or @xmath13 .
the quantum numbers of the narrow exotic resonance @xmath13 have been determined to be @xmath34 @xcite .
one plausible interpretation is an admixture of a @xmath35 molecule and a conventional charmonium with the same @xmath36 , the yet - unseen @xmath14 @xcite .
the @xmath14 component may have a substantial decay rate to @xmath37 because of no obvious conflict in quantum numbers and observations of di - pion transitions between @xmath38 states in the bottomonium system . in case
that @xmath13 is not a mixed state and hence @xmath14 is a physically observable state , its decay to @xmath32 would still be expected .
its mass is predicted to be about 3920 mev/@xmath39 , assuming that it lies between @xmath40 and the @xmath41 that is interpreted as @xmath42 by pdg @xcite . using the @xmath43 modes
, we report on the inclusive branching fractions ( @xmath44 ) of @xmath45 decays and the exclusive reconstruction of multi - body @xmath2 decays to @xmath46 in order to search for still - undiscovered intermediate states .
we use a data sample of @xmath16 @xmath47 events collected with the belle detector @xcite at the kekb asymmetric - energy @xmath19 collider operating at the @xmath18 resonance @xcite .
the belle detector is a large - solid - angle spectrometer , which includes a silicon vertex detector ( svd ) , a 50-layer central drift chamber ( cdc ) , an array of aerogel threshold cherenkov counters ( acc ) , time - of - flight scintillation counters ( tof ) , and an electromagnetic calorimeter ( ecl ) comprised of 8736 csi(tl ) crystals located inside a superconducting solenoid coil that provides a 1.5 t magnetic field .
an iron flux return located outside the coil is instrumented to detect @xmath48 mesons and identify muons ( klm ) . the detector is described in detail elsewhere @xcite .
two inner detector configurations were used .
a first sample of @xmath49 @xmath47 events was collected with a 2.0 cm radius beam - pipe and a 3-layer svd , while the remaining @xmath50 @xmath47 pairs were collected with a 1.5 cm radius beam pipe , a 4-layer silicon detector and modified cdc ( the cathode part of the cdc replaced by a compact small cell - type drift chamber ) @xcite .
we reconstruct inclusive @xmath46 from @xmath2 decays . to suppress continuum background , we exploit the @xmath18 decay topology . for the events passing the belle standard hadronic event selection @xcite
, we require the ratio of the second to zeroth fox - wolfram moment @xcite to be less than 0.5 .
charged tracks are required to originate from the vicinity of the interaction point ( ip ) : the distance of closest approach to the ip is required to be within 3.5 cm along the beam direction and within 1.0 cm in the transverse plane .
photons are reconstructed from the energy deposition in the ecl by requiring no matching with any extrapolated charged track . to further avoid photons coming from neutral hadrons
, we reject the photon candidate if the ratio of the energy deposited in the central array of 3@xmath513 ecl cells to that deposited in the enclosing array of 5@xmath515 cells is less than 0.85 .
we use evtgen @xcite with qed final state radiation by photos @xcite for the generation of monte - carlo ( mc ) simulation events .
a geant - based @xcite mc simulation is used to model the response of the detector and determine the efficiency of the signal reconstruction .
the @xmath52 meson is reconstructed via its decays to @xmath53 ( @xmath54 = @xmath55 or @xmath56 ) and selected by the invariant mass @xmath57 . for the di - muon mode ,
@xmath58 is given by the invariant mass @xmath59 ; for the di - electron mode , the four - momenta of all photons within 50 mrad with respect to the original direction of the @xmath60 or @xmath61 tracks are included in @xmath62 to reduce the radiative tail .
the reconstructed invariant mass of the @xmath52 candidates is required to satisfy 2.95 gev@xmath63 gev@xmath64 or 3.03 gev@xmath65 gev@xmath64 . for the selected @xmath52 candidates ,
a vertex - constrained fit is applied to the charged tracks and then a mass - constrained fit is performed to improve its momentum resolution .
the @xmath0 and @xmath1 candidates are reconstructed by combining a @xmath52 candidate with a photon having an energy larger than 100 mev .
to reduce combinatorial background coming from @xmath66 , we use a likelihood function that distinguishes an isolated photon from @xmath67 decays using the photon - pair invariant mass , the photon laboratory - frame energy , and the laboratory - frame polar angle with respect to the beam direction @xcite . we reject both photons of a pair whose @xmath67 likelihood probability is larger than 0.3 . applying this cut , combinatorial background is reduced by 56.9% ( 59.1% ) with a signal loss of 26.5% ( 39.9% ) for @xmath1 ( @xmath0 ) . to identify the signal
, we use the distribution of the @xmath68 invariant mass @xmath69 and extract the signal yield from a binned maximum likelihood fit .
the signal of @xmath46 is described by a double - sided crystal ball function @xcite , which accommodates the tails of the mass distribution .
the function s left ( right ) side tail parameters @xmath70 ( @xmath71 ) and @xmath72 ( @xmath73 ) are fixed to the values obtained from mc simulated events . for @xmath74
, all other shape parameters are floated in the fit whereas , for @xmath75 , they are fixed using the mass difference ( @xmath76 ) from ref .
@xcite and the resolution ratio between @xmath0 and @xmath1 , @xmath77 , determined from mc simulations .
the combinatorial background component is modeled with a third - order chebyshev polynomial .
distribution of the @xmath78 decays in data .
the curves show the signal ( cyan dash - dotted for @xmath0 and red dashed for @xmath1 ) and the background component ( green dash - double - dotted for combinatorial ) as well as the overall fit ( blue solid ) .
the lower plot shows the pull of the residuals with respect to the fit.,width=340,height=226 ] figure [ fig : incdata_fit ] shows the fit of the @xmath79 distribution for @xmath80 and @xmath81 decays in the range of [ 3.297 , 3.697 ] gev@xmath64 .
the fit returns a reduced @xmath82 of 1.3 with a p - value of 0.0123 and a yield of @xmath83 events for the @xmath0 and @xmath84 events for the @xmath1 , where the errors are statistical .
the reconstruction efficiencies for the inclusive @xmath74 and @xmath75 decays are estimated to be 24.2% and 25.9% , respectively .
the efficiency is estimated using simulated multi - body @xmath2 decays , @xmath85 , where the number of pions @xmath86 varies from 0 to 4 over the entire @xmath87 range ; it is averaged with proper weighting according to the distribution of @xmath87 in data .
we use the 2014 world - average values @xcite for secondary daughter branching fractions @xmath88 , @xmath89 , and @xmath90 .
we use the 89 @xmath91 off - resonance data sample taken at 60 mev below the @xmath18 resonance to estimate the contribution of @xmath46 particles that do not arise from @xmath2 meson decays . from the fit to the @xmath79 distribution for that sample , we obtain @xmath92 ( @xmath93 ) signal events for @xmath0 ( @xmath1 ) , corresponding to @xmath94 ( @xmath95 ) signal events for @xmath80 ( @xmath81 ) after proper scaling to the integrated luminosity at the @xmath18 resonance .
the scaled @xmath0 and @xmath1 continuum yields are subtracted from the on - resonance yields .
one also expects a contribution from `` feed - down ''
@xmath96 decays where the @xmath46 is from the cascade @xmath97 . to determine the rate for direct decays to the @xmath46 states
, we subtract this feed - down contribution , which is estimated using @xmath98 and @xmath99 from ref .
@xcite .
the sources and estimates of the systematic uncertainties are summarized in table [ tab : syserr ] . a correction for small differences in the signal detection efficiency between mc and
data has been applied for the lepton identification requirements .
uncertainties in these corrections are included in the systematic error .
the @xmath100 and @xmath101 ( @xmath102 @xmath55 or @xmath56 ) samples are used to estimate the lepton identification correction .
the uncertainty of the probability density function ( pdf ) shapes are obtained by varying all fixed parameters by @xmath103 , fitting with different binning , and using a fourth - order polynomial for the background , then adding the changes in the yield in quadrature to get the systematic uncertainty .
we perform a fit to the data by including the @xmath104 component and find its statistical significance to be 1.7@xmath105 .
we further add the signal yield difference for @xmath0 or @xmath1 with respect to the original fit to the pdf systematic uncertainty .
based on this , we get an uncertainty of 3.1% ( 7.9% ) for @xmath74 ( @xmath75 ) . the uncertainties due to the secondary branching fractions are also taken into account . the uncertainty on the track finding efficiency is found to be 0.35@xmath106 per track by comparing the data and mc for @xmath107 decay , where @xmath108 and @xmath109 here one of the @xmath110 is allowed not to be reconstructed explicitly
for @xmath111 , systematic uncertainty is estimated to be @xmath112 . the uncertainty on the photon identification is estimated to be 2.0% from sample of radiative bhabha events .
the systematic uncertainty associated with the difference of the @xmath67 veto between data and mc is estimated to be 1.2% from a study of the @xmath113 sample . the potential bias to extract signal yields of the @xmath46
is estimated by the mc from variation of the efficiency for the different decay modes bin by bin in the @xmath114 distribution . the efficiency change due to the unknown @xmath46 polarization
is estimated using the @xmath115 signal mc samples by varying the polarization over the allowed range .
the sum of these two effects is 4.0% .
we measure the feed - down - contaminated branching fractions @xmath116 and @xmath117 to be @xmath118 and @xmath119 , respectively , where the first ( second ) error is statistical ( systematic ) .
after subtracting the feed - down contribution , we obtain the pure inclusive branching fractions @xmath116 = @xmath120 and @xmath117 = @xmath121 . in both cases ,
the systematic uncertainty dominates .
we estimate the inclusive branching fractions according to the formula : @xmath122 here , @xmath123 is the obtained signal yield , @xmath124 is the estimated off - resonance contribution , @xmath125 is the reconstruction efficiency , @xmath126 is the number of @xmath2 mesons in the data sample and @xmath44 is the branching fraction for the particular mode taken from @xcite .
.[tab : syserr]summary of systematic uncertainties in the @xmath45 branching fraction . [ cols="<,^,^ " , ] table [ tab : syserrall ] summarizes the systematic for each mode .
corrections for small differences in the signal detection efficiency between mc and data have been applied for the lepton and kaon identification requirements , as was done in the inclusive study .
in addition to the items commonly affecting the inclusive branching fractions measurements , we consider the following systematic uncertainty sources . in belle , dedicated @xmath127 samples are used to estimate the kaon ( pion ) identification efficiency correction . to estimate the correction and residual systematic uncertainty for @xmath128 reconstruction ,
@xmath129 samples are used . for @xmath67 , the efficiency correction and systematic uncertainty
are estimated from a sample of @xmath130 decays .
the errors on the pdf shapes are obtained by varying all fixed parameters by @xmath103 and taking the change in the yield as the systematic uncertainty .
lcccccccccccc|c & + mode & @xmath54 & @xmath131 & @xmath110 & tracking & @xmath132 i d & secondary @xmath44 & @xmath128 & @xmath67 & @xmath67 veto & @xmath125 & pdf & @xmath111 & total + + @xmath0 & 2.1 & 1.0 & 1.0 & 1.4 & 2.0 & 3.6 & &&1.2 & 1.0 & 1.4 & 1.4 & 5.6 + @xmath1 & 2.1 & 1.0 & 1.0 & 1.4 & 2.0 & 3.7 & & &1.2 & 2.4 & 3.2 & 1.4&6.7 + + @xmath0 & 2.1 & & 1.0 & 1.8 & 2.0 & 3.6 & 0.7 & & 1.2 & 1.0 & 0.7 & 1.4&5.5 + @xmath1 & 2.1 & & 1.1 & 1.8 & 2.0 & 3.7 & 0.7 & & 1.2 & 2.2 & 9.1 & 1.4&10.8 + + @xmath0 & 2.1 & 1.0 & & 1.1 & 2.0 & 3.6 & & 2.2 & 1.2 & 1.0 & 1.7&1.4 & 5.9 + @xmath1 & 2.1 & 1.0 & & 1.1 & 2.0 & 3.7 & & 2.2 & 1.2 & 2.4 & 191 & 1.4&191.1 + + @xmath0 & 2.1 & 1.1 & 2.3 & 1.8 & 2.0 & 3.6 & & & 1.2 & 1.0 & 2.6&1.4 & 6.5 + @xmath1 & 2.1 & 1.2 & 2.3 & 1.8 & 2.0 & 3.7 & & & 1.2 & 2.2 & 2.0&1.4 & 6.6 + + @xmath0 & 2.1 & & 2.3 & 2.1 & 2.0 & 3.6 & 0.7 & & 1.2 & 1.0 & 8.1&1.4 & 10.1 + @xmath1 & 2.1 & & 2.3 & 2.1 & 2.0 & 3.7 & 0.7 & & 1.2 & 2.3 & 31.5&1.4 & 32.1 + + @xmath0 & 2.1 & 1.1 & 1.1 & 1.4 & 2.0 & 3.6 & & 2.2 & 1.2 & 1.1 & 3.6&1.4 & 6.9 + @xmath1 & 2.1 & 1.2 & 1.1 & 1.4 & 2.0 & 3.7 & & 2.2 & 1.2 & 2.9 & * * & 1.4 & 7.5 + table [ tab : final_results ] summarizes the studied exclusive decays of @xmath2 to @xmath133 decays . for the first time , we observe the @xmath134 , @xmath9 , @xmath8 , @xmath10 , @xmath135 and @xmath136 decay modes .
we find that in three - body decays the @xmath1 is more likely to be produced in association with higher @xmath137 resonances ; in contrast , decays to @xmath0 are accompanied predominantly by the @xmath138 .
the same phenomenon is observed in the four - body production of @xmath1 and @xmath0 from @xmath2 decays .
no strong hint for any narrow resonance ( less than 5 mev width ) is seen in the @xmath139 and @xmath140 distributions . if one adds the measured branching fraction in this paper ( excluding the obtained u.l . ) , we obtain @xmath141 with @xmath142 to be @xmath143 , which corresponds to a @xmath144 fraction of the measured @xmath3 . using @xmath145 @xcite , this accounts for @xmath146 of @xmath2 mesons decaying into @xmath80 . similarly , @xmath147 with @xmath148 is @xmath149 , corresponding to @xmath150 of the inclusive @xmath5 .
for the treatment of the uncertainty , no correlation is assumed and the uncertainty is the sum of the systematic and statistical uncertainties in quadrature .
we measured the feed - down - contaminated @xmath116 and @xmath117 of @xmath118 and @xmath151 , respectively , where the first ( second ) error is statistical ( systematic ) . after subtracting the @xmath152 feed - down contributions
, we find the pure inclusive branching fractions @xmath116 and @xmath117 of @xmath120 and @xmath153 , respectively . here , the systematic uncertainty dominates . for inclusive production of @xmath46
, we measure the ratio @xmath154 of @xmath155 .
we observe the @xmath134 decay mode for the first time , with @xmath156 signal events and a significance of 8.7@xmath105 , along with evidence for the @xmath9 decay mode , with 76@xmath15715 signal events and a significance of @xmath158 . in four - body
decays , we observe the @xmath10 , @xmath10 , @xmath11 , and @xmath12 decay modes for the first time and report on measurements of their branching fractions .
we find that @xmath1 production , in contrast with @xmath0 , increases with a higher number of multi - body @xmath2 decays : @xmath159 for @xmath160 decay ( [email protected] ) is almost twice that measured in the @xmath161 decay mode ( @xmath162 ) .
we observe that the @xmath1 is more often accompanied by higher @xmath137 resonances , in contrast to the @xmath0 that is dominantly produced with the lower @xmath137 resonance .
all previous studies @xcite were limited to @xmath163 , while our study suggests that @xmath1 is preferentially produced with higher @xmath137 resonances .
clearly , to study @xmath1 production in @xmath2 decays , it is important to avoid considering solely the lower @xmath137 resonances .
suppression in two - body @xmath2 decays is found to be due to the factorization hypothesis @xcite . in our search for @xmath164 and @xmath14 ,
we determine an u.l . on the product of branching fractions @xmath165 [ @xmath166 @xmath167 @xmath168 [ @xmath169 at the 90% c.l . the negative result for our searches is compatible with the interpretation of @xmath13 as an admixture state of a @xmath35 molecule and a @xmath14 charmonium state .
we thank the kekb group for the excellent operation of the accelerator ; the kek cryogenics group for the efficient operation of the solenoid ; and the kek computer group , the national institute of informatics , and the pnnl / emsl computing group for valuable computing and sinet4 network support . we acknowledge support from the ministry of education , culture , sports , science , and technology ( mext ) of japan , the japan society for the promotion of science ( jsps ) , and the tau - lepton physics research center of nagoya university ; the australian research council ; austrian science fund under grant no .
p 22742-n16 and p 26794-n20 ; the national natural science foundation of china under contracts no . 10575109 , no . 10775142 , no .
10875115 , no .
11175187 , and no . 11475187 ; the chinese academy of science center for excellence in particle physics ; the ministry of education , youth and sports of the czech republic under contract no .
lg14034 ; the carl zeiss foundation , the deutsche forschungsgemeinschaft and the volkswagenstiftung ; the department of science and technology of india ; the istituto nazionale di fisica nucleare of italy ; the wcu program of the ministry of education , national research foundation ( nrf ) of korea grants no .
2011 - 0029457 , no .
2012 - 0008143 , no . 2012r1a1a2008330 , no .
2013r1a1a3007772 , no .
2014r1a2a2a01005286 , no . 2014r1a2a2a01002734 , no .
2015r1a2a2a010032 + 80 , no .
2015h1a2a1033649 ; the basic research lab program under nrf grant no .
krf-2011 - 0020333 , center for korean j - parc users , no .
nrf-2013k1a3a7a06056592 ; the brain korea 21-plus program and radiation science research institute ; the polish ministry of science and higher education and the national science center ; the ministry of education and science of the russian federation and the russian foundation for basic research ; the slovenian research agency ; the basque foundation for science ( ikerbasque ) and the euskal herriko unibertsitatea ( upv / ehu ) under program ufi 11/55 ( spain ) ; the swiss national science foundation ; the national science council and the ministry of education of taiwan ; and the u.s.department of energy and the national science foundation .
this work is supported by a grant - in - aid from mext for science research in a priority area ( `` new development of flavor physics '' ) , for scientific research on innovative areas ( `` elucidation of new hadrons with a variety of flavors '' ) , and from jsps for creative scientific research ( `` evolution of tau - lepton physics '' ) .
the double - sided crystal ball function is defined as : + @xmath170^{-n_l } , \\~~~~~~~~~~~~~ \text{if}~ ( x-\mu)/\sigma \leq -\alpha_l \\~~~~~~~~=~ n_0 \left ( n_r/|\alpha_r| \right)^{n_r } \exp \left ( - |\alpha_r|^2/2 \right ) \times \\~~~~~~~~~~~~~~~~~~~~~~~~~~~~~\left [ n_r/|\alpha_r| - |\alpha_r| + ( x - \mu)/\sigma\right]^{-n_r } , \\~~~~~~~~~~~~~ \text{if}~ ( x-\mu)/\sigma \geq \alpha_r \\~~~~~~~=~ n_0 \exp \left [ - ( x - \mu)^2/(2\sigma^2 ) \right ] , ~~\text{otherwise}. $ ] + here @xmath171 is the normalization , @xmath105 is the standard deviation , @xmath56 is the mean , @xmath70 ( @xmath71 ) , @xmath72 ( @xmath73 ) are shape parameters for left ( right ) tail .
the logarithmic gaussian is parameterized as : + @xmath172 @xmath173\}^2/(2\sigma^2_0)}$ ] where @xmath174 , @xmath175 and @xmath176 . here
, @xmath171 is the normalization , @xmath105 is the standard deviation , @xmath177 is the mean and @xmath178 is the asymmetry . | we report inclusive and exclusive measurements for @xmath0 and @xmath1 production in @xmath2 decays .
we measure @xmath3= @xmath4 and @xmath5= @xmath6 . for the first time
, @xmath1 production in exclusive @xmath2 decays in the modes @xmath7 and @xmath8 has been observed , along with first evidence for the @xmath9 decay mode . for @xmath0 production ,
we report the first observation in the @xmath10 , @xmath11 and @xmath12 decay modes . using these decay modes
, we observe a difference in the production mechanism of @xmath1 in comparison to @xmath0 in @xmath2 decays .
in addition , we report searches for @xmath13 and @xmath14 in the @xmath15 decay mode .
the reported results use @xmath16 @xmath17 events collected at the @xmath18 resonance with the belle detector at the kekb asymmetric - energy @xmath19 collider . |
the bicep2 collaboration reports the signal of gravitational waves in the b - mode power spectrum in spatially flat spacetimes@xcite .
these tensor perturbations ( gravitational waves ) arises from infaltionary in the early universe . as is well known ,
gravitational waves propagate freely in the expanding universe with this assumption that the universe is perfect fluid @xcite and also the cosmological generation of gravitational waves was considered by l.p.grishchuk @xcite and a.a.starobinsky@xcite . even though in wmap nine - year data report the results
have been interpreted as an evidence in the support of flat universe , in no way do the data rule out the case of @xmath2@xcite . by choosing the maximally extended de - sitter metric as unperturbed background
, we found the evolution of tensor modes fluctuation in curved spacetime @xcite .
we compute only the tensor mode fluctuation between the time of inflation and present specially from a temperature @xmath3 when electron positron annihilation is substantially complete and neutrinos have decoupled from matter and radiation .
this paper is organized as follows . in sec.ii , we present the neutrino contribution to the anisotropic inertia tensor @xmath4 so that for calculation of high accuracy it is necessary to use the boltzmann equations of kinetic theory for neutrinos in phase space and then get the integro - differential equation of tensor mode fluctuation in closed cosmology . in sec.iii , we discuss short wavelengths re - entering the horizon during the radiation dominated phase because modes are time - independent when leaving the horizon . in sec.iv , we present the gw s equation in matter dominated era and discuss the effects of free streaming neutrinos on it . in last section
we conclude with discussing and comparing the results of flat @xcite,@xcite and closed cosmology .
in general the tensor mode gravitational perturbation takes the form @xmath5 and the tensor fluctuation in curved spacetime satisfies@xcite @xmath6 for which the dot stands for derivative with respect to the ordinary time , @xmath7 is curvature constant and @xmath8 is anisotropic inertia tensor . in the perfect fluids and independent of the background space time
the interaction of tensor modes with matter and radiation is zero @xcite , but there is a traceless transverse term in the anisotropic tensor so @xmath9 is only the anisotropic part of the stress tensor that is the spatial part of the perturbed energy - momentum tensor @xmath10 or @xmath11 where @xmath12 is unperturbed pressure .
the components of the perturbed metric in cartesian coordinate are @xcite @xmath13 if we consider only tensor perturbation so that @xmath14 the anisotropic inertia tensor is the sum of the contributions from photons and neutrinos but photons have a small contribution to the anisotropic inertia due to a short mean free time and neutrinos traveling without collisions when temperature dropped about @xmath15 so neutrino distribution function in phase space has a form @xmath16 where @xmath17 is individual neutrinos and anti neutrino trajectories . in the absence of collisions terms , boltzmann equation for neutrinos would be @xmath18 where @xmath19 and @xmath20 are the rate of the change of momentum and the rate of change of the coordinate respectively , so the above relation will be @xmath21 @xmath22 in the start of free streaming has the form of the ideal gas : @xmath23\end{aligned}\ ] ] where @xmath24 is the number of types of neutrinos and separately antineutrinos and @xmath25 is the boltzmann constant . with a small perturbation to the metric , the neutrino disturbation function gets a small difference from its equilibrium form as @xmath26 if the background spacetime is non - flat@xmath27 so @xmath28 , @xmath29 and @xmath30 are of independent variable @xmath31 by @xmath32 and @xmath33 and @xmath34 .
to first order in metric and density perturbation eq.([1 ] ) will be @xmath35 with @xmath36 , the relativistic boltzmann equation for the perturbation @xmath37 in curved spacetime will be @xmath38 we use a dimensionless intensity perturbation @xmath39 , defined by @xmath40 where @xmath41 is the number of species of neutrinos and antineutrinos and @xmath42 and also with @xmath43 the boltzmann equation ( [ 2 ] ) becomes @xmath44 the above equation can be have a solution in the following form @xmath45 and also we define @xmath46 as @xmath47 where @xmath48 is a stochastic parameter for the single non - decaying mode with discrete wave number @xmath49 and the helicity @xmath50 and @xmath51 is the corresponding polarization tensor . with @xmath52 and working in @xmath53 , eq.([3 ] ) becomes an equation for @xmath54 : @xmath55 and we can find a solution of above equation with green function method as @xmath56 the green function is obtained as @xmath57 in this case @xmath58 is @xmath59 \theta(\tau^{\prime}q-2\tau q ) e^{i\tau^{\prime } q(\mu^{\prime}+ 2\tau^{\prime}q)}\nonumber\\\end{aligned}\ ] ] in the tensor mode , the only nonvanishing component is @xmath60 : @xmath61 this is the neutrino contribution of the anisotropic inertia tensor @xmath62 : @xmath63 \nonumber\\&\times & \int _ { -1}^{+1}d\mu^{\prime } \int_{0}^{\tau q } d\tau^{\prime}[-2\dot{d}_{q}(\tau^{\prime } ) ] e^{i\tau^{\prime } q(\mu^{\prime}+ 2\tau^{\prime}q)}\end{aligned}\ ] ] in the right hand side , the second and third terms compared with the first term are too small so by ignoring them , the integro - differential equation of gravitational waves in the presence of inertia tensor of neutrinos becomes @xmath64\dfrac{\sin \tau^{'}q}{\tau^{'}q } e^{2i\tau^{\prime 2 } q^{2}}\end{aligned}\ ] ] the right hand side of the equation ( [ 7 ] ) is more complicated than the case of flat spacetime . in the next section
we will calculate the decay of gravitational waves in the radiation dominated era .
we found a complete solution of differential equation for tensor perturbation in the closed universe@xcite .
we begin by neglecting the anisotropic inertia tensor @xmath4 and then we consider @xmath65 so with this approximation , the field equation ( [ 4 ] ) governing the fourier components of the tensor components of @xmath66 becomes @xmath67 where @xmath68 and @xmath69 is discrete . to treat the evolution of @xmath70 , it is convenient to change the independent variable @xmath71 to @xmath72 , where @xmath73 is the value of the robertson - walker scale factor at matter - radiation equality . from friedmann equation
we have @xmath74 where @xmath75 is the expansion rate at matter - radiation equality , @xmath76 and @xmath77 are vacuum and curvature energy respectively when radiation and matter are equal . to investigate a short enough wavelength to have re - entered the horizon during the radiation dominated era , we consider @xmath78 the equation([6 ] ) with the change of variable from @xmath79 to @xmath80 is @xmath81 where @xmath82 and @xmath69 is discrete number . for @xmath83
the solution is @xmath84 and general solution is @xmath85 the solution is similar to the flat case but @xmath86 depends on discrete number @xmath69 in closed universe .
the fraction of the total energy density in neutrinos is @xmath87 where @xmath88 and the third term in the denominator is too small and can be ignored . in the presence of the anisotropic inertia tensor with change @xmath89 instead of @xmath79 and using the friedmann equation @xmath90 , the gravitational wave equation([7 ] ) in radiation dominated era becomes @xmath91 when @xmath92 the homogeneous solution is constant value of @xmath93 then @xmath94 so the gravitational waves equation becomes @xmath95 the solution is identical to the flat spacetime but for @xmath96 when the tensor modes are deep inside the horizon , the integro - differential equation ( [ 8 ] ) becomes @xmath97 and its general solution is @xmath98 where @xmath93 is constant and @xmath99 is the cosine integral as @xmath100
. deep inside of horizon when @xmath101 , the right hand side of the eq .
( [ 101 ] ) becomes negligible and the solution approaches a homogeneous solution as @xmath102 for large @xmath103 , @xmath99 and the third term of the general solution tend to zero so @xmath104
. also a numerical solution of eq.([101 ] ) shows that @xmath105 follows the @xmath106 solution pretty accurately until @xmath107 when the perturbation enters horizon [ as compared with the solution @xmath108 for @xmath106 ] and thereafter the solution rapidly approaches to the @xmath110 .
so then the effect of neutrino damping reduces the tensor amplitude by the factor of @xmath111 in the closed cosmology while in the flat case the factor was @xmath112 @xcite .
hence the tensor contribution to the temperature multipole coefficient @xmath113 and the whole of the @xmath114 polarization multipole coefficient @xmath115 will be @xmath116 less than they would be without damping due to free -streaming neutrinos . therefore in the radiation dominated era and in closed cosmology , the amplitude of the gravitational waves in presence of neutrinos will be less than the flat case or neutrinos have a greater effect on the damping of gravitational waves in closed cosmology as shown in fig.1 .
to investigate the tensor perturbation may enter horizon after the matter energy density becomes important , we consider @xmath117 so from ( [ 10 ] ) we have @xmath118 where @xmath76 is vacuum energy at the equality of radiation and matter .
using @xmath119 and @xmath120 equation ( [ 6 ] ) becomes @xmath121 where @xmath122 and @xmath69 is discrete number .
this equation is different from gravitational waves equation in flat cosmology @xcite . whatever the value of @xmath123 , the general solution is @xmath124-c_{n}^{0}[\dfrac{\kappa}{y } \sin(\dfrac{\kappa}{y})+\cos(\dfrac{\kappa}{y})]\end{aligned}\ ] ] where @xmath125 and @xmath126 are constant .
the solution is not oscillating in @xmath127 and the neutrinos have no effect on the gravitational waves .
also we impose the condition that the solution should approach the constant of the value @xmath126 for @xmath128 so the solution will be @xmath129\end{aligned}\ ] ] similar to the flat spacetime , all wavelength to take the gravitational wave amplitude in matter - dominated era are to be given by multiplying the above solution with the factor @xmath130 : @xmath131\end{aligned}\ ] ] where @xmath130 is @xmath132 for @xmath133 and @xmath134 for @xmath127 because the damping effect is equal to zero for @xmath127 .
therefore we can consider @xmath135 and amplitude of gravitational waves will be reduced by the factor @xmath130 .
at first we obtained the relativistic boltzmann equation for perturbation @xmath136 in spatially closed cosmology , the perturbation on the metric background creates a disturbance in the number density of neutrinos and then we investigate the effects of anisotropic inertia tensor on the damping of gravitational waves in the radiation and matter dominated era .
amplitude of gravitational waves is a sensitive function of discrete wave number . during the radiation dominated era when
the tensor perturbation re - entered the horizon , the effect of neutrinos on gws in flat spacetime is weaker than the closed case and the evolution is independent of any cosmological parameters . in closed cosmology , in matter - dominated era the equation of gravitational waves is different from the flat case and the propagating of the gravitational waves will be different from the flat case and the effect is less for wavelengths that enter the horizon at later times . the evolution of gws at later times after the radiation - matter equality depends on vacuum energy in the equality of matter and radiation
. 99 e.m .
lifshitz , zh . eksp.teor.phys . 16,587 ( 1946 ) ; l.p .
grishchuk , zh .eksp .
67 , 825 ( 1974)[sov .
phys.jetp 40,409(1975 ) ] ; l.h .
ford and l.parker , phys .
rev.d 16.1601 ( 1977 ) l.p .
grishchuk , jetp 40 , 409(1974 ) .
a.a.starobinsky , jetp lett .
30 , 682(1979 ) .
e. komatu and et .
al , apjs , 192 : 18 , ( 2011 ) .
s. weinberg , phys .
d * 69 * , 023503 ( 2004 ) .
[ astro - ph/0306304 ] .
b.b.stefanek , w.w.repko,phys.rev.d , 88 , 083536(2013 ) s.weinberg , gravitation and cosmology(weily , new york,1972 ) .
v. a. rubakov , m. sazhin , and a. veryaskin , phys .
115b , 189(1982 ) .
r. fabbri and m.d .
pollock , phys .
125b , 445 ( 1983 ) .
abbott and m. b. wise , nuclear physics b244 , 541 ( 1984 ) .
a. a.starobinsky , sov .
11 , 133 ( 1985 ) .
s. weinberg , cosmology ( oxford university press , new york , 2008 ) .
d.dicus and w.repko , phys.rev.d , 72 , 088302(2005)[arxiv:0509096[astro - ph ] ] | we derive an integro - differential equation for propagation of cosmological gravitation waves in spatially closed cosmology whereas the traceless transverse tensor part of the anisotropic stress tensor is free streaming neutrinos ( including antineutrinos ) , which have been traveling essentially without collision since temperature dropped below about @xmath0 .
we studied the short wavelengths and long wavelengths of gravitational waves ( gws ) that enter the horizon in closed spacetime .
the solution shows that the anisotropic stress reduces the squared amplitude by @xmath1 for wavelengths that enter the horizon during radiation - dominated phase and this reduction is less for the wavelength that enter the horizon at later times . at the end
we compare the results to the flat case and then we investigated the dependence of the evolution of gws on cosmological parameters . |
suppressed noise in two - mode multi - particle systems known as `` spin squeezing '' introduced by kitagawa and ueda @xcite is an essential tool in quantum metrology protocols @xcite .
the interferometric schemes utilizing this effect cover broad area of possible physical systems , ranging from collective spins of neutral atoms interacting by collisions @xcite , atoms interacting with light by faraday rotation and ac - stark shift @xcite , atoms interacting by rydberg blockade @xcite , polarized light @xcite , to bose - einstein condensates ( bec ) in double - well potentials ( bosonic josephson junctions ) @xcite .
typically , the preparation of spin squeezed states is based on nonlinear inter - particle interactions . in terms of the collective `` spin '' operators @xmath0 ,
the procedures have been classified as `` one - axis twisting '' ( oat ) with a term @xmath1 , and `` two - axis counter - twisting '' ( tact ) with a term @xmath2 , the tact being shown to be more efficient to produce highly squeezed states @xcite .
recently , a scheme has been proposed to combine a sequence of oat and spin rotations to an effectively tact procedure @xcite .
efficient preparation of spin - squeezed states has become an objective of various optimized procedures @xcite . here
i show that any quadratic interaction in the collective spin can be described by means of a twisting tensor , encompassing the oat and tact as special cases .
equations of motion for the first and second moments in the gaussian approximation are used to show how squeezing is generated in various cases of the twisting tensor . at the initial stage ,
the maximum squeezing rate only depends on the difference between the maximum and minimum eigenvalues of the twisting tensor . for certain times ,
deviations from the optimum squeezing rate can be compensated by suitable rotations .
the results are applicable for optimizing strategies of interferometric measurements with various nonlinear media .
the paper is organized as follows . in sec .
[ secham ] the system hamiltonian and equations of motion are derived , in sec .
[ secphys ] possible schemes for physical realization are mentioned , in sec .
[ secrate ] the rate of squeezing generation is studied , in sec .
[ secsolution ] approximate solutions of the equations of motion are given , in sec .
[ secoptimum ] the conditions for generating squeezing at maximum rate are found , and a conclusion is given in sec .
[ secconcl ] .
consider a two - mode bosonic system described by annihilation operators @xmath3 and @xmath4 with total number of particles @xmath5 conserved .
the dynamics can be expressed by operator @xmath6 defined as @xmath7 with @xmath8 .
the components of @xmath6 satisfy the angular momentum commutation relations @xmath9=ij_z$ ] , @xmath10=ij_x$ ] , and @xmath11=ij_y$ ] .
let the hamiltonian be composed of @xmath3 , @xmath4 , @xmath12 , and @xmath13 such that in each term the same number of creation and annihilation operators occurs ( total number of particles is conserved ) , and the highest power of each operator is 2 .
the hamiltonian then can be written as @xmath14 where @xmath15 and @xmath16 transform as vectors and @xmath17 transforms as a tensor under o(3 ) rotations . here
@xmath18 , and the einstein summation is used . in eq . (
[ ham ] ) , @xmath19 is a linear or quadratic function of the total particle number , generating an unimportant overall phase .
let us call @xmath20 the _ twisting tensor _ and note that the special case of @xmath21 for @xmath22 or @xmath23 , @xmath24 corresponds to the oat scenario , and the case @xmath25 , @xmath26 otherwise , corresponds to the tact scenario of @xcite . since @xmath27 , addition of an arbitrary multiple of unit matrix to @xmath20
can be absorbed in the unimportant term @xmath19 .
therefore , any diagonal @xmath20 in which one element is exactly in the middle of the remaining two elements also corresponds to the tact .
hamiltonian ( [ ham ] ) specified by three parameters @xmath28 as @xmath29 , @xmath30 , @xmath31 , with @xmath32 otherwise , corresponds to the lipkin - meshkov - glick ( lmg ) @xcite model which was introduced as a solvable model of atomic nucleus and serves as a paradigm to study quantum phase transitions @xcite . using the heisenberg equations of motion ,
@xmath33 $ ] , and calculating the mean values of the operators , we arrive at the equations for @xmath34 @xmath35 , \label{dotj}\end{aligned}\ ] ] where @xmath36 is the variance tensor , @xmath37 the equations for @xmath38 can be obtained in a similar way , however , in this case mean values of cubic terms @xmath39 occur .
our approximation is based on the assumption that the distribution of the components @xmath16 is close to gaussian for which all higher moments are functions of the first and second moments . in particular , we express the third moments as @xmath40 thus we find @xmath41 where @xmath42 is the levi - civita symbol . equations ( [ dotj ] ) and ( [ dotv ] ) form a closed set of 9 equations for 9 dynamical variables describing rotational and squeezing properties of the system .
note that a special case of this set for oat with @xmath43 and @xmath26 for @xmath44 has been studied in @xcite where the influence of the variances @xmath45 on the first moments @xmath46 in eq .
( [ dotj ] ) has been interpreted as the `` bogoliubov backreaction '' .
a simple scheme realizing nontrivial @xmath20 is in fig .
[ f - scheme1](a ) .
two optical resonators are crossed and their fields are mixed by a balanced beam splitter . in each of the four branches a kerr medium induces a phase shift proportional to the intensity of the field .
thus , e.g. , light passing through the medium in branch @xmath3 picks up the phase @xmath47 .
if the round - trip duration is @xmath48 , the hamiltonian can be written as @xmath49 a special case of @xmath50 , and @xmath51 reduces the hamiltonian to @xmath52 note that already this simplest form of interaction covers all the main categories of spin squeezing : oat ( @xmath53 or @xmath54 ) , tact ( @xmath55 or @xmath56 ) , or more general twistings ( other relations between @xmath57 and @xmath58 ) , as well as the lmg model [ @xmath59 , @xmath60 .
although in the last expression the rotation frequency depends on @xmath5 , rotating terms of various @xmath61 independent of @xmath5 can be introduced by shifting the positions of the mirrors and/or by tuning the parameters of the beam splitter .
any more general form of the twisting tensor , including off - diagonal terms , can be achieved by chaining the beam splitters and nonlinear zones as in fig .
[ f - scheme1](b ) .
the model is general : rather than optical resonators one can assume , for example , two bosonic traps with several josephson junctions and position - dependent nonlinearities induced by feshbach resonances .
recently , a scheme with a ring bec trap with spatially modulated nonlinearity has been proposed to realize tact and other squeezing regimes as well as the lmg model @xcite .
depending on the particular physical realization , a specific decoherence or loss mechanism will limit performance of the scheme .
for instance , in the optical scheme the absorption may become dominant , whereas in the scheme discussed in @xcite we anticipate the inelastic atomic collisions to represent the most important limitation . detailed discussion of the influence of particle losses and thermal noise for the oat bec schemes have been given in @xcite .
expanding the model to cover general twisting interactions , eqs .
( [ dotj ] ) and ( [ dotv ] ) would then be generalized based on the corresponding master equation .
these problems will be studied in a subsequent work .
let us first choose the coordinate system such that the state is centered at the pole of the bloch sphere with @xmath62 . using ( [ dotv ] )
we find @xmath63 , @xmath64 and @xmath65 while expressing the variance matrix @xmath66 , @xmath67 as the rotated diagonal matrix of principal variances @xmath68 , where @xmath69 thus we find @xmath70v_{\pm } , \end{aligned}\ ] ] where @xmath71 is the orientation angle of the squeezed state .
the optimum rate occurs for @xmath71 satisfying @xmath72 for which @xmath73 where @xmath74 is the optimum squeezing rate . to find the maximum squeezing rate and optimum variance orientation for arbitrary location of the state
, we have to transform the components of the twisting tensor . for simplicity
, we choose the coordinate system oriented such that @xmath20 is diagonal .
two angles , @xmath75 and @xmath76 , determine the direction of the state as shown in fig .
[ f - koule]b . on calculating the elements of @xmath20 in the new
coordinates one finds @xmath77 ^ 2 \nonumber \\ & & + \left .
4(\chi_x-\chi_y)^2\cos^2 \vartheta \cos^2 \varphi
\sin^2 \varphi \right\}^{1/2 } , \label{rateqgen}\end{aligned}\ ] ] and @xmath78 \nonumber \\ & & \times \left [ ( \chi_y-\chi_x)\cos \vartheta \sin 2\varphi \right]^{-1},\end{aligned}\ ] ] where for the diagonal @xmath20 we used @xmath79 , etc . from eq .
( [ rateqgen ] ) one can find for which directions @xmath80 the squeezing rate is maximum and for which it is zero .
let us first consider the general case when the three eigenvalues of @xmath20 are all different ( fig . [ f - koule]c - f ) , e.g. , @xmath81 .
then there are four points where @xmath82 , all at the equator of the bloch sphere , @xmath83 , and @xmath84 .
the maximum of @xmath85 is achieved at the poles at @xmath86 and @xmath87 where @xmath88 , and the optimum orientation of the squeezing ellipse is exactly half - way between the @xmath89 and @xmath90 directions .
a special case of this situation is tact with @xmath91 with symmetrical squeezing geometry ( fig .
[ f - koule]c , d ) .
generally , the maximum achievable squeezing rate only depends on the difference between the maximum and minimum eigenvalues of the twisting tensor .
if @xmath20 is degenerate with , say @xmath92 , @xmath93 ( oat , fig .
[ f - koule]a , b ) , there are two zeros of @xmath85 located at @xmath83 , @xmath94 , and the maximum is achieved along the meridian @xmath95 with @xmath96 .
for simplicity we choose the coordinate system such that the initial state is centered at the pole of the bloch sphere ( @xmath97 ) oriented such that the off - diagonal term @xmath98 vanishes .
let us further assume that the frequency components @xmath99 and @xmath100 are chosen such that the state is kept centered at the pole ( this would be simply @xmath101 if @xmath102 , but otherwise a nontrivial expression for @xmath103 has to be used to compensate for the bogoliubov backreaction ) .
we introduce scaled variables @xmath104 , @xmath105 and @xmath106 as @xmath107 and get a closed set of equations @xmath108v_{xy } , \\ \frac{dv_{yy}}{d\tau } & = & 2\left [ \tilde \omega - ( \chi_{x}-\chi_{z } ) j \right]v_{xy } , \label{dvyydt } \\
\frac{dv_{xy}}{d\tau } & = & \tilde \omega \left ( v_{xx } -v_{yy } \right ) \nonumber \\ & & + j \left [ ( \chi_{z}-\chi_{x } ) v_{xx } - ( \chi_{z}-\chi_{y } ) v_{yy } \right ] , \label{dvxydt } \\
\frac{dj}{d\tau } & = & \frac{1}{n}(\chi_{x}-\chi_{y})v_{xy } .\end{aligned}\ ] ] in the limit of @xmath109 , the derivative @xmath110 approaches zero and one can take @xmath111 and solve the equations with the initial condition of the spin coherent state @xmath112 , @xmath113 .
let us assume @xmath114 and define @xmath115 in the special case of @xmath116 we find on solving eqs .
( [ dvxxdt])([dvxydt ] ) @xmath117 , \\
v_{yy } & = & \frac{1}{2\delta \chi_y}\left[(\delta
\chi_x + \delta \chi_y ) \cosh ( \delta \chi \tau ) \right
. \nonumber \\ & & \left .
+ \delta \chi_y -\delta
\chi_x \right ] , \\ v_{xy } & = & \frac{\delta \chi_x + \delta \chi_y}{2 \sqrt{\delta \chi_x \delta \chi_y } } \sinh ( \delta \chi \tau ) .
\label{eqvxy}\end{aligned}\ ] ] the squeezing parameter @xmath118 defined as the ratio of the minimum variance of the uncertainty ellipse of the state and the variance of the spin coherent state @xcite is @xmath119 using the results of eqs .
( [ eqvxx])([eqvxy ] ) we get @xmath120 ^{1/2 } \right\}. \label{xi2gen}\end{aligned}\ ] ] the results are illustrated in fig . [ f - graf]a . in the graphs ,
all lines were calculated by numerical solution of the schrdinger equation , the lines marked with symbol `` @xmath121 '' coinciding with the analytical result of eq .
( [ xi2gen ] ) .
two special cases are worth mentioning , first , in the oat with @xmath122 ( green broken line with symbol `` @xmath121 '' in fig .
[ f - graf]a ) the squeezing is @xmath123 for short times this expression drops linearly as @xmath124 , and for long times it approaches zero as @xmath125 .
the second special case is tact with @xmath126 ( blue full line with symbol `` @xmath121 '' in fig .
[ f - graf]a ) when we get @xmath127 .
although at the beginning for the two cases the squeezing evolves at the same rate given by the difference of the biggest and smallest eigenvalues of @xmath20 , for longer times it drops to zero much faster in the tact case . one should keep in mind that for longer times these results can only be used as long as the used approximations are valid ( the deviation of the exact values for various finite @xmath5 from these approximate solutions can be seen in fig [ f - graf]a ) .
so far the special case of @xmath116 has been considered . to generate squeezing at the maximum rate , one has to keep the state optimally oriented with respect to the main twisting axes , so that @xmath128 ( see eq .
( [ tan2a ] ) with @xmath129 ) .
this leads to @xmath130 and @xmath131 it follows that @xmath132 , and from eqs .
( [ dvxxdt ] ) and ( [ dvyydt ] ) that the optimum rotation frequency should satisfy @xmath133 thus , for the exact tact with @xmath134 no rotation is needed to achieve the optimum squeezing rate . for any other values of the twisting parameters one needs to keep the variance ellipse optimally oriented by means of suitable rotation frequency .
the evolution then follows from eqs .
( [ dvxxdt])-([dvxydt ] ) as @xmath135 and @xmath136 if the system starts in the spin coherent state with @xmath137 , and @xmath113 , one finds @xmath138 .
\end{aligned}\ ] ] these results are illustrated in fig .
[ f - graf]b .
note that the gaussian approximation with finite @xmath5 works for relatively short times , after which the state undergoes an @xmath139-shape deformation and the squeezing is deteriorated ( inset of fig .
[ f - graf]b ) .
these results hint for which parameters it might be suitable to apply the additional rotation during the squeezing preparation stage , e.g. , the data of @xcite ( @xmath140 ) suggest a possible room for further optimization by this means .
the aim of this paper was to offer a unified tensor approach to all quadratic squeezing schemes that have so far been treated separately , such as one - axis twisting or two - axis counter - twisting .
the main results are derivation of a closed set of equations governing the first and second moments in gaussian approximation , and showing the role of eigenvalues of the twisting tensor . at the early stages of squeezing ,
the most relevant parameter is the difference between the maximum and minimum eigenvalues of the twisting tensor which determines the squeezing rate . for longer times , most efficient
squeezing is achieved if the middle eigenvalue halves the interval between the extreme ones ( tact ) . in other cases , for certain times the imbalance of the middle eigenvalue
can be compensated by suitable rotation .
the approach is suitable for two - mode systems with quadratic nonlinearities , such as two - mode optical resonators with kerr media , bec in structured traps , collective atomic spins , etc . apart from covering various squeezing scenarios important mostly for quantum metrology and interferometry
, it is also relevant for the lmg model studied as a paradigm of quantum phase transitions .
further generalization to cover losses and decoherence @xcite and various squeezing optimization strategies will be the subject of a forthcoming work .
stimulating discussions with k. k. das , m. kol , and k. mlmer are acknowledged .
this work was supported by the european social fund and the state budget of the czech repuplic , project cz.1.07/2.3.00/30.0041 .
i. bouchoule and k. mlmer , phys .
rev . a * 65 , * 041803 ( 2002 ) ; t. opatrn and k. mlmer , phys . rev .
a * 86 , * 023845 ( 2012 ) ; l. i. r. gil , r. mukherjee , e. m. bridge , m. p. a. jones , and t. pohl , phys .
* 112 , * 103601 ( 2014 ) .
a. kuzmich , k. mlmer , and e. s. polzik , phys .
* 79 , * 4782 ( 1997 ) ; b. julsgaard , j. sherson , j. i. cirac , j. fiurek , and e. s. polzik , nature * 432 , * 482 ( 2004 ) ; l. b. madsen and k. mlmer , phys .
a * 70 , * 052324 ( 2004 ) .
b. julia - diaz , e. torrontegui , j. martorell , j. g. muga , and a. polls , phys .
a * 86 , * 063623 ( 2012 ) ; a. yuste , b. julia - diaz , e. torrontegui , j. martorell , j. g. muga , and a. polls , phys .
a * 88 , * 043647 ( 2013 ) ; t. caneva , s. montangero , m. d. lukin , and t. calarco , arxiv:1304.7195 ( 2014 ) .
y. li , y. castin , and a. sinatra , phys .
lett . * 100 * , 210401 ( 2008 ) ; b. opanchuk , m. egorov , s. hoffmann , a. i. sidorov , and p. d. drummond , europhys . lett . * 97 * , 50003 ( 2012 ) ; a. sinatra , y. castin , and e. witkowska , europhys . lett .
* 102 * , 40001 ( 2013 ) . | a unified tensor description of quadratic spin squeezing interactions is proposed , covering the single- and two - axis twisting as special cases of a general scheme .
a closed set of equations of motion of the first moments and variances is derived in gaussian approximation and their solutions are discussed from the prospect of fastest squeezing generation .
it turns out that the optimum rate of squeezing generation is governed by the difference between the largest and the smallest eigenvalues of the twisting tensor .
a cascaded optical interferometer with kerr nonlinear media is proposed as one of possible realizations of the general scheme . |
during the last years there has been great interest in applications of statistical physics to financial market dynamics .
a variety of agent - based models have been proposed over the last few years to study financial market dynamics @xcite . in particular
, spin models as the most popular models of statistical mechanics have been applied to describe the dynamics of traders in financial markets by several researchers @xcite .
a particularly simple model of a stock market in the form of a spin model motivated by the ising model has been proposed recently @xcite , in order to simulate the dynamics of expectations in systems of many agents .
the model introduces a new coupling that relates each spin ( agent ) to the global magnetization of the spin model , in addition to the ferromagnetic ( ising ) couplings connecting each spin to its local neighborhood .
the global coupling effectively destabilizes local spin orientation , depending on the size of magnetization . the resulting frustration between seeking ferromagnetic order locally ,
but escaping ferromagnetic order globally , causes a metastable dynamics with intermittency and phases of chaotic dynamics .
in particular , this occurs at temperatures below the critical temperature of the ising model . while the model exhibits dynamical properties which are similar to the stylized facts observed in financial markets , a careful interpretation in terms of financial markets is still lacking .
this is the main goal of this paper . in particular ,
@xcite treat the magnetization of the model as price signal which , however , is unnatural when deriving a logarithmic return of this quantity as these authors do . as a result ,
small magnetization values cause large signals in the returns with an exponent of the size distribution different from the underlying model s exponent
. let us here consider bornholdt s spin model in the context of a stock market with heterogeneous traders .
the aim of this paper is ( i ) to interpret the magnetization of the spin model in terms of financial markets and to study the mechanisms that create bubbles and crashes , and ( ii ) to investigate the statistical properties of market price and trading volume .
the new elements in the model studied in this paper are the explicit introduction of two groups of traders with different investment strategies , fundamentalists and interacting traders , as well as of a market clearing system that executes trading at matched book . given these conditions ,
the market price is related to the sum of fundamental price and magnetization , and the trading volume is simply given by the magnetization .
we also show that the model is able to explain the empirically observed positive cross - correlation between volatility and trading volume .
finally , we observe that the model reproduces the well - known stylized facts of the return distribution such as fat tails and volatility clustering @xcite , and study volatilities at different time - scales .
let us consider a stock market where a large stock is traded at price @xmath1 .
two groups of traders with different trading strategies , _ fundamentalists _ and _ interacting traders _ , participate in the trading . the number of fundamentalists @xmath2 and the number of interacting traders @xmath3 are assumed to be constant .
the model is designed to describe the stock price movements over short periods , such as one day . in the following
, a more precise account of the decision making of each trader type is given .
fundamentalists are assumed to have a reasonable knowledge of the fundamental value of the stock @xmath4 .
if the price @xmath1 is below the fundamental value @xmath5 , a fundamentalist tends to buy the stock ( as he estimates the stock to be undervalued ) , and if the price is above the fundamental value , a fundamentalist tends to sell the stock ( then considered as an overvalued and risky asset ) .
hence we assume that fundamentalists buying or selling order is given by : @xmath6 where @xmath2 is the number of fundamentalists , and @xmath7 parametrizes the strength of the reaction on the discrepancy between the fundamental price and the market price . during each time period , an interacting trader may choose to either buy or sell the stock , and is assumed to trade a fixed amount of the stock @xmath8 in a period of trading .
interacting traders are labeled by an integer @xmath9 .
the investment attitude of interacting trader @xmath10 is represented by the random variable @xmath11 and is defined as follows : if interacting trader @xmath10 is a buyer of the stock during a given period , then @xmath12 , otherwise he sells the stock and @xmath13 .
now let us formulate the dynamics of the investment attitude of interacting traders in terms of the spin model @xcite .
let us consider that the investment attitude @xmath14 of interacting trader @xmath10 is updated with a heat - bath dynamics according to @xmath15 where @xmath16 is the local field of the spin model , governing the strategic choice of the trader .
let us consider the simplest possible scenario for local strategy changes of an interacting trader .
we assume that the decision which an interacting trader makes is influenced by two factors , local information , as well as global information .
local information is provided by the nearby interacting traders behavior . to be definite ,
let us assume that each interacting trader may only be influenced by its nearest neighbors in a suitably defined neighborhood .
global information , on the other hand , is provided by the information whether the trader belongs to the majority group or to the minority group of sellers or buyers at a given time period , and how large these groups are . the asymmetry in size of majority versus minority groups
can be measured by the absolute value of the magnetization @xmath17 , where @xmath18 the goal of the interacting traders is to obtain capital gain through trading .
they know that it is necessary to be in the majority group in order to gain capital , however , this is not sufficient as , in addition , the majority group has to expand over the next trading period .
on the other hand , an interacting trader in the majority group would expect that the larger the value of @xmath17 is , the more difficult a further increase in size of the majority group would be .
therefore , interacting traders in the majority group tend to switch to the minority group in order to avert capital loss , e.g. to escape a large crash , as the size of the majority group increases .
in other words , the interacting trader who is in the majority group tends to be a risk averter as the majority group increases . on the other hand , an interacting agent who is in the minority group tends to switch to the majority group in order to gain capital .
an interacting agent in the minority group tends to be a risk taker as the majority group increases . to sum up ,
the larger @xmath17 is , the larger the probability with which interacting traders in the majority group ( interacting traders in the minority group , respectively ) withdraw from their coalition . following @xcite , the local field @xmath16 containing the interactions discussed above is specified by @xmath19 with a global coupling constant @xmath20 .
the first term is chosen as a local ising hamiltonian with nearest neighbor interactions @xmath21 and @xmath22 for all other pairs .
we assume that the interacting - traders excess demand for the stock is approximated as @xmath23 let us leave the traders decision - making processes and turn to the determination of the market price .
we assume the existence of a market clearing system . in the system
a _ market maker _ mediates the trading and adjusts the market price to the market clearing values .
the market transaction is executed when the buying orders are equal to the selling orders .
the balance of demand and supply is written as @xmath24 + b\ ; n\ ; m(t ) = 0 .
\label{eqn : a7}\ ] ] hence the market price and the trading volume are calculated as @xmath25 and @xmath26 using the price equation ( 8) we can categorize the market situations as follows : if @xmath27 , the market price @xmath28 is equal to the fundamental price @xmath5 .
if @xmath29 , the market price @xmath30 exceeds the fundamental price @xmath5 ( _ bull _ market regime ) .
if @xmath31 , the market price @xmath28 is less than the fundamental price @xmath5 ( _ bear _ market regime ) . using ( 8) , the logarithmic relative change of price , the so - called log - return , is defined as @xmath32 let us consider for a moment that only fundamentalists participate in trading .
then in principle the market price @xmath1 is always equal to the fundamental price @xmath5 .
this implies that the so - called _ efficient market hypothesis _ holds .
following the efficient market model by @xcite then the fundamental price @xmath4 follows a random walk .
since the continuous limit of a random walk is a gaussian process , the probability density of the log - return , defined as @xmath33 , is normal .
for real financial data , however , there are strong deviations from normality . as we discuss here , including both fundamentalists and interacting traders to coexist in the market , offers a possible mechanism for excessive fluctuations such as bull markets and bear markets . to investigate the statistical properties of the price and the trading volume in the spin model of stock markets , we will assume for simplicity that the fundamental price is constant over time .
in the new framework developed so far we see that the dynamics of the log - return corresponds to the linear change in absolute magnetization of the spin model @xcite .
typical behavior of the such defined return @xmath34 as well as the magnetization @xmath35 is shown in figures 1(a ) and 1(b ) . here , a 101 * 101 lattice of the general version of the model as defined in eq .
( 8) and with the spins updated according to ( 5 ) is shown .
it is simulated at temperature @xmath36 with couplings @xmath37 and @xmath38 , using random serial and asynchronous heat bath updates of single sites . in figure 1(a )
the intermittent phases of ordered and turbulent dynamics of the log - return are nicely visible .
qualitatively , this dynamical behavior is similar to the dynamics of daily changes of real financial indices , as for example the dow jones index shown in figure 1(c ) . to some degree
, these transitions of the return can be related to the magnetization in the spin model .
figure 1 ( b ) illustrates that the bull ( bear ) market @xmath39 ( @xmath40 ) becomes unstable , and the transition from a metastable phase to a phase of high volatility occurs , when the absolute magnetization @xmath17 approaches some critical value . noting that trading volume is defined as @xmath41
, this suggests that also some critical trading volume exists near the onset of turbulent phases .
this is in agreement with the empirical study by @xcite who found the empirical regularities : ( i ) positive correlation between the volatility and the trading volume ; ( ii ) large price movements are followed by high trading volume .
the origin of the intermittency can be seen in the local field @xmath42 eq .
( 5 ) representing the external influences on the decision - making of interacting - trader @xmath10 . in particular , the second term of @xmath42 tends to encourage a spin flip when magnetization gets large .
thus each interacting - trader frequently switches his strategy to the opposite value if the trading volume gets large . as a consequence ,
the bull ( bear ) market is unstable and the phase of the high volatility appears .
the metastable phases are the analogue of speculative bubbles as , for example , the bull market is defined as a large deviation of the market price from the fundamental price .
in fact it is a common saying that `` it takes trading volume to move stock price '' in the real stock market .
a typical example is the crash of oct .
1987 , when the dow jones industrial average dropped 22.6% accompanied by an estimated @xmath43 shares that changed hands at the new york stock exchange alone @xcite .
fama @xcite has argued that the crash of oct .
1987 at the us and other stock markets worldwide could be seen as the signature of an efficient reassessment of and convergence to the correct fundamental price after the long speculative bubble proceeding it .
it is interesting to investigate how long the bull or bear markets last from the point of view of practical use .
figure 2 shows the distribution of the bull ( bear ) market durations that is defined as the period from the beginning of a bull ( bear ) market @xmath44 to the end of the bull ( bear ) market @xmath45 . in other words ,
the bull ( bear ) market duration means the period from a point of time that the market price exceeds the fundamental price to the next point of time that the market price falls short of the fundamental price . in the model one
observes that the bull ( or bear ) market durations are power law distributed with an exponent of approximately @xmath46 . as shown in the previous works @xcite and @xcite the simple spin model considered here reproduces major stylized facts of real financial data .
the actual distribution of log - return @xmath34 has _ fat tails _ in sharp contrast to a gaussian distribution ( figure 3 ) .
that is , there is a higher probability for extreme values to occur as compared to the case of a gaussian distribution .
recent studies of the distribution for the absolute returns @xmath47 report power law asymptotic behavior , @xmath48 with an exponent @xmath49 between about 2 and 4 for stock returns .
figure 4 shows the cumulative probability distribution of the absolute return that is generated from the model .
the observed model exponent of @xmath50 lies in the range of empirical data .
let us consider a time - scale @xmath53 at which we observe price fluctuations .
the log - return for duration @xmath53 is then defined as @xmath54 .
volatility clustering as described in the previous section is this observable defined for an interval @xmath53 ranging from several minutes to more than a month or even longer . in this intermediate and strongly correlated regime of time - scales ,
volatilities at different time - scales may show self - similarity @xcite .
self - similarity in this context states that volatilities @xmath55 at different time - scales @xmath53 are related in such a way that that the ratio of volatilities at two different scales does not statistically depend on the coarse - graining level .
thus daily volatility is related to weekly , monthly volatilities by stochastic multiplicative factors .
the self - similarity has been shown to be equivalent to scaling of moments under some conditions @xcite@xcite@xcite .
scaling occurs when @xmath56 , where @xmath57 is the scaling function which is related to the statistical property of the multiplicative factors @xcite .
figure 6 ( a ) depicts the presence of self - similarity in actual data of the nyse stock index .
though it is not straightforward to relate time - scales between the simulation and real data , it is interesting to look at how the volatilities at different time - scales behave in the regime where the volatility clustering is valid . figure 6 ( b ) shows the scaling of moments from the data of log - returns calculated for different scales in time - steps of the simulation .
we observe a range where self - similarity dominates and that this property is broken at some time - scale , which corresponds to the scale where volatility clustering as seen in figure 5 deviates from a power - law of the autocorrelation function .
this observation is encouraging as it might help relate the time - scales of simulations and real markets .
in this paper we have considered the spin model presented in @xcite in the context of a stock market with heterogeneous traders , that is , fundamentalists and interacting traders .
we have demonstrated that magnetization in the spin model closely corresponds to trading volume in the stock market , and the market price is determined by magnetization under natural assumptions . as a consequence
we have been able to give a reasonable interpretation to an aperiodic switching between bull markets and bear markets observed in bornholdt s spin model . as a result
, the model reproduces main observations of real financial markets as power - law distributed returns , clustered volatility , positive cross - correlation between volatility and trading volume , as well as self - similarity in volatilities at different time - scales .
we also have found that scaling is observed in the distribution of transition times between bull and bear markets .
although the power law scaling of the distribution has never been examined empirically on short time scales , the power law statistics showed here is not only an interesting finding theoretically but presumably also useful to measure the risk of security investments in practice .
the empirical study will be left for future work .
k. sznajd - weron and r. weron , a simple model of price formation , int . j. mod .
c * 13 * 115 ( 2002 ) .
t. yamano , bornholdt s spin model of a market dynamics in high dimensions , int .
j. mod .
c * 13 * 89 ( 2002 ) .
g. iori , a microsimulation of traders activity in the stock market : the role of heterogeneity , agents interactions and trade frictions , www.arxiv.org/abs/adap-org/9905005 ( 2001 ) ; journal of economic behavior and organization , vol .
49 , no . 1 to appear in 2002 .
fama , perspectives on october 1987 , or , what did we learn from the crash ? , in r. barro , et .
al . , black monday and the future of financial markets , dow jones - irwin , inc . , homewood , illinois . 71 ( 1989 ) . , defined as change in magnetization @xmath35 in the spin model .
( b ) magnetization @xmath35 of the spin model .
( c ) for comparison with ( a ) , the log - return of the dow jones daily changes 1896 - 1996 is shown.,height=793 ] -th order ) of volatilities @xmath59 at time - scales @xmath53 .
( a ) for a stock in nyse with @xmath53 in minutes .
( b ) for the simulation result with @xmath53 in time - steps . in both plots
, @xmath60 ranges from 0.5 ( cross ) to 4.0 ( triangles downward ) with increment 0.5.,width=529 ] | the dynamics of a stock market with heterogeneous agents is discussed in the framework of a recently proposed spin model for the emergence of bubbles and crashes .
we relate the log returns of stock prices to magnetization in the model and find that it is closely related to trading volume as observed in real markets .
the cumulative distribution of log returns exhibits scaling with exponents steeper than @xmath0 and scaling is observed in the distribution of transition times between bull and bear markets . , econophysics , stock market , spin model , volatility 89.90.+n , 02.50.-r , 05.40.+j |
after the experimental realization of bose - einstein condensation in vapors of alkali atoms @xcite , the understanding of their behavior under rotation became the main focus of theoretical and experimental work .
experimental observation of the vortex states in a two - component system@xcite , and the realization of the vortex states in a stirred one - component bose - einstein condensate @xcite confirmed the expectations that sufficiently large condensates exhibit bulk superfluid properties by supporting their rotation above the critical angular frequency by the feynman - onsager vortex state@xcite .
the theoretical approaches to the bose - einstein condensed gases commonly distinguish between the two regimes of the strong and the weak interparticle interactions , which correspond to the conditions @xmath10 and @xmath11 , respectively , where @xmath12 is the number of particles in the gas , @xmath13 is the scattering length , and @xmath14 is the oscillator length defined in terms of the oscillator frequency @xmath1 and the boson mass @xmath15 ( for a review see e.g. @xcite . ) in the weak coupling regime , where the scale of the variation of the condensate wave - function is of the same order as the size of the cloud , quantum effects play an essential role and the system behaves much like an atomic nucleus ( see e. g. ref .
@xcite ) . in the strong coupling regime and for sufficiently large condensate , the condensate is well described in terms of hydrodynamics . under these conditions the time - dependent gross - pitaevskii ( g - p ) equation@xcite for the ( complex ) condensate wave function reduces to two ( real ) hydrodynamic equations for the density and velocity of the condensate at zero temperature .
one consequence of the assumption that the system is sufficiently large is that the quantum pressure term can be ignored compared with the interaction terms .
another is that the thomas - fermi approximation@xcite is valid and the equilibrium kinetic energy density is negligible compared to the potential energy density .
this approximation also ensures that the thermodynamic quantities like the pressure or the chemical potential of the gas are well defined at any local point of the gas cloud .
finally , the coherence length of the condensate @xmath16 , with @xmath17 being the number density of condensate particles , becomes sufficiently small in this limit to allow for a description of a vortex core as a singularity .
the superfluid hydrodynamic approach has been used in the studies of the collective excitations of the condensate clouds@xcite the scissors mode oscillations@xcite , the moment of inertia of a rotating condensate@xcite , mainly in the irrotational regime ; see however ref .
. the critical angular velocity of condensate , @xmath18 , at which the creation of a vortex is energetically favorable , scales , in units of the trap frequency , as @xmath19 , where @xmath20 is the size of the cloud ; the numerical coefficient depends on the details of the geometry of the trap@xcite . for a sufficiently large condensate the ground state would correspond to coarse - grained rigid - body rotation supported by a vortex lattice . in this case , the irrotational constraint on the average superfluid velocity @xmath21 is replaced by the rigid - body rotation condition @xmath22 .
the purpose of this paper is the study of the oscillations of the trapped bose condensates under rigid - body rotation .
we shall not discuss the modes related to the vortex structure itself ( e.g. tkachenko modes ) .
we shall confine ourselves to the zero - temperature limit , in which case the effects of mutual friction due to the interaction of the thermal excitations with the vortex state , and dissipation due to the viscosity of the thermal component both can be neglected .
below , we shall apply the tensor virial method originally developed for the study of equilibrium and stability of rotating liquid masses bound by self - gravitation@xcite .
the tensor virial method transforms the local hydrodynamical equations into global virial equations that contain the full information on the structure and stability of a system as a whole .
the method is especially useful for studying perturbations of incompressible uniform ellipsoids from equilibrium , in which case the each perturbed virial equation yields ( in the absence of viscous dissipation ) a different set of normal modes .
the extension to compressible flows is straightforward in the case of adiabatic perturbations . moreover ,
when the equilibrium is sustained by an external confining potential , fluid perturbations do not back - react ( change ) the confining potential itself , and the tensor virial method can be extended to non - uniform compressible flows .
the description is particularly useful for gases with polytropic equations of state @xmath23 , where @xmath24 is the pressure , @xmath25 is the density , and @xmath26 is the adiabatic index . in that case , we shall see that all of the modes of the gas can be found from the tensor virial equations .
( we illustrate this only for the second harmonic modes of oscillation in this paper , but the extension to higher order modes is straightforward , and will be discussed elsewhere . ) for an interacting bose gas the pressure is a non - analytic function of the diluteness parameter @xmath27 .
however the equation of state to the leading order in the small parameter @xmath28 can be written in a polytropic form : @xmath29 , with @xmath30 .
since the zero temperature equations of motion of a trapped rotating condensate turn out formally identical to the corresponding equations of motion of a ordinary ( non - superfluid ) liquid , our results might be of significance in a broader context .
on the other hand , present results can serve as a starting point for an extension to finite temperatures where , in addition to the superfluid , the fluid of normal quasiparticle excitations plays a role .
the paper is organized as follows . in
2 we derive the first and second order virial equations for trapped bose - condensed gases in the hydrodynamic regime .
the equilibrium shape of rotating condensates and the second harmonic modes of oscillations in axisymmetric traps are discussed in 3 .
4 discussed the second harmonic modes of oscillations in non - axisymmetric traps .
our results are summarized in
the appendix gives a brief derivation of the hydrodynamic equations from the g - p equation .
the euler equation for a condensate in a harmonic , arbitrary deformed trap , can be written as ( t+u_j x_j ) u_i =- px_i- + + 2_ilmu_l_m , [ eq : euler ] where the latin subscripts denote coordinate directions ; @xmath25 , @xmath24 , and @xmath31 are the density , pressure , and velocity of condensate ; ( summation over repeated indexes is assumed ) .
the external harmonic trapping potential is _ tr ( ) = _ 0 ^ 2_i=1 ^ 3 a_i x_i^2 , where @xmath32 are the dimensionless deformation parameters , @xmath1 is the frequency of the harmonic oscillator in the trap .
( [ eq : euler ] ) is written in a frame rotating with angular velocity @xmath33 relative to some inertial coordinate reference system .
its derivation from the g - p equation for the condensate wave - function is given in the appendix .
our fundamental assumption is that the condensate undergoes a rigid - body rotation , which is supported by an array of vortices . in this case , as explained in the appendix , the equations of motion are formally identical to those of a ordinary compressible ( non - superfluid ) liquid . starting from eq .
( [ eq : euler ] ) we can construct a set of virial equations of various order .
suppose that the condensate occupies a volume @xmath34 .
taking the zeroth moment of eq .
( [ eq : euler ] ) amounts to integrating over @xmath34 ; doing so , we obtain ( _ vd^3xu_i ) & = & 2_ilm_m_vd^3xu_l+(^2_ij-_i_j ) _
vd^3xx_j + & & -_vd^3x .
[ eq : v1 ] the first order virial equation describes the center of mass motion of the condensate , and is trivial as these motions correspond to a uniform translation of the system as a whole . taking the first moment of eq .
( [ eq : euler ] ) results in the second order virial equation ( _ vd^3xx_j u_i ) & = & 2_ilm_m(_v d^3xx_ju_l ) + ( ^2 - _ 0 ^ 2 a_i ) i_ij-_i_ki_kj + & & + 2t_ij+_ij , [ eq : v2 ] where i_ij_vd^3xx_ix_j , _ vd^3xp , _ ij_vd^3xu_iu_j .
consider the variation of the second order virial equation under the influence of perturbations .
the eulerian variations of various terms in eq .
( [ eq : v2 ] ) are straightforward@xcite and we find _ vd^3x u_i x_j = ( ^2 -_0 ^ 2 a_i ) v_ij - _ i_k v_kj + 2 _ ilm_m _ v d^3x u_l x_j+2_ij+ _ ij , where for a lagrangian displacement @xmath35 v_ij i_ij = _ v d^3x ( _ i x_j + _ j x_i ) .
the tensors @xmath36 are manifestly symmetric in their indexes , as is evident from their definition .
it is useful to define also their non - symmetric parts as v_i;j = _ v d^3x _
i x_j , such that @xmath37 .
when there are no condensate motions in the unperturbed state in the rotating frame ( @xmath38 ) , the second order virial equation becomes & = & 2_ilm _ m dv_l;jdt + ( ^2 -_0 ^ 2 a_i ) v_ij-_i_kv_kj + _ ij .
[ eq : v3 ] the tensor virial eq .
( [ eq : v3 ] ) can be used for the study of small amplitude oscillations of a trapped condensate . for time - dependent lagrangian displacements of the form ( x_i , t ) = ( x_i ) e^t eq .
( [ eq : v3 ] ) becomes ^2 v_i;j - 2_ilm _ m v_l;j = ( ^2 -_0 ^ 2 a_i ) v_ij-_i_kv_kj + _ ij .
[ eq : v4 ] eq .
( [ eq : v4 ] ) contains all the second harmonic modes of the rotating condensate in a arbitrary harmonic trap .
next , we specialize eq .
( [ eq : v2 ] ) to the case of axisymmetric traps , @xmath39 and assume , without loss of generality , @xmath40 .
the assumption of the axisymmetry is , in practice , an approximation as a perfectly axisymmetric trap will not exert a torque on the condensate .
our assumption is that the deviations from the axisymmetry , which drive the rotation of the condensate , are small compared with the deformation of the trap @xmath41 . in principle , once the condensate is brought to rotation the axisymmetry can be restored .
consider a equilibrium state in which the condensate rotates uniformly with the rotation vector along the @xmath42 axis of cartesian system of coordinates . then eq .
( [ eq : v2 ] ) reduces to ^2 ( i_ij - _ i3 i_3j ) -_0 ^ 2 a_i i_ij = -_ij .
[ eq : eqbr ] in axisymmetric traps with @xmath43 denoting the ellipsoidal semi - axis along the rotation , the two remaining semi - axis of the ellipsoid , @xmath44 and @xmath45 , are equal .
the diagonal component of eq .
( [ eq : eqbr ] ) provides the relation between the rotation frequency , the deformation parameters along the @xmath42 direction ( both are fixed in an experiment ) and the ratio of the semi - axis of the resulting spheroidal figure : [ eq : figures ] = ( 1- ) .
depending whether @xmath46 or @xmath47 the condensate shape is either oblate or prolate .
note that the rotation frequency is bounded from above , @xmath48 , and for @xmath49 the equilibrium figures are always oblate . a striking consequence of the eq . ( [ eq : figures ] ) , when @xmath50 , is an equilibrium figure which is a rotating sphere@xcite .
although we had to assume small deviations from axisymmetry to make the condensate rotate , this assumption is a prerequisite for producing rotation , but not an intrinsic property of the rotating condensate .
the equilibrium density profile of the condensate , for a polytropic equation of state , can be obtained by a direct integration of the unperturbed limit of eq .
( [ eq : euler ] ) : [ profile ] ( ) = _ 0 \{1- } ^ , where @xmath51 is the central density of the cloud , the dots in the argument of the @xmath52-function ( which insures that the condensate density is positive ) stand for the expression in the curly brackets . here
we have retained the parameter @xmath53 to incorporate the case of non - axisymmetric traps which is treated in the next section .
the effect of the rotation is the stretching of the profile of the condensate in the equatorial plane due to the centrifugal potential . according to eq .
( [ profile ] ) , the density is constant on concentric ellipsoids in the unperturbed rotating background .
note that the normalization of the wave - function of the condensate implies _ vd^3x ( ) = m , where @xmath54 is the total mass in the cloud .
next consider perturbations from the equilibrium state of uniform rotation , with the spin vector along the @xmath42 axis .
surface deformations related to various modes can be classified by corresponding terms of the expansion in surface harmonics labeled by indexes @xmath55 .
second order harmonic deformations correspond to @xmath4 with five distinct values of @xmath15 , @xmath56 .
the 18 equations represented by eq .
( [ eq : v4 ] ) separate into two independent subsets which are odd and even with respect to the index 3 .
the corresponding oscillation modes can be treated separately .
these modes correspond to the surface deformations with @xmath57 and represent relative shearing of the northern and southern hemispheres of the spheroid .
the components of eq .
( [ eq : v4 ] ) , which are odd in index 3 , are : [ mac_modes:1.1 ] ^2 v_3;1 & = & - a_3 v_31 , + [ mac_modes:1.2 ] ^2 v_3;2 & = & - a_3 v_32 , + [ mac_modes:1.3 ] ^2 v_1;3 - 2 v_2;3 & = & -v_13 + ^2 v_13 , + [ mac_modes:1.4 ] ^2 v_2;3 + 2 v_1;3 & = & -v_23 + ^2 v_23
. we sum the eqs .
( [ mac_modes:1.1 ] ) , ( [ mac_modes:1.3 ] ) and ( [ mac_modes:1.2 ] ) , ( [ mac_modes:1.4 ] ) , respectively , and use the symmetry properties of @xmath36 combined with eqs . ( [ mac_modes:1.2 ] ) , ( [ mac_modes:1.4 ] ) .
this results in the relations & & ( ^2 + a_3 + 1 -^2 ) v_13 - 2 ( ^2
+ a_3 ) v_23 = 0 , [ mac_modes:1.5 ] + & & ( ^2 + a_3 + 1 - ^2 ) v_23 -2(^2 + a_3 ) v_23 = 0 .
[ mac_modes:1.6 ] the characteristic equation can be factorized by substituting @xmath58 and we find [ eq : rslt1 ] ^3 - 2 ^ 2 - ( 1 + a_3 - ^2 ) + 2 a_3 = 0 .
the roots are give by [ sig1 ] _ 1 = + ( s_++s_- ) , _ 2,3 = - ( s_+ + s_- ) ( s_+-s_- ) , where [ sig_coff ] s_^3 & = & ( 1 - 2 a_3 - ) ^1/2 .
three complementary modes follow from eqs .
( [ sig1])-([sig_coff ] ) via the replacement @xmath59 .
it can be verified that the condition ^2 -^3 0 , is satisfied for any @xmath60 and @xmath61 , therefore all three roots are real .
the real frequencies of the transverse - shear modes are shown in fig .
1 as a function of @xmath60 .
note that in the zero - temperature limit , to which present analysis is restricted , these modes are purely real , i. e. represent undamped oscillations .
these modes correspond to @xmath62 and the motions in this case are confined to the planes parallel to the equatorial plane .
the components of eq .
( [ eq : v4 ] ) , which are even in index 3 , are : [ mac_modes:2.1 ] ^2v_3;3&= & - a_3 v_33 , + [ mac_modes:2.2 ] ^2v_1;1 - 2v_2;1 & = & + ( ^2 - 1 ) v_11 , + [ mac_modes:2.3 ] ^2v_2;2 + 2v_1;2 & = & + ( ^2 - 1 ) v_22 , + [ mac_modes:2.4 ] ^2v_1;2 - 2v_2;2 & = & ( ^2 - 1 ) v_12 , + [ mac_modes:2.5 ] ^2v_2;1 + 2v_1;1 & = & ( ^2 - 1 ) v_21 .
we add eqs .
( [ mac_modes:2.4 ] ) and ( [ mac_modes:2.5 ] ) , and subtract eqs .
( [ mac_modes:2.2 ] ) and ( [ mac_modes:2.3 ] ) to find the following coupled equations : [ mac_modes:2.6 ] ( v_11-v_22 ) -4 v_12 & = & 0 , + [ mac_modes:2.7 ] v_12 + ( v_11-v_22 ) & = & 0 .
the characteristic equation for the toroidal modes is [ mac_modes:2.8 ] & & ^2 + 4 ^2 ^ 2 = 0 , which is factorized by writing @xmath63 .
two solutions are then [ eq : sig_tor ] _ 1,2=. there are two complementary modes which are found by substituting @xmath64 for @xmath65 .
since the rotation frequency is bounded ( @xmath61 ) the toroidal modes are stable independent of the magnitude of the deformation in the equatorial plane .
for the same reason these modes are undamped in the zero - temperature limit considered here . to find the pulsation modes , which correspond to @xmath5 , we first add the eqs .
( [ mac_modes:2.2])-([mac_modes:2.3 ] ) , and then use the eq .
( [ mac_modes:2.1 ] ) to eliminate @xmath66 in the result . in this manner
we find that [ mac_modes:3.1 ] ( ^2/2-^2 + 1)(v_11+v_22 ) + 2 ( v_1;2-v_2;1 ) -(^2 + 2 a_3 ) v_33=0 .
subtracting eqs .
( [ mac_modes:2.5 ] ) and ( [ mac_modes:2.4 ] ) ( and discarding the @xmath67 root ) one finds [ mac_modes:3.2 ] & & ( v_1;2-v_2;1 ) - ( v_11+v_22 ) = 0 . eqs .
( [ mac_modes:3.1 ] ) and ( [ mac_modes:3.2 ] ) can be further combined to a single equation : [ mac_modes:3.3 ] & & ( ^2 + 2 ^2 + 2 ) ( v_11+v_22 ) -2(^2 + 2 a_3 ) v_33 = 0 .
it is instructive first to consider the case where the superfluid is incompressible . then the eq .
( [ mac_modes:3.3 ] ) should be supplemented by the divergence free condition + = 0 .
[ div ] again , writing @xmath58 , we find for the square of frequency of the pulsation mode in the incompressible limit _ 0
^ 2=(+)^-1 ( 1+^2 + 2 a_3 ) .
[ incompuls ] as @xmath68 is always positive , these modes correspond to undamped stable oscillations . for compressible fluids we need the variation of the pressure tensor , which for adiabatic perturbations can be written as = ( -1 ) d^3x _ i _ i p , where we assumed a polytropic equation of state @xmath69 ; the polytropic index for a bose gas to leading order in the parameter @xmath70 is equal 2 .
to evaluate the gradient of the pressure we turn to the euler eq .
( [ eq : euler ] ) in the unperturbed limit , = - + , [ eq : euler : eq ] where the enthalpy @xmath71 is defined by @xmath72 and is simply @xmath73 for @xmath74 . substituting the explicit expression for the trapping potential
, we find that the gradient of the enthalpy is a linear function of the coordinates and , hence , the variation of the pressure tensor can be expressed in terms of virials @xmath75 : = -(-1)2[(v_11+v_22)(1-^2)+a_3v_33 ] .
[ delpi ] any of the equations which are even in the index 3 now can be used to close the system of equations for @xmath76 and the virial combination @xmath77 . substituting eq .
( [ delpi ] ) for @xmath78 in , e.g , eq . ( [ mac_modes:2.1 ] ) , one finds [ mac_modes:3.4 ] [ ^2+(+1)a_3]v_33+(-1)(1-^2)(v_11+v_22)=0 . eqs .
( [ mac_modes:3.3 ] ) and ( [ mac_modes:3.4 ] ) completely determine the unknown virials ; the characteristic equation for the pulsation modes , which is quadratic in @xmath79 , is ^4+^2[(a_3 + 2 - 2 ^ 2)+a_3 + 4 ^ 2 ] + 2a_3[(3-^2)+3 ^ 2 - 1]=0 . on substituting @xmath58 ,
the solution of the resulting quadratic equation becomes [ eq : puls ] ^2_&= & + & & .
it is easy to see that there are only unstable modes if @xmath80 ; otherwise all modes are stable . in particular , all modes are stable for @xmath7 .
note , too , that there are twice as many solutions as were found in the incompressible case ; eq . (
[ incompuls ] ) can be recovered from eq .
( [ eq : puls ] ) by taking the @xmath81 limit of @xmath82 .
the origin of the additional root , @xmath83 , may be traced to the @xmath79 dependence of eq .
( [ mac_modes:3.4 ] ) , which only reduces to the incompressibility condition , eq .
( [ div ] ) , in the @xmath81 limit if it is assumed that @xmath84 .
the square of the pulsation modes as a function of the @xmath60 is plotted in fig . 3 for @xmath85 , the polytropic index of a dilute bose gas .
when the symmetry with respect to the rotation axis is broken , the equilibrium constraint eq .
( [ eq : figures ] ) needs to be supplemented by a relation fixing the semi - axis ratios in the plane orthogonal to the spin axis .
the diagonal components of eq .
( [ eq : eqbr ] ) provide the triangle relations which determine the non - axisymmetric equilibrium figure a_1 ^ 2(^2-_0 ^ 2a_1 ) = a_2 ^ 2(^2-_0 ^ 2a_2 ) = -a_3_0 ^ 2 a_3 .
these simultaneous constraints can be written in a equivalent form [ eq : nonfigures ] = ( 1- ) , = ( a_2- ) .
note that in addition to the upper bound on the rotation frequency set by the first relation ( as in the case of the axisymmetric traps ) , the second relations places a lower bound on the deformation in the plane orthogonal to the spin axis : @xmath86 .
given the experimentally controlled values of @xmath65 , @xmath60 and @xmath87 , relations ( [ eq : nonfigures ] ) determine , in a unique manner , the semi - axis ratios of the resulting figure .
the non - axisymmetric modes can be found from eq .
( [ eq : v4 ] ) in a manner similar to the axisymmetric modes ; however , now the degeneracy in indexes 1 and 2 should be relaxed .
the oscillation modes separate into two non - combining groups which have even or odd parity with respect to the index 3 .
below , we shall treat these modes separately . among the four components of eq .
( [ eq : v4 ] ) , which are odd in index 3 , three are identical to eqs .
( [ mac_modes:1.1 ] ) , ( [ mac_modes:1.2 ] ) , and ( [ mac_modes:1.3 ] ) under non - axisymmetric conditions ; the component which is modifed reads [ jac_modes:1.4 ] ^2 v_2;3 + 2 v_1;3 & = & -a_2v_23 + ^2 v_23 .
summing eqs .
( [ mac_modes:1.1 ] ) , ( [ mac_modes:1.3 ] ) , and ( [ mac_modes:1.2 ] ) , ( [ jac_modes:1.4 ] ) , we arrive at & & ( ^2 + a_3 + 1 -^2 ) v_13 - 2 ( ^2 + a_3 ) v_23 = 0 , [ jac_modes:1.5 ] + & & ( ^2 + a_3 + a_2 - ^2 ) v_23 -2(^2 + a_3 ) v_23 = 0 .
[ jac_modes:1.6 ] the sixth order characteristic equation derived from this algebraic system is ^6 + ^4 + ^2 + 4 a_3 ^ 2 ^2=0 .
the corresponding modes appear as conjugate pairs , i.e. there are only 3 distinct modes .
the real parts of these modes ( the modes are purely real ) are show in fig .
3 as a function of @xmath60 for several values of @xmath65 and fixed ratio @xmath88 .
the even parity components of the eq .
( [ eq : v4 ] ) which are given by the eqs .
( [ mac_modes:2.1 ] ) , ( [ mac_modes:2.2 ] ) , and ( [ mac_modes:2.4 ] ) remain unchanged when axisymmetry is relaxed ; the remainder equations read [ jac_modes:2.3 ] ^2v_2;2 + 2v_1;2 & = & + ( ^2 - a_2 ) v_22 , + [ jac_modes:2.4 ] ^2v_2;1 + 2v_1;1 & = & ( ^2-a_2 ) v_21 . in the incompressible limit
these equations should be supplemented by the condition ( [ div ] ) ( with a obvious modification of the second term ) . in the compressible case the variations of the pressure become = -(-1)2[(1-^2)v_11 + ( a_2-^2 ) v_22+a_3v_33 ] .
[ delpi2 ] using this relation in eq .
( [ mac_modes:2.1 ] ) we find [ eq : delpi_bis ] v_33 + ( -1)=0 . equations ( [ mac_modes:2.1 ] ) , ( [ mac_modes:2.2 ] ) , ( [ mac_modes:2.4 ] ) and ( [ jac_modes:2.3]),([jac_modes:2.4 ] ) can be manipulated to the the following set v_11 -v_22 - 4v_12=0 , + v_11 + v_22 - 2(^2 + 2a_3 ) v_33 + 4(v_1;2-v_2;1)=0 , + ( ^2 + 1 + a_2 - 2 ^ 2)v_12 + ( v_11-v_22)=0 , + ^2(v_1;2-v_2;1 ) -(v_11+v_22)+(1-a_2)v_12=0 , which should be supplemented with eq .
( [ eq : delpi_bis ] ) .
the corresponding characteristic equation is of order 8 and has been solved numerically .
the results for the real parts of the modes ( which are purely real ) are shown in fig .
4 as a function of @xmath60 for several values of @xmath65 for andx fixed ratio @xmath88 .
we have analyized the hydrodynamic oscillations of bose - condensed atomic clouds at zero temperature in the thomas - fermi approximation .
the equilibrium shape of the cloud in a axisymmetric trap represent either a prolate or oblate spheroid of revolution , which , for a particular choice of the rotation rate and trap potential , degenerates into a rotating sphere .
the rotation frequency of the condensate is bounded from above by the characteristic frequency of the harmonic oscillator in a given trap ( @xmath89 ) .
we have also analyized non - axisymmetric , triaxial ellipsoidal figures , which admit equilibrium solutions under additional constrain on the deformation in the plane orthogonal to the rotation axis @xmath90 .
small amplitude oscillations have been derived for linear perturbations from the rigidly rotating equilibrium background state .
the oscillations in axisymmetric traps , which are related to the transverse - shear and toroidal modes are found to be stable for all values of the trap deformation and its rotation frequency .
these modes represent undamped oscillations ( all eigenfrequencies are real ) in the absence of dissipation , i.e. , they are dynamically stable
. a dynamical instability against the quasi - radial pulsation mode can arise when @xmath91 . otherwise the system is stable , including the limit of the incompressible superfluid ( @xmath92 , independent of the shape of equilibria . for a bose gas with
@xmath85 the above stability condition is satisfied ; hence , these modes represent stable oscillations independent of the rotation rate and deformation .
numerical results for the oscillations in non - axisymmetrical traps show that , both , the even and odd partity second harmonic modes of oscillations of triaxial clouds are stable for the adiabatic index @xmath85 relevant to the interacting dilute bose gas . our results for axisymmetric traps ( in which case analytical results are available ) reduce to those of ref .
@xcite when the rotation frequency is set to zero and @xmath93 . indeed , from eqs .
( [ eq : rslt1 ] ) , ( [ eq : sig_tor ] ) and ( [ eq : puls ] ) , which describe to the oscillations with @xmath94 , @xmath95 and @xmath96 , respectively , for @xmath97 , we find ( l = 2 , m = 1 ) & = & , + ( l = 2 , m = 2 ) & = & + ( l = 2 , m = 0 ) & = & 2 + , which coincide with eqs . ( 22 ) , ( 23 ) and ( 24 ) of ref .
@xcite on making the appropriate changes in the parameters describing the trapping potential .
it is instructive also to compare the results above to the classical analysis of the equilibrium and stability of the self - gravitating fluids@xcite in view of speculations@xcite that intense off - resonant laser beams can give rise to a gravitational - type potential between the condensate particles leading to self - bound configurations .
self - gravitating systems , in particular the axisymmetric figures , are stable against transverse - shear and pulsation modes in the incompressible limit ; the same is found in the above model of trapped rotating condensates . however , in the compressible case , the self - gravitating fluids are unstable against the pulsation modes whenever the adiabatic index @xmath98 ( the precise value of the critical @xmath26 depends on the rotation rate ) and are stable otherwise . in the present case , the system is unstable when @xmath99 , where the rotation frequency covers the range @xmath100 .
another major difference between the two systems is their stability against the toroidal modes .
the axisymmetric self - gravitating fluids are unstable dynamically ( i.e. in the absence of dissipation ) beyond the point @xmath101 , as the deformation is increased . at a smaller deformation @xmath102 , these oscillations become neutral , which is a prerequisite of the onset of secular ( i.e. driven by the viscosity ) instability . in contrast , the toroidal modes of the trapped condensates are always dynamically stable and there are no neutral points within the allowed parameter space where equilibrium figures exist . the present model can be extended for a study of the higher order ( @xmath103 ) harmonic oscillations by constructing higher order virial equations as well as to finite temperatures , in which case the viscosity of thermal excitation , hence secular instabilities , and mutual friction between the vortex lattice state and the excitations play a role .
we acknowledge the generous hospitality and support of the european center for theoretical studies in nuclear physics and related areas ( ect@xmath104 ) , trento , were this work was carried out .
we are grateful to professor b. mottelson for a helpful conversation .
acknowledges the support of the nederlandse organisatie voor wetenschappelijk onderzoek via the stichting voor fundamenteel onderzoek der materie .
i.w . acknowledges partial support from nasa .
as is well - known , dilute bose gases can be described in the mean - field approximation in terms of the gross - pitaevskii ( g - p ) theory . the latter theory ( in analogy to the phenomenological theories of superconductivity and of superfluid he@xmath105 near the @xmath106-point ) , is based on a ginzburg - landau type functional for the wave - function @xmath107 of the coherent state , whose variation provides the equation of motion for @xmath107 . in the stationary case
this functional for a dilute bose gas has the well - known form e ( ) = d^3x , where @xmath108 .
the g - p equation is obtained by taking functional derivative with respect to @xmath109 subject to the constarint that the particle number @xmath12 is constant . the extremum condition @xmath110 gives [ gp ] - * * ( ) + ( ) = ( ) , where the lagrangian multiplier @xmath111 has the meaning of the chemical potential of particles . the time - dependent generalization of ( [ gp ] ) follows on the assumption that the temporal variations of @xmath107 should be described by a first order equation which , by analogy with the quantum mechanics , is written as [ tdgp ] i = -**(,t ) + ( , t ) .
on writing @xmath112 the superfluid density and velocity can be expressed as [ subst ] ( , t ) = m(,t)^2 = m(,t)^2 , ( , t ) = ( , t ) .
the real part of eq .
( [ tdgp ] ) after deviding by @xmath113 , applying a gradient , and multiplying the result by @xmath25 becomes [ eq : preeuler ] ( + ) + ( - + _
tr+ u_0 ^ 2 ) = 0 , which is the euler equation for the condensate in the zero - temperature , inviscid limit .
it differs from the analogous equation for the ordinary fluids by the `` quantum pressure '' term @xmath114 .
note that any constant term can be added to the second bracket , for example , the chemical potential in the ground state . the imaginary part of eq .
( [ tdgp ] ) , on multiplying by @xmath115 , leads to the mass conservation equation [ eq : euler_rho ] + = 0 .
when gradients of the @xmath107 are small , the `` quantum pressure '' term in eq .
( [ eq : preeuler ] ) can be dropped and it reduces to [ eq : euler2 ] ( + ) = - p - _ tr where @xmath116 . note that _ formally _ eq . ( [ eq : euler2 ] )
is identical to the euler equation in the ordinary hydrodynamics ; the distinctive feature of the superfluid is that the superflow is irrotational @xmath117 in general ( the special case of the rotating superfluid , when the analogy becomes perfect , is discussed below ) .
( [ eq : euler2 ] ) can be derived also starting from the momentum conservation equation : [ eq : momentum ] + _ ik= _ tr , where _ ik = + p_ik , @xmath24 is the pressure , c.c
. stands for complex conjugate . the right - hand - side of eq .
( [ eq : momentum ] ) is the external force per unit volume .
on writing @xmath112 and using the relations ( [ subst ] ) eq .
( [ eq : momentum ] ) becomes v_i + ( v_i v_k + p_ik)= _ i_tr .
the time derivative of @xmath25 can be eliminated in terms of eq .
( [ eq : euler_rho ] ) .
if we use for the pressure of the dilute bose gas the relation p = and keep the leading order term in the diluteness parameter , we arrive again at eq .
( [ eq : euler2 ] ) .
if the condensate is rotating at a angular velocity which is larger than the critical one @xmath18 , its energy is minimized via creation of vortices ; then the curl of the last relation in ( [ subst ] ) is non - zero , rather the phase of the superfluid order parameter changes by @xmath118 around a path which encircles vortex lines [ curlv ] = ( ) = _ j ^(2 ) ( -_j ) , where @xmath119 is a unit vector along a vortex line , @xmath120 is the radius vector of a vortex line in the plane orthogonal to the vector @xmath119 , and @xmath121 is a two - dimensional delta function in this plane . if @xmath122 the macroscopic hydrodynamic equations involve only _ course - grained _ quantities , which are averages over a large number of vortices ( i.e. over scales much larger than the size of a single vortex ) . the right - hand side of eq .
( [ curlv ] ) then becomes proportional to the density of the vortex lines , @xmath123 , since the continuum limit of vortex distribution implies @xmath124 .
the left - hand side of eq .
( [ curlv ] ) in the course - grained limit gives @xmath125 , since the energy is minimized by a superflow which mimics a rigid body rotation ( this minimization is carried out e.g. , in e.g. ref .
@xcite ) . writing eq .
( [ eq : euler2 ] ) in the frame rotating uniformly with the angular velocity @xmath33 amounts to adding to the right - hand side of this equation the centrifugal potential @xmath126 and the coriolis acceleration @xmath127 ( here @xmath128 is the superfluid velocity in the rotating frame ) . with this substitution
we recover eq .
( [ eq : euler ] ) .
note that the analogy to the euler equation for a uniformly rotating ordinary fluid now is complete . for _ intrinsic _ quantities which characterize the thermodynamic state of the fluid ( for example pressure and density ) the lagrangian and eulerian variations do not vanish in general .
if we use the relation between the lagrangian and eulerian variations @xmath129 ( here @xmath130 is a lagrangian displacement ) , the conservation of the mass of any fluid element implies that @xmath131 the first variations of the integrals of the various functions @xmath132 , that characterize the fluid flows , are given by @xmath133 for @xmath134 we see that the first eulerian variation of the density integral vanishes , @xmath135 as a consequence of the conservation of fluid mass .
then , for any _ extrinsic attribute _ of the fluid @xmath136 , which characterizes its dynamics ( for example velocity ) we have @xmath137 on taking into account that for such dynamical variables @xmath138 .
the latter rule should be applied for calculations of the eulerian variations of the virial equations .
a detailed account of the rules for constructing the variations of the virial equations is given in ref .
@xcite , chap .
13 - 15 .
it is worthwhile to note that kerr s solution to the equations of the general theory of relativity , which describes a rotating black hole , is the only other known example of an equilibrium figure which is a rotating sphere . in the present case ,
however , the equilibrium spherical figure is maintained due to the exact cancellation of the effect of the centrifugal potential by the external trapping potential .
d. odell , s. giovanazzi , g. kurizki , and v. m. akulin , phys .
lett * 84 * , 5687 ( 2000 ) .
a. sedrakian and i. wasserman , phys .
d * 63 * , 0240xx ( 2001 ) .
i. m. khalatnikov , _ introduction to the theory of superfluidity _
( addison wesley , new york , 1989 ) . | the tensor - virial method is applied for a study of oscillation modes of uniformly rotating bose - einstein condensed gases , whose rigid body rotation is supported by an vortex array .
the second order virial equations are derived in the hydrodynamic regime for an arbitrary external harmonic trapping potential assuming that the condensate is a superfluid at zero temperature .
the axisymmetric equilibrium shape of the condensate is determined as a function of the deformation of the trap ; its domain of stability is bounded by the constraint @xmath0 on the rotation rate ( measured in units of the trap frequency @xmath1 . )
the oscillations of the axisymmetric condensate are stable with respect to the transverse - shear and toroidal modes of oscillations , corresponding to the @xmath2 , @xmath3 surface deformations .
the eigenfrequencies of the modes are real and represent undamped oscillations .
the condensate is also stable against quasi - radial pulsation modes ( @xmath4 , @xmath5 ) , and its oscillations are undamped , if the superflow is assumed incompressible . in the compressible case
we find that for a polytropic equation of state , the quasi - radial oscillations are unstable when @xmath6 , and are stable otherwise .
thus , a dilute bose gas , whose equation of state is polytropic with @xmath7 to leading order in the diluteness parameter , is stable irrespective of the rotation rate . in non - axisymmetric traps , the equilibrium constrains the ( dimensionless ) deformation in the plane orthogonal to the rotation to the domain @xmath8 with @xmath9 .
the second harmonic oscillation modes in non - axisymmetric traps separate into two classes which have even or odd parity with respect to the direction of the rotation axis .
numerical solutions show that these modes are stable in the parameter domain where equilibrium figures exist . |
dynamic properties of electronic excitations in artificial and biological molecular aggregates are determined by the intra- and intermolecular interactions . due to resonant intermolecular interactions , electronic excitations of the aggregate form collective exciton states exhibiting coherent relationship between molecular excited states @xcite .
interactions with intramolecular vibrations reflect the specificity of molecular constituents of the aggregate and usually manifest themselves as vibrational sub - bands in electronic absorption or fluorescence spectra @xcite .
depending on the relative strength of these interactions , the induced excitations in molecular aggregates may lead to a host of photoinduced dynamics : from coherent and incoherent energy migration to reorganization of the surrounding environment @xcite . to resolve the details of the excitation evolution time - resolved spectroscopy techniques
have been developed .
femtosecond coherent two - dimensional ( 2d ) spectroscopy is one of the most versatile techniques @xcite .
2d spectroscopy has been already applied by considering the excitation dynamics in various molecular aggregates , e. g. photosynthetic pigment - protein complexes , polymers , tubular aggregates , quantum dots @xcite .
the inherent complexity of the 2d spectra was disclosed in all these cases .
for instance , the long - lasting oscillatory features , initially attributed to electronic quantum coherences , have been resolved @xcite .
possible vibronic / vibrational origin of some of these beats has been proposed in a number of recent studies @xcite .
thus , it has been proved that coupling to multiple intramolecular vibrational modes is essential to determine the spectral properties of the most chromophore molecules resulting in complex behavior of the 2d spectra @xcite . since the role of the coherence
is considered to be an important factor by defining the excitation dynamics in molecular aggregates the task of distinguishing electronic and vibrational character of coherences is essential and at the same time very challenging issue . to distinguish between the origins of the spectral oscillating behavior a molecular dimer ( md ) as the simplest model of the vibronic molecular aggregate
is considered .
this work is a natural extension of our previous comparative study of coherent 2d spectra of vibrational monomer and electronically - coupled dimer ( ed ) @xcite .
here we demonstrate that coherent oscillations of electronic and vibrational character are mixed due to intermolecular interactions
. however , they exhibit distinct oscillation patterns and different behavior with respect to the energetic disorder .
as usual for resonant spectroscopy applications let us assume that a constituent chromophore molecule of the aggregate is characterized by two levels corresponding to electronic ground and excited states and the electronic transitions are coupled to a single nuclear coordinate of high - frequency intramolecular vibrational mode with frequency @xmath0 .
this coupling can be effectively represented in terms of the displaced harmonic oscillator model @xcite .
the electronic excitation results in the shift along the dimensionless coordinate @xmath1 of the potential energy surface defined in the ground state as @xmath2 . assuming the heitler - london approximation @xcite the hamiltonian of the dimer is given by @xmath3|\mathrm{g}\rangle\langle\mathrm{g}|\nonumber \\ & + \left[\varepsilon_{1}+v(q_{1}-d_{1})+v(q_{2})\right]|\mathrm{e}_{1}\rangle\langle\mathrm{e}_{1}|\nonumber \\ & + \left[\varepsilon_{2}+v(q_{1})+v(q_{2}-d_{2})\right]|\mathrm{e}_{2}\rangle\langle\mathrm{e}_{2}|\nonumber \\ & + j\left[|\mathrm{e}_{1}\rangle\langle\mathrm{e}_{2}|+|\mathrm{e}_{2}\rangle\langle\mathrm{e}_{1}|\right]\nonumber \\ & + \left[\varepsilon_{1}+\varepsilon_{2}+v(q_{1}-d_{1})+v(q_{2}-d_{2})\right]|\mathrm{f}\rangle\langle\mathrm{f}|,\label{eq : hcoord}\end{aligned}\ ] ] where @xmath4 is the electronic common ground state
, @xmath5 corresponds to the electronic excited state of the @xmath6-th molecule of the dimer ( @xmath7 ) ; @xmath8 is the doubly - excited state of the dimer , which corresponds to electronic excitations of both monomers .
term @xmath9 represents the total kinetic energy of the vibrational harmonic oscillators coupled to both molecules , @xmath10 is the displacement value of the vibrational potential in the electronically excited state , @xmath11 is the energy difference between the potential minima of @xmath4 and @xmath5 states of the @xmath6-th monomer , @xmath12 is the resonance interaction . since @xmath13 with @xmath14 being the so - called reorganization energy , this difference in the potential energies determines the coupling of the electronic excitation to molecular vibrations . the latter is usually characterized by the dimensionless huang - rhys factor @xmath15 .
dependence of the energies of the singly - excited states of the molecular dimer on the resonant coupling ( a , black solid lines ) in the case of small huang - rhys factors @xmath16 . a clear separation of electronic excited states ( denoted as @xmath17 and @xmath18 , respectively ) and vibronic states ( denoted as @xmath19 , where @xmath20 the number of vibrational quanta ) is observerd . the energies of a purely electronic dimer @xmath21 are represented by the red dashed lines .
( b ) the absorption spectrum of the molecular vibronic dimer with coupling @xmath22 is also shown , which corresponds to the energy states indicated by the vertical gray line in ( a ) .
positions of the transitions contributing to the absorption spectrum are indicated by vertical lines . ] in the case of no displacement ( @xmath23 ) the electronic excitations are not coupled to molecular vibrations and the electronic part of the hamiltonian is easily diagonalized giving analytical expressions of the singly - excited state energies of the electronic dimer , @xmath24 , where @xmath25 and @xmath26 .
dependence of the exciton energies of the electronic dimer on the coupling is shown in fig .
[ fig : dependence - of - the]a by the red dashed lines .
the remaining vibrational hamiltonian corresponds to the harmonic oscillator which can be easily quantized .
the ground - state hamiltonian in eq .
( [ eq : hcoord ] ) is diagonal by definition independently on the value of @xmath27 and contains a ladder of vibrational states with energies spaced by @xmath0 . the singly - excited state block of the hamiltonian is not diagonal in this basis and depends on the value of @xmath27 .
after numerical diagonalization , if a sufficient number of vibrational states @xmath28 is taken into account , the excitation energies are found approximately at @xmath29 and @xmath30 ( @xmath31 ) and virtually constitute two ladders of equally - spaced states , denoted by @xmath32 and @xmath33 in fig .
[ fig : dependence - of - the]a . in the energy spectrum , the avoided crossing regions , where the electronic splitting matches vibrational quantum , emerge at particular values of the parameters when the most pronounced mixing of state character occurs @xcite . for the larger huang - rhys factors ( @xmath34
is used in fig .
[ fig : dependence - of - the]a ) the deviation of the vibrational excitation energies from the harmonic ladder structure becomes significant , especially in the vicinity of the avoided crossing regions .
we denote the electronic ground state as the state where both molecules are in their ground states with arbitrary vibrational excitations , @xmath35 .
index @xmath36 here indicates the @xmath36-th vibrational excitation of the first molecule and @xmath37 denotes the vibrational excitation of the second molecule .
@xmath38 denotes the state of the first molecule being in the vibronic excited state maintaining the @xmath36-th vibration quantum and the second molecule being in the @xmath37-th vibrational ground state ; analogously , @xmath39 corresponds to the state where the second molecule is vibronically excited .
doubly - excited electronic states @xmath40 are constructed in a similar way .
thus , the site basis of our model is complete with respect to vibronic states since one - particle vibronic states ( such as @xmath41 ) and two - particle vibronic states ( @xmath42 @xmath43 ) are included . in terms of these definitions the vibronic set of eigenstates for the singly- and doubly - excited electronic states ( state vectors @xmath44 and @xmath45 , correspondingly ) can be obtained by using the relevant transformations @xmath46 and @xmath47 .
the corresponding transformation coefficients @xmath48 and @xmath49 are acquired from the diagonalization of the singly - excited and doubly - excited blocks of the hamiltonian ( [ eq : hcoord ] ) and provide us with the information about the eigenstate composition which allows the estimation of the amount of mixing between vibrational and electronic states .
the transition dipole moments between the ground and singly - excited states as well as between singly- and doubly - excited states in the vibronic eigenstate basis are then given by @xmath50 and @xmath51 where @xmath52 and @xmath53 are the transition dipole moments of the monomers .
we will consider 2d signals of the molecular dimer , which are calculated using the perturbative system - response function theory in the vibronic eigenstate basis using the impulsive limit of the laser pulses , i.e. assuming that the laser pulse spectrum covers all frequencies .
more information regarding the details of calculations can by found elsewhere @xcite .
for the sake of simplicity , we assume the pure dephasing as the only mechanism responsible for the homogeneous lineshape formation .
the energetic disorder ( uncorrelated fluctuations of site energies @xmath11 ) is considered to be responsible for the inhomogeneous broadening that is taken into account by averaging over ensemble ( 1000 realizations of independent simulations ) with the gaussian distribution with standard deviation @xmath54 of excitation energies for every molecule .
we consider the molecular dimer with huang - rhys factors equal for both monomers , @xmath55 .
we set vibrational frequency @xmath0 as the reference parameter and choose the resonant coupling strength @xmath56 .
we also choose that electronic site energies are separated by the same value as vibrational frequency , @xmath57 . as values typical for aggregates , we consider @xmath58 and @xmath34 , thus resulting in @xmath59 @xmath60 , which is approximately equal to the excitonic splitting of the electronic dimer @xmath61 .
hence , the excitonic splitting is off - resonant with respect to the vibronic resonance ( compare @xmath18 , @xmath62 and @xmath63 levels fig . [
fig : dependence - of - the]a ) .
we assume that the strengths of the transition dipole moments of the monomers are equal and constitute the inter - dipole angle @xmath64 .
the dephasing rate determining the homogeneous linewidth is set to @xmath65 and central absorption frequency is chosen @xmath66 . for numerical calculations we choose @xmath67 ; calculations with more vibrational states show no evident changes in any simulated spectroscopic signals for the used set of parameters .
the deviation from a harmonic ladder of vibronic sates appears , but is insignificant for the model considered here , as the singly - excited state energy departure from the ideal harmonic progression is less than 5% and the chosen parameters ensure that we are away from the avoided - crossing regions ( see fig .
[ fig : dependence - of - the]a ) .
the electronic interaction with the high - frequency vibrational mode is almost indistinguishable in the absorption spectrum for the chosen small value of the huang - rhys factor as the only evidence of vibronic content is a sole weak shoulder in absorption at @xmath68 ( fig .
[ fig : dependence - of - the]b ) . for
the larger huang - rhys factor the stronger vibrational progression would evidently appear in the spectrum .
however , consideration of such an aggregate instead would not change conclusions of the further discussion appreciably
. transition amplitudes @xmath69 between the ground state manifold of the vibrational states and the manifold corresponding to the singly - excited states signify possible interaction configurations .
the " vibronic content " in a specific electronic transition can be quantified by the transformation coefficients combined as @xmath70 .
the maximum value of this quantity ( @xmath71 ) indicates the pure electronic character of the state , while @xmath72 reflects that the corresponding transition is vibronic .
transitions originating from the zero - vibrational state , i. e. @xmath73 are described in table [ tab : dipoles ] .
the two most significant transitions correspond to electronic - only transitions with @xmath74 , while the other transitions are of dominant vibrational character ( @xmath75 ) . [
cols="^,^,^,^,^,^,^,^,^,^,^ " , ] the total ( a sum of rephasing and non - rephasing ) and only rephasing 2d spectra of the molecular dimer at the population time @xmath76 , with gaussian disorder @xmath77 @xmath78 ( upper panels ) .
solid ( dashed ) contour lines for positive ( negative ) signals are drawn in 10% intervals .
time dependences of the amplitudes of the upper ( 1 - 2 ) and lower ( 2 - 1 ) cross - peaks of the rephasing signal are shown in the lower panels ( left and right , respectively ) .
green no disorder , red
@xmath79 , black
@xmath80 . ]
the real parts of rephasing and total 2d spectra contain three positive features : two diagonal peaks and a clearly distinguishable cross - peak ( 2 - 1 ) below the diagonal reflecting the coherent resonance coupling between the molecules ( fig .
[ fig : time - dependence - of ] ) .
the higher cross - peak ( 1 - 2 ) is less visible due to the overlap with the excited state absorption ( esa ) contribution .
apparent simple structure of the spectra the complicated pattern of various overlapping contributions .
they can be sorted into : i ) _ stationary contributions _ , denoting the excitation pathways ,
where the aggregate is in a population state during delay time @xmath81 ; ii ) _ oscillating contributions _ , reflecting the coherent electronic or vibronic excitation dynamics in the excited state or vibrational dynamics in the ground state .
the multitude of the later ones in case of the molecular dimer system is illustrated in fig .
[ fig : diagrams_smallhr]a .
temporal dynamics of oscillating contributions can be represented by the time dependent traces of certain spectral regions .
such traces of upper ( 1 - 2 ) and lower ( 2 - 1 ) cross - peaks of the rephasing signal are presented in the lower panel of fig .
[ fig : time - dependence - of ] .
the oscillations can also be characterized by studying the so - called oscillation fourier maps which are constructed by applying the fourier transform to the real part of the rephasing spectrum @xmath82 with respect to the delay time @xmath81 @xcite , @xmath83 this allows us to directly identify the phase and amplitude of oscillations in the different spectral regions .
such oscillation maps of the 2d spectra at frequencies corresponding to @xmath84 and @xmath85 are presented in figures [ fig : diagrams_smallhr]b and [ fig : diagrams_smallhr]c in case of no disorder ( @xmath86 ) and substantial disorder ( @xmath87 ) , respectively .
the oscillation map at @xmath85 involves two states with the most pronounced electronic character ( see table [ tab : dipoles ] ) and is the same as the purely electronic coherence map of an excitonic dimer @xcite , i.e. oscillations in the rephasing part appear only in the cross - peaks . as we find in the disordered system ( @xmath87 ) the oscillating patterns are completely dominated by the vibrational frequency @xmath0 , while the oscillations with the electronic gap frequency @xmath59 are at least 20 times weaker and , thus , their contribution is negligible .
the oscillation map of the disordered vibronic dimer is highly asymmetric , which is mostly due to the ground state bleaching ( gsb ) contribution ( fig .
[ fig : diagrams_smallhr]a ) .
it is known that the phase of spectral oscillations of the electronic - only systems is 0 at the maxima of the peaks and changes continuously when going away from the peak ( rises in direction outwards the diagonal ) . for the monomer coupled to a single mode of high - frequency vibration
, the phase of either @xmath88 or @xmath89 can be obtained depending on the coupling strength @xcite . however , in congested spectra any phase relationships can be obtained if the spectral overlap is substantial . especially for the mixed system , when both resonance coupling and vibronic interactions are taken into account . for example
, it can be seen in the oscillation map in fig .
[ fig : diagrams_smallhr]c that at the center of the lower diagonal peak the phase is detuned by @xmath90 from the value of @xmath89 , which would be expected for an isolated oscillating peak @xcite .
( a ) arrangement of the the most significant ( 90% of the total amplitude ) oscillatory contributions ( diamonds denote gsb , squares se and circles esa ) in the 2d spectrum of the molecular dimer .
the oscillations are at vibrational frequency @xmath0 ( open symbols ) or electronic frequency @xmath59 ( full gray symbols ) and the plus / minus sign at each peak denotes the phase of the oscillation .
( b ) the fourier maps are represented by the amplitude and phase of oscillations as contour lines and peak color , respectively , at frequencies @xmath0 and @xmath59 when @xmath86 .
( c ) the oscillation map at frequency @xmath0 when @xmath87 .
the map at @xmath59 frequency is not shown since it has a negligible amplitude . ]
the parameter set used in calculations describes a very general electronic - vibronic system .
the chosen values @xmath91 do not correspond to any special case since the monomer energies remain different , @xmath92 .
the chosen absolute values of parameters are typical for many molecular aggregates of interest , including the photosynthetic pigment - protein complexes .
similar parameters have already been used by investigating the vibronic transition dipole moment borrowing and coherence enhancement effects in molecular dimer systems @xcite .
the introduced model captures both limiting cases of purely vibrational and purely electronic model systems that are usually considered separately . indeed , by assuming either the huang - rhys factors or the resonance interaction to be zero both limiting cases can be obtained .
two prominent frequencies are present in the simulated oscillatory dynamics of the 2d spectrum of the vibronic molecular system .
evidently , one frequency corresponds to the electronic energy gap @xmath59 , while the other is equal to the vibrational frequency @xmath0 .
these coherences are between the states of mixed electronic - only and electronic - vibronic character , respectively ( table [ tab : dipoles ] ) .
it is noteworthy to mention , that a very similar amount of mixing is obtained using even simpler description with only one - particle vibronic states and @xmath93 included @xcite .
the mixed character might indirectly appear more signifficant if the diagonal disorder is comparable to the gap between states of different character .
then , the states of different character become overlapping leading to more mixing @xcite .
however , the distribution of the transition frequencies not the amount of mixing of particular state defines the oscillations .
thus , beatings with rather electronic and rather vibrational character still result in distinct oscillation maps and therefore the origin of the oscillation can be identified .
strength of oscillations strongly depends on system inhomogeneity . from simulations ,
this can clearly be seen in the lower panel of fig .
[ fig : time - dependence - of ] , where time dependences of the peak value for different values of the gaussian disorder ( @xmath86 , @xmath94 and @xmath95 ) are presented . the initial intensive oscillations with the @xmath59 frequency decay
rapidly when the disorder is increased and the only dynamics observed at longer delay times correspond to the @xmath0 beats .
strengths of the oscillations at different frequencies are represented by the corresponding values of the amplitude in the oscillation maps .
therefore , the value of the amplitude maximum of the oscillation map is used to evaluate the strength of the corresponding oscillation .
the amplitude of electronic oscillations decays sharply with the disorder , while dependence of the amplitude of vibronic / vibrational coherences is much more flat ( the amplitudes for the stimulated emission ( se ) , excited state absorption ( esa ) and ground state bleaching ( gsb ) contributions are presented in fig .
[ fig : amplitudes_diagrams ] ) .
when disorder is absent , signatures of electronic coherences mostly coming from the se contribution are at least 5 times larger than those of vibrational character , while for @xmath96 the amplitude of electronic coherences is negligible ( the corresponding oscillation map for @xmath97 can not be resolved ) .
separation of coherences of electronic and vibrational character is very significant as our results demonstrate that besides vibrational beats the electronic beats could be in principle observed and distinguished for the weakly disordered systems .
however , the electronic beats rapidly decay in time in a gaussian fashion ( for the gaussian disorder ) as @xmath98 . whereas vibrational beats will prevail for longer times
. maximal amplitude of the oscillation , taken from the oscillation maps as a function of disorder for gsb ( diamonds ) , se ( squares ) and esa ( circles ) contributions at different beating frequencies ( @xmath0 and @xmath59 ) . ] note that in the hamiltonian used ( eq .
[ eq : hcoord ] ) no transformation of canonical variables into symmetric and antisymmetric components is performed that would allow us to decouple the schrdinger equations of the excited and ground states .
therefore , presented model contains states which are not separated to diabatic or adiabatic ones as has been done in the theoretical work of jonas and coworkers addressing coherences in fmo @xcite .
also , since we show the separation of beats of electronic and vibrational character for arbitrary system parameters , the resonant condition of the mixed character coherences ( for example , when electronic energy gap or the chromophore energy difference is equal to the vibrational frequency ) is just a special case of our model .
it supports the assumptions about coherences in fmo where such resonant conditions are met electronic coherences at initial times could be of the same order as those of the vibronic origin due to disorder of approximately @xmath99 @xmath100 ; electronic beats will decay in a short time ( ~200 fs ) while vibrational beats would persist over the long time . for higly disordered and uncorrelated systems electronic coherences are not likely to be significant at all and the observed coherent dynamics are due to the ground - state vibrations .
it is also important to consider transition dipole moment orientations @xmath101 listed in table [ tab : dipoles ] .
the orientations of the transitions to the `` mixed '' states are different for each state and are also different from the transition dipole moments of monomer transitions .
this implies that coherences involving arbitrary states ( of electronic or vibrational character ) might survive the measurements with polarization schemes , devised to suppress all but electronic coherence signals @xcite . on the other hand , these polarization schemes
can then be used to distinguish between purely vibrational ( localized on one molecule ) and mixed origin coherences .
results presented here are related to the assumption that the vibrational frequencies are not affected by inhomogeneities , which induce the disorder of electronic transition energy .
this is often the case as vibrational resonances are less sensitive to the electrostatic configuration of the environment than the delocalized electronic wavefunctions .
vibrational coherences , hence , decay on a timescale of vibrational dephasing , which usually is in the order of a picosecond .
it should be also noted that when vibrational frequency @xmath0 matches the electronic energy gap @xmath59 , the quantum mechanical mixing complicates the whole picture .
then damping of the resulting electronic - vibrational coherence will define the decay timescale of the oscillations in the spectra .
coherent dynamics of coupled molecules can be effectively sorted out by the modeling of 2d spectra .
coupling between the multitude of vibrational states on different molecules results in a manifold of mixed states .
when vibrational frequency and the electronic energy gap are off resonance this manifold constitutes two ladders of equally - spaced and well - resolved states .
two types of beats having either electronic or vibrational character can then be distinguished by the use of oscillation maps , constructed from the sequences of time - resolved 2d spectra .
inhomogeneous disorder effects the coherences of different state character differently .
the amplitude of the electronic character beatings , caused by the coherences in excited states , is dramatically reduced by the disorder and consequently electronic coherences are quickly dephased .
vibrational character beatings stem from the ground and excited state contributions and depend weakly on the disorder , assuring their long - time survival .
this research was partially funded by the european social fund under the global grant measure .
v. b. acknowledges support by project " promotion of student scientific activities " ( vp1 - 3.1-mm-01-v-02 - 003 ) from the research council of lithuania .
d. z was supported by the swedish research council .
d. zigmantas , e. l. read , t. manal , t. brixner , a. t. gardiner , r. j. cogdell , and g. r. fleming .
two - dimensional electronic spectroscopy of the b800-b820 light - harvesting complex .
, 103(34):1267212677 , 2006 .
g. s. engel , t. r. calhoun , e. l. read , t. k. ahn , t. manal , y. c. cheng , r. e. blankenship , and g. r. fleming .
evidence for wavelike energy transfer through quantum coherence in photosynthetic systems .
, 446:782 , 2007 .
jaroslaw sperling , alexandra nemeth , jrgen hauer , darius abramavicius , shaul mukamel , harald f. kauffmann , and franz milota .
excitons and disorder in molecular nanotubes : a 2d electronic spectroscopy study and first comparison to a microscopic model .
, 114(32):81798189 , 2010 .
daniel b. turner , krystyna e. wilk , paul m. g. curmi , and gregory d. scholes . comparison of electronic and vibrational coherence measured by two - dimensional electronic spectroscopy .
, 2(15):19041911 , 2011 .
gitt panitchayangkoon , dmitri v. voronine , darius abramavicius , justin r. caram , nicholas h. c. lewis , shaul mukamel , and gregory s. engel . direct evidence of quantum transport in photosynthetic light - harvesting complexes .
, 108(52):2090820912 , 2011 .
vivek tiwari , william k. peters , and david m. jonas .
electronic resonance with anticorrelated pigment vibrations drives photosynthetic energy transfer outside the adiabatic framework .
, 110(4):12031208 , 2013 .
tom manal , niklas christensson , vladimr luke , franz milota , oliver bixner , harald f. kauffmann , and jrgen hauer .
system - dependent signatures of electronic and vibrational coherences in electronic two - dimensional spectra . , 3(11):14971502 , 2012 .
a. w. chin , j. prior , r. rosenbach , f. caycedo - soler , s. f. huelga , and m. b. plenio .
the role of non - equilibrium vibrational structures in electronic coherence and recoherence in pigment - protein complexes .
, 9(2):113118 , february 2013 .
daniel b turner , rayomond dinshaw , kyung - koo lee , michael belsley , krystyna e wilk , paul mg curmi , and gregory scholes .
quantitative investigations of quantum coherence for a light - harvesting protein at conditions simulating photosynthesis . , 14:48574874 , 2012 .
t. r. calhoun , n. s. ginsberg , g. s. schlau - cohen , y - c .
cheng , m. ballottari , r. bassi , and g. r. fleming . quantum coherence enabled determination of the energy landscape in light - harvesting complex ii .
, 113:1629116295 , 2009 .
gabriela s. schlau - cohen , tessa r. calhoun , naomi s. ginsberg , elizabeth l. read , matteo ballottari , roberto bassi , rienk van grondelle , and graham r. fleming .
pathways of energy flow in lhcii from two - dimensional electronic spectroscopy . , 113(46):1535215363 , 2009 . | coherent dynamics of coupled molecules are effectively characterized by the two - dimensional ( 2d ) electronic coherent spectroscopy . depending on the coupling between electronic and vibrational states , oscillating signals of purely electronic , purely vibrational or mixed character
can be observed with the help of oscillation maps , constructed from the time - resolved 2d spectra .
the amplitude of the beatings caused by the electronic coherence is heavily affected by the energetic disorder and consequently the electronic coherences are quickly dephased .
beatings with the vibrational character weakly depend on the disorder , assuring their long - time survival .
we show that detailed modeling of 2d spectroscopy signals of molecular aggregates provides direct information on the origin of the coherent beatings .
# 1 |
in hierarchical merging scenarios dwarf galaxies are thought to be the bricks from which larger galaxies were assembled .
the study of the star formation history ( sfh ) and of the chemical evolution of presently existing dwarf galaxies is then extremely important for a proper understanding of the formation and evolution of the larger systems that this type of galaxies may have contributed to build in the past .
colour magnitude diagrams ( cmds ) reaching the faint main sequence turn - off ( to ) of the oldest stellar components are the most traditional and reliable way to derive the sfh of any individual dwarf galaxy .
however , these detailed studies , possible so far mainly for the dwarf members of the local group ( lg ) , require very time consuming observations . the rr lyrae variables , being about 3 magnitudes brighter than coeval to stars ( @xmath7 gyrs ) , are much easier to observe .
they offer an excellent tool for tracing the oldest stellar populations , and therefore the epoch of galaxy formation , in composite systems such as the resolved lg dwarf galaxies .
recent work by several groups has lead to the discovery of rr lyrae stars in increasing numbers of lg dwarf galaxies ( e.g. leo i and ii , ic 1613 , fornax , and vi , ngc6822 : @xcite , just to mention a few of them ) . an early stellar population , nearly coeval to the old galactic globular clusters , has been found in the majority of lg galaxies , irrespective of their star formation histories .
this indicates that all lg dwarfs started forming stars at an early epoch , @xmath813 gyr ago ( e.g. held et al .
2000 ) .
the sculptor dwarf spheroidal galaxy is no exception to this general trend .
226 rr lyrae stars and 3 anomalous cepheids have been detected in a 15@xmath0 area around the galaxy centre by @xcite who published multi - epoch photometry for all of them .
@xcite also found that the period distribution of the rrab stars shows a sharp cut - off at @xmath9 d implying a metallicity of [ fe / h]@xmath10 ( on the zinn & west 1984 scale ) , and that the dispersion of the average @xmath11 magnitudes is most likely due to the metallicity spread exhibited by the stars in this galaxy .
similarly , @xcite found @xmath12[fe / h]@xmath13 with a large dispersion from about @xmath14 to @xmath15 dex ( on the metallicity scale by @xcite ) , from the fourier decomposition of the light curves of the rrab stars .
indeed , sculptor dwarf spheroidal ( dsph ) has long been known , from photometric studies , to have a large metallicity spread and bimodality in the metallicity and spatial distribution of its horizontal branch ( hb ) stars @xcite . very few spectroscopic determinations of the metal abundance of sculptor stars existed so far ( @xcite , 2003 ) .
@xcite , from high resolution spectroscopy of red giants , find [ fe / h ] values in the range from @xmath16 to @xmath17 dex , confirming a large metallicity spread in sculptor .
two distinct ancient populations showing an abrupt change in the [ fe / h ] distribution at about 12 arcmin from the galaxy center are found in sculptor by @xcite based on flames@vlt low resolution spectroscopy and wfi imaging of the galaxy .
the two components have also different spatial distribution and velocity dispersion .
in this paper we present metal abundance determinations based on multi - slit low resolution spectroscopy obtained with fors2 at the vlt for more than a hundred rr lyrae stars in sculptor .
the variables are spread over a 15@xmath0 area around the galaxy centre , thus being almost coincident with the internal region of sculptor where @xcite find segregation of red hb stars .
the knowledge of the _ metal distribution _ of the rr lyrae population allows to put important constraints on the early star formation and chemical evolution histories of the host galaxy , by removing the age - metallicity degeneracy : the earliest measurable data point for the chemical enrichment history of sculptor can be determined with accuracy .
observations and data reduction are discussed in section 2 .
metal abundances are presented in section 3 .
the luminosity - metallicity relation and the radial velocities we determined for the rr lyrae stars in sculptor are discussed in sections 4 and 5 , respectively . in section 6 the metallicity distribution of the sculptor
rr lyrae stars is compared with that of other old stellar populations in the galaxy .
a summary and some final remarks in section 7 close the paper .
observations of 110 variable stars ( 107 rr lyrae , 1 anomalous cepheid , 1 binary system , and 1 variable of unknown type ) in sculptor and of rr lyrae stars in 4 galactic globular clusters ( namely : m 2 , m 15 , ngc 6171 , and ngc 6441 ) were carried out using fors2 ( focal reducer / low dispersion spectrograph 2 ) , mounted on the eso very large telescope ( paranal , chile ) .
the data were collected in service mode during the period july 29 to august 5 , 2003 .
typical seeing values during the observations were in the range 0@xmath1871@xmath19 and on average of about 1@xmath20 .
we used the mxu ( mask exchange unit ) configuration , that allows to observe simultaneously many objects with more freedom in choosing the location , size and shape of individual slitlets with respect to the standard mos mode .
the detector is a mosaic of two mit ccds with 15 @xmath21 m pixel size .
spectra were collected using the blue grism gris_600b covering the 3450 - 5900 wavelength range , with a dispersion of 50 mm@xmath6 with slits 1@xmath22 wide , and usually 14@xmath22 long to allow for sky subtraction . with this configuration
, each pixel corresponds to 0.75 .
an effort was made to cover for each star the relevant wavelength range ( @xmath8 3900 - 5100 ) containing both the caii k and the hydrogen balmer lines up to h@xmath23 .
we have used an instrumental set - up similar ( i.e. same spectral range , resolution , and typical s / n ) to that employed in our study of the rr lyrae stars in the lmc @xcite , so that rr lyrae variables in some calibrating gcs are already available ( namely in clusters ngc 1851 , ngc 3201 , and m 68 ) .
exposure times on the sculptor variables were of 31 min , as an optimal compromise between s / n and time resolution of the light curve of the rr lyrae targets .
we employed 9 masks in sculptor , 2 in ngc 6441 , 1 in m2 , 1 in ngc 6171 and 1 in m15 .
the 9 sculptor fields were slightly overlapped , so that for 25 variables we have more than 1 spectrum .
a detailed log of the observations is given in table [ t : tablelog ] where n is the number of variable stars observed in each mask .
the complete listing of the variables observed in sculptor is provided in table [ t : sculptor ] where we have adopted @xcite identification numbers .
their location on a @xmath24 map of the central region of sculptor dsph galaxy is shown in fig .
[ f : fatte ] .
finding charts corresponding to the nine @xmath25 fors2 subfields are given in appendix a. centre of field coordinates are provided in table [ t : tablelog ] .
equatorial coordinates for all our targets can be found in table 2 of @xcite .
lcccrr field & ra & dec & date & exptime & n + & ( jd2000 ) & ( jd2000 ) & ( ut ) & ( sec ) + m15 & 21:30:13.4 & 12:11:50.4 & 2003 - 07 - 31 & 150 & 9 + m2 & 21:33:40.4 & -00:50:45.7 & 2003 - 08 - 05 & 150 & 6 + ngc6171 & 16:32:34.6 & -13:06:46.8 & 2003 - 07 - 29 & 150 & 6 + ngc6441a & 17:50:19.9 & -37:05:21.9 & 2003 - 07 - 29 & 300 & 5 + ngc6441b & 17:50:24.5 & -37:00:06.3 & 2003 - 07 - 30 & 300 & 7 + scl1 & 01:00:30.7 & -33:38:18.4 & 2003 - 08 - 01 & 1860 & 15 + scl2 & 01:00:08.7 & -33:38:18.5 & 2003 - 08 - 01 & 1860 & 18 + scl3 & 00:59:51.2 & -33:38:18.5 & 2003 - 08 - 02 & 1860 & 15 + scl4 & 01:00:34.8 & -33:42:43.1 & 2003 - 08 - 02 & 1860 & 15 + scl5 & 01:00:10.3 & -33:42:43.5 & 2003 - 08 - 02 & 1860 & 19 + scl6 & 00:59:49.1 & -33:42:45.0 & 2003 - 08 - 03 & 1860 & 18 + scl7 & 01:00:32.0 & -33:47:18.5 & 2003 - 08 - 03 & 1860 & 15 + scl8 & 01:00:09.5 & -33:47:15.5 & 2003 - 08 - 05 & 1860 & 19 + scl9 & 00:59:47.8 & -33:47:53.8 & 2003 - 08 - 05 & 1860 & 13 + [ t : tablelog ] data reduction was performed using the standard iraf routines .
images have been trimmed , corrected for bias and for the normalized flat field .
then we used the iraf command _
lineclean _ to reduce the contamination by cosmic rays .
up to 19 spectra were present in each pointing , and were extracted with the optimal extraction and automated cleaning options switched on . the sky contribution was subtracted making use of the slit length .
the contamination of targets from nearby stars was reduced to a minimum , except for a few objects . for each science
mask a hecdhg lamp was acquired , and used to calibrate in wavelength the spectra , each one covering a different spectral range , depending on the target position .
not less than 10 lines of the calibration lamp were visible for each aperture , and the resulting dispersion solutions have r.m.s . of about 0.03 .
further cleaning of cosmic rays hits and bad sky subtractions was done using the clipping option in the iraf _ splot _ task .
[ f : figspectra ] shows examples of the final spectra .
we obtained spectra for 110 of the variable stars identified in sculptor by @xcite , and for 25 of them we have multiple observations .
these authors published photometry in the @xmath11 band for all our targets .
periods and epochs of maximum light were determined from their time series data ( kindly made available by dr .
j. kaluzny ) , using the period search package graphycal analyzer of time series ( gratis , di fabrizio 1999 , clementini et al .
the new ephemerides are provided in table [ t : sculptor ] .
we found that our periods are in general slightly different from those published by @xcite ; differences are in most cases around the fourth or fifth digit .
however , there are a number of cases where our periods and type classifications significantly differ from those of @xcite who published aliases of the periodicities preferred here .
all these objects have been flagged in the last column of table [ t : sculptor ] , where we provide comments on individual stars .
we used our new periods to phase the spectra of our target stars in sculptor since the scatter in the light curves appears to be significantly reduced than using the published values .
phases corresponding to the heliocentric julian day ( hjd ) at half exposure are listed in column 3 of table [ t : sculptor ] ; for the double - mode rr lyrae stars they correspond to the first overtone pulsation period .
we also computed intensity - averaged luminosities for all the variables in our study , that are given in column 5 of table [ t : sculptor ] . based on our study of the light
curves our sample contains 62 _ ab-_type , 40 _
c-_type , 3 confirmed and 2 suspected _ d-_type rr lyrae stars , 1 anomalous cepheid , 1 suspected binary system , and 1 variable of unknown type .
precise and homogeneous metal abundances for the target stars in sculptor were measured using the revised version of the @xmath26 method @xcite devised by @xcite .
we do not actually measure @xmath27s values , but rather estimate metallicities for individual variables by comparing the strength of the h lines and of the k ca ii line with analogous data for variables in gcs of known metallicity .
a detailed description of our method can be found in @xcite .
a summary of the technique , and an update of the calibration procedures are provided in appendix b , to which the interested reader is referred to for details .
spectral line indices measured for the variables in sculptor following @xcite , are given in columns from 8 to 12 of table [ t : sculptor ] . and
@xmath28 spectral indices for the variable stars in sculptor .
solid lines represent the mean lines for m68 and ngc1851 from @xcite.,width=332 ] the correlation between @xmath29 and @xmath28 spectral indices is shown in fig .
[ f : fig2 ] . in this figure
the solid lines represent the mean relations of the calibrating clusters m68 and ngc1851 ( see appendix b ) .
the variables in sculptor fall almost entirely below the mean line of ngc1851 indicating that they are metal - poorer than this cluster ( [ fe / h]@[email protected] ) , and also extend below the mean line of m68 , showing that there is a number of variables in sculptor metal - poorer than [ fe / h]@[email protected] .
metallicity indices @xmath32 defined according to equations ( 1 ) , ( 2 ) and ( 3 ) in appendix b , are listed in column 13 of table [ t : sculptor ] .
individual metallicities were derived from the calibration equations ( 4 ) and ( 5 ) in appendix b in the zinn & west ( 1984 , hereafter zw84 ) and in the carretta & gratton ( 1997 , hereafter cg97 ) metallicity scales separately ; they are listed in columns 14 and 15 of table [ t : sculptor ] . for different observations of the same star
the [ fe / h ] values are averages of all the available measurements . from the objects with multiple observations we estimate that errors of individual abundance determinations are 0.15 and 0.16 dex in zw84 and cg97 metallicity scales , respectively .
the average metal abundance of the 107 rr lyrae stars in our sample is [ fe / h][email protected]@xmath330.03 ( r.m.s.=0.26 dex ) with a total metallicity range of @xmath34[fe / h]@xmath3 in zw84 scale , and an offset of about 0.2 dex to higher metallicity in cg97 scale .
the star - to - star scatter inferred from the r.m.s .
dispersions is about 0.19 - 0.23 dex , hence larger than typical measurement errors .
if we adopt 0.23 dex as the measured metallicity spread , and quadratically subtract a [ fe / h ] measurement error of 0.16 dex , we obtain @xmath35 dex as our estimate of the intrinsic spread in the metal abundance of the sculptor rr lyrae stars .
the observed metallicity distribution of the sculptor rr lyrae stars ( uncorrected for the measurement errors ) in the zw84 metallicity scale is shown in fig .
[ f : fig3hbw ] .
variable stars are divided by type .
we find that the different types follow the same metallicity distributions .
as well known , cg97 scale provides systematically higher metallicities than zw84 . in the following
we will adopt zw84 values , unless explicitly noted .
however , independently of the adopted metallicity scale there are only very few metal - rich stars in our rr lyrae sample . based on the period distribution of the _ ab-_type pulsators , @xcite
conclude that the bulk of the rr lyrae stars in sculptor have metallicities equal to or lower than [ fe / h]@xmath36 .
indeed , there are only 26 stars with [ fe / h]@xmath37 in our sample .
this is also the metallicity at which @xcite find the break between metal - poor and metal - rich old populations in sculptor .
we will come back to this point in section 6 .
@xmath38 } & { \rm [ fe / h ] } & ~v_{r } & { \rm notes}\\ { \rm ( a ) } & { \rm ( b ) } & { \rm ( b ) } & { \rm ( b ) } & { \rm ( b ) } & { \rm ( c ) } & & & & & & & & { \rm zw } & { \rm cg } & { \rm km s^{-1 } } & \\
\noalign{\smallskip } \hline \noalign{\smallskip } 33 & { \rm c } & 0.309071 & 9168.9087 & 20.110 & 6.7675 & 0.088 & 0.002 & 0.352 & 0.347 & 0.300 & 0.333 & -0.280 & -2.42 & -2.26 & ~89 & \\ 37 & { \rm ab } & 0.508709 & 9162.9151 & 20.189 & 6.7675 & 0.230 & 0.161 & 0.168 & 0.199 & 0.164 & 0.177 & 0.239 & -1.96 & -1.78 & ~68 & \\ 59 & { \rm c } & 0.359681 & 9168.9359 & 20.176 & 6.7675 & 0.073 & 0.073 & 0.277 & 0.253 & 0.226 & 0.252 & 0.035 & -2.14 & -1.97 & 111 & { \rm ( d)}\\ 71 & { \rm c } & 0.434922 & 9228.8270 & 20.015 & 6.7675 & 0.589 & 0.104 & 0.196 & 0.159 & 0.176 & 0.177 & -0.127 & -2.28 & -2.12 & 105 & \\ 321 & { \rm ab } & 0.601783 & 9167.9003 & 20.137 & 4.7959 & 0.616 & 0.234 & 0.120 & 0.133 & 0.148 & 0.134 & 0.386 & -1.77 & -1.57 & 101 & \\ & & & & & 6.7675 & 0.892 & 0.224 & 0.172 & 0.141 & 0.172 & 0.162 & 0.528 & & & ~99 & \\ 368 & { \rm c } & 0.358962 & 9184.8696 & 20.133 & 4.7959 & 0.727 & 0.096 & 0.269 & 0.249 & 0.245 & 0.254 & 0.252 & -1.95 & -1.76 & 108 & \\ 439 & { \rm ab } & 0.496115 & 8809.6429 & 20.146 & 4.7959 & 0.658 & 0.215 & 0.141 & 0.161 & 0.152 & 0.152 & 0.401 & -1.81 & -1.62 & 109 & \\
737 & { \rm c / d ? } & 0.338614 & 8809.8469 & 20.061 & 4.7959 & 0.612 & 0.083 & 0.308 & 0.294 & 0.252 & 0.285 & 0.328 & -1.88 & -1.69 & ~96 & { \rm ( e)}\\ 763 & { \rm c } & 0.293114 & 8830.8675 & 20.139 & 3.7485 & 0.615 & 0.128 & 0.280 & 0.239 & 0.176 & 0.232 & 0.370 & -1.84 & -1.65 & 102 & \\ 786 & { \rm c } & 0.279088 & 8809.7457 & 20.179 & 3.7485 & 0.048 & 0.072 & 0.362 & 0.377 & 0.328 & 0.356 & 0.601 & -1.64 & -1.43 & 111 & \\ 793 & { \rm c } & 0.359921 & 8809.9129 & 20.135 & 3.7485 & 0.339 & 0.118 & 0.268 & 0.264 & 0.228 & 0.253 & 0.442 & -1.91 & -1.72 & 107 & \\ & & & & & 4.7959 & 0.249 & 0.092 & 0.253 & 0.250 & 0.230 & 0.244 & 0.150 & & & 140 & \\ 803 & { \rm ab } & 0.573658 & 8809.8060 & 20.061 & 3.7485 & 0.394 & 0.222 & 0.135 & 0.131 & 0.167 & 0.144 & 0.392 & -1.82 & -1.63 & ~92 & \\ 811 & { \rm c } & 0.363134 & 8809.8517 & 20.131 & 3.7485 & 0.098 & 0.092 & 0.308 & 0.283 & 0.268 & 0.287 & 0.420 & -1.93 & -1.75 & 131 & \\ & & & & & 4.7959 & 0.982 & 0.090 & 0.265 & 0.229 & 0.232 & 0.242 & 0.120 & & & 140 & \\ 860 & { \rm c } & 0.359517 & 9158.9314 & 20.206 & 4.7959 & 0.071 & 0.140 & 0.299 & 0.208 & 0.185 & 0.231 & 0.464 & -1.76 & -1.56 & 129 & { \rm ( f)}\\ 1139 & { \rm ab } & 0.518930 & 8809.5982 & 20.060 & 3.7485 & 0.245 & 0.075 & 0.299 & 0.262 & 0.269 & 0.277 & 0.194 & -2.00 & -1.82 & 127 & \\
1164 & { \rm ab } & 0.606190 & 8809.6323 & 20.185 & 3.7485 & 0.367 & 0.261 & 0.152 & 0.113 & 0.113 & 0.126 & 0.473 & -1.75 & -1.56 & 126 & \\
1261 & { \rm ab } & 0.630359 & 8809.4200 & 20.172 & 3.7485 & 0.912 & 0.323 & 0.106 & 0.134 & 0.137 & 0.125 & 0.782 & -1.48 & -1.26 & 148 & \\
1439 & { \rm c } & 0.356043 & 9168.9196 & 20.152 & 6.8001 & 0.952 & 0.115 & 0.292 & 0.230 & 0.199 & 0.240 & 0.324 & -1.88 & -1.70 & 104 & { \rm ( d)}\\ 1457 & { \rm ab } & 0.717799 & 8827.8527 & 19.919 & 6.8001 & 0.915 & 0.085 & 0.341 & 0.295 & 0.243 & 0.293 & 0.398 & -1.82 & -1.63 & ~74 & \\ 1491 & { \rm c } & 0.357866 & 8809.7866 & 20.135 & 6.8001 & 0.727 & 0.153 & 0.350 & 0.225 & 0.223 & 0.266 & 0.854 & -1.41 & -1.20 & ~84 & { \rm ( g)}\\ 1519 & { \rm c } & 0.356859 & 9169.9083 & 20.173 & 6.7675 & 0.414 & 0.127 & 0.228 & 0.193 & 0.182 & 0.201 & 0.161 & -2.03 & -1.85 & ~79 & \\ 1546 & { \rm ab } & 0.531239 & 9174.9206 & 20.151 & 6.8001 & 0.737 & 0.143 & 0.076 & 0.158 & 0.129 & 0.121 & -0.153 & -2.31 & -2.14 & ~92 & \\
1555 & { \rm ab } & 0.527243 & 8823.8298 & 20.089 & 6.7675 & 0.112 & 0.078 & 0.312 & 0.301 & 0.286 & 0.300 & 0.372 & -1.84 & -1.65 & ~81 & \\
1558 & { \rm c / d?/bl } & 0.243016 & 8833.9116 & 20.203 & 6.8001 & 0.006 & 0.102 & 0.296 & 0.250 & 0.198 & 0.248 & 0.267 & -1.65 & -1.45 & ~94 & { \rm ( h)}\\ & & & & & 6.7675 & 0.872 & 0.175 & 0.259 & 0.265 & 0.225 & 0.250 & 0.913 & & & 107 & \\ 1566 & { \rm ab } & 0.570272 & 8809.9982 & 20.053 & 6.8001 & 0.266 & 0.177 & 0.175 & 0.210 & 0.193 & 0.193 & 0.458 & -1.77 & -1.57 & 107 & \\
1823 & { \rm c } & 0.298462 & 9189.8649 & 20.255 & 4.7959 & 0.373 & 0.152 & 0.294 & 0.230 & 0.235 & 0.253 & 0.743 & -1.52 & -1.30 & 107 & \\
1830 & { \rm ab } & 0.517855 & 8809.2712 & 20.224 & 3.8132 & 0.181 & 0.145 & 0.180 & 0.165 & 0.160 & 0.168 & 0.089 & -2.09 & -1.92 & 112 & \\
1875 & { \rm ab } & 0.499773 & 8823.9437 & 20.028 & 3.8132 & 0.400 & 0.188 & 0.153 & 0.151 & 0.113 & 0.139 & 0.174 & -2.02 & -1.84 & ~95 & { \rm ( f)}\\ 1899 & { \rm ab } & 0.646664 & 9182.9074 & 19.978 & 4.7959 & 0.201 & 0.071 & 0.240 & 0.219 & 0.210 & 0.223 & -0.137 & -2.29 & -2.13 & ~74 & \\
1910 & { \rm ab } & 0.572828 & 9189.9080 & 20.083 & 3.8132 & 0.172 & 0.111 & 0.183 & 0.185 & 0.177 & 0.182 & -0.059 & -2.11 & -1.93 & ~97 & \\ & & & & & 6.8001 & 0.386 & 0.209 & 0.133 & 0.128 & 0.118 & 0.126 & 0.208 & & & ~77 & \\
1914 & { \rm ab } & 0.570540 & 8829.8781 & 20.220 & 6.8001 & 0.084 & 0.195 & 0.120 & 0.106 & 0.194 & 0.140 & 0.213 & -1.87 & -1.68 & ~98 & \\ & & & & & 6.7675 & 0.027 & 0.248 & 0.137 & 0.133 & 0.138 & 0.136 & 0.478 & & & ~78 & \\
1930 & { \rm ab } & 0.611160 & 9188.9162 & 20.251 & 3.8132 & 0.626 & 0.141 & 0.251 & 0.235 & 0.206 & 0.231 & 0.473 & -1.79 & -1.59 & 103 & \\ & & & & & 6.8001 & 0.514 & 0.165 & 0.220 & 0.170 & 0.190 & 0.193 & 0.380 & & & 119 & { \rm ( g)}\\ 1932 & { \rm ab } & 0.506044 & 9174.9206 & 20.155 & 4.7959 & 0.848 & 0.275 & 0.143 & 0.152 & 0.155 & 0.150 & 0.731 & -1.53 & -1.32 & 118 & \\ & & & & & 6.8001 & 0.809 & 0.257 & 0.228 & 0.141 & 0.113 & 0.160 & 0.716 & & & ~90 & \\ 1940 & { \rm ab } & 0.692975 & 9224.7748 & 20.117 & 4.7959 & 0.312 & 0.332 & 0.127 & 0.137 & 0.091 & 0.118 & 0.769 & -1.49 & -1.28 & ~98 & \\ & & & & & 6.7675 & 0.157 & 0.333 & 0.109 & 0.126 & 0.117 & 0.118 & 0.769 & & & 104 & \\
1941 & { \rm c } & 0.365674 & 9166.8670 & 20.152 & 6.7675 & 0.689 & 0.157 & 0.258 & 0.253 & 0.250 & 0.254 & 0.789 & -1.47 & -1.26 & 107 & \\
1943 & { \rm ab } & 0.551149 & 9169.9100 & 20.146 & 3.8132 & 0.045 & 0.203 & 0.172 & 0.131 & 0.127 & 0.143 & 0.281 & -1.54 & -1.32 & ~89 & \\ & & & & & 6.8001 & 0.464 & 0.248 & 0.235 & 0.222 & 0.184 & 0.214 & 1.159 & & & 103 & { \rm ( i)}\\ 1997 & { \rm ab } & 0.626766 & 8823.9437 & 20.136 & 3.8132 & 0.625 & 0.179 & 0.165 & 0.101 & 0.138 & 0.134 & 0.097 & -2.08 & -1.91 & 125 & \\
2058 & { \rm ab } & 0.503415 & 9226.9053 & 20.238 & 3.8132 & 0.610 & 0.133 & 0.224 & 0.212 & 0.204 & 0.214 & 0.288 & -1.92 & -1.73 & ~96 & \\ 2423 & { \rm c } & 0.358540 & 9167.9003 & 20.010 & 3.8132 & 0.328 & 0.202 & 0.170 & 0.149 & 0.157 & 0.159 & 0.372 & -1.84 & -1.65 & 120 & { \rm ( f)}\\ 2424 & { \rm c } & 0.348780 & 8809.8058 & 20.145 & 4.7959 & 0.527 & 0.166 & 0.215 & 0.227 & 0.178 & 0.207 & 0.482 & -1.75 & -1.55 & 118 & \\
\noalign{\smallskip } \hline \end{array}\ ] ] @xmath38 } & { \rm [ fe / h ] } & ~v_{r } & { \rm notes}\\ { \rm ( a ) } & { \rm ( b ) } & { \rm ( b ) } & { \rm ( b ) } & { \rm ( b ) } & { \rm ( c ) } & & & & & & & & { \rm zw } & { \rm cg } & { \rm km s^{-1 } } & \\
\noalign{\smallskip } \hline \noalign{\smallskip } 2450 & { \rm ab } & 0.617962 & 8810.1279 & 20.071 & 2.7656 & 0.886 & 0.252 & 0.141 & 0.141 & 0.126 & 0.136 & 0.500 & -1.84 & -1.65 & 113 & \\ & & & & & 3.8132 & 0.581 & 0.238 & 0.123 & 0.135 & 0.133 & 0.130 & 0.383 & & & ~92 & \\ & & & & & 4.7959 & 0.172 & 0.192 & 0.136 & 0.157 & 0.146 & 0.146 & 0.239 & & & 108 & \\ 2455 & { \rm ab } & 0.636250 & 9184.8700 & 20.156 & 3.7485 & 0.413 & 0.247 & 0.160 & 0.135 & 0.112 & 0.136 & 0.467 & -1.76 & -1.56 & 113 & \\ 2458 & { \rm c } & 0.357686 & 8809.6823 & 20.156 & 2.7656 & 0.446 & 0.118 & 0.260 & 0.208 & 0.187 & 0.218 & 0.199 & -1.73 & -1.53 & ~85 & \\ & & & & & 3.8132 & 0.375 & 0.201 & 0.228 & 0.208 & 0.205 & 0.214 & 0.802 & & & 102 & \\ 2470 & { \rm ab } & 0.693447 & 8809.4598 & 20.073 & 3.8132 & 0.245 & 0.168 & 0.183 & 0.220 & 0.207 & 0.203 & 0.467 & -1.76 & -1.56 & ~94 & \\
2482 & { \rm c } & 0.365865 & 9167.8931 & 20.191 & 4.7959 & 0.233 & 0.108 & 0.337 & 0.250 & 0.279 & 0.289 & 0.600 & -1.64 & -1.44 & 135 & { \rm ( f)}\\ 2502 & { \rm
ab } & 0.487474 & 9185.8606 & 20.238 & 4.7959 & 0.419 & 0.123 & 0.268 & 0.300 & 0.258 & 0.275 & 0.645 & -1.60 & -1.39 & 140 & { \rm ( f)}\\ 2545 & { \rm ab } & 0.674059 & 8809.0126 & 20.016 & 2.7656 & 0.104 & 0.168 & 0.224 & 0.231 & 0.220 & 0.225 & 0.644 & -1.60 & -1.39 & 116 & \\
2566 & { \rm ab } & 0.583517 & 9190.8800 & 20.234 & 4.7959 & 0.018 & 0.180 & 0.248 & 0.235 & 0.247 & 0.243 & 0.901 & -1.38 & -1.15 & ~88 & \\ 2627 & { \rm ab } & 0.575091 & 9166.8885 & 20.130 & 4.7959 & 0.733 & 0.192 & 0.136 & 0.157 & 0.146 & 0.146 & 0.239 & -1.96 & -1.78 & 108 & \\
2689 & { \rm ab } & 0.511352 & 8809.5668 & 20.139 & 3.7485 & 0.805 & 0.270 & 0.155 & 0.206 & 0.159 & 0.173 & 0.909 & -1.37 & -1.15 & ~99 & \\ 3004 & { \rm ab } & 0.715496 & 8809.5055 & 20.044 & 2.7656 & 0.990 & 0.187 & 0.199 & 0.183 & 0.192 & 0.191 & 0.514 & -1.72 & -1.52 & 129 & \\
3016 & { \rm c } & 0.360325 & 8809.8443 & 20.117 & 2.7656 & 0.220 & 0.139 & 0.279 & 0.227 & 0.172 & 0.226 & 0.414 & -1.81 & -1.61 & 131 & { \rm ( d)}\\ 3019 & { \rm ab } & 0.732965 & 9162.9250 & 19.978 & 3.7485 & 0.472 & 0.202 & 0.112 & 0.094 & 0.138 & 0.115 & 0.109 & -2.07 & -1.90 & 105 & \\
3024 & { \rm ab / bl ? } & 0.572818 & 9223.7730 & 20.059 & 3.7485 & 0.052 & 0.179 & 0.197 & 0.229 & 0.216 & 0.214 & 0.641 & -1.61 & -1.40 & 143 & { \rm ( f)}\\ 3043 & { \rm ab } & 0.623908 & 8810.0604 & 20.192 & 2.7656 & 0.650 & 0.178 & 0.201 & 0.199 & 0.204 & 0.201 & 0.529 & -1.70 & -1.50 & 129 & \\ 3044 & { \rm d } & 0.354271 & 8810.0791 & 20.126 & 3.7485 & 0.044 & 0.092 & 0.319 & 0.283 & 0.227 & 0.276 & 0.355 & -1.86 & -1.67 & 133 & { \rm ( l)}\\ 3104 & { \rm d } & 0.356984 & 8809.9225 & 20.209 & 2.7656 & 0.982 & 0.243 & 0.182 & 0.171 & 0.169 & 0.174 & 0.745 & -1.51 & -1.30 & 124 & { \rm ( l)}\\ 3113 & { \rm ab } & 0.593260 & 8809.6681 & 20.098 & 2.7656 & 0.049 & 0.121 & 0.296 & 0.244 & 0.256 & 0.265 & 0.556 & -1.79 & -1.59 & 130 & \\ & & & & & 3.7485 & 0.706 & 0.216 & 0.152 & 0.134 & 0.130 & 0.139 & 0.319 & & & 116 & \\ 3125 & { \rm ab } & 0.533130 & 8809.9264 & 20.120 & 2.7656 & 0.220 & 0.201 & 0.179 & 0.160 & 0.154 & 0.164 & 0.404 & -1.81 & -1.62 & 134 & \\
3143 & { \rm d } & 0.354531 & 9184.9270 & 20.188 & 2.7656 & 0.607 & 0.170 & 0.276 & 0.236 & 0.251 & 0.255 & 0.914 & -1.36 & -1.14 & 140 & { \rm ( l)}\\ 3302 & { \rm ac } & 1.346056 & 9227.8700 & 18.560 & 4.8256 & 0.506 & 0.295 & 0.101 & 0.116 & 0.124 & 0.114 & 0.553 & -1.78 & -1.58 & ~81 & \\ & & & & & 6.8001 & 0.975 & 0.089 & 0.289 & 0.294 & 0.251 & 0.278 & 0.340 & & & ~84 & \\
3318 & { \rm ab } & 0.640243 & 9169.8898 & 20.148 & 4.8256 & 0.530 & 0.221 & 0.086 & 0.108 & 0.131 & 0.108 & 0.165 & -2.02 & -1.85 & ~79 & \\ 3319 & { \rm ab / bl } & 0.564984 & 9223.8400 & 20.012 & 4.8256 & 0.703 & 0.229 & 0.172 & 0.148 & 0.153 & 0.158 & 0.524 & -1.71 & -1.51 & ~89 & \\ 3320 & { \rm c } & 0.282469 & 9192.9040 & 20.119 & 4.8256 & 0.954 & 0.067 & 0.341 & 0.338 & 0.300 & 0.326 & 0.403 & -1.94 & -1.76 & ~84 & \\ & & & & & 6.8001 & 0.944 & 0.039 & 0.358 & 0.338 & 0.290 & 0.329 & 0.115 & & & ~66 & \\ 3346 & { \rm c } & 0.357547 & 9168.9247 & 20.125 & 4.8256 & 0.857 & 0.088 & 0.244 & 0.216 & 0.189 & 0.216 & -0.043 & -2.02 & -1.84 & ~96 & \\ & & & & & 6.8001 & 0.379 & 0.141 & 0.214 & 0.243 & 0.203 & 0.220 & 0.394 & & & 120 & \\ 3365 & { \rm ab } & 0.668088 & 9235.8929 & 20.105 & 6.8001 & 0.807 & 0.196 & 0.126 & 0.157 & 0.119 & 0.134 & 0.187 & -2.01 & -1.82 & ~99 & \\
3413 & { \rm c } & 0.359537 & 9167.9120 & 20.132 & 4.8256 & 0.607 & 0.194 & 0.247 & 0.206 & 0.181 & 0.211 & 0.727 & -1.72 & -1.52 & ~90 & \\ & & & & & 6.8001 & 0.098 & 0.103 & 0.299 & 0.230 & 0.220 & 0.250 & 0.285 & & & ~85 & \\ 3710 & { \rm bin ? } & 0.473839 & 9192.9030 & 19.856 & 3.8132 & 0.070 & 0.046 & 0.215 & 0.205 & 0.182 & 0.201 & -0.420 & -2.56 & -2.42 & ~73 & { \rm ( f)}\\ & & & & & 4.8256 & 0.207 & 0.077 & 0.201 & 0.159 & 0.194 & 0.185 & -0.270 & & & ~85 & \\ & & & & & 6.8001 & 0.374 & 0.029 & 0.159 & 0.186 & 0.140 & 0.162 & -0.646 & & & ~89 & \\ 3827 & { \rm ab } & 0.587708 & 9226.7339 & 20.135 & 6.8001 & 0.644 & 0.212 & 0.147 & 0.071 & 0.122 & 0.114 & 0.152 & -2.04 & -1.86 & ~67 & \\
3834 & { \rm c } & 0.375237 & 9189.9150 & 20.003 & 3.7809 & 0.125 & 0.169 & 0.179 & 0.158 & 0.172 & 0.170 & 0.245 & -1.84 & -1.65 & ~96 & \\ & & & & & 4.8256 & 0.909 & 0.164 & 0.204 & 0.205 & 0.184 & 0.198 & 0.400 & & & ~85 & \\ & & & & & 6.8001 & 0.171 & 0.195 & 0.184 & 0.164 & 0.184 & 0.177 & 0.463 & & & ~90 & \\ 3862 & { \rm c } & 0.294092 & 9188.9188 & 20.219 & 3.7809 & 0.622 & 0.087 & 0.307 & 0.318 & 0.268 & 0.298 & 0.442 & -1.78 & -1.58 & 121 & \\ 3907 & { \rm ab } & 0.583204 & 9169.9000 & 20.203 & 3.8132 & 0.683 & 0.269 & 0.126 & 0.093 & 0.137 & 0.119 & 0.464 & -1.76 & -1.56 & 112 & \\
4263 & { \rm c } & 0.284768 & 8820.7399 & 20.229 & 2.7333 & 0.861 & 0.208 & 0.284 & 0.283 & 0.268 & 0.279 & 1.493 & -0.85 & -0.60 & 111 & \\
4277 & { \rm c } & 0.306303 & 8809.7972 & 20.172 & 2.7333 & 0.136 & 0.113 & 0.332 & 0.254 & 0.251 & 0.279 & 0.576 & -1.79 & -1.60 & 115 & \\ & & & & & 3.8132 & 0.661 & 0.100 & 0.245 & 0.277 & 0.239 & 0.253 & 0.284 & & & ~96 & \\ 4291 & { \rm c } & 0.387992 & 8810.1654 & 20.057 & 2.7656 & 0.282 & 0.185 & 0.209 & 0.238 & 0.202 & 0.217 & 0.700 & -1.64 & -1.43 & 105 & \\ & & & & & 3.8132 & 0.982 & 0.186 & 0.206 & 0.191 & 0.180 & 0.192 & 0.514 & & & 106 & \\
4308 & { \rm c } & 0.358956 & 8809.7147 & 20.123 & 2.7656 & 0.364 & 0.112 & 0.229 & 0.240 & 0.170 & 0.213 & 0.122 & -1.83 & -1.64 & 118 & \\ & & & & & 3.8132 & 0.283 & 0.169 & 0.230 & 0.220 & 0.226 & 0.225 & 0.649 & & & 130 & \\
\noalign{\smallskip } \hline \end{array}\ ] ] @xmath39 rcccccccccccrccclp0.5ccccccccccccccccc & type & p & epoch & v & hjd & & k & h _ & h _ & h _ & h & m.i . &
[ fe / h ] & [ fe / h ] & v_r & notes + ( a ) & ( b ) & ( b ) & ( b ) & ( b ) & ( c ) & & & & & & & & zw & cg & km s^-1 & + 4313 & ab & 0.731064 & 8820.6952 & 20.111 & 2.7333 & 0.298 & 0.231 & 0.202 & 0.150 & 0.152 & 0.168 & 0.617 & -1.63 & -1.42 & 119 & + 4385 & ab & 0.487413 & 8809.8949 & 20.191 & 2.7656 & 0.550 & 0.200 & 0.108 & 0.190 & 0.183 & 0.160 & 0.376 & -1.87 & -1.68 & 117 & + & & & & & 3.8132 & 0.699 & 0.076 & 0.328 & 0.298 & 0.253 & 0.293 & 0.311 & & & 103 & + 4689 & ab & 0.639202 & 8809.9119 & 20.024 & 2.7656 & 0.843 & 0.178 & 0.150 & 0.137 & 0.183 & 0.157 & 0.221 & -1.98 & -1.79 & 108 & + 4747 & ab & 0.591979 & 9166.9255 & 20.181 & 2.7333 & 0.246 & 0.168 & 0.234 & 0.201 & 0.198 & 0.211 & 0.534 & -1.70 & -1.50 & 120 & + 4780 & ? & 0.391270 & 8810.1007 & 19.958 & 2.7656 & 0.164 & 0.252 & 0.061 & 0.081 & 0.103 & 0.082 & 0.160 & -2.03 & -1.85 & 110&(m ) + 4785 & ab & 0.506109 & 8809.4911 & 20.139 & 2.7656 & 0.935 & 0.069 & 0.267 & 0.179 & 0.136 & 0.194 & -0.283 & -2.42 & -2.27 & 63 & + 4786 & ab & 0.537290 & 9226.9054 & 20.102 & 2.7333 & 0.365 & 0.263 & 0.146 & 0.172 & 0.185 & 0.168 & 0.817 & -1.45 & -1.23 & 138 & + 4793 & ab & 0.559834 & 8809.1465 & 20.059 & 2.7333 & 0.839 & 0.131 & 0.163 & 0.158 & 0.142 & 0.155 & -0.068 & -2.23 & -2.06 & 133 & + 5000 & c & 0.323261 & 9166.8850 & 20.158 & 4.8256 & 0.543 & 0.136 & 0.192 & 0.222 & 0.182 & 0.199 & 0.207 & -1.99 & -1.81 & 93 & + 5049 & ab & 0.648752 & 8825.8195 & 20.076 & 4.8256 & 0.392 & 0.234 & 0.151 & 0.115 & 0.160 & 0.142 & 0.444 & -1.78 & -1.58 & 96 & + 5065 & c & 0.380212 & 9222.7846 & 19.989 & 3.7809 & 0.931 & 0.101 & 0.322 & 0.275 & 0.231 & 0.276 & 0.440 & -1.74 & -1.54 & 125 & + & & & & & 4.8256 & 0.678 & 0.176 & 0.251 & 0.197 & 0.167 & 0.205 & 0.547 & & & 116 & + 5105 & ab & 0.556682 & 9166.8930 & 20.154 & 4.8256 & 0.847 & 0.255 & 0.162 & 0.108 & 0.168 & 0.146 & 0.588 & -1.65 & -1.45 & 132 & + 5123 & c & 0.325323 & 8810.9530 & 20.162 & 4.8256 & 0.327 & 0.074 & 0.325 & 0.323 & 0.312 & 0.320 & 0.441 & -1.78 & -1.59 & 95 & + 5155 & ab & 0.567457 & 8823.8298 & 20.090 & 4.8256 & 0.611 & 0.241 & 0.081 & 0.104 & 0.167 & 0.117 & 0.316 & -1.89 & -1.70 & 108 & + 5330 & ab & 0.645766 & 8809.3652 & 20.136 & 3.7809 & 0.976 & 0.210 & 0.170 & 0.202 & 0.214 & 0.195 & 0.704 & -1.55 & -1.34 & 110&(f ) + 5343 & ab / bl ? & 0.546964 & 9164.9231 & 20.181 & 4.8256 & 0.149 & 0.190 & 0.158 & 0.147 & 0.162 & 0.156 & 0.282 & -1.92 & -1.74 & 108 & + 5359 & ab & 0.670952 & 8809.7021 & 19.972 & 3.7809 & 0.375 & 0.121 & 0.108 & 0.115 & 0.150 & 0.124 & -0.247 & -2.39 & -2.23 & 93 & + 5364 & c & 0.394203 & 9168.9185 & 19.943 & 3.7809 & 0.624 & 0.044 & 0.238 & 0.268 & 0.231 & 0.246 & -0.248 & -2.39 & -2.23 & 135&(f ) + 5376 & c & 0.389520 & 9187.8882 & 20.079 & 3.7809 & 0.315 & 0.074 & 0.224 & 0.167 & 0.164 & 0.185 & -0.287 & -2.42 & -2.27 & 137 & + 5382 & ab & 0.595937 & 9191.9233 & 20.171 & 3.7809 & 0.708 & 0.249 & 0.121 & 0.136 & 0.111 & 0.123 & 0.389 & -1.83 & -1.63 & 99 & + 5393 & c & 0.322741 & 9166.9255 & 20.081 & 3.7809 & 0.587 & 0.153 & 0.221 & 0.185 & 0.183 & 0.196 & 0.313 & -1.89 & -1.71 & 144 & + 5492 & ab & 0.528786 & 8833.9145 & 20.161 & 3.7809 & 0.065 & 0.171 & 0.222 & 0.205 & 0.190 & 0.206 & 0.515 & -1.72 & -1.52 & 115 & + 5710 & c & 0.355803 & 9227.7927 & 20.132 & 3.7809 & 0.997 & 0.133 & 0.234 & 0.205 & 0.171 & 0.203 & 0.219 & -1.98 & -1.79 & 141 & + 5714 & c & 0.292913 & 8809.7224 & 20.221 & 2.7333 & 0.747 & 0.113 & 0.267 & 0.280 & 0.228 & 0.258 & 0.434 & -1.79 & -1.59 & 84 & + 5723 & ab & 0.566020 & 9166.9255 & 20.250 & 2.7333 & 0.801 & 0.248 & 0.138 & 0.135 & 0.147 & 0.140 & 0.505 & -1.73 & -1.53 & 112 & + 5724 & ab & 0.498508 & 8809.8649 & @xmath31 & 2.7333 & 0.400 & 0.009 & 0.326 & 0.333 & 0.298 & 0.319 & -0.265 & -2.40 & -2.25 & 104&(f ,
n ) + 5747 & ab & 0.559848 & 8833.9264 & 20.040 & 2.7333 & 0.388 & 0.176 & 0.172 & 0.145 & 0.153 & 0.156 & 0.206 & -1.99 & -1.81 & 126 & + 5751 & c & 0.397328 & 8809.6162 & 20.083 & 2.7333 & 0.756 & 0.112 & 0.267 & 0.191 & 0.208 & 0.222 & 0.177 & -1.99 & -1.81 & 97 & + & & & & & 3.7809 & 0.393 & 0.072 & 0.340 & 0.276 & 0.245 & 0.287 & 0.233 & & & 118 & + 5773 & ab & 0.508760 & 8823.9310 & 20.223 & 3.7809 & 0.931 & 0.084 & 0.272 & 0.256 & 0.223 & 0.250 & 0.116 & -2.07 & -1.89 & 91 & + 5802 & ab & 0.514632 & 8810.9385 & 20.244 & 3.7809 & 0.788 & 0.215 & 0.107 & 0.121 & 0.142 & 0.123 & 0.221 & -1.98 & -1.79 & 150 & + 6031 & c & 0.326878 & 9168.9123 & 20.124 & 2.7333 & 0.709 & 0.172 & 0.244 & 0.216 & 0.203 & 0.221 & 0.637 & -1.61 & -1.40 & 140&(f ) + 6048 & ab & 0.626791 & 9224.8920 & 20.168 & 2.7333 & 0.964 & 0.167 & 0.229 & 0.221 & 0.220 & 0.223 & 0.623 & -1.62 & -1.41 & 117 & + 6050 & c & 0.305175 & 8809.7002 & 20.121 & 2.7333 & 0.227 & 0.151 & 0.248 & 0.228 & 0.180 & 0.219 & 0.462 & -1.76 & -1.57 & 139 & + 6085 & c & 0.361528 & 9235.7159 & 20.146 & 2.7333 & 0.808 & 0.165 & 0.293 & 0.243 & 0.228 & 0.255 & 0.869 & -1.40 & -1.18 & 139 & + @xmath39 identifiers are from @xcite epochs are @xmath312440000 . along with types , periods , and mean magnitudes ( in intensity average ) they were redetermined from the study of the light curves based on data from @xcite . in a number of cases ( marked with notes ) they differ significantly from values published in @xcite .
hjds are @xmath312452850 , they correspond to the hjd at half exposure .
@xcite classifies star 59 , 1439 , and 3016 as suspected double - mode rr lyrae stars , with periodicities of 0.35968/0.4837 ; 0.35604/0.47809 ; and 0.36033/0.48401 respectively .
our study of the light curves does not confirm these findings , we think these stars are monoperiodic _ c-_type variables with noisy light curves .
we think this is a double - mode rr lyrae , we list the first overtone period .
classification and periods differ significantly from @xcite who published aliases of the periodicities preferred here
. noisy spectrum .
we think this rr lyrae could either be a double - mode or be affected by the blazhko effect ( bl , blazhko 1907 ) .
g band visible .
@xcite classifies these variables as double - mode rr lyrae stars respectively with periodicities of 0.35427/0.47546 ( 3044 ) , 0.35699/0.47975 ( 3104 ) , and 0.35453/0.47630 ( 3143 ) .
variable star of unknown type . incomplete light curve .
metal abundances for the ab - type rr lyrae stars in sculptor have been estimated by @xcite using the parameters of the fourier decomposition of the light curves and the @xcite method .
derived metallicities are in the range from @xmath40 to @xmath14 dex , with an average value of [ fe / h]@xmath411.5 dex in @xcite metallicity scale . based on the frequency analysis of @xcite data
, @xcite also identified 15 confirmed and 3 suspected double mode pulsators in sculptor and estimated their metal abundance using the pulsation equations .
he found that , similarly to the rrab s , the bulk of the double - mode rr lyrae stars in sculptor have [ fe / h]@xmath42 , with only two rrd s ( stars 1168 and 5354 in @xcite ) at [ fe / h]=@xmath43 .
the metallicity distributions of rrab and rrd stars separately are shown in fig .
5 of @xcite , and can be compared with our distribution in fig .
[ f : fig3hbw ] .
we recall that the @xcite metallicity scale is on average 0.2 dex more metal rich than the zw84 scale at [ fe / h]@xmath42 .
this partially accounts for the difference between average values in @xcite and in our distribution in fig . 4 , but there still is a residual difference of 0.17 dex between the average values , and our distribution appears to cover a metallicity range larger than in @xcite .
we determined metallicities for six of the rrd variables discovered by @xcite , three of which are actually only suspected rrd s . from our re - analysis of the light curves we think that these stars are monoperiodic c - type rr lyraes with noisy light curves .
the comparison between individual metallicity values is shown in table [ t : tabrrd ] , where the range given in column 4 was evaluated from the metallicity distribution of the rrd stars in fig .
5 of @xcite .
unfortunately we did not observe the two most metal poor rrd s in @xcite . [ cols=">,^,^,^,^,^ " , ] [ t : tabbin ] the luminosity - metallicity relation of the sculptor rr lyrae stars is shown in fig .
[ f : fig3 t ] .
it is based on 105 stars covering the metallicity range [ fe / h ] from @xmath44 to @xmath45 , and we used different symbols for the various types of rr lyrae stars .
all variables seem to follow the same luminosity - metallicity relation independent of type .
@xcite found that the lmc double - mode rr lyrae stars are offset to brighter luminosities in the luminosity - metallicity plane and explain this evidence with the lmc rrd s being more evolved than the single - mode pulsators .
the lack of a similar difference in luminosity between single and double - mode sculptor rr lyrae stars suggests that in this galaxy also the single - mode variables are evolved .
a more striking difference is the slope of the luminosity - metallicity relation , significantly shallower than that obtained for the lmc variables by @xcite values . ] . on the other hand
, we note that also the scatter in the average apparent luminosity of the rr lyrae stars in sculptor is less than half of that observed for the variables in the lmc : @xmath46(sculptor)=0.07 mag , compared to the 0.15 - 0.16 mag in the lmc ( @xcite ) .
this scatter is almost entirely accounted for by photometric errors and the dispersion in metallicity of the sculptor rr lyrae stars , thus indicating that these variables have a very similar degree of evolution off the zero age horizontal branch .
the shallow slope of the luminosity - metallicity relation in sculptor could thus be explained with the sculptor rr lyrae s being all rather evolved stars arising from the old metal poor population that also gives origin to the blue hb in the galaxy . some support to
this suggestion arises from the comparison of the average luminosity of the rr lyrae stars with that of non variable horizontal branch stars in sculptor . from the color - magnitude diagram of sculptor , values of @xmath47 20.18 - 20.20 mag , @xmath48 20.20 and @xmath49 20.29 mag
are obtained by @xcite , babusiaux , gilmore , & irwin ( 2005 ) , and @xcite , respectively .
this is at least @xmath500.06 - 0.08 mag fainter than the average luminosity of the rr lyrae stars which , for the 106 rr lyrae with full light curve coverage , is @xmath51=20.127 with a dispersion of @xmath52=0.072 ( this value agrees well with the estimate by @xcite , based on the total sample of 226 rr lyrae stars in sculptor ) .
radial velocities were measured from the spectra of the sculptor rr lyrae stars , and are given in column 16 of table [ t : sculptor ] .
multiple observations of the same stars show that our estimates have typical errors of about @xmath53 km s@xmath6 , with no systematic differences for different masks .
this error includes the contribution of measurement uncertainties , errors related to the centering of the stars in the slit , and uncertainties due to the poor sampling of the radial velocity curves of the variable stars .
the average heliocentric radial velocity of all our variable stars in sculptor is @xmath54=@xmath55 1.9 km s@xmath6
( r.m.s.=19.9 km s@xmath6 , 110 stars , having preliminarily averaged individual values for stars with multiple observations ) .
the average radial velocity of the rr lyrae stars alone is @xmath54=@xmath56 km s@xmath6 ( r.m.s.=19.8 km s@xmath6 , 107 stars ) .
these values are in excellent agreement with the estimate obtained from k - giants in sculptor by queloz , dubath & pasquini ( 1995 : @xmath54=@xmath57 km s@xmath6 ) and are consistent with the value of @xmath54=@xmath58 km s@xmath6 previously derived by armandroff & da costa ( 1986 ) .
the good agreement with the literature values suggests that the offcentering problems noted for the calibrating cluster variables ( see appendix b ) do not seem to affect the variables in sculptor and that the undersampling of the variable star s radial velocity curves does not significantly bias our estimates of the average @xmath59 value . to further check this point we extracted from the database of the galactic field rr lyrae stars analyzed with the baade - wesselink method ( @xcite ) template radial velocity curves of _ ab- _ and _ c-_type rr lyrae stars with metal abundance comparable to that of the variable stars in sculptor , and estimated phase - dependent radial velocity corrections for each spectrum of rr lyrae star .
using this procedure we found that the average correction to apply to the radial velocity mean value of the 107 rr lyrae stars in our sample is less than @xmath80.1 km s@xmath6 and can be safely neglected .
the difference between r.m.s .
scatter of the rr lyrae stars ( 19.8 km s@xmath6 ) and typical measurement errors ( 15 km s@xmath6 ) implies an intrinsic radial velocity dispersion of @xmath52=12.9 km s@xmath6 for the rr lyrae stars .
this value is larger than the @xmath60 km s@xmath6 and @xmath61 km s@xmath6 found for the k - giants by armandroff & da costa ( 1986 ) and @xcite , and for sculptor metal rich red giant stars by tolstoy et al .
( 2004 ; @xmath62=@xmath63 km s@xmath6 ) , but is consistent with the @xmath64 km s@xmath6 dispersion observed by these same authors for the metal poor red giants in sculptor .
this result gives further support to the hypothesis that the rr lyrae stars in sculptor arise from the old , metal - poor population giving origin to the galaxy blue - horizontal branch , although our value for their velocity dispersion needs to be confirmed by higher resolution spectroscopy .
the average metallicity , metal abundance distribution and range in metal abundance spanned by the rr lyrae stars can be compared with the analogous quantities for other old stellar components in sculptor , namely with the metallicity spread inferred from the width of the red giant branch ( e.g. @xcite ) , and with the abundances directly measured for red giants in sculptor by @xcite and @xcite , respectively . the metallicity distribution in fig .
[ f : fig3hbw ] shows that the rr lyrae stars in sculptor cover a full metallicity range of about 1.6 dex , which however reduces to @xmath8 1 dex if the single most metal rich star in the sample is discarded .
this range is larger than inferred from the spread of the red giant stars ( @xmath65 } \simeq$ ] 0.6 dex , @xcite ; @xmath500.8 dex , @xcite ; @xmath500.7 dex , @xcite ) , and is consistent with the spectroscopic study of red giants in sculptor by geisler et al .
( 2005 : @xmath501.1 dex ) and @xcite in the galaxy inner region ( see upper panel of their fig .
@xcite find that the ancient stellar component ( @xmath66 10 gyr old ) in sculptor is divided into two distinct groups having different metal abundance , kinematics and spatial distribution .
the metal - rich population is concentrated within the @xmath67 degree central region , has metallicities in the range @xmath68[fe / h]@xmath69 , velocity dispersion of @xmath62=7@xmath331 km s@xmath6 and is related to the sculptor red horizontal branch .
the metal - poor population is more spatially extended , has metallicities in the range @xmath70[fe / h]@xmath71 , velocity dispersion of @xmath72=11@xmath331 km s@xmath6 and is related to the sculptor blue horizontal branch .
our rr lyrae stars are located in the central region of sculptor , virtually coincident with the region where @xcite find segregation of red hb stars . however , their metallicity distribution is dominated by the metal - poor objects with an average value of [ fe / h][email protected] .
the 26 stars with [ fe / h]@xmath731.7 are marked by boxes in fig .
[ f : fatte ] .
their average luminosity @xmath74>-1.7}\rangle$]=20.154 ( @xmath52=0.060 , 26 stars ) is marginally fainter that the average of the remaining 80 stars with [ fe / h]@xmath751.7 ( @xmath74\leq -1.7}\rangle$]=20.118 , @xmath52=0.075 , 80 stars ) , as expected given the shallow slope of the sculptor rr lyrae luminosity - metallicity relation , and is anyway brighter than both the blue and red hbs of the non variable stars , thus indicating that they are evolved objects , like their metal - poor counterpart .
the average radial velocity of the two samples is only marginally different : @xmath76=117.5 ( r.m.s.=18.5 , 26 stars ) for the metal - rich sample and @xmath76=107.0 ( r.m.s.=19.6 , 80 stars ) for the metal - poor stars .
the r.m.s .
scatters are very similar and , once deconvolved for the measurement errors ( 15 km s@xmath6 ) , lead to very similar velocity dispersions of 10.8 and 12.6 km s@xmath6 in agreement with the velocity dispersion measured by @xcite for the metal - poor component associated to sculptor blue hb .
low resolution spectra obtained with fors2 at the vlt have been used to measure individual metal abundances [ fe / h ] and radial velocities for 107 rr lyrae stars in the sculptor dwarf spheroidal galaxy .
metallicities were derived using a revised version of the @xmath27s method @xcite .
the rr lyrae stars in sculptor are predominantly metal - poor with an average metal abundance of [ fe / h]=@xmath1 ( r.m.s.=0.26 dex ) on the zw84 metallicity scale , and only a few outliers having metallicities larger than @xmath77 dex .
the observed metallicity dispersion is larger than the observational errors , thus showing that these variables have a real metallicity spread .
the rr lyrae stars in sculptor are found to follow a luminosity - metallicity relation with a slope of 0.09 mag dex@xmath6 , which is shallower than in the lmc @xcite .
this is explained with the sculptor variable stars being rather evolved from the zero age hb , as also supported by their brighter luminosity compared to the non variable hb stars .
from our spectra we measured an intrinsic velocity dispersion 12.9 km s@xmath6 for the rr lyrae stars , which appears to be in agreement with the dispersion derived by @xcite for metal - poor red giants associated to the blue - hb stars in sculptor .
all these evidences suggest that our rr lyrae sample , and the rr lyrae stars in sculptor in general , are connected to the blue - hb population and arise from the first burst of star formation that produced the galaxy blue metal - poor hb .
they allow to trace and distinguish this older component in the internal regions of the galaxy which are otherwise dominated by the metal - rich and younger red - hb stars .
indeed , this component only produced a few , if any , of the sculptor rr lyrae stars since average luminosity and kinematic properties of the few metal - rich objects in our sample suggest that they are more likely the high metallicity tail of the metal - poor star population rather than objects associated with sculptor red - hb stars .
a special thanks goes to j. kaluzny for providing us the time series photometry of the sculptor variables . we thank the anonymous referee for his / her comments and suggestions .
this research was funded by miur , under the scientific projects : 2002028935 , `` stellar populations in the local group '' ( p.i .
: monica tosi ) and 2003029437 , `` continuity and discontinuity in the milky way formation '' ( p.i . : raffaele gratton ) .
99 armandroff t.e . ,
da costa g.s . , 1986 , aj , 92 , 777 babusiaux c. , gilmore g. , irwin m. , 2005 , mnras , 359 , 985 bersier d. , wood p.r . , 2002 , aj , 123 , 840 cacciari c. , clementini g. , fernley j. , 1992 , apj , 396 , 219 carretta e. , gratton r.g . , 1997 ,
a&as , 121 , 95 clement c.m . , et al . , 2001 , aj , 122 , 2587 clement c.m . , shelton i. , 1997 , aj , 113 , 171 clementini g. , et al .
2000 , aj , 120 , 2054 clementini g. , gratton r.g . , bragaglia a. , carretta e. , di fabrizio l. , maio m. , 2003b , aj , 125 , 1309 clementini g. , gratton r.g .
, bragaglia a. , ripepi v. , martinez fiorenzano a.f . , held e.v . , carretta e. , 2005 , apjl , in press ( astro - ph/0508079 ) clementini g. , held e.v . ,
baldacci l. , rizzi l. , 2003a , apj , 588 , l85 coutts c.m . , sawyer hogg h. , 1971 , publ .
david dunlap obs . 3 , 61 dolphin a.e . , et al . ,
2001 , apj , 550 , 554 da costa g.s . ,
apj , 285 , 483 di fabrizio l. 1999 , laurea degree thesis , university of bologna fernley j. , a&a , 284 , l16 geisler d. , smith v.v . , wallerstein g. , gonzales g. , charbonnel c. , 2005 , aj , 129 , 1428 gratton r.g . , bragaglia a. , clementini g. , carretta e. , di fabrizio l. , maio m. , taribello e. , 2004 , aa , 421 , 937 harbeck d. , et al . , 2001 ,
aj , 122 , 3092 held e.v .
, clementini g. , rizzi l. , momany y. , saviane i. , di fabrizio l. , 2001 , apj , 562 , l39 held e.v . , saviane i. , momany y. , carraro g. , 2000 , apj , 530 , l85 hurley - keller d. , mateo m. , grebel e.k . , 1999 ,
apj , 523 , l25 jones r.v . ,
carney b.w . , storm j. , latham d.w . , 1992 , apj , 386 , 646 jurcsik j. , 1995 , aca , 45 , 653 jurcsik j. , kovcs g. , 1996 , aa , 312 , 111 lee j .- w . , carney b.w . , 1999 ,
aj , 117 , 2868 liu t. , janes k. a. , 1990 , apj , 354 , 273 kaluzny j. kubiak m. , szymaski m. , udalski a. , krzemiki , mateo m. , 1995 , a&as , 112 , 407 kovcs g. , 2001 , a&a , 375 , 469 majewski s. , siegel m.h . , patterson r.j . , rood r.t . , 1999 , apj , 520 , l33 norris j. , bessell m.s . , 1978 , apj , 225 , l49 preston g.w . , 1959 , apj , 130 , 507 pritzl b.j . , armandroff t.e . , jacoby g.h . , da costa g.s . , 2002 , aj , 124 , 1464 queloz d. , dubath p. , pasquini l. , 1995 , a&a , 300 , 31 rizzi l. 2003 , research doctorate thesis , university of padua rizzi l. , held e. v. , bertelli g. , saviane i. , 2004 , mmsai , 75 , 110 siegel m. , majewski s. , 2000 , aj , 120 , 284 silbermann n.a . ,
smith h.a . , 1995 ,
aj , 110 , 704 skillen i. , fernley j. , stobie r.s . ,
jameson r.f . , 1993 , mnras , 265 , 301 tolstoy e. , irwin m.j . ,
cole a.a . , pasquini l. , gilmozzi r. , gallagher j.s . , 2001 , mnras , 327 , 918 tolstoy e. , venn k. a. , shetrone m. , primas f. , hill v. , kaufer a. , szeifert t. , 2003 , aj , 125 , 707 tolstoy e. , et al . , 2004 ,
apj , 617 , l119 zinn r. , west m.j . , 1984 ,
apjs , 55 , 45
we present in this section finding charts ( fig.s 6a , b , c ) for all the variable stars we observed in sculptor .
they correspond to the nine @xmath25 fors2 subfields used to map the central @xmath78 area of the galaxy .
centre of field coordinates are provided in table [ t : tablelog ] . in each map north
is up and east to the left , and variables are identified according to @xcite identifiers .
the rr lyrae stars are marked by open circles ( in red in the electronic edition of the journal ) , other types of variables by ( blue ) open squares .
the complete listing of the variables observed in sculptor is provided in table [ t : sculptor ] .
equatorial coordinates for all our targets can be found in table 2 of @xcite .
following the procedure devised by @xcite , line indices for rr lyrae stars are computed from the spectra shifted to rest wavelength by directly integrating the instrumental fluxes in spectral bands centered on the ca ii k , h@xmath79 , h@xmath80 , and h@xmath23 lines ( see table 2 and fig .
10 of @xcite for the definition of the spectral bands ) .
then a @xmath28 index is defined as the average of the indices of the 3 hydrogen lines , and @xmath29 as the index of the ca ii k line .
metallicities are derived by comparing the @xmath28 and @xmath29 indices measured for the target stars to the same quantities for variables in a number of globular clusters of known metal abundance .
the calibration of the line indices of the sculptor variables in terms of metal abundances [ fe / h ] was obtained using rr lyrae stars in the newly observed clusters m15 , m2 and ngc6171 , and in the calibrating clusters used in @xcite , namely m68 , ngc1851 and ngc3201 . for all these clusters , precise metal abundances are available on both the zw84 and cg97 metallicity scales .
ngc6441 , a metal rich cluster ( [ fe / h]=@xmath81 according to zw84 metallicity scale ) having rr lyrae stars with anomalously long periods , was not used for calibration purposes and will be discussed elsewhere ( see @xcite ) .
correlation between @xmath29 and @xmath28 spectral indices for the calibrating clusters m15 , m2 , and ngc6171 observed in the present run .
solid lines represent the mean lines for the calibrating clusters m68 , ngc3201 , and ngc1851 , from @xcite .
the calibrating clusters define mono - metallic correlations in the @xmath29 _ versus _
@xmath28 plane .
these relations are shown in fig .
7 where filled squares represent stars in m15 , m2 and ngc6171 , and solid lines are the mean relations defined by m68 , ngc1851 and ngc3201 taken from @xcite .
rr lyrae stars in m15 fall precisely on the mean line of m68 and those in m2 closely follow the mean line of ngc3201 , confirming the good agreement between metal abundances of these two pairs of clusters .
stars in ngc6171 generally fall slightly above the mean line of ngc1851 , in agreement with the slightly higher metallicity of ngc6171 with respect to ngc1851 . according to @xcite the mean relations drawn in fig .
7 by m68 and ngc1851 are : @xmath82 ( valid for @xmath28 between 0.12 and 0.40 ) for m 68 , and : @xmath83 ( valid for @xmath28 between 0.04 and 0.34 ) for ngc 1851 .
we thus defined metallicity index @xmath32 the quantity : @xmath84 where @xmath29 is the ca ii k line index of the star , and @xmath85 are derived entering the @xmath28 index measured for the star into equations ( 1 ) and ( 2 ) .
@xmath32 values derived by this procedure for the rr lyrae stars in the calibrating clusters analyzed in the present paper are listed in column 10 tables [ t : m15 ] , [ t : m2 ] , [ t : ngc6171 ] .
the calibration of the metallicity index in terms of metal abundance [ fe / h ] was derived by computing average @xmath86 values for the m15 , m2 and ngc6171 variables from the individual @xmath32 s in tables [ t : m15 ] , [ t : m2 ] , [ t : ngc6171 ] and from the @xmath32 values in column 9 of tables 3 , 4 , and 5 of @xcite for m68 , ngc1851 and ngc3201 , and correlating these @xmath86 s with the metal abundances of the clusters on the zw and cg metallicity scales , respectively .
the average @xmath86 values and their dispersions are summarized in columns 6 and 7 of table [ t : calibration ] along with the cluster metallicities ( and their uncertainties ) in the two metallicity scales ( columns from 2 to 5 ) .
the correlation between [ fe / h ] and @xmath86 values is very well represented by linear regressions with scatter typically within the error of measure .
these linear regressions are described by the following equations : @xmath87_{zw}}= 0.882 \langle m.i.\rangle - 2.170\ ] ] @xmath87_{cg}}= 0.941 \langle m.i.\rangle - 2.000\ ] ] @xmath88 } & { \rm [ fe / h ] } & v_{r}~~\\ { \rm ( a ) } & { \rm ( -2400000 ) } & & { \rm ( b ) } & & & & & & & { \rm zw } & { \rm cg } & { \rm km s^{-1}}\\ \noalign{\smallskip } \hline \noalign{\smallskip } { \rm v10 } & 52851.6744 & { \rm c } & 0.781 & 0.126 & 0.195 & 0.192 & 0.147 & 0.178 & 0.025 & -2.15 & -1.98 & -83~~\\ { \rm v12 } & 52851.6744 & { \rm ab } & 0.645 & 0.159 & 0.148 & 0.155 & 0.122 & 0.142 & 0.027 & -2.15 & -1.98 & -92~~\\ { \rm v17 } & 52851.6744 & { \rm d } & 0.429 & 0.123 & 0.181 & 0.151 & 0.158 & 0.163 & -0.073 & -2.23 & -2.07 & -96~~\\ { \rm v19 } & 52851.6744 & { \rm ab } & 0.592 & 0.090 & 0.272 & 0.247 & 0.208 & 0.242 & 0.121 & -2.06 & -1.89 & -122~~\\ { \rm v23 } & 52851.6744 & { \rm ab } & 0.024 & 0.118 & 0.218 & 0.208 & 0.184 & 0.203 & 0.108 & -2.08 & -1.90 & -118~~\\ { \rm v25 } & 52851.6744 & { \rm ab } & 0.492 & 0.167 & 0.141 & 0.107 & 0.126 & 0.125 & -0.015 & -2.18 & -2.01 & -99~~\\ { \rm v29 } & 52851.6744 & { \rm ab } & 0.967 & 0.146 & 0.170 & 0.145 & 0.133 & 0.149 & -0.002 & -2.17 & -2.00 & -114~~\\ { \rm v42 } & 52851.6744 & { \rm c } & 0.818 & 0.102 & 0.228 & 0.171 & 0.163 & 0.187 & -0.092 & -2.25 & -2.09 & -99~~\\ { \rm v50 } & 52851.6744 & { \rm c } & 0.298 & 0.079 & 0.272 & 0.253 & 0.225 & 0.250 & 0.078 & -2.10 & -1.93 & -104~~\\ \noalign{\smallskip } \hline \end{array}\ ] ] lccccccccccccp0.5cccccccccccc + & hjd & type & & k & h _ & h _ & h _ & h & m.i . &
[ fe / h ] & [ fe / h ] & v_r + ( a ) & ( -2400000 ) & & ( b ) & & & & & & & zw & cg & km s^-1 + & 52856.7503 & ab & 0.758 & 0.194 & 0.221 & 0.219 & 0.193 & 0.211 & 0.722 & -1.53 & -1.32 & -38 + v15 & 52856.7503 & c / d & 0.345 & 0.097 & 0.319 & 0.308 & 0.281 & 0.303 & 0.583 & -1.66 & -1.45 & -25 + v26^c & 52856.7503 & c & 0.428 & 0.196 & 0.224 & 0.223 & 0.177 & 0.208 & 0.716 & -1.54 & -1.33 & -21 + v32^c & 52856.7503 & c & 0.350 & 0.152 & 0.266 & 0.257 & 0.249 & 0.257 & 0.774 & -1.49 & -1.27 & -34 + v7 & 52856.7503 & ab & 0.621 & 0.203 & 0.221 & 0.217 & 0.238 & 0.225 & 0.923 & -1.36 & -1.13 & -36 + v3^d & 52856.7503 & ab & 0.191 & 0.253 & 0.150 & 0.144 & 0.147 & 0.147 & 0.585 & -1.65 & -1.45 & -15 + @xmath89 } & { \rm [ fe / h ] } & v_{r}~~\\ { \rm ( a ) } & { \rm ( -2400000 ) } & & { \rm ( b ) } & & & & & & & { \rm zw } & { \rm cg } & { \rm km s^{-1}}\\ \noalign{\smallskip } \hline \noalign{\smallskip } { \rm v2 } & 52849.5952 & { \rm ab } & 0.746 & 0.330 & 0.113 & 0.128 & 0.117 & 0.119 & 0.771 & -1.49 & -1.27 & -77~\\ { \rm v4 } & 52849.5952 & { \rm c } & 0.385 & 0.220 & 0.283 & 0.241 & 0.219 & 0.247 & 1.280 & -1.04 & -0.80 & -91~\\ { \rm v6 } & 52849.5952 & { \rm c } & 0.570 & 0.214 & 0.256 & 0.284 & 0.205 & 0.248 & 1.239 & -1.08 & -0.83 & -87~\\ { \rm v7 } & 52849.5952 & { \rm ab } & 0.683 & 0.399 & 0.124 & 0.129 & 0.142 & 0.132 & 1.234 & -1.08 & -0.84 & -86~\\ { \rm v9 } & 52849.5952 & { \rm c } & 0.793 & 0.201 & 0.277 & 0.238 & 0.231 & 0.249 & 1.133 & -1.17 & -0.93 & -96~\\ { \rm v21 } & 52849.5952 & { \rm c } & 0.774 & 0.228 & 0.267 & 0.295 & 0.222 & 0.262 & 1.504 & -0.84 & -0.59 & -77~\\ \noalign{\smallskip } \hline \end{array}\ ] ] phases were derived from the hjd of observation , the periods published by @xcite and the epochs we estimated from the study of the light curves published from the same authors .
v2 was not observed by @xcite , for this star the adopted ephemerides , from @xcite , may be no longer valid .
@xmath90 } & { \rm err } & { \rm [ fe / h ] } & { \rm err } & \langle m.i.\rangle & { \rm err } & \langle { \rm [ fe / h]}\rangle & \sigma & \langle { \rm [ fe / h]}\rangle & \sigma & \langle v_{r}\rangle ~~ \\ & { \rm zw } & & { \rm cg } & & & & { \rm this~paper } & & { \rm this~paper } & & { \rm km~s^{-1}}\\ & & & & & & & { \rm zw~scale } & & { \rm cg~scale } & & \\ \noalign{\smallskip } \hline \noalign{\smallskip } { \rm ngc\,1851 } & -1.36 & 0.09 & -1.03 & 0.06 & 1.007 & 0.028 & -1.36\pm 0.02^{\mathrm{a}}&0.10&~-&-&-~~\\ { \rm ngc\,3201 } & -1.61 & 0.12 & -1.24 & 0.03 & 0.710 & 0.030 & -1.56\pm 0.02^{\mathrm{a}}&0.08&~-&-&-~~\\ { \rm ngc\,4590~(m68 ) } & -2.09 & 0.11 & -2.00 & 0.03 & 0.000 & 0.014 & -2.09\pm 0.01^{\mathrm{a}}&0.05&~-&-&-~~\\ { \rm
ngc\,6171 } & -0.99 & 0.06 & -0.95 & 0.04 & 1.194 & 0.098 & -1.12\pm 0.09~&0.21&-0.88\pm 0.09&0.22 & -86 \pm 3\\ % { \rm ngc\,6441 } & -0.59 & 0.15 & - & - & 1.768 & 0.094 & -0.74\pm 0.11~&0.38&-0.47\pm 0.12&0.41 & -1 \pm 7\\ { \rm ngc\,7078~(m15 ) } & -2.15 & 0.08 & -2.02 & 0.04 & 0.020 & 0.025 & -2.15\pm 0.02~&0.06&-1.98\pm 0.02&0.07&-103 \pm 4\\ { \rm ngc\,7089~(m2 ) } & -1.62 & 0.07 & -1.31 & 0.04 & 0.717 & 0.052 & -1.54\pm 0.05~&0.11&-1.33\pm 0.05&0.12 & -28 \pm 4\\ \noalign{\smallskip } \hline \end{array}\ ] ] averages of the values in tables 3,4 , and 5 of @xcite corrected downward respectively by 0.10 , 0.07 and 0.03 dex to account for systematic differences between zw84 and the metallicity scale adopted in @xcite .
individual metal abundances for the rr lyrae stars in the calibrating clusters analyzed in the present paper were derived from the above calibration equations .
they are listed in tables [ t : m15 ] , [ t : m2 ] , [ t : ngc6171 ] respectively , while the mean metallicities derived from the averages of these individual values and their respective dispersions are given in table [ t : calibration ] . the last column in each of tables [ t : m15 ] , [
t : m2 ] , and [ t : ngc6171 ] gives individual radial velocities measured from the spectra of the stars observed in the calibrating clusters .
averages of these values are listed in table [ t : calibration ] .
we note that these average radial velocities differ somewhat from the literature values , particularly for ngc6171 , suggesting the presence of systematic offsets possibly caused by offcentering of the cluster variables in the slit .
these systematic differences are not found in the case of the sculptor variables , for which the average radial velocity we measured is perfectly consistent with the literature values ( see section 5 ) .
this suggests that we are simply seeing an effect of the small samples in the galactic clusters , while the one in sculptor is large enough to average away the offcentering effects . | low resolution spectroscopy obtained with fors2 at the very large telescope ( vlt ) has been used to measure individual metal abundances ( [ fe / h ] ) for 110 variable stars , including 107 rr lyrae stars and 1 anomalous cepheid , and trace the metal distribution of the oldest stellar component in the sculptor dwarf spheroidal galaxy .
the rr lyrae stars are spread over a 15@xmath0 area around the galaxy centre .
their metallicities have an average value of [ fe / h]=@xmath1 ( r.m.s .
0.26 dex ) and cover the metallicity range @xmath2[fe / h]@xmath3 ( on the scale of @xcite ) , but there is only one variable having [ fe / h]@xmath4 . the star - to - star scatter is larger than typical errors on individual metallicities ( @xmath5 dex ) , indicating a real spread in metal abundances .
the radial velocities measured from the rr lyrae spectra have a dispersion of 12.9 km s@xmath6 .
this value is consistent with the dispersion derived by @xcite for metal - poor red giants associated to the blue horizontal branch stars in sculptor . along with the metallicity distribution
these results suggest that most of the rr lyrae stars in sculptor arise from the same burst of stellar formation that produced the metal - poor component giving origin to the galaxy blue horizontal branch .
the metal - rich red horizontal branch population found to be centrally concentrated only produced a few ( if any ) of the rr lyrae stars in our sample .
the spectroscopic metallicities were used along with the apparent luminosities to study the luminosity - metallicity relation followed by the rr lyrae stars in sculptor , for which we derive a shallow slope of 0.09 mag / dex .
this result can be due to a high level of evolution off the zero age horizontal branch of the rr lyrae stars in this galaxy , again in agreement with their origin from the blue horizontal branch population .
[ firstpage ] [ lastpage ] stars : abundances stars : evolution stars : population ii stars : variables : other galaxies : individual : sculptor dwarf spheroidal |
seeking localized solutions to ( non- ) linear differential equations has a very long history , starting probably with the famous observation by john scott russell of a solitary wave in the union canal @xcite .
since then , many a research were made .
an interest to the theoretical study of localized solutions of linear equations was renewed after the discovery of lasers and further progress in technologies of emitting ultra - short pulses . nowadays , there are also numerous potential applications of such solutions , for example , for the localized low - loss energy transmission , communication , medical imaging or remote sensing .
the state of the art in this field , along with its history , is presented in recent books @xcite . in free space ( or non - dispersive media ) the problems of wave propagation are usually described using the wave equation ( we ) [ wave - eq ] ^2_t- = 0 , where @xmath0 , etc . ,
and @xmath1 is three dimensional laplace operator . here and throughout the paper we put the constant of speed of light equal to unity , @xmath2 . on the other hand ,
the waves in dispersive medium are often described by the klein
gordon fock equation ( kgfe ) [ kgfe ] ( + m^2 ) u = 0 , where @xmath3 is a mass parameter having different physical meaning in different systems .
the kgfe is important in studying of electromagnetic waves in the isotropic cold collisionless plasma @xcite , waves of charge density in drude metals , or the high - frequency acoustic waves in the gas of charged particles treated in the hydrodynamic approach @xcite .
a relatively recent spike in activity in investigating localized solutions was promoted by the paper by brittingham @xcite , who outlined a new one parametric family of beam - like localized explicit solutions to the we which he named the focus wave mode ( fwm , which we also call the gaussian beam ) .
this solution is localized in the gaussian way along a straight line in space and has infinite energy .
this solution can not be obtained by the separation of variables in coordinates @xmath4 and time .
however introduction of new variables [ albe ] = z -t , = z + t , where @xmath5 is coordinate along the propagation axis , enables one to do this .
the kgf analogue of the brittingham fwm for the we was given by ziolkowski in @xcite by the separation of variables ( [ albe ] ) .
in @xcite he also suggested to seek other solutions as weighted superpositions over a free parameter of fwm and in doing so he obtained a solution of finite energy and a power
law localization .
another way of construction of highly localized solutions , the so called bidirectional representation , is based on the fourier integral in new variables ( [ albe ] ) .
taking the fourier weight in the proper way besieres , shaarawi and ziolkowski @xcite , donelly and ziolkowski @xcite found new solutions with finite energy of we and kgfe respectively .
all highly localized solutions both for the we and for kgfe spread propagating .
there were found however the so called undispersive solutions , which propagate without spreading , but have a power - law localization from their moving amplitude maximum . for the we these are the bessel beams , solutions found by durnin @xcite by separation of variables in initial space - time coordinates , and the x waves found by weighted superposition of bessel beams @xcite . for the kgfe solutions the same property
is possessed by mackinnon s solution @xcite which can be found by combination of the separation of variables and the lorentz transformations . the particle like exact solution to the we which decreases exponentially in all directions away from a point moving along a straight line was first presented by kiselev and perel in @xcite .
perel and sidorenko @xcite considered this solution from the point of view of wavelet analysis , investigated numerically the uncertainty relation for this solution and found explicitly its fourier transform . in @xcite
the solution was also analyzed from the point of view of complex sources .
it was shown that it can be generated by a pulse source moving with the speed of wave propagation .
an integral representation of this solution in terms of gaussian beams due to ideas of @xcite was given by perel and fialkovsky in the paper @xcite which was however mainly devoted to the kgfe . in @xcite
it was also suggested a class of explicit exponentially localized packet - like solutions for kgfe and investigated some of their properties .
one of solutions from the class obtained in @xcite coincides with one from @xcite .
kiselev , plachenov and chamorro - posada @xcite created astigmatic beam - like and packet - like solutions of the we .
finally , we shall also mention the work by overfelt @xcite who got a class of solutions which generalizes gaussian beams of @xcite and the bessel beams of @xcite and which have better localization near the propagation axis than gaussian beams . in besieres ,
shaarawi and ziolkowksi @xcite there was suggested a new method of design of solutions of three dimensional kgfe reducing them to a solution of one dimensional kgfe equation with new time and coordinate containing an arbitrary function . on this way they found the counterparts of the gauss
bessel pulses @xcite and some other solutions .
intrerested reader can find more detialed refences and discussion of various constructions of the localized solutions both for the we and the kgfe in books @xcite and reviews @xcite .
the present work is a continuation of works @xcite and @xcite .
we present nonseparable solutions with exponential localization , both beam - like and particle - like ones .
the former solutions have gaussian localization near a straight line and traveling wave - fronts .
the latter ones in addition to localization near a line are localized in a gaussian manner near a point moving along this line .
the structure of the paper is following . in the next section
we revisit known results on exact exponentially localized solutions for the wave equation . by doing so
, we also construct the astigmatic generalizations of the known beam like and particle like solutions for the we .
in section [ kgfeqn ] we apply the developed methods to the construction of the multidimensional astigmatic solutions to the klein gordon fock equation .
we proceed by asymptotic investigation of the kgf solutions in section [ asyinv_sect ] where we consider both small / moderate and large times regimes and discuss the choice of parameters which enable us to govern the localization properties of solutions .
we conclude our research by obtaining the fourier transformation of all constructed families of solutions in section [ fourier - analysis ] , and by providing some final remarks and numerical studies of the solutions in section [ discu_sec ] . in appendix [ asymptotic - kgf ] we present the asymptotic investigation of obtained solutions in the fourier domain .
all our results are valid for the space time with any number of spatial dimensions .
in this section we construct a generalizations of the known localized exact solutions to the we _ t^2 - _ n = 0 , [ wave - eqn ] where @xmath6 stands for @xmath7 dimensional laplacian operator . by doing
so we also revive the necessary techniques to be used also for the klein gordon fock equation . in constructing particle like solutions of the wave equation
, we follow the idea by ziolkowski @xcite and seek such solutions in the form of a superposition of gaussian beams .
the exponentially localized solutions we are focusing on may be considered as `` relatively undistorted progressive waves '' by courant and hilbert @xcite , or `` nondispersive waves '' by hillion @xcite .
they are of the form of a ray series which comprises one term only = g(,t)f ( ) , [ rel - dist]where @xmath8 is a solution of the eikonal equation for we ( _ t)^2 - ( ) ^2 = 0 , [ eikon ] function @xmath9 depends on the form of @xmath8 and satisfies two equations ( _ t)(_tg_0 ) - , g_0 + g_0 = 0 , ( _ t g)^2 - ( g)^2 = 0 , [ transp2 ] and @xmath10 is an arbitrary function . here
@xmath11 .
the focus wave modes by brittingham @xcite belonging to the so called bateman hillion class are based on the following eikonal @xcite = + , = ( _ , z ) , [ theta ] where @xmath12 , @xmath13 and @xmath14 are defined in ( [ albe ] ) , and @xmath15 is an arbitrary positive constant .
the function @xmath9 in this case is of the form @xmath16 , @xmath17 . choosing the arbitrary function @xmath10 as a pure exponent @xcite @xmath18 where @xmath19 is a positive parameter we obtain fwm @xcite , or gaussian beam , which reads = .
[ britkis ] we separated imaginary and real parts in the exponent to stress exponential localization of the solution near @xmath5-axis , with @xmath20 we denote the width of the beam in all transverse directions .
such axisymmetric solutions are called the _
stigmatic _ ones .
one of the obvious but far reaching generalizations of the solution is achieved by formal linear transformation of the transverse coordinates in the eikonal function @xmath8 ( see , e.g. , @xcite ) [ theta ] = + ( _ , ( ) _ ) , where @xmath21 is a @xmath22 complex matrix depending on @xmath14 , whose properties are to be discussed later .
astigmatic generalization of can be obtained by putting @xmath23 and is given by @xmath24 where @xmath25 and @xmath26 are arbitrary constants .
the we ( [ wave - eqn ] ) is indeed resolved by ( [ wavebeam ] ) if @xmath27 satisfies the bernoulli equation @xmath28 and therefore reads ( ) = _ 0(*e*+_0)^-1 , [ gammasol ] where @xmath29 is the unity @xmath30 matrix and @xmath31 constant non - degenerate one .
function @xmath32 has no singularities if @xmath33 does not have nonzero real eigenvalues , and it has the gaussian localization around the axis @xmath5 if the matrix @xmath34 is positively defined .
both conditions are fulfilled if @xmath33 has a positive definite imaginary part .
indeed , the regularity condition is fulfilled since @xmath35 for all nonzero vectors @xmath36 ( real or complex ) including eigenvectors of @xmath33 and hence all its eigenvalues have positive imaginary parts . to show the localization of solution for all values of @xmath14 we note that once @xmath37 has negative definite imaginary part , @xmath38 is also negative definite matrix
. then @xmath39 stays positively defined for all @xmath40 .
we remind here that positive definiteness of @xmath41 and negative definiteness of @xmath42 are equivalent : @xmath43 , where @xmath44 . the axisymmetrical stigmatic beam ( [ britkis ] )
may be obtained from ( [ wavebeam ] ) if @xmath45 ( @xmath46 and @xmath47 are real constants ) .
the width of the gaussian curve in the transverse direction depends in this case on the propagation direction @xmath5 and time . for fixed time
the width has a minimum which is called a waist .
the simplest non - axisymmetrical solutions , the _ aligned simple astigmatic _ ones , correspond to a diagonal @xmath21matrix : @xmath48 with @xmath49 and @xmath50 being real constants , @xmath51 .
when @xmath27 can be diagonalized by a orthogonal rotation of axes the solutions are called _ rotated simple astigmatic_. in two dimensional case it is given by [ g0-matr ] = * u * * * * u*^-1 , = ( cc & - + & ) , = ( cc & 0 + 0 & ) .
the level surfaces of the modulus of aligned ( and rotated ) simple astigmatic solution in the transverse space are ellipsoids ( ellipses if the whole space is three dimensional ) .
the directions of the main axes of ellipses , defined by the ( constant ) eigenvectors of @xmath21 , do not depend on time and the coordinate @xmath5 .
the value of the waist of the gaussian curve as well as its position , is different for different axes .
the case of _ general astigmatism _ is caractrezed by the dependence on time and @xmath5 of the direction of the main axes of the localization ellipsoid .
it happens when the eigenvectors of @xmath21 ( which are constant due to ) are complex and do not coincide with those of @xmath52 and @xmath53 .
the real eigenvectors of the latter matrices do depend on @xmath14 .
then the main axes of the ellipsoids ( or hyperboloids ) of constant phase and of modulus levels rotate with time and/or coordinate @xmath5 .
the absolute value of the total angle of rotation is equal to @xmath54 for both of them ( see @xcite or @xcite ) .
there is no definition of the waist in the case of general astigmatism .
arnaud and kogelnik realised @xcite that in two dimensions a general astigmatic solution can be obtained by assigning a complex value to @xmath55 in . indeed , ( [ wavebeam ] ) is still a solution of the wave equation ( [ wave - eqn ] ) in this case but the eigenvectors of @xmath52 and @xmath53 will be different .
the solution with such @xmath21 is localized in the neighborhood of the @xmath5axis if @xcite @xmath56 are positive and @xmath57 .
we note , that the smaller is @xmath58 , the closer is the solution to a simple astigmatic one .
all the above discussion of the astigmatic properties is equally applicable to the particle like solutions of the we of the next section , and to all solutions of the kgfe considered in section [ kgfeqn ] .
we seek the particle - like solutions of eq .
( [ wave - eqn ] ) in the form of a superposition of gaussian beams @xmath32 obtained in ^()_p ( , , _ ) = _
0^d f^ ( ) ( ) _ b ( , , _ , ) , [ fnuwave ] where @xmath59 is a particular function depending on the parameter @xmath60 .
we put f^ ( ) ( ) = ^--1 e^-(+^2/ ) , [ fnu ] where @xmath60 , @xmath61 , and @xmath62 are arbitrary constants , @xmath63 , @xmath64 . such particular choice of the spectral function @xmath65 is motivated by consideration of the fourier transformation of one of the previously known solution to the we , given below in ( [ wavepaexam ] ) . for the first time this spectral weight appears in @xcite in regard of developing the so called ` bidirectional ' representation for the solutions of the we .
it can easily be shown that ( [ fnuwave ] ) is reduced to an integral representation of the macdonald function @xmath66 @xcite ( also called the modified bessel function of the second kind ) [ tabl - in ] _
0^ x^l-1 dx = 2 @xmath67^l/2 k_l@xmath68 , which is valid if @xmath69 , @xmath70 .
we put @xmath71 , @xmath72 @xmath73 and obtain from ^()_p ( x , y , z , t ) = c_p s^k_(s ) , s= 2@xmath74 ^ 1/2 , [ wavepa ] where @xmath8 is given by ( [ theta ] ) , and @xmath75 can be expressed through numerical parameters of ( [ wavebeam ] ) as @xmath76 but can also be treated as an arbitrary numerical constant .
it is worth noting that @xmath77 satisfies the eikonal equation ( [ eikon ] ) for wave equation ( [ wave - eqn ] ) .
for particular case of @xmath78 and diagonal @xmath79 , the formula ( [ wavepa ] ) yields a family of axisymmetric solutions first presented in @xcite . for @xmath80
it gives ^(1/2)_p ( x , y , z , t ) = ( -i)^-1 , [ wavepaexam ] which was first obtained in @xcite .
it corresponds to a following choice of @xmath10 in : @xmath81 , were @xmath82 is a positive real constant .
all of the solutions have a gaussian localization in vicinity of a point running with wave velocity along a straight line provided the free parameters satisfy for certain relations , similar to those obtained in @xcite .
however their detailed investigation is out of scope of the current paper .
finally we note , that all the solutions of the form of ( [ wavepa ] ) can be interpreted as part of ` arbitrary waveform solutions ' ( [ rel - dist ] ) with @xmath83 and @xmath84 . apart from being
more constructive the integral representation ( [ fnuwave ] ) used here with the weight function ( [ fnu ] ) will be important in construction of the solution for the kgfe , where no waveform freedom is available .
operating only in the space time domain we construct now both beam like and particle like solutions of the multidimensional klein gordon - fock equation ( _ t^2-_n+m^2)u=0 [ kgfn ] where @xmath85 is the laplace operator in @xmath7 spatial dimensions . to elaborate a particle like solution @xmath86 of the klein
fock equation as a superposition of the beam like solutions @xmath87 , we shall first construct the latter ones . in doing so
, we shall consider the solution ( [ wavebeam ] ) of the we in spacial dimension increased by one , @xmath88 , and calculate its fourier transform with respect to @xmath89 u_b(t , z , _ ) _ -^d _ b(t , z , _
, ) e^-i m , [ fourier ] where @xmath32 is given by ( [ wavebeam ] ) in @xmath90 dimensions .
we shall further assume that the enlarged @xmath91 matrix @xmath92 is such that ( _ ^,()_^ ) = ( _ , ( ) _ ) + ^2/(-i_m ) [ zeta ] with a positive constant @xmath93 , and @xmath21 is a @xmath30matrix as before .
then @xmath94 taking the integral and introducing a new numerical constant , @xmath95 , we obtain a non - axisymmetric generalization of solution found in @xcite u_b = c_b i m s_b , s_b = m - , [ kgfbeam ] and @xmath8 is defined in .
the absolute value of this solution does not depend on @xmath96 moving with a unit velocity along the @xmath5axis in the negative direction , and @xmath97 is exponentially localized in transversal directions provided the matrix @xmath21 satisfies the conditions discussed in the section [ wave - equation ] .
as it follows from ( [ gammasol ] ) , @xmath98 for @xmath99 , and thus localization degree around @xmath5axis is decreasing with time and the solution becomes more axisymmetric .
apart from this , the absolute value of the pre - exponential factor is also decreasing as @xmath100 with @xmath99 .
thus , at every given moment of time the solution ( [ kgfbeam ] ) has a gaussian localization along the transversal coordinates and power - law localization along the longitudinal ones .
the total energy of the beam is infinite .
thus , the considered solution is indeed a gaussian beam , one can also call it a focus wave mode for the klein gordon - fock equation . acting along the lines of subsection [ wave - equation].[part_we ] we seek the particle like solutions of the kgfe in the form of a superposition of the beam like solutions @xmath87 u^()_p ( , , _ ) = _ 0^d f^ ( ) ( ) u_b ( , , _ ; ) .
[ superposkgf ] by choosing @xmath59 as ( [ fnu ] ) and plugging it into ( [ superposkgf ] ) , we immediately recognize the same integral representation for the modified bessel function ( [ tabl - in ] ) with @xmath101 and @xmath102 , and arrive at u_p^(-1/2)=c_p @xmath103^ k_@xmath104 , [ kgfpackets ] where we use notation for the complex phase function s_p = @xmath105 ^ 1/2 , [ s_p ] here @xmath106 , @xmath107 and @xmath108 can be treated as new independent parameters .
the square root with the positive real part is assumed in ( [ s_p ] ) .
the constructed family of solutions @xmath86 is nonaxisymmetric mutlidimensional generalizations of the solutions obtained for @xmath78 and @xmath109 in @xcite .
we stress here that is indeed applicable for .
first , @xmath110 is positive since @xmath111 and @xmath14 is real .
secondly , @xmath112 is positive as well , because @xmath113 and @xmath114 , as it follows from its definition ( [ theta ] ) and the fact that the imaginary part of the quadratic form @xmath115 is positively defined by assumption .
we note that for @xmath116 this solution transforms to the localized solution of the wave equation @xmath117 ( [ wavepa ] ) in the account that @xmath118 changes with @xmath3 in such a way that @xmath119 .
let us consider now some particular examples of the constructed solutions .
the modified bessel function reduces to elementary functions if @xmath120 is half - integer . for the values @xmath121 and @xmath122 , we have @xmath123 and the formula ( [ kgfpackets ] ) yields [ simkgf2 ] u^(0)_p = c_2 , u^(-1)_p = c_3 , where @xmath124 and @xmath125 are numerical constants .
we further note , that in two - dimensional space ( i.e. with @xmath126 ) we can reduce the second solution of ( [ simkgf2 ] ) to a function depending on one variable @xmath127 only u^(-1)_p ( s_p ) = c_3 .
[ simplekgf ] in this case , the phase @xmath127 ( [ s_p ] ) is given by s_p = = , with @xmath128 , @xmath129 .
we note that such @xmath127 can be interpreted as a distance in the ( euclidian ) space time with imaginary time @xmath130 where @xmath131 , @xmath132 . from this point of view , ( [ simplekgf ] ) can be thought of as a point source solution @xmath133 of the equation [ helm ] _
3 g - m^2 g = ( * r * ) , where @xmath134 is the three dimensional dirac delta function , @xmath135 is the laplacian in three dimensions .
so , our solution of the kgfe in two spatial dimensions is the point source solution of the elliptic equation ( [ helm ] ) in three dimensions which is analytically continued to the complex plane .
finally , the solutions and may be treated from the point of view of ray method approach discussed for kgfe by maslov @xcite .
phase functions @xmath136 and @xmath127 satisfy the eikonal equation [ eikon - kgfe ] ( _ t s)^2 - ( s)^2 - 1 = 0 and solutions @xmath87 and @xmath86 may be regarded as ray expansions @xmath137 comprising a single term .
we prove here that formula ( [ kgfpackets ] ) gives a family of particle - like solutions of the klein - gordon - fock equation . to do so
, we first note that in all cases when @xmath138 , @xmath139 we can use the following asymptotic expression of the modified bessel function [ as - bes ] k_(m s_p ) _ |s_p| e^-m s_p @xmath140
. then all the solutions of our family behave as u_p^(-1/2 ) _
|s_p| c_p .
[ pack - estim2 ] basing on this expression we will develop in what follows asymptotical expansions of .
we show now that the solution ( [ kgfpackets ] ) has exponential decay at spatial infinity , i.e. for @xmath5 , @xmath141 and finite times , and therefore has a finite energy . to prove the applicability of expression ( [ pack - estim2 ] ) for @xmath142 and fixed time , first we give the estimate from below for the absolute value of @xmath127 |s_p| = |@xmath143@xmath144|^1/2 ( ^2 + ^2)^1/4 ( + |_|^2 h ( ) ) ^1/2= [ abss2 ] @xmath145 at the fisrt line of ( [ abss2 ] ) we have used that @xmath146 and introduced the notation @xmath147 while to proceed to the second line we further used that [ gam - b1 ] = ^-1*e * - and the fact that the imaginary part of @xmath21 is positively defined .
together with the continuity and boundedness of @xmath148 as a function of @xmath14 and @xmath149 , it allows us to conclude that there exists a constant @xmath150 such that for any @xmath14 and @xmath149 @xmath151 this justifies the last inequality of ( [ abss2 ] ) and thus proves that @xmath152 does indeed grow with @xmath153 and/or @xmath141 . on the other hand , since @xmath154 and @xmath155 for any finite @xmath14 we have @xmath156 @xmath157 where @xmath158 and thus [ args1 ] ||= || < /2- .
this already proves applicability of the asymptotical expression ( [ pack - estim2 ] ) .
furthermore , assuming that @xmath159 is big enough we deduce that @xmath160 while @xmath161where we again used ( [ gam - b1 ] ) .
this tells us that for @xmath153 , the arguments of the factors @xmath162 and @xmath163 in @xmath127 have opposite signs and ( at least partially ) cancel each other .
then the estimate ( [ args1 ] ) can be further strengthen [ args2 ] ||=|| < /4 .
the estimates ( [ abss2 ] ) , ( [ args2 ] ) and the formula ( [ pack - estim2 ] ) show that the absolute value of any solution @xmath164 decreases exponentially with growing coordinates , and therefore all these solutions have finite energy .
we intend to show now that for some relation ( to be discussed below ) between the mass @xmath3 and the solution parameters @xmath62 , @xmath118 the solution ( [ kgfpackets ] ) is a wave packet with the gaussian envelop moving with group speed .
first we assume that the coordinates and time are small enough ( in what follows we clarify the formal meaning of the smallness ) for the square root in ( [ s_p ] ) to be expanded in taylor series up to the terms of the second order in time and coordinates [ s - mod - t1 ] s_p = m ( 1 + - + ) ( 1 - + + ) .
we use @xmath13 instead of @xmath8 in the last term because we are interested only in the terms quadratic in coordinates and time . however , we postpone expanding @xmath27 till we work out the applicability conditions . simplifying ( [ s - mod - t1 ] ) and collecting the terms we obtain [ s - mod - t ] s_p = @xmath166 , v_gr=. [ v_gr ] we choose for definiteness that @xmath167 .
it ensures that the solution propagates forward along the @xmath5axis , i.e. that @xmath168 .
we notice now that the expansion ( [ s - mod - t1 ] ) requires the conditions @xmath169 which are reduced under assumption that @xmath167 to the following ( actually , stronger ) ones [ cond - t - mod-3 ] |t| , , _ ^2 ( ) where @xmath170 is an appropriate matrix norm , e.g. euclidian . the possible expansion of the @xmath27 depends on the range of values of @xmath165 .
first we note that if _ 0 ^-1 [ g0_cond ] then @xmath171 and we can expand @xmath27 in the following way [ hel - lam ] ( ) ( * e*+_0)^-1 = _ 0@xmath172 . the same is true if ( [ g0_cond ] ) is not satisfied but @xmath165 is small enough for the condition @xmath171 to be fulfilled .
if both conditions are not meet , we can still expand @xmath173 by using [ hel - lam ] ( ) = = @xmath174 the latter expansion is valid if @xmath175 ^ -1 ( 2 t ) 1 .
[ cond - t - mod-1 ] thus , the dependence on @xmath165 can be important even for small times , @xmath176 , if the @xmath177 is large enough .
thus we conclude , that if the conditions and either one of the and are valid , then we may use the expansion of @xmath127 and the asymptotics ( [ as - bes ] ) of the modified bessel function to obtain the asymptotics of the packet ( [ kgfpackets ] ) [ pack - mod - t ] u_p^(-1/2 ) _ p a ( ) , where we used the dimensionless coordinates , time and mass [ non - dim ] = t/ , = z/ , = / , p = m .
all the characteristics of the asymptotic @xmath178 @xmath179 @xmath180 , @xmath181 and @xmath182 in ( [ pack - mod - t ] ) are expressed in terms of non - dimensional parameters @xmath82 , @xmath183 and @xmath184 as follows = p @xmath185/2 , = p @xmath186/2 , v_gr = / , [ def - o - k ] [ pack - mod - t - ds ] _ ^2 = , _
^2 = , = 0,&|t |_0 1 + 2 t , & otherwise here @xmath187 .
the amplitude factor @xmath188 and correction term in the phase @xmath189 read [ pack - mod - t - a ] a = c_p , g = ( * e * _ , ( ) * e * _ ) .
the solution @xmath87 describes a wave with frequency @xmath190 and wave number @xmath191 , which propagates along the @xmath5 axis and has the gaussian envelop moving with group velocity @xmath180 .
the localization near the @xmath5 axis is determined by the @xmath192 which depends on the orientation of @xmath149 ( the result of astigmatic nature of the considered solution ) and time .
note that for @xmath193 ( while still being much less then @xmath118 ) the width @xmath192 starts growing linearly with time .
we call this regime as moderate times one
. it can only show up when @xmath194 , otherwise only two asyptotic regimes can be identified for our solution : small times or large ones .
now we will check , that the formulae - describe correctly the field up to the distances where the packet becomes exponentially small . to this end
we first compare ( by order of magnitude ) the longitudinal width of the packet , @xmath195 , with the distance from the point @xmath196 , where ( [ pack - mod - t ] ) is still applicable as defined by the second condition ( [ cond - t - mod-3 ] ) .
this distance is @xmath197 with account of . secondly , the transverse width of the packet @xmath198 should be inside the zone determined by the third condition of ( [ cond - t - mod-3 ] ) .
thus , it must hold that [ cond1 - 2 ] _ , _ ( ( ) ) ^-1/2 .
thirdly , we demand that our solution must travel according to ( [ pack - mod - t ] ) on distances which are much larger then its longitudinal width @xmath181 , i.e. , [ cond-3 ] _ _ max~v_gr .
the two widths @xmath181 and @xmath182 contain @xmath82 in the denominator .
therefore all of the conditions ( [ cond1 - 2 ] ) and ( [ cond-3 ] ) are satisfied if @xmath183 is fixed and @xmath199 .
the last condition ( [ cond-3 ] ) is the more restrictive . in terms of the parameters @xmath200 and @xmath82 it reads [ mod - t - cond - t ] @xmath201 ^ 2,i.e .
finally , we may compare our results in the limit @xmath116 with the formulas for the packet like ( stigmatic ) solution for the wave equation obtained in ( 37 ) of @xcite .
in doing so we must put @xmath202 , where @xmath203 is a constant used in @xcite ( compare with the note after ( [ kgfpackets ] ) ) and also @xmath204 and thus obtain @xmath205 in complete agreement with @xcite .
let us find the asymptotics of ( [ kgfpackets ] ) for large times and large distances @xmath206 .
we assume that @xmath207 for @xmath208 , but @xmath209 is fixed .
the asymptotics of @xmath21 ( [ gammasol ] ) and @xmath8 ( [ theta ] ) are as follows [ help - t ] _ t - + o(t^-3 ) , _
t -t - + o(t^-1 ) , here we used that @xmath210 .
then the @xmath127 ( [ s_p ] ) can be expanded as [ as - s - help ] s_p= ( ) ^1/2(1 + i ) ^1/2 @xmath211 ^ 1/2 _ t ( ) ^1/2 @xmath212 = i
t + + + + o(t^-1 ) .
[ s_p_as ] in the first line here we took the square root with the positive real part .
thus , @xmath213 for @xmath208 and we can use the asymptotics ( [ as - bes ] ) of @xmath214 , and ( [ pack - estim2 ] ) for the whole solution .
now , we suppose that @xmath215 . introducing for the sake of brevity new variables [ no - nk ] = 1/ ,
=/ , we rewrite the solution ( [ pack - estim2 ] ) in the form that allows for its further simplification @xmath216 which we obtained by using that @xmath217 for large @xmath218 due to @xmath219 .
we used also that @xmath220 , and the fact that @xmath221 . to get the latter equality
we note that for @xmath222 we get @xmath223 , where @xmath224 is positive .
the branch of the square root @xmath225 is fixed by the asymptotics for @xmath226 : @xmath227 .
we do not specify the branch of the square root in @xmath228 and introduce the argument [ delta - g ] _ = - .
when @xmath229 , the modulus of the second exponent in ( [ t - as - pack ] ) has a sharp maximum and we will use quadratic approximation of @xmath55 in its vicinity .
we seek its position in the spherical coordinate system , i.e. , @xmath230 , @xmath231 @xmath232 .
we note that the first term in @xmath55 has nonnegative real part , thus its least value is equal to zero when @xmath233 or @xmath234 .
the two other terms of @xmath55 are mutually inverse ( up to a factor of @xmath235 ) , thus their sum reaches its least value equal to unity if + _
z=. [ no ] together with the condition @xmath236 this gives us [ extrem ] * v*_=0,v_z = v_gr,= /m , = k / m , where @xmath191 and @xmath190 are defined in .
it is easy to check that @xmath237 is incompatible with for @xmath238 , @xmath239 .
finally , we obtain [ lar - t - pack ] u_p^(-1/2 ) ( t , _ gr ) .
we have used here that @xmath240 , @xmath241 and that the derivative @xmath242 is zero . the notation for @xmath243
was introduced in .
the widths of the packet @xmath244 and @xmath245 are expressed through the second derivatives of @xmath246 as follows @xmath247^{-1 } = \frac{(1-v_{gr}^2)^2}{p } = \frac{p^3}{\omega^4 } , \label{lar - t - del1 } \\ & & \delta_{\vartheta}^2(t ) = \(p \im \phi''_{\vartheta\vartheta } \)^{-1 } = \frac{\tau ( 1-v_{gr}^2)}{p\,v_{gr}^2 ( { \bf e}_{\perp},-\im\g_0^{-1 } { \bf e}_{\perp } ) } = \frac{p}{{\cal k}^2 } \frac{\tau}{({\bf e}_{\perp},-\im\g_0^{-1 } { \bf e}_{\perp } ) } , \label{lar - t - del2}\end{aligned}\ ] ] where @xmath248 , @xmath249 , @xmath250 .
it is worth mentioning here , that due to the fact that @xmath251 is positively defined , so is @xmath252 , and thus @xmath253 is positive . according to ( [ lar - t - pack ] )
the field is concentrated in the intersection of a cone and a spherical annulus .
the width of the annulus increases with time and may be estimated as @xmath254 .
we will require that the speed of the packet center exceeds the speed of the packet enlarging .
the angle of the cone does not depend on time and we can assume that it is small .
finally we have conditions [ cond - loc ] _ v v_gr , _ 1 , which can be written in the simplest stigmatic case , @xmath255 , as follows p , p /.
these conditions are the more restrictive the closer are @xmath118 and @xmath62 to each other . if @xmath199 for fixed other parameters the localization is more pronounced . if @xmath256 the localization both in angle and along the propagation direction is better . in the case of general astigmatism
the term @xmath257 should stand in the last inequality instead of @xmath258 .
now we turn to the applicability conditions of the obtained formulas .
time @xmath165 will be considered large if expansions of @xmath21 and @xmath8 ( [ help - t ] ) could be limited to their first terms . for that we will require |t| /|1+v_z| , |t| ( 1-v^2 ) /|1+v_z| .
[ cond1 ] for the expansion of @xmath127 ( [ s_p_as ] ) being valid we additionally need 1 , = 1 . [ cond2 ] both conditions ( [ cond1 ] ) and ( [ cond2 ] ) contain @xmath259 , but we can substitute it with @xmath180 by recalling that @xmath260 according to ( [ cond - loc ] ) .
the group velocity itself can be expressed in terms of @xmath118 and @xmath62 by using ( [ v_gr ] ) .
two conditions ( [ cond2 ] ) are reduced to just one then , @xmath261 .
taking into account that @xmath262 we replace ( [ cond1 ] ) by a stronger inequality .
combining the two , we obtain [ lar - t - defin ] |t| 4 _ 0 ^ -1 / ,
let us introduce a fourier transformation relevant to the problem in hand f[f](,_,k_z ) = _ ^d+2 dt d^d _ d z f(t,_,z ) e^i(t- _ _ -k_z z ) , [ cal f ] here @xmath263 denotes the cartesian components of the wave vector , @xmath264 , perpendicular to @xmath265 , @xmath266 is the number of transversal dimensions .
the inverse transformation is @xmath267(\o,\bk_\perp , k_z)\ , e^{-i(\o t- \bk_\perp \br_\perp - k_zz)}.\ ] ] performing the fourier transformation of multidimensional we ( [ wave - eqn ] ) or kgfe ( [ kgfn ] ) one obtains the following equation in terms of generalized functions ( ^2-^2(_)-k_z^2 ) f[f](,_,k_z ) = 0 , [ eqfour ] here @xmath268 for we and @xmath269 for kgfe .
any solution of ( [ eqfour ] ) must be representable as f[f](,_,k_z ) = ( ^2-^2(_)-k_z^2 ) f ( _ , k_z ) [ hat f ] with a suitable well behaved function @xmath270 .
it is this function @xmath271 which we call the fourier image in what follows .
constructing particular solutions to the we or kgfe , we are free to choose any particular subspace of the surface @xmath272 in the phase space . for instance , in @xcite it was considered a solution of the from f[f](,_,k_z ) = _ ( k)@xmath273 @xmath274 [ ziolk ] where @xmath275 is an arbitrary ( real ) parameter ( to avoid conflict of notation we changed the original notation of @xcite ) , and
@xmath276 is an arbitrary weight function .
now we obtain the fourier image both for the beam like solutions @xmath32 , @xmath87 and particle like ones @xmath277 , @xmath86 .
apart from revealing the connection of our solutions with aforementioned ones , it will also be used in constructing asymptotic expansions in appendix [ asymptotic - kgf ] .
the fourier transform of the solution ( [ wavebeam ] ) of the we is given by @xmath278(\omega , k)=c_b \int dt\ , dz\ , d^d\br_\perp\ , e^{i(\o t- k_z z-\bk_\perp\br_\perp ) } \ , \sqrt{\det\g}\,e^{i\eta(\a+(\br_\perp,\g\br_\perp))}\ ] ] @xmath279 @xmath280 the definition of @xmath281 see in ( [ delta - g ] ) .
we recall that all the eigenvalues of the matrix @xmath282 have positive real part , and thus the last integral is convergent . substituting the last formula into ( [ phi_b_k_def ] ) and taking into account
that @xmath283 we have @xmath284(\bk_\perp , k_z)= c_b e^ { i\delta_{\g } } \(\pi/\eta\)^{d/2 } e^{-\tfrac{i}{4\eta } ( \bk_\perp,\g^{-1}_0\bk_\perp ) } \int dt\ , e^{i\o t - it(\eta+\tfrac{1}{4\eta } \bk_\perp^2 ) } \int dz\ , e^{-iz(k_z-\eta+\tfrac{1}{4\eta } \bk_\perp^2)}\ ] ] @xmath285 using the properties of the delta functions we can rewrite it finally as [ _ b](_,k_z ) = _ b @xmath286
@xmath287 , [ f_b_k_fn ] @xmath288 we remind that @xmath26 is a free parameter of our solution , along with @xmath33 .
the fourier image is defined now as ( assuming @xmath289 ) _
b(_,k_z ) = _ b @xmath286 .
[ phi_b_k_fn ] if compared with the considerations of @xcite ( see eq .
( [ ziolk ] ) ) we can see that in our case @xmath290 and @xmath291 this shows that our spectral function has more variables and less symmetries depending on @xmath263 and @xmath265 separately . for obtaining the fourier image of a particle like solution
@xmath292 we employ @xmath293 @xmath294 then , we can write @xmath295 now we shall construct the fourier transform of the kgf solution ( [ kgfbeam ] ) .
we accomplish it by acting similar to the section [ kgfeqn][beamskgf ] .
first we increase the number of dimensions by one , @xmath296 , and put the additional momenta component equal to mass , @xmath297 .
thus , we have to substitute everywhere in ( [ phi_b_k_fn ] ) @xmath266 by @xmath90 and @xmath263 by @xmath298 @xmath299 at the same time we must assume similar to ( [ zeta ] ) that @xmath300 where @xmath301 is a @xmath302 matrix , and @xmath303 is @xmath22 one as before .
thus , the fourier image of the beam like solution is u_b(_^m , k_z ) = _ b @xmath286 [ u_b_k_fn ] , @xmath304 here it is assumed that @xmath305 . for obtaining the fourier image of the particle like solution
@xmath306 we perform the integral transformation ( [ superposkgf ] ) of the @xmath307 @xmath308 then we arrive at @xmath309 where we used the notation of the previous section , @xmath106 , and put @xmath310 .
we notice a remarkable difference in localization properties of the fourier images for kgfe and we : in @xmath311 as compared with @xmath312 it is the absence the exponential suppression of small @xmath313 via terms of the type of @xmath314 .
in the present paper we have elaborated four families of explicit exact exponentially localized solutions to the wave equation ( [ wavebeam ] ) , ( [ wavepa ] ) , and to the klein gordon
fock one ( [ kgfbeam ] ) , ( [ kgfpackets ] ) . the families ( [ wavebeam ] ) , ( [ kgfbeam ] ) represent beam like solutions localized exponentially near a ray , while ( [ wavepa ] ) and ( [ kgfpackets ] ) are particle like ones localized exponentially near a point moving with group velocity along one of the axis .
all of the presented solutions are astigmatic multi - dimensional generalizations of those obtained before by the authors @xcite , as well as by other researches @xcite , etc .
unlike most of the others works , we performed all the analysis in space
time domain , which proved to be both convenient and efficient .
focusing on the particle like solutions of the kgf equation which are somewhat less studied in the literature , we investigated in detail the asymptotic properties of the central result of our work the particle like solutions to the kgf equation distinguishing several regimes : small times , moderate times and large times .
we also presented explicit fourier transformation of all constructed solutions and confirmed our asymptotic consideration obtained in space
time domain by investigating the fourier integral .
now we summarize briefly the contents of the section [ asyinv_sect ] .
the constructed solutions contain several parameters : @xmath315 @xmath62 , @xmath33 ( or @xmath15 in the stigmatic case ) .
if the non - dimensional mass [ par - p ] p = m is large the solution behaves as a packet with the gaussian envelop filled with oscillations which on the axis of the packet has the wave number @xmath191 and the frequency @xmath316 , and moves with the group speed @xmath317 .
like behaviour takes place for all times .
below we discuss properties of the field in different regimes . for consequent times
@xmath318 as a function of @xmath319 and @xmath5 for @xmath320 ( upper pictures ) and as function of @xmath321 and @xmath319 for @xmath322 ( lower pictures ) .
convenient normalization for the @xmath323 is chosen , the coordinartes and time are in the units of mass @xmath3 .
see text for the values of all parameters.,width=480 ] for small times , @xmath324 ( assuming that @xmath325 ) the solution behaves according to with @xmath326 .
this regime is characterized by complete absence of any distortion during the propagation .
the longitudinal width as well as transversal one , is time independent . in the transversal direction
the astigmatic properties are practically frozen and do not depend neither on time , nor on propagated distance .
the localization ellipse is defined by @xmath33 itself ( compare with large times ) .
the maximum of the propagated distance is of the order of @xmath327 .
this regime is exemplified at the fig.[small_t_pic ] for the following values of parameters ( in uits of mass @xmath3 ) @xmath328 , @xmath329 , @xmath330 , @xmath331 , @xmath332 , @xmath333 .
moderate times are characterized by condition @xmath334 .
the solution in this case can be described rather good by the asymptotic formula with @xmath335 .
the distortion of the solution in this regime is twofold .
first of all , the absolute value of the solution is decreasing linearly with time due to the dependence on @xmath165 of the prefactor @xmath188 via @xmath336 . secondly , the transverse width of the solution @xmath337 also grows linearly with time .
both these features are clearly visible on the fig.[moder_t_pic ] , where the absolute value of @xmath86 is plotted at the same values of the parameters as before .
we also note that at this stage the astigmatic properties can already be seen the localization ellipse is slowly rotating . for consequent times
@xmath338 as a function of @xmath319 and @xmath5 for @xmath320 ( upper pictures ) and as function of @xmath321 and @xmath319 for @xmath322 ( lower pictures ) .
convenient normalization for the @xmath323 is chosen , the coordinartes and time are in the units of mass @xmath3 .
see text for the values of all parameters.,width=480 ] for the large time regime @xmath165 satisfies conditions ( [ lar - t - defin ] ) .
the amplitude of the solution also decreases with time in this case , as it follows from .
the packet is concentrated in the intersection of a cone and a spherical annulus as it is seen from .
the width of the annulus increases linearly with time as @xmath339 , while the angular width of the cone is given by @xmath245 , see , .
their connection with the widths at small times is [ lar - t - par - widths ] _
v^2 = p _ ^4 , _
we see that the larger is the transverse width for small time the narrower is a cone .
this property reflects the uncertainty principle .
large - time behaviour is presented on the fig.[large_t_pic ] , where the absolute value of @xmath86 is plotted over quite large period of time at the same values of the parameters as before .
the astigmatic properties are already frozen at this regime , the axis of the localisation ellipse are rotated to the full angle @xmath54 as compared with its position at @xmath340 .
the latter fact can be understood if one notes that the localization ellipse for large times is defined by @xmath37 , see , contrary to the case of small times where it is defined by @xmath33 . under inversion ,
the smaller eigenvalue ( the ellipse axis ) becomes the bigger one , so , the localization ellipse effectively rotates by an angle of @xmath341 as compared to small times . from a practical point of view
, it can be more convenient to characterize the solutions not by choosing the initial parameters @xmath118 , @xmath62 and @xmath15 , but by specifying their asymptotic properties the wave number , group velocity and localization widths either at small times , or at large ones .
the former are unambiguously expressed through the latter as we see from .
a legitimate practical problem is to find for a given kgfe ( i.e. for a given value of @xmath3 ) a solution with particular values of , e.g. , @xmath190 and the width of the packet for small times @xmath181 . from
we can deduce then the value of @xmath82 , which must satisfy @xmath342 , if we wish the asymptotics be applicable next , we calculate the wave number , @xmath343 and the group speed , @xmath344 . knowing @xmath82 and @xmath180 we can derive both the product @xmath345 by and the ratio @xmath346 , which together gives us all the parameters of the desired solution but its astigmatic properties .
we can deduce the latter by choosing , for instance , the transversal width at small times , @xmath337 .
now all the parameters for the solution are known .
parameter @xmath120 can not be derived by considering the asymptotic properties in the highest order .
we have however some restrictions to be satisfied , if we wish our solution possesses good localization properties [ p - k - cond ] p ^2 , p^3 ( ) ^2 , p ^2 .
the first condition makes a longitudinal width of the packet for small and moderate times smaller than the distance where asymptotics works , see . the second and the third ones concern the large - time asymptotics , they originate from .
the second condition ensures that the speed of increasing of the longitudinal width of the packet is smaller then the group speed .
the third condition means that the angle of the cone is small .
all of these conditions can be satisfied , for example , if we assume that @xmath347 , @xmath348 .
then @xmath349 is of order of @xmath350 . if we take parameters in such a way that @xmath351 as well , the restriction conditions will be satisfied . , for consequent times @xmath352 as a function of @xmath319 and @xmath5 for @xmath320 ( upper pictures ) and as function of @xmath321 and @xmath319 for @xmath322 ( lower pictures ) . convenient normalization for the @xmath323 is chosen , the coordinartes and time are in the units of mass @xmath3
. see text for the values of all parameters.,width=480 ] we expect that our results may be useful for prediction of waves propagation in media with dispersion .
the obtained solution have non - zero angular momentum @xcite which is studied intensively for the wave equation in context of manipulating of nanoparticles .
the investigation of this momentum for the waves in dispersive media is a very appealing for the future research .
another possible application concerns two - dimensional solutions of kgfe
. such solution may be a base for further design of localized solutions of the dirac equations which may find application in prediction and modelling of waves in epitaxial graphen .
this work was supported in part by fapesp ( i.v.f . ) and rfbr grant 140200624 ( m.v.p . ) .
any solution of the kgf equation can be written as a fourier integral @xmath353 where @xmath354 , and @xmath355 .
it is convenient to introduce new dimensionless variables @xmath356 @xmath357 .
we assume that the modulus of the fourier transform @xmath358 has a sharp maximum and @xmath358 can be written in the form [ four - int ] u ( ) = a ( ) , p 1,i.e .
we assume that @xmath55 has stationary point in the minimum of its real part . the formula ( [ as - t1 ] )
may be rewritten in the form suitable for the analysis by the method of the steepest descent now [ int - steep ] u(,t ) = _
^n d^n a ( ) , where [ phase - int - steep ] ( ) = ( ) + i ( - ) , = /t ,
= m t /p , = .
the main term of asymptotics of the integral ( [ int - steep ] ) for @xmath199 reads * ? ? ?
* @xmath359 where @xmath360 is the determinant of the jakobi matrix @xmath361 calculated in the saddle point @xmath362 , i.e. , the matrix of the second derivatives of @xmath363 with respect to @xmath364 .
the saddle point @xmath362 should be found from the equation [ saddle - p - mod - t ] ( _ * ) = ( _ * ) + i ( ( _ * ) - ) = 0 . for small times ( in comparison with @xmath365 ) we seek the saddle point as an expansion @xmath366 and get ( _ * ^(0 ) ) = 0 , _ * ^(1 ) = - ( _ 0 ) ^-1(_gr - ) , _ gr ( _ * ^(0 ) ) , where @xmath367 is the matrix of the second derivatives of @xmath55 with respect to @xmath364 calculated in the point @xmath368 .
corrections are to be taken into account in the formula ( [ as - t22 ] ) only in the exponential term containing the large parameter @xmath82 [ exp - psi ] ( _ * ) ( ^(0 ) _ * ) + i ( ^(0 ) _ * - ^(0 ) _ * ) - @xmath369 - ^2 @xmath370 . substituting ( [ exp - psi ] ) in the ( [ as - t22 ] ) , neglecting the correction terms in the amplitude and recalling that @xmath371 we obtain @xmath372 this formula can be applied to the exact solution @xmath86 presented in section [ loc - part - kgf ] . its fourier image ( [ u_p(k ) ] ) can be given in the form ( [ four - int ] ) as follows @xmath373 it is easy to check that formula ( [ as - t22-fin ] ) with account of ( [ amp - int ] ) and ( [ phase - int ] ) gives ( [ pack - mod - t ] ) .
now we turn to the large time behaviour and assume that @xmath374 in ( [ as - t1 ] ) changes slowly as compared with the oscillatory term .
thus we are able to proceed with stationary phase method and obtain @xcite @xmath375 here @xmath376 is the solution of the equation @xmath377 where @xmath378 .
it is easy to check that @xmath379 , @xmath380 where @xmath381 and @xmath382 are given by ( [ no - nk ] ) .
by @xmath383 in ( [ as-1-t ] ) we denote the @xmath384 matrix of second derivatives of @xmath385 with respect to components of @xmath386 calculated in the point @xmath387 .
it is easy to check that _ * = - = = .
[ parav - l - t1 ] formula ( [ as-1-t ] ) demonstrates complicated dependence of @xmath388 and @xmath165 through @xmath389 .
it reads @xmath390 after substitution of @xmath391 from ( [ u_p(k ) ] ) with account of ( [ no - nk ] ) we obtain the formula which is in agreement with previously found formula ( [ lar - t - pack ] ) .
it is important to note that the obtained formula can not be applied when @xmath116 .
it is due to the fact that the second derivative of the phase function from ( [ as-1-t ] ) in this case tends to zero , thus the region of validity of the asymptotic ( [ as - l - t3 ] ) is approaching spacial infinity .
the stationary phase method which we used is not applicable in this case .
lu - greenleaf , _ nondiffracting x waves - exact solutions to free - space scalar wave equation and their finite aperture realizations _
, in ieee transactions on ultrasonics , ferroelectrics , and frequency control , vol .
i ( 1992 ) , p. 19 .
perel , i.v .
fialkovsky , _ exact exponentially localized solutions to the klein - gordon equation _
, journal of mathematical sciences , vol .
3994 - 4000 ( 7 ) , kluwer academic publishers ( engl . transl . from zapiski nauch .
pomi , 245 , p.187 - 198 , 2001 ) , arxiv:0711.4143 [ hep - th ] i. m. besieres , a. m. shaarawi , and l. p. ligthart , `` a note on dimension reduction and finite energy localized wave solutions to the klein - gordon and scalar wave equations .
part i : fmw - type , '' j. electromagn . waves appl . , 14(5 ) , 593 - 610 ( 2000 ) .
a. m. shaarawi , i. m. besieres , and r. w. ziolkowski , `` a novel approach to synthesis of nondispersive wave packet solutions to the klein - gordon and dirak equations , '' j. math
. phys . * 31*(10 ) , 2511 - 2519 ( 1990 ) . | highly localized explicit solutions to multidimensional wave and klein gordon fock equations are presented .
their fourier transform is also found explicitly .
solutions depend on a set of parameters , and demonstrate astigmatic properties .
asymptotic analysis for large and moderate time shows that constructed solutions have gaussian localisation near a point moving with the group speed . |
in the most recent _ fermi _ large area telescope ( lat ) catalog of 1873 @xmath3gev sources , over 1170 have statistically reliable counterpart identifications at lower energy @xcite . the bulk of these identifications are blazars and spin - powered pulsars , with young pulsars ( both radio - selected and geminga - like @xmath2-ray selected ) dominating at low galactic latitude .
an additional population of millisecond pulsars extends to high latitude among the blazar - associated sources .
identification progress is especially impressive for the bright , well localized sources that have been studied since early in the _ fermi _ mission .
here we update @xcite , selecting ` bright ' sources from the 2fgl catalog as those with @xmath4 detection significance and time - averaged energy flux @xmath5 there are 249 such sources .
all but six presently have lower energy identifications : blazars , pulsars and a few binaries .
the handful of sources _ not _ yet associated with one of these source classes provide the best prospect for new types of @xmath2-ray emitter .
we have initiated a campaign to characterize these unidentified sources .
the first result from this effort is the discovery of dramatic optical and x - ray variability for the high latitude ( @xmath6 ) source j2339@xmath00533 , implying that it is a short period ` black - widow'-type binary millisecond pulsar ( msp ) @xcite .
here we report progress on the next high latitude unidentified source in this set , 2fgl j1311.7@xmath03429 ( hereafter j1311 ) at @xmath7 . with a @xmath8 detection significance ( the highest among the 2fgl unidentified sources ) and an energy flux of @xmath9 ( the second brightest unided 2fgl )
, this is a top candidate for follow - up .
the ` variability index ' value 19.5 and ` curvature significance'=6.3 mark this as a steady source with a substantial spectral cut - off : a prime pulsar candidate .
it has been searched for @xmath10hz @xmath2-ray pulsations ( eg . * ? ? ? * ) and for radio pulses down to @xmath11ms periods @xcite , with no detection .
thus it is unlikely to be an isolated young pulsar or a radio - loud msp .
we describe here an optical campaign to identify and characterize a counterpart .
initial exposures of the j1311 error region were taken with the minimo camera at the 3.6 m wiyn telescope on feb .
17 - 18 , 2012 .
exposures were 180s using the gunn filter set ( @xmath12 on 2/17 ; @xmath13 on 2/18 ) .
observations were perforce at airmass @xmath14 , seeing was variable and conditions were non - photometric .
a more extensive image sequence was taken with the soar optical imager ( soi ) at the 4.2 m soar telescope on march 21 - 23 , 2012 . here
120s exposures using sdss filters were taken , covering one color each night ( @xmath15 on 3/21 , @xmath16 on 3/22 and @xmath17 on 3/23 ) .
the camera was binned @xmath18 , allowing fast 11s read - out and high observation efficiency .
again image quality was highly variable ( partly due to loss of wavefront sensor control of the primary ) .
we searched image archives for exposures covering j1311 , finding a sequence ( program 086.d-0388 ) of 170s vlt / vimos exposures taken on march 4 ( @xmath19 $ ] ) and march 11 , 2011 ( @xmath20 $ ] ) .
vimos is an array camera with 2@xmath21 gaps between the four ccd chips , so the five pointings were offset to cover the full j1311 error region . in examining the images for variable objects
, we paid particular attention to sources coincident with detections in archival x - ray data ( see below ) .
one such star was blue and highly variable on short timescales .
the nomad position of this likely counterpart is 13 11 45.741 , -34 30 29.96 ( 2000.0 ) .
while we ensured that this star was covered in all minimo and soi exposures , we found that the counterpart was in the array gap for the first vimos pointing on march 3 , 2011 and at the extreme edge of a chip , suffering major vignetting , in the second pointing on march 11 , 2011 . to quantify the variations
, we performed simple aperture photometry of the counterpart and surrounding field stars in all frames .
since the bulk of the imaging was in the sdss filter set , we attempted to reference all photometry to this scale .
the minimo data were calibrated using @xmath22 images in sdss fields at low airmass ; we transferred this calibration to stars in the j1311 field .
unfortunately , the non - photometric wiyn conditions and pass - band differences between the sdss and gunn sets make the substantial atmospheric extinction corrections for j1311 uncertain .
this is not a severe problem for the field stars , which were mostly late type with @xmath23 .
however the target s color is as blue as @xmath24 and varies appreciably with magnitude .
thus while the relative photometry for each color sequence has very small statistical errors , there may be a systematic offset between the wiyn and soar points as large as 0.2mag , especially in @xmath25 .
for the vimos images , we used nomad field stars to establish zero points in b and v. nomad - detected stars were generally saturated in the vimos i images , so we established the i scale using stars from the denis catalog .
we show the photometry for each night in figure 1 , converting to fluxes and plotting against estimated binary phase .
the soi @xmath26 data show dramatic quasi - sinusoidal variability by a factor @xmath27 with a period of @xmath28h . in @xmath29 an initial flux minimum
was followed by a large @xmath30 flare , with rapid fluctuations during return to quiescence .
the @xmath26 fluxes also show variability , especially near maximum .
the wiyn data are sparse and have larger statistical errors , but do follow the periodic modulation .
most wiyn fluxes appear 10 - 20% lower than soar at the corresponding phase , suggesting an underestimate of @xmath31mag for the extinction correction .
however , given the apparent non - periodic variability this can not be firmly concluded .
the vlt fluxes from march 11 pass through two minima ; those from march 4 show one minimum , but appear to have @xmath32jy additional flux .
9.3truecm -0.7truecm the intermittent flaring activity of j1311 compromises simple periodogram or fourier methods for refining the period .
however , the combination of soar , wiyn and vlt exposures give photometry information on 1 - 3 , 7 , @xmath1135 , and @xmath11350 day time scales , which suffices to resolve possible aliases .
after referencing all observation mid - point times to the solar system barycenter , we can constrain the phase of the 2012 epoch minima to @xmath33 , giving a best fit period of @xmath34s .
the epoch of ( quiescent ) maximum light associated with our soar @xmath35 measurements ( presumably pulsar superior conjunction , see below ) is barycentric mjd@xmath36 .
the approximately sinusoidal nature of the light curve and very large modulation amplitude can be best interpreted in terms of strong companion heating .
the inferred heating luminosity is appreciably above any x - ray flux ( see below ) ; we infer that accretion power is not dominant and that the heating source is invisible in the optical and faint in the x - ray .
we conclude that this is a ` black - widow ' type pulsar binary and is the likely counterpart of the @xmath2-ray source . in this
it is remarkably similar to the black widow candidate counterpart for j2339@xmath00533 .
but the @xmath37 shorter orbital period is unprecedented ; this is an extreme black widow system .
7.2truecm -0.1truecm 2fgl j1311.7@xmath03429 has been observed by both the _ cxo _ and _ suzaku _ satellites .
one exposure from each facility is available in the nasa archives , additional exposure has been taken but is not publicly available .
the counterpart was detected in both data sets and well localized by _
cxo_. we analyzed both observations for x - ray spectrum and variability .
the _ suzaku _ observation ( obsid 804018010 ; kataoka , pi ) started on august 04 2009 ( mjd 55047.2290 ) and had 33.4ks of on - source time over 95ks ( @xmath38 binary orbits ) with exposure in xis0,1,3 .
a comparably bright source lies @xmath39 from the target , so we extracted source counts from a @xmath40 radius aperture .
figure 2 ( left ) shows the combined xis count rate , both direct aperture counts ( histogram ) and the background - subtracted aperture - corrected count rate ( points ) .
the vertical dotted lines mark the phases of maximum optical light ; we see that orbital modulation is difficult to measure since @xmath41 is so close to that of satellites in low earth orbit .
a clear x - ray flare is seen 5000s after the observation start .
an acis - i exposure of 19.8ks live - time ( obsid 11790 ; cheung , pi ) was taken starting mjd 5527.6652 . unfortunately , as for vlt observation 1 , the counterpart was directly in the gap between acis - i chips .
however , the _ cxo _ dither did provide some ( @xmath1125% effective live - time ) exposure and a source coincident with the optical counterpart was well detected with 60 counts .
there is no strong evidence for secular x - ray variability on 3000s timescales .
the exposure only covers @xmath42 orbits .
figure 2 ( right ) shows the folded x - ray light curves from the _ suzaku _ and _ cxo _ data sets .
after the early flare in the _ suzaku _ data is excluded there is only weak evidence for orbital modulation ; note there is an exposure gap for phases @xmath43 .
similarly the _ cxo _ data show no strong peak or eclipse . in both light curves
there seems to be a slight excess at phases 0.7 - 1.1 .
this may indicate additional weaker flaring activity near superior conjunction .
only crude spectral constraints can be obtained with the limited counts , especially after selecting for _ suzaku _ photons outside of the flare period . if we fit an absorbed power law to the x - ray data we find that @xmath44 is essentially unconstrained ( @xmath45 ) but consistent with the limited extinction in this high @xmath46 direction .
we therefore fix at the @xmath47 inferred from the dust maps ( heasarc nh tool ) , and fit for the power law index . using ciao the _ cxo _ data fit to @xmath48 .
xspec fits to the _ suzaku _ xis0,3 give @xmath49 .
fixing at @xmath50 , we obtain a quiescent unabsorbed flux for the source of @xmath51 . a deeper exposure , allowing a better light curve and a orbit - dependent spectral study is clearly of interest .
9.15truecm -0.7truecm
figure 3 shows two periods of the soi light curves folded on the ephemeris of 2.1 .
the large quasi - sinusoidal modulation shows that the light curve is dominated by heating of the secondary .
the typical @xmath52 near @xmath53 suggests @xmath54k . by phase @xmath55 the color has reddened to @xmath56 ( @xmath57k ) . at flux minimum ( @xmath58 )
large ( up to @xmath59 ) fluctuations dominate . since relatively small fluctuations can overwhelm the quiescent flux , these prevent any meaningful measurement of color at flux minimum .
we have attempted to model the ` quiescent ' light curves with the elc code @xcite .
a formal fit is not indicated as there are many positive ` flare ' excursions from any reasonable light curve .
however comparison with a range of models shows that a few of the parameters are significantly constrained .
for example , matching the characteristic @xmath60 at maximum requires heating by a primary luminosity ( modeled here as an isotropic ` x - ray ' flux ) of @xmath61 = 36.0 \pm 0.3 $ ] . also , to obtain adequate modulation while avoiding an x - ray eclipse , requires @xmath62 .
at this point other parameters are poorly constrained , although large mass ratios and a near - roche - lobe filling secondary are preferred .
however , the basic elc model seems inadequate for a good fit ; a particular challenge is to reproduce the relatively flat maxima and broad ` shoulders ' in the @xmath25 and @xmath35 curves at @xmath63 , 0.75 .
this may be due to limitations of the atmosphere tables .
one can substantially improve the fit by modifying the limb - darkening law , but for blackbody surface spectra broad light curves ( with adequate heating ) require unphysical negative limb darkening coefficients .
another , more attractive , possibility is that the distribution of surface heating differs from the isotropic form assumed here ; the pulsar may primarily heat a portion of the companion star .
an illustrative set of curves is shown in figure 3 , assuming a mass ratio @xmath64 , orbital semi - major axis @xmath65 , inclination @xmath66 , heating flux @xmath67 and a @xmath68k companion reaching 0.99 of its roche lobe radius .
the observed @xmath29 flux evidently has a substantial added flux even before the bright flare , at the epoch observed .
at this short orbital period , the mean density in the secondary roche lobe is @xmath69 , so the secondary could be a main sequence star of @xmath70 . to reach this period ,
however , the system very likely passed through a common envelope phase , and would have a helium - rich core .
the strong heating and near - roche lobe filling suggest a sub - stellar secondary , but additional observations are needed to determine the state of the secondary core and envelope . a very rough estimate of the source distance can be made from the @xmath71 color and @xmath72 flux at maximum , if we assume that the star nearly fills its roche lobe at @xmath73 .
such color indicates @xmath114,000k ( type @xmath11b6 ) , corresponding to main sequence @xmath74 and @xmath75 . correcting for the effective area and the small ( @xmath76 ) galactic extinction in this direction , we get @xmath77kpc .
this is likely an overestimate as the secondary should be h - poor , and the temperature of the viewed hemisphere is far from uniform , even at maximum .
if we adopt the view that the x- and @xmath2-ray emissions come from an energetic pulsar , we can make some additional estimates .
@xcite finds that pulsars have @xmath78 , so from the observed x - ray flux we would infer a spin - down luminosity @xmath79 .
gamma - ray pulsars are also observed to follow a heuristic luminosity law @xmath80 @xcite .
this is related to the observed flux as @xmath81 where the beaming - dependent correction factor is @xmath82 @xcite .
thus the observed @xmath83 gives a distance - dependent estimate for the spindown power of @xmath84 .
this is consistent with the x - ray luminosity estimate when @xmath85kpc .
however , if we match @xmath86 to the elc - estimated heating flux a distance of @xmath87kpc is preferred .
the nature of the intermittent ` flaring ' emission is at present unclear .
since the bands were observed sequentially , we do not even know the characteristic color .
all we can infer are large increases above the quiescent flux ( up to @xmath30 in the optical ; @xmath88 in the x - ray , reaching a peak flux @xmath89 ) .
these are concentrated toward the heated side of the companion , but visible at all phases .
shocks in the wind stripping the companion could be the origin .
however , since we do not know of such flaring activity on other well - studied black widow binaries , it is tempting to associate the activity with the unusually short @xmath41 of j1311 . as has been noted ( eg .
* ) , low mass stars in close binaries often show enhanced coronospheric activity since tidal locking enforces rapid rotation , driving dynamo - generated magnetic fields .
the other main ingredient for high magnetic activity is convection .
this may suggest a low stellar - type mass for the companion .
if sub - stellar , perhaps deep differential heating may drive the convection . a spectroscopic search for coronal line emission , or spectra during an outburst , could resolve the origin .
additional data can greatly improve estimates of the system properties . simultaneous multicolor photometry covering many periods and flaring events
can help us understand the temperatures of the various emitting regions . a complete spectroscopic study of the secondary can place important constraints on the compact object mass and probe the state of the secondary photosphere .
spectra can also improve knowledge of the compact object s period , orbital epoch and semi - major axis , each of which are crucial for decreasing the parameter space to be searched for a @xmath2-ray pulses .
of course , a @xmath2-ray or radio pulsar ephemeris would produce a qualitative change in our knowledge of the system properties .
we are pursuing all of these follow - up investigations .
we close with a comment on @xmath2-ray pulsar population in the bright sample .
figure 4 shows these sources in the curvature - variability plane .
the pulsar - blazar separation is excellent ; the only pulsar in the blazar region is the crab , whose pwn contributes the variable power - law component .
the pulsar region contains a few other source types : high mass binaries , a globular cluster , snr , but each may harbor spin - powered pulsars .
the unidentified sources , too , have significant curvature and low variability .
thus it is unsurprising that two of the six unidentified sources are good candidates for binary msp , although they lack radio detection .
the other unidentified sources may be similar or new source types .
an interesting question is whether j2339@xmath00533 and j1311@xmath03429 represent the first of a new source class radio quiet msp .
it is not yet clear that we need to appeal to such novelty : for both the optical gives evidence of a strong evaporative wind .
it is entirely possible that radio emission is directed at earth , but the pulse is scattered and dispersed to undetectability by this wind .
thus it remains possible that the msp @xmath2-ray beam lies within the radio emission zone i.e. that all @xmath2-ray msp are radio detectable . even if these two sources represent a harbinger of a new source class we emphasize that this source class must be small , at least for the bright population , in contrast to some suggestions in the literature @xcite .
the statistics in our sample clearly make this case of the 41 young pulsars in this set only 17 ( @xmath90 ) are radio selected .
in contrast there are 12 msp in the bright sample , all radio pulsars . thus even _
if _ all six unidentified sources prove to be radio - quiet msp the radio loud subset will be _ at least _
2/3 of the @xmath2-ray msp , twice the fraction of the young pulsars .
thus , while j1311@xmath03429 may have joined the ` black widow ' club of pulsars evaporating their companions , it is not yet clear if there is a ` radio quiet msp ' club for it to join .
and if such a society does exist , it is very select , indeed .
the soar telescope is a joint project of : conselho nacional de pesquisas cientificas e tecnologicas cnpq - brazil , the university of north carolina at chapel hill , michigan state university , and the national optical astronomy observatory .
abdo , a.a .
2009a , apjs , 183 , 46 abdo , a.a .
et al . 2009b , science , 325 , 840 abdo , a.a .
2010 , apjs , 187 , 460 .
becker , w. 2009 , assl , 357 , 91 harding , a.k . ,
usiv , v.v . , & muslimov , a.g .
2005 , apj , 622 , 531 kong , a. et al .
2012 , apj , 747 , l3 .
nolan , p.l .
2012 , apjs , 199 , 31 orosz , j.a . &
hauschildt , p.h . 2000 ,
aa , 364 , 265 ray , p.s .
2012 , proc .
2011 fermi symposium richards , m.t . &
albright , g.e .
1993 , apjs , 88 , 199 romani , r.w . & shaw , m.s .
2011 , apj , 743 , l26 takata , j , wang , & y. cheng , k.s . 2011 ,
mnras , 415 , 1827 watters , k.p .
2009 , apj , 695 , 1289 | we have found an optical / x - ray counterpart candidate for the bright , but presently unidentified , _ fermi _ source 2fgl j1311.7@xmath03429 .
this counterpart undergoes large amplitude quasi - sinusoidal optical modulation with a 1.56h ( 5626s ) period .
the modulated flux is blue at peak , with @xmath114,000k , and redder at minimum .
superimposed on this variation are dramatic optical flares .
archival x - ray data suggest modest binary modulation , but no eclipse . with the @xmath2-ray properties
, this appears to be another black - widow - type millisecond pulsar .
if confirmation pulses can be found in the gev data , this binary will have the shortest orbital period of any known spin - powered pulsar .
the flares may be magnetic events on the rapidly rotating companion or shocks in the companion - stripping wind . while this may be a radio - quiet millisecond pulsar , we show that such objects are a small subset of the @xmath2-ray pulsar population . |
for any graph @xmath1 , let @xmath17 , @xmath18 and @xmath19 be respectively the set of the edges , the set of the vertices and the set of the leaves of @xmath1 . a * weighted graph *
@xmath0 is a graph @xmath1 endowed with a function @xmath20 . for any edge @xmath21 , the real number @xmath22
is called the weight of the edge . if all the weights are nonnegative ( resp .
positive ) , we say that the graph is * nonnegative - weighted * ( resp . *
positive - weighted * ) , if all the weights are nonnegative and the ones of the internal edges are positive , we say that the graph is * internal - positive - weighted*. throughout the paper we will consider only simple finite graphs . for any subgraph @xmath3 of @xmath1
, we define @xmath4 to be the sum of the weights of the edges of @xmath3 .
let @xmath23 be a nonnegative - weighted graph .
for any distinct @xmath24 , we define @xmath25 more simply , we denote @xmath26 by @xmath27 for any order of @xmath7 .
we call the @xmath28 the * @xmath8-weights * of @xmath9 and we call a @xmath8-weight of @xmath9 for some @xmath8 a * multiweight * of @xmath9 . if @xmath29 is a subset of @xmath18 and we order in some way the @xmath8-subsets of @xmath30 ( for instance , we order @xmath31 in some way and then we order the @xmath8-subsets of @xmath30 in the lexicographic order with respect the order of @xmath31 ) , the @xmath8-weights with this order give a vector in @xmath32
. this vector is called @xmath8 * -dissimilarity vector * of @xmath33 .
equivalently , if we do nt fix any order , we can speak of the * family of the @xmath8-weights * of @xmath33 .
we can wonder when a family of real numbers is the family of the @xmath8-weights of some weighted graph and of some subset of the set of its vertices . if @xmath31 is a finite set , @xmath34 and
@xmath35 , we say that a family of positive real numbers @xmath36 is * p - graphlike * ( resp .
nn - graphlike , ip - graphlike ) ) if there exist a positive - weighted ( resp .
nonnegative - weighted , internal - positive - weighted ) graph @xmath0 and a subset @xmath31 of the set of its vertices such that @xmath37 for any @xmath38 @xmath8-subset of @xmath30 .
if the graph is a positive - weighted tree , we say that the family is * p - treelike * ; if the graph is a positive - weighted tree and @xmath39 , we say that the family is * p - l - treelike * ( and analogously for nonnegative - weighted or positive - internal - weighted trees ) .
the first contribute to the characterization of the graphlike families of numbers is due to hakimi and yau ; in 1965 , they observed that a family of positive real numbers @xmath40 is p - graphlike if and only if the @xmath41 satisfy the triangle inequalities . in the same years
, also a criterion for a metric on a finite set to be nn - treelike was established , see @xcite , @xcite , @xcite : let @xmath42 be a set of positive real numbers satisfying the triangle inequalities .
it is p - treelike ( or nn - l - treelike ) if and only if , for all @xmath43 , the maximum of @xmath44 is attained at least twice . in terms of tropical geometry ,
the theorem above can be formulated by saying that the set of the @xmath45-dissimilarity vectors of weighted trees with @xmath12 leaves and such that the internal edges have negative weights is the tropical grassmanian @xmath46 ( see @xcite ) . for @xmath47 also the case of not necessarily nonnegative weights has been studied . for any weighted graph @xmath0 and for any @xmath48
, we define @xmath49 to be the minimum of @xmath50 for @xmath51 a simple path joining @xmath52 and @xmath53 .
again , we call such numbers `` @xmath45-weights '' . in 1972
hakimi and patrinos proved that a family of real numbers @xmath40 is always the family of the @xmath45-weights of some weighted graph and some subset @xmath54 of its vertices . in @xcite , bandelt and steel
proved a result , analogous to buneman s one , for general weighted trees , precisely they proved that , for any set of real numbers @xmath55 , there exists a weighted tree @xmath56 with leaves @xmath57 such that @xmath58 for any @xmath59 @xmath45-subset of @xmath60 if and only if , for any @xmath61 , we have that at least two among @xmath62 are equal . for higher @xmath8
the literature is more recent . in 2004 ,
pachter and speyer proved the following theorem ( see @xcite ) . *
( pachter - speyer)*. let @xmath63 with @xmath64 and @xmath65 .
a positive - weighted tree @xmath56 with leaves @xmath57 and no vertices of degree 2 is determined by the values @xmath66 , where @xmath38 varies in @xmath67 . in @xcite
, the authors gave the following characterization of the ip - l - treelike families of positive real numbers : * ( herrmann , huber , moulton , spillner)*. if @xmath68 , a family of positive real numbers @xmath69 is ip - l - treelike if and only if the restriction to every @xmath70-subset of @xmath60 is ip - l - treelike . besides they studied when a family of positive real numbers is ip - l - treelike in the case @xmath71
. finally , in @xcite and @xcite , for any weighted tree @xmath56 , for @xmath72 and for any distinct leaves @xmath73 , the author defines @xmath74 to be the sum of the lengths of the edges of the minimal subtree joining @xmath75, .... ,@xmath76 and gives an inductive characterization of the families of real numbers indexed by the subsets of @xmath60 of cardinality greater or equal than @xmath45 , that are the families of the multiweights of a tree with @xmath12 leaves and the set of its leaves .
let @xmath77 with @xmath78 .
in this paper we study the problem of the characterization of the families of positive real numbers , indexed by the @xmath8-subsets of an @xmath12-set , that are p - graphlike .
as we have already said , the case @xmath47 has already been studied by hakimi and yau . here
we examine the case @xmath16 , both for trees ( section 3 ) and graphs ( section 4 ) , see theorems [ thm : trees ] , [ thm : trees2 ] , [ thm : graphs ] and [ thm : morevertices ] .
@xmath79 let @xmath80 .
@xmath81 for any @xmath82 with @xmath83 , let @xmath84= \{1 , ... , n\}$ ] .
@xmath81 for any set @xmath31 and @xmath85 , let @xmath86 be the set of the @xmath8-subsets of @xmath31 .
@xmath81 throughout the paper , we will consider only simple finite graphs .
for any @xmath87 , let @xmath88 denote the edge joining @xmath89 and @xmath90 .
@xmath79 for simplicity , the vertices of graphs will be often named with natural numbers . in the figures
the names of the vertices will be written in bold calligrahic in order to avoid confusion with the weights .
[ intest ] let @xmath23 be a positive - weighted graph .
we say that a connected subgraph of @xmath1 , @xmath91 , realizes @xmath28 , or it is a * subgraph realizing * @xmath28 if @xmath92 are vertices of @xmath91 and @xmath93 .
observe that it is necessarily a tree , so we will call it also * subtree realizing @xmath28*. let @xmath31 be a finite set . if @xmath41 for @xmath94 are positive real numbers and there exist a nonnegative weighted graph @xmath95 and a subset @xmath31 of @xmath18 , such that @xmath96 for any @xmath94 , then we say that @xmath97 realizes the family @xmath36 and , following @xcite , we call the vertices in @xmath31 * external vertices * and the other vertices * internal vertices*. [ rem : triangleinequality ] let @xmath98 be a positive - weighted graph ; then , for any @xmath99 with @xmath100 and @xmath101 , we have the `` triangle inequality '' @xmath102 let @xmath3 and @xmath103 be two subgraphs of @xmath1 realizing respectively @xmath104 and @xmath105 : the union of these subgraphs is a connected subgraph , @xmath106 , whose set of vertices contain @xmath107 and @xmath108 and then @xmath59 .
so we have : @xmath109
in this section we want to examine when a family of positive real numbers @xmath111 \choose n-1 } } $ ] is treelike .
let @xmath112 . given a family of real numbers @xmath111 \choose n-1 } }
$ ] , we denote @xmath113 by @xmath114 for any @xmath115 $ ] . [ thm : trees ] let @xmath112 and let @xmath111 \choose n-1 } } $ ] be a family of positive real numbers .
( a ) there exists a positive - weighted tree @xmath116 with at least @xmath12 _ vertices _ , @xmath57 , and such that @xmath117 for any @xmath118 if and only if @xmath119-\ { i\ } } d_{\hat{j}}\ ] ] for any @xmath120 $ ] and at most one of the inequalities ( [ disn-2 ] ) is an equality .
\(b ) there exists a positive - weighted tree @xmath116 with at least @xmath12 _ leaves _ , @xmath57 , and such that @xmath121 for any @xmath115 $ ] if and only if @xmath122 -\{i\ } } d_{\hat{j}}\ ] ] for any @xmath120 $ ] .
\(a ) @xmath123 let @xmath124 be a positive weighted tree and let @xmath125 \subset v(t)$ ] .
we want to show that for any @xmath126 $ ] @xmath127 -\{k\ } } d_{\hat{j } } ( { \cal t}).\ ] ] let @xmath128 be a subtree of @xmath129 realizing @xmath130 ; obviously it is a tree with set of leaves contained in @xmath84 $ ] .
let @xmath131 be a subtree of @xmath129 realizing @xmath132 ; obviously it is a tree with set of leaves contained in @xmath133 .
observe that @xmath134 is a subgraph of @xmath135 ( in fact , for every @xmath136 , the path between @xmath137 and @xmath138 in @xmath134 must coincide with the path between @xmath137 and @xmath138 in @xmath128 , since @xmath129 is a tree ) . for every @xmath139 ,
let @xmath140 be the weight of the subgraph @xmath141 .
the inequality ( [ n-2hati ] ) becomes @xmath142 -\{k\ } } ( d_{1, ... ,n } ( { \cal t } ) -a_j),\ ] ] which is equivalent to @xmath143 -\{k\ } } a_j \leq d_{1, ... ,n}({\cal t } ) + ( n-2 ) a_k,\ ] ] which is true , since , obviously , @xmath144 -\{k\ } } a_j \leq d_{1, ... ,n } ( { \cal t})$ ] .
now we want to prove that at most one of the inequalities ( [ disn-2 ] ) is an equality .
suppose @xmath145 -\{k\ } } d_{\hat{j } } ( { \cal t}).\ ] ] then @xmath143 -\{k\ } } a_j = d_{1, ... ,n } ( { \cal t } ) + ( n-2 ) a_k.\ ] ] since @xmath144 -\{k\ } } a_j \leq d_{1, ...
,n}({\cal t})$ ] , we get @xmath146 and @xmath147 -\{k\ } } a_j .\ ] ] thus @xmath148 is a star tree with leaves @xmath149 and center @xmath8 and @xmath150 if and only if @xmath151 .
it is easy to check that in this case only one of the inequalities ( [ disn-2 ] ) is an equality .
@xmath152 we consider two cases : the case where all the inequalities ( [ disn-2 ] ) are strict and the case where exactly one of the inequalities ( [ disn-2 ] ) is an equality .
@xmath153 first let us suppose that all the inequalities ( [ disn-2 ] ) are strict .
let @xmath154 be the star tree with @xmath84 $ ] as set of leaves and center @xmath155 and , for @xmath156 , let @xmath157 -\{k\ } } d_{\hat{j}}- ( n-2 ) d_{\hat{k}}}{n-1 } .\ ] ] for any @xmath158 , we have that @xmath159 is equal to the sum of the weights of all the edges but the edge @xmath160 .
therefore @xmath161 -\{k\ } } \;\ ; \left ( \sum _ { j \in [ n ] -\{h\ } } d_{\hat{j}}- ( n-2 ) d_{\hat{h } } \right ) \right]=\ ] ] @xmath162 -\{k\ } } \;\ ; \left(\sum_{j \in [ n ] } d_{\hat{j}}- ( n-1 ) d_{\hat{h } } \right ) \right]=\ ] ] @xmath163 } d_{\hat{j } } \right ) \;\;- \;\ ; ( n-1 ) \left ( \sum_{h \in [ n ] -\{k\ } } d_{\hat{h } } \right ) \right ] = d_{\hat{k}}.\ ] ] @xmath153 now let us suppose that exactly one of the inequalities ( [ disn-2 ] ) is an equality , precisely @xmath164 -\{r\ } } d_{\hat{j}}.\ ] ] let @xmath154 be the star tree with @xmath165 as set of leaves and center @xmath137 and , for @xmath166 -\{r\}$ ] , let @xmath167 -\{k\ } } d_{\hat{j}}- ( n-2 ) d_{\hat{k}}}{n-1 } .\ ] ] we want to show that , for any @xmath139 , we have that @xmath168 .
let @xmath169 .
thus @xmath159 is equal to the sum of the weights of all the edges but the edge @xmath170 .
therefore @xmath161 -\{r , k\ } } \;\ ; \left(\sum_{j \in [ n ] -\{h\ } } d_{\hat{j}}- ( n-2 ) d_{\hat{h } } \right ) \right]=\ ] ] @xmath162 -\{r , k\ } } \;\ ; \left ( \sum_{j \in [
n ] } d_{\hat{j}}- ( n-1 ) d_{\hat{h } } \right ) \right]=\ ] ] @xmath171 } d_{\hat{j } } \right ) \;\;- \;\ ; ( n-1 ) \left ( \sum_{h \in [ n ] } d_{\hat{h } } -d_{\hat{r } } - d_{\hat{k}}\right ) \right]=\ ] ] @xmath172 } d_{\hat{j } } + ( n-1)d_{\hat{r } } + ( n-1)d_{\hat{k } } \right]= d_{\hat{k}},\ ] ] where the last equality holds because , by ( [ eqr ] ) , we have that @xmath173 . besides , for any @xmath169 , @xmath174 where the last equality holds again by ( [ eqr ] ) .
\(b ) @xmath175 we can argue as in the analogous implication in ( a ) . in this case , since @xmath176 are leaves , all the @xmath177 are nonzero , so we get the strict inequalities ( [ strictn-2 ] ) .
@xmath178 let @xmath154 be the star tree with @xmath54 as set of leaves and center @xmath155 and , for @xmath166 $ ] , let @xmath179 -\{k\ } } d_{\hat{j}}- ( n-2 ) d_{\hat{k}}}{n-1}.\ ] ] we can easily prove that @xmath180 for any @xmath156 .
now we examine the case where there are no internal vertices ( see definition [ intest ] ) .
[ thm : trees2 ] let @xmath112 and let @xmath111 \choose n-1 } } $ ] be a family of positive real numbers .
there exists a positive - weighted tree @xmath116 with exactly @xmath12 vertices , @xmath181 , and such that @xmath182 for any @xmath115 $ ] if and only if the following three conditions hold : \(i ) @xmath183 for any @xmath184 $ ] and at most one of the inequalities ( [ disn-2bis ] ) is an equality .
\(ii ) either one of the inequalities ( [ disn-2bis ] ) is an equality or the maximum in @xmath185 } $ ] is acheived at least twice .
\(iii ) the maximum in @xmath186 } $ ] is acheived at most @xmath187 times . @xmath123 ( i )
the inequality follows from the analogous implication of theorem [ thm : trees ] .
( ii ) , ( iii ) since @xmath188 , at least one of the vertices of @xmath129 must have degree greater or equal than @xmath45 . besides observe that , if a vertex @xmath189 of @xmath129 has degree greater or equal than @xmath45 , then @xmath190 , which is greater or equal than @xmath191 for every @xmath192 .
so @xmath193 is the maximum of @xmath194 } $ ] .
thus , if in @xmath129 there are at least two vertices of degree greater or equal than @xmath45 , then the maximum in @xmath195 } $ ] is acheived twice .
if in @xmath129 there is only one vertex , @xmath8 , of degree greater or equal than @xmath45 , then @xmath129 is a star tree with center @xmath8 and we can check easily that only one of the inequalities ( [ disn-2bis ] ) is an equality .
therefore we have proved ( ii ) . to prove ( iii ) , observe that in @xmath129 there are at least two leaves , @xmath137 and @xmath138 .
since they are leaves , we have : @xmath196 for any @xmath8 vertex of degree at least @xmath45 .
so the maximum in @xmath197 } $ ] is acheived at most @xmath187 times and we have proved ( iii ) .
@xmath152 suppose ( i),(ii ) , and ( iii ) hold . * if at least one ( and then , by assumption , exactly one ) of the inequalities ( [ disn-2bis ] ) is an equality , let us say @xmath198 , let @xmath154 be the star tree with @xmath165 as set of leaves and center @xmath137 and , for @xmath156 , @xmath169 , let @xmath199 we can easily prove that @xmath180 for any @xmath156 .
* if all the inequalities ( [ disn-2bis ] ) are strict , then , by assumption , the maximum in @xmath200 \choose n-1 } } $ ] is acheived at least twice .
so we can suppose @xmath201 for some @xmath202 .
by ( iii ) we have that @xmath203 .
let us divide the set @xmath204 into two nonempty subsets : @xmath205 and @xmath206 .
+ let @xmath207 be the weighted tree with leaves @xmath208 in fig .
[ treenvert ] such that + @xmath209 for @xmath210 , + @xmath211 for @xmath212 , + @xmath213 for @xmath214 .
+ so the path between @xmath215 and @xmath216 has weight @xmath217 .
observe that the weights of the edges of @xmath56 are positive , in fact @xmath218 for @xmath219 and @xmath220 which is positive because we are in the case where all the inequalities ( [ disn-2bis ] ) are strict .
+ let @xmath221 .
then @xmath222 is the sum of the weights of all the edges of @xmath56 , thus : @xmath223 + let @xmath224 .
then @xmath222 is the sum of the weights of all the edges of @xmath56 but the edge @xmath225 if @xmath226 or the edge @xmath227 if @xmath228 . hence @xmath229
in this section we want to examine when a family of positive real numbers @xmath111 \choose n-1 } } $ ] is graphlike .
first we consider the case of graphs with exactly @xmath12 vertices .
obviously we can suppose that , in ( [ primon-22 ] ) , @xmath236 } d_{\hat{j}}$ ] .
therefore @xmath237 -\{i , i_1 , .... , i_{k+1}\ } } d_{\hat{j}}.\ ] ] the inequality above and condition ( i ) imply at once the inequality ( [ primon-22 ] ) .
let @xmath241 be a subtree of @xmath1 realizing @xmath242 .
observe that we can construct an injective map @xmath243 sending a vertex @xmath89 to an edge incident with @xmath89 ( consider @xmath241 as rooted tree with @xmath215 as root and send any leaf to the unique incident edge , then delete these edges and repeat the procedure until you arrive at the root ) . observe that @xmath244 for @xmath245 , in fact : let @xmath246 be the subgraph given by the union of a subtree realizing @xmath114 and of @xmath247 ; the subgraph @xmath246 is connected ( because the ends of @xmath248 are @xmath52 and a vertex among @xmath249 different from @xmath52 , thus a vertex of the subtree realizing @xmath250 ) ; besides @xmath251 ; hence @xmath252 . therefore @xmath253
@xmath254 we can suppose @xmath256 for @xmath257 . obviously , to prove that the inequalities ( [ n-2 ] ) are strict , it suffices to prove that @xmath258 we have already proved that @xmath259 for @xmath245 , in particular for @xmath260 ; besides we know that @xmath261 .
thus we get : @xmath262 @xmath263 let @xmath9 be the weighted graph defined in the following way : consider the complete graph with vertices @xmath281 and weights of the edges equal to @xmath279 and then , for any @xmath278 , draw an edge joining @xmath52 and @xmath215 and an edge joining @xmath52 and @xmath8 , both with weight @xmath282 .
\2 ) we prove the statement by induction on @xmath12 .
precisely we prove , by induction on @xmath12 ( with @xmath287 as base case ) , that if ( i ) holds and the maximum in @xmath255}$ ] is acheived only once , then there exists a weighted graph @xmath288 with exactly @xmath12 vertices such that : observe also that @xmath325 for @xmath326 .
therefore also in the set @xmath327 } $ ] , the maximum is acheived only once ( by @xmath328 ) .
besides the @xmath319 for @xmath329 satisfy ( [ n-2 ] ) with @xmath330 instead of @xmath12 , in fact : obviously it suffices to prove ( [ n-2 ] ) when the first member is the maximum , that is @xmath331 , so it suffices to prove that @xmath332 which is equivalent to @xmath333 that is @xmath334 which follows from ( [ n-2 ] ) .
therefore , by induction assumption , there exists a weighted graph @xmath335 with vertices @xmath336 such that : @xmath337 for any @xmath338 and any subgraph realizing @xmath339 ( which is the maximum of the @xmath340 ) has @xmath215 as vertex ( so @xmath339 is equal to @xmath341 ) .
let @xmath9 be the graph obtained from @xmath342 by adding an edge @xmath343 incident with @xmath215 with weight @xmath344 ; call @xmath12 the other end of @xmath343
. observe that any subgraph realizing @xmath345 has @xmath215 as vertex by the construction of @xmath9 .
we want to show that @xmath230 for any @xmath346 .
obviously , @xmath347 @xmath348 besides , for @xmath349 , @xmath350 the case of graphs with internal vertices is more difficult : in this case condition ( [ n-2 ] ) is not necessary : see [ fig : examplenon2 ] for an example of a graph with four external vertices such that the @xmath351-dissimilarity vector does nt satisfy condition ( [ n-2 ] ) . here
we study the case with internal vertices only in the case @xmath287 .
it seems difficult to generalize the result to @xmath12 greater than @xmath352 .
[ thm : morevertices ] let @xmath353 ; as usual we will denote them respectively by @xmath354 .
there exists a weighted graph @xmath355 with @xmath356 $ ] subset of @xmath18 such that @xmath357 for @xmath358 if and only if condition ( i ) implies @xmath365 for any distinct @xmath360,$ ] so we get : @xmath366 @xmath367 @xmath368 hence @xmath369 obviously condition ( i ) does not imply the strict triangle inequalities : in fact , if we take @xmath370 , condition ( i ) is satisified .
_ proof of theorem [ thm : morevertices ] .
_ @xmath175 let @xmath0 be a weighted graph and let @xmath356 \subset v(g)$ ] .
let @xmath371 for any distinct @xmath372 $ ] .
the subgraph realizing @xmath373 is necessarily a tree with two or three leaves . for any
@xmath374 $ ] , we choose a subtree realizing @xmath375 .
we call @xmath376 the weights of the branches of the subtree realizing @xmath377 if it is a tree with three leaves , @xmath378 .
if it is a tree with two leaves , for instance if it is a tree with leaves @xmath215 and @xmath351 , we call @xmath379 the weight of the path between @xmath215 and @xmath380 , @xmath381 the weight of the path between @xmath45 and @xmath382 and we set @xmath383 and analogously in the other cases ( see [ fig : a1,a2,a3 ] ) . in an analogous way we call @xmath384 the weights of the branches of the subtree realizing @xmath385 , we call @xmath386 the weights of the branches of the subtree realizing @xmath387 , and , finally , we call @xmath388 the weights of the branches of the subtree realizing @xmath389 . if the tree realizing @xmath403 has three leaves and the tree realizing @xmath406 has two leaves , we get : @xmath407 if the leaves of the tree realizing @xmath404 are @xmath408 , @xmath409 if the leaves of the tree realizing @xmath406 are @xmath410 , @xmath411 if the leaves of the tree realizing @xmath404 are @xmath412 . finally , if both the tree realizing @xmath413 and the tree realizing @xmath404 have two leaves , we get : @xmath414 if the leaves of the tree realizing @xmath404 are @xmath408 and the leaves of the tree realizing @xmath413 are @xmath415 and analogously we can argue in the other cases .
that is , @xmath1 is a graph with vertices @xmath422 and , for any @xmath423 $ ] , we join the vertices @xmath352 and @xmath424 with an edge of weight @xmath216 and , for any @xmath423 , j \in [ 3]-\{i\}$ ] , we join the vertices @xmath52 and @xmath425 with an edge of weight @xmath426 . | let @xmath0 be a positive - weighted simple finite graph , that is , let @xmath1 be a simple finite graph endowed with a function @xmath2 from the set of the edges of @xmath1 to the set of the positive real numbers . for any subgraph @xmath3 of @xmath1 , we define @xmath4 to be the sum of the weights of the edges of @xmath3 . for any @xmath5 vertices of @xmath1 , let @xmath6 be the minimum of the weights of the subgraphs of @xmath1 connecting @xmath7 .
the @xmath6 are called @xmath8-weights of @xmath9 . given a family of positive real numbers parametrized by the @xmath8-subsets of @xmath10 , @xmath11 , we can wonder when there exist a weighted graph @xmath9 ( or a weighted tree ) and an @xmath12-subset @xmath13 of the set of its vertices such that @xmath14 for any @xmath15 .
in this paper we study this problem in the case @xmath16 . |
as a result of gravitationally driven shocks and compression , majority of the baryons in the most massive , collapsed objects in the universe such as groups and clusters of galaxies exist in the form of hot plasma with temperature of @xmath3 k. these hot baryons manifest themselves by strong diffuse x - ray sources in the sky within the framework of bremsstrahlung emission , which are directly detectable with current x - ray instruments . on the other hand , the hot electrons in the virialized massive dark halos also scatter the passing cosmic microwave background ( cmb ) photons , giving rise to a subtle change in the cmb spectrum , known as the sunyaev - zeldovich ( sz ) effect ( sunyaev & zeldovich 1972 ) . unlike the x - ray emission of the hot gas in massive halos , which shows a strong dependence on the detailed structures and enrichness history of the hot gas as well as the distances of the host halos from us ,
the sz effect is completely determined by the total thermal energy of the gas intrinsic to the systems .
therefore , it is well suited for studies of distant , massive clusters , which may provide a stringent constraint on cosmological parameters especially for the mean matter density @xmath4 and the normalization of matter power spectrum @xmath5 ( e.g. barbosa et al .
1996 ; molnar , birkinshaw & mushotzky 2002 ; komatsu & seljak 2002 ) .
a more practical and powerful utilization of the sz effect nowadays is perhaps to statistically measure the angular power spectrum of the sz signals at various scales .
this allows us to gain the weak sz signals at high statistical significance levels without resolving individual groups and clusters .
indeed , recent detection of the excess power relative to primordial cmb anisotropy at arcminute scales has been attributed to the statistical signals of the thermal sz effect ( dawson et al .
2001 ; bond et al .
. a number of ongoing and future experiments will soon be able to provide a more robust measurement of the sz angular power spectrum at high multipoles or small angular scales .
there are many other effects that should be corrected for before the sz angular power spectrum is used for cosmological purposes .
these include nongravitational effect on the distribution and evolution of the hot gas inside halos ( da silva et al .
2000 , 2001 ; seljak , burwell & pen 2001 ; xue & wu 2001 ; holder & carlstrom 2001 ; zhang , pen & wang , 2002 ; komatsu & seljak 2002 ; zhang & wu 2003 ; etc . ) , possible asymmetric shape of dark halos ( cooray 2000 ; lee & suto 2003 ; piffaretti , jetzer & schindler 2003 ) , contaminations of radio point sources ( cooray et al . 1998
; lin , chiueh & wu 2002 ; holder 2002 ; rubin - martn & sunyaev 2002 ) , etc . despite large uncertainties ,
many attempts have been made over the past few years towards the evaluation of these effects on the sz cluster counts and angular power spectrum measurements . in this work , we will explore the contamination of radio sources in clusters in the measurement of the thermal sz angular power spectrum , namely , cluster radio halos and galaxies . in theory of hierarchical clustering ,
massive dark halos like clusters of galaxies form by gravitational aggregation of individual low - mass objects .
sub - halo mergers will generate a considerably large number of energetic electrons that can be accelerated by cluster magnetic fields over cluster regions , giving rise to the radio halos due to synchrotron emission ( buote 2001 and reference therein ) .
regardless of the debate on the acceleration mechanism for relativistic electrons in clusters , about @xmath6 of nearby , rich clusters exhibit radio halos , and this fraction may increase with redshift if mergers were more frequent in the past ( enlin & rttgering 2002 ) .
furthermore , there is a strong observational evidence that cluster radio halos show a morphology similar to the diffuse x - ray emission from the hot intracluster gas , and the corresponding radio power is correlated strongly with the x - ray luminosity and temperature of clusters ( liang et al .
2000 ; govoni et al .
2001a , b ) .
this enables one to estimate the distribution and evolution of cluster radio halos based on the well - determined x - ray properties of clusters .
for example , one can easily derive the radio luminosity function of clusters at different wavebands and redshifts ( enlin & rttgering 2002 ) .
together with the radio surface brightness profiles of clusters motivated by empirical formula ( e.g. @xmath7 model ) for the x - ray surface brightness distribution of clusters , one can estimate the radio background from the diffuse radio halos of clusters .
existence of diffuse radio halos constitutes a major source of contaminations in the measurement of the thermal sz effect toward clusters .
since radio halos always contribute a positive signal superimposed on the cmb sky , the decrements of the cmb signals due to the thermal sz effect of the hot intracluster gas at low frequencies would be compensated , while the increments of the sz signals at high frequencies will be correspondingly enhanced .
the crossover point occurs at @xmath8 ghz .
moreover , the amplitude of such a modification to the thermal sz signals depends critically on observing frequencies .
this arises because the radio flux of cluster halos as a result of synchrotron emission has a steep spectrum of roughly @xmath9 .
consequently , at very high frequencies the cluster radio halos may become hardly visible , and no modification to the sz signals from the cluster radio halos is needed .
however , with the decrease of observing frequencies the sz decrements and the diffuse radio emission of cluster halos may appear to be comparable , and eventually the latter would become the dominant component in the microwave sky .
this will lead to a nonnegligible correction to the sz measurement ( enlin 2002 ) especially in the central regions of clusters , because the radio emission is more centrally concentrated than the thermal sz effect if the radio emission is assumed to follow the x - ray emission .
therefore , it deserves a close investigation of the influence of the cluster radio halos on the measurement of the sz angular power spectrum .
another source of contaminations in the sz measurements is radio point sources , which include radio galaxies inside clusters and tracing large - scale structures of the universe . in a similar way to cluster radio halos
, cluster radio galaxies also compensate the decrements of the thermal sz intensity in the rayleigh - jeans region . on the other hand ,
the poisson distribution and clustering of radio galaxies on large scales would enhance the radio fluctuations , resulting in a positive contribution to both primary cmb and sz power spectra .
the role of these field radio galaxies has been well addressed by several authors ( e.g. franceschini et al . 1989 ; toffolatti et al . 1998 ) , and the resulting contamination in the measurement of cmb anisotropies is found to be generally small or can be somehow removed .
nevertheless , a significant fraction of the sz signals could be diluted by the unsubtracted radio sources inside clusters ( cooray et al . 1998
; lin , chiueh & wu 2002 ) . using a toy model of radio point source subtraction , in which a certain central region is removed from all clusters , holder ( 2002 ) showed that the true sz power at @xmath10 can be underestimated by as much as @xmath11 at @xmath12 ghz . a more realistic model for the radial distribution of radio galaxies in clusters
is thus needed to refine this estimation . in this paper ,
using a purely analytical approach , we perform a quantitative study of the contamination of cluster radio halos and galaxies in the measurement of the angular power spectrum of the thermal sz effect .
it is hoped that such a study may clarify the issues of to what extent one has to take the correction of cluster radio sources into account in the cosmological applications of the thermal sz angular power spectrum . throughout the paper we will work with a flat cosmological model of @xmath13 and @xmath14 .
the change in cmb temperature @xmath15 in direction @xmath16 due to the thermal sz effect of the hot electrons in massive clusters is simply @xmath17 where @xmath18 , @xmath19 and @xmath20 are the number density and temperature of electrons in clusters , respectively , and the integral is performed along the line of sight @xmath21 .
the angular power spectrum of the temperature fluctuation @xmath22 can be separated into the poisson term @xmath23 and the clustering term @xmath24 : @xmath25 since the clustering term @xmath24 is several orders of magnitude smaller than the poisson term @xmath23 at small angular scales @xmath26 where the sz power becomes dominant in the cmb sky , we will neglect the contribution of source clustering to the sz angular power spectrum . in the flat sky approximation
, the poisson term can be calculated by ( e.g. cole & kaiser 1988 ) @xmath27 in which @xmath28 denotes the comoving volume per unit redshift and per steradian , and @xmath29 is the fourier transform of @xmath30 : @xmath31 where @xmath32 is the bessel function of zero order .
we adopt the standard press & schechter ( 1974 ) formalism for the mass function of clusters @xmath33 where @xmath34 is the mean mass density of the universe , @xmath35 is the linear over - density of spherical collapse that virialized at redshift @xmath36 , and @xmath37 is the present rms mass fluctuation within the top - hat filter @xmath38 : @xmath39 in which @xmath40 is the fourier representation of the window function .
we parameterize the power spectrum of fluctuation @xmath41 , and take the fit given by bardeen et al .
( 1986 ) for the transfer function of adiabatic cdm model @xmath42 with the shape parameter @xmath43 .
the primordial power spectrum is assumed to be the harrison - zeldovich case @xmath44 .
the amplitude in the power spectrum is fixed using the rms fluctuation on an 8 @xmath45 mpc scale , @xmath5 , which will be taken to be @xmath46 in this study .
over the past few years , diffuse radio halos have been detected in a few tens of nearby , rich clusters .
they often extend to a distance of @xmath47 mpc from cluster centers , and have regular shape , low surface brightness and steep radio spectrum @xmath9 .
moreover , observations suggest that there is a tight correlation between the radio power @xmath48 and the x - ray luminosity ( @xmath49 ) of host clusters ( liang et al . 2000 ; govoni et al .
2001b ) , which can be well described by the following fitting formula ( enlin & rttgering 2002 ) @xmath50 in which @xmath51 , @xmath52 , and @xmath53 is the present hubble constant in units of @xmath54 km s@xmath55 mpc@xmath55 . in order to set up a link between the x - ray luminosity and total mass of the cluster ,
we first employ the best - fit @xmath49-@xmath56 relation of xue & wu ( 2000 ) , based on the hitherto largest sample of nearby x - ray clusters , @xmath57 it has been shown that such an @xmath49-@xmath56 relation exhibits no significant evolution at least out to @xmath58 ( rosati , borgani & norman 2002 and references therein ) .
we then convert the x - ray temperature into the total cluster mass using virial theorem ( bryan & norman 1998 ) : @xmath59 in which @xmath60^{1/3},\ ] ] @xmath61 is the hubble constant in units of 100 km s@xmath55 mpc@xmath55 , @xmath62 is the normalization factor , which will be taken to be @xmath63 in this work , and @xmath64 is the overdensity of the virialized dark halos with respect to the critical density of the universe at redshift @xmath36 .
equivalently , one may also employ the empirical @xmath49-@xmath65 relation ( reiprich & bhringer 1999 ) incorporating with a simple model of cosmic evolution for dark halos .
consequently , for a given halo of mass @xmath65 one is able to estimate its radio power at 1.4 ghz through @xmath66 radio observations indicate that cluster radio halos usually have a steep power index , which is also well motivated by synchrotron emission process .
we adopt a spectral power index of @xmath67 ( e.g. giovannini et al .
1993 ) to convert the radio power at 1.4 ghz to the radio emission at other wavelengths .
furthermore , several observations have also reveled a similarity of the radio and x - ray morphologies in clusters ( deiss et al .
1997 ; liang et al . 2000 ; govoni et al . 2001a ; feretti et al . 2001 ) .
this enables us to model the radio surface brightness of clusters using the empirical fitting formula for x - ray surface brightness profile , i.e. , the @xmath7 model , @xmath68 in which the normalization factor can be determined by the total radio power @xmath69 at a given wavelength @xmath70 .
meanwhile , the electron number density in clusters follows @xmath71 here we introduce the universal baryon fraction @xmath72 to fix the central electron density @xmath73 : @xmath74 , where @xmath75 is the total gas mass within @xmath76 .
the gaseous halo ( also the radio halo ) is truncated at the virial radius @xmath76 defined by @xmath77 where @xmath78 is the critical mass density of the universe at @xmath36 .
we adopt a constant value of @xmath79 for the @xmath7 parameter , and assume that the core radius @xmath80 of gas distribution is proportional to the size of the host halo : @xmath81 .
we specify the proportionality constant such that @xmath82 @xmath45 mpc for a massive cluster of @xmath83 at @xmath84 ( cf . komatsu & kitayama 1999 ) .
finally , radio halos are preferentially found in clusters that demonstrate violent mergers and substructures , which provide an accelerating mechanism for the relativistic particles responsible for radio emission ( e.g. buote 2001 ) . yet , the fraction of clusters that possess giant radio halos is still uncertain .
following enlin & rttgering ( 2002 ) , we adopt a constant radio halo fraction of @xmath85 .
nevertheless , we will also demonstrate the result for a fraction of unity for comparison . recall that cluster mergers may occur more frequently in the past than at the present . in the rayleigh - jeans region ,
the presence of radio halos can dilute the thermal sz signals towards clusters .
the significance of the effect depends critically on working frequencies .
recall that radio emission of clusters has a steep power spectrum of @xmath86 . at very low frequencies ,
the decrements of the cmb temperature due to the sz effect could be completely compensated ( enlin 2002 ) by cluster radio halos . here
we demonstrate the radio halo contamination using an observing frequency of @xmath12 ghz .
the change in the cmb temperature is related to the intensity fluctuations by @xmath87 where @xmath88 .
@xmath89 is composed of two components , the thermal sz signals @xmath90 and the cluster radio surface brightness @xmath91 , @xmath92 or equivalently , we can separated the temperature fluctuation into @xmath93 if the fraction of cluster radio halos is @xmath94 , then the observed sz power spectrum at small angular scales is @xmath95,\ ] ] where @xmath96 is the fourier transform of @xmath97 . in order to highlight the effect of cluster radio halos on the sz angular power spectrum , we demonstrate in figure 1 the relative correction to the thermal sz angular power spectrum @xmath98 , along with the contributions of clusters in different redshift ranges . at @xmath99 ,
the radio halo contamination in the sz angular power spectrum arises mainly from clusters at @xmath100 .
high - redshift clusters come into effect only at very small angular scales @xmath101 , while large angular scales are dominated by low - redshift clusters . in figure 2
we show the contributions of cluster radio halos with different cluster masses to the thermal sz angular power spectrum . at large angular scales below @xmath102 ,
the radio contamination is governed by nearby , massive clusters , which can be easily removed from the sz map because of their rarity . at smaller angular scales beyond @xmath103 ,
the contamination arises mainly from moderately rich clusters at higher redshifts @xmath104 .
contributions from poor clusters and groups of masses less than @xmath105 @xmath106 are negligible because of the absence of radio halos in these systems . to provide a quantitative illustration of the modifications due to cluster radio halos
, we list in table 1 the relative corrections @xmath107 for a set of @xmath108 evaluated by equation ( 20 ) , assuming the radio halo fractions of @xmath85 and @xmath109 , respectively .
we have also shown in table 1 how our estimates are affected by the uncertainties associated with the measurements of radio power , x - ray luminosities and total masses of clusters . to do this
, we simply vary the radio power @xmath69 , x - ray luminosity @xmath110 and normalization factor @xmath111 in the virial theorem by @xmath0 , @xmath0 and @xmath11 , respectively , which are roughly representative of the uncertainties in current determinations of these quantities .
it turns out that the modifications to the thermal sz angular power spectrum from cluster radio halos are only minor , and the maximum correction is no more than @xmath112 for @xmath113 even in the extreme case of where @xmath109 and @xmath114 .
major uncertainty in our calculation arises from the mass estimate of clusters characterized by the normalization factor @xmath111 , which could differ by a factor of @xmath115 among different simulations and observations ( e.g. bryan & norman 1998 ; huterer & white 2002 ) . 0.2truein 0.2truein cccccc @xmath94 & @xmath116 & @xmath117 & @xmath118 & @xmath119 + @xmath6 & @xmath120 & @xmath121 & @xmath122 & @xmath123 + 1 & @xmath124 & @xmath125 & @xmath126 & @xmath127 +
a critical issue in the estimate of confusion of cluster radio galaxies in the measurement of the sz angular power spectrum is to properly handle the distribution of cluster radio galaxies .
this includes the luminosity function and radial profile of radio galaxies in clusters .
we adopt the univariate radio luminosity function given by ledlow & owen ( 1996 ) , based on a statistical complete sample of 188 radio galaxies in abell clusters within @xmath128 and at @xmath129 ghz .
the univariate radio luminosity function @xmath130 can be well fitted by a broken power law @xmath131 where @xmath132 is the total number of galaxies brighter than @xmath133 within 0.3 abell radius of the cluster center , and @xmath134 are proportionality constants with ( @xmath135 ) and ( @xmath136 ) for @xmath137 whz@xmath55 and @xmath138 whz@xmath55 , respectively . in order to convert the above radio luminosity function at @xmath129 ghz into the ones ( @xmath139 ) at other frequencies , we use the spectral index distribution fitted by lin , chiueh , & wu ( 2002 ) for a sample of 64 radio sources in 56 clusters observed by cooray et al .
( 1998 ) , in which the mean spectral index is @xmath140 .
this allows us to take the dispersion of spectral index into account .
the luminosity - weighted cluster radio luminosity function takes the form of @xmath141 where @xmath142 and the break power @xmath143 has a weak dependence on frequency : log@xmath144 .
we assume that the number density profile of galaxies in clusters follows the king model @xmath145 the surface number density is thus @xmath146 .
a good approximation of the core radius is @xmath147 . in order to fix the central density @xmath148
, we introduce the so - called halo occupation distribution , @xmath149 .
we adopt the best fit analytic formulae of sheth & diaferio ( 2001 ) for the spiral [ @xmath150 and elliptical [ @xmath151 galaxies based on the gif simulations ( kauffmann et al .
1999 ) : @xmath152 ^ 2 } \nonumber\\ n_{\rm gal , e}(m)&=&(m / m_{\rm e})^{\alpha_e } e^{-(2\times10^{11}m_{\sun}/m)^2 } \nonumber\\ n_{\rm gal}(m)&=&n_{\rm gal , s}(m)+n_{\rm gal , e}(m ) , \end{aligned}\ ] ] where @xmath153 , @xmath154 , @xmath155 , and @xmath156 .
another parameter we need to fix before we proceed to our numerical calculation is @xmath132 , the galaxies brighter than @xmath133 within 0.3 abell radius @xmath157 . instead of evaluating @xmath132 directly ,
we turn to the fraction of galaxies with absolute magnitude brighter than @xmath158 , @xmath159 .
this can be achieved by employing the luminosity functions of galaxies in clusters , @xmath160 and @xmath161 .
we adopt the best fit schechter forms in the @xmath162 band for early and late type galaxies derived from the sloan digital sky survey ( goto et al .
2002 ) . the best fit parameters in goto et al .
( 2002 ) are based on galaxies within @xmath163 @xmath164 mpc , which is only slightly larger than @xmath165 .
we convert the @xmath162 magnitude to the standard johnson - morgan - cousins system using @xmath166 ( fukugita et al .
1996 ) and then assign the colors to early and late type galaxies according to the tabulated values given by fukugita , shimasaku & ichikawa ( 1995 ) .
moreover , we choose a lower magnitude cutoff of @xmath167 to guarantee the validity of the above adopted halo occupation distribution equation ( 25 ) which accounts for galaxies brighter than @xmath168 . taking all these factors into account
, we have @xmath169 the total radio power of all galaxies in a cluster can be obtained by extrapolating the radio luminosity function of ledlow & owen ( 1996 ) determined with @xmath165 to the whole cluster region out to virial radius , which yields @xmath170 \int p_{\nu } ( dn_{\rm radio}/d\log p_{\nu } ) d\log p_{\nu } \nonumber\\ & = & f_{\rm gal } n_{\rm gal}(m ) \int p_{\nu } f_{\nu } d\log p_{\nu}.\end{aligned}\ ] ] the total surface brightness of all radio galaxies measured at frequency @xmath171 in direction @xmath16 toward a cluster of mass @xmath65 and redshift @xmath36 is thus @xmath172 \left[\frac{(1+z)p_{\rm tot,\nu}}{4\pi d_l^2(z ) } \right ] \nonumber\\ & = & \frac{\sigma_{\rm gal}({\mbox{\boldmath $ \theta$}})f_{\rm gal } \int p_{\rm \nu}f_{\rm\nu}d{\rm log}p_{\rm \nu}}{4\pi(1+z)^3},\end{aligned}\ ] ] where @xmath173 and @xmath174 are the angular diameter distance and luminosity distance to the cluster , respectively .
this adds to the sz signals such that the observed radio intensity is @xmath175 .
the corresponding angular power spectrum reads @xmath176 where @xmath177 is the fourier transform of @xmath178 , and the temperature fluctuation @xmath179 is related to @xmath180 following equation ( 17 ) . in a similar way to figure 1 ,
we plot in figure 3 the relative corrections to the sz angular power spectrum @xmath98 because of the presence of radio galaxies in clusters , together with the contributions of cluster radio galaxies in different redshift ranges . at @xmath10 ,
the distant radio galaxies at @xmath104 , which could be regarded as the unresolved radio sources , dominate the correction term .
we list in table 2 the relative corrections for a set of small scales , together with the uncertainties corresponding to @xmath0 variation of the radio luminosity function .
it turns out that the confusion of unresolved radio galaxies in clusters in the measurement of the thermal sz effect is less than @xmath181 on all scales beyond @xmath1 .
this is somewhat smaller than the estimate ( @xmath11 ) of holder ( 2002 ) , who evaluated the effect of radio point sources by removing a certain central region ( @xmath182 ) from all clusters .
the latter could simultaneously lead to an underestimate of the thermal sz effect in the centers of all clusters especially for high - redshift clusters . 0.2truein 0.2truein cccccc redshift & @xmath116 & @xmath117 & @xmath118 & @xmath119 + @xmath183 & @xmath184 & @xmath185 & @xmath186 & @xmath187 + @xmath188 & @xmath189 & @xmath190 & @xmath191 & @xmath192 +
contamination of radio sources has been a major concern in the measurement of thermal sz effect , which include radio halos and galaxies of clusters , and clustering of radio galaxies tracing large - scale structures of the universe .
the latter comes into effect only on large angular scales and appears to be relatively small as compared with the thermal sz effect ( toffolatti et al .
we have thus concentrated in this paper on the role of cluster radio halos and galaxies .
the effect of cluster radio halos on the measurement of the sz angular power spectrum has previously remained unknown , while the confusion of the cluster radio galaxies has been recently shown to be very significant ( holder 2002 ) .
using an analytic model for both radio sources and dark matter of clusters , we have performed a quantitative analysis of the extent to which cluster radio halos and galaxies may contaminate the measurement of thermal sz angular power spectrum .
we have found that while the cluster radio sources can indeed cause some confusion in the measurement of sz angular power spectrum , the relative correction @xmath107 is well within @xmath0 at @xmath1 for an observing frequency of @xmath12 ghz .
the modification becomes even smaller for high - frequency observations .
we thus conclude that there is probably no need to take the confusion of cluster radio sources into account in the thermal sz measurement , if our observing frequencies are chosen to be @xmath2 ghz , unless the thermal sz power spectrum is used for the purpose of precision cosmology .
we gratefully acknowledge torsten enlin for useful discussion and help , and an anonymous referee for useful suggestions .
this work was supported by the national science foundation of china , under grant no .
10233040 , and the ministry of science and technology of china , under grant no .
nkbrsf g19990754 .
barbosa , d. , bartlett , j. g. , blanchard , a. , & oukbir , j. 1996 , , 314 , 13 bardeen , j. m. , bond , j. r. , kaiser , n. , & szalay , a. s. 1986 , , 304 , 15 bond , j. r. , et al .
2003 , , in press ( astro - ph/0205386 ) buote , d. a. 2001 , , 553 , l15 bryan , g. l. , & norman , m. l. 1998 , , 495 , 80 carlberg , r. g. , et al .
1996 , , 462 , 32 cole , s. , & kaiser , n. 1988 , , 233 , 637 condon , j. j. , et al .
1998 , , 115 , 1693 cooray , a. r. 2000 , , 313 , 783 cooray , a. r. , et al .
1998 , , 115 , 1388 cooray , a. , & melchiorri , a. 2002 , , 66 , 083001 da silva , a. c. , barbosa , d. , liddle , a. r. , & thomas , p. a. 2000 , , 317 , 37 da silva , a. c. , et al .
2001 , , 561 , l15 david , l. p. , et al .
1993 , , 412 , 479 dawson , k. s. , et al .
2001 , , 553 , l1 deiss , b. m. , reich , w. , lesch , h. , & wielebinski , r. 1997 , , 321 , 55 dunlop , j. s. , & peacock , j. a. 1990 , , 247 , 19 enlin , a. t. , & rttgering , h. 2002 , , 396 , 83 enlin , a. t. 2002 , , 396 , 17 feretti , l. 1999 , in the universe at low radio frequencies , iau symp . 199 ( astro - ph/0006379 ) feretti , l. , fusco - femiano , r. , giovannini , g. , & govoni , f. 2001 , , 373 , 106 franceschini , a. , toffolatti , l. , danese , l. , & de zotti , g. 1989 , , 344 , 35 fukugita , m. , shimasaku , k. , & ichikawa , t. 1995 , , 107 , 945 giovannini , g. , et al .
1993 , , 406 , 399 goto , t. , et al .
2002 , , 54 , 515 govoni , f. , enlin , t. a. , feretti , l. , & giovannini , g. 2001a , , 369 , 441 govoni , f. , et al . 2001b , , 376 , 803 herbig , t. , & readhead , a. c. s. 1992 , , 81 , 83 holder , g. p. 2002
, , 580 , 36 holder , g. p. , & carlstrom , j. e. 2001 , , 558 , 515 huterer , d. , & white , m. 2002 , , 578 , l95 kauffmann , g. , colberg , j. m. , diaferio , a. , & white , s. d. m. 1999 , , 303 , 188 king , i. r. 1962 , , 67 , 471 komatsu , e. , & kitayama , t. 1999 , , 526 , l1 komatsu , e. , & seljak , u. 2002 , , 336 , 1256 ledlow , m. j. , & owen , f. n. 1996 , , 112 , 9 lee , j. , & suto , y. 2003 , , 585 , 151 liang , h. , hunstead , r. w. , birkinshaw , m. , & andreani , p. 2000 , , 544 , 686 lin , h. , chiueh , t. , & wu , x .-
2002 , preprint ( astro - ph/0202174 ) loeb , a. , & refregier , a. 1997 , , 476 , l59 mason , b. s. , et al .
2003 , , in press ( astro - ph/0205384 ) molnar , s. m. , birkinshaw , m. , & mushotzky , r. f. 2002 , , 570 , 1 mushotzky , r. f. , & scharf , c. a. 1997 , , 482 , l13 padin , s. , et al .
2001 , , 549 , l1 piffaretti , r. , jetzer , p. , & schindler , s. 2003 , , 398 , 41 pointecouteau , e. , et al .
2001 , , 552 , 42 press , w. h. , & schechter , p. 1974
, , 187 , 425 reiprich , t. h. & bhringer , h. 1999 , , 320 , 296 rosati , p. , borgani , s. , & norman , c. 2002 , , 40 , 539 rubio - martn , j. a. , & sunyaev , r. a. 2002 , , submitted ( astro - ph/0211430 ) seljak , u. , burwell , j. , & pen , u .- l .
2001 , , 63 , 063001 sheth , r. k. , & diaferio , a. 2001 , , 322 , 901 slee , o. b. , roy , a. l. , & andernach , h. 1996 , australian j. phys . , 49 , 977 sunyaev , r. a. , & zeldovich , y. b. 1972 , comm .
space phys . , 4 , 173 toffolatti , l. , et al .
1998 , , 297 , 117 white , s. d. m. , navarro , j. f. , evrard , a. e. , & frenk , c. s. 1993 , , 336 , 429 xue , y .- j .
, & wu , x .-
2000 , , 538 , 65 xue , y .- j .
, & wu , x .-
2001 , , 552 , 452 zhang , p. , pen , u .- l . , &
wang , b. 2002 , , 577 , 555 zhang , y .-
y , & wu , x .- p .
2003 , , 583 , 529 | we present a quantitative estimate of the confusion of cluster radio halos and galaxies in the measurement of the angular power spectrum of the thermal sunyaev - zeldovich ( sz ) effect . to achieve the goal
, we use a purely analytic approach to both radio sources and dark matter of clusters by incorporating empirical models and observational facts together with some theoretical considerations .
it is shown that the correction of cluster radio halos and galaxies to the measurement of the thermal sz angular power spectrum is no more than @xmath0 at @xmath1 for observing frequencies @xmath2 ghz .
this eliminates the concern that the sz measurement may be seriously contaminated by the existence of cluster radio sources . |
recently , various low - energy properties arising from anharmonic ion oscillations have attracted much attention .
an ion located at the center of an over - sized cage oscillates with large amplitude and thus the anharmonic terms in the potential energy play an important role .
indeed , these anharmonic oscillations are observed in @xmath7-pyrochlore,@xcite filled - skutterudite,@xcite and clathrate compounds .
@xcite in metallic systems , such anharmonic oscillations interact with conduction electrons , and due to their large amplitude , the electron - phonon coupling constant becomes large .
the large anharmonicity and the strong electron - phonon coupling lead unusual relaxation of conduction electrons,@xcite anomalous nuclear magnetic relaxation time,@xcite the sound velocity anomalies,@xcite and the strong coupling superconductivity.@xcite these anharmonic oscillations in these systems have also been studied theoretically .
dahm and ueda discussed the anomalous temperature dependence of the resistivity@xcite and the nmr relaxation time@xcite observed in kos@xmath1o@xmath2 by using a single - site anharmonic phonon model , employing the self - consistent gaussian approximation for the quartic term of the ion displacement.@xcite recently , yamakage and kuramoto extended this approach to the lattice problem in the same level of gaussian approximation.@xcite these works can explain the softening of optical - phonon frequency as temperature decreases observed in many compounds.@xcite the anharmonic ion oscillations had been discussed also as a possible mechanism of high transition temperature superconductivity , particularly for high @xmath8 cuprates.@xcite as for the systems of anharmonic ion oscillations in over - sized cages , importance of contributions of low - energy pr ion oscillations to the superconductivity is pointed out in pros@xmath9sb@xmath10.@xcite recently , chang , _
et al_.,@xcite discussed the superconductivity in kos@xmath1o@xmath2 using strong coupling theory of @xmath11-wave superconductivity@xcite also employing the gaussian approximation@xcite for anharmonicity . in this paper , we focus on the @xmath7-pyrochlore compounds @xmath12os@xmath1o@xmath2 ( @xmath12=k , rb , or cs ) .
monovalent @xmath12-cations are located at the center of the os@xmath10o@xmath13 cages and form a diamond lattice structure .
lattice dynamics was investigated by neutron - scattering experiments,@xcite and a low - energy optical phonon is observed at about 3 mev in the k compound .
this is basically k - cation oscillations , and the same phonon is also observed at around 5 - 7 mev in the rb and cs compounds .
the root - mean - square amplitude of the k oscillation turns out to be 0.12 - 0.14 at zero temperature from the elastic neutron scattering@xcite while much smaller values are reported for rb and cs compounds.@xcite the first - principle calculations indicate that the k ion potential has large anharmonicity and is very shallow along [ 111 ] and three other equivalent directions.@xcite it is also calculated that the first excited state has the excitation energy of @xmath148 k and k - cation oscillation amplitude is as large as 1 at zero temperature .
these values are quite different from the experimental data , indicating that some parameters in the ion potential are not so realistic . in the present paper
, we will first systematically analyze the effects of anharmonicity of the ion potential on the ion dynamics in kos@xmath1o@xmath2 .
the @xmath7-pyrochlore compounds reveal superconducting phase transition and the transition temperature @xmath8 s are 9.6 k , 6.3 k and 3.3 k for @xmath12=k , rb , and cs , respectively .
these values of @xmath15 are inversely related to the @xmath12-cation size , _
@xmath8 is the highest for the smallest ion : potassium .
the symmetry of the gap function is considered to be fully gapped @xmath11-wave.@xcite it has been also shown that the electron - phonon coupling is large.@xcite thus , it is expected that the conduction electrons on the cages strongly interact with the anharmonic oscillations of ion inside the cage in these systems . for kos@xmath1o@xmath2 , in addition to superconducting transition
, there exists a first - order structural transition at @xmath16 k. at this transition , no sign of symmetry breaking has been observed.@xcite the oscillation of k - cation seems less anharmonic below @xmath17 as indicated by electrical resistivity , specific heat jump at @xmath8 in magnetic fields and the mean free path estimated from the upper critical field @xmath18.@xcite recently , we have proposed that this is driven by a sudden change in k - cation oscillation amplitude driven by intersite ion interactions.@xcite the amplitude of the low - energy excited state with @xmath19 symmetry jumps at @xmath17 , but this does not change the @xmath20 point - group symmetry .
the main purpose of this paper is to clarify the properties of anharmonic oscillations in tetrahedral symmetry , which is the local symmetry for @xmath12-cations in @xmath12os@xmath1o@xmath2 , and the superconductivity mediated by these anharmonic oscillations . in order to fully take into account the anharmonicity and anisotropy
, we will solve the three - dimensional schrdinger equation for an anharmonic oscillator in the tetrahedral symmetry . using these exact phonon eigenstates
, we will then discuss the strong coupling superconductivity assuming an @xmath11-wave pairing .
this paper is organized as follows . in sec .
[ anharmo ] , we will calculate the energy spectrum of the anharmonic potential problem and discuss various thermodynamic and dynamical quantities and their dependence on temperature .
anharmonic effects on debye - waller factor will also be discussed .
section [ scs ] is devoted to the discussions for strong coupling superconductivity mediated by the anharmonic ion oscillations discussed in sec .
[ anharmo ] . in sec .
[ discuss ] , we will discuss the relevance of the present results to the @xmath7-pyrochlore compounds .
we also apply our theory to discuss the changes in phonon dynamics at the first - order transition at @xmath17 .
finally sec .
[ conclusion ] is a summary of this paper .
in this paper , we investigate an anharmonic oscillation of k ion in kos@xmath1o@xmath2 .
our model is anharmonic local phonons at each lattice point and we assume the local symmetry is tetrahedral one which corresponds to the case of k - site symmetry in kos@xmath1o@xmath2 . in tetrahedral symmetry , in addition to
spherical and cubic fourth order terms , there exists a third - order anharmonic term which breaks inversion symmetry and the hamiltonian is given by @xmath21 where @xmath22 is the real - space displacement of the ion from the equilibrium position .
@xmath23 is the mass of the ion . throughout this paper
we set @xmath24 , where @xmath25 is the mass of electron , except for discussions in sec .
[ discusschemi ] , and this corresponds to the mass of k ion . @xmath26 and @xmath27 are coefficients of second- and third - order terms in the ion potential , while @xmath28 and @xmath29 are isotropic and cubic fourth - order terms , and we ignore the higher - order potential terms of @xmath30 . to study dependence on potential parameters , it is useful to rewrite hamiltonian ( [ eqh ] ) into a dimensionless form by renormalizing displacement and energy by their units . as for the energy unit , we choose the energy of harmonic phonon corresponding the second - order term @xmath31 .
the unit of the length is chosen as @xmath32 . here , @xmath33 is the bohr radius @xmath34 . in these units ,
hamiltonian ( [ eqh ] ) is transformed to @xmath35 where @xmath36 . @xmath0 and @xmath37 and @xmath38 are dimensionless constants and @xmath39 is introduced to vary the second - order term for later purpose , but @xmath40 for the main part of this paper . in secs .
[ scs ] and [ discuss ] , we will use different values of @xmath39 to examine the effects of changes in the second - order term.@xcite throughout this paper we set @xmath41 k which corresponds to the energy scale of the optical - phonon frequency observed in @xmath7-pyrochlore compounds,@xcite and thus @xmath42 . in the limit of @xmath43 , hamiltonian ( [ h00 ] ) can be diagonalized by using creation and annihilation operators as , @xmath44 where @xmath45 with @xmath46 , @xmath47 or @xmath48 .
eigenstates are labeled by three occupation numbers as @xmath49 and @xmath50 . for nonzero @xmath0 , @xmath37 , and @xmath38 , we diagonalize hamiltonian ( [ h00 ] ) numerically in the restricted hilbert space spanned by @xmath51 with @xmath52 .
in this paper , we use @xmath53 which corresponds to the hilbert space with 12341 states and check the convergence by comparing the results for @xmath54 including 23426 states .
we employ this approach rather than the conventional self - consistent gaussian approximation@xcite .
this is because the conventional approximation does not work for the potential , eq .
( [ h ] ) since the third - order term can not be decoupled as in the fourth - order terms .
it is also important that , as will be shown later , energy differences between adjacent eigenstate multiplets are not the same , and this property can not be described by the self - consistent gaussian approximation . in fig .
[ fig - spec ] , we show a low - energy part of the energy spectra of hamiltonian ( [ h00 ] ) as a function of the third - order anharmonicity @xmath0 for @xmath55 and @xmath56 .
the ground state is always nodeless in the space @xmath57 and therefore singlet , while the first excited states are triplet , corresponding to @xmath11 and @xmath58 orbitals for the case of isotropic potential . as @xmath0 increases
, the energy of the lowest singlet excited state ( hereafter , this state will be referred to as @xmath59 state ) decreases and shows an anticrossing with the ground state at @xmath60 .
this kind of anti - crossing behavior does not occur in the one - dimensional anharmonic potential problem : @xmath61.@xcite interestingly , near @xmath62 the five lowest - energy states are well separated from other states in the energy spectra .
thus , the validity of our previous five - state toy model is justified around this crossover region.@xcite for @xmath63 , the excited triplet states , which are @xmath58-wave like , are nearly degenerate with the ground state .
this means that these four states are localized away from the origin in the four directions : @xmath64 $ ] , @xmath65 $ ] , @xmath66 $ ] , and @xmath67 $ ] . for the illustration
, we show the potential form along @xmath68 $ ] , @xmath69 $ ] , and @xmath70 $ ] directions in fig.[fig - pot ] .
it is clear that an off - center potential local minimum emerges along [ 111 ] direction as @xmath0 increases .
the energy of the first excited state from the ground state @xmath3 is small , owing to smallness of the quantum tunneling probability between different valleys .
these four states form @xmath71 orbitals and their energy can be well described by considering the quantum tunneling of the ion between the four potential minimum positions . hereafter
, we call these states in this parameter region ( @xmath63 ) as _ quantum tunneling states_. for @xmath55 and @xmath56 .
@xmath11 , @xmath58 , @xmath72 , and @xmath73 denote the approximate orbital symmetry , which is exact in the isotropic harmonic potential case.,scaledwidth=44.0% ] , [ 110 ] , and [ 111 ] directions for various @xmath0 . @xmath55 and @xmath56 .
@xmath74 , where @xmath75 being the unit vector along each direction.,scaledwidth=42.0% ] we now investigate the temperature dependence of the oscillation amplitude of these anharmonic ions .
thermodynamic average of observable @xmath12 is calculated by the formula : @xmath76 where @xmath77 is the eigen state of hamiltonian ( [ eqh ] ) with eigenenergy @xmath78 and @xmath79 is its boltzmann weight @xmath80 , and @xmath81 is the temperature . in the summation in eq .
( [ aav ] ) , we discard states with @xmath82 , since the high - energy parts of our energy spectrum is not correct due to the cutoff @xmath83 and the high - energy part does not matter as long as the low - temperature properties of the system are concerned . in fig [ fig - x2xyz ]
, we show the temperature dependence of fluctuations @xmath84 and @xmath85 .
note that since @xmath19 is invariant quantity in tetrahedral symmetry , this does not vanish except for @xmath86 . as the temperature decreases , both @xmath87 and @xmath88 decrease and saturate to finite values for @xmath89 . on the other hand , for @xmath63
, these quantities increase at low temperatures suggesting the quantum tunneling state . and ( b ) @xmath6 . @xmath55 and @xmath56 . , scaledwidth=40.0% ] to see the change with @xmath0 more clearly
, we calculate the ion density @xmath90 at position @xmath57 . here
@xmath91 is the wave function at position @xmath92 which is normalized as @xmath93 . in figs .
[ fig - dens](a)-[fig - dens](d ) , we show @xmath94 along [ 111 ] direction , @xmath95 , for several temperatures .
@xmath96 in other directions has monotonic temperature dependence and is peaked at @xmath97 as in fig .
[ fig - dens](a ) . as @xmath0 increases , a part of the weight of @xmath94 shifts to the position around @xmath98 .
interestingly near @xmath62 [ fig . [ fig - dens](c ) ] , the temperature dependence of @xmath94 at @xmath98 , is non - monotonic , _
i.e. , _ as @xmath81 decreases , @xmath99 increases first , but decreases below @xmath100 . in fig .
[ fig - dens](d ) , the @xmath101 is suppressed at low temperatures reflecting the nature of the quantum tunneling states , since the potential minimum away from the origin is deep enough as shown in fig .
[ fig - pot ] .
along @xmath95 for various @xmath0 .
@xmath55 and @xmath56.,scaledwidth=35.0% ] now we discuss the debye - waller factor , which contains information about the amplitude of ion oscillation and it is observed by elastic neutron - scattering experiments.@xcite consider an atom whose position is represented by @xmath102 in a unit cell and denote its displacement by @xmath103 .
then , the debye - waller factor at the scattering wave vector @xmath104 for this atom @xmath105 is given by @xmath106= \langle \exp(-i{\bf q \cdot u_d})\rangle .
\label{dw0}\end{aligned}\ ] ] in the harmonic approximation , eq .
( [ dw0 ] ) reduces to @xmath107 using @xmath108 , the static structure factor @xmath109 is represented as @xmath110,\end{aligned}\ ] ] where @xmath111 is the averaged scattering length of the atom at position @xmath112 . when anharmonicity is not negligible , eq .
( [ dw1 ] ) is not sufficient and we must take into account higher - order terms . up to the third order in @xmath113 , we obtain @xmath114 as an example , let us consider k atoms in kos@xmath1o@xmath2 , which constitute a diamond sub - lattice .
there are two sites in the unit cell , @xmath115 and @xmath116 . here
@xmath117 is the lattice constant . from the symmetry arguments , we obtain @xmath118 in the diamond lattice structure , @xmath119 and @xmath120 .
using eqs .
( [ w1 ] ) and ( [ w2 ] ) and setting @xmath121 , we obtain @xmath122 \bigg\}.\label{eq - fq}\end{aligned}\ ] ] thus , the effective @xmath123 becomes @xmath124 \big|.\nonumber\\\end{aligned}\ ] ] it is important to note that since the @xmath125 depends on the wave vector as @xmath126 , the effect of anharmonicity is anisotropic .
for example , at @xmath127 , @xmath128 is given by @xmath129 \bigg\},\nonumber\\ \label{w111}\end{aligned}\ ] ] whereas at @xmath130 , we obtain @xmath131 the third - order contribution @xmath132 also modifies the extinction rule of the structure factor . according to eq .
( [ eq - fq ] ) , @xmath133 at @xmath134 and also at @xmath135 due to the interference factor @xmath136 $ ] when @xmath125 is set to be zero . for @xmath137 ,
@xmath132 is finite and this leads to nonvanishing @xmath125 at @xmath138 but @xmath139 at @xmath140 , since @xmath125 is proportional to @xmath126 .
therefore , @xmath141 is nonvanishing at @xmath138 while remains zero at @xmath134 .
we show the calculated @xmath142 for several values of @xmath0 as a function of temperature in fig .
[ fig - debyewaller ] . in fig .
[ fig - debyewaller](a ) , the first term in eq .
( [ w111 ] ) is shown and @xmath142 for @xmath143 and @xmath144 are shown in ( b ) and ( c ) , respectively .
it is easily seen that , for small @xmath0 , @xmath142 in figs .
[ fig - debyewaller](b ) and [ fig - debyewaller](c ) are essentially the same as @xmath145 in ( a ) , while , for larger @xmath0 , @xmath142 is enhanced ( suppressed ) for @xmath143 in ( b ) [ @xmath144 in ( c ) ] . for other directions of @xmath146 with @xmath147 , @xmath148 does not include the contribution of the third order term @xmath149 and thus , @xmath150 .
figure [ fig - dwqdep ] shows the wave - vector dependence of @xmath151 for several temperatures and @xmath0 .
note that @xmath152 at @xmath153 vanishes due to the extinction rule mentioned above , while the intensity at @xmath138 appears for nonzero @xmath0 although its strength is weak .
concerning @xmath0 dependence , @xmath141 for @xmath154 is much suppressed for large @xmath155 and low temperatures , which is originated from the large displacement for @xmath63 . as for the temperature dependence , @xmath141 increases as @xmath81 decreases for @xmath86 and @xmath156 , while decreases for @xmath157 .
this temperature dependence of @xmath141 for @xmath63 is also understood from the temperature dependence of @xmath128 in fig .
[ fig - debyewaller ] , where @xmath128 for larger @xmath63 increases as @xmath81 decreases . we note that by careful analysis of the neutron data , as we have demonstrated the effects of the third - order term on the debye - waller factor , we can obtain useful information about anharmonicity in the real materials . in kos@xmath1o@xmath2 , since there are a lot of kinds of atoms such as the oxygen and osmium , our results presented in this section can not directly be compared with the neutron data of kos@xmath1o@xmath2 .
it is desired to carry out high resolution neutron scattering experiments using the single crystal and analyze the results by taking into account the anharmonicity . and @xmath56 .
( a ) @xmath158 .
( b ) @xmath159 for @xmath143 .
( c ) @xmath159 for @xmath144 .
inset in ( c ) : schematic potential form along the [ 111 ] direction in ( b ) and ( c ) for large @xmath0 cases.,scaledwidth=40.0% ] dependence of @xmath141 for @xmath55 and @xmath56 at temperatures ( a ) @xmath160 , ( b ) @xmath161 , and ( c ) @xmath162 . for each of @xmath104 ,
@xmath141 values for @xmath86 , @xmath156 , and @xmath163 are shown from left to right . for @xmath164
, @xmath141 is always zero.,scaledwidth=50.0% ] now let us investigate the dynamics of the anharmonic ion oscillation . in fig .
[ fig - spectralfunc ] , we show the phonon density of states ( dos ) @xmath165 , which is calculated from the correlation function @xmath166 , @xmath167 the dos is given by its imaginary part , @xmath168 .
notice @xmath169 in the tetrahedral symmetry .
we set the phenomenological broadening parameter @xmath170 .
although one can try more sophisticated analyses by introducing dissipation by electron - phonon couplings or couplings with other degrees of freedom to calculate the imaginary part of the self - energy @xmath171,@xcite it is sufficient at this stage to restrict ourselves in the phenomenological level as long as characteristic properties of the dos are concerned . for various @xmath0 s fixing @xmath55 and @xmath56 .
the data correspond to @xmath17210 , 9.2 , 8.4 , 7.6 , 6.8 , 6.0 , 5.2 , 4.4 , 3.6 , 2.8 , 2.0 , 1.8 , 1.6 , 1.4 , 1.2 , 1.0 , 0.8 , 0.6 , 0.4 , 0.2 , and 0.1 , from bottom to top and each curve is shifted up by 0.005 .
energy eigenvalues relative to the energy of the ground state are shown by short bars in each panel.,scaledwidth=48.0% ] figure [ fig - spectralfunc](a ) shows the phonon dos @xmath165 for the case @xmath86 and the eigenenergies relative to the ground state energy indicated by small bars .
it is clearly seen that the position of the peak decreases as the temperature decreases and this softening is consistent with the study based on self - consistent gaussian approximations.@xcite the single peak in @xmath165 is actually constituted of many lorentzian peaks each of which has a width of @xmath170 . in this case of @xmath86 , the transition matrix elements are sizable only between neighboring `` multiplets '' of the corresponding harmonic system . because of the fourth - order anharmonicity , @xmath37 and @xmath38 , the degeneracy of each multiplet is only approximate and slightly lifted , and also the energy separation between multiplets increases with increasing energy .
since the distribution of lorentzian peak positions is quite smooth and the energy dependence of boltzmann weight is monotonic , they finally form a single broad peak .
its position shifts from a low energy at low @xmath81 where only the ground state is an important initial state to a higher energy at higher @xmath81 where higher excited states contribute more importantly . for @xmath143
, qualitatively different behaviors appear as shown in figs .
[ fig - spectralfunc](b)-[fig - spectralfunc](f ) .
the transition matrix elements are now finite for many more pairs of eigenstates @xmath173 for nonzero @xmath0 .
this is because the point - group symmetry changes from cubic @xmath174 to tetrahedral @xmath20 and the eigenstates @xmath77 do not have definite parity related to inversion symmetry @xmath175 . in this point - group symmetry , for example , @xmath176 and @xmath177 are bases of the same @xmath178 irreducible representation . as a result ,
the matrix element such as @xmath179 is nonvanishing ( here `` @xmath11 '' means an isotropic @xmath11-wave state , _
@xmath180 in @xmath20 point group ) .
in particular , the ground state has transition matrix elements much larger than in the case of @xmath86 .
thus , even at the lowest temperature , more than one peaks are present in @xmath165 for @xmath143 , while the @xmath86 case has no additional visible high - energy peak in @xmath165 as shown in fig .
[ fig - spectralfunc](a ) , since the ground state has nonvanishing matrix elements only with the odd - parity excited states and their magnitude is small . for clarity ,
let us concentrate on a few specific peaks to understand their origin for @xmath137 .
@xmath165 for the lowest temperature @xmath181 in fig .
[ fig - spectralfunc](c ) shows four peaks below @xmath182 . as it is easily checked by comparing the energy eigenvalues indicated by bars in fig .
[ fig - spectralfunc ] , the positions of these peaks correspond to the excitation energies . actually , the lowest three peaks in fig .
[ fig - spectralfunc ] ( c ) correspond to the transitions to the @xmath183 states which can be traced back to the @xmath58-wave sates in the one - phonon multiplet , the @xmath72-wave states in the two - phonon multiplet , and the @xmath58- or @xmath73-wave states in the three - phonon multiplet in the harmonic oscillator ( see fig .
[ fig - spec ] where the label of the multiplets in the harmonic oscillator are indicated ) .
there are also peaks or shoulder - like structures originating from the transition between the @xmath59 and @xmath183 states for @xmath143 . among them
, the transition between the first excited @xmath58-wave states to @xmath59 state is clearly seen in figs .
[ fig - spectralfunc](b ) and [ fig - spectralfunc](c ) .
since the energy difference between these excited states @xmath184 is smaller than that between the ground state and the first excited state @xmath3 , these shoulder - like structures appear at energies lower than the main peak position which corresponds to the transition from the ground state and the first excited states for the intermediate temperatures . in fig .
[ fig - spectralfunc](d ) , @xmath184 is so small that the structure at @xmath185 is not clearly visible but the slope @xmath186 near @xmath187 is enhanced . in figs .
[ fig - spectralfunc](e ) and [ fig - spectralfunc](f ) , the first excited states are almost degenerate with the ground state , so that the lowest energy peak in the figures does not correspond to the transition to the first excited states but that between the first excited states and @xmath59 state .
in this section , we will discuss the strong coupling theory of superconductivity@xcite in which the attractive force between electrons is mediated by the anharmonic phonons discussed in the previous section .
the main issue of this section is the effect of the phonon anharmonicity on the superconducting transition temperature @xmath8 .
we assume that the phonon dos is given by @xmath165 obtained in sec .
[ correlationfunc ] , _
i.e. , _ we ignore the renormalization due to the electron - phonon coupling .
generally , electron - phonon couplings lead to frequency renormalization and a finite life time of phonons , both of which are described in the renormalization of @xmath165 .
this renormalization is interesting and generally should be included , but the present calculation reproduces correct qualitative behaviors and we leave full self - consistent calculation as a future problem .
it is noted that the frequency renormalization is taken into account such that the potential parameters are chosen to reproduce a renormalized frequency .
the superconducting transition temperature @xmath8 is obtained by applying the conventional strong coupling theory of @xmath11-wave pairing@xcite to our anharmonic phonon system coupled to isotropic electron gas . throughout this section ,
we use matsubara formalism , which is efficient to determine @xmath8 .
matsubara formulation has also advantage that the phenomenological broadening factor @xmath188 used in sec .
[ correlationfunc ] is not necessary . following the conventional theory of strong coupling superconductivity,@xcite the transition temperature @xmath8 for isotropic s - wave gap @xmath189
is determined by the gap equation , @xmath190 where the electron green s function is given by @xmath191 and the normal self - energy is given by @xmath192 here @xmath193 is the phonon green s function . @xmath194 and @xmath195 are the fermionic and bosonic matsubara frequencies , respectively .
@xmath196 is proportional to the square of the electron - phonon coupling constant times electron dos at fermi energy . in our model , it is a frequency - independent quantity and is set as @xmath197 k / @xmath198 , which corresponds to @xmath199 in our units of energy and length . as in the conventional theory@xcite , the normal self - energy , eq .
( [ sig ] ) is essentially given by the second - order perturbation theory and @xmath200 is analytically given by@xcite @xmath201\big|^{-1}.\end{aligned}\ ] ] to determine @xmath8 , we define @xmath202 and solve eigenvalue problem numerically @xmath203 with decreasing temperature , @xmath204 increases and @xmath8 is given by the condition : @xmath205 . before discussing the effects of phonon anharmonicity on @xmath8 ,
we show in fig .
[ fig - tscharmo ] @xmath8 in the harmonic system as a function of the first excited state energy @xmath206 .
the dash line is @xmath8 calculated by mcmillan s formula@xcite , which is valid in the weak - coupling regime , while the solid line is that by allen and dynes,@xcite which gives a good estimation of @xmath8 in the extremely strong coupling regime .
let us summarize two analytic formulae for the two opposite limits .
the mcmillan s formula is given by @xmath207 where the coulomb pseudo - potential is not taken into account . here ,
the dimensionless coupling constant @xmath4 is given by @xmath208 and the characteristic energy scale @xmath209 is given by @xmath210 the allen - dynes formula is given by @xmath211 where @xmath212 . in the case of harmonic phonons ,
@xmath4 can be analytically calculated and is given by @xmath213.@xcite this means the smaller @xmath3 is , the larger @xmath4 is realized . the mcmillan s formula ( [ mcmillan ] )
indeed reproduces @xmath8 for large @xmath3 as expected , while for small @xmath3 , @xmath8 approaches the value of the allen - dynes formula ( [ allen ] ) ( @xmath214 ) as shown in fig .
[ fig - tscharmo ] . note that for the harmonic phonons , @xmath215 .
this relation is valid only for harmonic oscillators as will be discussed below .
vs the phonon energy @xmath3 for the harmonic case @xmath43 .
the dashed line corresponds to @xmath216 given by eq .
( [ mcmillan ] ) and the solid line represents @xmath217 given by eq .
( [ allen]),scaledwidth=44.0% ] now , we discuss the effects of anharmonicity of phonon dynamics on @xmath8 .
figure [ fig - tsc ] shows the dependence of @xmath8 on the third - order anharmonic term @xmath0 of the ion potential , eq .
( [ h ] ) for @xmath218 , @xmath219 , and @xmath220 , at fixed @xmath55 .
all the three cases exhibit a pronounced peak around the crossover value @xmath62 as discussed in sec .
[ anharmo ] for each @xmath38 , and @xmath8 s are strongly suppressed in the quantum tunneling states , _
i.e. , _ for large @xmath63 .
the solid lines in the figure represent the mcmillan s formula @xmath221 [ eq . ( [ mcmillan ] ) ] .
the agreement with the mcmillan s formula indicates that @xmath8 is qualitatively given by the conventional theory of strong coupling superconductivity .
note that @xmath4 and @xmath209 vary with the temperature in the presence of anharmonicity and @xmath222 are calculated by using these values evaluated at @xmath223 . for @xmath224
, however , the discrepancy becomes large as shown in the inset of fig .
[ fig - tsc ] , and the effects of anharmonicity and the nature of the extreme strong coupling regime appear , which will be discussed later in sec .
[ exe ] .
vs @xmath0 for @xmath55 and @xmath2250 , 0.01 , and 0.03 .
the solid lines show @xmath8 s of mcmillan s formula ( [ mcmillan ] ) .
@xmath226 s for each of @xmath38 are indicated by arrows .
inset : zoom up in the large @xmath0 part.,scaledwidth=44.0% ] in sec .
[ coupl ] , we discuss the origin of the peak structure based on the mcmillan formula.@xcite the mcmillan s formula is expressed with the two parameters , @xmath4 and @xmath209 , as shown in eq .
( [ mcmillan ] ) . in fig .
[ fig - lam ] , we show the @xmath0 dependence of @xmath4 and @xmath209 calculated at @xmath8 together with @xmath3 .
it is noticeable that @xmath4 shows a steep increase above @xmath226 and @xmath209 is almost the same as @xmath3 for @xmath227 . from these facts
, one can easily understand that @xmath8 calculated by eq .
( [ mcmillan ] ) decreases in the quantum tunneling states . the peak structure in @xmath8 in fig .
[ fig - tsc ] is a result of the competition between the suppression of @xmath228 and the enhancement of @xmath4 .
thus , the peak structure is realized at @xmath229 , where @xmath4 is large enough and simultaneously @xmath209 is not vanishingly small . , @xmath209 , and @xmath3 vs @xmath0 for @xmath218 , @xmath219 , and @xmath220 .
the open symbols represent @xmath4 and the filled symbols with a solid line represent @xmath209 . lines without symbols represent @xmath3 .
@xmath4 and @xmath209 are calculated at @xmath223.,scaledwidth=44.0% ] in an extreme strong coupling regime @xmath230 , allen and dynes showed that @xmath8 is approximately given by eq .
( [ allen ] ) rather than the mcmillan s formula ( [ mcmillan]).@xcite when eq .
( [ allen ] ) is applied to our local phonon problem , we need care about the quantity @xmath231 . as pointed out by hardy and flocken,@xcite this quantity
is related to the f - sum rule and this implies @xmath8 is universal in any potential and depends only on the mass of the ion and @xmath196 as shown in fig .
[ fig - tscharmo ] .
however , this is not true in our anharmonic phonon system . from numerical calculations , instead of @xmath232 , we find that @xmath8 for @xmath233 is approximately given by @xmath234 we note that the origin of large @xmath4 for @xmath63 is the strong suppression of @xmath3 for @xmath63 and other high - energy states do not play an important role .
thus , @xmath235 overestimate the energy scale of @xmath8 .
these facts support the use of @xmath3 instead of the averaged frequency @xmath235 .
, @xmath184 vs @xmath3 for @xmath56 and five sets of @xmath39 and @xmath37 .
the open ( filled ) symbols represent @xmath184 ( @xmath4 ) .
@xmath3 is controlled by varying @xmath0.,scaledwidth=44.0% ] and @xmath236 vs @xmath3 for @xmath56 and five sets of @xmath39 and @xmath37 .
the open ( filled ) symbols represent @xmath4 ( @xmath236 ) .
, scaledwidth=44.0% ] as we noted before , there is a crossover at @xmath62 to the quantum tunneling states and this means there are five states at low - energy region of the spectra as shown in fig .
[ fig - spec ] . in the harmonic case
, @xmath4 is basically determined by the matrix element @xmath237 between the ground state @xmath238 and the threefold - degenerate first excited states , and the corresponding excitation energy @xmath3 . since
four low - energy excited states are nearly degenerate with the ground state around @xmath62 , the singlet excited state @xmath239 also has a noticeable matrix element @xmath240 .
this provides another channel of attractive interaction and contributes to @xmath8 .
it is important to note that this additional contribution never appears in the one - dimensional model and is a direct consequence of the crossover of the ground state at @xmath62 . in fig .
[ fig - lambda - es - delta ] , we show @xmath4 and the energy difference between @xmath59 state and the @xmath58-wave like first excited states : @xmath241 . they are calculated with varying @xmath0 , and plotted as a function of the energy difference @xmath3 between the first excited state and the ground state .
@xmath4 increases roughly as @xmath242 for small @xmath3 .
@xmath184 shows minimum around @xmath243 and this reflects the anti - crossing in the energy spectra . corresponding to this
, the coupling constant @xmath236 related to @xmath59 state shows a maximum around @xmath244 as shown in fig .
[ fig - lam_i ] . here
, we define @xmath236 as @xmath245 where the @xmath246 summation is taken over the three first excited states .
note that the smaller @xmath184 is , the larger @xmath236 is obtained . depending on the magnitude of @xmath236 , total @xmath4 has a small bump around the @xmath244 as a function of @xmath3 .
this observation clearly shows that there is a new channel of interaction via the excitations between the first excited states and the @xmath59 state . as we discussed in sec .
[ additional ] , the enhancement of @xmath236 influences the total @xmath4 and also affects @xmath8 itself . in sec .
[ devi ] , we derive a formula which includes the contribution of @xmath236 . in order to take into account this contribution in eq .
( [ allen2 ] ) , we assume the dos given by @xmath247 .
\label{f12}\end{aligned}\ ] ] since the enhancement of @xmath236 takes place in a finite but extremely strong coupling regime , the total @xmath4 is large and we follow the discussion by allen and dynes to estimate the lower bound of @xmath8 , by assuming a trial gap function @xmath248.@xcite using the form ( [ f12 ] ) , we obtain the lower bound of @xmath8 as @xmath249 ^ 2 + 4(\lambda_p+\lambda_{s'})\delta^2{\delta'}^2},\nonumber\\ \label{solution0}\end{aligned}\ ] ] in the limit of @xmath250 eq .
( [ solution0 ] ) reduces to eq .
( [ allen2 ] ) : @xmath251 .
although the more sophisticated gap function can give the almost correct coefficient @xmath252 instead of @xmath253,@xcite it is enough to consider the simplest trial gap function in order to examine the contributions of the @xmath59 state to @xmath8 . in eq .
( [ solution0 ] ) , we assume @xmath254 corresponds to the coupling constant originating from the transition between the ground state and @xmath58-wave like states ( first excited states ) and @xmath255 is the part from @xmath59 state [ eq .
( [ lamdas ] ) ] . since @xmath256
, we can obtain a simpler approximate form of eq .
( [ solution0 ] ) .
this is obtained by simply replacing @xmath4 in eq .
( [ allen2 ] ) by @xmath257 , @xmath258 .
\label{solution2}\end{aligned}\ ] ] figure [ fig - scaling ] shows the @xmath4 dependence of @xmath259 for two typical sets of parameters in log - log scale .
as shown with the dashed line , the allen - dynes formula shows a straight line . for small @xmath4 ,
the calculated values agree with the mcmillan s formula , while for extremely large @xmath4 they approach the line of the allen - dynes formula . for the intermediate @xmath260 ,
however , the results show a bending and this bending is larger in fig .
[ fig - scaling](b ) than in fig .
[ fig - scaling](a ) . our two - channel formula ( [ solution0 ] )
can describe the bending as indicated by the solid line .
although the lower bound of @xmath8 , eq . ( [ solution0 ] ) is not in very good agreement with the calculated @xmath8 , it captures the overall behavior of @xmath8 for @xmath260 .
it is also noted that , although the mcmillan s formula ( [ mcmillan ] ) shows an increase in @xmath261 for large @xmath4 , the increase is too steep .
the present two - component analysis confirms the existence of the additional channel of the interaction near the crossover point in this system
. formulas ( [ mcmillan ] ) , ( [ allen ] ) , ( [ solution0 ] ) , and ( [ solution2 ] ) vs @xmath4 . formula ( [ mcmillan ] ) with @xmath209 replaced by @xmath3 is also plotted for comparison .
( a ) @xmath262 , @xmath55 , and @xmath56 .
( b ) @xmath40 , @xmath263 , and @xmath56.,scaledwidth=50.0% ]
let us now discuss the implications of our calculations to compare with characteristic properties experimentally observed in @xmath7-pyrochlore compounds .
one point is the chemical trends among the three member compounds , and we discuss why the potassium compound has the highest @xmath8 of superconductivity .
another point is about the question why the superconducting phase is not so much affected by the isomorphic structure transition at @xmath17 in the phase diagram .
we will also propose a possible change in the k - oscillation profile at @xmath17 and the effects of the transition on superconductivity . among the three @xmath7-pyrochlore compounds @xmath12os@xmath1o@xmath2 ( @xmath12=k , rb or cs ) , the k compound has the highest @xmath264=9.6 k of superconductivity and the strongest anharmonicity in the @xmath12-cation oscillation dynamics as observed in the neutron - scattering experiments : the debye - waller factor of k ion is much smaller than of rb and cs compounds@xcite and the softening of the low - energy phonon peak is also the strongest in kos@xmath1o@xmath2.@xcite the rb compound has the next strongest anharmonicity and the second highest @xmath265=6.3 k , while the cs compound has the weakest anharmonicity and the lowest @xmath266=3.3 k. the ratio of @xmath8 is approximately 3:2:1 .
thus , the anharmonicity in the ion dynamics and the value of @xmath8 are related .
let us examine this point in our results and also check if one can explain , at least qualitatively , the trends of other important quantities , electron mass enhancement and phonon energy .
a crucial difference among the three compounds is the size of @xmath12-cation ; the k ion has the smallest size , followed by rb , and cs is the biggest ion .
since the size of the surrounding os@xmath10o@xmath13 cage essentially does not change , the k ion has the largest space inside the cage,@xcite leading to strongly anharmonic oscillations .
this reflects in different shapes of the @xmath12-cation potential , as shown in the calculation by kune , _
et al._,@xcite although the point - group symmetry is common . in our theory ,
the compound - dependent potential shapes are modeled by adjusting parameters in the potential , eq .
( [ eq : pot ] ) or ( [ eq : potential ] ) . in the following , we will show that @xmath8 values and the ion dynamics in the three compounds are naturally explained by appropriate choice of potential parameters . throughout sec .
[ discusschemi ] , we will use some parameters explicitly shown with physical dimensions if necessary .
let us first consider the trend of the value @xmath267 among the three compounds , recalling the anharmonicity appearing in the debye - waller factor and the phonon energy observed in the neutron experiments .
@xcite as shown in fig .
[ fig - dens ] , the anharmonicity grows with @xmath267 .
this implies that the @xmath267 value is the smallest for cs , then for rb , and the largest for k. note that the largest ratio is still smaller than or at most equal to the crossover value @xmath268 , since the observed k - cation density profile does not change its peak position from the equilibrium position.@xcite secondly , we consider the trend of the second - order potential parameter @xmath39 .
figure [ fig - varomega0 ] shows the @xmath0 dependence of @xmath8 for several values of @xmath39 .
the first - principle calculation of the ion potential clearly shows that @xmath39 is very small for k while larger for rb and the largest for cs.@xcite this trend is consistent with the fact that the k compound has the strongest anharmonicity and the highest @xmath8 as far as @xmath89 , as shown in fig .
[ fig - varomega0 ] .
it is also important to note that the peak position @xmath269 becomes smaller as @xmath39 decreases .
vs @xmath0 for five values of @xmath39 fixing @xmath55 and @xmath56.,scaledwidth=50.0% ] by keeping these variations in @xmath267 and @xmath39 in mind , let us discuss our results of @xmath3 , @xmath4 at @xmath223 and @xmath8 itself for each of the three compounds . in order to make discussion simple , we assume the electron phonon coupling and the electron density of states are the same for the three compounds and @xmath270 k / @xmath198 . the atomic mass of each @xmath12-cation is @xmath271 , @xmath272 , and @xmath273 , respectively . we have adjusted potential parameters , eq .
( [ eq : pot ] ) , for each compound to reproduce the phonon energy and the effective ion oscillation variance @xmath274 determined from the debye - waller factor.@xcite as shown in sec .
[ debye - waller ] , @xmath275 generally contains the contribution of the third - order fluctuations @xmath276 in addition to @xmath277 .
however , since the experimental data is the @xmath104-averaged value of @xmath278 , it is sufficient to consider the second - order fluctuations @xmath279 . since our potential , eq .
( [ eq : pot ] ) or ( [ eq : potential ] ) , includes four parameters , we fix @xmath280 k / @xmath281 and @xmath282 k / @xmath283 for simplicity and vary @xmath26 . it is noted that @xmath26 is expected to become smaller as the size of the alkali cation decreases , and , for fixed @xmath27 , @xmath267 becomes smaller as @xmath26 increases . table [ tbl-1 ] shows the list of the basic quantities , @xmath3 , @xmath4 , and @xmath8 for the three compounds .
let us first discuss @xmath8 .
it is quite sure that one needs also to include a high - energy phonon , which plays a role to increase the @xmath8 by about 3 - 5 k for all the three compounds , to obtain @xmath8 consistent with the experimental results .
@xmath284 is one calculated with including an additional phonon with energy @xmath285 k. we have set the corresponding dimensionless coupling constant @xmath286 in the calculations .
the results of @xmath284 are quantitatively consistent with the experimental values @xmath287 k , @xmath288 k , and @xmath289 k. here , we can reproduce the experimental values of @xmath8 by assuming the same @xmath196 , @xmath290 and @xmath291 for all the three compounds and we conclude that the origin of the difference in @xmath8 for the three @xmath7-pyrochlore compounds is due to the difference mainly in the anharmonicity of the alkali cation oscillations.@xcite now , let us discuss the electron effective mass . in our theory , the electron mass enhancement factor is related to @xmath292 as @xmath293 .
@xmath4 s obtained from experiments and previous theoretical studies are @xmath2941.6 - 2.4 , @xcite for k , @xmath2951.0 - 1.3 , @xcite for rb and @xmath296,@xcite for cs , respectively .
our estimation of @xmath297 with @xmath4 in table [ tbl-1 ] is qualitatively consistent with these values , although we have not fit @xmath8 for each compound .
the experimental values of specific - heat coefficient @xmath298 is @xmath299 , @xmath300 , and @xmath301 mj / mol k@xmath198 while the band calculations predicted @xmath302 mj / mol k@xmath198 for all the three compounds.@xcite thus , the experimental mass enhancement factor @xmath303 in kos@xmath1o@xmath2 is 1.5 - 1.7 times larger than in rbos@xmath1o@xmath2 and csos@xmath1o@xmath2@xcite . in our calculation , @xmath304 and @xmath305 , which are semi - quantitatively consistent with the experimental values . from our results ,
the enhanced mass enhancement in kos@xmath1o@xmath2 is attributed to the proximity to the crossover to the quantum tunneling state and also the large oscillation amplitude , which effectively enhances the electron - phonon coupling . in order to obtain the correct value of the mass enhancement ( @xmath306 for kos@xmath1o@xmath2 and @xmath307 for the other two )
, we should also take into account the electron - electron interactions , but this is beyond the scope of the present study and we leave it for a future problem .
nevertheless , it is important to note that @xmath4 is significantly enhanced near @xmath62 by anharmonic oscillations and that this is the central reason why the mass enhancement in kos@xmath1o@xmath2 is much larger than in rb and cs compounds . as for the phonon frequency @xmath3 ,
optical modes related to the @xmath12-cation oscillations are observed at energy around 5 - 7 mev at room temperature in the inelastic neutron scattering@xcite and their energies show strong softening as the temperature decreases .
this corresponds to our results shown in fig .
[ fig - spectralfunc ] . for kos@xmath1o@xmath2 , the phonon energy at @xmath308 k is about 3.1 mev , which is quite small compared with the value of 5.5 mev at @xmath309 k.@xcite furthermore , the low - energy phonon peaks in kos@xmath1o@xmath2 are rather broad compared with rb and cs compounds,@xcite which suggests the proximity to the crossover point .
the specific - heat experiment@xcite and the photo - emission spectroscopy data@xcite also support these results . in our calculation for kos@xmath1o@xmath2 with parameters in table [ tbl-1 ] ,
the peak position in the phonon spectrum is about 70 k @xmath310 6 mev at @xmath309 k and this shifts to @xmath311 k as @xmath81 decreases .
this means the potential parameters in table [ tbl-1 ] can reproduce not only @xmath8 and @xmath3 but also the temperature dependence of the spectra . as for the chemical trends , we have calculated superconducting transition temperature for the three @xmath7-pyrochlore compounds adjusting parameters in the potential , eq .
( [ eq : pot ] ) , and we can reproduce the phonon energy and amplitude related to the average debye - waller factor . we have obtained quantitatively consistent values of @xmath8 with the experimental ones , using the same electron - phonon coupling , the electron dos and also the same additional high - energy phonon for all the three compounds . .local potential parameters and @xmath312 at @xmath313 , @xmath4 at @xmath223 , @xmath3 , @xmath8 , and @xmath284 for the three @xmath7-pyrochlore compounds .
the other parameters are @xmath314 , @xmath315 k / @xmath283 , and @xmath316 . [ cols="^,^,^,^,^,^,^,^,^",options="header " , ] [ tbl-1 ] as mentioned in sec .
[ intro ] , kos@xmath1o@xmath2 exhibits an isomorphic first - order transition at @xmath16 k.@xcite in sec .
[ tptran ] , let us investigate the effects of this transition on the phonon dynamics , using a set of constraints offered by the experimental results .
we also discuss its effects on superconductivity .
electric resistivity shows a concave temperature dependence at high temperatures , and this is attributed to the strong coupling to phonons with strong anharmonicity.@xcite at the magnetic field @xmath317 t , the resistivity is suddenly suppressed by 25 % at @xmath17 , and shows a different @xmath81-dependence proportional to @xmath318 at @xmath319.@xcite this indicates that the electron - phonon scattering processes are reduced significantly at @xmath17@xcite .
this is also consistent with the reduction in the specific - heat jump at @xmath8 in magnetic fields above 8 t.@xcite in the magnetic field - temperature phase diagram , @xmath17 remains essentially insensitive to magnetic field @xmath320 as shown in fig .
[ fig - schematichc2 ] .
the upper critical magnetic field @xmath18 is suppressed below @xmath17 , but extrapolating to the region above @xmath17 , it seems that it also vanishes at the position very close to @xmath321 . and @xmath17 , respectively .
the dashed line indicates the extrapolation of low-@xmath81 @xmath18 to the higher - temperature region.,scaledwidth=44.0% ] this fact can not be explained if the characteristic energy scale , @xmath209 or @xmath3 is common across @xmath17 : the extrapolated @xmath18 line should vanish below @xmath321 since @xmath4 is reduced below @xmath17 as discussed above .
therefore , the above fact implies the enhancement of the characteristic energy scale @xmath3 or @xmath209 below @xmath17 .
indeed , chang , _
et al._,@xcite assumed a slightly increased einstein energy to fit the specific heat data and our previous study also predicted the increase in the oscillation energy.@xcite let us discuss the changes in @xmath4 , @xmath3 , @xmath322 , and @xmath323 at @xmath17 based on our results , modeling the isomorphic transition by a sudden change in the two potential parameters , @xmath39 and @xmath0 .
this change corresponds to the variation in the mean field part of inter - site ion interactions , which was discussed in ref .
34 and also changes in the oxygen positions and the lattice constant.@xcite two parameters are chosen under the constraints ( i ) two different parameter sets above and below @xmath17 lead to the same @xmath8 , and ( ii ) @xmath4 is smaller below @xmath17 , both of which are the experimental constraints , and we also assume ( iii ) no significant change in the electron band structure across the transition and thus @xmath196 is also unchanged , and ( iv ) @xmath89 for both below and above @xmath17 implied by the result of the electron - density profile obtained by the x - ray and the neutron experiments @xcite as discussed before . in the following ,
we discuss the change across @xmath17 based on these constraints .
dependence of characteristic quantities , when @xmath8 is fixed to 0.05 by tuning @xmath0 .
( a ) @xmath4 , @xmath3 , and @xmath0 .
( b ) the second and third - order moments of ion oscillation @xmath5 , and @xmath88 .
@xmath55 and @xmath56.,scaledwidth=44.0% ] and the density map on the section with ( 110 ) plane at @xmath324 .
( a ) @xmath325 , @xmath326 and ( b ) @xmath40 , @xmath327 . ( c ) @xmath96 along [ 111 ] direction corresponding to ( a ) and ( b ) , ( d ) along [ 001 ] and ( e ) along [ 110 ] direction .
s in ( d ) and ( e ) are not shown since @xmath329 along these directions.,scaledwidth=44.0% ] in fig .
[ fig - discussion ] , characteristic quantities are shown for fixed @xmath37 and @xmath38 with varying @xmath39 and simultaneously @xmath0 such that those give the same @xmath330 . since @xmath89 which is the assumption ( iv ) , @xmath8 is well approximated by the mcmillan formula ( [ mcmillan ] ) , in which @xmath8 is determined by the two factors @xmath331 and @xmath4 .
firstly , it is noted that @xmath4 increases as @xmath39 increases .
one might expect a suppression of @xmath4 as @xmath39 increases , but the key point is that we simultaneously tune @xmath0 to fix @xmath8 unchanged and this means larger @xmath0 is necessary at larger @xmath39 .
this increase in @xmath0 overcomes the competing effect of the increase in @xmath39 , and finally @xmath4 increases .
secondly , the energy of the first excited state @xmath3 increases as @xmath39 decreases .
this is naturally understood by noting that @xmath3 becomes large for smaller @xmath0 .
thus , from the constraint ( ii ) the above results imply that the phonon energy is enhanced below @xmath17 and , indeed , this is consistent with the previous theoretical studies.@xcite most interestingly , the second- and third - order fluctuations @xmath312 and @xmath88 change oppositely as shown in fig .
[ fig - discussion](b ) .
as @xmath39 decreases , @xmath312 increases for the most of the @xmath39 range , while @xmath323 decreases monotonically .
this is because small @xmath39 corresponds to small @xmath0 as explained above .
these results indicate that @xmath312 is slightly enhanced below @xmath17 while @xmath323 is suppressed . in order to illustrate these anisotropic fluctuations , the density isosurface for @xmath332 and the density ( color ) map
are shown for two different parameters at @xmath324 in fig .
[ fig-3dfig ] : for ( a ) and ( b ) , @xmath325 and @xmath333 , and @xmath326 and @xmath334 , respectively , and these two sets give the same @xmath335 , and we choose the two parameter sets to emphasize the change in the oscillation profile . as expected from the values of @xmath0 , the case ( a ) has the smaller @xmath336 and the larger @xmath337 , while @xmath338 and @xmath339 for the case ( b ) .
one can see that the density isosurface shows more anisotropic character in the case ( b ) , where the `` spikes '' sticking out along four [ 111 ] directions are sharper than those in the case ( a ) .
this aspect reflects in the values of @xmath277 and @xmath6 . indeed , @xmath277 at @xmath223 is larger @xmath340 in ( a ) than @xmath341 in ( b ) , while @xmath323 at @xmath223 is smaller @xmath342 in ( a ) than @xmath343 in ( b ) .
figures [ fig-3dfig](c)-[fig-3dfig](e ) show @xmath96 along [ 111 ] , [ 001 ] , and [ 110 ] directions , respectively , for the same two data .
it is clearly seen that @xmath323 in ( b ) is larger than that in ( a ) as shown in ( c ) , and @xmath312 in ( a ) is larger than in ( b ) as shown in ( d ) and ( e ) .
these changes in the anisotropy of density distribution is a characteristic nature of the first - order isomorphic transition at @xmath17 .
this transition is isomorphic , consistent with the experiments,@xcite because the amplitude of anisotropy changes but it does not break the point group , translational , and any other symmetries .
our results suggest that @xmath312 is slightly enhanced while @xmath323 is suppressed below @xmath17 in kos@xmath1o@xmath2 .
recent high - resolution neutron - scattering experiment shows the increase in @xmath312 across @xmath17 as @xmath81 decreases,@xcite which is consistent with our result .
furthermore , although it is naively expected that an increase in @xmath312 corresponds to the enhancement of @xmath4 , this does not necessarily hold in this system , since the value of @xmath4 is sensitive to the value of @xmath0 , especially near the crossover point @xmath226 .
our results show that the value of @xmath323 plays more important role for the enhancement of @xmath4 than @xmath312 . in order to detect this change in the density distribution across @xmath17 ,
it is important to perform detailed neutron - scattering experiments and to analyze the results with taking into account the third - order term in debye - waller factor as discussed in sec .
[ debye - waller ] .
it is worthwhile to examine the @xmath104 dependence of the debye - waller factor to extract the anisotropy of the oscillations , since @xmath104-averaged debye - waller factor hinders the anisotropy and anharmonicity . finally , we comment on the possibility of re - entrant superconducting transition .
experimentally , hiroi _
et al._,@xcite observed a re - entrant superconducting transition around @xmath344 t as shown in fig .
[ fig - schematichc2 ] . this can be understood from @xmath345 in the strong coupling theory of superconductivity : @xmath346 for the clean limit.@xcite in our analysis , the @xmath4 decreases and @xmath209 increases below @xmath17 , while the @xmath8 is approximately the same for the potential parameters above and below @xmath17 .
thus , @xmath345 for the low-@xmath81 parameter set is smaller than that for the high-@xmath81 set .
therefore , we expect a @xmath347 curve similar to that shown in fig .
[ fig - schematichc2 ] .
this result does not change when we use the formula for the dirty limit.@xcite
we have investigated in the present paper anharmonic phonons and strong coupling superconductivity in @xmath7-pyrochlore compounds , @xmath12os@xmath1o@xmath2 ( @xmath12=k , rb , or cs ) .
first , we have solved the schrdinger equation of the three - dimensional anharmonic phonon in tetrahedral symmetry .
the main issue is the importance of the third - order anharmonicity @xmath348 in the ion potential allowed for this symmetry .
we have determined the energy spectrum of the anharmonic phonon as a function of the third - order anharmonicity @xmath0 .
we have found that there exists a crossover of the ground state to the quantum tunneling state for @xmath63 .
we have pointed out non - monotonic temperature dependence of ion density profile near the crossover point @xmath226 .
secondly , we have calculated the transition temperature of superconductivity @xmath8 mediated by these anharmonic phonons .
we have found that the enhancement of @xmath8 near @xmath62 .
its @xmath0 dependence can be well fitted by the mcmillan formula at @xmath227 .
for @xmath349 we have found that there exists an additional channel of pairing interaction , which turns out to originate from low - energy excited states appearing at @xmath350 .
we have analyzed its contribution and derived an approximated formula of @xmath8 in the strong coupling limit .
we have also discussed the chemical trends of @xmath8 in the @xmath7-pyrochlore family @xmath12os@xmath1o@xmath2 ( @xmath12=k , rb , or cs ) .
the main difference among the three members are different values of @xmath267 and the second - order term in the ion potential @xmath39 . by assuming the same electron - phonon coupling constant @xmath196 and the same high - energy phonon for all the three compounds , the differences in @xmath8 and the energy of the first excited phonon states @xmath3
have been quantitatively explained only by the difference in the local anharmonic potential .
finally , we have discussed the effect of the first - order transition observed in kos@xmath1o@xmath2 .
the changes in the density distribution of k cation at the first - order transition have been discussed based on the experimental data obtained so far .
especially , we have found that @xmath277 and @xmath323 change differently across @xmath17 .
our results suggest the increase in the first excited phonon energy @xmath3 and the reduction in the dimensionless electron - phonon coupling constant @xmath4 across @xmath17 . while the latter is consistent with the experimental results ,
the former has not been observed and this is our prediction for the experiments . in order to detect the increase in @xmath3 ,
it is important to carry out the high - resolution inelastic neutron scattering experiments .
we hope our results shed lights in @xmath7-pyrochlore compounds both on the strong coupling superconductivity and on the anharmonic oscillations observed .
the authors would thank t. dahm , k. ueda , j. yamaura and z. hiroi for grateful discussions .
they also acknowledge the international workshop `` _ new developments in theory of superconductivity _ ''
held at institute for solid state physics , university of tokyo , during june 22-july 10 , 2009 for providing an opportunity of discussion with other participants .
this work is supported by kakenhi ( grants no .
19052003 and no .
20740189 ) and also by the next generation super computing project , nanoscience program , from the mext of japan .
99 z. hiroi , s. yonezawa , and y. muraoka , j. phys .
. jpn . * 73 * , 1651 ( 2004 ) .
z. hiroi , s. yonezawa , j. yamaura , t. muramatsu and y. muraoka , j. phys .
jpn . * 74 * , 1682 ( 2005 ) .
z. hiroi , s. yonezawa , y. nagao and j. yamaura , phys .
b * 76 * , 014523 ( 2007 ) .
t. hasegawa , y. takasu , n. ogita , m. udagawa , j. i. yamaura , y. nagao , and z. hiroi , phys .
b * 77 * , 064303 ( 2008 ) .
t. goto , y. nemoto , k. sakai , t. yamaguchi , m. akatsu , t. yanagisawa , h. hazama , k. onuki , h. sugawara , and h. sato , phys .
b * 69 * , 180511(r ) ( 2004 ) .
k. kaneko , n. metoki , h. kimura , y. noda , t. d. matsuda , and m. kohgi , j. phys .
. jpn . * 78 * , 074710 ( 2009 ) .
h. kotegawa , h. hidaka , t. c. kobayashi , d. kikuchi , h. sugawara , and h. sato , phys .
rev . lett . * 99 * , 156408 ( 2007 ) .
b. c. sales , b. c. chakoumakos , r. jin , j. r. thompson , and d. mandrus , phys .
b * 63 * , 245113 ( 2001 ) .
i. zerec , v. keppens , m. a. mcguire , d. mandrus , b. c. sales , and p. thalmeier , phys .
. lett . * 92 * , 185502 ( 2004 ) .
h. matsumoto , t. mori , k. iwamoto , s. goshima , s. kushibiki , and n. toyota , phys .
b * 79 * , 214306 ( 2009 ) .
y. kasahara , y. shimono , t. shibauchi , y. matsuda , s. yonezawa , y. muraoka , and z. hiroi , phys .
lett . * 96 * , 247004 ( 2006 ) .
m. yoshida , k. arai , r. kaido , m. takigawa , s. yonezawa , y. muraoka , and z. hiroi , phys .
lett . * 98 * , 197002 ( 2007 ) .
y. nakai , k. ishida , h. sugawara , d. kikuchi , and h. sato , phys .
b * 77 * , 041101(r ) ( 2008 ) . y. nemoto , t. yamaguchi , t. horino , m. akatsu , t. yanagisawa , t. goto , o. suzuki , a. dnni , and t. komatsubara , phys . rev .
b * 68 * , 184109 ( 2003 ) .
i. ishii , t. fujita , i. mori , h. sugawara , m. yoshizawa , k. takegahara , and t. suzuki , j. phys .
* 78 * , 084601 ( 2009 ) .
y. nakanishi and m. yoshizawa ( private communications ) .
m. brhwiler , s. m. kazakov , j. karpinski , and b. batlogg , phys .
b * 73 * , 094518 ( 2006 ) .
i. bonalde , r. ribeiro , w. brmer - escamilla , j. yamaura , y. nagao , and z. hiroi , phys .
lett . * 98 * , 227003 ( 2007 ) .
y. shimono , t. shibauchi , y. kasahara , t. kato , k. hashimoto , y. matsuda , j. yamaura , y. nagao , and z. hiroi , phys .
* 98 * , 257004 ( 2007 ) .
y. nagao , j. yamaura , h. ogusu , y. okamoto , and z. hiroi , j. phys . soc
. jpn . * 78 * , 064702 ( 2009 ) .
j. chang , i. eremin , p. thalmeier , and p. fulde , phys .
b * 76 * , 220510(r ) ( 2007 ) .
j. chang , i. elemin , and p. thalmeier , new j. phys .
* 11 * , 055068 ( 2009 ) .
t. dahm and k. ueda , phys .
lett . * 99 * , 187003 ( 2007 ) .
a. yamakage and y. kuramoto , j. phys .
soc . jpn . *
78 * , 064602 ( 2009 ) .
k. sasai , k. hirota , y. nagao , s. yonezawa , and z. hiroi , j. phys .
soc . jpn . *
76 * , 104603 ( 2007 ) .
r. galati , c. simon , p. f. henry , and m. t. weller , phys .
b * 77 * , 104523 ( 2008 ) .
h. mutka , m. m. koza , m. r. johnson , z. hiroi , j. i. yamaura , and y. nagao , phys .
b * 78 * , 104307 ( 2008 ) .
n. m. plakida , v. l. aksenov , and s. l. drechsler , europhys . lett . * 4 * , 1309 ( 1987 ) .
j. r. hardy and j. w. flocken , phys .
lett . * 60 * , 2191 ( 1988 ) . v.
h. crespi and m. l. cohen , phys .
b. * 48 * , 398 ( 1993 ) .
g. d. mahan and j. o. sofo , phys .
b * 47 * , 8050 ( 1993 ) .
s. manalo , h. michor , g. hilscher , m. brhwiler and b. batlogg , phys .
b * 73 * , 224520 ( 2006 ) .
j. kune , t. jeong , and w. e. pickett , phys .
b * 70 * , 174510 ( 2004 ) ; j. kune and w. e. pickett , physica b * 378 * -*380 * , 898 ( 2006 )
. k. hattori and h. tsunetsugu , j. phys .
jpn * 78 * , 013603 ( 2009 ) . in those cases of @xmath351 ,
we keep the same energy unit of 44 k , instead of renormalization of energy .
this simplifies our discussions about @xmath39 dependence in later sections .
s. w. lovesey , _ theory of neutron scattering from condensed matter _
( clarendon press , oxford , 1984 ) , vol . 1 .. k. hattori and k. miyake , j. phys .
jpn . * 76 * , 094603 ( 2007 ) .
m. takechi and k. ueda , j. phys .
. jpn . * 78 * , 024604 ( 2009 ) .
g. m. eliashberg , zh .
. fiz . * 38 * , 966 ( 1960 ) [ sov .
jetp * 11 * , 696 ( 1960 ) ] .
d. j. scalapino , j. r. schrieffer , and j. w. wilkins , phys . rev . * 148 * , 263 ( 1966 ) . a. a. abrikosov , l. p. gorkov , and i. ye .
dzyaloshinskii , _ quantum field theoretical methods in satistical physics _ , 2nd ed .
( pergamon press , oxford , 1965 ) .
g. bergmann and d. rainer , z. phys . *
263 * , 59 ( 1973 ) .
w. l. mcmillan , phys .
rev . * 167 * , 331 ( 1968 ) .
p. b. allen and r. c. dynes , phys .
b * 12 * , 905 ( 1975 ) . when @xmath4 is expressed by the quantities with the dimension , @xmath352 , which is , indeed , dimensionless .
j. yamaura , s. yonezawa , y. muraoka , and z. hiroi , j. solid state chem .
* 179 * , 336 ( 2006 ) .
k. sasai , m. kofu , r. m. ibberson , k. hirota , j. yamaura , z. hiroi , and o. yamamuro , j. phys . : condens .
matter * 22 * , 015403 ( 2010 ) .
the result in ref .
33 suggests a slightly negative value of @xmath39 for kos@xmath1o@xmath2 .
however , the first excited state then has an energy too small compared with the experimental estimate from the specific heat data ( ref .
3 ) and the neutron scattering ( refs . 25 and 27 ) .
correspondingly , calculated @xmath312 is too large compared with the experimental results and its temperature dependence is qualitatively different from the experimental one ( refs . 25 and 26 ) . in this paper
, we consider the cases of @xmath353 for simplicity .
we hope that more detailed information on the ion potential is provided by the neutron scattering experiments in the future . in this point
whether the additional phonon is acoustic or optical mode does not make an essential difference . to improve estimation of @xmath8 , we need the detailed information of each compound on the electron - phonon coupling @xmath196 , @xmath291 , @xmath290 , and other parameters . z. hiroi , j. yamaura , s. yonezawa , and h. harima , physica c * 460 * -*462 * , 20 ( 2007 ) . note that this value of the mass enhancement factor includes not only the contributions of electron - phonon interactions but also those of electron - electron interactions .
t. shimojima , y. shibata , k. ishizaka , t. kiss , t. tagoshi , a. chainani , t. yokoya , x .- y .
wang , c. t. chen , s. watanabe , j. yamaura , s. yonezawa , y. muraoka , z. hiroi , t. saitoh and s. shin , phys .
lett . * 99 * , 117003 ( 2007 ) .
j. p. carbotte , rev .
phys . * 62 * , 1027 ( 1990 ) . | we investigate three - dimensional anharmonic phonons in tetrahedral symmetry and superconductivity mediated by these phonons .
three - dimensional anharmonic phonon spectra are calculated directly by solving schrdinger equation and the superconducting transition temperature is determined by using the theory of strong coupling superconductivity assuming an isotropic gap function . with increasing the third - order anharmonicity @xmath0 of the tetrahedral ion potential , we find a crossover in the energy spectrum to a quantum tunneling regime .
we obtain strongly enhanced transition temperatures around the crossover point .
we also investigate the anharmonic effects on the debye - waller factor , the phonon spectral functions and the density profile , as a function of the anharmonicity @xmath0 and temperature .
the isomorphic first - order transition observed in kos@xmath1o@xmath2 is discussed in terms of the first excited state energy @xmath3 , and the coupling constant @xmath4 in the strong coupling theory of superconductivity .
our results suggest the decrease in @xmath4 and increase in @xmath3 below the first order transition temperature .
we point out that the change in the oscillation amplitude @xmath5 and @xmath6 characterizes this isomorphic transition .
the chemical trends of the superconducting transition temperature , @xmath4 , and @xmath3 in the @xmath7-pyrochlore compounds are also discussed . |
stern and gerlach showed in their famous experiment in 1922 that a beam of silver atoms is transversely split into spin - up and spin - down components by an inhomogenous magnetic field @xcite . after this discovery , the question was raised whether a corresponding experiment could also be conducted with an electron beam . for charged particles ,
however , the spin - separating mechanism is hindered by the influence of the lorentz force ( in combination with the uncertainty principle ) , as was pointed out by bohr and pauli .
more precisely , they argued that it is impossible to separate an electron beam into spin components by means of experiments based on the concept of classical particle trajectories @xcite .
this conclusion was later turned into a paradigm about the impossibility of a stern - gerlach spin polarizer for free electrons and has entered into the textbook literature .
first attempts to overcome this paradigm have been undertaken in the late 1990s .
the principle feasibility of separating spins in an electron beam by the inhomogenous magnetic field of two antiparallel electric currents was analyzed in terms of classical trajectories @xcite .
partial splitting of spin components by a unidirectional inhomogenous @xmath0 field was shown by means of combined classical and quantum calculations @xcite .
complete spin separation through a `` longitudinal '' stern - gerlach effect , where the electron beam moves along the magnetic field axis and is split in this direction , was obtained from a full quantum treatment @xcite .
significant progress towards a stern - gerlach - like spin polarizer for electrons has been achieved in very recent years .
it has been shown theoretically that spin - polarization of electron beams may be created by magnetic gratings which are either formed by microscopic current loops @xcite or solid - state nanostructures @xcite . while in the first case , the diffraction pattern consists of characteristic spin doublet lines , in the latter case , the spin polarization occurs a talbot length away from the exit surface of the grating . besides , a mach - zehnder interferometer , which combines standing laser waves with static magnetic fields from solenoids , has been proposed to exert magnetic phase control on electron beams @xcite . by exploiting fine - tuned spin - dependent phase shifts , a transverse spin separation into two components similar to the original stern - gerlach experiment is predicted .
moreover , a spin - polarizing wien filter for electron vortex beams was described which relies on multipolar electric and magnetic fields with cylindrical symmetry and a conversion of orbital to spin angular momentum @xcite .
very recently , two proposals have been put forward how spin polarization of free electron beams may be attained by exploiting the kapitza - dirac ( kd ) effect @xcite .
previous studies of spin - dependent kd scattering have always found `` symmetric '' spin effects : the probabilities to scatter an incident spin - up electron into a spin - down state and vice versa coincide @xcite .
consequently , when an unpolarized electron beam is incident , the scattered part of the beam will remain unpolarized because the spin of each single electron has flipped . this symmetry is broken in @xcite by considering an electron moving along the axis of a monochromatic standing laser wave of circular polarization .
the two different spin orientations along the beam axis are found to precess with slightly different rabi frequencies , allowing to divide the electron beam longitudinally into two spin - polarized portions by tailoring the interaction time . a transverse separation of the spin components can be achieved by an interferometric arrangement of three pairs of linearly polarized laser beams @xcite . in the first stage of this setup ,
the electron beam is divided into two portions by scattering from a bichromatic laser field in a three - photon kd process .
afterwards , the partial beams are superposed coherently , leading to two outgoing electron beams with a high degree of spin polarization
. the original kd effect refers to ( spin - independent ) electron scattering on a standing wave of light which can be formed by two counterpropagating waves of the same frequency @xcite . by absorbing one photon from one of the laser beams and emitting another photon of opposite momentum into the counterpropagating beam ( stimulated compton scattering ) ,
the electron is elastically scattered . a theoretical generalization to strong laser fields
has been developed @xcite . the effect was experimentally observed both in the bragg and the diffraction regimes @xcite .
related experiments studied kd scattering on atomic beams @xcite . for reviews on the subject , we refer to @xcite .
moreover , recent studies on spin - insensitive kd scattering may be found in @xcite .
in the present paper , we continue our investigations @xcite of spin - dependent kd effects in bichromatic laser fields .
our main goal is to reveal the influence exerted by the polarizations of the counterpropagating laser waves on the electron spin dynamics . to this end
, the time - dependent dirac equation is solved by combined analytical and numerical tools .
in particular , we demonstrate pronounced spin dynamics at relatively low laser intensities in mixed setups which combine laser waves of circular and linear or elliptical polarization . instead ,
when both waves are circular - polarized , spin dynamics are shown to emerge at sufficiently high field strengths only .
our paper is organized as follows . in sec .
we perform an analytical treatment of spin - dependent bichromatic kd scattering , which is valid in the perturbative regime of low field strengths . in sec .
iii we extend our consideration to higher laser intensities by way of numerical simulations based on the time - dependent dirac equation .
our conclusions are summarized in sec .
iv . for notational convenience
, we set @xmath1 in this paper .
the product of two four - vectors @xmath2 and @xmath3 is denoted as @xmath4 . besides , feynman slash notation is employed for the four - product with dirac @xmath5-matrices .
throughout this paper , we shall consider the following setup .
an electron of charge @xmath6 is scattered from two counterpropagating laser beams , which consist of a ( `` strong '' ) right - traveling component @xmath7 with wave vector @xmath8 and a ( `` weak '' ) left traveling component @xmath9 with wave vector @xmath10 . the momentum of the incident electron is chosen to lie in the @xmath11-plane , i.e. @xmath12 , with @xmath13 . by absorbing two photons from the strong mode and emitting one photon into the weak mode , the electron is bragg scattered into the mirrored momentum state with @xmath14 , thereby fulfilling the laws of energy and momentum conservation . the question is to which extent the electron spin plays a role in this scattering process , depending on the polarization geometry of the laser waves .
the short - time behavior of the scattering probability can be obtained by a consideration within the third order of time - dependent perturbation theory .
such kind of analysis was carried out in @xcite for the case when both waves are linearly polarized along the @xmath15-axis .
the calculation is straightforward , but rather tedious . in the same paper
it was shown ( see app .
b therein ) that the same result may be obtained from an alternative treatment which describes the electrons by volkov states which are dressed by the strong field mode @xmath7 .
the remaining effect of the weak mode @xmath9 can then be treated as a first - order perturbation .
this approach has turned out to be rather convenient .
it shall be followed in this section .
regarding the kd effect as stimulated compton scattering , we start from the @xmath16 matrix describing multiphoton compton scattering @xcite @xmath17 here , @xmath18 denotes the dirac - volkov state for the incoming electron dressed by the field @xmath7 .
the latter is given by a generic plane wave of the form @xmath19 in radiation gauge , with the wave four - vector @xmath20 .
the exponential factor contains the classical action , with the field - dependent term @xmath21 d\phi \label{eqn : volkov_action}\ ] ] the free dirac spinors @xmath22 are taken from @xcite , with the spin quantized along the @xmath23 axis .
accordingly , @xmath24 represents the dirac - volkov state for the scattered electron .
the electron four - momenta are given by @xmath25 and @xmath26 , where @xmath27 .
the momenta are assumed to be nonrelativistic so that @xmath28 .
note that , even though we are interested in nonrelativistic electron momenta , a treatment beyond pauli theory is necessary for laser field polarizations different from linear @xcite .
we express the vector potential as @xmath29 , with a polarization four - vector @xmath30 satisfying @xmath31 and @xmath32 . without loss of generality , we can choose @xmath33 to be real , but note that @xmath34 might still be a complex number .
its absolute value @xmath35 ranges from one for linear polarization to zero for circular polarization .
a corresponding notation is employed for the counterpropagating wave , whose amplitude , polarization , wave four - vector and frequency are @xmath36 , @xmath37 , @xmath38 and @xmath39 , respectively . the space - time integration in can be performed via a fourier series expansion , according to @xmath40 \nonumber\\ & = & \sum_{n \in \mathbb{z } } \tilde{j}_n \left ( \alpha_p - \alpha_{p ' } , \rho \left ( \beta_p - \beta_{p ' } \right ) ; e^{i\delta } \right ) \nonumber\\ & & \times e^{-i n k_1 \cdot x - 2 i \left ( \beta_p - \beta_{p ' } \right ) k_1 \cdot x}\ , \label{eqn : s_phase } \end{aligned}\ ] ] with the abbreviations @xmath41 besides , the generalized bessel functions @xmath42 with the ordinary bessel functions @xmath43 , have been used . in correspondence ,
the @xmath16-matrix for the electron transition from @xmath44 to @xmath45 by absorbing two photons from @xmath7 and emitting one into @xmath9 can be expressed as @xmath46 \tilde{j}_1 e^{i \left ( p ' - p - k_1 \right ) \cdot x } \right ) u_{p , s } \nonumber\\ & \approx & \frac{i e}{2 } t \bar{u}_{p ' , s ' } \left [ a_2
\tilde{j}_2 \bar{\slashed{\epsilon}}_2 - \frac{e a_1 a_2}{4 c } \tilde{j}_1 \left ( \frac{\slashed{\epsilon}_1 \slashed{k}_1 \bar{\slashed{\epsilon}}_2}{k_1 \cdot p ' } + \frac{\bar{\slashed{\epsilon}}_2 \slashed{k}_1 \slashed{\epsilon}_1}{k_1 \cdot p } \right ) \right ] u_{p , s } \nonumber\\
\label{eqn : s1 } \end{aligned}\ ] ] here , @xmath47 defines the component of @xmath48 which describes absorption of one photon from this wave .
similarly , @xmath49 describes one - photon emission into @xmath9 , where @xmath50 . for notational simplicity
, the arguments of the generalized bessel functions have been omitted .
note that in eq .
we have restricted ourselves to the resonant contribution , which agrees with the bragg condition .
the integration has thus produced a factor of @xmath51 , with the formal interaction time @xmath52 and quantization volume @xmath53 .
the spinor - matrix products in eq .
read @xmath54 and @xmath55 u_{p , s } \right)_{s ' , s } \nonumber\\ & \approx & \frac{2 \vec{\epsilon}_1 \cdot \vec{\epsilon}_2^{~*}}{m c } + \frac{4 i \omega}{m^2 c^3 } \left ( \vec{\epsilon}_1 \times \vec{\epsilon}_2^{~ * } \right)\cdot \vsigma \label{eqn : matrix2 } \end{aligned}\ ] ] the notation @xmath56 on the left - hand sides in eqs . and is meant to indicate @xmath57 matrices whose entries are indexed by the spin quantum numbers .
furthermore , from the taylor series of the generalized bessel functions , we obtain @xmath58 the applicability of these approximations requires , in particular , that the momentum component @xmath59 is sufficiently small . putting all pieces together , the @xmath16-matrix adopts the form @xmath60 with the spin- and polarization - dependent rabi matrix @xmath61 besides , in the second step , the dimensionless field amplitudes @xmath62 , which are commonly used in strong - field atomic physics , have been introduced .
equation represents the perturbative limit of the three - photon kd process , which assumes @xmath63 .
the polarization states of the counterpropagating laser waves enter through the rabi matrix . for various combinations of linearly ,
circularly and elliptically polarized waves , the specific form of @xmath64 is indicated in table [ tab : s_matrix ] .
.[tab : s_matrix ] specific forms of the rabi matrix @xmath64 [ see eq . ] for various polarizations configurations of the laser waves .
the usual definition @xmath65 is applied . [
cols=">,^,<,^,^,^",options="header " , ]
in order to confirm our analytical predictions and to extend the consideration to higher field intensities beyond the leading order of perturbation theory , we have conducted numerical simulations .
the time - dependent dirac equation @xmath66 in the presence of the combined laser fields @xmath67 $ ] was solved by runge - kutta methods .
a slowly varying envelope function @xmath68 is introduced here to model the switching on and off of the laser field in the simulation . the turn - on and turn - off phases
are sin@xmath69-shaped and comprise five cycles of the fundamental laser period . this way , the time evolution of the spinor wave function of the electron is obtained , starting from an initial state with longitudinal momentum @xmath13 and spin projection either up ( @xmath70 ) or down ( @xmath71 ) . below
, these initial states and the corresponding scattered states with @xmath14 shall be symbolically denoted as @xmath72 , @xmath73 , @xmath74 , and @xmath75 . in the regime of laser parameters under consideration
, the outgoing electron state after the interaction with the laser fields can be expanded into these four states . while states with other longitudinal momenta are important during the interaction and taken into account in the simulation , their contributions to the final electron state are neglibibly small .
note besides that the value of the initial transverse momentum @xmath59 is suppressed in the symbolic notation .
this is justified because the canonical momentum in this direction is conserved , so that @xmath76 . for reasons of practical feasibility ,
our numerical computations assume high - intensity laser fields with frequencies in the x - ray domain .
radiation sources with corresponding characteristics are , in principle , available through high - harmonic emission from plasma surfaces @xcite or high - power x - ray free - electron laser ( xfel ) facilities , such as the european xfel ( hamburg , germany ) or the linac coherent light source ( stanford , california ) @xcite . at present , the latter is able to generate brilliant x - ray pulses with up to @xmath77kev photon energy , @xmath78w / cm@xmath69 intensity , and @xmath80fs duration .
the scenario , when both waves are linear - polarized along the @xmath15 axis , was considered within pauli theory in @xcite , focussing on incident electron momenta with @xmath81 . in this case , the rabi frequency @xmath82 results from eq .
, in agreement with eq .
( 16 ) in @xcite .
this frequency describes the flopping dynamics between the resonantly coupled momentum states @xmath44 and @xmath45 . as an illustration ,
the outcome of a numerical simulation for this polarization geometry is depicted in fig .
[ lin - lin ] .
it shows the dependence of the final - state electron population on the interaction time , i.e. , every plotted data point corresponds to a full interaction with switching on and off the fields .
a rabi oscillation between the states @xmath72 and @xmath75 takes place .
the amplitude @xmath83 of these oscillations does not reach the maximal value of 1 , though .
the population of the scattered state rather follows the behavior @xmath84 with the oscillation frequency @xmath85 .
the fact that the rabi oscillations are not fully developed can be attributed to an intrinsic detuning which arises when the field intensities are not so small @xcite .
the oscillation amplitude in fig . [ lin - lin ] , following from dirac theory , amounts to @xmath86 .
we point out that a slightly smaller amplitude of @xmath87 would result from a numerical solution of pauli s equation .
this indicates that some relativistic effects are present in the detuning which arises for the parameters in fig .
[ lin - lin ] .
we emphasize that the same kind of rabi oscillation would arise if the electron was incident with opposite spin instead .
then the flopping dynamics would occur between the states @xmath73 and @xmath74 .
the corresponding scattering probabilities for up@xmath88down and down@xmath88up are the same in the scenario where both laser waves are linear - polarized , as was mentioned in the introduction . the other polarization geometries shall be discussed in the subsequent sections .
( green solid line ) and @xmath89 ( black dashed line ) for kd scattering from bichromatic counterpropagating laser waves of linear polarization , as a function of the interaction time @xmath52 .
the field parameters are @xmath90ev , @xmath91ev and @xmath92ev , corresponding to a combined laser intensity of @xmath93 .
the electron is incident with longitudinal momentum of @xmath94 and transverse momentum of @xmath95.,scaledwidth=45.0% ] when studying spin effects in kd scattering , it is suggestive to consider fields of circular polarization . in this case , the photons carry a definite spin along the beam axis ( helicity ) which might be transfered to the electron during the combined emission - absorption process .
indeed , it has recently been shown that pronounced spin dynamics may arise in two - photon kd scattering processes which occur in a monochromatic standing laser wave of circular polarization @xcite ; see also @xcite for related studies . in general
, however , spin - flip transitions in circular - polarized fields have to compete with spin - preserving transitions because in the nonrelativistic language of pauli theory the @xmath96 interaction term can not be avoided . as a result
, spin flips are expected to be suppressed , unless the electron moves along the laser beam axis @xcite . in this subsection
, we analyze spin effects in bichromatic kd scattering when the strong mode @xmath7 is circularly polarized .
it has been speculated that distinct spin effects could arise when an electron is kd scattered from bichromatic laser fields which both are circularly polarized @xcite .
the underlying idea is that , in a three - photon process , the net absorption of spin angular momentum from the field can be one unit of @xmath97 .
this spin transfer could be compensated by a spin flip of the electron along the laser beam axis .
while this idea is very appealing , the classical simulations , which were carried out in @xcite based on the bargman - michell - telegdi equation , could not reveal spin flips , though .
our analytical calculations confirm that , to leading order , there is no spin flip involved in the three - photon kd effect .
rather , as shown in table [ tab : s_matrix ] , the rabi matrix is given by @xmath98 and thus independent of the spin .
this results holds for two counterpropagating waves which are circularly polarized in the same sense . according to table [ tab : s_matrix ] , qualitatively the same conclusion holds , when only the strong wave is circular - polarized , whereas the weak mode is linear - polarized .
when , instead , counterrotating circular - polarized laser waves are applied , no scattering occurs at all .
this is in line with the outcome of @xcite regarding two - photon kd processes .
nevertheless , the analytical predictions of the previous section only apply to the low - intensity regime . by performing a systematic series of numerical simulations
, we found that pronounced spin dynamics in corotating circular - polarized fields may arise in a specific range of parameters .
an example is shown in fig .
[ circ - circ ] . for an incident electron state @xmath73 with zero transverse momentum , a rabi oscillation dynamics with the @xmath74 state appears .
remarkably , the electrons are only scattered when they are initially in the spin - down state , but not when they are in the spin - up state .
the effective scattering term must therefore be proportional to @xmath99 .
we note besides that the measured oscillation frequency does not follow the @xmath100scaling of eq . .
our results therefore suggest that the scattering occurs in a nonperturbative domain , is driven highly non - linear , and the number of involved photons can not be discriminated easily as before due to substantial detuning .
( blue dash - dotted line ) and @xmath101 ( red dotted line ) for kd scattering from bichromatic counterpropagating laser waves with corotating circular polarization , as a function of the interaction time @xmath52 .
the field parameters are @xmath90ev , @xmath102ev .
the combined laser intensity is @xmath103 .
the initial electron momentum components are the same as in fig .
[ lin - lin].,scaledwidth=45.0% ] a summary of our parameter scan is illustrated by the schematic diagram in fig .
[ dyke - circ ] . in analogy with the case when both fields are linear - polarized ( see fig . 7 in @xcite ) , a dykelike structure separates various interaction regions . in the current case of two corotating circular - polarized fields , however , in the perturbative region far below the dyke
no spin - dependent kd scattering is found .
rabi oscillations with sizeable amplitude rather occur in the region just above the dyke , where the field intensities are relatively high .
[ see eq . ] in the @xmath104@xmath105 plane , indicating various parameter regimes , when both waves are corotating circular - polarized .
the dyke summit is located approximately at @xmath106ev ( order of magnitude).,scaledwidth=40.0% ] in the previous subsection we have seen that a strong ( i.e. fundamental ) field mode of circular polarization is only beneficial for spin effects in bichromatic kd scattering when the field intensities are sufficiently high .
when , instead , the weak ( i.e. second - harmonic ) field mode is circularly polarized , the analytical considerations in sec .
predict pronounced spin dynamics already at low laser intensities where perturbation theory applies .
this prediction is confirmed by a numerical solution of the dirac equation .
figure [ lin - circ-0 ] shows the rabi oscillations between an incident electron state @xmath72 and the scattered spin - flipped state @xmath75 when the weak mode is circular - polarized and the strong - mode linear - polarized .
a vanishing transverse electron momentum is assumed .
because of detuning effects @xcite , the oscillations are not fully developed but reach a large amplitude of @xmath107 . note
that , when the electron was initially in state @xmath73 , no spin - flipping transitions are found in our simulations .
in fact , the electron will remain in its initial state , then .
this is in accordance with the rabi matrix being proportional to @xmath108 , see table i. thus , in contrast to the case of two linearly polarized laser waves in sec .
iii.a , the symmetry between the up@xmath88down and down@xmath88up transition probabilities is broken in the present field configuration @xcite .
( green solid line ) and @xmath89 ( black dashed line ) for kd scattering from a laser field consisting of a linear - polarized wave with @xmath109ev , @xmath91ev and a counterpropagating circular - polarized wave with @xmath110ev , @xmath92ev .
the combined laser intensity is @xmath93 .
the electron is incident with longitudinal momentum of @xmath111 and transverse momentum of @xmath95.,scaledwidth=45.0% ] the clean picture of a two - state flopping dynamics changes dramatically when the electron is incident with nonzero transverse momentum .
an example is depicted in fig .
[ lin - circ - px ] .
now the electron probability density is distributed over all four states , leading to a rather complex evolution with the interaction time .
one can see that , in the limit of small @xmath52 , the initial population is transfered not only into the spin - flipped @xmath75 state but also into the spin - preserving @xmath74 state .
this is because the rabi matrix contains an additional term now , which is spin - independent and proportional to @xmath59 ( see table i ) . due to this term
, population can also be shifted from the @xmath75 state to the @xmath73 state , which accordingly starts to occur with some time delay , as fig .
[ lin - circ - px ] illustrates . in a real experiment , the detrimental effect caused by nonzero @xmath59
can hardly be avoided because the incident electron beam will always have some transverse momentum width . in light of this ,
the combination of linear and circular polarized laser waves is not yet optimal . but with @xmath112 .
due to the nonzero transverse momentum , also the electron states @xmath74 and @xmath73 are populated .
the corresponding occupation probabilities are shown by the red dotted and blue dash - dotted lines , respectively.,scaledwidth=45.0% ] interestingly , the robustness of the setup can be largely enhanced , when a small degree of ellipticity is imprinted on the strong mode . by a suitable choice of the polarization vectors , the @xmath59-dependent term in the rabi matrix
can be removed [ see eq . ] .
this is achieved , for example , when the strong mode possesses an ellipticity degree of 20% .
as indicated in table i , for this specially tailored field geometry , the rabi matrix becomes proportional to @xmath108 , even for nonvanishing values of @xmath59 .
an example of the population dynamics in the optimized setup is shown in fig .
[ ell - circ ] , with @xmath113 . while the same value of transverse momentum led to a very complicated behavior in fig .
[ lin - circ - px ] , in the present case the detrimental impact of @xmath59 is balanced by the properly chosen ellipticity of the strong mode . as a result ,
a clean rabi oscillation between the two states @xmath72 and @xmath75 is obtained , reaching an amplitude of about @xmath114 .
we verified numerically that the interaction in this field configuration is indeed independent of @xmath59 , at least for @xmath115 and the other parameters chosen as in fig .
[ ell - circ ] .
note that , when the electron was initially in state @xmath73 , no spin - flipping transitions are found in our simulations . , but the fundamental laser wave is elliptically polarized with polarization vector given by @xmath116 . as a consequence , only the states @xmath72 and @xmath75 couple with each other.,scaledwidth=45.0% ] in the context of the current section , a recent study on pair production in very strong laser fields is worth mentioning . as was shown in @xcite ,
spin - polarized electron - positron pairs can be created in two counterpropagating , monochromatic laser beams of elliptical polarization .
spin - dependent three - photon kd scattering in bichromatic fields , which both are linearly polarized , was discussed in @xcite .
as already mentioned above , symmetric spin effects arise in this case , where the scattering probabilities for up@xmath88down and down@xmath88up are the same .
nevertheless , a spin polarizer can be formed when three pairs of linearly polarized laser fields are combined in an interferometric setup .
then , one is sensitive to the quantum phase generated by the spin - dependent bichromatic kd effect , which forms the first step in the spin - polarizing beam splitter ( see fig .
[ fig : skizze ] ) . the setup may be viewed as three consecutive ponderomotive gratings , from which the incident electron beam diffracts .
they are given by the formulas @xmath117 for a bichromatic wave ( first stage of the interferometer ) and @xmath118 for a monochromatic standing laser wave ( second and third stage of the interferometer ) .
upon scattering from the spin - dependent potential , the incident electron beam is split into two portions , whose spin and momentum are entangled .
afterwards , both partial beams are reflected by a first interaction with the spin - independent potential and , finally , coherently mixed by another interaction with a monochromatic standing laser wave . as a result
, the outgoing electron beam is split into its spin components along the laser magnetic field .
scheme of the spin - polarizing interferometric beam splitter , as introduced in @xcite .
an incident electron beam undergoes three stages of kd scattering . as a result
, the outgoing electron beam is separated into its spin components @xmath119 and @xmath120 along the laser magnetic field , in close analogy to the stern - gerlach effect.,scaledwidth=40.0% ] details on the derivation of the spin - dependent effective ponderomotive potential in eq . are provided in @xcite .
it relies on a magnus expansion of the corresponding pauli - hamiltonian .
we would like to point out that the potential given in eq .
( 20 ) of @xcite contains an errorenous sign . in this form
, the equation would hold for a negative value of the unit charge @xmath121 .
the mistaken sign , though , does not affect the results presented there .
the relative position of the standing wave nodes can be adjusted by the phase shift parameter @xmath122 in the potential .
it turns out that the choice of this parameter has a strong impact on the performance of the spin polarizer . in @xcite ,
the phase parameter was chosen as @xmath123 in order to reach an optimized polarization degree of 77% for each of the outgoing beam portions . in fig .
[ degree ] we show how the polarization degree depends on this parameter .
[ see eq .
] of the polarization degree of the outgoing electron beam with positive momentum , as obtained by the spin - polarizing interferometric beam splitter of ref .
the black crosses present the numerical data , whereas the dotted line is an analytical curve of the form @xmath124 , fitted over the parameters @xmath125 and @xmath126 .
the outgoing portion of the electron beam with negative momentum has the opposite degree of polarization.,scaledwidth=48.0% ] our findings , that the polarization degree @xmath127 strongly depends on @xmath122 and that the maximally achievable value of @xmath127 remains below 100% , may be understood qualitatively as follows .
the field - induced detuning arising in the bichromatic kd effect has a twofold negative impact on the performance of the beam splitter .
first , the reduced rabi oscillation amplitude leads to an uneven splitting of the beam in the first stage .
this results in imperfect polarization because the destructive interference in the final stage for either spin states can not be complete .
second , because of the intrinsic detuning , the ( mean ) energies of the incident momentum mode and the scattered momentum mode slightly differ during the interaction . as a consequence , a relative offset developes between the quantum phases accumulated by both modes .
this is harmful for the spin polarization achieved after their coherent superposition in the final stage of the interferometer .
we argue that , by a proper choice of @xmath122 , this second detrimental effect can be compensated because @xmath122 enters into the same quantum phases .
this is why the resulting degree of polarization is modulated by the phase parameter , as depicted in fig . [ degree ] .
we have shown that electron spin dynamics in bichromatic kd scattering are very sensitive to the polarization states of the laser beams , whose frequency ratio was chosen as 2 . this way , previous investigations of the subject @xcite have been extended . when the fundamental laser beam is circular - polarized , no spin effects were found to leading order .
the latter corresponds to a three - photon process , where two photons of the fundamental frequency are absorbed and a second - harmonic photon is emitted ( or vice versa ) .
however , at higher field intensities , where a perturbative description does not apply , spin - flip transitions can occur .
they exhibit characteristic rabi oscillations as a function of the interaction time .
this effect was explicitly demonstrated for the configuration where both laser waves are circular - polarized .
spin dynamics can be controlled very effectively when a second - harmonic beam of circular polarization is combined with a fundamental laser beam of linear polarization .
this setup , which works at low field intensities , can be used as a longitudinal spin filter for electrons moving along the beam axis ( similarly to the scheme proposed in @xcite ) .
transverse momentum components of the electrons would exert a detrimental impact , though , allowing for spin - independent scattering events .
the setup can be optimized by utilizing a fundamental laser beam with a small degree of ellipticity .
if the latter is properly chosen with @xmath128% , the spin filter is effective for electrons with nonzero transverse momentum , as well . from an incident electron beam satisfying the bragg condition ,
only , say , spin - up electrons are scattered ( and thereby flipped to the spin - down state with respect to the laser beam axis ) whereas spin - down electrons go through undeflected . the laser field configuration thus acts as a spin polarizer , splitting the electron beam according to its spin composition and creating two outgoing partial beams with parallel spin orientation . for comparison
, we note that the interferometric spin polarizer proposed in @xcite produces outgoing partial beams with opposite spin orientations along the magnetic field direction , in close analogy with the stern - gerlach effect @xcite . in summary ,
bichromatic kd scattering enables spin polarization of free electron beams . by properly chosing the polarization geometry of the laser waves
, the spin dynamics can be controlled in order to optimize the desired spin properties of the output beams .
this study was supported by sfb tr18 of the german research foundation ( dfg ) under project no .
b11 . for recent papers related to the coupling between orbital and spin angular momentum for relativistic electrons , we refer to a. g. hayrapetyan , o. matula , a. aiello , a. surzhykov , and s. fritzsche , phys .
lett . * 112 * , 134801 ( 2014 ) ; m. wen , h. bauke , and c. h. keitel , sci .
* 6 * , 31624 ( 2016 ) .
p. h. bucksbaum , d. w. schumacher , and m. bashkansky , phys .
lett . * 61 * , 1182 ( 1988 ) ; p. l. gould , g. a. ruff , and d. e. pritchard , _ ibid . _ * 56 * , 827 ( 1986 ) ; p. j. martin , b. g. oldaker , a. h. miklich , and d. e. pritchard , _ ibid . _ * 60 * , 515 ( 1988 ) ; s. eilzer , h. zimmermann , and u. eichmann , _ ibid . _ * 112 * , 113001 ( 2014 ) .
see , e.g. , c. rdel , d. an der brgge , j. bierbach , m. yeung , t. hahn , b. dromey , s. herzer , s. fuchs , a. g. pour , e. eckner _
lett . * 109 * , 125002 ( 2012 ) ; m. cerchez , a. l. giesecke , c. peth , m. toncian , b. albertazzi , j. fuchs , o. willi , and t. toncian , _ ibid .
_ * 110 * , 065003 ( 2013 ) . | spin - dependent kapitza - dirac scattering of electron beams from counterpropagating bichromatic laser waves in various polarization geometries is studied .
the corresponding scattering probabilities are obtained by analytical and numerical solutions of the time - dependent dirac equation , assuming a field frequency ratio of 2 . when the fundamental field mode is circular - polarized
, we show that spin dynamics are generally suppressed at low intensities , but can become distinct at high intensities .
conversely , when a linearly or elliptically polarized fundamental mode is combined with a second harmonic of circular polarization , strong spin effects arise already at low field intensities . in particular ,
a polarization configuration is identified which acts as a spin filter for free electrons . |
one of the recurring themes in surface photochemistry on metal surfaces is the almost ubiquitous contribution that photoelectrons make to the observed photochemical processes . in particular ,
many phenomena have been interpreted as being the consequence of _ hot photoelectrons_- those photoelectrons having energies between the fermi and the vacuum levels of the system . in nanosecond
timescale photochemistry on noble metal substrates , the system has sufficient time for many cycles of electron excitation , scattering and decay , so that the resultant electron energy distribution that results has a form @xmath1 at energies well above @xmath2@xcite .
this rather smooth energy distribution can interact with the relevant electron affinity levels of the adsorbate system .
if one is particularly interested in the dissociation of molecules via dissociative electron attachment ( dea ) , then one would expect that the convolution of the @xmath3 with the @xmath4 will yield a flux of desorbing species that arises from a relatively wide range of photoelectron energies ( roughly the width of @xmath5 for energies a few ev above @xmath2 ) .
one can inquire if there are mechanisms that will modify this nascent energy distribution in particular , can the photoelectron energy distributions be narrowed to be more selective in the dissociation of adsorbed species . a related question has to do with electron transport in thin molecular films and how the electrons in these films can interact with other species- again , are there physical systems that will be highly selective in the interactions between co - adsorbed molecules ?
one approach was first detailed on theoretical grounds by rous@xcite who considered how an incoming electron might couple with the electron affinity levels of a @xmath6 molecule near a metal surface to vibrationally excite the molecule .
it was shown by application of a kkr calculation that the electron affinity level @xmath7 would display a strong enhancement by resonance with electron image states of the surface system . in order to form such an image state and have an image state lifetime sufficient to yield a significant enhancement ,
one requirement is that the surface have a bandgap at the relevant electron energy ( typically e@xmath8 - 0.5ev relative to @xmath9 for the _ n_=1 state ) , which does indeed occur for a number of low index metal surfaces .
a practical complication is that these image states on clean metal surfaces have been found to be very sensitive to contamination effects rather small concentrations of some adsorbates on metal surfaces have been found to open decay channels that suppress the image state @xcite .
a different approach is to take advantage of image states that can be formed at dielectric surfaces- these were first described in detail by cole and cohen@xcite and were subsequently observed using microwave spectroscopy on _ l_-he
surfaces@xcite . more recent work on image states at dielectric thin films has been in the context of characterizing the energies and lifetimes of these states using femtosecond two - photon photoemission ( fs-2ppe ) spectroscopy . in general terms ,
molecules having negative electron affinities ( i.e. repulsive for low energy electrons ) in thin layers form a bandgap that encompasses the image state energies ( i.e. the energy region just below @xmath9 ) . in qualitative terms , the decay of these electron image states is dominated by the transmission probability through the barrier , a barrier that increases the lifetimes roughly exponentially with film thickness .
there is now a substantial body of work on the effect of thin films interacting with intrinsic metal image states for both the case of available conduction band states at the image state energies ( image state resonance , e.g. kr and xe on cu(100)@xcite ) and pure image electron barrier layers ( e.g. thin alkane layers@xcite ) on various metal surfaces .
there has also been work on coupling of a negative ion resonance for o@xmath10 with an image state resonance formed for xe / cu(111 ) , in which the o@xmath10 anion state couples to the xe conduction band , with a strong layer dependence for the anion state lifetime@xcite . for the present case , a schematic image potential at a 4ml _
n_-hexane on cu(110 ) is shown in fig .
[ image_states ] .
we have used the modified dielectric continuum model of hotzel _
et al_@xcite to describe the image potential at the _ n_-hexane vacuum interface , and physical values similar to those chosen in recent fs-2ppe studies of this system ( for _ n_-hexane , ea=-0.20ev , @xmath11)@xcite .
the lowest lying image states have binding energies of 0.36ev ( _ n_=1 ) and 0.11ev ( _ n_=2 ) . in the case of the cu(110 )
substrate used in this work there is no intrinsic surface bandgap near @xmath12 at these energies , so that image state electrons that tunnel through the _
n_-hexane barrier layer will simply enter cu conduction band states . in the experiments described below
, a submonolayer of ` detector ' molecules ( ch@xmath0br or ch@xmath0i ) are adsorbed on top of the _ n_-hexane layer .
if the dissociative attachment cross section for these adsorbed molecules has a substantial overlap with the image state energy and the image state has sufficient lifetime , we expect that there would be image state mediated dea observed .
[ image_states ] also shows that the largest spatial overlap between the image states and an adsorbed molecule will occur for the _ n_=1 image state for higher _
n _ image states , the wavefunction becomes increasingly weighted away from the surface . from fs-2ppe experiments
, it is also found that the _ n_=1 image state has a substantially higher population than the _ n_=2 and higher states@xcite , so one would expect dissociation to occur primarily from the _
n_=1 state . for ch@xmath0br and ch@xmath0i in the gas phase
, dea occurs for low energy electrons with a threshold electron energy related to the curve crossing of the neutral ground state potential with the anion state@xcite .
the photodissociation of these halomethanes at metal surfaces and the role of photoelectron mediated dea has been studied by a variety of workers@xcite .
when these molecules are adsorbed at a metal or dielectric surface , the anion state is shifted downward in energy and the threshold electron energy for dea decreases and the survival probability for the anion increases , so that in general the peak dea cross section is at lower electron energy and has a substantially larger magnitude .
these effects are well known and have been previously reported in the literature ( e.g. see ref .
@xcite ) . in the case of ch@xmath0br and ch@xmath0i
, the lowering of the anion potential by @xmath13 will result in the threshold electron energy for dea at or close to 0ev and the peak dea cross section at very low electron energy . at the surface , these correspond to electron energies below the vacuum level@xcite , and can be accessed by the hot photoelectrons discussed above and , as will be discussed in the present work , will also have energetic overlap with electron image states at the surface .
experiments were performed using an ultra - high vacuum surface science and photochemistry apparatus that has been described previously@xcite . we have utilized both cu(100 ) and cu(110 ) single crystal substrates in this work .
the samples are heated to 915k by electron bombardment for cleaning and cooled to @xmath892k using liquid nitrogen , the temperature at which the experiments were performed .
sample cleanliness and order is monitored by auger electron spectroscopy and low energy electron diffraction measurements .
neutral products from surface photodissociation travel 185 mm to pass through a 4 mm diameter aperture to a differentially pumped extrel quadrupole mass spectrometer with an axial electron bombardment ionizer .
the sample to ionizer distance is 203 mm .
ions created in the ionizer then travel to the quadrupoles and are mass selected , in the present experiments using m / q=15amu ( ch@xmath14 ) .
selected ions are detected by a conversion dynode and electron multiplier in pulse counting mode and the pulses are then passed through a fast preamplifier .
the particle arrival pulses are time recorded using a multichannel scaler ( mcs ) that begins counting prior to the initiating laser pulse , with a typical time - of - flight ( tof ) spectrum using 1000 1.0@xmath15s time bins , and summing the counts from 1000 laser pulses
. the laser pulses ( @xmath85ns duration ) are produced by a small excimer laser ( mpb psx-100 ) operating at 20hz .
both unpolarized and linearly polarized light has been used in this work . to create polarized light
, the beam passes through a birefringent mgf@xmath10 crystal to separate p- and s - polarized light , which can then be directed at the sample .
unless otherwise stated , p - polarized laser pulses were used in the present study .
various laser wavelengths are produced by selection of different laser gases : 193 nm ( arf ) , 248 nm ( krf ) and 308 nm ( xecl ) .
pulse energies varied for the different laser wavelengths , but the fluence on the sample were less than 1mj/@xmath16 .
the laser pulses were collimated using a 4 mm diameter aperture and were unfocused on the sample .
the laser light is incident upon the sample at an angle of 45@xmath17 from the mass spectrometer axis when the cu sample is aligned to collect mass fragments along the surface normal , the light is incident at 45@xmath17 .
angular distributions of the photofragments were measured by rotating the sample with respect to the mass spectrometer axis , which also changes the incident angle of the laser light . in the present experiments
the effect of changing the incident light angle is small and its consequence on photoelectron generation is well understood .
the liquids used for dosing in the present work : _ n_-hexane ( aldrich @xmath1899.5% ) and ch@xmath0i ( aldrich 99.5%)- were transferred to a glass and teflon gas - handling system and degassed by freeze - pump - thaw cycles .
the ch@xmath0br ( aldrich 99.5% ) was used as delivered from a lecture bottle .
room - temperature vapor pressures were used for gas dosing through a precision leak valve that backfilled the chamber .
gas doses are described using uncorrected ion gauge readings .
gas doses that correspond to monolayer coverages were calibrated using temperature programmed desorption measurements or from prior experience with the selected molecule in other experiments .
a set of tof spectra using 248 nm ( @xmath19=5.0ev ) laser pulses for ch@xmath0br adsorbed on cu(110 ) are shown in fig .
[ ch3br_tof_comparison ] . in the gas - phase
, ch@xmath0br has a low cross section for direct photodissociation at this wavelength , so it is expected that dea by photoelectrons will be the dominant dissociation mechanism .
this expectation is supported by the observed tof spectra . in fig .
[ ch3br_tof_comparison](a ) the tof spectrum for a submonolayer of ch@xmath0br adsorbed on a 4ml _ n_-hexane layer on cu(110 ) shows the single narrow feature that is ascribed to the ct - dea via an image state intermediate .
that this feature is caused by photoelectrons is supported by the spectrum of fig .
[ ch3br_tof_comparison](b ) , in which the substrate photoelectron yield is reduced due to the addition of atomic iodine to the cu surface , which forms a c(2x2 ) cu(110)-i surface@xcite .
this iodization largely suppresses the ch@xmath0br dissociation signal , though a small peak at the same flight time is just visible . that this small feature is observed at the same flight time ( i.e. same ch@xmath0 translational energy )
supports the notion of an image state intermediate at the _ n_-hexane surface , since one would expect that the photoelectron energy distribution would be altered by the iodization . in general , changes in the photoelectron energy distribution
will be reflected in the tof spectrum if the dea occurs via a resonance at an energy where the photoelectron distribution is altered . in figs .
[ ch3br_tof_comparison](c ) and ( d ) the ch@xmath0br image state mediated dissociation to differs from that of 1ml ch@xmath0br on clean cu(110 ) ( fig .
[ ch3br_tof_comparison](c ) ) and 2ml ch@xmath0br on cu(110)-i ( fig .
[ ch3br_tof_comparison](d ) ) . in these latter spectra ,
the peak of the tof spectrum is shifted and the tof peaks are broader than that of fig . [
ch3br_tof_comparison](a ) . the absolute width in energy ( fwhm ) for the peak in fig .
[ ch3br_tof_comparison](a ) is 0.18ev , while for figs [ ch3br_tof_comparison](c ) and ( d ) the widths are 0.28 and 0.24ev respectively .
the relative peak widths ( @xmath20 ) are 0.32 , 0.62 and 0.64 respectively , again highlighting the distinctly narrower translational energy distribution for dissociation via the image state electron mechanism .
in particular , comparison of the spectra of figs [ ch3br_tof_comparison](a ) and ( d ) for ch@xmath0br on cu(110)-i is telling , since this ch@xmath0br layer is known to be quite well orientationally ordered and can display very narrow tof features@xcite . at 248 nm , it appears that the width of the tof feature for ch@xmath0br / cu(110)-i in fig .
[ ch3br_tof_comparison](d ) is dominated by the range of photoelectron energies that cause dea , resulting in ch@xmath0 fragments of more widely varying kinetic energy .
the narrow tof spectrum of fig . [ ch3br_tof_comparison](a ) for ch@xmath0br / hexane / cu(110 ) is the first piece of evidence that the dea is dominated by a very narrow range of incident electron energies , due to coupling with an image state that has captured a photoelectron which is then coupled to the ch@xmath0br affinity level for dea .
time - of - flight spectra for ch@xmath0br using 193 nm ( @xmath19=6.40ev ) are shown in fig .
[ ch3br_193nm_comparison ] .
in contrast to the 248 nm case , at 193 nm ch@xmath0br is known to have a large cross section for direct photodissociation in the gas - phase@xcite and at surfaces@xcite .
it is therefore somewhat surprising to find that for 0.1ml of ch@xmath0br on top of 4ml of _
n_-hexane on cu(110 ) that the same tof distribution is seen using 193 nm light as for 248 nm ( fig .
[ ch3br_193nm_comparison](a ) ) .
the peak in this tof spectrum is at exactly the same flight time as for the 248 nm case , showing the same energy release in ch@xmath0 translation in spite of the 1.40ev higher photon energy . that the position and width of this tof feature is unchanged shows that the electrons responsible for this dissociation have the same energy distribution and is further support for the image state mediated mechanism of photodissociation . at 193 nm , as the ch@xmath0br coverage in increased , the total ch@xmath0 yield increases but a new feature at 44@xmath21 flight time emerges ( fig .
[ ch3br_193nm_comparison](b ) ) and at higher ch@xmath0br coverage , dominates the spectrum .
this feature is associated with the 193 nm direct photolysis of ch@xmath0br , as it peaks at the same flight time as that seen in a previous study@xcite- a representative tof spectrum is shown in fig .
[ ch3br_193nm_comparison](c ) for comparison . thus even at a wavelength
at which ch@xmath0br has a large direct photodissociation cross - section , the image state mediated dissociation dominates at low coverage .
this is due to the large effective cross section for this process as the photoelectrons that are promoted to the image state are free - electron like parallel to the _ n_-hexane surface , and can cause dea in adsorbed species that are relatively distant from the location of the initial excitation . for example , for a trilayer of _
n_-pentane on ag(111 ) , the measured @xmath22=1 image state lifetime is 17.6ps ( and increasing by roughly @xmath23 per added monolayer)@xcite , so that such an @xmath24=1 image state electron with @xmath80.2ev of kinetic energy would have a mean free path length of roughly 4.6@xmath15 m . at higher ch@xmath0br coverages ( approaching 1 monolayer and above )
, the image state mediated process persists but the direct photodissociation mechanism becomes more competitive .
when the thickness of the _ n_-hexane barrier " layer is varied for a fixed submonolayer of ch@xmath0br , the ch@xmath0 yield is found to vary as shown in fig .
[ ch3br_vary_hexane_dose ] . from temperature programmed desorption ( tpd ) measurements made for _
n_-hexane on cu(110 ) and cu(100 ) , the completion of the first monolayer occurs at approximately 4.5l _
n_-hexane dose .
if it is assumed that the sticking coefficient is roughly constant for subsequent layers under our conditions ( borne out by tpd measurements ) , we observe from fig .
[ ch3br_vary_hexane_dose ] that very little ch@xmath0 yield is obtained for 2ml or less of _ n_-hexane .
the ch@xmath0 yield in fig .
[ ch3br_vary_hexane_dose ] is observed to increase significantly as the _
n_-hexane coverage increases beyond 2ml and reaches a maximum at approximately 4ml _ n_-hexane thickness . at coverages beyond 4ml
, the ch@xmath0 yield is found to decrease rapidly .
similar measurements made using the cu(100 ) substrate showed essentially identical behavior little or no ch@xmath0 yield from 2ml or less of _ n_-hexane and a rapid rise in yield to a maximum near 4ml _ n_-hexane coverage .
the decreasing ch@xmath0 yield for film thicknesses beyond 4ml is not due to changes in the ch@xmath0 angular distribution , as substantial angular changes were not observed .
it is possible that the adsorbed submonolayer is increasingly subsumed in the
_ n_-hexane layers for higher coverage ( e.g. at defects ) , though we do not see changes in the tof spectrum peak shapes ( i.e. translational energy distribution ) or angular distributions .
when cu(100)-i or cu(100)-cl substrates are used , there are much lower yields ( e.g. see fig . [ ch3br_tof_comparison](b ) ) but the coverage for maximum ch@xmath0 yield is observed to occur in the range of 12ml _ n_-hexane . in studies of image state lifetimes of similar molecular systems using fs-2ppe , it is found that the lifetime increases rapidly ( roughly exponentially ) with increasing layer thickness .
the cu(110 ) substrate used in the present work has no surface bandgap at the image state energies , so the decay of the populated image states will be dominated by the electron tunneling through the _ n_-hexane barrier . in contrast , the cu(100 ) surface has a prominent s - p bandgap around @xmath25 for energies relevant to the image states . that we observe essentially identical photodissociation behavior on both surfaces
indicates that the metal surface intrinsic bandgap is not playing a significant role in these experiments .
the low yield of ch@xmath0 for 2ml or less of _ n_-hexane suggests that the image state lifetimes are too short for the image state electrons to localize on the ch@xmath0br adsorption sites , and that for 3 to 4ml _ n_-hexane coverage , the probability to locate and dissociate the ch@xmath0br becomes favorable .
another possibility is that the dissociating ch@xmath0br@xmath26 temporary anion is quenched more rapidly for these thin _
n_-hexane films , presumably by overlap with unoccupied metal states ( i.e. autodetachment ) .
experience from previous dea photodissociation experiments would suggest that this is likely more important on the single monolayer thickness films . the optimal @xmath84ml _
n_-hexane thickness for dea from co - adsorbed ch@xmath0br seems to represent a compromise between the formation and population of the image state and the image state lifetime . for thicker _ n_-hexane films ( @xmath274ml )
the decreased ch@xmath0 yield is most likely due to a decrease in the initial population of the _
vacuum interface image state , which becomes increasingly less probable as the requisite photoelectrons generated in the metal substrate can not tunnel through the thick _ n_-hexane layers .
one more possibility to consider is that the ch@xmath0br dissociation is due to dea from an interface electron state _ inside _ the _ n_-hexane layer ( similar to states identified in ar thin films@xcite ) , and the rapid rise in the 2ml4ml region is due to the exclusion of this state from the thinner films . sustaining such
a state requires a surface bandgap for the metal substrate at the relevant energies and we observe no differences between the behavior on cu(100 ) ( which has such a gap at @xmath25 ) and cu(110 ) ( which does not ) .
also when experiments were performed on cu(100)-i or cu(100)-cl the maximum signal was seen for much thinner films ( 12ml ) so we do not believe that such an interface electron state in _ n_-hexane is responsible for the observed dissociation . if the observed photodissociation of adsorbed ch@xmath0br is mediated by an image state electron intermediate , then one expectation would be that this photodissociation would be very efficient for even a small submonolayer of the dissociating species .
the image state electron at the thin film vacuum interface , though constrained in the surface normal direction , would be expected to be mobile parallel to the surface as described above .
the data of fig .
[ vary_ch3br_on_hexane ] support this notion- there is a very steep increase in the observed ch@xmath0 yield for submonolayer ch@xmath0br on the 4ml _
n_-hexane on cu(110 ) system using 248 nm light . from tpd and photochemistry experiments on clean cu(110 ) , we have found that a 6.5l dose of ch@xmath0br corresponds to one monolayer coverage , and for adsorption on _ n_-hexane , one would anticipate that the sticking coefficient would be , at best , the same , so that the observed increase in the region of 02l dose corresponds to roughly 00.3ml coverage range . at higher coverages ( above @xmath80.5ml ) the ch@xmath0 angular distribution becomes broader , so the observed roughly constant yield at higher doses measured in the surface normal direction in fig .
[ ch3br_vary_hexane_dose ] is not entirely reflective of the total yield , however the steep increase in yield at low coverage is not affected by this . from image state studies at clean metal surfaces with intrinsic surface bandgaps ( e.g. using inverse photoemission ) , a sensitivity of the image state to surface contamination has been noted@xcite
. this surface contamination sensitivity does not seem to carry over to the image states created at thin _
n_-hexane surfaces , the characteristic image state dissociation is observable even at monolayer coverages of the dissociating species ( for example , see fig .
[ ch3br_193nm_comparison](b ) ) .
in addition to ch@xmath0br , we have also observed the same image state mediated dissociation occur for adsorbed ch@xmath0i submonolayers .
ch@xmath0i has a larger cross section for dea in the gas - phase than ch@xmath0br does , with a peak cross section of @xmath27100@xmath28@xmath16 for @xmath2960mev@xcite . in the adsorbed state
, ch@xmath0i has a large dea cross section for very low energy electrons ( @xmath30@xmath16 at e@xmath80ev when adsorbed on 10ml of kr)@xcite , with the peak dea cross section having apparently shifted to an energy below the vacuum level due to image interactions with the surface .
a tof spectrum obtained at 248 nm for 0.1ml ch@xmath0i on 4ml _
n_-hexane on cu(110 ) is shown in fig .
[ ch3i_tof_comparison](a ) . in this case , the ch@xmath0i tof distribution is narrower than that seen for ct - dea of adsorbed ch@xmath0i in other cases . if the ch@xmath0i coverage is increased , and p - polarized 248 nm light is used , photodissociation by both ct - dea and by direct photodissociation can be seen ( fig .
[ ch3i_tof_comparison](b ) ) .
similar to the case for ch@xmath0br at 193 nm , direct photodissociation can be seen in the small fast peak at 45@xmath21 flight time .
another contrast can be seen in fig .
[ ch3i_tof_comparison](c ) which shows a tof spectrum obtained from 2ml ch@xmath0i that is well ordered on a cu(110)-i substrate at 308 nm ( @xmath31 ) using s - polarized light . under these conditions ,
the tof spectrum is dominated by the dea mechanism@xcite from subvacuum level photoelectrons .
the tof spectrum for the ch@xmath0i on _
n_-hexane is noticeably narrower than that obtained at the longer wavelength on cu(110)-i .
although for ch@xmath0i we observe that the ch@xmath0 tof distributions are not as narrow as those found for ch@xmath0br , they are still narrower than those observed in other situations .
one possible explanation for the slightly broader ch@xmath0 tof distribution for ch@xmath0i , as compared to ch@xmath0br is that ch@xmath0i does have a substantially larger intrinsic ( i.e. not image state mediated ) dea cross section , so that there may be an additional contribution to the observed tof signal from a broader range of dea photoelectron energies . that dea can dominate the ch@xmath0i dissociation at 248 nm is in itself surprising , since the cross section for direct photolysis at this wavelength is quite large .
similar to the case for ch@xmath0br at 193 nm , for small submonolayer coverages of ch@xmath0i at 248 nm , the image mediated dea dominates the photodissociation .
direct photolysis is found to be increasingly important at higher ch@xmath0i coverage , especially when p - polarized light is present since the oriented ch@xmath0i has a large cross section in this case@xcite .
it has also been found that all of the same general findings made for ch@xmath0br are also found for ch@xmath0i the pronounced yield variation for varying _
n_-hexane spacer thickness , the large observed yields at low ch@xmath0i coverage and very similar angular distributions of the ch@xmath0 photofragments .
the findings made for ch@xmath0i complement those made for ch@xmath0br and lend further support to the proposed image state mediated dissociation mechanism .
the energetics of the dissociation by low energy electrons and the contrast between the case for ch@xmath0br and ch@xmath0i are further illustrated by the tof data shown in fig .
[ hires_tof_comparison ] . for rapid dissociation via dea for a free ch@xmath0x molecule , the translational energy of the ch@xmath0 fragment can be understood from : @xmath32 where the mass ratio accounts for momentum conservation for a dissociating free molecule , @xmath33 is the kinetic energy of the incident electron , @xmath34 is the electron affinity of the halogen atom ( @xmath35ev , @xmath36ev)@xcite , @xmath37 is the methyl halogen bond dissociation energy from the ground state ( @xmath38ev , @xmath39ev)@xcite , @xmath40 is the additional stabilization energy for the halogen anion at the surface and @xmath41 accounts for internal excitation of the ch@xmath0 fragment upon dissociation .
if we assume that @xmath41 is similar for both ch@xmath0br and ch@xmath0i and is near zero for the fastest ch@xmath0 fragments , then @xmath421.0ev gives @xmath43 consistent with the observed values .
this seems reasonable given that one would expect @xmath44 to be of the order of 0.8ev@xcite , by analogy with electron scattering measurements for n@xmath10 on thin rare gas films and @xmath450.2ev is reasonable .
based on these values , the expected tof _ differences _ for ch@xmath0 from ch@xmath0br and ch@xmath0i would be roughly 6.0@xmath21 , which is slightly larger than the 5.0@xmath21 shift we observe in fig .
[ hires_tof_comparison ] .
it would be expected that the solvation energies for the br@xmath26 will be somewhat larger than that for i@xmath26 , though the precise magnitude of this difference is not known for the halogen anions at a hexane surface a difference of @xmath80.04ev would make the values consistent with the observed flight time differences . another interesting observation from fig .
[ hires_tof_comparison ] is that there is no detectable shift between the tof spectra from 3ml and 5ml _ n_-hexane coverages for either ch@xmath0br of ch@xmath0i .
this implies that @xmath46 is essentially constant over the range of _
n_-hexane coverages used for ch@xmath0x dissociation . for example , between 3ml and 5ml _ n_-hexane films the _ n_=1 image state energy rises ( from calculations such as that shown in fig .
[ image_states ] ) from -0.40ev to -0.33ev but @xmath44 would be expected to decrease@xcite by @xmath80.090ev due to the increased anion to metal surface distance , for an expected change in tof of + 0.5@xmath15s . within the resolution of our tof spectra
we do not detect any shifts in either the leading edge or the centroid of the tof distributions .
it is also worth noting that while the contribution of higher energy @xmath22=2 image state electrons would yield significantly faster ch@xmath0 fragments ( by @xmath47s ) , the relative number of these is much smaller@xcite than for the @xmath22=1 image state , which seems to be primarily due to the much smaller spatial overlap of the @xmath22=2 wavefunction with the population of hot photoelectrons that are transported to the metal interface .
the photodissociation of ch@xmath0br and ch@xmath0i adsorbed on top of thin _
n_-hexane layers has been found to be caused by hot photoelectrons generated in the metal substrate .
the nascent photoelectron energy distribution is modified by the _
n_-hexane layers and some of these photoelectrons are temporarily trapped at the _ n_-hexane vacuum interface in the @xmath22=1 image state .
these image state electrons can interact with the co - adsorbed ch@xmath0x and cause dissociation via the dea mechanism .
the ch@xmath0 photofragments detected in the tof experiment have a translational energy distribution that is insensitive to the incident photon energy and the yields of ch@xmath0 are indicative of the image state mediated mechanism .
this mechanism is likely to be viable for a number of molecular systems , as a wide variety of molecules have the requisite negative electron affinity and bandgap to create such image states , and there are also many molecules which undergo various chemical processes due to interactions with very low energy electrons .
the author would like to thank the natural sciences and engineering research council ( nserc ) of canada for financial support of this work .
thanks also to jeff hnybiba , robert vogt and masresha berhanu , who provided assistance with the experiments as a part of their nserc undergraduate student research award tenure in the laboratory . | the 248 nm and 193 nm photodissociation of submonolayer quantities of ch@xmath0br and ch@xmath0i adsorbed on thin layers of _ n_-hexane indicate that the dissociation is caused by dissociative electron attachment from sub - vacuum level photoelectrons created in the copper substrate .
the characteristics of this photodissociation translation energy distributions and coverage dependences show that the dissociation is mediated by an image potential state which temporarily traps the photoelectrons near the _ n_-hexane vacuum interface , and then the charge transfers from this image state to the affinity level of a co - adsorbed halomethane which then dissociates . |
the invention of the laser in 1960 s facilitated exploration of various nonlinear phenomena in optics , such as harmonic generation @xcite , multi photon absorption @xcite , wave mixing @xcite , optical kerr effect @xcite , stimulated raman scattering @xcite and so on .
today nonlinear optics plays an important role in the developing of new light sources @xcite , nonlinear spectroscopy @xcite , multiphoton microscopy @xcite , all - optical signal processing @xcite , quantum information @xcite and many other applications @xcite .
optical parametric oscillation / amplification ( opo / opa ) @xcite which transfers energy from the driving field to the signal and idler waves is an example of parametric processes caused by the nonlinear light - matter interaction .
opa led to development of tunable radiation sources throughout the infrared , visible , and ultraviolet spectral regions @xcite .
such sources yield amplification of coherent radiation at lower - frequencies @xmath1 and @xmath2 by pumping nonlinear optical crystals with intense laser light at the sum frequency @xmath0@xmath4this process is known as the sum combination resonance .
it has been discovered in electronic circuits in 1950 s @xcite and later on found applications in optics @xcite .
similar resonance occurs in the free electron laser in which the low frequency plasma and laser waves are excited by the sum frequency undulator wave in the electron rest frame @xcite .
it also appears when electromagnetic waves interact with optical phonons in dielectric crystals @xcite . in the literature
, another type of combination resonance , namely the difference combination resonance , has been discussed in mechanical systems since the 1960 s ( see , e.g. , @xcite ) .
it can occur when the driving frequency matches the frequency difference between two normal frequencies @xmath5which yields amplification of @xmath1 and @xmath2 .
combination resonances affect dynamic stability of structures and appear in various systems having multiple degrees of freedom .
recently , the difference combination resonance has been experimentally demonstrated in an electronic circuit in the radio frequency range and proposed as a mechanism of generation of high frequency coherent radiation ( e.g. xuv or x - rays ) utilizing a low frequency ( e.g. , infrared ) driving field @xcite .
such a parametric amplifier , which operates by quantum amplification by superradiant emission of radiation , is called the qaser for short .
in contrast to opos , qaser generates light at higher frequencies than @xmath0 . here
we study the possibility of generation of high frequency coherent radiation using the difference combination resonance in nonlinear metamaterials .
metamaterials are artificial inhomogeneous structures composed of periodic subwavelength , polarizable elements
. they can give rise to negative refractive index , near - zero permittivity or permeability of the macroscopic response in the microwave , terahertz and even optical regions @xcite and have led to a variety of fascinating phenomena such as a prefect lens and invisible cloaks @xcite .
inclusion of nonlinear elements into metamaterials offers unique advantages to enhance nonlinearities by local - field amplification @xcite and to provide particular constructive configurations under the phase - matching condition for highly efficient nonlinear processes @xcite .
anomalous transmission properties of zero - permittivity channels can be used to boost kerr nonlinearities @xcite .
effect of nonlinearity in combination with the unique linear properties has been recently investigated in connection with new frequencies generation @xcite , solitary wave propagation @xcite , self tuning and hysteresis transition @xcite and parametric amplification @xcite .
study of parametric amplification in such materials has demonstrated that if for one of the waves the refractive index is negative ( directions of the wave vector and the poynting vector are opposite ) the phase matching leads to a possibility of loss compensation and distributed - feedback parametric oscillation with no cavity @xcite . in such system
the backward - propagating wave provides an automatic feedback mechanism and yields dynamics analogous to the quasi - phase matched mirror - less opo @xcite .
here we investigate parametric amplification in metamaterials caused by the difference combination resonance .
we find that when for the idler ( frequency @xmath1 ) and the signal ( frequency @xmath2 ) waves the refractive index has opposite signs they both can be amplified by a strong driving field with frequency @xmath6 ( see fig .
[ mfig1 ] ) . we show that the light amplification mechanism in negative index materials is equivalent to a combination resonance in a system of two coupled parametric oscillators with the opposite sign of masses .
we also find that the difference combination resonance can be achieved in materials with positive refractive index but strong anomalous dispersion .
namely , it can occur if at one of the generated frequencies @xmath1 or @xmath2 the group velocity is opposite to the phase velocity .
this condition is somewhat easier to realize than the negative refractive index .
photonic crystals @xcite and fiber bragg gratings @xcite are examples of materials with strong anomalous dispersion .
electromagnetically induced transparency ( eit ) is a popular technique to decrease absorption near a resonance in the region of strong dispersion . using the eit technique ,
the group velocity of light as low as meters per second or even completely stopped light was demonstrated in ultracold gases hau99,liu01 , hot gases @xcite and in a very cold solid turu02 .
negative group velocity can be also obtained in the kerr medium , which possesses a large nonlinear refractive index and long relaxation time , such as cr - doped alexandrite , ruby and gdalo@xmath7 @xcite .
parametric amplification process due to the difference combination resonance holds promise for a new kind of higher frequency convertor with tunable wavelength that makes use of metamaterials with negative refractive index or strongly dispersive media .
generation of high frequencies by a parametric resonance in systems with several normal modes can be understood in a simple example of two classical oscillators with masses @xmath8 and @xmath9 and ( positive ) frequencies @xmath10 and @xmath2 that are weakly coupled with a periodically varying strength .
the hamiltonian of such a system reads@xmath11using the hamilton equations of motion and rescaling the coordinate @xmath12 we obtain@xmath13@xmath14where @xmath15and@xmath16describe coupling between oscillators . for harmonic modulation of the coupling strength with frequency @xmath0
we write @xmath17where @xmath18 is the modulation amplitude . according to eq .
( [ xx ] ) , the coupling between oscillators is symmetric ( @xmath19 ) if @xmath20 and antisymmetric ( @xmath21 ) if the masses @xmath8 and @xmath9 have opposite signs @xcite .
next we assume that modulation frequency obeys the condition of the combination parametric resonance @xmath22making a slowly varying amplitude approximation , namely writing @xmath23@xmath24where @xmath25 and @xmath26 are slowly varying envelope functions on a time scale @xmath27 and keeping only the resonant terms , eqs .
( [ x1 ] ) and ( [ x2 ] ) reduce to@xmath28@xmath29which yield the following equation for @xmath25@xmath30solutions for @xmath31 exponentially grow with time , @xmath32 , provided @xmath33 .
namely , for symmetric coupling ( @xmath34 ) , the solution grows if @xmath35 ( the sum combination resonance ) . for antisymmetric coupling ( @xmath36 )
there is exponential grow provided @xmath6 ( the difference combination resonance ) . for both resonances
the gain per unit time @xmath37 is given by@xmath38 the sum combination resonance is used for generation of frequencies lower than @xmath0 .
this is the case for optical parametric oscillators that convert an input laser wave with frequency @xmath0 into two output waves of lower frequency ( @xmath1 and @xmath2 ) by means of the second order nonlinear optical interaction . on the other hand
, the difference combination resonance can be used for the generation of higher frequencies .
this has been recently demonstrated in nonreciprocally coupled electronic circuits in the radio frequency domain @xcite . as an illustration of combination resonances , in figs .
[ mfig2]-mfig4 we plot energy of the first @xmath39 and the second @xmath40 oscillator as a function of time for off - resonance ( fig .
[ mfig2 ] ) and on - resonance ( figs .
[ mfig3 ] and [ mfig4 ] ) modulation obtained by numerical solution of eqs .
( [ x1 ] ) and ( [ x2 ] ) . in simulations
we take @xmath41 and @xmath42 .
initially the first oscillator is excited while the second oscillator is at rest .
if @xmath43 does not satisfy the resonance condition ( [ r1 ] ) the oscillators do not exchange their energy , as shown in fig .
[ mfig2 ] .
however , if , e.g. , the condition of the difference combination resonance is satisfied , @xmath6 , the oscillator s energies exponentially grow for antisymmetric coupling ( see fig .
[ mfig3 ] ) .
this is the essence of the qaser amplification mechanism @xcite .
if the coupling is symmetric , the difference combination resonance yields an efficient energy exchange between two oscillators without exponential grow , as illustrated in fig .
[ mfig4 ] . this can be used in various applications , e.g. , for fast control of propagation of @xmath44-rays through crystals @xcite . in fig .
[ mfig5 ] we plot the gain @xmath37 as a function of the modulation frequency @xmath0 for symmetric ( @xmath45 ) and antisymmetric ( @xmath46 ) coupling between oscillators . for symmetric coupling the system undergoes the sum combination resonance at @xmath47 , but no difference combination resonance . on the other hand , for the antisymmetric coupling between oscillators there is the difference combination resonance at @xmath6 , but no resonance at @xmath35 .
the possibility of a difference combination resonance in materials with negative refractive index can be demonstrated in the simple example of magneto - optical effect . in a magneto - optic material
the presence of an externally applied magnetic field @xmath48 causes a change in the permittivity tensor @xmath49 .
the tensor becomes anisotropic with complex off - diagonal components .
the off - diagonal components of @xmath50 describe coupling between two perpendicular polarizations of an electromagnetic wave .
such coupling can be periodically modulated if the externally applied magnetic field is a harmonic function of time @xmath51 . here
we show that the problem of propagation of electromagnetic waves through the gyromagnetic medium can be reduced to the two coupled parametric oscillators discussed in the previous section .
evolution of the electromagnetic field in the medium is described by maxwell s equations@xmath52we assume that the wave propagates along the @xmath53-axis and the electromagnetic field of the wave has only @xmath54 and @xmath55 components which depend on @xmath56 and @xmath53 .
then the maxwell s equations ( [ me ] ) reduce to@xmath57@xmath58if the absorption losses can be neglected , the permittivity tensor is a hermitian matrix @xcite and we can write@xmath59@xmath60where the off - diagonal component @xmath61 to first order , is proportional to the applied external magnetic field @xmath48 that lies along the @xmath53-axis @xcite .
periodic variation of @xmath48 with time yields periodic modulation of @xmath62 , @xmath63 .
taking time derivative from eqs .
( [ me2 ] ) and using eqs .
( [ me1 ] ) , ( [ me3 ] ) and ( [ me4 ] ) we obtain @xmath64@xmath65change of functions@xmath66yields@xmath67@xmath68one can look for a solution of these equations in the form @xmath69 which results in the equations of two coupled harmonic oscillators with time dependent coupling@xmath70@xmath71where@xmath72@xmath73and@xmath74 eqs .
( [ x6 ] ) and ( [ x7 ] ) are identical to the equations of the coupled parametric oscillators ( [ x1 ] ) and ( [ x2 ] ) investigated in the previous section and , hence , they display the same combination resonances when the modulation frequency satisfies the condition @xmath75 .
namely , if @xmath76 the coupling between the two field polarizations @xmath54 and @xmath55 is symmetric and there is the sum combination resonance at @xmath35 which yields generation of lower frequencies @xmath1 and @xmath2 . this is the case for the positive index materials and is an operation mechanism of the optical parametric amplifiers .
the coupling between the two field polarizations is antisymmetric if @xmath77 and @xmath78 have opposite signs . to make frequencies in eq .
( [ x8 ] ) real the corresponding components of the permittivity tensor also should have opposite signs .
namely , if @xmath79 and @xmath80 then we must have @xmath81 and @xmath82 . under these conditions
there is the difference combination resonance at @xmath6 which yields generation of higher frequencies @xmath1 and @xmath2 .
this case is similar to the two coupled oscillators with the opposite sign of masses discussed in the previous section . according to eq .
( [ x8 ] ) , generated waves at frequencies @xmath1 and @xmath2 have the same wave vector @xmath83 , but orthogonal polarizations .
the values of @xmath1 and @xmath2 differ due to the anisotropy of the medium ( @xmath84 ) . using combination resonance
one can also generate waves with different wave vectors @xmath85 and @xmath86 if the driving field wave vector @xmath87 compensates for the momentum mismatch @xmath88 .
moreover , if we take into account frequency dispersion the expression for the gain @xmath37 is modified .
straightforward calculations similar to those we present in the next section yield @xmath89 where @xmath90 and @xmath91 are the group and phase velocities of the waves with frequencies @xmath92 .
( [ g1 ] ) shows that the difference combination resonance can also occur in materials with positive refractive index if at one of the generated frequencies the group and the phase velocities are opposite to each other .
this requires medium with strong anomalous dispersion .
we discuss these questions in details in the next section in which we consider combination resonances in nonlinear dispersive medium assuming that waves have the same linear polarization .
using nonlinear crystals the optical parametric oscillators generate light at frequencies @xmath1 and @xmath2 satisfying the condition of the sum combination resonance @xmath93 , where @xmath3 is the frequency of the strong driving ( pump ) field @xmath94 . here
we investigate the possibility of difference combination resonance in such materials . in the present model
we assume that medium is spatially uniform , isotropic and polarization is a nonlinear function of the electric field , while magnetization is a linear function of the magnetic field .
we also assume that the driving field and generated waves propagate along the @xmath95-axis and have the same polarization . under these conditions
the maxwell s equations reduce to@xmath96we write the total electric field as@xmath97where the strong driving field is assumed to be fixed@xmath98and @xmath99 @xmath100 . in nonlinear isotropic medium the total displacement field in the absence of dispersion
can be written as @xmath101@xmath102where we omitted small terms of the second order in @xmath103 .
@xmath104 is the linear permittivity of the material and @xmath105 is the nonlinear susceptibility .
if there is frequency dispersion it is convenient to write the maxwell s equations ( [ d1 m ] ) and the constitutive relations in terms of the field fourier components , namely@xmath106@xmath107@xmath108@xmath109 , \label{d4}\]]where we disregarded irrelevant term proportional to @xmath110 , took into account that fourier transform of @xmath111 gives @xmath112 and for simplicity assumed that the nonlinear susceptibility @xmath105 has no dispersion . in our considerations the wavenumber @xmath113 is real while frequency @xmath114 has small imaginary part @xmath37 which determines the wave amplification or attenuation per unit time .
( [ d3 ] ) and ( [ d4 ] ) into eq .
( [ d2 ] ) yields the following equation for @xmath115 and @xmath116@xmath117 \left ( e_{1}(\omega , k)+e_{2}(\omega , k)\right)\]]@xmath118 = 0 .
\label{d5}\ ] ] next we assume that @xmath119@xmath120where @xmath121 and @xmath92 obey the relations@xmath122@xmath123since the driving field propagates through the same medium its frequency @xmath43 and the wavenumber @xmath124 satisfy the similar equation@xmath125eqs .
( [ s1a ] ) and ( [ s2a ] ) are consistent with eqs .
( [ s3])-(s5 ) provided@xmath126where @xmath127 and @xmath128 stands for @xmath129 and @xmath130 , etc .
we look for a solution of eq .
( [ d5 ] ) in the form@xmath131where @xmath37 is the gain per unit time . keeping the resonant terms we obtain @xmath132 \delta ( \omega \pm \omega _ { 1}-ig)\delta ( k\pm k_{1})=0 , \label{h1}\]]@xmath133 \delta ( \omega -\omega _ { 2}-ig)\delta ( k - k_{2})=0 .
\label{h2}\]]the factor in front of the delta functions in eq .
( [ h1 ] ) must be calculated at @xmath134 and @xmath135 , while the factor in eq .
( [ h2 ] ) should be estimated at @xmath136 and @xmath137 . making a taylor expansion one can write@xmath138 = \pm ig\frac{2c\omega _ { 1}n_{1}}{v_{g1}% } , \label{g2}\]]@xmath139 = -ig\frac{2c\omega _ { 2}n_{2}}{v_{g2 } } , \label{g3}\]]where @xmath140 is the refractive index ( @xmath141 is positive for @xmath104 , @xmath142 and negative for @xmath104 , @xmath143 ) , @xmath144is the group velocity of the wave at frequency @xmath114 and @xmath145 is the speed of light in vacuum . for brevity we introduced notations @xmath146 and @xmath147 . taking into account eqs .
( [ g2 ] ) and ( [ g3 ] ) , and keeping the leading terms , eqs .
( [ h1 ] ) and ( [ h2 ] ) reduce to@xmath148@xmath149eqs .
( [ d8 ] ) and ( [ d9 ] ) give the following expression for the gain@xmath150where @xmath151 is the phase velocity at frequency @xmath92 . in the time domain solution ( [ d11 ] )
reads @xmath152this solution exponentially grows with time if @xmath37 is real ( and positive ) .
thus , in order to have amplification the right hand side of eq .
( [ s12 ] ) must be positive . for materials with @xmath153
there is gain for the upper sign , that is when @xmath35 .
such materials yield the sum combination resonance and are used in conventional optical parametric amplifiers . in order to achieve light amplification at the difference combination resonance @xmath6 ( the lower sign in eq .
( [ s12 ] ) ) we must have@xmath154 conditions ( [ c0 ] ) and ( [ c1 ] ) can be fulfilled , e.g. , if the crystal has a negative refractive index at the frequency @xmath1 ( @xmath155 and @xmath156 ) , but a positive refractive index at the frequency @xmath157 ( @xmath158 and @xmath159 ) .
in addition , eq .
( s12 ) shows that the difference combination resonance can also occur in nonlinear materials with positive refractive index if at one of the generated frequencies @xmath1 or @xmath2 the group velocity @xmath160 is opposite to the phase velocity @xmath161 .
this condition is somewhat easier to achieve than the negative refractive index .
it requires strong anomalous dispersion , however , material does not need to have both magnetic and electric responses .
in our analysis we have assumed that medium is spatially uniform .
an interesting question appears in this connection .
namely , if a material is nonuniform can this eliminate the requirement of having a negative refractive index to achieve the difference combination resonance in medium with weak dispersion ?
in appendix a we investigate this question for materials described by the local constitutive relations @xmath162 and @xmath163 .
based on a general consideration we show that the difference combination resonance in nonlinear weakly dispersive materials is possible only if for one of the generated frequencies both the magnetic permeability @xmath164 and dielectric permittivity @xmath104 are negative in some region of space .
in this paper we consider parametric generation of coherent radiation in nonlinear medium produced by a strong driving field . in conventional nonlinear crystals with a positive refractive index such parametric mechanism yields generation of smaller frequencies due to the sum combination resonance ( see eq .
( [ s1 ] ) ) which is the operating principle of optical parametric amplifiers . here
we find that in left - handed materials the difference combination resonance described by eq .
( [ d1 ] ) can occur if for one of the generated frequencies the refractive index is negative .
such a resonance yields excitation of frequencies higher than @xmath0 .
we show that this mechanism is similar to a combination resonance in a system of two coupled parametric oscillators with periodically modulated coupling strength .
the difference combination resonance occurs when the coupling between oscillators in the equation of motion is antisymmetric [ see eqs .
( [ x1 ] ) and ( [ x2 ] ) ] .
such antisymmetric coupling is realized in the hamiltonian system if the two oscillators have the opposite sign of masses .
if coupling is symmetric ( both masses are positive ) the system has the sum ( rather than the difference ) combination resonance .
we also find that generation of higher frequencies by means of the difference combination resonance is possible in nonlinear materials with positive refractive index but strong anomalous dispersion .
in such a medium the resonance can occur if at one of the generated frequencies the group and phase velocities are opposite to each other .
anomalous dispersion is easier to realize than the negative refractive index .
thus , highly dispersive nonlinear optical systems are an attractive tool for applications in generation of high frequency coherent radiation .
the sum combination resonance is a robust phenomenon which occurs in various areas of physics .
in contrast , the difference combination resonance is hard to achieve because it requires special conditions , e.g. , nonreciprocal coupling .
recently such a resonance has been experimentally demonstrated in nonreciprocally coupled resonant rlc circuits at radio frequencies @xcite .
it has also been proposed for generation of high frequency coherent radiation in atomic medium by driving the atomic ensemble with a much smaller frequency @xcite .
the proposed new kind of light amplifier ( called the qaser ) , contrary to a laser , does not need any population of atoms in the excited state .
the amplification mechanism of the qaser is governed by the difference combination resonance which occurs when the driving field frequency matches the frequency difference between two normal modes of the coupled light atom system .
the present paper links the qaser amplification mechanism and parametric processes that can occur in negative index nonlinear materials and strongly dispersive media .
the difference combination resonance in such systems opens a perspective for development of new kind of light sources that , contrary to conventional opos , emit coherent radiation of tunable wavelengths with the frequency higher than the input frequency .
we thank prof .
wolfgang schleich for useful discussions .
we gratefully acknowledge support of the national science foundation grants phy-1241032 ( inspire creativ ) , phy-1205868 , and the robert a. welch foundation ( awards a-1261 and a-1547 ) .
here we investigate conditions of the combination resonance in nonlinear medium assuming that magnetic permeability @xmath165 and dielectric permittivity @xmath166 depend on coordinates .
this allows us to include nonuniform dielectric structures into consideration . for simplicity
we assume that dispersion is weak which allows us to do calculations in the time domain .
maxwell s equations ( [ me ] ) yield @xmath167we write the total electric field as@xmath168where@xmath169@xmath170@xmath171and @xmath37 is a small gain .
we assume that the driving field frequency @xmath172 satisfies the condition of the combination resonance taking into account that nonlinear susceptibility @xmath105 is small and disregarding the fast oscillating terms we obtain the following equations for @xmath174 and @xmath175@xmath176@xmath177where @xmath178 , @xmath179 and @xmath180 , @xmath181 are the values of @xmath104 and @xmath164 at the frequencies @xmath1 and @xmath2 respectively . multiplying eq .
( [ a1 ] ) by @xmath182 and eq .
( [ a2 ] ) by @xmath183 yields@xmath184@xmath185subtracting these equations from each other , taking into account that @xmath186and integrating over space we find@xmath187@xmath188 = 0 . \label{a0}\]]taking imaginary part of eq .
( [ a0 ] ) , we then obtain@xmath189 = 0 .
\label{a5}\]]here the upper ( lower ) sign corresponds to the sum ( difference ) combination resonance .
( [ a5 ] ) implies that for the difference combination resonance ( lower sign @xmath192 ) the integrals @xmath193 and @xmath194 must have opposite signs . if , e.g. , @xmath195 then eq . ( [ a6 ] )
yields that @xmath196 in some region of space .
thus , we obtain that the difference combination resonance in nonlinear weakly dispersive materials is possible only if for one of the generated frequencies both the magnetic permeability @xmath197 and dielectric permittivity @xmath104 are negative in some spatial volumes .
one should note that calculations presented here assume local ( in time and space ) constitutive relations @xmath198 and @xmath199 .
thus , the results obtained are valid if one can disregard frequency and spatial dispersion . | we demonstrate the possibility of generation of coherent radiation with tunable frequencies higher than the frequency of the driving field @xmath0 in a nonlinear medium utilizing the difference combination resonance that occurs when @xmath0 matches the difference of the frequencies of the two generated fields @xmath1 and @xmath2 .
we find that such a resonance can appear in materials which have opposite signs of refractive index at @xmath1 and @xmath2 .
it can also occur in positive refractive index materials with strong anomalous dispersion if at one of the generated frequencies the group and phase velocities are opposite to each other .
we show that the light amplification mechanism is equivalent to a combination resonance in a system of two coupled parametric oscillators with the opposite sign of masses .
such a mechanism holds promise for a new kind of light source that emits coherent radiation of tunable wavelengths by an optical parametric amplification process with the frequency higher than @xmath3 . |
in collisionally ionized or x - ray photoionized plasmas , high - energy electrons or photons lead to the production of @xmath0-vacancy ionic states which then decay via sequential emission of single or multiple electrons and/or photons .
the exact strengths of these competing processes determine fundamentally important quantities of the plasma such as the ionization balance and the observed spectra of emitted and/or absorbed photons .
hence , interpreting the properties of these plasmas requires accurate atomic physics calculations for the various autoionization and radiative rates .
here we are interested in assessing the accuracy of the available data base that provides such computed ( or inferred ) auger rates and fluorescence yields to the astrophysics community .
the accuracy of these atomic data are crucial to the interpretation of the spectra of photoionized plasmas such as are found in x - ray binaries and active galactic nuclei .
these data are also important for supernova remnants ( snrs ) under conditions of non - equilibrium ionization ( nei ) . two of the more widely used spectral codes for modeling photoionized plasmas are cloudy @xcite and xstar @xcite .
a commonly used code for modeling nei in snrs is that of @xcite .
these all in turn rely on the table of electron and photon emission probabilities compiled by .
this comprehensive data base considers the sequential multiple electron and/or photon ejections for all stages of all @xmath0-vacancy ions in the periodic table up through zinc . in order to produce such a massive array of numbers ,
however , certain approximations , questionable from a purely theoretical atomic physics standpoint , were invoked .
first , the only rigorously computed atomic rates were taken from the early works of @xcite for _ singly - charged ions _ , which furthermore neglected configuration interaction ( ci ) and spin - orbit effects .
due to the limited computational resources available at the time , and the approximations thus needed to perform such calculations , even these can not be considered as reliable as those that can be carried out with today s state - of - the - art capabilities .
second , these singly - ionized results were then scaled along entire isoelectronic sequences , assuming constant autoionization rates and oscillator strengths ; this approximation is least valid for near - neutrals .
a third approximation used by is that the electron and photon emission yields were computed using radiative and autoionization rates that were configuration averaged over possible terms , and the fluorescence yield was then given as a ratio of the averaged radiative rate to the sum of the averaged radiative and averaged auger rate .
for modeling purposes , however , this is incorrect ; the actual required value is the average of the term - specific yields - an average of ratios rather than a ratio of averages . in other words ,
the relative probability of producing each specific inner - shell - vacancy term , and its subsequent term - specific decay , needs to be considered , and this was not done correctly for the data compiled by ( see also @xcite for a further discussion of this importance ) . in this paper
, we investigate the validity of the above three approximations in order to assess the accuracy of the resultant data base of . to this end
, we first study the simplest @xmath0-vacancy system that can radiate via a @xmath2 dipole - allowed transition .
this is the removal of a @xmath0 electron from the ground - state b - like sequence , or rather the @xmath3 be - like inner - shell excited sequence , which is investigated in the next section , and which is further simplified by the fact that only one electron , or one photon , can be emitted .
we follow in section [ sec : f ] with a study of the simplest closed-(outer)shell case of f - like ions , corresponding to @xmath0 vacancies from the ne - like sequence .
a summary of our findings and concluding remarks are then given in section [ sec : conclusion ] .
inner - shell @xmath0 vacancy of a be - like ion , whether by photoionization or electron - impact ionization of b - like ions ( or by photoexcitation or electron - impact excitation of be - like ions ) , results in either the @xmath4 state or the @xmath5 state . from an independent particle perspective ,
in ls coupling , the following competing decay processes can then occur : @xmath6 that is , the @xmath0-vacancy state can either fluoresce , if it is in the @xmath7 state , with a radiative rate @xmath8 , or autoionize , from either state , with a total state - dependent rate @xmath9 , yielding free electrons denoted by @xmath10 .
( if left in the @xmath11 state the ion does not fluoresce - we consider ci and spin - orbit effects in the next section ) .
the radiative rate @xmath8 in atomic units ( 1 a.u .
@xmath12 ) is related to the dimensionless emission oscillator strength @xmath13 by @xmath14 where @xmath15 is the emitted photon energy in a.u .
( 1 a.u . of energy
@xmath16 ev ) and @xmath17 is the fine structure constant .
here we define the emission oscillator strength as the absolute value of the oscillator strength from the upper @xmath18 term @xmath19 to the lower @xmath20 term @xmath21 : @xmath22 and can thus be related to the absorption oscillator strength @xmath23 from the lower term @xmath21 to the upper term @xmath19 via @xmath24 where @xmath25 and @xmath26 are the statistical weights of the initial and final be - like terms , respectively .
oscillator strengths are more convenient quantities to use along isoelectronic sequences because they exhibit certain bounds .
since the absorption oscillator strength is bounded by @xmath27 , where @xmath28 is the number of @xmath0 electrons , the emission oscillator strength is bounded by @xmath29 ( for the present cases ) , and is a well - behaved function of the nuclear charge @xmath1 .
in fact , if the hydrogenic approximation is valid , i.e. , if the nuclear potential dominates over the interelectronic repulsive potential , then the emission oscillator strength is _ independent _ of @xmath1 , and the same is true for the autoionization rate @xmath30 .
such an approximation is valid for highly - charged ions but not for lower - charged species .
the fluorescence yield @xmath31 , from a given inner - shell vacancy state , is a measure of the relative probabilities of the radiative and autoionization decay pathways and is defined as @xmath32}\ \ . \label{eqfy}\end{aligned}\ ] ] thus it only depends on the squared transition energy @xmath33 and the ratio of the autoionization rate to the emission oscillator strength @xmath34 . in the hydrogenic approximation ,
these scale respectively with nuclear charge as @xmath35 and @xmath36 ( i.e. , independent of @xmath37 ) , where @xmath38 is the asymptotic ionic charge seen by the outer - most electron of the be - like ion @xcite . with these scaling properties ,
the expected behaviors at low-@xmath1 and high-@xmath1 are @xmath39 and @xmath40 ( provided @xmath41 ) , respectively . as pointed in section [ sec : be ] , both the @xmath7 and @xmath11 terms can be populated after @xmath0 photoionization or electron - impact ionization .
following @xcite , and using the sudden approximation , we have determined that the probability of populating each term can be deduced by considering the squared recoupling coefficient @xmath422p\left(^2p\right)\left\vert \left[\left(1s2p\right)\left(^{\left\{2{\cal s}+1\right\}}p\right)\right ] 1s\left(^2p\right)\right\rangle\right\vert\right.^2 & = & \left\vert\left(-1\right)^{1+{\cal s}}\left(2{\cal s}+1\right)^\frac{1}{2 } \left\ { \begin{array}{ccc } { \cal s } & \frac{1}{2 } & \frac{1}{2 } \\ 0 & \frac{1}{2 } & \frac{1}{2 } \end{array}\right\}\right\vert^2 \nonumber \\ & = & \frac{2{\cal s}+1}{4}\ , \label{eqrecoup}\end{aligned}\ ] ] where @xmath43 for the @xmath7 state and @xmath44 for the @xmath11 state .
this means that the states are populated according to their statistical weights , and the @xmath7 state is populated with a probability of @xmath45 .
( in general , there also should be a recoupling coefficient involving the orbital angular momenta of the three electrons in eq .
[ eqrecoup ] ; however , the @xmath46 values for two of the electrons orbital momenta reduces the coefficient to unity for the present case . )
we have also verified this computationally by performing r - matrix photoionization calculations using the wigner - eisenbud r - matrix method @xcite . using both approaches
we find that in intermediate coupling , the states are also populated according to their statistical weights ( a similar expression to eq .
[ eqrecoup ] involving the total angular momentum values @xmath19 for each electron can be obtained ) .
considering the relative populations of the @xmath3 @xmath0-vacancy states , the desired quantity for plasma modeling purposes is the configuration - average fluorescence yield .
if ci and spin - orbit effects are neglected , this can be defined as an average over ls single - configuration ( sc ) terms as @xmath47 where fluorescence from the @xmath11 state is zero so that the asymptotic behavior at large @xmath1 is @xmath45 .
ci and spin - orbit effects modify this behavior , however .
the largest ci effect is the intrashell mixing @xmath48 , where the mixing fraction @xmath49 is essentially term independent and @xmath1 independent for nonrelativistic calculations - it varies between 0.067 for and 0.053 for .
this mixing affects the computed emission oscillator strength @xmath13 and autoionization rate @xmath30 at the near neutral end of the sequence , but changes the high-@xmath1 fluorescence yield by less than 10% .
the more important ci effect is that the admixture of the @xmath50 configuration in the @xmath11 term allows it to radiate to the @xmath51 state .
this @xmath52 radiative rate is about a factor of 20 smaller than the @xmath53 rate , so it only increases the fluorescence yield by a few percent at low @xmath1 .
as @xmath1 increases , however , eventually even this reduced radiative rate dominates the autoionization rate , giving @xmath54 the spin - orbit interaction also affects the computed fluorescence yield , primarily by mixing the @xmath55 and @xmath56 levels . the mixing fraction , while only about @xmath57 at @xmath58 , has a @xmath59 dependence , and eventually becomes quite significant , reaching @xmath60 at @xmath61 . as a result ,
the `` @xmath56 '' level ( this is now just a label used to indicate the dominant term of a level ) has an increased fluorescence yield , and we get that the intermediate coupling ( ic ) , configuration - averaged fluorescence yield , including ci , behaves as @xmath62 thus we see that ci and the spin - orbit interaction each cause an increase in the computed fluorescence yield as @xmath1 is increased .
the approach of for this particular be - like series was to neglect spin - orbit and ci effects , and to assume that the hydrogenic approximation is valid throughout the series .
furthermore , they used configuration - averaged values for the autoionization rate and emission oscillator strength , which were computed by @xcite , and the experimental values of @xmath15 from @xcite , to obtain the ratio @xmath63 required for determining @xmath31 using eq .
this is not the same as the desired configuration - averaged fluorescence yield @xmath64 in eq .
[ eqca ] - the ratio of the averages does not equal the average of the ratios : @xmath65 } & \ne & \sum_{{\cal s}=0,1}\left({2{\cal s}+1\over 4}\right)\left[{a_r(^{2{\cal s}+1}p)\over a_r(^{2{\cal s}+1}p)+a_a(^{2{\cal s}+1}p)}\right ] \ . \label{eqne}\end{aligned}\ ] ] we first address the accuracy of the computed autoionization rates and emission oscillator strengths in the next subsection , and then address the validity of the hydrogenic approximation in the following subsection .
fluorescence yields are presented in the last subsection , where the incorrect averaging and neglect of ci and spin - orbit effects by are addressed . in order to calculate the transition matrix elements appearing in the expressions for the radiative and autoionization rates
@xcite , it is first necessary to produce atomic wave functions .
@xcite used the herman - skillman approximation in determining the ( single - configuration ) wave functions , whereby _ all _ electrons ( i.e. , the @xmath0 , @xmath66 , @xmath67 , and continuum ones ) are eigenfunctions of a common central potential ; as stated by @xcite , this `` neglect(s ) ... exchange and correlation effects . '' furthermore , this potential @xmath68 is approximated by `` a series of straight lines '' in order to yield piece - by - piece analytic whitakker functions .
here we are concerned with the validity of these approximations , given that more rigorous calculations can be easily performed using today s state - of - the - art technologies . for the present study , we use the program autostructure @xcite , which generates slater - type @xmath0 , @xmath66 , @xmath67 , and distorted - wave continuum orbitals . in order to compare with the results of @xcite for , and with
as we scale from @xmath58 to @xmath61 , we first performed single - configuration ls calculations . for the more rigorous calculations that we compare to other theoretical results and that we recommend as the definitive data
, we also included ci - @xmath69 for the inner - shell vacancy state and @xmath70 for the final radiative decay state - and spin - orbit effects .
the two accessible continua were described as @xmath71 and @xmath72 , where @xmath10 denotes a continuum distorted wave .
given atomic wave functions , @xcite computed the _ configuration average _ ( ca ) radiative and partial autoionization rates in eqs . 1 - 3 .
the emission oscillator strength given is thus @xmath73 whereas for the total autoionization rate , the ca rates for the processes in eqs .
2 - 3 were used , that is , @xmath74 where @xmath75 ^ 2 \right\}\nonumber\\ & & + \frac{3}{4}\left\ { 2\pi\left[r_0(1s,\epsilon p,2s,2p)\right]^2 \right\ } \ . \label{eqaa1new}\end{aligned}\ ] ] and @xmath76 ^ 2\ , \label{eqaa2new}\end{aligned}\ ] ] since our calculations indicate that @xmath77 . here
@xmath78 is a slater integral of multipole @xmath79 @xcite , and @xmath10 represents the outgoing @xmath80-wave continuum electron orbital .
( the expressions in eqs .
[ eqaa1new ] and [ eqaa2new ] are equivalent to those in eq
. 6 of @xcite for inequivalent electrons and single-@xmath81 orbital occupation , considering the different continuum normalization used by @xcite ) .
note that the partial rate @xmath82 in eq .
[ eqaa1new ] is greatly suppressed relative to the @xmath83 rate due to a near cancellation of monopole and dipole slater integrals .
( indeed , it was due to this near cancellation of slater integrals that @xcite explained why the inner - shell photoexcited @xmath18 resonance in preferentially decayed - by two orders of magnitude - to the @xmath84 channel , compared to the @xmath85 channel . )
thus the configuration average partial rate @xmath86 will be larger than the partial rate @xmath82 , and hence the configuration average total rate @xmath87 will be larger than @xmath88 .
since we are interested in computing @xmath89 , which requires @xmath90 and @xmath91 , we have converted the reported values from @xcite to the @xmath7 values ( the slater integrals were also given in that work ) .
we get the following values @xmath92 which compare fairly well with our results obtained using autostructure : @xmath93 in summary , we find that the earlier results for of @xcite are consistent with ours . however , for the astrophysical plasma modeling purposes we have in mind , one really requires the configuration average fluorescence yield , not the ratio of the averaged radiative and total rates used by @xmath94 due to @xmath8 and @xmath30 scaling as @xmath35 and @xmath36 , respectively , in the hydrogenic approximation .
equation [ eqxikm ] differs from the correct @xmath64 given in eq .
first , we have @xmath95 due to the near cancellation in the @xmath7 @xmath96 partial autoionization rate , so at low @xmath1 , where @xmath97 , we have @xmath98 .
second , when ci and spin - orbit effects are ignored , as they were in , the fluorescence yields differ asymptotically by a factor of 4 , @xmath99 as can be seen by comparing eqs .
[ eqca ] and [ eqxikm ] .
of course , ci needs to be included for all @xmath1 , whereas spin - orbit mixing needs to be included at higher @xmath1 , and both @xmath100 and @xmath101 as @xmath102 .
however , in the intermediate @xmath1 range , it can be shown that the results are still larger than the icci results . in order to assess the validity of scaling the results along the isoelectronic series , we computed both the @xmath7 autoionization rate @xmath30 and emission oscillator strength @xmath13 for all be - like ions up through zinc , first neglecting spin - orbit effects . in fig .
1 , it is seen that neither of the two is independent of the nuclear charge @xmath1 at the lowest stages of ionization - the emission oscillator strength increases by about @xmath103 in going toward the highly - ionized regime whereas the autoionization rate more than doubles .
furthermore , by choosing the scale so that our two quantities coincide for , it is seen that the important ratio @xmath34 appearing in eq .
[ eqfy ] increases by roughly 25% by the time is reached .
thus the assumption of pure hydrogenic scaling by alone introduces an uncertainty at the highly - charged end of this series . due to the stronger @xmath1 dependence at the near - neutral end , together with the greater sensitivity to the atomic basis used in this region
, we recommend that if scaling along an isoelectronic sequence is to be performed , the better starting point would be at the highest @xmath1 desired , extrapolating the rates to lower @xmath1 members .
of course , given the ease of determining atomic rates with modern computing capabilities , the most reliable approach is to calculate the fluorescence yield directly rather than resort to questionable scaling methods .
while the assumption of hydrogenic scaling introduces an @xmath104 inaccuracy in @xmath34 , the initial quantity being scaled in - the ratio of averages rather than the average of ratios - is really not the desired quantity to be scaled in the first place .
together , these approximations lead to an uncertain prediction for the fluorescence yield . in fig . 2 ( and table 1 )
, we compare various results for @xmath31 along the be - like sequence , where it can be seen that our single - configuration ls results differ greatly from those of , especially at higher @xmath1 ; here , especially , their results are expected to differ from the correct single - configuration values due to their incorrect asymptotic value given by eq .
[ eqxikm ] .
a more disturbing result was found when we tried to repeat their calculations , i.e. , when we used eq .
[ eqfy ] , with the ratio of @xmath105 taken from @xcite , and the energies @xmath15 taken from @xcite .
whereas these scaled results exhibit a smooth monotonic increase with nuclear charge @xmath1 , those of are somewhat irregular , showing unphysical dips , and do not agree with what we tried to reproduce , given their stated method .
either way , the results of , or our scaled ones using the results of @xcite , initially underestimate our results at lower @xmath1 , and then overestimate our ( lssc ) results by almost a factor of 3 for the highest @xmath61 . to our knowledge ,
there have been two other calculations for the fluorescence yields of some members of the be - like sequence : those of @xcite using the hullac codes @xcite and those of @xcite using a multiconfiguration dirac - fock ( mcdf ) method . in both cases , ci and spin - orbit effects were included .
here we do the same , first adding the important @xmath106 ci discussed earlier to the ls calculations in order to see that this effect increases the @xmath61 fluorescence yield by about 30% .
then when spin - orbit effects ( and other higher - order , relativistic effects ) are included in our intermediate coupling calculation , there is a further increase in the fluorescence yield by about 20% more . in comparison with the other two calculations along this series
, there is overall good agreement with these ic results .
we turn now to the simplest closed-(outer)shell case of a @xmath0-vacancy in f - like ions , giving the @xmath107 state which decays as @xmath108 again , only one photon , or one electron , can be emitted , which simplifies the analysis considerably ( when spin - orbit effects are considered , the final ionic term in eq . [ eqarf ] is fine structure split into the ground @xmath109 level and the metastable @xmath110 level ) .
since this is a closed - shell system , the herman - skillman method for the important @xmath67 electrons is expected to be more accurate than for . indeed , as stated by @xcite , `` in stripping away electrons ( in reducing to a closed - shell system ) ,
... we should be increasing the applicability of the common central - field approximation . ''
furthermore , there is only one @xmath111 @xmath0-vacancy state , rather than the two we had for the be - like sequence , and no other intrashell configurations to ci mix with , so we do not need to consider population of non - fluorescing states by ci or spin - orbit mixing , nor do we have to consider configuration averaging issues .
consequently , a single configuration ls coupling calculation is sufficient to determine accurate @xmath30 , @xmath13 , and @xmath31 values for the @xmath112 term . as a result
, the computed values of the autoionization rate and emission oscillator strength given by @xcite agree quite well with our values , as seen in fig . 3 and table 2 . however , both of these values depend on the internuclear charge @xmath1 , giving a ratio @xmath34 that increases by about a factor of @xmath113 in going from to .
thus the scaled fluorescence yield @xmath31 , using eq .
[ eqfy ] , the ratio @xmath34 from @xcite , and @xmath15 from @xcite , increases relative to the actual computed value , as is seen in fig . 4 and table 2 .
the more troublesome news in this figure is the actual tabulated values of - their values do not follow our attempt at reproducing those results , but rather tend to follow our computed values , except for certain unphysical dips .
nevertheless , the results reported by for f - like ions are not plagued by as many uncertainties as those for be - like ions .
we also see in fig . 4 that the hullac results are in good agreement with our present ones ( the results reported earlier by @xcite only considered fluorescence into the @xmath109 level , which includes only 4 of all 6 magnetic sublevels of the @xmath114 configuration ; therefore , those values must be multiplied by about 3/2 to account for fluorescence into the two @xmath110 sublevels as well .
furthermore , the earlier hullac result for f@xmath115 was erroneously listed incorrectly , and here we have given the actual computed value that should have appeared ) .
the inaccuracies we have discovered in the reported results of for be - like ions are as follows : 1 . the computed atomic data for are used in the form @xmath105 , that is
, the radiative and autoionization rates have been averaged over the @xmath7 and @xmath11 configurations , whereas the desired quantity for plasma modeling applications is @xmath116 and is not the same thing , differing qualitatively and quantitatively , especially in the asymptotic high-@xmath1 limit .
the hydrogenic scaling assumed is invalid .
the autoionization rates , the emission oscillator strengths , and even the ratio @xmath34 are not independent of nuclear charge @xmath1 .
3 . the tabulated data of do not seem to follow the results we obtain when we try to reproduce their stated method using eq .
[ eqfy ] , with @xmath105 from @xcite , and @xmath15 from @xcite .
4 . the calculations of neglected ci and spin - orbit effects as they scaled to higher @xmath1 . for f - like ions , items 1 and 4 are not issues since there is only one inner - shell vacancy term .
however , points 2 and 3 still apply for the f - like sequence . for plasma modeling purposes ,
we recommend our @xmath116 for the be - like sequence and our @xmath31 for the f - like sequence . in conclusion , we propose that , given the many uncertainties discovered , the entire data base of should be reevaluated . while we have focused on systems that can emit only one photon or one electron ,
their comprehensive tabulation also includes data for ions with @xmath117 shells occupied ; these can emit multiple electrons and/or photons through numerous cascading channels , compounding the inaccuracies we have discovered .
we would like to thank e. j. mcguire for careful reading of , and extremely helpful comments on , an earlier version of this manuscript .
twg , cnk , ktk , and oz were supported by nasa space astrophysics research and analysis program grant nag5 - 10448 .
eb was supported by the yigal - alon fellowship and by the gif foundation under grant # 2028 - 1093.7/2001 .
the work of mhc was performed under the auspices of us department of energy by the university of california , lawrence livermore national laboratory , under contract number w-7405-eng-48 .
dws was supported in part by nasa space astrophysics research and analysis program grant nag5 - 5261 and nasa solar physics research , analysis , and suborbital program grant nag5 - 9581 .
rllcccccccc 5 & 0.1519 & 0.1045 & 0.0014 & 0.0006 & 6.751 & 0.0006 & & & 0.0011 & 0.0011 + & 0.1508 & 0.1692 & & & & + 6 & 0.1712 & 0.1194 & 0.0032 & 0.0019 & 10.349 & 0.0013 & & 0.0024 & 0.0025 & 0.0025 + 7 & 0.1859 & 0.1328 & 0.0061 & 0.0052 & 14.721 & 0.0027 & & 0.0045 & 0.0048 & 0.0048 + 8 & 0.1972 & 0.1442 & 0.0106 & 0.0096 & 19.866 & 0.0049 & & 0.0079 & 0.0083 & 0.0083 + 9 & 0.2062 & 0.1540 & 0.0168 & 0.0154 & 25.753 & 0.0081 & & 0.0128 & 0.0132 & 0.0133 + 10 & 0.2134 & 0.1622 & 0.0250 & 0.0229 & 32.379 & 0.0128 & 0.0209 & 0.0191 & 0.0199 & 0.0201 + 11 & 0.2193 & 0.1693 & 0.0352 & 0.0352 & 39.782 & 0.0192 & & & 0.0285 & 0.0287 + 12 & 0.2243 & 0.1754 & 0.0474 & 0.0424 & 47.924 & 0.0276 & 0.0414 & 0.0377 & 0.0390 & 0.0393 + 13 & 0.2285 & 0.1808 & 0.0612 & 0.0484 & 56.806 & 0.0384 & 0.0538 & & 0.0514 & 0.0518 + 14 & 0.2320 & 0.1854 & 0.0761 & 0.0768 & 66.465 & 0.0518 & 0.0685 & & 0.0653 & 0.0658 + 15 & 0.2351 & 0.1896 & 0.0916 & 0.1102 & 76.862 & 0.0681 & & & 0.0805 & 0.0812 + 16 & 0.2378 & 0.1933 & 0.1073 & 0.1446 & 88.034 & 0.0874 & 0.0984 & & 0.0965 & 0.0974 + 17 & 0.2402 & 0.1965 & 0.1225 & 0.1656 & 99.982 & 0.1100 & & & 0.1129 & 0.1141 + 18 & 0.2423 & 0.1995 & 0.1369 & 0.1671 & 112.664 & 0.1357 & 0.1273 & 0.1237 & 0.1295 & 0.1309 + 19 & 0.2442 & 0.2022 & 0.1502 & 0.1626 & 126.122 & 0.1644 & & & 0.1458 & 0.1478 + 20 & 0.2459 & 0.2046 & 0.1623 & 0.1984 & 140.348 & 0.1958 & 0.1569 & & 0.1616 & 0.1646 + 21 & 0.2475 & 0.2068 & 0.1732 & 0.2963 & 155.342 & 0.2298 & & & 0.1769 & 0.1813 + 22 & 0.2488 & 0.2089 & 0.1828 & 0.3438 & 171.107 & 0.2658 & & & 0.1916 & 0.1982 + 23 & 0.2501 & 0.2108 & 0.1912 & 0.3838 & 187.645 & 0.3033 & & & 0.2058 & 0.2154 + 24 & 0.2513 & 0.2125 & 0.1985 & 0.4214 & 204.991 & 0.3419 & & & 0.2194 & 0.2333 + 25 & 0.2523 & 0.2141 & 0.2049 & 0.4562 & 223.108 & 0.3810 & & & 0.2327 & 0.2518 + 26 & 0.2533 & 0.2156 & 0.2105 & 0.4903 & 241.998 & 0.4200 & 0.2394 & 0.2633 & 0.2457 & 0.2713 + 27 & 0.2542 & 0.2169 & 0.2153 & 0.5267 & 261.659 & 0.4584 & & & 0.2585 & 0.2916 + 28 & 0.2551 & 0.2182 & 0.2194 & 0.5836 & 282.129 & 0.4960 & & & 0.2712 & 0.3125 + 29 & 0.2559 & 0.2194 & 0.2230 & 0.6215 & 303.333 & 0.5322 & & & 0.2840 & 0.3339 + 30 & 0.2566 & 0.2205 & 0.2261 & 0.6322 & 325.310 & 0.5668 & & & 0.2969 & 0.3553 + rllccccc 10 & 0.2159 & 0.1056 & 0.0147 & 0.0182 & 31.184 & 0.0169 & 0.0215 + & 0.216 & 0.0948 & & & & & + 11 & 0.2286 & 0.1164 & 0.0214 & 0.0263 & 38.723 & 0.0258 & + 12 & 0.2406 & 0.1273 & 0.0298 & 0.0346 & 47.035 & 0.0376 & 0.0380 + 13 & 0.2515 & 0.1377 & 0.0402 & 0.0397 & 56.081 & 0.0526 & 0.0493 + 14 & 0.2615 & 0.1477 & 0.0528 & 0.0449 & 65.864 & 0.0712 & 0.0630 + 15 & 0.2705 & 0.1571 & 0.0679 & 0.0634 & 76.422 & 0.0936 & + 16 & 0.2786 & 0.1659 & 0.0855 & 0.0875 & 87.720 & 0.1197 & 0.0983 + 17 & 0.2859 & 0.1741 & 0.1058 & 0.1019 & 99.795 & 0.1497 & + 18 & 0.2926 & 0.1817 & 0.1286 & 0.1305 & 112.646 & 0.1832 & 0.1443 + 19 & 0.2987 & 0.1888 & 0.1540 & 0.1253 & 126.276 & 0.2199 & + 20 & 0.3042 & 0.1954 & 0.1818 & 0.1505 & 140.682 & 0.2592 & 0.2001 + 21 & 0.3093 & 0.2016 & 0.2118 & 0.2073 & 155.863 & 0.3004 & + 22 & 0.3139 & 0.2073 & 0.2437 & 0.2411 & 171.820 & 0.3429 & + 23 & 0.3182 & 0.2127 & 0.2771 & 0.2751 & 188.552 & 0.3860 & + 24 & 0.3221 & 0.2177 & 0.3116 & 0.3068 & 206.093 & 0.4289 & + 25 & 0.3258 & 0.2224 & 0.3469 & 0.3386 & 224.395 & 0.4710 & + 26 & 0.3291 & 0.2269 & 0.3825 & 0.3692 & 243.504 & 0.5118 & 0.4041 + 27 & 0.3324 & 0.2309 & 0.4180 & 0.3942 & 263.386 & 0.5513 & + 28 & 0.3353 & 0.2348 & 0.4531 & 0.4438 & 284.040 & 0.5883 & + 29 & 0.3380 & 0.2385 & 0.4874 & 0.4734 & 305.465 & 0.6230 & + 30 & 0.3406 & 0.2420 & 0.5207 & 0.4758 & 327.735 & 0.6554 & | we have investigated the accuracy of the @xmath0-vacancy fluorescence data base of resulting from the initial atomic physics calculations and the subsequent scaling along isoelectronic sequences . in particular ,
we have focused on the relatively simple be - like and f - like @xmath0-vacancy sequences .
we find that the earlier atomic physics calculations for the oscillator strengths and autoionization rates of _ singly - charged _ and are in sufficient agreement with our present calculations .
however , the substantial charge dependence of these quantities along each isoelectronic sequence , the incorrect configuration averaging used for , and the neglect of spin - orbit effects ( which become important at high-@xmath1 ) all cast doubt on the reliability of the data for application to plasma modeling . |
many different mechanisms have been suggested to explain the observed late - time acceleration of the expansion of the universe @xcite .
differentiating between these options is one of the main challenges facing cosmologists today .
we can try to build up the evidence for different mechanisms by examining the evolution of the universe in two key ways : measuring the background geometry and measuring structure formation within it .
the geometrical evolution of the universe can most easily be measured using two primary techniques : we can use supernovae as standard candles or galaxy clustering as a standard ruler to make precise measurement of cosmological expansion . although supernovae were first to confirm the accelerated expansion of the universe at high statistical significance @xcite following analyses of early galaxy surveys @xcite , supernovae surveys are now limited by systematic rather than statistical errors @xcite .
the possibility of using galaxy clustering to provide a standard ruler has become increasingly important since the baryon acoustic peak was detected @xcite in galaxy power spectra measured from the 2df galaxy redshift survey ( 2dfgrs ; @xcite ) and the sloan digital sky survey ( sdss ; @xcite ) .
using only the baryon acoustic oscillation ( bao ) component of the galaxy clustering signal makes constraints robust to non - linear effects , and has already been exploited to produce interesting constraints on cosmological models @xcite .
galaxy surveys provide complementary information about the build - up of large - scale structure through redshift space distortions ( rsd ) .
these arise because we do not observe true galaxy positions , but instead infer distances from measured redshifts .
coherent comoving galaxy velocities due to the growth of structure therefore lead to measurable anisotropic clustering @xcite .
rsd have been measured from the 2dfgrs and sdss using techniques based on both correlation functions and power - spectra @xcite , and have recently been detected at higher redshift @xcite .
in fact , for fundamental reasons , rsd are independent of galaxy bias : galaxies act as test particles in the matter flow , so their motion is independent of galaxy properties .
we can therefore measure the matter velocity field at the locations of the galaxies and , provided that galaxy positions are representative within the velocity field , this gives an unbiased measurement of @xmath5 where @xmath6 , the logarithmic derivative of the linear growth rate , @xmath7 , with respect to the scale factor @xmath8 , and @xmath9 quantifies the amplitude of fluctuations in the matter density field .
many techniques have been proposed for measuring rsd , including using multipoles of the correlation function @xcite , or the power spectrum @xcite , based on the plane - parallel approximation for rsd @xcite .
the geometry of the system can be easily incorporated into an analysis of the correlation function as the separation of each pair can be split into radial and angular components .
it is more difficult to perform such a decomposition when measuring the power spectrum , although various techniques have been suggested , decomposing the density field into a basis of spherical harmonics and radial functions ( e.g. @xcite ) .
as well as allowing for the geometry to a survey and rsd within it , we need to allow for two effects : 1 .
rsd are degenerate with the angular anisotropy of power spectrum caused by the alcock - palczinski effect , so measurements of growth and geometry from the same survey will be correlated , and it is sensible to undertake a combined analysis @xcite .
as noted by @xcite , the full rsd operator has extra terms compared to the plane - parallel one and results in a redshift - space power spectrum which is not diagonal in @xmath10 .
these extra terms are important for galaxy pairs separated by wide angles and need to be included in any analysis that uses wide - angle data @xcite .
in this paper we concentrate on the second of these effects . @xcite and
@xcite proposed a method of computing the redshift - space correlation function which does not assume a plane - parallel approximation and is applicable to arbitrarily large angles . in their treatment
the redshift - space correlation function depends not only on the redshift - space pair separation @xmath11 and cosine of the angle of the pair with respect to the line - of - sight @xmath12 , but also on the separation angle between two galaxies in a pair @xmath13 , which does not have to be small .
they were able to express the redshift - space correlation function explicitly trough a real - space correlation function and cosmological parameters . given the precision that future surveys will achieve and the fact that they will cover large fractions of the sky , these effects will need to be corrected .
the wide - angle effects can be subdivided in `` purely wide - angle '' and `` mode - coupling '' terms : `` purely wide - angle '' effects correct plane - parallel predictions accounting for the fact that the separation angle is non - zero , `` mode - coupling '' terms in addition account for the fact that galaxy pairs coherently moved from the high - density to low - density regions .
these latter terms vanish if the initial real - space distribution of galaxies is uniform in distance , and both terms vanish in the plane - parallel limit due to the symmetry of the system .
both terms are of the same order and mode - coupling tends to smooth out features in the power spectrum such as the baryon acoustic oscillations ( bao ) .
given the importance and , particularly , the quality of rsd constraints expected from forthcoming surveys such as the baryon oscillation spectroscopic survey ( boss ; @xcite ) , surveys resulting from proposals for wide - field multi - object spectrographs on 4 m telescopes such as bigboss @xcite and satellite missions such as euclid @xcite , it is timely to revisit this problem , and to test wide - angle rsd theory .
in this paper we use numerical simulations to extensively test the dependence of rsd constraints on the angular separation of galaxy pairs .
this paper is organised as follows : in sec .
[ sec : rsd ]
we briefly review the theory of wide - angle rsd ; in sec .
[ sec : rsdmeas ] we describe a time - efficient method of getting a low - noise measurement of wide - angle correlation function from a mock hv catalog ; in sec .
[ sec : results ] we present our results and discuss all the steps necessary to match measured correlation function with theoretical predictions ; in sec .
[ sec : disc ] we conclude and discuss how the results will affect real survey measurements .
galaxy positions are measured in a redshift - space , and differ from the real - space positions because of the contributions from peculiar velocities , @xmath14 where @xmath15 is the redshift space position , @xmath16 is the real space position and @xmath17 the radial component of peculiar velocity .
since the total number of galaxies in real and redshift spaces is the same , the number @xmath18 of galaxies observed in a volume element @xmath19 of redshift space is related to the real space number density @xmath20 by : @xmath21 the observed redshift space galaxy overdensity @xmath22 at position @xmath15 can be related to the redshift - space selection function @xmath23 , @xmath24 in contrast , an unbiased estimate of the true galaxy overdensity @xmath25 at position @xmath16 is given by : @xmath26 where @xmath27 is the expected galaxy distribution in real - space the expected number density of unclustered galaxies at position @xmath16 given the selection criteria of the survey .
we distinguish here between the measured redshift - space selection function @xmath28 and the true selection function @xmath27 even though the two are the same at linear order .
we now assume that the expected density is a function of the radial component of the position only , which we denote @xmath29 in redshift - space and @xmath30 in real - space .
thus the relation between the observed redshift space overdensity @xmath31 and the true overdensity @xmath32 is : @xmath33 s^2 ds = \bar{n}^r(r ) [ 1+\delta^r({\mathbf{r } } ) ] r^2 dr.\ ] ] using eq .
( [ eq : nd ] ) we obtain : @xmath34 \left(1 + \frac{\partial v}{\partial r } \right)^{-1 } \\
\nonumber & \left(1 + \frac{v}{r } \right)^{-2 } \frac{\bar{n}(r)}{\bar{n}(r+v\hat{r})},\end{aligned}\ ] ] which gives , to linear order : @xmath35 where @xmath36 usually the @xmath37 term in the jacobian is omitted because , in the linear regime , it gives rise to a @xmath38 term , that would tend to zero at large distances .
however , @xcite have argued that , for wide angles , the @xmath39 term is of the same order as the @xmath40 term .
using the exact jacobian , and following @xcite , we can express the linear overdensity as : @xmath41 \delta(k),\ ] ] so the correlation function is : @xmath42 \nonumber \\ & \left [ 1 + \frac{f}{3 } + \frac{2f}{3 } l_2(\hat{r}_2 \hat{k } ) + \frac{i\alpha f}{r_2 k } l_1(\hat{r}_2 \hat{k } ) \right ] , \label{eq : xi1}\end{aligned}\ ] ] where @xmath43 are the legendre polynomials of order @xmath44 and @xmath45 is the real - space matter power spectrum .
the third terms in the brackets are the ones responsible for the wide - angle effects , while the fourth terms are the ones responsible for the mode - coupling .
note that the @xmath46 and @xmath47 terms in the denominator depend on the angular separation of the galaxies . setting @xmath48 as the angle between the vector to the first galaxy in a pair @xmath49 , @xmath50 to be the vector connecting galaxies in a pair , and @xmath51 to be the angle between vector to the second galaxy in a pair @xmath52 and @xmath50
, we can use the sine rule ( see fig .
[ fig : triangle ] ) to express @xmath46 and @xmath47 as @xmath53 tripolar spherical harmonics are the most natural basis for the expansion of a function that depends on three directions @xcite so , as suggested in @xcite and @xcite , we expand eq .
( [ eq : xi1 ] ) using a subset of them that are proportional to the zero angular momentum : @xmath54 where @xmath55 are normalized spherical functions , multiplied by the @xmath56 wigner symbol . the redshift space correlation function
is then written as : @xmath57 where @xmath58 are a series of coefficients that depend on @xmath59 , @xmath60 and @xmath61 , and they can be divided into wide - angle components ( given in @xcite & @xcite ) that do not depend on the third term in eq .
( [ eq : del ] ) , and mode - coupling components ( given in @xcite ) ; the plane - parallel approximation emerges as a limit when @xmath62 .
, and galaxies @xmath63 and @xmath64 .
[ fig : triangle ] ] as described in @xcite , for any pair of galaxies we can fully analyse the problem within a plane formed by the two galaxies and the observer ; the effect of redshift - space distortions only depends on the geometry within this plane . in the coordinate system described above , the vectors connecting
the observer with the galaxies are @xmath49 and @xmath52 , and the separation vector between the galaxies is @xmath50 ; @xmath65 are the angles between @xmath66 and @xmath50 . in the plane
parallel approximation @xmath67 .
this description of the triangle is shown in fig .
[ fig : triangle ] . at this point
we still have the freedom to choose a line - of - sight to the galaxy _ pair _ : we follow the standard approach and choose the line of sight as a direction bisecting the angle formed by the two galaxies , which we assume is given by @xmath68 .
the angle between the angle bisector and @xmath50 is denoted @xmath69 , and we define @xmath70 .
we therefore have that @xmath71 , and @xmath72 .
this is shown in fig .
[ fig : triangle ] . with this choice of coordinate system the redshift - space correlation function can be expanded as : @xmath73 with coefficients given in @xcite .
( [ eq : xiwa ] ) reduces to the plane - parallel result @xcite in the limit of @xmath74 .
a standard method for analysing a n - body simulation to test wide - angle redshift - space distortions would be to : 1 .
locate an observer within the output , 2 .
translate all galaxies from real into redshift - space based on this observer , 3 .
sample from these galaxies based on desired radial distribution , 4 .
split pairs into bins in @xmath69 , @xmath13 , @xmath11 and counts pairs , 5 . estimate the correlation function .
this approach mimics that of an actual survey analysis , creating a mock galaxy catalogue .
however , if we want to measure the effect of redshift - space distortions for galaxy pairs with a particular angular separation , the method is not optimal as only a small fraction of the galaxy pairs analysed will have this angular separation .
in addition , we have to perform the full procedure for every radial galaxy distribution that we wish to analyse , and this distribution will limit the pair - density that can be selected from the simulation . in order to rapidly increase the signal - to - noise , we adopt a different procedure , allowing the origin to move so that each galaxy pair can be analysed as if it was observed with the required angular separation
. this procedure can be summarised as follows : 1 .
decide on the value of @xmath13 for which we wish to analyse pairs , 2 .
take each galaxy pair from the simulation with real - space separation @xmath75 , 3 . for each pair
randomly choose @xmath76 $ ] , 4 .
choose the location of the origin giving this @xmath12 and @xmath13 , 5 .
move galaxies according to their expected redshift - space distortion , 6 .
weight the pair by a function of @xmath77 to match desired distribution , 7 .
split pairs into bins in @xmath69 , @xmath11 and counts pairs , 8 .
estimate the correlation function .
the added complexity of including redshift - space distortions on a pair - by - pair basis is outweighed by the ability to obtain more pairs at the desired angular separation .
note that @xmath12 in step ( iii ) is constrained and can not have any value within @xmath78 $ ] because @xmath79 results in geometrically impossible triangles ( see , fig .
[ fig : triangle ] ) .
the distributions of triangles in @xmath12 and @xmath11 for a galaxy distribution with a power law selection in @xmath30 can be calculated analytically , as we now demonstrate .
the redshift - space correlation function in eq .
( [ eq : xiwa ] ) depends on the selection function @xmath80/\partial r$ ] , where @xmath81 is the number density of galaxies in real space and @xmath16 is the position of the galaxy from the observer as in section [ sec : rsd ] . if the galaxy distribution is uniform ( @xmath82 , @xmath83 , both independent of @xmath84 ) , then the probability of finding a galaxy in some region is proportional to the volume of that region .
therefore if we randomly pick a galaxy in the catalog , there will be on average @xmath85 galaxies in a small volume element @xmath86 which is distance @xmath87 away from that galaxy within an angular slice @xmath88 . in the plane - parallel approximation , where all galaxies are assumed to lie along the same direction from the observer ,
the distribution of galaxy pairs will scale as @xmath89 where @xmath90 is the number of pairs with separation @xmath91 and cosine of the angle with the line of sight in the @xmath92 interval .
if we pick galaxy pairs separated by a fixed opening angle from the uniform spatial distribution of galaxies ( step ( i ) in the procedure described above ) , for a large enough angle the distribution of pairs will not follow eq .
( [ eq : dn ] ) . for galaxy pairs with a fixed half - angle @xmath13
( see , fig .
[ fig : triangle ] ) , the likelihood of finding one galaxy at position @xmath93 and another at @xmath94 is @xmath95 , @xmath96 .
since the two likelihoods are independent , the joint probability of finding that pair of galaxies is : @xmath97 using eqs .
( [ eq : g1 ] ) and ( [ eq : g2 ] ) we can rewrite eq .
( [ eq : nxx ] ) in terms of @xmath69 and @xmath11 as : @xmath98 after completing steps ( ii ) and ( iii ) in the previous subsection , we will have a distribution of galaxy pairs that is uniform in @xmath99 and scales as @xmath100 with separation .
this distribution of galaxy pairs does not correspond to @xmath101 . if we want to compare our data with the correlation function computed from eq .
( [ eq : xiwa ] ) for @xmath101 , we have to weight our galaxy pairs ( step ( iv ) in the previous subsection ) by an additional factor of @xmath102 to get a distribution given by eq .
( [ eq : dnwa ] ) . this procedure can be applied to the case of arbitrary @xmath103 .
if , for example , the galaxy number density scales as a power - law @xmath104 , then @xmath105 , and the distribution of galaxy pairs with a fixed opening half - angle drawn from this distribution will be : @xmath106 we therefore weight all pairs recovered from the simulation with a weight given by : @xmath107 where @xmath108 is the separation of the galaxies in redshift - space , @xmath109 is the real - space value of @xmath69 for the chosen galaxy pair and @xmath110 is the redshift - space value .
we divide the real - space distribution by the redshift - space equivalent in order to normalise the weights to give no net change in the expected pair distribution .
when transformed into redshift space , the real - space galaxy distribution is `` washed out '' by the random component of galaxy velocities , so we need to convolve with the random velocity distribution .
this would normally not be a problem as we would estimate the galaxy distribution directly in redshift - space .
however , using the above procedure we have set the real - space distribution of galaxies , so we need to take care when modelling the redshift - space distribution . to correct for this effect
, we first estimate the velocity dispersion @xmath111 from the catalog , assuming that the random velocities are drawn from a gaussian distribution .
we then convolve the initial distribution in real space with this gaussian .
for example , for @xmath112 , the probability of finding a galaxy at distance @xmath30 away from the observer is @xmath113 . in the unclustered redshift space this will transform into : @xmath114,\end{aligned}\ ] ] where f is a hypergeometric function . for a specific case of @xmath115 , eq .
( [ eq : ps ] ) results in : @xmath116 where @xmath117 is an error function .
the distribution of galaxy pairs for @xmath115 will be : @xmath118 where @xmath119 is given by eq .
( [ eq : psa2 ] ) and @xmath120 and @xmath121 can be expressed in terms of @xmath15 and @xmath69 via eqs .
( [ eq : g1 ] ) and ( [ eq : g2 ] ) , after the substitution of @xmath122 with @xmath123 . , and @xmath115 .
open circles ( with 1@xmath124 errors ) were calculated using the redshift - space model for the galaxy distribution .
open diamonds used the real - space model .
the solid line shows the prediction from section [ sec : rsd ] .
[ fig : xi_cmpr_gauss ] ] when modelling the theoretical correlation function , @xcite argue that the difference between @xmath125 and @xmath126 is small .
while this is true when creating a model correlation function , it is not true , and we need to use the @xmath125 instead of @xmath127 , when we model the data .
this is shown in fig .
[ fig : xi_cmpr_gauss ] , where we plot @xmath128 calculated for @xmath115 from a mock catalog , assuming that the galaxy distribution follows the one expected from either the real - space or redshift - space calculation .
see section [ sec : results ] for more details of the calculation . .
the shaded circles around each galaxy are exclusion zones within which the redshift - space distortions is comparable with the distance to the galaxy . o1 and o2 mark two possible positions for an origin that would give angular galaxy separation of @xmath129 . [
fig : excl ] ] galaxies that are close to the origin can cause problems as , for such galaxies , the redshift - space displacement can be larger than the distance to the galaxy .
this problem is exacerbated because we do not include the velocity of the observer in our calculation . in extreme situations , naively applying the expected redshift - space distortion would place the galaxy on the opposite side of an observer . in order to avoid such problems ,
we only include galaxy pairs where both galaxies are more than @xmath130 away from the origin , where @xmath111 is the 1d velocity dispersion of the galaxy population .
for the hv simulation @xmath131 .
[ fig : excl ] shows , for galaxies @xmath63 and @xmath64 , the loci of positions at which the origin could be placed for fixed @xmath13 .
the circles mark the exclusion zones .
if one of the galaxies in a pair is inside the exclusion zone in real space , we do not include that pair when we estimate the correlation function .
this exclusion is tracked when we calculate the expected galaxy distribution .
we apply the procedure outlined in sec . [
sec : rsdmeas ] to analyse wide - angle redshift - space distortions within the @xmath132cdm hubble volume ( hv ) simulation @xcite .
the @xmath132cdm hv simulation , covering a @xmath133 box , assumes a cosmological model with @xmath134 , @xmath135 , @xmath136 , @xmath137 , @xmath138 , @xmath139 , & @xmath140 .
we do not apply any galaxy bias , simply poisson sampling the matter particles to give our `` galaxy distribution '' ; the inclusion of a bias model would not alter the conclusions of this work .
we use the periodic nature of the numerical simulation to eliminate boundaries from our pair counts .
this means that , by using the above weighting scheme and allowing for the removal of galaxies close to the origin , the expected number of galaxy pairs in the absence of clustering @xmath141 can be calculated analytically .
we can therefore use the natural estimator @xmath142 @xcite , where @xmath143 is the measured number of galaxy - galaxy pairs .
we have performed 100 runs , each based on a sample of @xmath144 galaxies drawn from the @xmath145 output of the hv @xmath132cdm simulation . for each unique pair of galaxies within the sample
we have selected two locations for the origin , one where the galaxies subject a separation angle of @xmath146 , chosen to match figure 1 of @xcite , and one origin where @xmath147 , twice this angular separation , to emphasise the wide angle effects on pairs separated by a very wide angle .
we include all pairs with redshift - space separation @xmath148 , and each pair was weighted as described in section [ sec : rsdist ] , based on its real - space separation and angle to the line of sight .
given that , as one can see from eq .
( [ eq : xi1 ] ) , the mode - coupling terms depend strongly on the radial galaxy distribution , for both the aperture angles we selected samples with different radial galaxy distributions , in order to test our methodology in different cases ; we chose @xmath149 : * @xmath150 corresponds to the case of a galaxy distribution that has equal density in radial bins of equal width . in this case
there are no mode - coupling terms in eq .
( [ eq : xi1 ] ) , so this corresponds to the `` pure wide angles '' case .
* @xmath151 corresponds to a final real - space distribution of galaxy pair separations with probability density function proportional to @xmath152 .
this is the same distribution obtained by randomly sampling pairs from the simulation , so the @xmath11 factor drops out of eq .
( [ eq : weight ] ) .
note that , for clarity , we do not present results from this value of @xmath103 as they are very similar to , and overlap results for @xmath150 .
the match between data and theory has the same quality as for the other values of @xmath103 .
* @xmath115 matches a distribution of galaxies that are uniformly distributed in volume , so equal volumes contain equal numbers of galaxies .
most planned surveys aim to observe galaxies with this radial distribution .
note that , for this galaxy distribution , the pair distribution goes as @xmath153 .
* @xmath154 represents a steeper radial distribution where more galaxies are found at larger distances .
this will increase the effect of the mode - coupling terms in the wide - angle redshift - space distortion formulae ( eq . [ eq : weight ] ) .
we have calculated errors by comparing outputs from 100 subsamples , that cover the same hv volume and so only contain the shot noise element .
these errors will therefore underestimate the true error , because they do not fully include the cosmic variance component .
however , the hv volume is @xmath133 , and we only consider pairs with @xmath155 , so we expect the shot noise to dominate the error budget .
even so , it is worth pointing out that our primary aim is to consider deviations from plane - parallel theory and that our proposed methodology works to match data with the full mode - coupling theory : the size of the errors is unimportant , provided they are far smaller than the differences between theories .
[ fig : xi_cmpr_papai ] shows the correlation function , calculated within a narrow bin in @xmath12 for @xmath146 and @xmath115 ; this is equivalent to figure 1 of @xcite .
we are able to fit the theoretical correlation function to the estimate from hv simulations for scales larger then @xmath156 ( we do not expect to match perfectly data with theory on smaller scales because we did nt model non - linearities ) . looking at plane - parallel , pure wide - angle and full mode - coupling theories , it is clear that only the full mode - coupling theory provides a good fit to the data .
[ fig : ratio ] shows the ratio between the correlation function predicted by the full mode - coupling theory and that for the plane - parallel case , as a function of @xmath13 , for a fixed value of @xmath157 , with @xmath115 ; from this plot we can see that even for small angles there is a non negligible difference . in order to analyse rsd , the measured correlation function is usually decomposed into legendre momenta , which contain all of the rsd signal @xcite .
some combination of these momenta is then used to constrain cosmological parameters through their effects on the growth of structure . to see how wide angle effects would modify these measurements we estimate first three even legendre momenta of the correlation function from hv mock catalog .
we measure : @xmath158 for @xmath159 . adopting the plane - parallel philosophy @xcite
, one might be tempted to interpret these functions as : @xmath160 where @xmath161 is a real space correlation function and @xmath162 .
in fact the functions in eq .
( [ eq : xitildadata ] ) are given by : @xmath163 where @xmath164 is a weight function that appears because of geometrical constraints and because the distribution of @xmath12 will not be uniform .
we do not correct for this weight in our measurement from simulations , and apply it to the theoretical calculations that we compare against .
+ + + fig .
[ fig : xil_cmpr ] presents the main results of this work , showing the application of our methodology to sample and analyse data , and comparing against the full mode - coupling predictions .
we plot @xmath128 , @xmath165 , @xmath166 for both the aperture angles and for radial galaxy density distributions corresponding to @xmath150 , @xmath115 and @xmath154 ; as we can see , the contribution of the mode coupling terms increase with @xmath103 , and as the radial galaxy distribution steepens , more galaxies are moved nearer by the rsd , leading to an increase in the small - scale correlation function .
momenta of the correlation function computed with the full mode coupling theory match remarkably well the data analysed as explained in sec .
[ sec : rsdmeas ] .
we also plot the plane - parallel prediction as a comparison , and this demonstrate how badly this approximation fails for galaxy pairs with wide angular separation .
the only parts where our methodology do not match the mode coupling theory are where the correlation multipoles are very small ( around @xmath167 ) or for small scales , where non - linearities become important ; we also have to take in account that , for the steeper galaxy distributions , we are upweighting results from the same number of pairs as the more shallow galaxy distributions so , although the signal is stronger , the relative error will be the same .
redshift - space distortion analyses are a powerful tool for cosmology , but incoming data need to be analysed very carefully in order to fully extract all available information . until now rsd analyses have concentrated on using the plane - parallel approximation ( with a few exceptions including @xcite ) ; this is almost correct if the survey is narrow , but when galaxy surveys with wide field of view data will be available , there will be a consistent number of galaxy pairs separated by wide angles and in this case the plane - parallel approximation fails .
this approximation arises from not taking in account some of the terms in the jacobian relating redshift- to real - space ; in this paper we considered its exact expression , and this causes additional , non - diagonal terms in the correlation function .
we then use the expansion of the correlation function in a base of tripolar spherical harmonics , as suggested in @xcite ; following the formalism developed in @xcite , we make predictions for the momenta of the correlation function for galaxy pairs at fixed angular separation , and we test them against data from the hubble volume simulation . in order to do this we have had to introduce a new methodology : rather than creating a single sample of galaxies in redshift - space from which we can count pairs , we have instead dynamically applied rsd on a pair - by - pair basis , choosing an origin for each . by including weighting functions in @xmath11 and @xmath12 we can match results from the more traditional approach .
this allows us to only consider galaxy pairs at particular values of @xmath13 . to show both the correctness of our methodology and the deviation from the plane - parallel situation we tested galaxies separated by two different fixed values of @xmath13 ; the mode coupling terms are also strongly dependent on the galaxy radial distribution
, so we have tested simulations with 4 different number density distributions , for both values of @xmath13 .
we show that taking in account the wide angle and mode coupling terms ( that are of the same order , as stated in @xcite ) give a clear deviation from the plane - parallel theory ; using the exact theory and our methodology , we can match the results of simulations , the agreement between data and theory being remarkable , especially considering how crude our modelling of the hv simulations is ( we use a measurement of the 3d real - space correlation function as our baseline model , and we do not include a correction for fingers - of - god type effects ) . for rsd measurements made within radial bins , it will be vital to match the theory to the exact distribution of galaxies observed . in a measurement of rsd from real data ,
the final result will be a weighted average over different opening angles , to take account for the fraction of galaxy pairs separated by a @xmath13 angle , so in wide surveys one has to discard the plane - parallel approximation , or face losing a considerable amount of information .
for this reason our methodology will be particularly useful for incoming and future redshift surveys , such as the baryon oscillation spectroscopic survey ( boss ; @xcite ) , bigboss @xcite , and euclid @xcite ; the importance of this methodology for current and future experiments will be considered in a subsequent paper @xcite .
ar is grateful for the support from a uk science and technology facilities research council ( stfc ) phd studentship .
wjp is grateful for support from the european research council and from the leverhulme trust and stfc .
ls acknowledges support from european research council , georgian national science foundation grant st08/4 - 442 and snsf ( scopes grant no 128040 ) .
ar would like to thank g. w. pettinari for the useful discussions .
simulated data was calculated and analysed using the cosmos altix 3700 supercomputer , a uk - ccc facility supported by hefce and stfc in cooperation with cgi / intel . | the analysis of redshift - space distortions ( rsd ) within galaxy surveys provides constraints on the amplitude of peculiar velocities induced by structure growth , thereby allowing tests of general relativity on extremely large scales .
the next generation of galaxy redshift surveys , such as the baryon oscillation spectroscopic survey ( boss ) , and the euclid experiment will survey galaxies out to @xmath0 , over @xmath1@xmath2 . in such surveys ,
galaxy pairs with large comoving separation will preferentially have a wide angular separation .
in standard plane - parallel theory the displacements of galaxy positions due to rsd are assumed to be parallel for all galaxies , but this assumption will break down for wide - angle pairs . @xcite and @xcite provided a methodology , based on tripolar spherical harmonics expansion , for computing the redshift - space correlation function for all angular galaxy pair separations . in this paper
we introduce a new procedure for analysing wide - angle effects in numerical simulations .
we are able to separate , demonstrate , and fit each of the effects described by the wide - angle rsd theory .
our analysis highlights some of the nuances of dealing with wide - angle pairs , and shows that the effects are not negligible even for relatively small angles .
this analysis will help to ensure the full exploitation of future surveys for rsd measurements , which are currently confined to pair separations less than @xmath3 out to @xmath4 .
[ firstpage ] cosmological parameters large scale structure of universe cosmology : observations methods : analytical |
as we look ahead to the end of the coming decade , we will be entering the era of the extremely large telescopes ( elts ) .
plans are well advanced for three elt projects with filled apertures well in excess of current optical - ir facilities : the giant magellan telescope ( gmt , with an equivalent diameter of 21.4 m ) , the thirty meter telescope ( tmt ) and the 42 m european elt ( e - elt ) . when coupled with sophisticated adaptive optics ( ao ) systems to correct for atmospheric turbulence
, the elts will provide us with unique views of stellar populations in the local volume .
although planning a decade ahead may seem far removed , the lead - in times on these ambitious projects is more akin to those commonly associated with space missions , i.e. construction planning , detailed science simulations , conceptual instrument designs and the financial planning are all well advanced on all three elt projects .
here we introduce the eagle phase a instrument study for the e - elt , highlighting its performance for observations of resolved stellar populations .
photometric methods are immensely powerful when applied to extragalactic stellar populations , but only via precise chemical abundances and stellar kinematics can we break the age - metallicity degeneracy , while also disentangling the populations associated with different structures , i.e. follow - up spectroscopy is required . over the past decade
the calcium triplet ( cat , spanning 0.85 - 0.87 @xmath1 m ) has become an increasingly used diagnostic of stellar metallicities and radial velocities in nearby galaxies , providing new views of their star - formation histories and sub - structure , e.g. the vlt - flames dart large programme @xcite .
however , 8 - 10 m class telescopes are already at their limits in pursuit of cat spectra of the evolved populations in galaxies at distances greater than @xmath2300 kpc , e.g. keck - deimos observations in m31 struggled to yield useful signal - to - noise below the tip of the red giant branch at @xmath3 @xmath4 21.5 @xcite . with its vast primary aperture and excellent angular resolution
, the e - elt will be _ the _ facility to unlock spectroscopy of evolved stellar populations in the broad range of galaxies in the local volume , from the edge of the local group , out towards the virgo cluster .
this will bring a wealth of new and exciting target galaxies within our grasp , spanning a broader range of galaxy morphologies , star - formation histories and metallicities than those available to us at present in the local group .
these observations can then be used to confront theoretical models to provide a unique view of galaxy assembly and evolution .
there are many compelling and ground - breaking targets for stellar spectroscopy with the e - elt including , in order of distance : * ngc 3109 and sextans a with sub - smc metallicities ( @[email protected]@xmath7 ) , both at 1.3 mpc . * the spiral dominated sculptor ` group ' at 2 - 4 mpc .
* the m83/ngc5128 ( centaurus a ) grouping at @xmath24 - 5 mpc . * ngc 3379 , the nearest normal elliptical at 10.8 mpc . * the virgo cluster of galaxies at 16 - 17 mpc , the nearest massive cluster .
in contrast to proposed e - elt observations of high - redshift galaxies , targets for cat spectroscopy are readily available .
for example , deep ground - based and _ hst _ imaging in galaxies in the local volume has begun to investigate their stellar populations @xcite , yet the stellar magnitudes are well beyond spectroscopy with existing facilities .
note that although we have focussed mostly on southern - hemisphere targets here , there are equally compelling northern hemisphere targets , including the m81 group , and deeper studies in m31 and m33 .
the eagle phase a study is a french - uk partnership to provide an advanced conceptual design of an ao - corrected , near - infrared spectrograph with multiple integral - field units ( ifus ) .
the baseline design is summarised in table 1 , and has been shaped by five top - level science topics : * physics and evolution of high - redshift galaxies * detection and characterisation of ` first light ' galaxies * galaxy assembly and evolution from stellar archaeology * star - formation , stellar clusters and the initial mass function * co - ordinated growth of black holes and galaxies eagle will employ multi - object adaptive optics ( moao , @xcite to provide significantly improved image quality for selected target fields within the focal plane .
this entails an array of six laser guide stars and five natural guide stars ( ngs ) to map the atmospheric turbulence .
the deformable mirror in the telescope ( m4 ) will be used to correct for the low - order wavefront error terms , with the high - order terms corrected by deformable mirrors in each science channel .
an integral part of the eagle project is the canary on - sky demonstrator on the william herschel telescope in la palma @xcite .
the consortium has calculated moao point - spread functions ( psfs ) which take into account real ngs configurations , illustrative of relatively good and poor performance given the spatial distribution and magnitude of the available guide stars . in the following section we summarise eagle observations of resolved stellar populations beyond the local group , which would be used to probe the assembly history and chemical evolution of the host galaxies .
.eagle baseline design .
the patrol field is the instrument field - of - view within which ifus can be configured to observe individual targets . [
cols="<,<",options="header " , ]
eagle provides a unique combination of abilities to harness the power of the e - elt for spectroscopy of resolved stellar populations .
the image quality from moao will be significantly better than that obtained from seeing - limited or ground - layer ao modes , enabling us to explore spatially - resolved , extragalactic stars across a wide field of more than five arcminutes .
a range of large programmes can be envisaged , each of which will help provide a fundamentally new view of stellar populations in the local universe ; from mapping of the sculptor galaxies , to deeper observations of the most luminous evolved stars in selected fields in galaxies at 10mpc and beyond .
tolstoy , e. et al .
, 2004 , _ apjl _ , 617 , 119 chapman , s. et al . , 2006 , _ apj _ , 653 , 255 rejkuba , m. et al . , 2005 , _
apj _ , 631 , 262 de jong , r. s. et al . , 2007 , _
iaus241 _ , p503 dalcanton , j. j. et al . , 2009 , _
apjs _ , 183 , 67 assmat f. , gendron e. , hammer f. , 2007 , _ mnras _ , 376 , 287 myers , r. m. et al . , 2008 , _
spie _ , 7015 , 6 puech , m. et al . , 2008 , _ mnras _ , 390 , 1089 munari , u. & castelli , f. , 2000 , _ a&as _
, 141 , 141 sarazin , m. , melnick , j. , navarrete , j. & lombardi , g. , 2008 , _ msngr _ , 132 , 11 | we give an overview of the science case for spectroscopy of resolved stellar populations beyond the local group with the european extremely large telescope ( e - elt ) .
in particular , we present science simulations undertaken as part of the eagle phase a design study for a multi integral - field - unit , near - infrared spectrograph .
eagle will exploit the unprecedented primary aperture of the e - elt to deliver ao - corrected spectroscopy across a large ( 38.5 arcmin@xmath0 ) field , truly revolutionising our view of stellar populations in the local volume . |
a resonance at @xmath5 is discovered in both the atlas @xcite and cms experiments @xcite , which is consistent with the standard model ( sm ) higgs boson @xcite .
this is one of the great experimental achievement in recent years . despite the great success of sm , there are still problems implying new physics beyond sm .
one of the simplest extensions is the two higgs doublet model ( 2hdm ) , with a second higgs doublet introduced by ref .
@xcite ( see also @xcite ) .
the higgs potential and the yukawa lagrangian corresponding to the 2hdm is not unique .
the one with a cp - invariant potential , and with the yukawa lagrangian such that one of the higgs doublet couples to the down sector of fermions while the other couples to the up , which is so called 2hdm - type - ii @xcite .
such a lagrangian can avoid tree - level flavor changing neutral currents ( fcncs ) @xcite .
there are 2 vacuum expectation values ( vev ) , denoted as @xmath6 , and @xmath7 in 2hdm , and 5 higgs bosons denoted as @xmath8 , @xmath9 , @xmath10 and @xmath11 . by studying the phenomenons induced by 2hdm , and comparing the results with the experimental data
, one can detect the signal of the existence of new physics , and fix the free parameters in the model .
the radiative leptonic decay of the heavy pseudoscalar meson with a massive lepton provides a good opportunity to study the 2hdm .
this process can be mediated both by @xmath12 bosons and @xmath11 bosons , however , the contribution of the charged higgs bosons is barely investigated before .
the contribution of 2hdm is usually small . to probe the signal , sufficient precision of the numerical results in the sm
is required . in this decay mode ,
the strong interaction is involved only in the initial hadronic state . as a result
, this decay mode has been investigated using the factorization approach @xcite , and results has been calculated up to the order @xmath13 in the sm @xcite . apart from that , with a massive lepton in the final state , the results are expected to be sensitive to the charged higgs bosons .
there are two free parameters in 2hdm associated with this decay mode , @xmath14 , and @xmath15 .
the branching ratios can be expressed as the function of @xmath16 . using the constraints on those parameters @xcite
, the branching ratios including the contribution of the charged higgs bosons can be obtained .
the results are shown in table [ brtot ] .
we find that , the decay mode @xmath17 is very sensitive to the charged higgs bosons , which makes it a good decay mode to probe the signal of 2hdm . on the other hand , the decay mode @xmath18
is barely affected by 2hdm , and should be excluded in the search of the charged higgs bosons .
this paper is organized as follows .
ii is a briefly review of the factorization in the sm .
the contribution of the 2hdm is discussed in sec .
iii . the numerical results and analysis
are contained in sec .
v is a summary .
the heavy pseudoscalar meson is constituted with a quark and an anti - quark , and one of the quarks is a heavy quark .
the feynman diagrams of the radiative leptonic decay of @xmath19 meson at tree level can be shown as fig .
[ tree1 ] . the contribution of fig . [ tree1 ] .
d is suppressed by a factor of @xmath20 , and can be neglected .
the amplitudes of fig .
[ tree1 ] .
a , b and c can be written as tree level amplitudes in sm , the double line represents the heavy quark propagator but not the hqet propagator . ]
@xmath21 where @xmath22 and @xmath23 are the momenta of the anti - quark @xmath24 and quark @xmath25 , @xmath26 , @xmath27 and @xmath28 are the momenta of photon , lepton and neutrino , @xmath29 denotes the polarization vector of photon , and @xmath30 .
the idea of factorization is to absorb the infrared ( ir ) behaviour into the wave - function , the matrix element can be written as the convolution of wave - function and hard kernel @xcite @xmath31 for simplicity , we denote @xmath32 using the wave - function defined in coordinate space @xmath33q_{\beta}(y)|\bar{q}^s(p_{\bar{q}}),q^s(p_q ) > \label{eq.1.4}\ ] ] where @xmath34 $ ] is the wilson line which can be written as @xcite @xmath35=\exp \left[ig_s\int _
y^x d^4z z_{\mu } a^{\mu}(z)\right]=\sum _ { \substack{n}}\frac{(ig_s)^n}{n!}\prod _ { \substack{i}}^n\int _ y^x d^4z_i z_{i\mu } a^{\mu}(z_i ) \label{eq.1.5 } \end{split}\ ] ] it has been proved that @xcite , the matrix element up to the order @xmath13 at one - loop can be factorized as @xmath36 with @xmath37 except for @xmath38 , all the other products are contribution of order @xmath39 , for clarity , we define @xmath40 , with @xmath41 represents order @xmath42 contribution , the coefficients are @xmath43 @xmath44 @xmath45 @xmath46 @xmath47 where @xmath48 is defined as : @xmath49 in the results given above , the large logarithms at order @xmath50 is already resummed .
the @xmath51 term is the contribution from @xmath52 , and it dose not have 1-loop hard scattering kernel , so we have @xmath53 the numerical result can be obtained by using the wave - function with the one obtained in ref .
@xcite @xmath54 with @xmath55 and the result is @xmath56 with @xmath57 where @xmath58 and @xmath59 come from the contribution of @xmath52 .
using @xmath60 we find @xmath61 and the decay amplitude can be written as @xmath62
the feynman diagrams at tree level of the radiative leptonic decay in 2hdm are shown in fig .
[ tree2 ] . the contribution of fig . [ tree2 ] .
d is suppressed by higgs propagator and neglected , the decay amplitudes of the others can be written as tree level amplitudes of charged higgs bosons .
the double line represents the heavy quark , and the dashed line , the charged higgs . ]
@xmath63 where @xmath64 is the mass of the charged higgs boson , and @xmath65 is a free parameter in 2hdm .
the leading contribution at order @xmath66 is @xmath67 , which will vanish .
this can be shown immediately , the matrix element can be written as @xmath68 at the leading order , we can replace @xmath69 by @xmath70 , and the matrix element can be factorized as @xcite @xmath71 contracting with @xmath72 , it will vanish .
in fact , both fig .
[ tree2 ] .
a and fig . [ tree2 ] .
b , and the qcd corrections of them will not contribute to the decay amplitude because of the symmetry of the wave - function of the pseudoscalar mason .
using the similar factorization procedure as in the sm , we find that , up to the order @xmath73 , the matrix element of fig . [ tree2 ] .
a and fig .
[ tree2 ] .
b with 1-loop qcd corrections can be factorized as @xmath74 the delta - function in the wave - function will replace the transmission momenta @xmath75 , @xmath76 to the on - shell momenta @xmath77 , @xmath23 . using the definition of the wave - function in eqs .
( [ eq.1.14 ] ) and ( [ eq.1.15 ] ) , in dirac representation , we find @xmath78\\ \end{split } \label{eq.2.5}\ ] ] with @xmath79 and @xmath80 the kinematic indicates that @xmath81 notice that , @xmath82 are either even functions of @xmath83 or @xmath84 , in this case , after the trace , all terms are vanished , or the odd function of @xmath83 or @xmath84 , then , all terms will vanish after the integral @xmath85 . however , @xmath86 will survive and will contribute to the decay amplitude . as the same case in the sm , @xmath86 also dose not receive contribution from 1-loop hard scattering kernel . as a result
, we find @xmath87 with @xmath88
the form factors defined in eq .
( [ eq.1.16 ] ) can be calculated using the integral defined in eq .
( [ eq.1.17 ] ) , and we evaluate this integral using @xcite @xmath89 the numerical results of the form factors are inconvenient to use when calculate the decay widths . for simplicity , we use some simple forms to fit the numerical results . inspired by the form factors in ref .
@xcite , the form factors are fitted as @xmath90 the results are more reliable at the region @xmath91 because we have neglected the higher order terms of @xmath92 . as we have done in ref .
@xcite , we choose the region @xmath93 to fit the parameters in eq .
( [ eq.3.2 ] ) .
the result of @xmath19 mason in the sm is @xmath94 for @xmath95 mason , there is an additional minus sign in @xmath96 and @xmath97 , and the result is @xmath98 the fitting of the form factors are shown in fig .
[ formb ] and fig .
[ formd ] .
fit of the form factors of @xmath99 .
the points ` @xmath100 ' and ` + ' are the numerical results for the form factors @xmath97 and @xmath96 , and the dotted and dashed curves are the fitted results using eq .
( [ eq.3.2 ] ) . ]
fit of the form factors of @xmath101 .
the points ` @xmath100 ' and ` + ' are the numerical results for the form factors @xmath97 and @xmath96 , and the dotted and dashed curves are the fitted results using eq .
( [ eq.3.2 ] ) . ] on the other hand , @xmath58 and @xmath59 are related to the decay constant , and we use the result in refs .
@xcite which is calculated using the same wave - function and the same parameters as what we use in this work .
the decay constants are @xmath102 the contribution of the 2hdm @xmath103 can be calculated using eq .
( [ eq.2.10 ] ) , and the result is @xmath104 with @xmath105 defined as @xmath106 , the branch ratios can be written as @xmath107 with @xmath108 just as the case of the pure leptonic decay @xcite , the interference term in the radiative leptonic decay is also found to be destructive .
we calculate @xmath109 , @xmath110 and @xmath111 separately . using the fitted result of @xmath96 , @xmath97 , the result of @xmath112 and @xmath103 , and using the cabibbo - kobayashi - maskawa ( ckm ) matrix elements @xcite , and mass of the leptons
@xmath113 we can obtain @xmath114 , @xmath115 and @xmath116 .
there are ir divergences in the radiative leptonic decays when the photon is soft or collinear with the emitted lepton .
theoretically this ir divergences can be canceled by adding the decay rate of the radiative leptonic decay with the pure leptonic decay rate , in which one - loop correction is included @xcite .
this is because the radiative leptonic decay can not be distinguished from the pure leptonic decay in experiment when the photon energy is smaller than the experimental resolution to the photon energy .
so the decay rate of the radiative leptonic decay depend on the experimental resolution to the photon energy @xmath117 which is denoted by @xmath118 .
the dependence of the branching ratios of @xmath19 meson in the sm on the resolution are listed in table [ bsm ] . and
for the same reason , @xmath110 and @xmath111 also depend on @xmath118 , the results of @xmath110 and @xmath111 of b meson are listed in table [ bab ] .
.the branching ratios with different photon resolution @xmath118 in the sm . [ cols="^,^,^,^,^,^",options="header " , ] and the ratio @xmath119 as a function of @xmath105 are shown in figs .
[ rhbmu ] , [ rhbtau ] and [ rhdmu ] .
@xmath119 as function of @xmath105 in the @xmath120 decay mode . ] @xmath119 as function of @xmath105 in the @xmath121 decay mode . ]
@xmath119 as function of @xmath105 in the @xmath122 decay mode . ]
we find that , the decay mode @xmath3 is very sensitive to the contribution of the charged higgs bosons .
for @xmath123 , for example , the branching ratios is suppressed by more than @xmath124 .
in this paper , we calculated the branching ratios of the radiative leptonic decay of the heavy pseudoscalar meson with a massive lepton .
the contribution of 2hdm - type - ii is included .
the sm contribution is obtained using the factorization procedure up to the order @xmath125 with one - loop correction .
the contribution of the charged higgs boson is also obtained by the factorization scheme , however , we find that only the diagram with the photon emitting from the lepton leg will contribute , which is an order @xmath126 contribution and dose not receive contributions from 1-loop hard scattering kernel
. the numerical results of the branching ratios are listed in table [ brtot ] , and dependence of @xmath119 on @xmath105 is shown in figs .
[ rhbmu ] , [ rhbtau ] and [ rhdmu ] .
we find that , the decay mode @xmath17 is sensitive to the contribution of the charged higgs in the 2hdm .
this decay mode is as sensitive as the pure leptonic decay of @xmath19 meson , which is estimated to be @xmath127 @xcite .
however , the result of the pure leptonic decay is derived at the leading order of @xmath66 , while our result is calculated up to the order @xmath128 . on the other hand , the branching ratios of this decay mode in the sm
is about @xmath129 , which is larger than @xmath130 and @xmath131 both are believed to be less than @xmath132 @xcite .
the decay mode @xmath120 is also important because the charged higgs can suppress the branching ratios as large as about @xmath133 .
it is more sensitive than the pure leptonic decay modes @xmath134 , which is estimated to be @xmath135 .
though the branching ratio of the radiative leptonic decay of the @xmath19 meson is much smaller , the result in the sm is more reliable than the case of the @xmath95 meson because the heavy quark mass @xmath136 is larger then @xmath137 .
we also find that , the decay mode @xmath138 is rarely affected by the charged higgs boson . on the other hand , the numerical result of this mode is also not sufficiently accurate in the sm because the mass of @xmath139 quark is not so heavy .
this decay mode should be excluded in the search of the charged higgs bosons . | in this work , we study the radiative leptonic decays of @xmath0 and @xmath1 mesons in the standard model and the two - higgs - doublet - model type ii . the results are obtained using the factorization procedure , and the contribution of the order @xmath2 is included .
the numerical results are calculated using the wave - function obtained in relativistic potential model . as a result , the decay mode @xmath3 is found to be sensitive to the effect of the charged higgs boson . using the constraint on the free parameters of the two higgs doublet model given in previous works
, we find the contribution of the charged higgs boson in the decay mode @xmath3 can be as large as about @xmath4
. 0.25 cm |
carbonate melts are liquids with remarkable physical and chemical properties that have recently received much attention @xcite . in addition to being present in earth s upper mantle , they have also been increasingly used in various technological applications , either as bulk material or in composites with solid oxides or metals .
molten carbonates are successfully used as electrolytes in _ molten carbonate fuel cells _ ( mcfc ) and are experimented in carbon capture and storage devices . yet , despite the importance of carbonate melts for both geophysical settings and industrial applications , a clear picture of the structure , dynamics and reactivity of molten carbonates at the molecular scale is still lacking . here , we set out to use molecular simulation methods to provide insight into the microscopic structure and dynamical properties of the @xmath2 eutectic mixture ( 62:38 mol% ) , which is of particular relevance for industrial applications as it is a standard electrolyte for molten carbonate fuel cells due to its high ionic conductivity and low eutectic temperature @xcite . _ first principles molecular dynamics _
( fpmd ) simulations are one possible choice of simulation technique in order to simulate the carbonate melts . relying on quantum chemical calculations of the electronic structure at the _ density functional theory _ ( dft ) level
, they do not require _ a priori _ knowledge of the molecular interactions in carbonate melts .
fpmd simulations have been used with great success in recent years to calculate the equation of state of molten @xmath3 @xcite and the speciation and transport of @xmath4 in @xmath3 @xcite . however , first principles techniques are computationally very expensive , and therefore inherently strongly constrained on the time scales and system sizes that can be investigated in that manner . on the other hand , molecular dynamics ( md ) based on
empirical force fields are less computationally demanding , but requiring careful parametrization . for our system of interest
, we note that there have been so far relatively few attempts to derive force fields for molten alkali carbonates @xcite .
therefore , in this work we propose an optimized force field for the @xmath2 eutectic mixture , based on both experimental data and results from fpmd simulations .
we then use this force field to study the thermodynamics , structure and dynamics of the molten mixture in the 9001100 k temperature range , at pressures up to 5 gpa .
we focus in particular on the dynamical properties including the self diffusion coefficients , the viscosity and the ionic conductivity .
in order to be able to provide reference structural information to optimize and validate our classical force field , we have first performed fpmd simulations on the @xmath2 ( 62:38 mol% ) eutectic mixture at the three target temperatures , @xmath5 k , @xmath6 k and @xmath7 k. we conduct these simulations using dft and the born oppheneimer dynamics .
we use the cp2k software package @xcite and in particular the quickstep algorithm @xcite , which employs a hybrid gaussian plane - wave method ( gpw ) @xcite for the electronic structure .
the core electrons are replaced by using norm conserving goedecker teter hutter ( gth ) pseudo - potentials @xcite . for all atomic species
we use a double - zeta valence plus polarization ( dzvp ) basis set optimized for molecules @xcite .
we cut off the electronic density at 700 ry and use a ` spline3 ` smoothing for applying the exchange correlation potential .
the latter is described thanks to the blyp functional @xcite , with added dispersive interactions through the dft - d3 scheme @xcite with a cutoff distance of 40 .
the system studied by first - principles simulations is composed of 128 @xmath8 , 159 @xmath9 and 97 @xmath10 , with periodic boundary conditions .
we adjust the box length in order to prepare the system at the experimental density at atmospheric pressure @xcite and we run the simulations in the @xmath11 ensemble .
the temperature is controlled by using the csvr thermostat @xcite with a time constant of 1 ps .
we employ a simulation time step of 0.5 fs and the total simulation time is 30 ps . the initial configurations had been obtained by classical md using the tissen and janssen force field @xcite .
relatively few attempts have been previously made to derive a force field for molten alkali carbonates @xcite . the force field developed by tissen and janssen @xcite for pure lithium , sodium and potassium carbonate salts , and
later also used to study mixtures @xcite , treats the carbonate molecule as rigid . in this simple force field ,
the interaction potential between the atoms is simply given by the sum of the coulombic potential and a born type repulsive term .
partial charges are used for the atoms composing the @xmath8 units , while formal charges are used for the cations .
costa and ribeiro @xcite have compared the results obtained using the force field by tissen and janssen to the ones produced by a force field in which the polarizability of the anion is added in the form of fluctuating partial charges , showing a large impact on the prediction of transport properties . here , in light of the length of the runs needed to converge the calculation of dynamical quantities such as diffusion coefficients , ionic conductivity and viscosity , we limit ourselves to a non polarizable force field .
the complexity of the force field used is however somewhat increased with respect the tissen and janssen model . for the derivation of our classical force field for @xmath2 ,
we start from the force field obtained for @xmath3 in ref . @xcite .
we consider flexible @xmath8 ions .
the c o pairs in carbonate interact via harmonic terms : @xmath12 with @xmath13 kj/(mol @xmath14 ) , @xmath15 , and where `` intra '' means that the sum is taken only for o in the same carbonate molecule of the corresponding c. the carbonate molecule is prevented from folding by adding exponential repulsion terms between o atoms of the kind : @xmath16 with @xmath17 kj / mol and @xmath18 .
the non bonded ( or intermolecular ) part of the force field includes a coulombic part , a repulsive term and a dispersive term , with its analytical form being : @xmath19 where `` nb '' means that the sum is taken only over non - bonded pairs of atoms .
we use the same partial charges as in ref .
@xcite , that is to say @xmath20 , @xmath21 and @xmath22 .
we had initially attempted to reoptimize these charges , but any significant change leads to drastic deterioration of the pressures and densities of the system when compared to experimental data .
we thus proceed to adjust the interaction parameters @xmath23 , @xmath24 and @xmath25 by manual tuning in order to reproduce at best both the experimental density by constant - pressure @xmath26 simulations and the pair radial distribution functions ( rdfs ) , as obtained from first - principles molecular dynamics ( fpmd ) simulations , by constant - volume @xmath11 simulations .
the best values of the parameters are reported in table [ tab:1 ] .
all parameters not shown ( @xmath27 , @xmath28 , @xmath29 , @xmath30 , @xmath31 , @xmath32 ) are equal to zero .
.non bonded force field parameters @xmath23 , @xmath24 for the repulsive and @xmath25 for the dispersive terms of the interaction potential , eq . [ eq : nb ] .
all pair parameters not shown are set to zero . [
cols="<,^,^,^",options="header " , ] [ tab : app3 ]
we thank michel cassir and virginie lair for fruitful discussions . we acknowledge funding by psl research university ( project coocar , grant anr-10-idex-0001 - 02 ) and
agence nationale pour la recherche ( project electrolith , grant anr-2010-blan-621 - 03 ) .
this work was performed using hpc resources from genci ( grant 2015 - 082309 ) .
g. j. janz .
_ thermodynamic and transport properties for molten salts : correlation equations for critically evaluated density , surface tension , electrical conductance and viscosity data .
_ j. phys .
chem . ref .
data * 17 * , suppl . 2 ( 1988 ) . | we use molecular dynamics simulations to study the thermodynamics , structure and dynamics of the li@xmath0co@xmath1k@xmath0co@xmath1 ( 62:38 mol% ) eutectic mixture .
we present a new classical non - polarizable force field for this molten salt mixture , optimized using experimental and first principles molecular dynamics simulations data as reference .
this simple force field allows efficient molecular simulations of phenomena at long timescales .
we use this optimized force field to describe the behavior of the eutectic mixture in the 9001100 k temperature range , at pressures between 0 and 5 gpa . after studying the equation of state in these thermodynamic conditions , we present molecular insight into the structure and dynamics of the melt .
in particular , we present an analysis of the temperature and pressure dependence of the eutectic mixture s self diffusion coefficients , viscosity and ionic conductivity . @preabstractspace3.0 |
the morphology , evolution and chemical composition of circumstellar disks and halos of pre - main - sequence intermediate - mass herbig ae / be stars is of key importance in the study of star and planet formation @xcite .
emission features of silicates and pahs are commonly though not always observed @xcite and in a few cases emission from surface - hydrogen - terminated diamonds is also detected .
this raises interesting and challenging questions as to how the diamonds form , why their emission is seen in only a few stars , and whether there is a link between circumstellar diamonds and the presence of nanodiamonds in meteorites .
the existence of organic pahs in the nebular environment also provides astrobiological interest . in this paper
we describe near - ir long - slit observations of one of the rare diamond c - h emitting objects , elias 1 .
we focus on the spectroscopy and spatial distribution of diamond ( c - h ) , pah and continuum emission from the circumstellar environment , and make comparison with the related object hd 97048 .
the pre - main - sequence star elias 1 ( v892 tau , elias 3 - 1 ) is a close binary with a separation of @xmath350 mas at pa @xmath350@xmath4 @xcite .
its spectral type is uncertain ranging through a0 @xcite , a6 with @xmath5@xmath6 = 3.9 mag @xcite , b9 with @xmath5@xmath6 = 8.85 mag @xcite , with an @xmath7 value of 10.5 also reported @xcite .
elias 1 has a very flat , possibly rising , sed in the near - ir to far - ir which is markedly different from that of hd 97048 @xcite , the two stars being classified as group ii and i , respectively .
the sed for elias 1 is taken by @xcite to suggest the possible existence of an envelope in addition to a disk , or a flared disk .
other elias 1 characteristics include x - ray @xcite and radio @xcite emission .
an emission feature in the _ c. _ 3.5@xmath0 m region for elias 1 was first reported in 1982 by @xcite and was subsequently shown to have two components near 3.41 and 3.52@xmath0 m with an additional pah feature at 3.29@xmath0 m @xcite .
the only previously detected example was hd 97048 @xcite .
subsequent spectroscopic observations of elias 1 have been made @xcite , including iso ( sws ) data @xcite .
the features at 3.43 and 3.53@xmath0 m remained unassigned until @xcite found a convincing correspondence between the astrophysical bands of elias 1 and hd 97048 and laboratory absorption spectra of the c - h stretching modes on hydrogen - terminated diamond nanocrystal films at @xmath3 1000 k @xcite .
iso spectra of elias 1 covering the range 2.5 - 13.5 @xmath0 m @xcite contain absorption features due to water and co@xmath8 ice , and emission from pahs , diamonds and silicates , the last of these being absent from the spectrum of hd 97048 . in this paper
we have chosen not to use the word ` nanodiamond ' in discussing the stellar emission features as their size is probably at least 50 nm @xcite in contrast to meteoritic nanodiamonds which are typically only 2 - 3 nm across @xcite .
we also remark that the diamonds observed are exclusively those that are h - terminated at the surface ; absence of hydrogen or its substitution by other chemical groups on the surface in general give spectra outside the region studied .
the only other herbig ae / be objects for which diamond emission is reported are bd@xmath940@xmath44124 @xcite and mwc 297 @xcite ; emission features attributed to diamond emission are also seen in the post - agb star hr 4049 @xcite but are relatively weak .
unresolved at mm wavelengths @xcite , observations of the spatial distribution of the circumstellar emission of elias 1 comprise a contour map of the inner 0.2@xmath2 @xmath10 0.2@xmath2 region which exhibits asymmetry along the binary pa of 50@xmath11 @xcite , and speckle interferograms at k , l ( 0@xmath11 and 90@xmath11 ) and l at 45@xmath11 @xcite and at j , h , k and l@xmath12 @xcite .
@xcite interpreted their data in terms of a core with a wavelength - dependent diameter of @xmath3 6 - 14 au , and a highly flattened halo elongated in the east - west direction with radius @xmath3 30 - 100 au in the form of a nearly edge - on circumstellar disk .
based on observations covering a wider range of wavelengths and at higher resolution , @xcite proposed that the east - west elongation is due to scattered light from bipolar ( nebular ) lobes . from the pa 90@xmath11 and 0@xmath11
data it was concluded that the components observed at j and h comprise a well - resolved narrow blue structure elongated east - west with a fwhm of _ c. _ 1@xmath2 and a second that is marginally resolved ( fwhm @xmath10.2@xmath2 ) , and that there is marginally resolved red structure at k and l@xmath13 which is more circular and symmetric . from a consideration of all the available data and a number of possible interpretations , @xcite proposed the existence of a bipolar nebula along the e - w axis as the origin of the e - w extension .
they also discussed the existence of a possible stellar companion , noting that their k and l@xmath12 data could be fitted by a close binary with separation oriented at pa 45@xmath11 or 135@xmath11 with separation @xmath140.1@xmath2 , together with some extended structure .
recent diffraction - limited speckle interferometric observations confirm that elias 1 is indeed a close binary system with these characteristics , the separation being 0.05@xmath2 with pa 50@xmath11 @xcite .
observations of the herbig ae / be star elias 1 were carried out between 20th july - 11th september 2004 at the 3.8 m united kingdom infrared telescope ( ukirt ) .
the 1 - 5 @xmath15 m imager - spectrometer ( uist ) was used for the long - slit spectroscopic observations .
the slit dimensions were 0.48@xmath2 x 120@xmath2 and the resolving power = 700 .
the short - l grism was used to obtain spectra in the wavelength region 2.905 - 3.638@xmath0 m in combination with an insb array of 1024 x 1024 pixels , giving a spatial resolution of 0.12@xmath2 pix@xmath16 .
the slit was centred on the star and orientated at position angles 0@xmath4 ( ns ) and 90@xmath4 ( we ) .
the observational details are listed in table 1 .
the sky background emission was subtracted from the standard and object frames and the object spectra were divided by the standard .
the point spread function ( psf ) of the observations is very consistent .
a further exposure along e - w was made on 27 july but due to poor seeing the psf was _ c. _ 0.7@xmath2 and was not included in this analysis .
the data were partially reduced by the orac - dr ( automated reduction ) pipeline for uist and further reductions were carried out with the figaro reduction package .
the spectra were analysed using the iraf reduction package ; the continuum was fitted and subtracted from the spectra and lorentzians were used to fit the emission bands .
.observing log [ cols="<,^ , < , < , < , < " , ]
in common with similar long - slit studies of red rectangle pah features ( song et al .
2003 ) , one of the objectives was to investigate any evolution in profile shape , central wavelength etc . of the pah c - h stretch and ( particularly ) the c - h diamond features with increase in offset from the star .
the summed 3.0 - 3.6@xmath0 m spectra for elias 1 along the n - s axis for the inner @xmath17 relative to the star are shown in fig 1 and are similar to the ukirt data of @xcite although the pf@xmath18 line of hydrogen which sits atop the 3.3@xmath0 m pah band is much weaker in our spectra .
the pf@xmath18 flux in our 2004 data is _
c. _ 0.2 x 10@xmath19 wm@xmath20@xmath21m@xmath16 compared with _ c. _ 1.0 x 10@xmath19 wm@xmath20@xmath21m@xmath16 @xcite .
the assignments of the c - h diamond features have been described elsewhere @xcite and are not reproduced here .
the best match between experimental absorption spectra and the astrophysical data is achieved for a temperature of _ c. _
1000 k @xcite .
laboratory data have revealed a strong dependence of the spectrum on particle size , indicating that the diamonds in elias 1 ( and hd 97048 ) are probably at least 50 nm in diameter and are likely formed in a cvd - type process @xcite .
m region recorded with slit alignment n - s .
the data were summed over the inner @xmath17 .
the data for e - w are the same within the signal - to - noise ratio . ]
as described in the following sub - section , the diamond distribution in elias 1 is confined to a region quite close to the star within @xmath148 au ( fwhm , gaussian fitted ) . by taking data extractions ` on - star ' at the point of maximum emission strength and
secondly at the greatest possible offset for which a reasonable signal - to - noise is still maintained , we searched for variation in the diamond features but did not find evidence for this to within 0.15% ( in @xmath22 ) , 20% ( in fwhm ) , and 10% ( in relative intensity ) .
the psf of _
c. _ 0.5@xmath2 is larger than desirable relative to the diamond emission spatial extent of fwhm @xmath10.34@xmath2 .
however , changes in the relative band strengths of the components , in peak wavelength or in profile shape that might occur due to change in temperature with offset , diamond size effects on the spectra , or in the relative emission contributions of the different c - h supporting diamond facets , would be discernible .
this result points to a rather specific set of formation conditions and/or surface chemistry , possibly not unrelated to the rarity of the appearance of the features . of the main pah emission bands ,
only the relatively weak 3.3@xmath0 m feature occurs in the spectral range recorded and is subject to contamination by telluric methane and water lines , as well as from the pf@xmath18 emission line of the star . as for the diamond emission bands
there was no measurable change in the feature profile with offset but a 20% change in width would have been detected .
slit alignments along the n - s and the e - w axes were selected in order to probe the spatial distribution of the pah ( 3.3@xmath0 m ) , diamond and continuum emission relative to the star .
although adaptive optics and larger mirror size ( and speckle interferometric observations ) offer a lower effective psf , high signal - to - noise data achieved with moderate exposure times and at high altitude on mauna kea provide an attractive option with relatively straightforward data reduction procedures .
determination of the psf was achieved using the standard ( point source ) star bs 1203 and gave a stable value of 0.53(2)@xmath2 ( see table 1 . )
the n - s and e - w spatial distributions as recorded for the pah ( 3.3@xmath0 m , continuum - subtracted ) , diamond ( 3.53@xmath0 m , continuum - subtracted ) and continuum ( 3.60@xmath0 m ) are given in figs 2 and 3 , respectively .
the set of diamond features near 3.43@xmath0 m follow the 3.53@xmath0 m distribution within signal - to - noise contraints , the measured spatial fwhm for 3.53@xmath0 m being 0.590@xmath2 ( ns ) and 0.603@xmath2 ( ew ) , and 0.603@xmath2 ( ns ) and 0.597@xmath2 ( ew ) for the 3.43@xmath0 m group .
figs 4 and 5 show the results when a richardson - lucy ( rl ) deconvolution algorithm was employed to deconvolve the profiles .
for the n - s slit position ( pa = 0@xmath4 ) in fig 4 the profiles for pah , diamond and continuum overlap very well for the inner @xmath23 , with the latter two overlapping for all offsets .
however , the pah distribution extends weakly about three times further , with slightly greater strength in the n direction . for the e - w slit position ( pa = 90@xmath4 ) shown in figure 5 the diamond and continuum distributions are the same within the signal - to - noise .
in contrast , the pah distribution is very broad and skewed slightly towards w. the deconvolved continuum - subtracted 3.53@xmath0 m elias 1 diamond feature shows a symmetrical distribution n , s , e and w of the central star .
this is consistent with the circular symmetric component described by @xcite for k and l@xmath13 who noted that the latter filter at 3.69 @xmath24 0.315 @xmath0 m includes the strong 3.42 and 3.53@xmath0 m features observed in our study and that are now known to be due to diamond emission .
the distribution for diamond emission has a fwhm of 0.34@xmath25 giving a gaussian fwhm diameter of @xmath148 au where a distance to elias 1 of 140 pc @xcite was adopted .
the continuum profile that is generally taken to arise from thermal emission by larger grains is very similar .
the pah distribution is extended relative to diamond and continuum emission along both slit axes but more so for e - w than n - s .
for e - w a signal is discernible to @xmath260.7@xmath25 or @xmath26100 au w and e relative to the star .
this long - slit spectroscopic result agrees well with the findings of @xcite who found a narrow e - w elongated blue halo in j and h with fwhm of about 1@xmath25 , the n - s extent only being about 0.2@xmath25 .
we conclude that the new data provide strong evidence for a pah - containing bipolar nebula oriented close to e - w but , given the asymmetry in the pah distribution along both e - w and n - s ( see figs 4 and 5 ) , tilted slightly with respect to the observer .
m ) - solid line , diamond ( 3.53@xmath0 m ) - dashed line , and continuum ( 3.60@xmath0 m ) - dot - dashed line . ]
m ) - solid line , diamond ( 3.53@xmath0 m ) - dashed line , and continuum ( 3.60@xmath0 m ) - dot - dashed line . ]
it is of interest to compare the results for elias 1 with those for hd 97048 .
there is a clear similarity in that they both exhibit pah , diamond and ir continuum emission , but they have very different seds and fall in different classifications being in group ii with a large ir excess ( elias 1 ) and in group i ( hd 97048 ) with a circumstellar disk seen pole - on @xcite .
based on speckle observations , @xcite reported that the 3.53@xmath0 m diamond emission of hd 97048 is not substantially extended , arising within 0.1@xmath2 of the star .
@xcite more recently described an adaptive optics high angular resolution ( _ c. _ 0.1@xmath2 ) long - slit spectral study of the spatial distribution of diamond and continuum emission for hd 97048 .
it was deduced that the fwhm sizes are consistent with predictions for a circumstellar disk oriented pole - on , with the diamond emission originating in an inner region . the diamond fwhm emission for hd 97048
is reported by @xcite to be 41 au and that of the continuum 32 au , which compare with our elias 1 results of @xmath10.34@xmath2 ( 48 au ) for both diamond and the continuum .
@xcite have reported spatially resolved 3.3@xmath0 m pah emission for hd 97048 which is more extended than the diamond and continuum emission and hence similar to our results for elias 1 .
the pah distribution data for hd 97048 is incomplete due to a noisy section near 3.3@xmath0 m falling in the + 0.0 - 0.2@xmath2 region and a value for the fwhm was not quoted .
however , from the negative offset data ( figure 3 of the paper ) and assuming a symmetrical spatial profile , a fwhm of @xmath360 au can be estimated .
an extended 3.3@xmath0 m distribution for hd 97048 is also consistent with a comparison between the iso - sws ( 20@xmath2 x 33@xmath2 ) spectrum and the spectra of habart et al .
( 2004 ) within 1@xmath2 of the star which showed the pah 3.3@xmath0 m band to be stronger relative to the diamond bands in the iso spectrum . resolved pah profiles in other herbig ae / be stars have been found and successfully modelled by @xcite . in general the 3.3@xmath0 m pah feature tends to be more confined spatially than the uir bands at 6.2 , 7.7 , 8.6 and 11.2@xmath0 m .
this is so for hd 97048 which extends _ c. _ 5 - 10@xmath25 in the mid - ir @xcite and from limited nulling interferometric observations is reported to be the case for elias 1 @xcite .
m ) - solid line , diamond ( 3.53@xmath0 m ) - dashed line , and continuum ( 3.60@xmath0 m ) - dot - dashed line , where the data for different nights ( see table 1 . ) are co - added and a rl deconvolution applied . ]
m ) - solid line , diamond ( 3.53@xmath0 m ) - dashed line , and continuum ( 3.60@xmath0 m ) - dot - dashed line , where a rl deconvolution has been applied . ]
we have conducted a long - slit study of elias 1 in the 3.2 - 3.6@xmath0 m region which has revealed a symmetrical distribution for the continuum ( 3.6@xmath0 m ) and diamond c - h emission features , and more extended pah 3.3@xmath0 m emission which is oriented along the e - w axis . given that for many objects the 3.3@xmath0 m pah emission is significantly less extended than for the uir bands at longer wavelengths
, it will be of interest to determine the mid - ir pah emission distribution using long - slit spectroscopy in order to probe the size and geometry of the bipolar nebula .
we thank the uk panel for the allocation of telescope time for the award of observing time on ukirt , pparc for a studentship to rt and epsrc for a studentship to jr . | we report long - slit spectroscopic observations of the herbig ae / be star elias 1 in the 3.2 - 3.6@xmath0 m region covering the c - h stretch emission features of hydrogen - terminated diamonds and pahs .
the data were recorded at ukirt using uist and yield information on the profiles and intensities of the bands as a function of offset along the n - s and e - w axes centred on the close binary .
the diamond and nearby ir continuum emission arises from a symmetrical inner core region ( @xmath10.34@xmath2 or 48 au ) .
the 3.3@xmath0 m pah emission is extended along the e - w axis up to _ c. _ 100 au each side of the star .
this result supports a suggestion of @xcite of an e - w oriented bipolar nebula in elias 1 .
[ firstpage ] stars : circumstellar matter stars : binaries stars : individual : v892 tau |
the challenge of x - rays in o stars was posed by einstein x - ray observatory observations of the carina nebula ( seward et al .
1979 ; seward and chlebowski , 1982 ) and cyg ob2 ( harnden et al . 1979 ) .
four particularly strong sources were found in tr14 and tr 16 : @xmath0 car itself , hd 93129a , b ( o3 if ) , hd 93162 = wr25 , a wolf - rayet star , and hd 93250 ( o3v ) .
all the bright o stars in the einstein field were detected in x - rays . among the clues to the source of x - rays was that l@xmath1 is approximately seven orders of magnitude smaller than l@xmath4 for all o stars .
the w - r star wr25 is unusually strong in x - rays compared with two other w - r stars in the region .
several x - ray studies have been made subsequently , including rosat images ( corcoran et al .
1995 ) and crowther et al .
( 1995a ) , both of whom confirmed that wr25 is unusually bright for a w - r star .
the region around @xmath0 car is of considerable interest .
@xmath0 car itself is arguably the most massive star in the galaxy , with a history of luminosity variations .
a hubble space telescope image revealed a bipolar structure in the center of the nebula .
the carina nebula contains the most massive stars in the galaxy with initial masses up to 120 m@xmath5 .
walborn ( 1971 ) invented the spectral classification
o3 " because several stars in this region are hotter than existing mk classifications .
recent studies include the massey and johnson study ( 1993 , mj93 hereafter ) of reddening from spectroscopy and photometry , and the proper motion study of cudworth et al .
( 1993 ) , which provides membership probability .
the clusters tr14 and tr16 provide a snapshot of evolution of the most massive stars known . the evolutionary sequence in order of decreasing mass goes from @xmath0 car through the o3 if star through o3 v stars . a particularly perplexing question in this scenario
is how the w - r star wr25 fits in .
this is discussed at some length by mj93 .
it should be more evolved than the o3 stars ; however , it is less luminous .
this statement , of course , depends on the bolometric corrections , which are difficult to determine for o stars and especially difficult for w - r stars .
wr25 is classified as wn7+abs by mj93 .
wn7+abs stars belong to the nitrogen sequence of w - r stars which are the most luminous and least chemically evolved as mass loss strips off outer layers of the original star .
they still retain some hydrogen , and the abs " designation indicates that there are absorption components in the upper balmer lines .
wn7+abs stars differ from of stars only in having a stronger wind . as noted by crowther et al .
( 1995b ) , wn7+abs stars are found only in the youngest clusters , in contrast to wn7 stars .
the scenario put forward by crowther et al .
( 1995 a and b ) , based on their spectroscopy , envisions that wn7+abs stars are part of the most massive sequence of stars , immediately following the evolutionary stage of of stars , but preceding wn7 stars .
( wn8 stars belong to a less massive sequence containing stars which have already been luminous blue variables or red supergiants . )
this scenario accords with the selection of objects in tr16 and tr14 ( wn7+abs , o3 if * , o3 v , and cooler stars ) .
tr 16 and tr 14 contain some of the youngest and most massive stars known , and have been well - studied . for the same reason , many of their parameters are extreme , and also controversial .
the distance , for instance , adopted by davidson and humphreys ( 1997 ) to @xmath0 car itself is 2.3 kpc .
mj93 find 3.2 kpc from extensive spectroscopy and photometry . among the problems in determining the distance
to tr 14 and tr 16 are the form of the reddening law , and whether the two clusters have the same age and distance .
walborn ( 1995 ) illustrates quantitatively how the reddening law affects the derived distance . for r = a@xmath6/e(b - v ) = 3.5 , his distance to tr 16 is 2.5 kpc .
crowther et al .
( 1995b ) use 2.6 kpc in their discussion of the wolf - rayet in tr14 , star wr25 .
we adopt 2.5 kpc for this study . given the dramatic appearance of the nebulosity in optical images of the cluster ,
the reddening within tr14 and tr16 is surprisingly uniform .
the values of e(b - v ) for the hot stars from mj93 have a scatter of approximately @xmath7 0.2 mag , centered approximately on 0.55 mag .
the form of the reddening law in the region , however , has been debated .
this is discussed further in section 6 .
bolometric corrections are similarly problematic in comparing very luminous hot stars with w - r stars . a good overview of x - ray studies of o stars and wolf - rayet stars in the pre - chandra era is provided by willis and crowther ( 1996 ) .
the search for correlations between x - ray luminosity and other parameters has been elusive , and is summarized by chlebowski ( 1989 ) .
the observations of @xmath0 car itself from the early chandra image have already been discussed by seward et al .
we now turn to the stars in the surrounding region .
the discussion will be broken into two papers .
the current paper discusses the source counts , the luminosity distributions , and hardness ratios .
the second paper will discuss the low - resolution spectra of the strongest sources .
the current paper is broken up into sections on the observation , detected sources , cool stars , luminosity distributions , x - ray to bolometric luminosity ratios , diffuse emission , upper limits , and hardness ratios .
the early chandra science image placed @xmath0 car at the focus of the i3 chip of the of the acis - i detector .
instrumental details about the chandra satellite are given in weisskopf et al .
unfortunately , the @xmath0 car observation was obtained during the period when the energy resolution of the detector was degrading .
it was also obtained at an instrumental temperature which has not been fully calibrated .
this means that the acis instrumental low - resolution spectra have uncertainties in the energy scale .
the intrinsic interest of the sources in this image , however , means that there is insight to be gained from an analysis of the image despite these limitations .
1 shows the image on the acis - i chips .
@xmath0 car itself is near the center of the image , with stars in trumpler 16 particularly prominent to the south .
trumpler 14 is in the extreme upper right .
the identification of several of the sources ( hot stars ) is provided in fig . 1
. the data were analysed using the cxc ciao ( chandra interactive analysis of observations ) software .
the data were filtered on standard asca grades and checked for times of poor aspect ( pointing stability ) ; none were found .
there were no significant background flares during the observation , giving a total good exposure time of 9.63 ksec .
sources were detected using the wavdetect program which performs a mexican hat wavelet decomposition and reconstruction .
wavdetect was run using scales of size 1 , 2 , 4 , 8 , and 16 pixels on an image which had been energy filtered to contain data from 300 to 6000 ev .
these choices maximized the number of total detections returned while minimizing the number of spurious detections .
the data were then split into three energy bands : 0.5 - 0.9 , 0.9 - 1.5 , and 1.5 - 2.04 kev .
an exposure map was computed for the observation to take into account vignetting and chip gaps .
a mean energy of 1.2 kev was used in generating the exposure map , which was normalized to its aimpoint value .
source counts were extracted using the regions supplied by wavdetect .
these counts were divided by the normalized mean exposure value , and background - subtracted using data from an annulus centered on the source region .
sources within the background region were subtracted prior to this computation and the area was adjusted accordingly .
the list of detected sources was culled in the following way to produce a list of significant sources .
the distance off - axis for each source was calculated and for this distance the size of the point spread function ( psf ) was determined .
see , for instance , fig .
6.4 in the chandra proposers observatory guide ( 2000 ) .
( specifically , a fit by j. grimes [ 2001 , private communication ] for 80% encircled energy at 1.5 kev was used . )
background counts were measured in two areas of the image and found to be 0.045 counts per square arcsec .
the background within the psf was calculated for each radius . for this background ,
the number of counts required for a 3 @xmath8 detection was computed using the approximate formula from gehrels ( 1986 , p. 339 , recommended approximation " ) .
using this approach , for instance , an on - axis source requires approximately 6 counts to be significant ; a source 7 arcmin off - axis requires approximately 12 counts .
the resulting sources are listed in tables 1 to 4 .
separated by their optical identifications into ob stars ( table 4 of mj93 ) and other sources .
the other sources in table 2 lists are predominately pre - main sequence stars .
sources within the @xmath0 car nebula have been omitted from the tables . as well as
total counts from 0.5 to 2.04 kev , counts were measured in 3 energy bands , 0.5 to 0.9 kev , 0.9 to 1.5 kev , and 1.5 to 2.04 kev , referred to as soft ( s ) , medium ( m ) , and hard ( h ) bands .
errors are computed by dmextract using poisson statistics .
tables 1 and 2 also list the radial distance r from the aimpoint ( in arcmin ) . tables 3 and 4 list quantities derived from source counts , again divided into ob stars and other sources .
hardness ratios have been calculated from the 3 energy bands : @xmath9 @xmath10 errors have been propagated from the errors on the energy bands .
in addition , fluxes and luminosities have been calculated from total count rates .
pimms was used to calculate a conversion from counts to energy .
the count rates were calculated on the basis of 9.630 ksec of good time .
a raymond - smith plasma model spectrum with a log temperature of 6.65 was used , a spectrum typical of a stellar source found in young associations .
the extinction is fairly uniform over the field . a column of n@xmath11 of 3 x
10@xmath12 @xmath13 was adopted , corresponding to e(b v ) = 0.52 mag , which is typical of the field ( mj93 ) . with these parameters , a conversion factor of 1.4
x 10@xmath14 ergs @xmath13 sec@xmath15 per count per sec was derived .
conversion factors differed by less than a factor of two for a range of log t from 6.4 to 7.1 .
luminosities were derived from fluxes using a distance of 2.5 kpc .
figure 2 shows the acis - i image of @xmath0 car overlaid with the detected sources .
extensive optical work has been done on the clusters in the vicinity of @xmath0 car , most recently by mj93 and cudworth , martin , and degioia eastwood ( 1993 ) .
table 4 in mj93 , the brightest and bluest stars near @xmath0 car , has been used to identify o and b stars with x - ray sources ( shown as circles ) .
sources which are not identified with o and b stars are shown as diamonds .
( note that the increase in the psf of sources which are farther off axis is not reflected in the plotting symbols . )
presumably most of the x - ray sources which are not o stars are pre - main sequence stars , although very rarely they could be background objects , or even extragalactic sources .
figure 3 shows the acis - i image of @xmath0 car with the o and early b stars from the mj93 list marked .
it is clear many of the x - ray sources are early type stars , but many are not .
furthermore , many of the stars on the o and early b star list are not x - ray sources .
figure 4 shows a plot of v magnitude vs spectral type of the o and early b stars from the list of mj93 .
( a few of the hot stars are not plotted because spectral types are not available . )
the observed v magnitude has not been corrected for reddening .
however , extinction is fairly uniform over the field , so the pattern is not changed much if v@xmath16 is used instead .
an exception to this statement is the o3 star at 10.78 mag .
this star was recognized by mj93 to be unusually heavily reddened ( mj 257 ) , and will be discussed further below .
the 06.5 star which is shown as not detected lies in the extreme upper right corner of the acis - i image ( fig . 1 or 2 ) , considerably off - axis , where a modest source could easily fail our detection criterion . with this exception ,
the dividing line between detections and nondetections is very sharp at just approximately b0 .
this confirms previous results for o stars .
there is one b1.5 star which is detected in x - rays .
a sole exception like this could easily result from a later companion .
it will be discussed in section 10.3 .
figure 5 shows the total x - ray counts ( table 1 ) as a function of spectral type for the hot stars .
the field corresponding to the chandra image of @xmath0 car was observed january 9 - 10 , 2001 using the 0.9 meter telescope at ctio .
the tek-2k ccd has a 13.5 field of view .
first , four pointings , arrayed in a cruciform pattern with @xmath0 carina near the edge of each field , were used to observe the entire field .
johnson b and v filters and a kron - cousins i filter were used , with exposures of 2 minutes each .
this gives a limiting magnitude of about 19 at v. we performed an additional exposure of 10 minutes with each filter , centered on @xmath0 carina .
this gave us a limiting magnitude of about 20.5 at v within 6.75 of the aimpoint .
basic data reduction was handled using the quadproc package within iraf to bias correct and flat - field the data .
fluxes for all standards and target stars were determined using the fitpsf tool within the iraf apphot package . using this tool fluxes
were derived from the analytic fit to each individual psf in lieu of simple aperture photometry .
stellar magnitudes and colors were put on the standard system using the photcal package .
there were three difficulties encountered during data analysis .
first , the region has a highly variable diffuse background .
second , @xmath0 carina and other bright stars make photometric measurements of stars in their proximity difficult . finally the region is quite crowded , and often faint stars have overlapping psfs .
optical data are presented in table 5 , ordered by source number from tables 2 and 4 .
problems are noted , as appropriate according to the code at the end of the table .
photometric errors were determined using fits to the standard stars and by comparing observed values of target stars observed multiple times in separate pointings .
this gives a more realistic assessment of the errors given the highly variable and confused nature of the field .
errors are less than 5% at v and less than 8.3 % in the colors down to the completeness limits .
these errors are somewhat conservative because a star near the center of one field is on the edge of the comparison field , thus the psfs may differ greatly . the central image with the x - ray sources overlaid is shown in figure 6 .
the resulting color magnitude diagram is shown in figure 7 . in this figure ,
x ray sources near trumpler 16 are shown as asterisks , those near trumpler 14 are shown as open circles .
other sources are shown as small dots .
there are 6219 stars plotted .
all stars detected in b , v , and i and brighter than our completeness limit in the deep center field exposure are plotted .
7 shows no significant difference between the age inferred for the trumpler 16 sources compared with those near trumpler 14 , although the statistics are small .
the figure demonstrates that the color magnitude diagram easily distinguishes stars of carina arm of the galaxy from nearby stars .
the x - ray sources are all in a band about 1 magnitude wide , located about one magnitude above the field stars in the carina arm .
the well defined sequence of field stars ( dots ) results because they , like the cluster stars , are concentrated in the carina arm .
heavy obscuration behind the clusters restricts background contamination .
the large spread among the stars is similar to that seen in orion and is common among stars less than 1 myr old when small differences in age can lead to large differences in luminosity .
about 100 stars in this region of the hr diagram are not detected in x - rays .
there are a few x - ray sources about 0.75 magnitudes above the upper envelope of the x ray population ; we interpret these as binary stars .
the reddest stars have v - i of about 2 .
given a@xmath6 = 1.95 , the dereddened v - i at the v=19 is 1.28 . at an age of 1 myr , this corresponds to a mass of about 1.15 m@xmath17 ( baraffe et al .
figure 8 shows the x - ray luminosity distribution of the o and early b stars ( table 3 ) .
9 shows the luminosity distribution for all sources which are not identified as ob stars ( table 4 ) , the pre - main sequence stars . a or b stars with later companions
might be included , as could any extra - galactic objects , but fig .
9 is heavily dominated by late type stars .
the luminosity distribution is distinctly different between these two groups .
most of the ob stars have luminosities greater than 10@xmath18 ergs s@xmath15 .
the cooler stars have luminosities strongly peaked at 10@xmath18 ergs s@xmath15 , with the luminosity distribution of cool stars ( fig .
9 ) truncated at the faint end because of the relatively shallow x - ray image . as discussed below in section 8
, the depth of the x - ray exposure allowed all o stars near the image center to be detected .
in this section we explore the efficiency of x - ray production from hot stars ( table 6 ) .
the first column is the optical identification , taken from table 4 of mj93 .
the second column is the x - ray source number from table 3 .
columns 3 and 5 are the v and e(b - v ) ( in magnitudes ) , again from mj93 .
column 4 is the spectral type .
the reader is referred to table 4 in mj93 for the comparison of spectral types from different sources and in some cases more complex spectral types .
recently walborn et al .
( 2002 ) have comprehensively reexamined the spectral classification of luminous stars .
they make small adjustments to some classifications , including introducing new classes o2 and o3.5 .
the new classifications for stars in table 5 are hd93129a : o2 if * , hd93129b : o3.5 v((f+ ) ) hd93250 : o3.5 v((f+ ) ) , hd93205 : o3.5 v((f+ ) ) , and hde303308 : o4 v((f+ ) ) .
this finer gradation would only mildly affect the figures below .
the sixth column is the bolometric magnitude , taken from table 3 in massey et al .
( 2001 ) where available , made fainter by 0.56 mag because of the different distance assumed .
( the x - ray to bolometric luminosity ratio which we use below , of course , does not depend on the distance . ) for other stars , bolometric corrections were computed from the same formulae ( massey et al .
the final column is the log of the ratio of x - ray luminosity to the bolometric luminosity .
some of the sources in table 6 have been identified in fig . 1 .
while the reddening through the cluster field is reasonably uniform ( mj93 ) , there been studies that have suggested that the reddening law is unusual ( e.g. herbst , 1976 ) .
other studies have refuted this , or concluded that only some objects have unusual reddening laws ( turner and moffatt , 1980 ; tapia et al .
1988 ) . one case where the reddening law does appear to be unusual is the w - r star hd 93162 = wr25 , although the unusual spectrum of the w - r star presents a problem in reddening determination .
crowther et al .
( 1995b ) discuss the flux distribution from iue through the ir and conclude that the reddening law for wr25 is unusual [ e(b - v ) = 0.66 with r = 4.6 ] .
this is the reddening that we adopt .
the unusual spectrum also complicates the determination of the bolometric correction .
we adopt the bolometric magnitude from crowther et al .
( 1995a ) , -10.1 mag .
( a small correction has been included for the difference between the distance they use and the one used here . )
there are a few unusual stars among the hot x - ray sources .
tr 16 - 244 is a faint star with an early spectral type ( o3-o4 if ) .
mj93 took particular care in their observations because of this , and because the star lies close to the bright w - r star hd 93162 .
tr 16 - 244 is marked in fig . 1 ,
and the w - r star is just to the right .
the reddening implied for tr 16 - 244 by the spectral type and photometry of mj93 is larger than other stars in the cluster , with e(b - v ) = 0.94 mag ( using the relations in massey et al .
2000 ) . tr16 - 244 and hd
93162 both lie very close to the band of obscuration in the visual image . however , the w - r star is between tr16 - 244 and the dust lane . while reddening is difficult to determine for w - r stars , the value found by crowther et al .
( 1995b ) for the total extinction in the visual of the wolf - rayet star , ( a@xmath6 = 3.0 mag ) is very close to that for tr16 - 244 .
thus , it is not clear whether the large reddening for those two stars is caused by the dust lane , or could have a circumstellar component ( for each star ) . for tr 16 - 244
, since the spectral type is the same as for hd 93129ab , the same temperature and bolometric correction were used in creating table 6 .
the resulting m@xmath4 is -8.6 mag , although there are obviously several sources of uncertainty in this number .
the other hot star with surprising photometry is tr16 - 22 .
the v magnitude is faint for a spectral type of o8.5 v , suggesting a high reddening .
tr16 - 22 is very close to the dust lane that cuts diagonally through the lower left of the image ( fig .
1 ) , so the foreground absorption may be particularly large for this star .
the spectrum mj93 show is similar to other o8.5 v spectra , except that the lines are slightly shallower and broader .
however , the value of q from the photometry gives rather unconvincing values for temperature and bolometric correction .
comparing the b - v and u - b values from mj93 with those of herbst ( 1976 ) and feinstein ( 1969 ) shows that the agreement between the latter two is much better .
for this reason , we adopt e(b - v ) = 0.78 mag from the mean b - v and q of herbst and feinstein . this provides a bolometric magnitude very close to the two o8.5 v stars tr16 - 115 and tr16 - 3 , so we adopt the bolometric magnitude from these stars ( -6.8 mag ) .
10 shows the l@xmath1/l@xmath4 ratio for the stars in table 6 as a function of spectral type .
an approximate adjustment to the x - ray luminosities of hd 93162 , tr 16 - 244 , and tr 16 - 22 of a factor 2.45 was made because of their larger reddening which would correspond to an e(b - v ) = 0.90 mag ( estimated using pimms ) .
as is known from previous studies , for most stars in the o and early b spectral range , the l@xmath1/l@xmath4 ratio remains constant at -7 .
there are 3 exceptions to this in fig .
the w - r star hd 93162 has been known since the x - ray discovery paper ( seward et al .
1979 ) to be unusually strong in x - rays as compared with other w - r stars .
this has been particularly surprising as it has been thought to be a single star .
binaries with components this massive are often colliding wind binaries and hence stronger x - ray sources than single stars .
however , the recent w - r catalogue by van der hucht ( 2001 ) lists for the first time a possible binary indicator , diluted emission lines " . a companion bright enough to be detected
photometrically would probably be luminous enough to have a wind which would enhance x - ray production .
tr16 - 22 ( source 55 ) also has an unusually large x - ray luminosity in comparison with its bolometric luminosity ( l@xmath1/l@xmath19 -5.8 ) .
as noted above , it has a larger than average absorption , possibly because it is on the edge of the dust lane .
however , the large x - ray luminosity suggests the source itself may have unusual characteristics beyond simple foreground reddening .
the x - ray luminosity of the source ( l@xmath1 = 7.9 @xmath7 0.5 x 10@xmath18 erg sec@xmath15 in table 3 before the adjustment for high reddening ) makes it one of the brightest o stars observed , approximately a factor of 10 larger than other o8.5 v stars .
the third source with a large l@xmath1/l@xmath4 ratio is tr16 - 11 , the b1.5 v star ( source 141 in table 1 ) .
fig 10 shows that a b star would not even be detected unless it is unusually strong in x - rays . the luminosity ( l@xmath1 = 1.4 @xmath7 0.5 x 10@xmath18 erg sec@xmath15 ) is within the range of the cool star distribution ( fig .
9 ) so a late companion is a possibility . mj93
display its optical spectrum , which looks very like others of similar spectral type , except that perhaps the hydrogen lines are slightly broader .
the spread in l@xmath1/l@xmath4 ratio in fig .
9 is particularly large among the o3 stars .
the two with the largest ratio are hd 93250 and tr 16 - 244 .
the x - ray production in cool stars was similarly investigated using the x - ray luminosity ( table 4 ) and the optical luminosity from the data in table 5 . in this case , the typical extinction for the clusters [ e(b v ) = 0.52 mag ] was assumed for all the cool stars , as was the standard distance .
bolometric corrections were interpolated table 15.3.1 in _ astrophysical quantities _ ( cox , 1998 ) .
the v - i ( johnson ) in that table was transformed to the kron - cousins system using the equations provided by bessell ( 1979 ) .
the resulting plot of log l@xmath1 and log ( l@xmath1/l@xmath4 ) as functions of v
i are are shown in figs .
11 and 12 . for comparison ,
the colors of a0v , f0 v , and k7 v stars ( from bessell , 1979 ) are shown with the reddening of the clusters .
note that in this relatively short exposure we only detect a fraction of the stars in the cluster .
for instance presumably many m stars are not detected .
the most informative comparison is with comparable figures for the orion nebula cluster ( onc , chandra results in feigelson , et al . , 2002 ; rosat results in gagne , et al .
stars in the two regions are similar in age , although those in tr 14tr 16 are slightly younger ( meynet , et al . , 1993 ; massey , et al . , 2001 ) .
tr14tr16 contain more massive stars than the onc .
some x - ray sources among a stars are detected in both clusters .
feigelson , et al .
argue that because the l@xmath1 distribution in late b and a stars is the same as for cooler stars , x - ray sources among late b and a stars can be explained by less massive companions .
11 shows the same result for tr 14tr 16 .
to illustrate diffuse emission in the field , a smoothed 0.3 - 7 kev image was created using the ciao adaptive smoothing tool csmooth .
the result is shown in fig .
13 , which indicates broad , diffuse emission from much of the nebula .
we investigated two circular regions 2.5@xmath20 in diameter marked in the figure ; one centered 4.0@xmath20 s of @xmath0 car and one centered 7.5@xmath20 s of hd93129 a / b .
the figure shows 7 unresolved sources in the first region , accounting for 30% of the emission after background subtraction .
there are no unresolved sources in the 2nd region . after subtracting bright stars ,
the residual above background in each region is @xmath21 750 counts , or @xmath21 150 counts arcmin@xmath22 .
this corresponds to a surface brightness of @xmath23 ergs @xmath13 s@xmath15 arcmin@xmath22 ( unabsorbed ) .
the luminosity of this region , if spherical , is 3 @xmath24 10@xmath25 ergs s@xmath15 pc@xmath26 in the orion nebula ( garmire et al .
2000 ) , one of the densest pms clusters known , the luminosity of x - ray emitting stars in 1 pc@xmath27 is @xmath28 ergs s@xmath15 .
either the carina region has a population of young stars denser than in most pre - main sequence regions or much of this emission is truely diffuse .
the x - ray spectra of these regions are thermal .
if the regions have depth approximately equal to angular extent , and the emission is due to diffuse , hot gas , over half the x - ray emission from the carina nebula originates in the diffuse component . a gas density of @xmath290.5 particles cm@xmath26 is required .
o star winds can easily supply the energy .
as is well known , b stars in general do not produce x - rays . because of the extensive optical work that has been done on the stars in the two clusters , they provide a good place to explore the exact x - ray boundary . to do this
, we have derived upper limits to the x - ray flux of the hot stars in table 4 of mj93 which are not detected in x - rays .
fig 14 shows the locations of these stars overlaid on the x - ray image . in tr 16 in the center of the image
, the circled locations appear to be free of sources . in tr 14 in the upper right ,
the situation is more complex .
tr 14 is a crowded region , and sources in that region may have source confusion because of the large psf off - axis .
we have adopted the criterion for source detection ( section 2 ) from gehrels ( 1986 ) .
it is parameterized as a function of the number of background counts , which is a function of the off - axis distance .
we have compared the measured counts at the positions of the o and b stars in mj94 ( table 4 ) which are not in table 1 with the detection level predicted by the gehrels criterion .
counts were measured in 2@xmath315 circles centered on these positions .
for the center of the field , this is a generous allowance for the psf ( 0@xmath3168 ) plus an uncertainty in the pointing of 1 " .
figure 15 shows that the measured counts are typically well below the detection criterion .
( note that the counts from which sources were selected for tables 1 - 4 were based on wavdetect results ; those in fig 13 are from extracted counts with no background subtraction . )
the detection criterion used by green et al .
( 1995 ) for a qso survey is c@xmath32 , where c@xmath33 and c@xmath34 are the counts in the source and background respectively . for the very low background in the center of a 1 ksec exposure , as seen in fig 15
, one background count corresponds to a detection limit of 4 source counts .
as summarized in fig 15 , the gehrels criterion provides a conservative upper limit for sources concealed in the measured counts .
further from the axis , the psf , and hence the included background counts , increase .
for comparison with the detection criterion , counts for locations greater than 7 off axis were measured in circles of 5 `` and 10 '' radius as well as 2@xmath315 ( not shown in fig 15 ) . for radii less than 10 off axis , the psf is less than 10 `` .
counts in the 10 '' circles for several of the off - axis locations are above the detection criterion . as discussed below , there are probably actually sources at several of these locations .
we note also that the tr 14 region is a dense and complex one , and it is highly likely that additional sources are included as the measurement area is increased . because of this complexity
, further discussion of the upper limits far off axis does not seem warranted . with these explorations in mind ,
we adopt the gehrels limit as the upper limit of counts at the location of early b stars .
the measured counts at the location of o and b stars from mj93 table 4 ( which are not detected sources ) is listed in table 7 . for each location ,
table 7 provides the number ( independent from the numbering in tables 1 to 4 ) , the position , the counts in the source and background areas , and the net counts , all with their errors .
counts are measured in a circle with a 5 `` radius ; background counts are measured in an annulus between 5 '' and 10 " radii .
table 7 lists the radius from the aim - point for each source and the spectral type ( where available from mj93 ) .
spectral types are abbreviated to the temperature portion for plotting purposes .
upper limits to the luminosity ( in ergs sec@xmath15 ) are computed for the gehrels criterion and the distance to the clusters .
fig 16 shows the _ net _ counts at the location of hot stars .
counts were measured in a 2@xmath315 circle around the stellar positions .
background counts were measured an annulus between 5 `` and 10 '' .
the scatter around measures for tr 16 near the center of the field is well balanced around zero , confirming that these are non - detections . for stars
further off axis , the 2@xmath315 circle , of course , does not include the full psf , and hence some source counts will be included in the region which defines the background " counts .
even though the measures will be over - corrected " for background , the tr 14 stars have predominately positive source counts .
this indicates that the central concentration in the psf is such that a number of tr 14 stars in fact are probably x - ray sources , even though they do not pass the detection criterion .
fig 17 shows the upper limit luminosity from table 6 for hot stars not detected in x - rays as a function of spectral type .
the two o stars which were not detected have comparatively high upper limits because they are at the edge of the image .
the counts corresponding to these upper limits are at or slightly above the lower envelope of the o star detections in fig 5 .
had they been at the image center , they might have been detected .
on the other hand , the b stars from the center of the image have a typical count upper limit ( log lum = 30.8 ergs sec@xmath15 ) which is below the counts of any detected hot star in fig 8 .
we can use fig 4 to examine the dividing line between detected x - rays and the lack of x - rays . with the exceptions noted above , all o
stars are detected .
for the b0 spectral types , two stars are detected and one is not .
cooler b stars are not detected with one exception .
this points to a very clean divide in spectral type at b0 .
it also provides us with a list of 14 stars in tr16 for which we can constrain the properties of a cool companion star .
this upper limit only overlaps with the weakest sources in the distribution of x - ray luminosities ( fig 9 ) .
while we would not yet have detected a companion comparable to the faintest x - ray sources among late type cluster members , we can say that the b stars do not have x - ray bright companions .
the hardness ratios of x - ray sources provide the simplest description of their spectra . in this section
, we explore the hardness ratios ( tables 3 and 4 ) as defined in section 2 for both hot stars and other x - ray sources .
we limit the sources examined to the brightest so the errors in the ratios are small enough so that they do not obscure any trends or comparisons .
we have examined the distribution of errors in the hardness ratios as a function of luminosity . for the o - b stars ( table 3 ) , if we use only sources with log l@xmath1 @xmath35 32.2 ergs sec@xmath15 , the uncertainty in the soft hardness ratio , hr@xmath36 is less than or equal to 0.1 ( with one exception ) ; for the hard hardness ratio , hr@xmath37 , the error is less than or equal to 0.15 .
the other stars ( table 4 , referred to as cool stars below ) are somewhat weaker in x - rays than the hot stars , so we use sources with log lum @xmath35 32.0 ergs sec@xmath15 .
this results in uncertainties less than 0.2 in both colors ( after two sources with large errors were removed from the sample ) .
our sample is thus limited to only the brightest sources .
fig 18 shows the resulting hardness ratio plot .
the hardness ratio created from the soft and medium bands ( 0.5 to 0.9 kev , 0.9 to 1.5 kev ) can be either positive or negative , i.e. , either band can have more counts whereas the hardness ratio from the medium and hard band ( 1.5 to 2.04 kev ) is always negative , i.e. the hard band always has fewer counts .
the hot stars fall in a diagonal sequence with a scatter consistent with the uncertainties in the colors .
there are only two cool stars with well determined hardness ratios , which are both are very hard in fig .
we have explored this locus further by creating hardness ratios in the same bands using pimms ( v3.2 ) . the basic model ( solid line in fig 18 )
is a raymond - smith model with solar abundance and a temperature range from log t = 6.0 to 7.2 . a neutral column density n@xmath11 = 3 x 10@xmath12 cm @xmath22 was used corresponding to the typical extinction for the field e(b - v ) = 0.52 mag .
for two temperatures , n@xmath11 was increased to 5 x 10@xmath12 cm @xmath22 , corresponding to e(b - v ) = 0.9 mag , an extreme value found for a few stars in the field .
the result of this difference in extinction is shown in fig 18 .
for comparison , we have computed hardness ratios for 0.2 solar abundance ( dashed line in fig .
we have also computed models for thermal bremsstrahlung ( kt from 0.2 to 1.0 kev ) .
fig 18 shows that a range of hrs ( the soft ratios ) can be obtained with any of the models .
however , the observed ratios for both hot and cool stars lie systematically at larger ( harder ) values of hrh than the computed values for either raymond - smith plasma or bremsstrahlung . although the bremsstrahlung ratios for the hottest temperature match the ratios for the hot stars , for cooler temperatures they are a poorer match than the raymond - smith models .
based on these models , the hrs values for the hot stars indicate a range of temperatures from approximately 3 x 10@xmath38 to more than 10@xmath39 . how are these x - ray hardness ratios related to the optical ( photospheric ) properties of the stars ? in fig 18 , the 3 hot stars with negative hr@xmath36 are hde 303308 [ e(b - v ) = 0.46 mag ; o3 v ] , hd93205 [ e(b - v ) = 0.40 mag ; o3 v ] , and -59@xmath402600 [ e(b - v ) = 0.51 mag ; o6 v ] . in other words ,
all three have reddening typical of the field .
the stars with hr@xmath36 near 0.1 are hd 93250 [ e(b - v ) = 0.49 mag ; o3 v ] and hd 93160 [ e(b - v ) = 0.31 mag ; o6 iii ] . interpreting the hardness ratios in terms of photospheric temperature arrives at the contradiction that the x - ray temperatures are hotter than the o3 v stars in the first group , although the photospheric temperatures are similar or cooler .
the four hot stars with distinctly harder hr@xmath36 are tr 16 - 22 ( o8.5 v , e(b - v ) = 0.78 mag ) , tr 16 - 244 ( o3 i , e(b - v ) = 0.94 mag ) , hd 93129ab ( o3 i , e(b - v ) = 0.55 mag ) , hd 93162 = wr 25 ( wn7 + abs , e(b - v ) = 0.66 mag ) .
none of these are normal main sequence objects .
hd 93129ab and tr 16 - 22 are both among the hottest stars known ( o3 ) , and also very luminous . the fact that hd 93129ab apparently has the normal reddening for tr 14 suggests that the very hard hrh ratio does not result from circumstellar material .
this is in contrast to tr 16 - 244 which has the same temperature and luminosity , but is more heavily reddened that the average cluster member .
both hd 93162 and tr 16 - 22 stand out in fig .
10 because they have a particularly large l@xmath1/l@xmath4 .
we already know that hd 93162 is unusual , since not only has it evolved to be a very early stage in the wolf - rayet sequence , it is also unusually x - ray luminous for a wolf - rayet star .
tr 16 - 22 , on the other hand , is cooler and fainter than stars evolving off the main sequence in the clusters .
its unusually hard and bright x - ray luminosity must be linked with some unusual property , such as a massive binary companion .
the two cool stars with well - determined hardness ratios are both among the hardest stars in the field .
interestingly , one is located within the range of the cool pre - main sequence stars in fig . 7 , but one is one of the x - ray bright a stars .
fig 19 confirms that the hrs ( the soft hardness ratio ) is not correlated with photospheric temperature indicated by spectral type .
the o3 stars , in particular , cover a very large range in hrs .
how do we interpret the hardness ratios of the x - ray sources in tr 14 and tr 16 ?
there seem to be two important considerations .
first , we are looking at a biased sample .
because we have restricted the sample to sources with reasonably well determined ratios , we are only looking at the most luminous sources .
second , both the hot and cool stars have more counts in the hard band than predicted from the simple pimms calculations .
a plausible interpretation for this is that we have used a one temperature model in the calculations , but the sources may actually have a second harder component .
this is frequently seen in spectral fits of stellar sources .
we note that the discussion of the individual sources above rules out single parameter explanations of the hardness based on photospheric temperatures or circumstellar reddening . in summary ,
unusually x - ray bright sources are surprisingly hard .
we have assembled the x - ray properties of sources in tr 14 and tr 16 , including several sources which are unusually strong or hard in x - rays .
we have already discussed properties such as reddening , spectral type and luminosity in relation to the x - ray properties .
one further property ( at least ) may be important in x - ray production , namely whether the star is single , a member of a close binary system , or a member of a wide binary system . in o stars , a binary companion with sufficient mass to drive a wind of its own may enhance x - ray production in via a colliding wind " . a good deal of work has been done to establish whether the stars in the carina clusters are single or multiple .
the most definitive study , with the highest resolution spectra , is that of penny et al .
they concluded that none of the six bright members of tr 14 are short - period spectroscopic binaries .
this is in contrast to an earlier study at lower resolution by levato et al .
( 1991b ) . however , the penny work , and a later study by garcia et al .
( 1998 ) , demonstrate that tr 14 is surprisingly deficient in short - period binaries among the hottest stars .
penny et al . suggest that the cluster has remained compact because the lack of binaries lowers the cluster heating via binary - binary interactions , although the results come from , of course , only the brightest stars .
velocity studies of 26 stars in tr 16 were carried out by levato et al .
( 1991a ) , using the same equipment used for the tr 14 study , with relatively low resolution ( 43 mm@xmath15 ) .
they found velocity variation in at least 50% of the stars studied .
the work by this team in tr 14 was superceded by higher resolution work , and the radial velocity variations they found in tr 16 need to be confirmed
. it would be quite surprising to have dramatically different binary fractions in the two clusters .
three of the x - ray sources in the acis field have recently been observed by the x - mega campaign to obtain optical spectroscopy of the hot stars which are x - ray sources .
colombo et al .
( 2001 ) have observed cpd -59@xmath412635 .
they found that it is a double - line spectroscopic binary and determined an orbit .
both components are massive and colombo et al .
determine spectral types of o8 v and o9.5 v. hd 93205 has been observed by morrell et al .
( 2001 ) , who report that it is another double - lined spectroscopic binary with o3 v and o8 v components .
cpd -59@xmath412636 is a triple system , with o7 v , o8v , and o9 v components ( albacete colombo et al .
components a and b are a binary with a period of 3.6 days .
component c is itself a binary with a period of 5.0 days .
cpd -59@xmath412603 = tr 16 - 104 has been observed by rauw et al .
they find it to be a triple system with o7 v , o9 v , and b0.2 iv components .
the close binary ( o7 v and o9 v stars ) has a period of 2.2 days , and short eclipses .
recently nelan et al .
( 2002 , private communication ) have resolved hd 93129 a with the hubble space telescope fgs ( fine guidance system ) .
this may affect the x - ray properties , although the two components are much more widely separated than the binaries in table 8 , so it is not clear that it does .
we have collected radial velocity information about the single / binary nature of the x - ray sources in the field in table 8 .
it shows the hot stars within the chandra field , the source of velocity information , the binary status , and the orbital period where available .
mason et al .
( 1998 ) provide a convenient summary of velocity work on a number of these stars and also speckle interferometry to sample wider binaries . in table 8
we have marked the velocity variations from levato et al .
( 1991a ) with : " to indicate that they need confirmation .
identifying multiple systems in hot stars with broad lines is not easy .
in addition to the radial velocity studies , as discussed above , van der hucht ( 2001 ) has suggested that the wolf - rayet star hd 93162 may be a binary on the basis of diluted emission lines .
penny ( 1996 ) has suggested that hd 93250 might be a binary on the basis of variable line widths .
this is in accord with the suggestion of a companion ( walborn , nichols - bohlin , and panek , 1985 ) based on unsaturated c iv line cores .
confirmation of the single / multiple status of hd 93250 is a very important question in the interpretation of the x - ray flux of this extremely luminous object .
table 8 can be used to compare what is known about the single / binary nature of x - ray sources that are unusually strong , using , for instance , l@xmath1/l@xmath4 in table 6 . for the wolf - rayet star hd 93162 ,
l@xmath1 is particularly strong , but evidence for a binary system is only indirect . among the o3 stars ,
hd 93250 is a strong , and also hard x - ray source , and again has only indirect evidence of a possible binary companion .
hd 93129ab is a strong and also hard x - ray source , although its l@xmath1/l@xmath4 is not remarkable .
there is no evidence for a close binary companion .
tr 16 - 22 is a remarkably strong x - ray source , but does not have radial velocity studies . on the other hand , hd 93205 , which is a double lined spectroscopic binary ( morrell et al .
2001 ) with an orbital period of 6.1 days and spectral types of o3 v and o8 v , is not a particularly strong x - ray source .
cpd -59@xmath40 2635 is also a double - lined spectroscopic binary ( colombo et al . 2001 ) , with a period of 2.3 days and spectral types o8 v and o9.5 v , and again is not a particularly strong x - ray source .
cpd -59@xmath40 2636 is a similar case of a multiple system with two short period binaries ( albacete colombo et al .
2002 ) , but an unremarkable x - ray flux . in summary ,
the two with particularly hard spectra ( hd 93129ab and hd 93250 ) may well not be binaries . on the other hand three sources in short period binaries made up of o star components ( hd 93205 , cpd -59@xmath40 2635 , and cpd -59@xmath40 2636 ) are not particularly remarkable x - ray sources .
the emerging picture is that while close binaries with two hot stars may produce a strong x - ray flux ( the colliding wind model ) , this does not seem to be the whole explanation for surprisingly strong x - ray sources .
clearly there is the standard problem that there is almost never enough information to prove that a star is single .
however , it is short period binaries with massive companions that should be colliding wind systems , and they are the easiest to find . with the incomplete velocity evidence that is available ,
strong sources and binary orbits are nearly anti - correlated .
we suspect this demonstrates that our knowledge of binary systems is incomplete .
hd 93162 = wr 25 , and hd 93250 are surprisingly strong in x - rays , but with only indirect suggestions of a binary companion .
should it turn out that the x - ray strength does come from a companion , this would provide an important way to get more complete statistics on the binary fraction in massive stars , since such strong x - ray sources are not overlooked .
chlebowski and garmany ( 1991 ) provide comparisons of x - ray flux and stellar parameters such as wind velocity and luminosity for an extensive list of o stars .
they find that in general binaries do not produce larger mass loss than single stars .
on the other hand , close binaries seem to produce more x - ray flux than either contact binaries or wider binaries .
one way to reconcile these trends with the apparent lack of correlation between binary status in table 8 and excess x - ray flux is that we are sampling regimes where different effects become dominant .
this is particularly likely for the group of extreme stars found in the carina nebula : the most massive , the most luminous .
for instance , for the most massive stars , extreme mass loss rates may dominate over binary - single trends and produce an unusually large x - ray flux . for stars less massive than o3 stars ,
the effect of a close binary with a massive companion may become evident in the x - ray flux . as summarized above
, there is no _ velocity _ evidence the hd 93250 is a short period binary system , although there are suspicions that it might have a companion from line - width variations and line depths .
however the recent reexamination by walborn et al .
stress line depth in the context of luminosity not companions .
this paper stresses that complications in the interpretation of these extreme stars because of binary companions must be understood using milky way examples .
even though there are many examples in the magellanic clouds , observations are typically contaminated by a surrounding cluster . if hd 93250 is single ( and there is no clear evidence at present against this ) , it is a pivotal object in understanding the x - ray production and evolution of very massive stars .
it is an extremely strong ( and also hard )
x - ray source in comparison with other o3 stars . among the tr 14 and tr 16 stars
it has a v magnitude second only to hd 93129ab ( mj93 ) .
the hardness ratio hrs , however , lies between the softer o3 stars and the harder but unusual and more luminous hd 93129ab and tr 16 - 244 .
if x - ray flux of hd 93250 is not affected by a companion , it provides a benchmark of an extremely luminous main sequence dwarf with a very high l@xmath1/l@xmath4 produced by the outer atmosphere processes of the star itself .
since it will presumably accelerate its mass loss as it evolves off the main sequence , it provides an important example of high x - ray production of a star before its mass loss becomes extreme .
there is another regime in which x - ray observations provide an important handle on the binary fraction in massive stars .
the upper limits to x - rays from b stars ( fig .
17 ) apply also to the x - ray flux from a cool companion . for instance , the upper limit to the luminosity ( log l@xmath1 = 30.8 ergs sec@xmath42 ) would pick up many of the cool star x - ray sources in the orion nebula cluster ( feigelson , et al .
in other words , a deeper image could rule out most cool companions . quantifying how what fraction of companions would be detected at a given sensitivity level
, however , is not easy .
we can estimate from fig 10 of feigelson , et al . that less than a quarter of the orion cool stars would be detected with the sensitivity limit of our @xmath0 car image ( including a small approximate correction to account for absorption in orion )
. however , fig . 12 in feigelson ,
impiles that companions of a and late b stars have a higher luminosity than the average cool star .
that is , for a given sensitivity limit , a larger fraction of companions would be detected than estimated from the luminosity distribution of late type stars . accurately identifying binary companions much less massive than the primary is very difficult to do , particularly for massive b stars with only a few broad lines to measure velocities from .
x - ray studies provide detection of this otherwise elusive population of companions .
the emerging picture of colliding wind binaries is that they include systems with widely separated components , as well as systems with close components .
wr 140 with an orbital period of 8 years is an example .
our summary of velocity information about binaries in table 8 , and the discussion in section 10.1 indicate that the relation between binary status and strong x - rays is not simple .
the recent xmm - newton spectrum of the wolf rayet star wr 110 ( skinner et al .
2002 ) suggests yet another complication in interpretation .
they find a cool soft component to the spectrum , as expected from a wind - shock model , and in addition , they find a hard x - ray component .
they examine a number of possibilities for producing this component and conclude that it can be produced by a wind and a close but as yet undetected companion . however ,
if the companion is not found , they suggest the possibility that the hard component is produced by the wolf rayet star itself .
the chandra image of @xmath0 car and the clusters tr 14 and tr 16 has provided a number of new results .
@xmath43 although the hottest stars have been known to be x - ray sources since the einstein observatory observations , this is the first time the population of x - ray sources associated with cooler stars has been discussed .
@xmath43 two new anomalously strong x - ray sources have been found among the hot star population .
tr 16 - 244 , a heavily reddened o3 i star , and tr 16 - 22 , a heavily reddened o8.5 v star have surprisingly large x - ray fluxes .
they are both close to the dust lanes which cross the sides of the cluster and which may be responsible for the above average reddening .
@xmath43 several stars have an unusually large l@xmath1/l@xmath4 .
hd 93162 , a wolf - rayet star , has long been known to have an even higher l@xmath1/l@xmath4 than other wolf - rayet stars .
there is some indication that it may be a binary , which could account for this .
tr 16 - 22 also stands out in l@xmath1/l@xmath4 , perhaps indicating a colliding wind binary , although there are no radial velocity studies to provide further information .
the third star with a large l@xmath1/l@xmath4 is tr16 - 11 , the b1.5 star .
the x - ray flux may be produced by a second cool star .
@xmath43 we have put upper limits on the ob stars in the field which we do not detect . for comparison , in this short exposure , if companions to the b stars had an x - ray luminosity distribution like the late type stars in the orion nebula cluster , we would detect only a fraction of the companions .
@xmath43 we discuss the hardness ratios of the brightest sources in bands 0.5 to 0.9 kev , 0.9 to 1.5 kev , and 1.5 to 2.04 kev , though the sample is limited to only the brightest sources .
hardness ratios are not correlated in a simple way with photospheric temperature ( spectral type ) .
both hot and cool stars lie parallel to a locus derived for an isothermal plasma , however there is an offset indicating extra counts in the highest energy band
. this could be due to a second component at a higher temperature , at least for these very luminous sources .
acknowledgments : it is a pleasure to thank everyone who worked on the chandra project for making this data possible .
we thank nuria huelamo ( mpe ) for obtaining optical images of the field .
useful comments were provided by d. gies .
we thank an anonymous referee for for comments which improved the presentation of the results .
financial assistance was provided from the chandra x - ray center nasa contract nas8 - 39073 2 & 10 45 16.55 & -59 43 36.9 & 3.0 & 84.7 & 9.1 & 43.5 & 6.6 & 33.9 & 5.7 & 7.3 & 2.6 + 5 & 10 45 12.89 & -59 44 19.3 & 3.5 & 114.0 & 10.5 & 50.7 & 7.1 & 56.1 & 7.4 & 7.2 & 2.6 + 8 & 10 45 8.23 & -59 40 49.2 & 0.6 & 32.6 & 5.7 & 12.6 & 3.6 & 20.0 & 4.5 & 0.0 & 0.0 + 9 & 10 45 6.74 & -59 41 56.4 & 1.0 & 15.4 & 4.0 & 9.8 & 3.2 & 5.8 & 2.5 & -0.2 & 0.1 + 10 & 10 45 5.90 & -59 40 5.8 & 1.0 & 232.1 & 15.3 & 141.4 & 11.9 & 73.7 & 8.6 & 17.0 & 4.1 + 22 & 10 44 45.04 & -59 33 54.0 & 7.5 & 1572.8 & 37.2 & 597.6 & 23.0 & 771.7 & 26.0 & 203.5 & 13.4 + 25 & 10 44 33.77 & -59 44 15.6 & 4.9 & 532.6 & 22.5 & 311.4 & 17.2 & 189.9 & 13.4 & 31.4 & 5.5 + 28 & 10 44 10.38 & -59 43 11.3 & 7.0 & 4989.9 & 67.3 & 733.2 & 25.9 & 3018.1 & 52.3 & 1238.5 & 33.5 + 39 & 10 45 20.61 & -59 42 51.0 & 2.8 & 19.9 & 4.4 & 8.2 & 2.8 & 10.8 & 3.2 & 0.9 & 1.0 + 44 & 10 45 12.74 & -59 44 46.2 & 3.9 & 61.3 & 7.6 & 30.0 & 5.3 & 26.9 & 5.0 & 4.4 & 2.0 + 45 & 10 45 12.26 & -59 45 0.6 & 4.1 & 64.4 & 7.9 & 30.0 & 5.4 & 28.1 & 5.2 & 6.3 & 2.5 + 55 & 10 45 8.24 & -59 46 7.2 & 5.1 & 226.4 & 14.7 & 43.0 & 6.7 & 130.9 & 11.1 & 52.4 & 7.0 + 58 & 10 45 5.86 & -59 43 7.6 & 2.1 & 27.0 & 5.2 & 14.9 & 3.9 & 12.1 & 3.5 & 0.0 & 0.0 + 59 & 10 45 5.80 & -59 45 19.9 & 4.3 & 21.7 & 4.9 & 10.6 & 3.5 & 7.3 & 2.8 & 3.8 & 2.0 + 83 & 10 44 47.37 & -59 43 53.3 & 3.5 & 52.2 & 7.2 & 37.1 & 6.1 & 11.1 & 3.3 & 4.1 & 2.0 + 88 & 10 44 41.78 & -59 46 56.5 & 6.5 & 179.9 & 13.6 & 101.6 & 10.3 & 71.1 & 8.4 & 7.2 & 2.9 + 96 & 10 44 32.31 & -59 44 31.2 & 5.3 & 108.9 & 10.1 & 80.4 & 8.7 & 27.6 & 5.1 & 0.8 & 1.0 + 102 & 10 44 13.12 & -59 43 10.3 & 6.7 & 142.7 & 12.2 & 17.3 & 4.9 & 85.6 & 9.4 & 39.7 & 6.2 + 103 & 10 43 57.73 & -59 32 53.2 & 11.7 & 2132.4 & 41.6 & 467.8 & 19.0 & 1240.4 & 31.7 & 424.2 & 19.0 + 135 & 10 44 30.33 & -59 37 26.6 & 5.5 & 25.9 & 5.4 & 11.3 & 3.6 & 11.8 & 3.6 & 2.8 & 1.7 + 141 & 10 44 22.62 & -59 39 25.3 & 5.4 & 13.3 & 4.8 & 1.2 & 2.7 & 8.5 & 3.3 & 3.5 & 2.0 + 145 & 10 44 8.96 & -59 34 33.0 & 9.5 & 408.2 & 21.1 & 153.1 & 12.8 & 212.4 & 15.2 & 42.7 & 7.1 + 151 & 10 44 38.27 & -59 32 55.8 & 8.7 & 112.6 & 15.1 & 68.1 & 11.1 & 37.0 & 9.0 & 7.5 & 4.8 + 1 & 10 45 22.01 & -59 40 52.5 & 2.3 & 23.7 & 4.9 & 3.9 & 2.0 & 15.8 & 4.0 & 4.0 & 2.0 + 3 & 10 45 16.25 & -59 41 41.0 & 1.7 & 18.1 & 4.2 & 3.0 & 1.7 & 13.1 & 3.6 & 2.0 & 1.4 + 4 & 10 45 14.80 & -59 40 7.1 & 1.7 & 12.9 & 3.6 & 3.9 & 2.0 & 6.9 & 2.6 & 2.0 & 1.4 + 6 & 10 45 11.79 & -59 41 33.8 & 1.2 & 27.7 & 5.3 & 6.9 & 2.6 & 14.8 & 3.9 & 6.0 & 2.5 + 7 & 10 45 8.57 & -59 40 16.7 & 1.0 & 27.9 & 4.9 & 8.1 & 2.6 & 16.3 & 3.7 & 3.5 & 1.7 + 11 & 10 45 4.21 & -59 40 10.7 & 0.9 & 15.9 & 4.0 & 4.9 & 2.2 & 9.0 & 3.0 & 2.0 & 1.4 + 12 & 10 45 3.89 & -59 38 19.9 & 2.7 & 20.9 & 4.5 & 4.2 & 2.0 & 13.6 & 3.6 & 3.1 & 1.7 + 14 & 10 45 2.93 & -59 42 25.9 & 1.4 & 13.8 & 3.7 & 1.8 & 1.4 & 11.1 & 3.3 & 0.9 & 1.0 + 16 & 10 44 52.79 & -59 43 19.1 & 2.6 & 37.1 & 6.1 & 7.7 & 2.8 & 21.3 & 4.6 & 8.1 & 2.8 + 17 & 10 44 52.27 & -59 41 55.1 & 1.7 & 37.2 & 6.2 & 1.5 & 1.4 & 26.8 & 5.2 & 8.8 & 3.0 + 18 & 10 44 50.39 & -59 40 17.1 & 1.8 & 21.3 & 2.8 & 5.3 & 1.4 & 13.3 & 2.2 & 2.7 & 1.0 + 19 & 10 44 48.42 & -59 41 13.8 & 1.9 & 14.3 & 3.9 & 1.4 & 1.4 & 7.0 & 2.6 & 5.9 & 2.5 + 20 & 10 44 46.15 & -59 42 28.8 & 2.6 & 7.2 & 2.6 & -0.1 & 0.1 & 4.2 & 2.0 & 3.1 & 1.7 + 21 & 10 44 45.30 & -59 39 18.9 & 2.9 & 15.8 & 4.0 & 2.7 & 1.7 & 11.1 & 3.3 & 2.0 & 1.4 + 23 & 10 44 44.25 & -59 42 33.5 & 2.9 & 21.9 & 4.6 & 4.0 & 2.0 & 12.7 & 3.5 & 5.3 & 2.2 + 24 & 10 44 42.68 & -59 41 23.4 & 2.7 & 18.8 & 4.4 & 2.8 & 1.7 & 9.9 & 3.2 & 6.1 & 2.4 + 26 & 10 44 29.38 & -59 41 44.2 & 4.4 & 29.2 & 5.5 & 4.7 & 2.5 & 18.5 & 4.2 & 6.1 & 2.5 + 27 & 10 44 27.04 & -59 44 39.3 & 5.8 & 19.9 & 4.6 & 5.5 & 2.7 & 9.2 & 3.0 & 5.3 & 2.2 + 29 & 10 45 43.74 & -59 41 47.6 & 5.1 & 17.8 & 4.5 & 3.8 & 2.3 & 7.2 & 2.8 & 6.9 & 2.7 + 30 & 10 45 38.41 & -59 42 7.5 & 4.5 & 11.8 & 3.5 & 0.7 & 1.0 & 4.8 & 2.2 & 6.2 & 2.5 + 31 & 10 45 28.83 & -59 43 48.1 & 4.2 & 14.1 & 4.2 & 2.0 & 2.0 & 9.3 & 3.2 & 2.7 & 1.7 + 32 & 10 45 28.54 & -59 42 46.7 & 3.6 & 32.4 & 5.5 & 6.1 & 2.5 & 20.8 & 4.4 & 5.5 & 2.2 + 33 & 10 45 28.51 & -59 41 55.3 & 3.3 & 11.9 & 3.5 & 0.7 & 1.0 & 11.3 & 3.3 & -0.1 & 0.1 + 34 & 10 45 28.02 & -59 41 38.6 & 3.1 & 19.1 & 4.4 & 4.0 & 2.0 & 14.1 & 3.7 & 1.0 & 1.0 + 35 & 10 45 25.61 & -59 43 50.8 & 3.9 & 8.9 & 3.0 & 1.8 & 1.4 & 2.0 & 1.4 & 5.2 & 2.2 + 36 & 10 45 24.49 & -59 42 20.8 & 2.9 & 6.9 & 2.6 & 1.9 & 1.4 & 5.1 & 2.2 & 0.0 & 0.0 + 37 & 10 45 20.88 & -59 45 49.1 & 5.2 & 12.9 & 4.3 & -0.1 & 2.1 & 6.9 & 2.8 & 6.1 & 2.5 + 38 & 10 45 20.61 & -59 41 32.6 & 2.2 & 15.9 & 4.0 & 1.8 & 1.4 & 9.1 & 3.0 & 5.1 & 2.2 + 40 & 10 45 16.57 & -59 42 14.6 & 2.0 & 10.0 & 3.2 & 4.0 & 2.0 & 4.0 & 2.0 & 2.0 & 1.4 + 41 & 10 45 16.24 & -59 43 16.7 & 2.7 & 10.4 & 3.3 & 3.6 & 2.0 & 4.9 & 2.2 & 1.9 & 1.4 + 42 & 10 45 15.01 & -59 40 21.7 & 1.6 & 8.1 & 2.5 & 2.7 & 1.4 & 5.4 & 2.0 & 0.0 & 0.0 + 43 & 10 45 14.54 & -59 42 46.8 & 2.2 & 13.8 & 3.6 & 2.1 & 1.4 & 8.5 & 2.8 & 3.2 & 1.7 + 46 & 10 45 11.04 & -59 44 57.5 & 4.0 & 18.3 & 4.4 & 7.0 & 2.8 & 6.9 & 2.7 & 4.3 & 2.0 + 47 & 10 45 11.02 & -59 42 39.2 & 1.9 & 17.3 & 4.1 & 2.7 & 1.7 & 11.4 & 3.3 & 3.2 & 1.7 + 48 & 10 45 10.68 & -59 42 54.0 & 2.1 & 9.2 & 3.2 & 2.6 & 1.7 & 4.7 & 2.2 & 2.0 & 1.4 + 49 & 10 45 10.42 & -59 44 9.5 & 3.2 & 8.7 & 3.0 & 1.7 & 1.4 & 6.1 & 2.5 & 0.8 & 1.0 + 50 & 10 45 10.33 & -59 40 28.6 & 1.0 & 18.1 & 2.6 & 7.8 & 1.7 & 5.2 & 1.4 & 5.2 & 1.4 + 51 & 10 45 9.89 & -59 40 18.5 & 1.1 & 6.5 & 2.2 & 1.3 & 1.0 & 5.2 & 2.0 & 0.0 & 0.0 + 52 & 10 45 9.76 & -59 42 45.5 & 1.9 & 8.4 & 2.8 & 2.1 & 1.4 & 4.2 & 2.0 & 2.1 & 1.4 + 54 & 10 45 8.35 & -59 38 47.3 & 2.3 & 6.7 & 2.6 & 2.0 & 1.4 & 4.8 & 2.2 & -0.1 & 0.1 + 57 & 10 45 6.53 & -59 40 15.5 & 0.9 & 8.8 & 2.8 & 2.1 & 1.4 & 6.7 & 2.5 & 0.0 & 0.0 + 60 & 10 45 5.72 & -59 43 43.8 & 2.7 & 22.6 & 4.8 & 7.7 & 2.8 & 14.0 & 3.7 & 0.9 & 1.0 + 61 & 10 45 5.01 & -59 37 33.1 & 3.5 & 18.9 & 4.4 & 6.8 & 2.6 & 10.0 & 3.2 & 2.0 & 1.4 + 63 & 10 45 3.62 & -59 43 20.5 & 2.3 & 9.4 & 3.2 & 7.6 & 2.8 & 1.8 & 1.4 & 0.0 & 0.0 + 64 & 10 45 2.18 & -59 41 34.4 & 0.6 & 6.6 & 2.7 & 0.7 & 1.0 & 5.0 & 2.2 & 1.0 & 1.0 + 68 & 10 45 1.10 & -59 39 17.4 & 1.8 & 8.2 & 2.8 & 4.0 & 2.0 & 4.2 & 2.0 & 0.0 & 0.0 + 69 & 10 45 1.08 & -59 38 53.4 & 2.2 & 6.2 & 2.2 & 1.3 & 1.0 & 5.0 & 2.0 & 0.0 & 0.0 + 70 & 10 45 0.99 & -59 45 15.6 & 4.2 & 50.4 & 7.0 & 27.3 & 5.2 & 19.9 & 4.4 & 3.2 & 1.7 + 71 & 10 44 57.96 & -59 47 9.5 & 6.1 & 51.3 & 7.4 & 2.6 & 2.9 & 25.7 & 5.0 & 22.9 & 4.6 + 72 & 10 44 58.08 & -59 42 31.3 & 1.6 & 13.3 & 2.5 & -0.3 & 0.1 & 6.7 & 1.7 & 6.8 & 1.7 + 73 & 10 44 57.07 & -59 38 26.7 & 2.7 & 15.8 & 4.0 & 10.8 & 3.3 & 4.0 & 2.0 & 1.0 & 1.0 + 74 & 10 44 54.83 & -59 43 51.6 & 3.0 & 8.2 & 3.0 & 3.7 & 2.0 & 3.6 & 2.0 & 0.9 & 1.0 + 75 & 10 44 54.53 & -59 42 43.1 & 2.0 & 9.0 & 3.0 & 1.0 & 1.0 & 7.0 & 2.6 & 1.0 & 1.0 + 76 & 10 44 53.75 & -59 43 24.2 & 2.7 & 6.7 & 2.7 & 1.7 & 1.4 & 5.0 & 2.2 & -0.1 & 0.1 + 77 & 10 44 53.25 & -59 41 35.9 & 1.4 & 6.9 & 2.6 & 0.9 & 1.0 & 4.9 & 2.2 & 1.0 & 1.0 + 78 & 10 44 51.40 & -59 44 6.9 & 3.4 & 7.9 & 2.8 & 0.7 & 1.0 & 7.2 & 2.6 & 0.0 & 0.0 + 79 & 10 44 51.14 & -59 39 38.0 & 2.1 & 8.0 & 2.8 & 2.0 & 1.4 & 4.0 & 2.0 & 2.0 & 1.4 + 80 & 10 44 50.14 & -59 44 43.1 & 4.0 & 14.4 & 3.8 & 8.3 & 2.8 & 5.0 & 2.2 & 1.0 & 1.0 + 81 & 10 44 50.13 & -59 42 38.8 & 2.3 & 7.5 & 2.6 & 2.1 & 1.4 & 4.4 & 2.0 & 1.1 & 1.0 + 82 & 10 44 50.05 & -59 34 29.0 & 6.8 & 69.7 & 8.5 & 32.0 & 5.8 & 27.5 & 5.3 & 10.2 & 3.2 + 84 & 10 44 47.30 & -59 41 16.2 & 2.1 & 11.6 & 3.5 & 1.7 & 1.4 & 7.9 & 2.8 & 2.0 & 1.4 + 85 & 10 44 46.54 & -59 34 12.4 & 7.2 & 90.8 & 10.1 & 19.3 & 5.3 & 49.3 & 7.2 & 22.2 & 4.7 + 86 & 10 44 43.98 & -59 39 55.5 & 2.7 & 13.8 & 3.7 & 2.8 & 1.7 & 6.0 & 2.5 & 5.0 & 2.2 + 87 & 10 44 41.96 & -59 42 15.9 & 3.0 & 11.7 & 3.5 & 1.9 & 1.4 & 6.7 & 2.7 & 3.1 & 1.7 + 89 & 10 44 41.73 & -59 42 54.5 & 3.3 & 9.6 & 3.2 & 0.4 & 1.0 & 8.1 & 2.8 & 1.0 & 1.0 + 90 & 10 44 41.34 & -59 44 8.8 & 4.2 & 13.9 & 3.9 & 6.4 & 2.7 & 6.7 & 2.7 & 0.8 & 1.0 + 91 & 10 44 41.15 & -59 38 18.9 & 3.9 & 21.2 & 4.7 & 4.3 & 2.2 & 11.8 & 3.5 & 5.2 & 2.2 + 92 & 10 44 39.88 & -59 43 58.5 & 4.2 & 7.7 & 2.8 & 1.6 & 1.4 & 5.0 & 2.2 & 1.0 & 1.0 + 93 & 10 44 38.08 & -59 41 31.4 & 3.3 & 6.6 & 2.7 & 3.6 & 2.0 & 3.0 & 1.7 & 0.0 & 0.0 + 94 & 10 44 38.06 & -59 43 26.2 & 4.0 & 21.9 & 5.0 & 4.1 & 2.5 & 16.1 & 4.1 & 1.7 & 1.4 + 95 & 10 44 32.84 & -59 43 6.8 & 4.4 & 10.9 & 3.5 & 1.4 & 1.4 & 5.4 & 2.5 & 4.1 & 2.0 + 97 & 10 44 28.55 & -59 41 45.3 & 4.5 & 7.8 & 2.8 & 4.9 & 2.2 & 3.0 & 1.7 & -0.1 & 0.1 + 98 & 10 44 28.49 & -59 39 19.7 & 4.8 & 45.7 & 6.8 & 5.4 & 2.7 & 26.7 & 5.1 & 13.5 & 3.6 + 99 & 10 44 27.80 & -59 45 22.0 & 6.2 & 40.0 & 6.3 & 21.5 & 4.7 & 17.6 & 4.1 & 0.9 & 1.0 + 100 & 10 44 26.28 & -59 42 41.0 & 5.0 & 18.2 & 4.4 & 2.8 & 2.0 & 9.2 & 3.0 & 6.2 & 2.5 + 101 & 10 44 25.34 & -59 41 11.5 & 4.8 & 9.6 & 3.2 & 6.6 & 2.7 & 1.9 & 1.4 & 1.0 & 1.0 + 104 & 10 45 47.17 & -59 47 50.9 & 8.7 & 19.8 & 14.1 & 16.3 & 10.4 & 2.4 & 8.5 & 1.1 & 4.0 + 105 & 10 45 41.58 & -59 38 20.4 & 5.5 & 20.0 & 4.5 & 7.0 & 2.7 & 10.0 & 3.2 & 3.1 & 1.7 + 106 & 10 45 37.19 & -59 48 6.6 & 8.2 & 46.0 & 14.9 & 7.1 & 10.4 & 23.6 & 8.5 & 15.3 & 6.4 + 107 & 10 45 28.01 & -59 43 19.8 & 3.8 & 7.5 & 3.0 & 0.9 & 1.4 & 4.7 & 2.2 & 2.0 & 1.4 + 108 & 10 45 26.40 & -59 40 29.1 & 2.9 & 10.6 & 2.6 & 2.9 & 1.4 & 7.7 & 2.2 & 0.0 & 0.0 + 109 & 10 45 16.84 & -59 46 29.3 & 5.7 & 13.9 & 4.3 & 7.8 & 3.4 & 4.1 & 2.3 & 2.0 & 1.4 + 110 & 10 45 13.47 & -59 44 4.4 & 3.3 & 7.4 & 2.8 & -0.6 & 0.2 & 6.0 & 2.5 & 1.9 & 1.4 + 111 & 10 45 13.18 & -59 35 21.8 & 5.8 & 12.8 & 3.8 & 3.9 & 2.3 & 3.7 & 2.0 & 5.2 & 2.2 + 112 & 10 45 11.09 & -59 32 58.4 & 8.1 & 46.7 & 7.8 & 20.0 & 5.3 & 24.6 & 5.5 & 2.1 & 1.8 + 116 & 10 45 3.27 & -59 42 18.1 & 1.3 & 6.7 & 2.7 & 2.9 & 1.7 & 2.8 & 1.7 & 0.9 & 1.0 + 119 & 10 44 59.28 & -59 44 59.8 & 4.0 & 13.1 & 3.8 & 6.2 & 2.7 & 4.9 & 2.2 & 2.0 & 1.4 + 120 & 10 44 57.92 & -59 39 59.6 & 1.3 & 7.1 & 2.0 & 1.6 & 1.0 & 3.6 & 1.4 & 1.9 & 1.0 + 121 & 10 44 55.97 & -59 44 48.7 & 3.9 & 12.3 & 2.5 & -0.9 & 0.2 & 8.8 & 2.0 & 4.4 & 1.4 + 122 & 10 44 52.95 & -59 46 34.8 & 5.7 & 19.4 & 4.5 & 2.0 & 2.0 & 9.7 & 3.0 & 7.8 & 2.6 + 123 & 10 44 52.18 & -59 45 14.9 & 4.4 & 13.8 & 4.4 & 6.0 & 3.0 & 6.3 & 2.9 & 1.6 & 1.4 + 124 & 10 44 51.75 & -59 43 5.6 & 2.5 & 8.1 & 2.8 & 4.0 & 2.0 & 4.1 & 2.0 & 0.0 & 0.0 + 125 & 10 44 47.39 & -59 35 7.0 & 6.3 & 12.8 & 3.8 & 9.3 & 3.2 & 3.7 & 2.0 & -0.1 & 0.1 + 126 & 10 44 46.38 & -59 36 3.2 & 5.5 & 9.7 & 3.2 & 2.5 & 1.7 & 6.3 & 2.5 & 0.9 & 1.0 + 127 & 10 44 46.18 & -59 44 49.8 & 4.4 & 10.2 & 3.3 & 4.0 & 2.3 & 4.1 & 2.0 & 2.1 & 1.4 + 128 & 10 44 41.83 & -59 40 47.4 & 2.8 & 8.2 & 3.0 & 0.2 & 1.0 & 8.0 & 2.8 & 0.0 & 0.0 + 129 & 10 44 41.16 & -59 44 22.8 & 4.4 & 10.8 & 3.5 & 2.0 & 1.8 & 5.6 & 2.5 & 3.2 & 1.7 + 130 & 10 44 36.68 & -59 44 23.9 & 4.8 & 11.5 & 3.5 & 5.7 & 2.5 & 4.0 & 2.0 & 1.8 & 1.4 + 131 & 10 44 36.40 & -59 44 10.7 & 4.6 & 17.0 & 4.7 & 9.6 & 3.6 & 5.8 & 2.7 & 1.7 & 1.4 + 132 & 10 44 35.70 & -59 43 51.1 & 4.5 & 8.3 & 3.2 & 1.7 & 1.8 & 5.9 & 2.5 & 0.8 & 1.0 + 133 & 10 44 34.95 & -59 44 5.6 & 4.7 & 15.2 & 4.9 & 5.7 & 3.4 & 8.4 & 3.2 & 1.1 & 1.4 + 134 & 10 44 33.93 & -59 37 33.5 & 5.1 & 12.2 & 3.8 & 0.6 & 1.4 & 5.4 & 2.5 & 6.2 & 2.5 + 136 & 10 44 29.73 & -59 43 33.1 & 4.9 & 19.3 & 4.7 & 3.6 & 2.5 & 11.8 & 3.5 & 4.0 & 2.0 + 137 & 10 44 28.58 & -59 40 51.5 & 4.4 & 14.9 & 4.1 & 4.6 & 2.5 & 7.4 & 2.8 & 3.0 & 1.7 + 138 & 10 44 24.76 & -59 39 58.5 & 5.0 & 24.5 & 3.8 & 4.3 & 1.7 & 16.5 & 3.0 & 3.7 & 1.4 + 139 & 10 44 23.72 & -59 41 15.0 & 5.0 & 25.9 & 5.5 & 3.8 & 2.7 & 15.3 & 4.0 & 6.8 & 2.7 + 140 & 10 44 22.86 & -59 44 32.8 & 6.2 & 16.4 & 4.4 & 2.6 & 2.3 & 12.0 & 3.5 & 1.8 & 1.4 + 142 & 10 44 19.94 & -59 42 30.2 & 5.7 & 12.1 & 4.0 & 3.8 & 2.5 & 5.4 & 2.7 & 2.9 & 1.7 + 143 & 10 44 17.13 & -59 41 56.5 & 5.9 & 11.5 & 3.8 & 1.4 & 1.8 & 6.2 & 2.7 & 3.9 & 2.0 + 144 & 10 44 9.15 & -59 37 17.9 & 7.8 & 37.7 & 7.7 & 1.7 & 3.9 & 23.5 & 5.6 & 12.6 & 3.6 + 146 & 10 44 2.66 & -59 39 44.9 & 7.8 & 67.2 & 7.6 & 10.6 & 4.1 & 30.7 & 4.9 & 25.9 & 4.0 + 147 & 10 45 37.02 & -59 47 3.1 & 7.3 & 22.0 & 6.7 & 2.7 & 4.3 & 11.6 & 4.2 & 7.7 & 3.0 + 148 & 10 45 36.57 & -59 44 11.7 & 5.2 & 10.2 & 3.6 & 1.8 & 2.0 & 4.3 & 2.3 & 4.1 & 2.0 + 149 & 10 45 23.52 & -59 45 45.5 & 5.3 & 13.4 & 5.1 & 3.2 & 3.4 & 3.6 & 2.7 & 6.6 & 2.7 + 150 & 10 45 7.31 & -59 44 47.1 & 3.8 & 9.2 & 3.3 & 4.0 & 2.3 & 3.2 & 2.0 & 2.0 & 1.4 + 152 & 10 44 38.23 & -59 44 35.8 & 4.8 & 11.8 & 3.6 & 9.8 & 3.2 & 2.5 & 1.7 & -0.5 & 0.2 + 153 & 10 44 36.47 & -59 33 40.2 & 8.1 & 22.7 & 6.8 & 6.2 & 4.0 & 15.1 & 4.9 & 1.4 & 2.4 + 154 & 10 44 33.34 & -59 32 16.1 & 9.6 & 33.7 & 7.6 & -2.0 & 3.3 & 25.8 & 5.8 & 9.9 & 3.7 + 155 & 10 44 29.37 & -59 32 18.6 & 9.8 & 38.0 & 8.9 & 4.5 & 4.7 & 24.6 & 6.3 & 9.0 & 4.1 + 156 & 10 44 29.43 & -59 48 20.9 & 8.5 & 76.1 & 13.4 & 42.5 & 10.9 & 30.0 & 7.0 & 3.6 & 3.3 + 157 & 10 44 24.76 & -59 45 55.6 & 6.9 & 16.0 & 5.6 & 11.7 & 4.7 & 3.1 & 2.7 & 1.3 & 1.4 + 158 & 10 44 12.47 & -59 34 26.5 & 9.3 & 106.1 & 13.7 & 14.8 & 7.0 & 56.8 & 9.8 & 34.6 & 6.5 + 159 & 10 44 12.81 & -59 33 41.8 & 9.8 & 61.2 & 13.1 & 10.9 & 6.7 & 25.2 & 9.2 & 25.2 & 6.5 + 160 & 10 44 6.61 & -59 36 13.4 & 8.7 & 52.3 & 10.9 & 9.3 & 6.5 & 27.7 & 7.5 & 15.3 & 4.5 + 161 & 10 44 4.89 & -59 32 36.8 & 11.2 & 69.4 & 15.5 & 10.4 & 7.6 & 34.3 & 11.0 & 24.7 & 7.9 + 162 & 10 44 2.94 & -59 31 40.3 & 12.1 & 74.0 & 12.5 & 5.8 & 6.3 & 46.8 & 9.2 & 21.4 & 5.7 + 163 & 10 44 0.70 & -59 34 41.6 & 10.2 & 35.8 & 13.6 & 4.7 & 7.3 & 18.6 & 9.8 & 12.5 & 5.9 + 164 & 10 45 19.68 & -59 32 19.0 & 9.0 & 49.8 & 9.8 & 10.6 & 7.0 & 23.6 & 5.6 & 15.6 & 4.1 + 165 & 10 44 49.38 & -59 47 30.2 & 6.7 & 17.7 & 8.9 & -1.7 & 6.4 & 16.7 & 5.7 & 2.7 & 2.3 + 166 & 10 44 26.25 & -59 37 39.7 & 5.8 & 11.3 & 4.2 & 6.0 & 3.0 & 3.7 & 2.5 & 1.6 & 1.4 + 167 & 10 44 16.45 & -59 33 9.0 & 9.9 & 49.1 & 10.3 & 11.9 & 6.1 & 25.1 & 7.0 & 12.1 & 4.4 + 2 & -0.12 & 0.14 & -0.65 & 0.20 & -12.91 & -13.88 & 31.96 & 30.99 5 & 0.05 & 0.12 & -0.77 & 0.17 & -12.78 & -13.82 & 32.09 & 31.06 8 & 0.23 & 0.22 & -1.00 & - & -13.33 & -14.08 & 31.55 & 30.79 9 & -0.26 & 0.32 & -1.07 & 0.62 & -13.65 & -14.24 & 31.22 & 30.64 10 & -0.31 & 0.08 & -0.63 & 0.14 & -12.47 & -13.66 & 32.40 & 31.22 22 & 0.13 & 0.03 & -0.58 & 0.04 & -11.64 & -13.27 & 33.23 & 31.61 25 & -0.24 & 0.05 & -0.72 & 0.09 & -12.11 & -13.49 & 32.76 & 31.39 28 & 0.61 & 0.02 & -0.42 & 0.02 & -11.14 & -13.01 & 33.73 & 31.86 39 & 0.14 & 0.28 & -0.84 & 0.39 & -13.54 & -14.20 & 31.33 & 30.67 44 & -0.06 & 0.16 & -0.72 & 0.23 & -13.05 & -13.96 & 31.82 & 30.91 45 & -0.03 & 0.16 & -0.64 & 0.22 & -13.03 & -13.94 & 31.84 & 30.93 55 & 0.51 & 0.10 & -0.43 & 0.09 & -12.49 & -13.67 & 32.39 & 31.20 58 & -0.11 & 0.24 & -1.00 & - & -13.41 & -14.12 & 31.47 & 30.75 59 & -0.18 & 0.30 & -0.31 & 0.38 & -13.50 & -14.15 & 31.37 & 30.73 83 & -0.54 & 0.18 & -0.46 & 0.32 & -13.12 & -13.98 & 31.75 & 30.89 88 & -0.18 & 0.09 & -0.82 & 0.15 & -12.59 & -13.71 & 32.29 & 31.17 96 & -0.49 & 0.12 & -0.94 & 0.26 & -12.80 & -13.84 & 32.07 & 31.04 102 & 0.66 & 0.13 & -0.37 & 0.11 & -12.69 & -13.75 & 32.19 & 31.12 103 & 0.45 & 0.03 & -0.49 & 0.03 & -11.51 & -13.22 & 33.36 & 31.65 135 & 0.02 & 0.27 & -0.61 & 0.35 & -13.43 & -14.11 & 31.45 & 30.77 141 & 0.75 & 0.53 & -0.41 & 0.39 & -13.72 & -14.16 & 31.16 & 30.71 145 & 0.16 & 0.07 & -0.66 & 0.08 & -12.23 & -13.52 & 32.65 & 31.36 + 151 & -0.30 & 0.15 & -0.66 & 0.26 & -12.79 & -13.66 & 32.09 & 31.21 1 & 0.60 & 0.29 & -0.60 & 0.29 & -13.47 & -14.15 & 31.41 & 30.73 + 3 & 0.62 & 0.32 & -0.73 & 0.34 & -13.58 & -14.21 & 31.29 & 30.66 + 4 & 0.28 & 0.38 & -0.55 & 0.43 & -13.73 & -14.28 & 31.14 & 30.59 + 6 & 0.36 & 0.27 & -0.43 & 0.28 & -13.40 & -14.12 & 31.48 & 30.76 + 7 & 0.34 & 0.24 & -0.64 & 0.28 & -13.39 & -14.15 & 31.48 & 30.72 + 11 & 0.29 & 0.33 & -0.64 & 0.39 & -13.64 & -14.24 & 31.24 & 30.64 + 12 & 0.53 & 0.30 & -0.63 & 0.31 & -13.52 & -14.19 & 31.36 & 30.69 + 14 & 0.73 & 0.37 & -0.84 & 0.39 & -13.70 & -14.27 & 31.17 & 30.61 + 16 & 0.47 & 0.24 & -0.45 & 0.23 & -13.27 & -14.06 & 31.60 & 30.82 + 17 & 0.89 & 0.26 & -0.50 & 0.21 & -13.27 & -14.05 & 31.61 & 30.82 + 18 & 0.43 & 0.23 & -0.67 & 0.26 & -13.51 & -14.39 & 31.36 & 30.49 + 19 & 0.66 & 0.46 & -0.08 & 0.34 & -13.68 & -14.25 & 31.19 & 30.62 + 20 & 1.07 & 0.71 & -0.15 & 0.45 & -13.98 & -14.42 & 30.89 & 30.46 + 21 & 0.61 & 0.35 & -0.70 & 0.36 & -13.64 & -14.24 & 31.23 & 30.64 + 23 & 0.52 & 0.31 & -0.41 & 0.29 & -13.50 & -14.18 & 31.37 & 30.70 + 24 & 0.56 & 0.36 & -0.24 & 0.31 & -13.57 & -14.20 & 31.31 & 30.67 + 26 & 0.59 & 0.27 & -0.51 & 0.26 & -13.37 & -14.10 & 31.50 & 30.77 + 27 & 0.25 & 0.33 & -0.27 & 0.32 & -13.54 & -14.18 & 31.33 & 30.70 + 29 & 0.32 & 0.40 & -0.03 & 0.33 & -13.59 & -14.19 & 31.29 & 30.69 + 30 & 0.75 & 0.58 & 0.13 & 0.37 & -13.77 & -14.30 & 31.11 & 30.58 + 31 & 0.64 & 0.42 & -0.54 & 0.38 & -13.69 & -14.22 & 31.18 & 30.65 + 32 & 0.54 & 0.24 & -0.58 & 0.25 & -13.33 & -14.10 & 31.54 & 30.77 + 33 & 0.89 & 0.40 & -1.01 & 0.42 & -13.76 & -14.30 & 31.11 & 30.57 + 34 & 0.56 & 0.30 & -0.86 & 0.35 & -13.56 & -14.20 & 31.31 & 30.67 + 35 & 0.06 & 0.65 & 0.45 & 0.47 & -13.89 & -14.36 & 30.98 & 30.51 + 36 & 0.47 & 0.48 & & - & -14.00 & -14.42 & 30.88 & 30.46 + 37 & 1.04 & 0.65 & -0.07 & 0.35 & -13.73 & -14.21 & 31.15 & 30.67 + 38 & 0.68 & 0.40 & -0.28 & 0.33 & -13.64 & -14.24 & 31.24 & 30.64 + 40 & 0.00 & 0.43 & -0.34 & 0.51 & -13.84 & -14.34 & 31.03 & 30.53 + 41 & 0.16 & 0.43 & -0.44 & 0.49 & -13.82 & -14.32 & 31.05 & 30.56 + 42 & 0.34 & 0.40 & - & - & -13.93 & -14.45 & 30.94 & 30.42 + 43 & 0.61 & 0.39 & -0.45 & 0.36 & -13.70 & -14.28 & 31.17 & 30.59 + 46 & -0.00 & 0.34 & -0.23 & 0.37 & -13.58 & -14.20 & 31.30 & 30.68 + 47 & 0.62 & 0.35 & -0.56 & 0.33 & -13.60 & -14.22 & 31.27 & 30.65 + 48 & 0.29 & 0.48 & -0.41 & 0.50 & -13.88 & -14.34 & 31.00 & 30.54 + 49 & 0.56 & 0.46 & -0.76 & 0.51 & -13.90 & -14.36 & 30.97 & 30.51 + 50 & -0.20 & 0.26 & 0.00 & 0.29 & -13.58 & -14.42 & 31.29 & 30.46 + 51 & 0.59 & 0.47 & - & - & -14.03 & -14.49 & 30.85 & 30.38 + 52 & 0.34 & 0.49 & -0.33 & 0.49 & -13.92 & -14.39 & 30.96 & 30.49 + 54 & 0.41 & 0.49 & -1.06 & 0.68 & -14.02 & -14.42 & 30.86 & 30.46 + 57 & 0.53 & 0.42 & - & - & -13.89 & -14.39 & 30.98 & 30.49 + 60 & 0.29 & 0.27 & -0.88 & 0.36 & -13.49 & -14.16 & 31.39 & 30.72 + 61 & 0.19 & 0.30 & -0.66 & 0.38 & -13.56 & -14.20 & 31.31 & 30.67 + 63 & -0.62 & 0.43 & - & - & -13.87 & -14.34 & 31.01 & 30.54 + 64 & 0.76 & 0.57 & -0.66 & 0.54 & -14.02 & -14.42 & 30.86 & 30.46 + 68 & 0.03 & 0.43 & - & - & -13.93 & -14.39 & 30.95 & 30.49 + 69 & 0.59 & 0.49 & - & - & -14.05 & -14.49 & 30.83 & 30.38 + 70 & -0.16 & 0.18 & -0.72 & 0.27 & -13.14 & -13.99 & 31.74 & 30.88 + 71 & 0.82 & 0.27 & -0.06 & 0.17 & -13.13 & -13.97 & 31.74 & 30.90 + 72 & 1.09 & 0.49 & 0.01 & 0.26 & -13.72 & -14.45 & 31.16 & 30.42 + 73 & -0.46 & 0.33 & -0.59 & 0.58 & -13.64 & -14.24 & 31.23 & 30.64 + 74 & -0.01 & 0.47 & -0.60 & 0.63 & -13.93 & -14.36 & 30.95 & 30.51 + 75 & 0.76 & 0.48 & -0.75 & 0.47 & -13.89 & -14.36 & 30.99 & 30.51 + 76 & 0.49 & 0.50 & -1.03 & 0.64 & -14.02 & -14.42 & 30.86 & 30.46 + 77 & 0.68 & 0.55 & -0.66 & 0.54 & -14.00 & -14.42 & 30.87 & 30.46 + 78 & 0.82 & 0.49 & - & - & -13.94 & -14.39 & 30.93 & 30.49 + 79 & 0.35 & 0.51 & -0.33 & 0.51 & -13.94 & -14.39 & 30.94 & 30.49 + 80 & -0.24 & 0.34 & -0.66 & 0.53 & -13.68 & -14.27 & 31.19 & 30.61 + 81 & 0.36 & 0.48 & -0.59 & 0.54 & -13.96 & -14.42 & 30.91 & 30.46 + 82 & -0.08 & 0.16 & -0.46 & 0.21 & -13.00 & -13.91 & 31.88 & 30.96 + 84 & 0.64 & 0.43 & -0.59 & 0.41 & -13.77 & -14.30 & 31.10 & 30.57 + 85 & 0.44 & 0.16 & -0.38 & 0.15 & -12.88 & -13.84 & 31.99 & 31.04 + 86 & 0.36 & 0.43 & -0.09 & 0.37 & -13.70 & -14.27 & 31.17 & 30.61 + 87 & 0.56 & 0.44 & -0.37 & 0.40 & -13.77 & -14.30 & 31.10 & 30.57 + 89 & 0.90 & 0.48 & -0.77 & 0.44 & -13.86 & -14.34 & 31.01 & 30.54 + 90 & 0.03 & 0.35 & -0.78 & 0.50 & -13.70 & -14.25 & 31.18 & 30.62 + 91 & 0.47 & 0.32 & -0.39 & 0.31 & -13.51 & -14.17 & 31.36 & 30.71 + 92 & 0.52 & 0.50 & -0.66 & 0.53 & -13.95 & -14.39 & 30.92 & 30.49 + 93 & -0.09 & 0.49 & - & - & -14.02 & -14.42 & 30.86 & 30.46 + 94 & 0.60 & 0.30 & -0.81 & 0.33 & -13.50 & -14.14 & 31.38 & 30.74 + 95 & 0.59 & 0.52 & -0.13 & 0.41 & -13.80 & -14.30 & 31.07 & 30.58 + 97 & -0.25 & 0.44 & -1.05 & 0.85 & -13.95 & -14.39 & 30.93 & 30.49 + 98 & 0.66 & 0.23 & -0.33 & 0.19 & -13.18 & -14.01 & 31.69 & 30.87 + 99 & -0.10 & 0.20 & -0.91 & 0.32 & -13.24 & -14.04 & 31.64 & 30.84 + 100 & 0.53 & 0.38 & -0.19 & 0.31 & -13.58 & -14.20 & 31.30 & 30.68 + 101 & -0.55 & 0.45 & -0.33 & 0.74 & -13.86 & -14.34 & 31.01 & 30.54 + 104 & -0.74 & 0.75 & -0.39 & 2.75 & -13.54 & -13.69 & 31.33 & 31.18 + 105 & 0.18 & 0.30 & -0.53 & 0.36 & -13.54 & -14.19 & 31.34 & 30.69 + 106 & 0.54 & 0.46 & -0.21 & 0.30 & -13.18 & -13.67 & 31.70 & 31.21 + 107 & 0.67 & 0.60 & -0.41 & 0.50 & -13.96 & -14.36 & 30.91 & 30.51 + 108 & 0.45 & 0.34 & - & - & -13.81 & -14.42 & 31.06 & 30.46 + 109 & -0.32 & 0.40 & -0.33 & 0.53 & -13.70 & -14.21 & 31.18 & 30.67 + 110 & 1.20 & 0.65 & -0.51 & 0.45 & -13.97 & -14.39 & 30.91 & 30.49 + 111 & -0.02 & 0.47 & 0.17 & 0.41 & -13.73 & -14.27 & 31.14 & 30.61 + 112 & 0.10 & 0.20 & -0.84 & 0.28 & -13.17 & -13.95 & 31.70 & 30.93 + 116 & -0.01 & 0.52 & -0.50 & 0.67 & -14.02 & -14.42 & 30.86 & 30.46 + 119 & -0.12 & 0.38 & -0.41 & 0.48 & -13.72 & -14.27 & 31.15 & 30.61 + 120 & 0.38 & 0.47 & -0.32 & 0.45 & -13.99 & -14.54 & 30.89 & 30.34 + 121 & 1.23 & 0.47 & -0.34 & 0.28 & -13.75 & -14.45 & 31.12 & 30.43 + 122 & 0.66 & 0.40 & -0.11 & 0.29 & -13.55 & -14.19 & 31.32 & 30.69 + 123 & 0.02 & 0.40 & -0.60 & 0.50 & -13.70 & -14.20 & 31.18 & 30.68 + 124 & 0.01 & 0.43 & - & - & -13.93 & -14.39 & 30.94 & 30.49 + 125 & -0.43 & 0.36 & -1.08 & 0.79 & -13.73 & -14.27 & 31.14 & 30.61 + 126 & 0.43 & 0.43 & -0.75 & 0.49 & -13.85 & -14.34 & 31.02 & 30.54 + 127 & 0.01 & 0.45 & -0.33 & 0.50 & -13.83 & -14.32 & 31.04 & 30.56 + 128 & 0.95 & 0.50 & - & - & -13.93 & -14.36 & 30.95 & 30.51 + 129 & 0.48 & 0.49 & -0.28 & 0.42 & -13.81 & -14.30 & 31.07 & 30.58 + 130 & -0.17 & 0.40 & -0.39 & 0.53 & -13.78 & -14.30 & 31.10 & 30.58 + 131 & -0.25 & 0.35 & -0.55 & 0.50 & -13.61 & -14.16 & 31.27 & 30.71 + 132 & 0.55 & 0.49 & -0.77 & 0.53 & -13.92 & -14.34 & 30.96 & 30.54 + 133 & 0.20 & 0.38 & -0.76 & 0.47 & -13.66 & -14.15 & 31.22 & 30.72 + 134 & 0.81 & 0.61 & 0.07 & 0.36 & -13.76 & -14.26 & 31.12 & 30.61 + 136 & 0.53 & 0.34 & -0.50 & 0.32 & -13.55 & -14.17 & 31.32 & 30.71 + 137 & 0.23 & 0.38 & -0.42 & 0.40 & -13.67 & -14.22 & 31.21 & 30.65 + 138 & 0.58 & 0.24 & -0.64 & 0.25 & -13.45 & -14.27 & 31.42 & 30.61 + 139 & 0.60 & 0.32 & -0.38 & 0.27 & -13.43 & -14.10 & 31.45 & 30.78 + 140 & 0.64 & 0.36 & -0.74 & 0.36 & -13.63 & -14.20 & 31.25 & 30.68 + 142 & 0.17 & 0.46 & -0.31 & 0.46 & -13.76 & -14.23 & 31.12 & 30.64 + 143 & 0.63 & 0.52 & -0.22 & 0.40 & -13.78 & -14.26 & 31.10 & 30.61 + 144 & 0.87 & 0.33 & -0.30 & 0.22 & -13.26 & -13.95 & 31.61 & 30.92 + 146 & 0.49 & 0.20 & -0.09 & 0.15 & -13.01 & -13.96 & 31.86 & 30.91 + 147 & 0.63 & 0.47 & -0.20 & 0.31 & -13.50 & -14.01 & 31.38 & 30.86 + 148 & 0.41 & 0.59 & -0.02 & 0.43 & -13.83 & -14.28 & 31.04 & 30.60 + 149 & 0.05 & 0.70 & 0.30 & 0.44 & -13.71 & -14.13 & 31.16 & 30.74 + 150 & -0.11 & 0.50 & -0.24 & 0.57 & -13.88 & -14.32 & 31.00 & 30.56 + 152 & -0.59 & 0.38 & -1.44 & 1.21 & -13.77 & -14.28 & 31.11 & 30.59 + 153 & 0.42 & 0.34 & -0.83 & 0.40 & -13.48 & -14.01 & 31.39 & 30.87 + 154 & 1.17 & 0.36 & -0.44 & 0.23 & -13.31 & -13.96 & 31.56 & 30.92 + 155 & 0.69 & 0.31 & -0.46 & 0.26 & -13.26 & -13.89 & 31.61 & 30.98 + 156 & -0.17 & 0.20 & -0.79 & 0.28 & -12.96 & -13.71 & 31.92 & 31.16 + 157 & -0.59 & 0.43 & -0.41 & 0.80 & -13.64 & -14.09 & 31.24 & 30.78 + 158 & 0.59 & 0.19 & -0.24 & 0.15 & -12.81 & -13.71 & 32.06 & 31.17 + 159 & 0.40 & 0.34 & 0.00 & 0.25 & -13.05 & -13.72 & 31.82 & 31.15 + 160 & 0.50 & 0.30 & -0.29 & 0.23 & -13.12 & -13.80 & 31.75 & 31.07 + 161 & 0.53 & 0.32 & -0.16 & 0.25 & -13.00 & -13.65 & 31.88 & 31.23 + 162 & 0.78 & 0.24 & -0.37 & 0.18 & -12.97 & -13.74 & 31.90 & 31.13 + 163 & 0.60 & 0.55 & -0.20 & 0.39 & -13.29 & -13.71 & 31.59 & 31.17 + 164 & 0.38 & 0.29 & -0.20 & 0.21 & -13.14 & -13.85 & 31.73 & 31.03 + 165 & 1.23 & 0.64 & -0.72 & 0.37 & -13.59 & -13.89 & 31.28 & 30.98 + 166 & -0.24 & 0.47 & -0.38 & 0.63 & -13.79 & -14.22 & 31.09 & 30.66 + 167 & 0.36 & 0.28 & -0.35 & 0.25 & -13.15 & -13.83 & 31.73 & 31.05 + lrrrrrrr 1 & 10:45:22.01 & -59:40:52.5 & 1.589 & 16.914 & 1.930 & + 3 & 10:45:16.25 & -59:41:41.0 & 0.421 & 13.377 & 0.770 & + 4 & 10:45:14.80 & -59:40:07.1 & & 18.452 & 1.734 & + 6 & 10:45:11.79 & -59:41:33.8 & 2.009 & 19.178 & 2.054 & + 7 & 10:45:08.57 & -59:40:16.7 & 1.106 & 15.410 & 1.350 & + 11 & 10:45:04.21 & -59:40:10.7 & & & & 3 + 12 & 10:45:03.89 & -59:38:19.9 & & & & 3 + 14 & 10:45:02.93 & -59:42:25.9 & 1.759 & 17.852 & 2.058 & 3 + 16 & 10:44:52.79 & -59:43:19.1 & 1.514 & 17.553 & 1.828 & 1b + 17 & 10:44:52.27 & -59:41:55.1 & 1.089 & 15.610 & 1.474 & + 18 & 10:44:50.39 & -59:40:17.1 & & 16.824 & 1.866 & 2 + 19 & 10:44:48.42 & -59:41:13.8 & 1.547 & 15.879 & 2.027 & 3 + 20 & 10:44:46.15 & -59:42:28.8 & & 20.269 & 2.948 & 1b + 21 & 10:44:45.30 & -59:39:18.9 & 1.185 & 15.605 & 1.578 & 2 + 23 & 10:44:44.25 & -59:42:33.5 & 0.305 & 12.750 & 0.602 & + 24 & 10:44:42.68 & -59:41:23.4 & & 18.059 & 2.079 & 2 + 26 & 10:44:29.38 & -59:41:44.2 & 1.538 & 18.729 & 1.744 & 1d + 27 & 10:44:27.04 & -59:44:39.3 & 1.279 & 16.151 & 1.487 & 3 + 29 & 10:45:43.74 & -59:41:47.6 & 1.941 & 17.759 & 2.455 & 1d + 30 & 10:45:38.41 & -59:42:07.5 & 0.556 & 13.354 & 1.080 & + 31 & 10:45:28.83 & -59:43:48.1 & & & & 1c + 32 & 10:45:28.54 & -59:42:46.7 & 0.815 & 16.646 & 0.920 & 1a + 33 & 10:45:28.51 & -59:41:55.3 & 0.390 & 14.180 & 0.717 & + 34 & 10:45:28.02 & -59:41:38.6 & 1.227 & 15.785 & 1.530 & + 35 & 10:45:25.61 & -59:43:50.8 & 1.368 & 16.536 & 1.742 & + 36 & 10:45:24.49 & -59:42:20.8 & 1.323 & 18.658 & 1.948 & + 37 & 10:45:20.88 & -59:45:49.1 & 1.665 & 17.220 & 2.244 & + 38 & 10:45:20.61 & -59:41:32.6 & 1.403 & 18.620 & 1.881 & 1b , 3 + 40 & 10:45:16.57 & -59:42:14.6 & 1.583 & 18.050 & 2.050 & + 41 & 10:45:16.24 & -59:43:16.7 & & & & 3 + 42 & 10:45:15.01 & -59:40:21.7 & 1.694 & 18.429 & 1.828 & 2 + 43 & 10:45:14.54 & -59:42:46.8 & 1.502 & 17.594 & 1.962 & 3 + 46 & 10:45:11.04 & -59:44:57.5 & 1.255 & 17.068 & 1.805 & 3 + 47 & 10:45:11.02 & -59:42:39.2 & 1.813 & 18.379 & 2.330 & + 48 & 10:45:10.68 & -59:42:54.0 & 1.305 & 17.163 & 1.616 & 1b + 49 & 10:45:10.42 & -59:44:09.5 & 1.323 & 19.380 & 2.242 & 1b + 50 & 10:45:10.33 & -59:40:28.6 & 1.267 & 16.670 & 1.450 & 2 + 51 & 10:45:09.89 & -59:40:18.5 & 1.354 & 16.890 & 1.565 & 3 + 52 & 10:45:09.76 & -59:42:45.5 & 1.326 & 16.533 & 1.734 & 1b , 3 + 54 & 10:45:08.35 & -59:38:47.3 & 0.338 & 12.482 & 0.636 & + 57 & 10:45:06.53 & -59:40:15.5 & & & & 3 + 60 & 10:45:05.72 & -59:43:43.8 & 1.272 & 15.084 & 1.658 & + 61 & 10:45:05.01 & -59:37:33.1 & 1.092 & 15.613 & 1.352 & 1d + 63 & 10:45:03.62 & -59:43:20.5 & & & & 3 + 64 & 10:45:02.18 & -59:41:34.4 & 1.430 & 17.094 & 1.833 & 3 + 68 & 10:45:01.10 & -59:39:17.4 & 1.353 & 17.376 & 1.646 & 1b , 2 + 69 & 10:45:01.08 & -59:38:53.4 & 0.725 & 18.364 & 0.529 & 3,2,6 + 70 & 10:45:00.99 & -59:45:15.6 & 0.980 & 14.345 & 1.041 & 1b + 71 & 10:44:57.96 & -59:47:09.5 & 1.379 & 16.818 & 1.919 & 1b + 72 & 10:44:58.08 & -59:42:31.3 & 1.047 & 17.836 & 1.518 & 1a 3 + 73 & 10:44:57.07 & -59:38:26.7 & 0.990 & 13.585 & 1.169 & 2 + 74 & 10:44:54.83 & -59:43:51.6 & 1.287 & 16.773 & 1.634 & + 75 & 10:44:54.53 & -59:42:43.1 & 1.414 & 17.837 & 1.932 & 1b + 76 & 10:44:53.75 & -59:43:24.2 & 1.084 & 15.985 & 1.621 & + 77 & 10:44:53.25 & -59:41:35.9 & 0.988 & 18.681 & 2.393 & 3 + 78 & 10:44:51.40 & -59:44:06.9 & & 19.439 & 2.490 & + 79 & 10:44:51.14 & -59:39:38.0 & & 20.681 & 3.615 & 2 + 80 & 10:44:50.14 & -59:44:43.1 & 1.379 & 18.640 & 1.822 & + 81 & 10:44:50.13 & -59:42:38.8 & & & &
1c + 82 & 10:44:50.05 & -59:34:29.0 & 1.429 & 16.784 & 1.732 & + 84 & 10:44:47.30 & -59:41:16.2 & & & &
1a + 85 & 10:44:46.54 & -59:34:12.4 & 0.422 & 13.161 & 0.662 & + 86 & 10:44:43.98 & -59:39:55.5 & 1.299 & 17.788 & 1.771 & 2 + 87 & 10:44:41.96 & -59:42:15.9 & 1.210 & 16.702 & 1.480 & 1d + 89 & 10:44:41.73 & -59:42:54.5 & 0.290 & 13.852 & 0.572 & 1d + 90 & 10:44:41.34 & -59:44:08.8 & 1.222 & 17.015 & 1.681 & 1b + 91 & 10:44:41.15 & -59:38:18.9 & 1.295 & 16.442 & 1.711 & 2 + 92 & 10:44:39.88 & -59:43:58.5 & 1.190 & 15.996 & 1.424 & + 93 & 10:44:38.08 & -59:41:31.4 & 1.539 & 16.985 & 1.750 & 3 + 94 & 10:44:38.06 & -59:43:26.2 & 1.059 & 15.370 & 1.364 & + 95 & 10:44:32.84 & -59:43:06.8 & 0.968 & 15.803 & 1.259 & + 97 & 10:44:28.55 & -59:41:45.3 & 1.241 & 16.900 & 1.641 & + 98 & 10:44:28.49 & -59:39:19.7 & & & & 5 + 99 & 10:44:27.80 & -59:45:22.0 & & & & 3 + 100 & 10:44:26.28 & -59:42:41.0 & 1.361 & 17.404 & 1.811 & + 101 & 10:44:25.34 & -59:41:11.5 & & 18.389 & 2.361 & + 104 & 10:45:47.17 & -59:47:50.9 & & & & + 105 & 10:45:41.58 & -59:38:20.4 & 0.832 & 14.133 & 0.896 & + 106 & 10:45:37.19 & -59:48:06.6 & & & & + 107 & 10:45:28.01 & -59:43:19.8 & 0.674 & 15.435 & 0.802 & 1a + 108 & 10:45:26.40 & -59:40:29.1 & 1.206 & 18.310 & 1.688 & 1b + 109 & 10:45:16.84 & -59:46:29.3 & 0.646 & 13.975 & 0.707 & + 110 & 10:45:13.47 & -59:44:04.4 & 1.596 & 18.379 & 1.747 & 3 + 111 & 10:45:13.18 & -59:35:21.8 & 1.640 & 17.813 & 2.049 & 1d + 112 & 10:45:11.09 & -59:32:58.4 & 0.686 & 13.369 & 0.671 & + 116 & 10:45:03.27 & -59:42:18.1 & & & & 3 + 119 & 10:44:59.28 & -59:44:59.8 & 1.571 & 18.136 & 2.019 & + 120 & 10:44:57.92 & -59:39:59.6 & 0.385 & 12.587 & 0.735 & 3,2 + 121 & 10:44:55.97 & -59:44:48.7 & 1.654 & 17.527 & 1.824 & 1b + 122 & 10:44:52.95 & -59:46:34.8 & & 19.483 & 2.478 & + 123 & 10:44:52.18 & -59:45:14.9 & 1.729 & 17.908 & 2.095 & 1d + 124 & 10:44:51.75 & -59:43:05.6 & 1.210 & 19.472 & 2.244 & 3 + 125 & 10:44:47.39 & -59:35:07.0 & 1.443 & 17.508 & 2.498 & 3 + 126 & 10:44:46.38 & -59:36:03.2 & 1.594 & 18.617 & 2.086 & + 127 & 10:44:46.18 & -59:44:49.8 & 1.418 & 17.517 & 1.710 & 1b,3 + 128 & 10:44:41.83 & -59:40:47.4 & 1.217 & 16.087 & 1.667 & 2 + 129 & 10:44:41.16 & -59:44:22.8 & 1.582 & 18.162 & 1.870 & + 130 & 10:44:36.68 & -59:44:23.9 & & 18.285 & 2.012 & 3 + 131 & 10:44:36.40 & -59:44:10.7 & 1.455 & 17.320 & 1.814 & 1b,3 + 132 & 10:44:35.70 & -59:43:51.1 & 1.519 & 17.361 & 1.850 & 3,2 + 133 & 10:44:34.95 & -59:44:05.6 & & 17.231 & 2.001 & 3,2 + 134 & 10:44:33.93 & -59:37:33.5 & 1.554 & 17.672 & 1.962 & 2 + 136 & 10:44:29.73 & -59:43:33.1 & 1.503 & 17.322 & 1.829 & + 137 & 10:44:28.58 & -59:40:51.5 & & 18.699 & 1.890 & 1b + 138 & 10:44:24.76 & -59:39:58.5 & 2.246 & 18.351 & 2.123 & + 139 & 10:44:23.72 & -59:41:15.0 & 1.304 & 16.867 & 1.710 & 1b + 140 & 10:44:22.86 & -59:44:32.8 & 1.642 & 17.819 & 2.160 & + 142 & 10:44:19.94 & -59:42:30.2 & 1.407 & 17.274 & 1.825 & 4,2 + 143 & 10:44:17.13 & -59:41:56.5 & & 17.507 & 2.238 & + 144 & 10:44:09.15 & -59:37:17.9 & & 17.744 & 2.070 & 4 + 145 & 10:44:08.96 & -59:34:33.0 & & & & 3 + 146 & 10:44:02.66 & -59:39:44.9 & & & & + 147 & 10:45:37.02 & -59:47:03.1 & & & & + 148 & 10:45:36.57 & -59:44:11.7 & 1.028 & 13.443 & 1.892 & + 149 & 10:45:23.52 & -59:45:45.5 & 2.081 & 20.192 & 2.745 & 1b + 150 & 10:45:07.31 & -59:44:47.1 & 1.417 & 18.856 & 1.949 & + 152 & 10:44:38.23 & -59:44:35.8 & 1.729 & 19.120 & 1.758 & 1d + 153 & 10:44:36.47 & -59:33:40.2 & & 19.105 & 1.716 & 4 + 154 & 10:44:33.34 & -59:32:16.1 & 1.282 & 17.219 & 1.660 & + 155 & 10:44:29.37 & -59:32:18.6 & & & & 3 + 156 & 10:44:29.43 & -59:48:20.9 & & & & 6 + 157 & 10:44:24.76 & -59:45:55.6 & & & & + 158 & 10:44:12.47 & -59:34:26.5 & 1.369 & 17.371 & 1.927 & 4 + 159 & 10:44:12.81 & -59:33:41.8 & & & & + 160 & 10:44:06.61 & -59:36:13.4 & & & & + 161 & 10:44:04.89 & -59:32:36.8 & & & & 7 + 162 & 10:44:02.94 & -59:31:40.3 & & & & 7 + 163 & 10:44:00.70 & -59:34:41.6 & 1.555 & 16.553 & 1.998 & 3 + 164 & 10:45:19.68 & -59:32:19.0 & & & & + 165 & 10:44:49.38 & -59:47:30.2 & 1.798 & 18.863 & 2.334 & 4 + 166 & 10:44:26.25 & -59:37:39.7 & 1.644 & 18.998 & 1.995 & 2 + 167 & 10:44:16.45 & -59:33:09.0 & & & & 3 + hd93129ab & 103 & 6.90 & o3 i & 0.55 & -11.5 & -6.72 + hd93250 & 22 & 7.41 & o3 v & 0.49 & -10.7 & -6.54 + hd93205 & 25 & 7.76 & o3 v & 0.40 & -10.1 & -6.76 + hd93160 & 145 & 7.88 & o6 iii & 0.31 & -9.4 & -6.60 + hd93162 & 28 & 8.11 & wn7 + abs & 0.66 & -10.1 & -5.40 + hde303308 & 10 & 8.19 & o3 v & 0.46 & -9.8 & -7.00 + hd93204 & 96 & 8.48 & o5 v & 0.41 & -9.0 & -7.02 + -59@xmath412600 & 88 & 8.65 & o6 v & 0.51 & -9.0 & -6.80 + -59@xmath412603 & 83 & 8.82 & o7 v & 0.46 & -8.5 & -7.14 + hde303311 & 151 & 9.03 & o5 v & 0.45 & -8.6 & -6.84 + -59@xmath412635 & 44 & 9.27 & o8.5 v & 0.54 & -8.0 & -6.86 + -59@xmath412641 & 2 & 9.28 & o6 v & 0.61 & -8.7 & -7.00 + -59@xmath412636 & 5 & 9.31 & o8 v & 0.60 & -8.3 & -6.72 + -59@xmath412628 & 8 & 9.49 & b0 v & 0.44 & -7.5 & -6.94 + hd93343 & 45 & 9.60 & o7 v & 0.56 & -8.0 & -6.84 + tr16 - 9 & 58 & 9.77 & o9.5 v & 0.56 & -7.2 & -6.89 + tr16 - 10 & 135 & 9.86 & b0 v & 0.57 & -6.9 & -6.80 + tr16 - 23 & 59 & 10.00 & o7 v & 0.69 & -7.5 & -7.12 + tr16 - 115 & 39 & 10.09 & o8.5 v & 0.46 & -6.8 & -6.88 + tr16 - 3 & 9 & 10.16 & o8.5 v & 0.52 & -6.9 & -7.02 + tr16 - 244 & 102 & 10.78 & o3 i & 0.94 & -8.6 & -6.34 + tr16 - 22 & 55 & 10.93 & o8.5 v & 0.78 & -6.8 & -5.82 + tr16 - 11 & 141 & 11.19 & b1.5 v & 0.43 & -4.9 & -6.28 + 1 & 10 43 59.96 & -59 32 26.3 & 30.0 & 5.5 & 61.0 & 7.8 & 9.6 & 6.1 & 11.8 & o6.5 & 31.613 + 2 & 10 44 26.54 & -59 41 02.8 & 3.0 & 1.7 & 15.0 & 3.9 & -2.0 & 2.2 & 4.7 & b1.5 & 30.925 + 3 & 10 44 38.73 & -59 48 14.3 & 6.0 & 2.4 & 14.0 & 3.7 & 1.3 & 2.7 & 7.8 & ... & 31.208 + 4 & 10 44 56.27 & -59 33 04.0 & 2.0 & 1.4 & 7.0 & 2.6 & -0.3 & 1.7 & 8.0 & b0.5 & 31.228 + 5 & 10 44 00.97 & -59 35 46.2 & 12.0 & 3.5 & 13.0 & 3.6 & 7.7 & 3.7 & 9.5 & o9 & 31.382 + 6 & 10 45 08.99 & -59 40 40.7 & 6.0 & 2.4 & 24.0 & 4.9 & -2.0 & 2.9 & 0.8 & ... & 30.813 + 7 & 10 44 40.38 & -59 41 49.0 & 2.0 & 1.4 & 6.0 & 2.4 & -0.0 & 1.6 & 3.0 & b1 & 30.843 + 8 & 10 45 11.22 & -59 41 11.4 & 6.0 & 2.4 & 16.0 & 4.0 & 0.7 & 2.8 & 1.0 & b1 & 30.814 + 9 & 10 44 32.94 & -59 40 26.1 & 4.0 & 2.0 & 12.0 & 3.5 & -0.0 & 2.3 & 3.9 & b1 & 30.880 + 10 & 10 44 00.14 & -59 33 10.1 & 19.0 & 4.4 & 45.0 & 6.7 & 4.0 & 4.9 & 11.3 & ... & 31.560 + 11 & 10 45 11.69 & -59 41 16.3 & 3.0 & 1.7 & 19.0 & 4.4 & -3.3 & 2.3 & 1.0 & ... & 30.814 + 12 & 10 45 03.20 & -59 40 12.8 & 4.0 & 2.0 & 33.0 & 5.7 & -7.0 & 2.8 & 0.8 & b0.5 & 30.813 + 13 & 10 44 30.56 & -59 41 40.6 & 3.0 & 1.7 & 39.0 & 6.2 & -10.0 & 2.7 & 4.2 & b1 & 30.896 + 14 & 10 44 54.12 & -59 41 29.6 & 9.0 & 3.0 & 25.0 & 5.0 & 0.7 & 3.4 & 1.3 & b1 & 30.815 + 15 & 10 45 02.16 & -59 42 01.3 & 6.0 & 2.4 & 16.0 & 4.0 & 0.7 & 2.8 & 1.0 & b1 & 30.814 + 16 & 10 45 16.53 & -59 39 57.0 & 3.0 & 1.7 & 10.0 & 3.2 & -0.3 & 2.0 & 2.0 & ... & 30.821 + 17 & 10 43 57.62 & -59 33 39.3 & 19.0 & 4.4 & 33.0 & 5.7 & 8.0 & 4.8 & 11.2 & b1 & 31.550 + 18 & 10 44 05.86 & -59 35 12.2 & 14.0 & 3.7 & 16.0 & 4.0 & 8.7 & 4.0 & 9.4 & b1 & 31.365 + 19 & 10 45 12.70 & -59 42 48.8 & 10.0 & 3.2 & 20.0 & 4.5 & 3.3 & 3.5 & 2.1 & b2 & 30.822 + 20 & 10 44 25.52 & -59 33 09.9 & 4.0 & 2.0 & 10.0 & 3.2 & 0.7 & 2.3 & 9.2 & b1.5 & 31.353 + 21 & 10 44 53.82 & -59 37 48.5 & 4.0 & 2.0 & 5.0 & 2.2 & 2.3 & 2.1 & 3.5 & b2 & 30.859 + 22 & 10 45 05.93 & -59 44 19.0 & 5.0 & 2.2 & 13.0 & 3.6 & 0.7 & 2.5 & 3.3 & b2 & 30.852 + 23 & 10 45 07.17 & -59 40 58.7 & 50.0 & 7.1 & 221.0 & 14.9 & -23.8 & 8.6 & 0.5 & ... & 30.813 + 24 & 10 44 29.47 & -59 38 38.0 & 3.0 & 1.7 & 11.0 & 3.3 & -0.7 & 2.1 & 4.9 & b1 & 30.943 + 25 & 10 44 37.22 & -59 40 01.6 & 8.0 & 2.8 & 5.0 & 2.2 & 6.3 & 2.9 & 3.5 & b0.5 & 30.860 + 26 & 10 44 13.85 & -59 42 57.1 & 3.0 & 1.7 & 15.0 & 3.9 & -2.0 & 2.2 & 6.6 & b0 & 31.080 + 27 & 10 45 22.14 & -59 37 38.8 & 10.0 & 3.2 & 11.0 & 3.3 & 6.3 & 3.4 & 4.1 & b2 & 30.891 + 28 & 10 45 15.05 & -59 43 23.4 & 4.0 & 2.0 & 13.0 & 3.6 & -0.3 & 2.3 & 2.8 & ... & 30.835 + 29 & 10 45 09.78 & -59 42 57.2 & 2.0 & 1.4 & 27.0 & 5.2 & -7.0 & 2.2 & 2.1 & b1 & 30.822 + 30 & 10 45 16.76 & -59 43 11.0 & 2.0 & 1.4 & 28.0 & 5.3 & -7.3 & 2.3 & 2.7 & ... & 30.834 + 31 & 10 45 22.62 & -59 42 37.0 & 5.0 & 2.2 & 11.0 & 3.3 & 1.3 & 2.5 & 2.9 & ... & 30.838 + 32 & 10 45 06.39 & -59 42 35.9 & 1.0 & 1.0 & 19.0 & 4.4 & -5.3 & 1.8 & 1.6 & ... & 30.817 + hd93129a & 4 & n & + hd93129b & 4 & n & + hd93250 & 1 & n & + hd93205 & 1,3 & y & 6.1 + hd93160 & 1 & n & + hd93162 & & & + hde303308 & 1 & y : & + hd93204 & 1 & n & + -59@xmath412600 & 1 & n & + -59@xmath412603 & 1 & y & 2.2 + hde303311 & 1 & n & + -59@xmath412635 & 2 & y & 2.3 + -59@xmath412641 & 1 & y : & 4.0 + -59@xmath412636 & 5 & y & 3.6 , 5.0 + -59@xmath412628 & 1 & y : & 5.8 + hd93343 & 1 & y : & + tr16 - 9 & 1 & n & + tr16 - 10 & 1 & y : & + tr16 - 23 & 1 & n & + tr16 - 115 & 1 & n & + tr16 - 3 & 1 & n & + tr16 - 244 & & & + tr16 - 22 & & & + tr16 - 11 & & & + willis , a. j. and crowther , p. a. 1996 , in wolf - rayet stars in the framework of stellar evolution " , 33rd liege int .
, eds j. m. vreux , a. detal , d. fraipont - caro , e. gosset , and g. rauw ( univ . de liege ) , p. 109 | the region around the @xmath0 car nebula has three ob associations , which contain a wolf - rayet star and several massive o3 stars .
an early chandra acis - i image was centered on @xmath0 car and includes trumpler 16 and part of trumpler 14 .
the chandra image confirms the well - known result that o and very early b stars are x - ray sources with l@xmath1 @xmath2 10@xmath3 l@xmath4 over an x - ray luminosity range of about 100 .
two new anomalously strong x - ray sources have been found among the hot star population , tr 16 - 244 , a heavily - reddened o3 i star , and tr 16 - 22 , a heavily - reddened o8.5 v star .
two stars have an unusually large l@xmath1/l@xmath4 : hd 93162 , a wolf - rayet star ( and possible binary ) , and tr 16 - 22 , a possible colliding wind binary . in addition , a population of sources associated with cool stars is detected . in the color - magnitude diagram , these x - ray sources sit above the sequence of field stars in the carina arm .
the ob stars are on average more x - ray luminous than the cool star x - ray sources .
x - ray sources among a stars have similar x - ray luminosities to cooler stars , and may be due to cooler companions .
upper limits are presented for b stars which are not detected in x - rays .
these upper limits are also the upper limits for any cool companions which the hot stars may have .
hardness ratios are presented for the most luminous sources in bands 0.5 to 0.9 kev , 0.9 to 1.5 kev , and 1.5 to 2.04 kev .
the available information on the binary nature of the hot stars is discussed , but binarity does not correlate with x - ray strength in a simple way . |
on one hand recent cosmological observations , and particuarly the hubble diagram for type ia supernovae @xcite have led to the idea that the universe may be dominated by a component with negative pressure @xcite and thus that today the universe is accelerating .
such a component can also , if one sticks to the prediction of inflation that @xmath3 , account for the `` missing energy '' . yet , many candidates have been proposed like a cosmological constant , a `` dynamical '' cosmological constant @xcite , cosmic strings @xcite or a spatially homogeneous scalar field rolling down a potential @xcite . on the other hand , potentials decreasing to zero for infinite value of the field have been shown to appear in particle physics models ( see e.g. @xcite ) . for instance , exponential potentials arise in high order gravity @xcite , in kaluza - klein theories which are compactified to produce the four dimensional observed universe @xcite or can arise due to non - perturbative effects such as gaugino condensation @xcite . inverse power law potentials can be obtained in models where supersymmetry is broken through fermion condensates @xcite .
this gives one more theoretical motivation to study the cosmological implications of a field with such potentials .
the cosmological solutions with such a field have first been studied by ratra and peebles @xcite ( see also @xcite ) who showed the existence and stability of scaling solutions in respectively a field , radiation or matter dominated dominated universe for a field evolving in an exponential and inverse power law potential . a complete study in the framework of barotropic cosmologies in the case of the exponential potential @xcite
show that the solutions were stable to shear perturbations and to curvature perturbations when @xmath4 , but that for realistic matter ( such as dust ) these solutions were unstable , essentialy to curvature perturbations .
liddle and scherrer @xcite made a complete classification of the field potentials and show that power law potentials also lead to scaling solutions ( i.e. to solutions such that the field energy density @xmath5 , behaves as the scale factor at a given power ) and the coupling of the field to ordinary matter has been considered in @xcite .
such solutions are indeed of interest in cosmology since they provide a candidate for a component with negative pressure .
cosmological models with a scalar field have started to be investigated @xcite for different kind of potentials as cosine potential @xcite , exponential potential @xcite and inverse power law potentials @xcite .
it has also been shown that the luminosity distance as a function of redshift @xcite or the behaviour of density perturbations in the dust era as a function of redshift @xcite can be used to reconstruct the field potential .
however , all these studies have been done under the hypothesis that the field is minimally coupled to the metric .
it is known that terms with such a non - minimal coupling @xmath6 between the scalar curvature @xmath7 and the field @xmath8 can appear when quantizing fields in curved spacetime @xcite , in multi - dimensional theories @xcite like superstring and induced gravity theories @xcite .
since these theories predict both the existence of scalar fields with potential or power - law potential and non - minimal coupling , it is of interest to study the influence of this coupling , and for instance the robustness of the existence and stability of scaling solutions .
the influence of such a coupling during an inflationary period and the existence of inflationnary attractors have yet been examined ( see e.g. @xcite ) . in this article , we study the stability of scaling solutions of a non - minimally coupled scalar field .
we first present ( [ ii ] ) the main notations and equations . after having , in
[ iii ] , briefly recalled the standard approach for determining the potentials that can give rise to such behaviour for a minimally coupled scalar field , we investigate the inverse power law potentials ( [ iv ] ) and the exponential potential ( v ) . in the latter case ,
we study the two limiting situations @xmath1 and @xmath9 and then have a heuristic discussion in the general case .
we assume that the universe is described by a friedmann - lematre model with euclidean spatial sections so that the line element reads @xmath10 \equiv { \bar g}_{\mu\nu}dx^\mu dx^\nu,\label{metrique_bgd}\ ] ] where @xmath11 is the scale factor and @xmath12 the cosmic time .
greek indices are running from 0 to 3 and latin indices from 1 to 3 .
the hubble parameter h is defined as @xmath13 , where a dot denotes a derivative with respect to @xmath12 .
we also introduce @xmath14 the covariant derivative associated to @xmath15 .
we assume that the matter content of this universe is composed of a perfect fluid and a homogeneous scalar field @xmath8 coupled to gravity and coupled to matter only through gravity .
the fluid energy density , @xmath16 , and pressure , @xmath17 , are related through the equation of state @xmath18 where @xmath19 refers to `` background '' .
the conservation of the energy - momentum of the fluid reduces to @xmath20 the scalar field @xmath8 evolves in a potential @xmath21 and its dynamics is given by the lagragian @xmath22\sqrt{-\bar g}d^4x , \label{action}\ ] ] where @xmath0 is the field - metric coupling constant ( @xmath23 corresponds to a minimally coupled field and @xmath24 to a conformally coupled field ) , @xmath7 is the scalar curvature of the spacetime .
no known fundamental principle predicts the functional form @xmath25 and we will assume that @xmath26 .
this is however the only choice that allows for a dimensionless @xmath0 .
the equation of evolution is then obtained by varying the action ( [ action ] ) with respect to the field which leads to the klein - gordon equation @xmath27 where @xmath28 . for an homogeneous field in the spacetime ( [ metrique_bgd ] )
, it reduces to @xmath29 the energy density and the pressure of this field are defined by @xmath30 and we define @xmath31 .
the energy - momentum tensor @xmath32 is obtained by varying the action ( [ action ] ) with respect to the metric @xmath33 and reads @xmath34 @xmath35 being the einstein tensor of the background metric .
this expression can be compared to standard results ( see e.g. @xcite ) .
note however that when the signature of the metric is @xmath36 , one has to change the sign of @xmath33 , @xmath37 , @xmath38 , and @xmath39 while @xmath32 and @xmath40 remain unaffected .
it is then straigthforward to check that the density and the pressure ( [ defprho ] ) of the scalar field are given by @xmath41 and that the conservation of the energy - momentum of the field ( @xmath42 ) reduces to the klein - gordon equation ( [ klein_gordon ] ) @xmath43 [ we have used that the scalar curvature and the einstein tensor are respectively given by @xmath44 , @xmath45 and @xmath46 .
the equation of state of the field is defined by @xmath47 the matter content being described , we can write the einstein equations which dictate the dynamics of the spacetime and , in our case , reduces to the friedmann equations @xmath48 with @xmath49 .
one equation of the set ( [ cons_fluide],[phiflu ] , [ friedmann],[friedmann2 ] ) is redundant due to the bianchi identities .
it will also be usefull to introduce the density parameter of a component @xmath50 as @xmath51 .
they must satisfy ( from equation [ friedmann ] ) the constraint @xmath52
in this section , we briefly recall the `` standard '' procedure to show that there exist scaling solutions and to determine the potentials that give rise to such solutions .
this presentation will also enable us to understand the differences with the more general case of non - minimally coupled scalar fields .
we follow the approach initiated by ratra and peebles @xcite and others @xcite where one , assuming a scaling form for @xmath5 , derives an equation for @xmath53 and then uses ( [ rhophi]-[pphi ] ) to deduce the associated potential . the equation ( [ cons_fluide ] ) of evolution of the background fluid density can be integrated to give @xmath54 where a subscript 0 refers to quantities evaluated at a given initial time .
we now look for scaling solutions , i.e. solutions such that @xmath55 since @xmath56 $ ] , we deduce that @xmath57 $ ] ( let us emphasize that this is a priori no longer true when @xmath58 ) . using equations ( [ rhophi]-[pphi ] ) , such a solution must satisfy @xmath59 now , the friedmann equation ( [ friedmann ] ) implies that the field should satisfy @xmath60 this can be integrated for different relative values of @xmath61 and @xmath62 . in that situation , ( [ dphidx ] )
leads to @xmath64 and then using ( [ phistate ] ) , to the potential @xmath65 this solution corresponds to the scalar field dominated universe of ratra and peebles @xcite and to their scaling solution in the case @xmath66 and @xmath67 ( i.e. radiation or matter dominated universe ) .
note that with such a potential , @xmath5 will , by construction , mimic the evolution of the background fluid and that we do not have to assume that @xmath68 .
note however that , if the scalar field has reached the attractor from very early time , @xmath5 behaves like radiation and thus contributes to a non - negligible part of the radiation content during the nucleosynthesis and it has been shown that it implies the constraint @xmath69 @xcite .
moreover , since @xmath70 in the matter era , such a field will not explain the supernovae measurements ( which seem to favor @xmath71 @xcite ) even if it can account for a substantial part of the dark matter . in this case
we have @xmath73 which can be integrated ( @xmath74 ) to give @xcite @xmath75}.\label{xpot}\ ] ] again , using ( [ phistate ] ) , we can deduce the potential @xmath76 @xmath77 being given by ( [ xpot ] ) . indeed , we only get the potential in a parametric form , but when @xmath78 ( i.e. when the perfect fluid drives the evolution of the universe , @xmath77 can be eliminated from ( [ xpot]-[xpot2 ] ) to give @xmath79 when @xmath66 and @xmath67 , we recover the ratra - peebles result @xcite as well as the liddle - scherrer result @xcite for all @xmath61 .
this parametric general form of the potential seems however not to have been exhibited before .
the former procedure can not be applied when the field is non - minimally coupled since it is impossible for instance to write a closed equation for @xmath80 as in ( [ dphidx ] ) . moreover , we are interested in a field evolving in a given potential .
we assume that the field evolves in an inverse power law potential and show that there exist scaling solutions , the stability of which is then studied .
we assume that the potential takes the form @xmath81 with @xmath82 being the planck mass .
the universe is dominated by the perfect fluid so that ( we assume @xmath83 ) @xmath84 redefining @xmath85 as @xmath12 and @xmath86 as @xmath8 , the klein - gordon equation ( [ klein_gordon ] ) becomes @xmath87 looking for a solution of this equations of the form @xmath88 , one obtains @xmath89 so that @xmath90 with @xmath91 .
this solution is only well - defined if @xmath92 one can then compute the energy and the pressure of this field by inserting this solution into ( [ rhophi]-[pphi ] ) and verify , after some algebra , that @xmath93 whatever the value of @xmath0 .
this shows that the scaling solution does not depend on the coupling in the sense that @xmath94 is independent of @xmath0 .
this relation generalises the one found for minimally - coupled scalar fields @xcite . as emphasized in the previous section
, the scaling solution does not depend on the coupling @xmath0 .
the stability of such a solution is known when @xmath23 @xcite , we now have to study it when @xmath58 .
following @xcite , we define the new set of variables @xmath95 where @xmath96 is the scaling solution ( [ scaling_plaw ] ) .
we set @xmath97 .
using @xmath98 , @xmath99 and @xmath100 , equation ( [ kg_law ] ) reduces to @xmath101 + \frac{12}{m}\left[\frac{4}{m}-1\right]\xi\right)\left(u - u^{-(\alpha+1)}\right)=0 . \label{eq_lambda}\ ] ] the scaling solution corresponds to the critical point @xmath102 . introducing @xmath103 and linearising around this critical point [ @xmath104
, we obtain @xmath105 + \frac{6(\alpha+2)}{m}\left[\frac{4}{m}-1\right]\xi\right ) & \left(1-\frac{6}{m}-\frac{4}{\alpha+2}\right ) \end{array } \right ) \left ( \begin{array}{l } \epsilon \\ v \end{array } \right).\ ] ] the eigenvalues , @xmath106 , of this system are @xmath107 } \label{racine_lambda}\ ] ] this expression reduces to the liddle - scherrer result @xcite when @xmath23 [ with @xmath108 and to the ratra - peebles result @xcite when either @xmath66 or @xmath67 . a necessary and sufficient condition for the critical point to be stable , is the negativity of the real part of the two eigenvalues . defining @xmath109 as @xmath110 the two solutions of ( [ racine_lambda ] ) are real when @xmath111 .
thus , * when the two eigenvalues are complex , @xmath112 and their real part is given by , @xmath113 and the scaling solution is stable if @xmath114 * when the two eigenvalues @xmath106 are real , their product is ( from [ eq_lambda ] ) @xmath115 + \frac{6(\alpha+2)}{m}\left[\frac{4}{m}-1\right]\xi\right)>0,\ ] ] because of the condition ( [ cond ] ) .
they are thus of the same sign , their sum being @xmath116 they will both be negative if @xmath117 and thus , the solution will be stable only if @xmath118 as in the case @xmath119 . in conclusion , we find that _ whatever _ the coupling constant @xmath0 , the scaling solution ( [ scaling_plaw ] ) will be stable if and only if @xmath114 the value of @xmath0 only determines the nature of the stable point , i.e. wether it is a stable spiral or a stable node .
this generalises the study by liddle - scherrer @xcite to a non - minimally coupled scalar field .
we now focuse on potentials of the form @xmath120 and work under the same assumptions as in [ iv ] . redefining @xmath12 and @xmath8 as in [ iv ]
, the klein - gordon equation now reads @xmath121 when the coeficient of @xmath8 vanishes , i.e. in a radiation dominated universe ( @xmath67 ) or when @xmath23 , one can find a special solution of the form @xmath122 .
we get that ( see [ exp ] with @xmath123 ) @xmath124 where the subscript @xmath125 refers to the scaling solution . in radiation era , this implies for instance that @xmath126 and the scalar field behaves like radiation .
indeed , this is a very special case since in that period @xmath127 and the field does not `` feel '' the non - minimal coupling and it evolves as if it were minimally coupled .
the complete study of these solution in function of the two parameters ( @xmath128 ) @xcite shows that when @xmath129 , the scaling solution @xmath96 is a stable node or spiral whereas , when @xmath130 , the late time attractor is the field dominated solution , which we do not consider here .
the convergence towards the solution @xmath96 is illustrated on figure [ fig1 ] .
now , in the most general case where @xmath131 , it is easy to realize that a solution of the form @xmath122 can not be solution of equation ( [ kg_exp ] ) .
let us first look at the effect of a small perturbation in @xmath0 in the sense that the potential term dominates over the coupling term in the klein - gordon equation .
for that purpose , we set @xmath133 the equation of evolution for @xmath134 can be deduced from the klein - gordon equation ( [ kg_exp ] ) @xmath135\hbox{e}^{2u-\lambda\phi_s}-\frac{12}{m } \left(\frac{4}{m}-1\right)\phi_s , \label{exp_int}\ ] ] where a prime denotes a derivative with respect to @xmath136 .
now , if we restrict to @xmath1 and linearise this equation using the expression ( [ sol_exp ] ) for @xmath96 , we obtain at zeroth order in @xmath0 @xmath137,\ ] ] the solution of which has the general form @xmath138.\label{toto}\ ] ] @xmath139 and @xmath140 are two independent solutions of the homogeneous equation . since @xmath141 , we deduce that if @xmath142 , the real parts of both @xmath143 and @xmath144 are negative , so that the two homogeneous solutions correspond to decaying modes and the particular solution is then an attractor ( if @xmath145 then @xmath143 and @xmath144 are real , otherwise they are complex and the homogeneous part will decay while oscillating ) . + indeed , this analysis is valid only as long as the potential term dominates over the coupling term in equation ( [ kg_exp ] ) , that is , as long as @xmath146 we recover that when @xmath147 or @xmath148 , @xmath149 and we are back to the minimally coupled case ( [ exp ] ) .
the behaviour of @xmath150 and @xmath151 in function of @xmath136 is shown on figure [ fig4 ] .
when @xmath1 , we see that , as expected , the solution is first dominated by the potential term but that , as time elapses the coupling term tends to become more dominant . in figure [ fig2 ] , we illustrate how the phase space trajectories are deformed due to the existence of this small coupling .
now , as long as @xmath152 , we can compute the equation of state of the field by inserting the particular part of the solution ( [ toto ] ) in ( [ rhophi]-[pphi ] ) which will take the form @xmath153 the difference in these two behaviours comes from the fact that @xmath154 if @xmath155 and @xmath156 when @xmath66 .
the exact forms of the functions @xmath157 , @xmath158 and @xmath159 can be obtained by doing an expansion of @xmath160 in @xmath161 . in figure
[ fig3 ] , we show the deviation of the equation of state from pure scaling and compare the former expansion to the numerical integration .
in conclusion we have that when @xmath1 and @xmath152 , the field converges towards a barotropic fluid .
let us now consider the case where initially the coupling term dominates over the potential term in the klein - gordon equation . at lowest order
, one has @xmath163 the solution of which is of the form @xmath164 when @xmath165 , since @xmath166 and @xmath167 , we deduce that if @xmath168 , the real part of both @xmath143 and @xmath144 are negative so that ( [ toto2 ] ) corresponds to two decaying modes [ it can be seen as a proof that the critical point @xmath169 is an attractor ] so that @xmath170 and the potential term will rapidly catch up the coupling term . when @xmath171 , @xmath172 and the real part of one of the two quantities , @xmath143 say , is positive when @xmath168 so that @xmath173 and the coupling term will dominates forever ( see figure [ fig4 ] ) and @xmath8 will behave as @xmath174 with @xmath175 for which , since the potential term is negligible , @xmath176 } { \alpha_+^2+\frac{24}{m}\xi\left[\alpha_++\frac{1}{m}\right]}.\ ] ] this solution has however to be excluded since one can check that it leads to @xmath177 .
the general case is more involved since we can not find any analytic solution to ( [ kg_exp ] ) .
first , when @xmath165 , we have seen that when either the potential or the coupling term dominates , the other slowly catch up .
we thus expect a late time solution which satisfies @xmath178 in figure [ fig4 ] , we plot the evolution of these two terms in the case where @xmath1 and in a more general case .
the two terms alternatively dominate and then converge to the same value . indeed ,
this is no proof .
+ we can however look for a general solution of the form @xmath179 with @xmath180 given by ( [ sol_exp ] ) and @xmath181 given by ( [ toto ] ) . inserting this expansion in the klein - gordon equation ( [ kg_exp ] ) , we obtain the hierarchy of equations @xmath182 the functions @xmath183 depends on all the solutions @xmath184 for @xmath185 .
all these equations have a solution of the form @xmath186 as in [ petit1 ] , the two homogeneous modes decay if @xmath142 and @xmath187 .
it can also be seen that @xmath188 will be a polynomial in @xmath136 of degre @xmath62 .
if the series @xmath189 converges then , @xmath190 from which we can conclude that if @xmath189 converges , there exists an attractor to the equation ( [ kg_exp ] ) given by ( [ tt2 ] ) . indeed , we can not demonstrate the convergence of this series in the general case . in figure [ fig6 ]
, we show the phase space trajectories showing the convergence towards this attractor .
when @xmath191 , the solution converges towards the attractor while oscillating so that the equation of state have wiggles and converges towards @xmath192 ( see figure [ fig6 ] ) .
when @xmath145 , this oscillatory behaviour does not appear whatever @xmath0 ( see figure [ fig6 ] ) .
+ as explained above , one has the solution ( [ sol_exp ] ) to the klein - gordon equation and this solution is an attractor ( see the phase analysis of figure [ fig5 ] ) and the field behaves as radiation .
the solution ( [ toto ] ) reduces to the two decaying modes . indeed
the derivation of ( [ t2 ] ) is no longer valid but , with the same method , one can show that , to first order in @xmath161 , @xmath193 where the two functions @xmath157 and @xmath158 are obained as in [ petit1 ] . note that for @xmath194 the @xmath0 terms of ( [ rhophi]-[pphi ] ) dominates the density and the pressure . in figure [ fig5 ] , we have plotted the phase space showing the convergence towards the attractor @xmath96 in the case of a conformally coupled scalar field and the evolution of its equation of state .
we integrate numerically equation ( [ kg_exp ] ) and use the poincar projection @xcite to represent the result in the plane @xmath195 @xmath196 this projection shrinks all the trajectory in the phase - space to the unit disk .
the point @xmath197 , @xmath198 and @xmath199 respectively correspond to ( @xmath200 ) , ( @xmath201 ) and ( @xmath202 ) .
note that the system is non - autonomous since equation ( [ exp_int ] ) depends on @xmath136 .
thus , curves can cross in the poincar representation but they will not cross in a 3d representation @xmath203 .
the following plots corresponds to the situation studied in the previous paragraphs .
we plot the evolution of the solution in the phase space @xmath195 and the evolution of the equation of state when @xmath1 .
to finish , we give some examples of evolution in the case @xmath2 .
in this article , we studied the influence of the coupling between the scalar curvature and the scalar field on the existence and stability of scaling solution of this field evolving in either an exponential or an inverse power law potential .
the motivation for considering such solutions is first that they can be a candidate for a matter component with negative pressure and second that they appear for a large class of potentials predicted by some theories of hight energy physics .
we first found a new parametric form of the potential that reduces to the inverse power law potential when the energy density of the fluid drives the evolution of the spacetime .
concerning the inverse power law potentials , we showed analytically that the existence and stability of a scaling solution does not depend on the coupling @xmath0 and that the equation of state of the field was always given by @xmath204 .
this generalises the work by ratra - peebles @xcite and liddle - scherrer @xcite .
the situation is more involved with exponential potentials since one can not find an analytic form for a scaling solution ( apart for a radiation dominated universe ) .
we then studied the effect of a small perturbation coupled to the scalar curvature and computed @xmath94 to first order in @xmath1 and showed that there always exists a time after which one can not neglect the effect of the coupling . in that limit
, we show that the equation of state was converging towards a barotropic equation of state and some numerical examples tend to show us that it should be the case whatever @xmath165 ( but this has not been demonstrated ) . indeed
, such potentials are not the most favored since they are constrained by nucleosynthesis to @xmath69 @xcite and can not explain ( when @xmath23 ) the supernovae measurements since @xmath70 .
note however that the convergence toward the barotropic equation @xmath192 is much slower when @xmath58 .
when @xmath171 , we have shown that there always exists a time after which the coupling term will dominate and thus that there exists a scaling solution but with a different equation of state ( as long as @xmath168 , and thus in a matter dominated era ) .
unfortunately , such a solution has to be rejected since it has negative energy . as pointed out by caldwell _
@xcite , a smooth time dependent field is unphysical since `` one has to take into account the back - reaction of the fluctuations in the matter components '' .
the cosmological consequences of such an inhomogeneous coupled scalar field in both exponential and inverse power law potentials ( such as the computation of the cmb anisotropies and the matter power spectrum ) will be presented later @xcite and our present study is only related to the implication of the homogeneous part of such a field .
it is a pleasure to thank pierre bintruy who raised my interest to the subject of quintessence , luca amendola , nuno antunes , nathalie deruelle and alessandro melchiorri for discussions , and ruth durrer and patrick peter for discussions and comments on the early versions of this text .
s. perlmutter _
et al . _ ,
nature * 391 * , 51 ( 1998 ) ; _ ibid .
_ astrophys .
j. * 493 * , 565 ( 1997 ) ; a.g .
_ to appear in astrophys
. j. [ astro - ph/9805201 ] ( 1998 ) i. zlatev , l. wang and p.j .
steinhardt [ astro - ph/9807002 ] ( 1998 ) k. coble , s. dodelson and j.a .
frieman , phys . rev . *
d55 * , 1851 [ astro - ph/9608122 ] ( 1997 ) d. spergel and u - l .
pen , astrophys .
j. * 491 * , l67 [ astro - ph/9611198 ] ( 1997 ) g. huey , l. wang , r. dave , r.r .
caldwell , p.j .
steinhardt [ astro - ph/9804285 ] ( 1998 ) p. bintruy , m.k .
gaillard and y - y .
wu , phys .
* b412 * , 288 [ hep - th/9702105 ] ( 1997 ) ; _ ibid _ nucl . phys . *
b493 * , 27 [ hep - th/9611149 ] ( 1997 ) p. bintruy [ hep - ph/9810553 ] ( 1998 ) j.d . barrow and s. cotsakis , phys . lett . *
b214 * , 515 ( 1988 ) ; d. wands , class . quant . grav . *
11 * , 269 ( 1988 ) . c. wetterich nuc
* b252 * , 668 ( 1985 ) b. de carlos , j.a .
casas and c. muoz , nuc .
b399 * , 623 ( 1993 ) b. ratra and p.j.e .
peebles , phys . rev . *
d37 * , 3406 ( 1988 ) d. wands , e.j . copeland and a.r .
liddle , ann .
sci . * 688 * , 647 ( 1993 ) c. wetterich , astron .
* 301 * , 321 [ hep - th/9408025 ] ( 1995 ) e.j .
copeland , a.r .
liddle and d. wands , phys .
* d57 * , 4686 [ gr - qc/9711068 ] ( 1997 ) a.p .
billyard , a.a .
coley and r.j .
van den hoogen , phys . rev . *
d58 * , 123501 [ gr - qc/9805085 ] ( 1998 ) a.r .
liddle and r.j .
scherrer [ astro - ph/9809272 ] ( 1998 ) s. carroll , phys . rev.lett . *
81 * , 3067 [ astro - ph/9806099 ] ( 1998 ) f. perrotta and c. baccigalupi , [ astro - ph/9811156 ] ( 1998 ) p.j.e .
peebles and a. vilenkin [ astro - ph/9810509 ] ( 1998 ) p.g .
ferreira and m. joyce , phys . rev .
* d58 * , 023503 [ astro - ph/9711102 ] ( 1997 ) p.t .
viana and a.r .
liddle , phys . rev . *
d57 * , 674 ( 1998 ) r.r .
caldwell , r. dave and p.j .
steinhardt phys .
80 * , 1582 [ astro - ph/9708069 ] ( 1998 ) d. huterer and m. turner , submitted to phys .
[ astro - ph/9808133 ] ( 1998 ) a.a .
starobinsky , [ astro - ph/9810431 ] ( 1998 ) n.d .
birrell and p.c.w .
davis , _ quantum fields in curved spacetime _ ( cambridge university press ) ( 1982 ) l.h .
ford , phys . rev . *
d35 * , 2955 ( 1987 ) k. maeda , class .
. grav . * 3 * , 233 ( 1986 ) f.s .
acetta , d.j .
zoller and m.s .
turner , phys . rev . *
d31 * , 3046 ( 1985 ) l. amendola , m. litterio and f. occhionero , j. mod . phys . *
a5 * , 3861 ( 1990 ) ; l. amendola , d. bellisai and f. occhionero , phys.rev . *
d47 * , 4267 ( 1993 ) ; t. futumase et k. maeda , phys . rev . *
d39 * , 399 ( 1989 ) .
m.s . turner and m. white , phys .
* d56 * , r4439 [ astro - ph/9701138 ] ( 1997 ) i.s .
gradshteyn and i.m .
rizhik , _ table of integrals , series and products _ , ed .
academic , n.y .
j. wainwright and g.f.r .
ellis , _ dynamical systems in cosmology _
( cambridge university press ) ( 1997 ) a. melchiorri and j - p .
uzan , work in progress | we study the existence and stability of cosmological scaling solutions of a non - minimally coupled scalar field evolving in either an exponential or inverse power law potential .
we show that for inverse power law potentials there exist scaling solutions the stability of which does not depend on the coupling constant @xmath0 .
we then study the more involved case of exponential potentials and show that the scalar field will assymptotically behaves as a barotropic fluid when @xmath1 .
the general case @xmath2 is then discussed an we illustrate these results by some numerical examples . |
le tournoi franais des jeunes mathmaticiennes et mathmaticiens @xmath1 est un tournoi de mathmatiques destin aux lves de lyce qui existe depuis 2011 .
il se distingue dautres comptitions comme les olympiades de mathmatiques en proposant des problmes ouverts do nt les noncs sont connus lavance et en tant organis par quipes .
guides par des encadrants , les quipes composes de 4 , 5 ou 6 lves ont environ trois mois pour rflchir aux problmes proposs .
lors du tournoi , les diffrentes quipes prsentent leurs rsultats et en discutent avec les autres quipes sous la forme de joutes mathmatiques ( voir lappendice par matthieu lequesne et le deuxime auteur ) .
les problmes proposs sont inhabituels pour la plupart des lves , car ils nadmettent pas , la connaissance du jury , de solution complte .
pour les quipes , il sagit donc de comprendre le problme , de rsoudre des cas particuliers , de reprer les difficults + le but de ce texte est de montrer quel point le @xmath1 peut amener les lves de lyce entreprendre une dmarche de recherche scientifique .
cette histoire date de ldition 2014 et commence par un post dans le blog mathmatique demmanuel kowalski , date du 29 juin 2012 .
celui - ci , intitul `` a bijective challenge '' , commenait comme suit : + _ tienne fouvry , philippe michel and myself have just finished a new paper , which is available on my web page and will soon be also on arxiv .
[ ] today i want to discuss a by - product that we found particularly nice ( and amusing ) .
it can be phrased as a rather elementary - looking challenge : given a prime number @xmath2 , and an element @xmath3 of @xmath4 which is neither 0 , 4 or -4 modulo @xmath2 , let @xmath5 be the number of solutions @xmath6 of the congruence @xmath7 the challenge is to prove , bijectively if possible , that @xmath8 [ ] this sounds simple and elegant enough that an elementary proof should exist , but our argument is a bit involved .
first , the number @xmath5 is the number of points modulo @xmath2 of the curve with equation above , whose projective ( smooth ) model is an elliptic curve , say @xmath9 , over @xmath10 .
then we checked using magma that @xmath9 and @xmath11 are isogenous over @xmath10 , and this is well - known to imply that the two curves have the same number of points modulo @xmath2 . _
[ ] the best explanation of this has probably to do with the relation between the family of elliptic curves and the modular curve @xmath12 ( a relation whose existence follows from beauville s classification of stable families of elliptic curves over @xmath13 with four singular fibers , as c. hall pointed out ) , but we did nt succeed in getting a proof of all our statements using that link .
in fact , we almost expected to find the identities above already spelled out in some corner or other of the literature on modular curves and universal families of elliptic curvers thereon , but we did not find anything .
+ cette question de fouvry , kowalski et michel a t reprise pour ldition 2014 du @xmath1 , ainsi que pour sa version internationale , litym .
bien entendu , on ne peut pas lancer des lves de lyce directement sur une question ouverte comme celle - ci .
cest pourquoi lnonc a t modifi de la faon suivante pour amener les lves la question de ces chercheurs sans pour autant la leur poser frontalement .
voici lnonc prcis de ce problme lors du @xmath1 : + _ soit @xmath2 un nombre premier impair . on considre lensemble @xmath14 des entiers modulo @xmath2 . on note
@xmath15 lensemble des lments inversibles de cet ensemble , _
i.e. _ les @xmath16 tels quil existe @xmath17 vrifiant @xmath18 .
pour @xmath19 , on note @xmath20 un tel lment @xmath21 ( remarquer quil est unique dans @xmath10 ) .
_ dans ce problme , on veut tudier lquation @xmath22 o @xmath23 et @xmath24 . on note @xmath5 le
nombre de solutions en @xmath25 et @xmath21 pour un @xmath3 fix . 1 .
quel est le cardinal de lensemble @xmath26 ?
et celui des ensembles @xmath27 on pourra commencer par les calculer explicitement pour des petites valeurs de @xmath2 .
calculer @xmath28 .
3 . pour des petites valeurs de @xmath2 , calculer explicitement @xmath5 pour tout @xmath3 .
existe - t - il des couples @xmath29 avec @xmath30 et @xmath31 tels que
@xmath32 pour tout @xmath24 ?
existe - t - il @xmath30 tel que @xmath33 pour tout @xmath34 ?
tudier dautres transformations possibles de @xmath3 laissant stable la valeur de @xmath35 .
gnraliser ces rsultats , @xmath36 pour @xmath37 quelconque ( on pourra commencer avec les puissances des nombres premiers ) , puis aux corps finis quelconques .
+ comme on pourra le remarquer , on sest permis de poser des questions _ a priori _ plus difficiles que celle pose par fouvry - kowalski - michel ( qui est cache dans la question 4 ) .
ceci est un facteur commun de tous les problmes poss au @xmath1 , le but tant que le problme soit ouvert mme pour les jurys qui valueront les lves lors du tournoi ( ces jurys sont composs de chercheurs , doctorants et tudiants de master ou de grandes coles ) .
+ ce qui fait lintrt de ce problme en particulier , cest quil est issu dune question pose par des vrais chercheurs en mathmatiques dautres chercheurs , donc toute solution devient _ ipso facto _ un rsultat de recherche ( certes pas de la plus haute importance , mais de la vraie recherche tout de mme ) .
et cest en mettant ensemble les diffrentes solutions des lves qui ont particip au @xmath1 2014 que lon a russi trouver une rponse lmentaire la question de fouvry - kowalski - michel .
grce au @xmath1 , des lycens franais ont donc fait un vritable petit apport la recherche mathmatique .
on prsente leurs arguments dans ce qui suit .
dans toute la suite @xmath2 dsigne un nombre premier impair .
pour tout @xmath38 , on note @xmath39 le nombre de solutions dans @xmath40 de lquation @xmath41 et @xmath5 le nombre de solutions dans @xmath42 de lquation @xmath43 notre but est de montrer le rsultat suivant
. [ thm ] pour tout @xmath44 , on a @xmath45 . la preuve du rsultat consiste en un comptage plus ou moins explicite des solutions .
on tablit dabord une borne suprieure sur les quantits @xmath5 , puis on montre que les quantits @xmath46 et @xmath5 sont congrues modulo @xmath2 .
ayant toutes ces informations , un simple argument de parit conclut laffaire .
+ on commence par donner une formule relativement explicite pour @xmath5 : on a @xmath47 on cherche donc calculer explicitement les @xmath48 .
lquation pour @xmath49 est quivalente @xmath50 , qui est une quation quadratique de discriminant @xmath51 .
elle a donc une solution si @xmath52 , deux si @xmath51 est un carr et aucune si @xmath51 nest pas un carr . on rappelle
donc la dfinition du symbole de legendre : @xmath53 sachant que le groupe multiplicatif @xmath15 est cyclique dordre @xmath54 , ce symbole peut aussi tre calcul par ce quon appelle le critre deuler : @xmath55 laide de ce symbole , on voit que lon peut exprimer les valeurs de @xmath56 comme suit : @xmath57 ces formules nous permettent de prouver deux rsultats lmentaires sur les valeurs @xmath5 .
le premier est la borne suprieure annonce plus haut , le deuxime concerne la parit de ces valeurs , ce qui sera utile la fin de la preuve .
[ majoration ] pour tout @xmath38 on a @xmath58 en appliquant lingalit de cauchy - schwarz on obtient @xmath59 o @xmath60 daprs .
pour calculer @xmath28 , on tudie la rpartition des valeurs de @xmath56 .
dune part , la formule nous dit que @xmath61 et @xmath56 vaut @xmath62 ou @xmath63 partout ailleurs .
dautre part on a @xmath64 donc @xmath56 prend @xmath65 fois la valeur @xmath62 , deux fois la valeur @xmath66 et @xmath67 fois la valeur @xmath63 .
on obtient donc @xmath68 [ parite ] @xmath5 est impair si et seulement si @xmath69 .
le cas @xmath70 dcoule immdiatement du lemme [ majoration ] .
supposons @xmath71 et soit @xmath72 : si @xmath73 est impair , alors @xmath74 daprs la formule , donc @xmath75 vaut @xmath76 ou @xmath77 .
ainsi , si @xmath78 alors tous les termes de la somme sont pairs donc @xmath5 est pair .
si @xmath79 alors le terme correspondant @xmath80 est impair et tous les autres sont pairs donc @xmath5 est impair
. le thorme de chevalley - warning est un rsultat classique de la thorie des corps finis .
soit @xmath81 un corps fini de caractristique @xmath2 et soit @xmath82 un polynme de degr @xmath83 coefficients dans @xmath81 .
alors , si @xmath84 , le nombre de solutions de lquation @xmath85 est congru @xmath62 modulo @xmath2 .
en particulier , si @xmath86 est une solution ( par exemple , si @xmath87 est homogne ) , alors il existe une solution non triviale .
la preuve de ce thorme , qui est bien classique ( cf . par exemple le premier chapitre de @xcite ) , consiste en une description astucieuse de lensemble des solutions qui utilise le petit thorme de fermat .
la mme technique nous permettra de dmontrer dans la suite le rsultat suivant : [ np(t ) congru a np(16/t ) ] pour tout @xmath44 on a @xmath88 .
lquation est quivalente @xmath89 donc @xmath90 vaut @xmath62 si @xmath91 est solution et @xmath66 sinon , daprs le petit thorme de fermat .
on peut ainsi compter les couples non - solutions modulo @xmath2 : @xmath92 en dveloppant le polynme et en remplaant @xmath3 par @xmath93 ( la bijection vidente @xmath94 donne @xmath95 ) , on obtient que @xmath96 et en rappelant que @xmath97 daprs le thorme de wilson , on obtient @xmath98 on a maintenant besoin du lemme suivant , cl dans la preuve de chevalley - warning .
soit @xmath99 .
alors @xmath100 vaut @xmath101 si @xmath54 divise @xmath102 , et @xmath62 sinon .
si @xmath54 divise @xmath102 , chacun des @xmath54 termes de la somme vaut @xmath66 daprs le petit thorme de fermat , do le rsultat .
sinon , soit @xmath103 un gnrateur du groupe multiplicatif @xmath40 . on a @xmath104
daprs le lemme , si le terme correspondant @xmath105 est non nul , alors @xmath106 et @xmath107 sont divisibles par @xmath54 , donc valent @xmath62 , @xmath54 ou @xmath108 .
si par exemple @xmath109 , alors @xmath110 et @xmath111 , et la contribution de ce terme vaut @xmath112 , et de mme lorsque @xmath113 , @xmath114 ou @xmath115 vaut @xmath54
. dans tous les autres termes non nuls , on a @xmath116 et @xmath117 do @xmath118 cest partir dici que notre preuve diffre de celle du thorme de chevalley - warning : alors que lhypothse sur le degr dans chevalley - warning garantit que tous les termes non triviaux sont nuls , il reste ici un certain nombre de termes contrler , le degr de tant plus lev que le nombre de variables .
on cherche donc regrouper les termes de selon la valeur de @xmath119 . on constate
que , @xmath54 tant pair , @xmath119 est toujours pair .
soit donc @xmath119 pair dans @xmath120\!]$ ] fix . on a @xmath121 o la dernire ligne utilise lgalit @xmath122 .
cest un cas particulier de lidentit de vandermonde , quon peut obtenir par exemple par double - comptage .
donc , en appliquant et respectivement @xmath3 et @xmath123 , on a dune part @xmath124 et dautre part @xmath125 ce qui donne , en changeant @xmath126 et @xmath127 , @xmath128 pour en dduire @xmath88 , il suffit donc de dmontrer que pour tout @xmath126 pair dans @xmath120\!]$ ] on a @xmath129 ou encore , en chassant les dnominateurs et en prenant des racines carres , @xmath130 ce qui se fait directement par rcurrence sur @xmath126 .
daprs la proposition [ np(t ) congru a np(16/t ) ] , on a @xmath88 , donc @xmath131 est divisible par @xmath2 .
de plus , on sait que @xmath5 et @xmath46 sont positifs et majors par @xmath132 daprs le lemme [ majoration ] , donc leur diffrence ne peut valoir que @xmath133 , @xmath62 ou @xmath2 .
enfin , @xmath5 est impair si et seulement si @xmath3 vaut @xmath134 ou @xmath135 daprs le lemme [ parite ] . les quantits @xmath5 et @xmath46 ont alors la mme parit et leur diffrence doit donc valoir @xmath62 , do le rsultat .
nous sommes partis des solutions prsentes par les deux quipes franaises qui participaient au tournoi international , do nt lune tait encadre par le premier auteur .
ces quipes taient donc les meilleures quipes du @xmath1 2014 , et elles ont eu loccasion dchanger avec les autres participants la comptition nationale .
cela fait de ces solutions un effort commun de plusieurs lves de trs haut niveau .
ainsi , les rsultats de la section [ section preliminaires ] sont entirement ds aux participants du tournoi .
ils ont t modifis dans ce texte pour gagner en brievet et rpondre spcifiquement lnonc original de fouvry - kowalski - michel .
comme il a t dmontr dans la section [ section coup de grace ] , ces rsultats suffisent pour rduire la preuve de lgalit en une preuve de congruence modulo @xmath2 .
cest l que lexprience a jou un rle , puisque cest leur encadrant ( _ i.e. _ le premier auteur ) qui a vu que les ides appliques dans chevalley - warning pouvaient tre utiles pour dmontrer une telle affirmation .
la section [ section chevalley - warning ] a t donc rdige par ses soins .
il est cependant important de remarquer que le niveau des techniques utilises dans cette dmonstration reste trs basique , y compris dans la section [ section chevalley - warning ] .
cest pourquoi on ne doute aucunement quelle aurait t la porte des lves si jamais leur encadrant leur avait suggr , tel un directeur de thse , daller lire la dmonstration de chavelley - warning pour y trouver des pistes .
malheureusement , cette ide nest venue lesprit du premier auteur quaprs le tournoi international .
+ enfin , nous tenons mentionner que , malgr le fait davoir trouv une preuve lmentaire la question de fouvry - kowalski - michel , le dfi nest pas pour autant compltement relev . en effet , ces chercheurs demandent une preuve non seulement lmentaire , mais _ bijective _ de lnonc du thorme [ thm ] , alors que notre preuve tablit seulement que les deux ensembles ont le mme cardinal .
dautres preuves non - bijectives ont dailleurs t trouves avant celle du @xmath1 , mais toujours avec des outils plus complexes .
il y a notamment : * la preuve mentionne sur le blog de kowalski , base sur des calculs informatiques sur des courbes elliptiques ; * une preuve ( non publie ) par david zywina , base sur ltude de certaines formes modulaires , comme il tait suggr la fin du post de kowalski ; * un rsultat encore plus gnral sur les reprsentations @xmath136-adiques du groupe fondamental dune courbe , d deligne et flicker ( cf .
@xcite ) , qui implique aussi lnonc .
aprs les courbes elliptiques , les formes modulaires , les reprsentations @xmath136-adiques et chevalley - warning , vous de trouver une meilleure preuve ! 10 counting local systems with principal unipotent local monodromy .
_ ann . of math .
178(3 ) , 921982 . algebraic twists of modular forms and hecke orbits .
25(2 ) , 580657 , 2015 . .
collection sup : `` le mathmaticien '' , no 2 .
presses universitaires de france , paris , 1970 .
le _ tournoi franais des jeunes mathmaticiennes et mathmaticiens _ ( @xmath1 ) est une comptition de mathmatiques qui propose des lves de lyce de travailler par quipe sur une srie de problmes ouverts pendant plusieurs mois . il est organis par lassociation animath et par une grande quipe danciens participants , aujourdhui tous en tudes suprieures .
+ cest david zmiaikou , jeune mathmaticien bilorusse , qui est lorigine du @xmath1 , ainsi que de sa version internationale , litym ( international tournament of young mathematicians ) .
pendant ses annes en tant que doctorant luniversit paris - sud , il a eu lide dorganiser un tournoi international limage des tournois de mathmatiques en bilorussie .
avec laide de bernardo da costa , galement doctorant la facult dorsay , et avec le soutien de lassociation animath , le premier tournoi international des jeunes mathmaticiens a eu lieu en 2009 dans les installations de luniversit avec la participation de six quipes provenant de quatre pays diffrents : bilorussie , bulgarie , france et russie . + en 2011 , deux ans aprs le succs du premier tournoi , un troisime doctorant
, igor kortchemski , sest ajout ce groupe pour crer la version franaise du tournoi , en partie pour mieux prparer les quipes franaises envoyes litym .
cest ainsi que le @xmath1 est n , accueillant pour sa premire dition quatre quipes venues de paris , versailles et strasbourg .
le @xmath1 sest bien dvlopp depuis . en effet ,
il accueillait dj dix - huit quipes en 2014 et , pour ldition 2015 , il a fait un grand saut en proposant pour la premire fois des tournois rgionaux aux quatre coins de la france .
+ le @xmath1 se compose aujourdhui de deux tapes : des tournois rgionaux rpartis partout en france , puis une finale nationale lcole polytechnique et lensta paristech .
une liste contenant entre 8 et 12 problmes est publie chaque anne en dcembre .
ainsi les lves peuvent regarder les problmes pendant leurs vacances et commencer travailler pendant le mois de janvier .
ils sont encadrs pour cela par des professeurs de lyce , mais aussi par des tudiants , doctorants ou chercheurs des universits voisines .
en 2017 , des tournois rgionaux sont prvus dans les villes de lille , lyon , paris , rennes , strasbourg et toulouse lors dun week - end pendant le mois davril .
la finale nationale aura lieu pour sa part au mois de mai .
+ lors des diffrents tournois , les quipes saffrontent par poules sous la forme de `` joutes mathmatiques '' .
dans celles - ci , chaque quipe occupe tour de rle trois fonctions : * _ prsenter la solution _ dun problme sur lequel ils ont travaill ; * _ critiquer la solution _ qui a t prsente : chercher les imprcisions , demander des explications et lancer un dbat constructif qui a lieu au tableau ; * _ valuer le dbat _ et ventuellement le prolonger sil y reste des choses intressantes discuter ; outre leur prestation orale , les lves sont aussi incits valuer les solutions crites par leurs camarades .
ils rdigent pour cela des notes de synthse sur la solution qui va tre prsente
. celles - ci sont galement values par le jury . les solutions sont soumises une semaine avant chaque tournoi .
peu aprs , un tirage au sort a lieu pendant lequel se dcide la formation des poules , les problmes qui seront prsents et lordre des prsentations .
+ on voit ainsi que , la diffrence des mathmatiques que les participants rencontrent au lyce et dans dautres comptitions , ce tournoi leur propose une vraie rencontre avec la recherche en mathmatiques . en effet : 1 .
les problmes proposs sont compltement ouverts puisque personne nen connat la solution gnrale , pas mme leurs auteurs .
les lves doivent donc chercher par eux - mmes en ayant droit toutes les ressources possibles , y compris des rsultats intermdiaires trouvs dans des livres , voire sur internet .
en cela , ils ont la mme libert quun chercheur dans son laboratoire face ses problmes mathmatiques quotidiens .
le travail se fait par quipes . on peut alors travailler
plusieurs sur un mme problme , en rpartissant les tches en fonction des comptences de chacun . on peut
se relire les uns les autres et on peut aussi travailler sur plusieurs problmes la fois .
les problmes recouvrant divers domaines des mathmatiques , lquipe entire simule ainsi un `` mini - laboratoire '' de recherche !
le rle de lencadrant est mettre en parallle avec celui dun directeur de thse . en effet ,
il est l pour aider les lves lorsquils se trouvent dans une impasse , pour leur donner des outils qui les aideront trouver la solution et pour leur donner des pistes de recherche , sans toutefois leur donner directement la solution .
les dbats qui ont lieu au tableau lors du tournoi nont rien envier aux discussions que les chercheurs ont tous les jours dans un laboratoire de mathmatiques . en effet ,
il sagit de quelquun qui prsente ses rsultats de manire concise et de son interlocuteur qui lui fait remarquer l o il a t imprcis , tout comme un chercheur qui prsente ses travaux au collgue du bureau voisin pour tre sr que tout semble correct .
le travail de relecture des solutions des autres quipes est similaire au travail des rapporteurs darticles de recherche . de plus , le fait de lire dautres solutions aide les lves comprendre ce qui a pu manquer en pdagogie dans leurs propres textes .
cest galement dans cette dynamique , en lisant de plus en plus darticles , quun chercheur apprend petit petit en crire de meilleurs .
le tournoi permet aussi les changes entre lves et chercheurs de premier niveau , lesquels participent de faon bnvole en tant que jurys . les prsidents de jury interagissent notamment avec les lves aprs le premier tour de chaque tournoi pour faire un compte - rendu des dbats afin de les faire progresser dans la prsentation de rsultats scientifiques .
en outre , depuis 2013 , les lves assistent des exposs donns par des chercheurs lors des crmonies douverture ou de clture .
+ bien que le rve des organisateurs du @xmath1 soit de donner une vraie exprience de recherche tous les lves de lyce , il faut avouer que ce tournoi sadresse aux meilleurs et plus motivs dentre eux . en effet ,
afin de pouvoir mener bien une rflexion pousse il est ncessaire que ces lves soient suffisamment laise avec les connaissances mathmatiques que requiert le programme de lyce .
le format des noncs est dj un dfi important en soi d lusage dun langage mathmatique qui est parfois un peu lourd , mme si tout est fait pour simplifier au maximum les concepts et ne pas alourdir les notations .
mais cela fait parti du dfi : ils ont du temps pour digrer ces nouveaux concepts , dfinitions et notations , ventuellement avec laide de leurs encadrants .
dailleurs , ce sont souvent ces problmes qui sont les plus plebiscits par les lves car ceux qui ont un nonc simple sont souvent les plus difficiles , tout comme en recherche ! + le tournoi sadresse donc tous les lycens de premire ou terminale scientifique ayant un bon niveau en mathmatiques mais surtout une grande curiosit et la volont daller chercher un peu plus loin que le programme scolaire .
ainsi , dans chaque lyce on pourrait constituer une quipe avec les meilleurs lves de premire ou terminale , encadrs par un ou deux professeurs .
on notera que lon a parfois vu des participants plus jeunes ( en seconde , voire au collge ) faire quipe avec des camarades plus gs , et galement des lves qui ntaient pas en filire scientifique mais nanmoins trs intresss par les mathmatiques .
parfois , certains tablissements possdent dj un club de mathmatiques anim par un enseignant et ce sont les lves du club qui forment une quipe .
le @xmath1 fournit ainsi ces clubs du contenu pour travailler pendant toute la premire moiti du deuxime smestre .
ces lycens se runissent ensuite une fois par semaine avec les professeurs pour changer sur leurs avances , et changent entre eux par mail , pendant les trois mois qui prcdent le tournoi ( la frquence peut sintensifier lapproche du tournoi ) .
+ ainsi , le public du tournoi ne se restreint pas aux tablissements les plus rputs , bien que ceux - ci restent encore sur - reprsents . il existe de nombreux exemples
pas particulirement rputs qui ont ralis de trs bonnes performances .
louverture des tournois rgionaux a permis daccueillir un public plus large et nous souhaitons continuer dans ce sens .
certaines quipes plus faibles sattaquent aux problmes avec leurs propres moyens , souvent moins thoriques et plus exprimentaux , mais vont parfois bien loin dans le tournoi . cela est d notamment au fait que les lves sapproprient trs bien les quelques concepts quils ont pu traiter et ont dvelopp
de bonnes capacits pdagogiques mises en valeur lors du dbat oral .
le principal facteur limitant reste certainement linvestissement horaire important que le tournoi demande de la part des encadrants .
ce sont en effet des professeurs qui dcident dencadrer bnvolement leurs lves et de sacrifier plusieurs soires et week - ends , souvent sans reconnaissance de leur hirarchie .
+ aprs un premier tournoi en 2011 avec seulement 17 participants , le @xmath1 a atteint un cap en 2014 avec 106 participants .
ceci a motiv la cration des tournois rgionaux en 2015 , lesquels rassemblent depuis plus de 150 participants .
le nombre de tournois na cess daugmenter : quatre tournois rgionaux en 2015 , puis cinq en 2016 et six en 2017 .
lobjectif long terme serait de parvenir en organiser un par acadmie .
+ ldition 2016 du tournoi , avec ses cinq tournois rgionaux , a runi des participants scolariss dans 16 acadmies distinctes , de nantes montpellier , en passant par nancy - metz .
ceci est possible grce au format du tournoi , qui permet de travailler en temps libre et distance .
dailleurs , des quipes mixtes inter - dpartamentales participent chaque anne . le @xmath1 se veut galement de plus en plus mixte avec une part de mathmaticiennes toujours croissante : de 12% de filles en 2012 , ce pourcentage est en augmentation constante et a atteint 23% en 2016 ( lanne 2011 nest pas reprsentative compte tenu
du nombre total de participants ) . | nous prsentons une dmonstration lmentaire dune question pose par fouvry , kowalski et michel sur les solutions de lquation @xmath0 sur un corps fini , en utilisant des rsultats obtenus par des lycens lors de leur participation au tournoi franais des jeunes mathmaticiennes et mathmaticiens .
2014 , ainsi que pour avoir servi de contact avec emmanuel kowalski par la suite .
nous remercions aussi ce dernier pour son soutien et pour avoir point les autres preuves existantes . ] |
due to geometrical frustration , transition metal spinel oxides with a general formula @xmath3o@xmath1 display a variety of unusual low - temperature properties .
the spin dynamics of these systems is usually described by a heisenberg antiferromagnet on the pyrochlore lattice .
this model is rather peculiar and its classical ground state is highly degenerate .
exact degeneracy can be lifted by various mechanisms , but the system still possesses many competing spin configurations with almost equal energies . as a result ,
when the temperature goes down , the system can evolve in a variety of ways : it can remain spin liquid down to the lowest temperatures due to quantum fluctuations , or choose a particular configuration either via the strongest order from disorder mechanism or through a structural phase transition which lowers the local symmetry of the lattice .
a peculiarity of transition metal spinels is that their magnetic ions often possess also an orbital degree of freedom .
this extra degree of freedom modulates the spin exchange and can at least partially lift the geometrical degeneracy of the underlying lattice .
however , the orbital degrees of freedom in geometrically frustrated lattices are by themselves frustrated , i.e. , many different orbital configurations have the same energy .
in this situation , the ordering in the orbital sector is coupled to the ordering in the spin sector , and the selection of a true ground state configuration becomes a non - trivial phenomena . of particular interest
is the orbital ordering in the @xmath4 systems on the pyrochlore lattice . in this work we study vanadium spinels of the type @xmath5v@xmath0o@xmath1 , where @xmath5 is a divalent ion like cd@xmath6 , zn@xmath6 , or mg@xmath6 . in these compounds ,
magnetically active v@xmath7 ions form a pyrochlore lattice and have two @xmath8 electrons in @xmath4-orbitals . due to strong hund s interaction ,
these two electrons form a state with @xmath2 .
all @xmath5v@xmath0o@xmath1 compounds show qualitatively similar structural and magnetic behavior , independently of what the divalent a ion is , and undergo two phase transitions a structural one and an antiferromagnetic one . we will be mainly discussing the physics of znv@xmath0o@xmath1 .
the structural transition occurs at a temperature @xmath9 50 k. @xcite below @xmath10 , the lattice shows a tetragonal distortion
the vanadium octahedra vo@xmath11 are uniformly flattened along the @xmath12 axis , and the symmetry is lowered from the cubic one to @xmath13 , which is the highest - symmetry tetragonal space group for the spinel structure .
the antiferromagnetic ( afm ) transition occurs at a slightly lower temperature @xmath14 of about 40 k. @xcite this temperature is significantly lower than the curie - weiss temperature @xmath15 , extracted from high temperature susceptibility , @xcite which underlines the importance of geometrical frustration .
the magnetic structure of av@xmath0o@xmath1 spinels at @xmath16 was first proposed by nizio @xcite and recently confirmed by reehius _
_ in ref .
along the diagonal @xmath17/[{\bar 1}10]$ ] direction in the @xmath18-plane , the ordering is antiferromagnetic + - + - ... , while along the two other diagonal directions @xmath19/[0{\bar 1}1]$ ] in the @xmath20-plane and @xmath21/[{\bar 1}01]$ ] in the @xmath22-plane , the spin ordering is in the form + + - - + + - - ... ( see fig . [ fig1 ] ) . at high temperatures
@xmath23 , inelastic neutron scattering data by lee _
et al . _
@xcite on the powder sample of znv@xmath0o@xmath1 still show strong low energy magnetic excitations which form a broad peak centered at @xmath24@xmath25 .
this broad peak is present also at @xmath26 , however it becomes asymmetric and shifts towards a smaller value of @xmath27 .
the asymmetry of the peak further increases in the magnetically ordered phase ( @xmath28 ) .
@xcite the value of @xmath27 and its temperature evolution can not be explained within a purely spin model , whose fundamental degrees of freedom are antiferromagnetic hexagonal spin loops with @xmath29@xmath25 , which is larger than the experimental @xmath27 .
the spin dynamics , however , can be understood if one assumes that spin degrees of freedom are affected by the orbital degrees of freedom .
in particular , the spatial asymmetry of the peak in the intermediate phase can be understood as the consequence of the fact that at @xmath30 , vanadium octahedra are flattened and @xmath18-orbital at each site is occupied .
this leads to a strong antiferromagnetic exchange between vanadium spins along @xmath18 direction , and , as a result , spin interactions become effectively one - dimensional .
strong one - dimensional spin fluctuations give rise , via spin - orbit coupling , to fluctuations of the occupations of @xmath22 or @xmath20 orbitals , causing the anisotropy of the neutron peak .
there were several theoretical attempts to understand the nature of the ground state of znv@xmath0o@xmath1 .
however , although it is widely accepted that orbital degrees of freedom play an important role , no consensus is reached yet about the type of the orbital ordering ( oo ) in the ground state . the first attempt to explain the physics of znv@xmath0o@xmath1 focused on the spin - lattice coupling mechanism , @xcite however it did nt explain why the structural and spin order occur at different temperatures .
tsunetsugu and motome @xcite later addressed this issue and related the presence of two separate phase transitions at @xmath14 and @xmath10 to the interplay between geometrical frustration and @xmath31 superexchange ( se ) interaction between v - ions .
the ground state orbital ordering suggested in ref .
consists of stacked @xmath32 planes with alternating orbital occupations @xmath33 and @xmath34 .
hereafter we label this orbital patterns as roo .
they also showed that this ordering of orbitals can partially remove magnetic frustration and explain experimentally observed ordered magnetic structure . on the other hand ,
tchernyshyov @xcite pointed out that the ground state obtained in ref .
is at odds with x - ray and neutron diffraction data , because it does not possess the required @xmath13 space symmetry .
he argued that the spin - orbit coupling should be included into consideration .
he considered a purely ionic model in which spin - orbit ( so ) coupling plays the major role and determines the orbital order in the tetragonal phase .
he suggested the following oo : one electron on each site occupies @xmath18 orbital , while the second electron is spreaded between @xmath22 and @xmath20 orbitals in such a way as to minimize the spin - orbit energy at each site .
hereafter we label this orbital pattern as coo .
recently r. valenti _
et al._@xcite found in the ab - initio dft calculations that a correct space symmetry can be actually obtained within the reasoning of ref . , if one includes into consideration an additional trigonal distortion .
still , the true ground state turns out to be the same as in ref . .
the ideas of refs . and were combined in the unique framework by s. di matteo _ et al . _ in ref . .
they proposed to construct a classical ground state phase diagram by considering se interaction and so coupling on equal footing .
they demonstrated that the so coupling is a relevant perturbation and favors the states with unquenched orbital momentum for any value of the coupling strength . they obtained a variety of phases and found that for reasonable values of se and so couplings , the ground state agrees with the experimentally observed one . in this paper
we extend the analysis of ref . and study the low energy excitations in vanadium spinels .
we find that the magnetic excitation spectrum strongly depends on the type of the oo , and that it is qualitatively different for the ground states with patterns that consist of real orbitals and those with the complex linear combination of orbitals , i.e. complex orbitals .
the former ground state is characterized by a quenched orbital angular momentum ( @xmath35 ) , while the later by a unquenched ( @xmath36 ) orbital angular momentum .
@xcite the difference in the magnetic excitation spectrum arises due to the fact that the magnetic moment of the vanadium ion is formed by both spin and orbital momentum , and fluctuations of @xmath37 contribute to the spectrum of magnetic excitations .
we argue that the measurement of magnetic excitations in neutron scattering experiments can shed light on the nature of the oo in the ground state .
this paper is organized as follows : in section ii we introduce the model appropriate for the description of the physical properties of znv@xmath0o@xmath1 . in section iii
we discuss the ground state and the magnetic excitations of the system when the orbital angular momentum is quenched .
we derive linear spin - wave theory of @xmath2 moments interacting on the @xmath38 pyrochlore lattice . in section
iv we discuss ground state and magnetic excitations of the system when the orbital angular momentum is unquenched .
we show that magnetic excitations for unquenched orbital angular momentum can be described in the framework of the magnetic excitonic model .
section v presents the conclusions .
some mathematical details are given in the appendices a and b.
the minimal model describing the low energy physics of vanadium spinel is given by @xmath39 the first term describes nearest neighbors ( @xmath40 ) super - exchange interactions between vanadium @xmath2 spins , arising from the virtual excitations @xmath41 .
these interactions can be written as : @xmath42 } o_{ij } \\ & -\sum_{<ij>}j_2{\big [ } 1-{\bf s}_i\cdot { \bf s}_{j}{\big ] } { \bar o}_{ij}~ , \end{array } \label{spinorb2}\end{aligned}\ ] ] where @xmath43 and @xmath44 are nearest neighbors , @xmath45 $ ] , @xmath46/[1 - 3\eta]$ ] , @xmath47/[1 + 2\eta]$ ] are coupling constants , @xmath48 is the overall energy scale ( @xmath49 and @xmath50 is the intra - orbital coulomb repulsion ) , @xmath51 is the normalized hund s exchange .
we consider only the largest the hopping term , associated with @xmath52-bonding @xcite .
such hopping is diagonal and non - zero only if the orbitals and the plane in which hopping occurs are of the same @xmath53 type ( @xmath54 , @xmath20 , @xmath18 ) . in this case
orbital contributions @xmath55 and @xmath56 are expressed in terms of projectors @xmath57 onto the occupied orbital state @xmath53 at site @xmath43 and @xmath44 : @xmath58 and @xmath59 . to describe the anisotropy and spin - orbit coupling term , we use the fact that , when the crystal field splitting between @xmath4 and @xmath60 orbitals is large , the @xmath4-electrons can be represented by an effective orbital angular momentum @xmath61 @xcite the anisotropy term ( the second term of the hamiltonian ( [ hamil ] ) ) ,
is then given by @xmath62 where @xmath12 is a constant .
this term describes the tetragonal distortion in the @xmath4 manifold .
we notice , that here for simplicity we neglect the trigonal distortion , which is small compared to the tetragonal one .
the spin - orbit coupling term ( the third term in ( [ hamil ] ) ) is given by @xmath63 where @xmath64 is so coupling constant .
note that the true angular momentum @xmath65 is related to an effective one as @xmath66 .
the parameters of the model can be estimated from the experiments .
the spectroscopy data @xcite yield the hund s exchange , @xmath67 ev and coulomb intra - orbital repulsion @xmath68 ev . the estimate of hopping matrix element from x - ray photoemission spectroscopy is @xmath69 ev.@xcite , so the energy scale is @xmath70 mev . since the so coupling constant is @xmath71 mev ( ref . ) , we can see that the super - exchange and the spin - orbit couplings are comparable and , therefore should be treated on equal footing .
first we discuss the ground state of the super - exchange hamiltonian alone ( eq . [ spinorb2 ] ) . without the anisotropy and spin - orbit terms the ground state orbital patterns
consist of only real orbitals : at each site two out of three @xmath4 orbitals @xmath72 , are occupied .
this type of orbital patterns is called real orbital order , as opposed to complex orbital order , when the orbital state is formed by a complex superposition of @xmath4 orbitals in such a way , that the gain of spin orbit interaction energy is maximized . depending on which orbitals are occupied , one obtains two types of interacting bonds and also non - interacting bonds .
consider for example a bond in the @xmath53-plane . if there is an electron occupying @xmath53 orbital only on one site of such bond ,
then the bond , which we label as @xmath73 , is weakly ferromagnetic ( fm ) , and is described by @xmath74 .
if both @xmath43 and @xmath44 sites of @xmath75 bond are occupied by @xmath53 electrons , then the bond , which we label as @xmath76 , is strongly antiferromagnetic ( afm ) .
the exchange coupling is then given by @xmath77 .
when neither @xmath43 nor @xmath44 site have @xmath53-orbital occupied , the bond is non - interacting .
one can easily demonstrate @xcite that for positive @xmath78 and @xmath79 , the lowest energy configuration corresponds to the state with four ferromagnetic @xmath73 bonds per tetrahedron .
there still exist two topologically different tetrahedral configurations with four @xmath80 bonds , characterized by different oo patterns .
one of them , the roo state with oo patterns proposed by tsunetsugu @xcite , is compatible with experimentally observed magnetic structure it yields afm chains running in [ 110 ] and [ 1@xmath810 ] directions ( see , fig.[fig1 ] ) .
the classical energy per site in the roo state is @xmath82 .
however , this state is actually not the true ground state of the super - exchange hamiltonian ( [ spinorb2 ] ) . the other state with 4 @xmath80 bonds ( the roo - i state in our notations ) , in which spins of each tetrahedron form a fully collinear up - up - up - down ( @xmath83 ) state , has the ground state energy @xmath84 , which is lower than that of the roo state .
this is the lowest classical energy that one can obtain within the manifold with quenched angular momentum at each site @xcite .
however , the magnetic ordering associated with the roo - i state is incompatible with the experimentally observed one : in roo - i two neighboring spins in @xmath18-plane are ferromagnetically aligned whereas the experimentally detected coupling in @xmath18-plane is antiferromagnetic .
this discrepancy between the roo - i ground state of the super - exchange hamiltonian ( [ spinorb2 ] ) and the experimental findings demonstrates the necessity of taking into account additional interactions we study the the anisotropy term and the spin - orbit coupling in the section iv . + one can also consider the model with both the super - exchange ( se ) and the jahn - teller ( jt ) couplings @xcite with a hope that the cooperative jt effect , which plays an important role in the structural transition , can also stabilize the roo type of orbital order .
this model has been studied in ref . by both the mean - field analysis and monte carlo simulations . here
we assume that the roo phase can be realized and in the next subsection we derive magnetic excitations spectrum for the corresponding effective spin model .
the starting point for the calculation of the magnon excitation spectrum is the classical neel ground state with antiferromagnetic spin chains in @xmath18-planes ( fig .
[ fig1 ] ) . this ground state has the magnetic unit cell with eight vanadium spins , which we denote as @xmath85 ( fig.[fig1 ] ) . after averaging the orbital operators we can re -
write the hamiltonian ( [ spinorb2 ] ) as @xmath86 where the first term describes the super - exchange along @xmath18-chain with @xmath87 , while the second term corresponds to the frustrated ferromagnetic inter - chain coupling , @xmath88 .
+ to describe the excitation spectrum of such eight - sublattice antiferromagnet , we introduce eight boson operators : @xmath89 .
we employ holstein - primakoff transformation , e.g. for up - spins @xmath90 and down - spins @xmath91 we have : @xmath92 in the linear spin wave approximation we substitute @xmath93 in the expressions above . performing fourier transformation @xmath94 , where @xmath95 is the number of lattice sites belonging to one sublattice
, we obtain the mean field hamiltonian for magnons @xcite @xmath96 where we introduced @xmath97 ) .
the matrices @xmath98 and @xmath99 are eight by eight matrices whose elements depend on the geometry of the lattice and the type of the magnetic ordering , tilde denotes the complex conjugation . the explicit expression for matrix elements
are presented in appendix a. the quadratic form is diagonalized using the generalized bogoliubov transformation . in the diagonal form , the hamiltonian takes the form : @xmath100 where the index @xmath101 runs from 1 to 8 , @xmath102 are magnon energies and @xmath103 are linear combinations of boson operators belonging to @xmath104 and @xmath105 .
we obtained the spin - wave excitation spectrum by solving eq .
( [ hdiag ] ) numerically .
the result is presented in fig .
[ fig2 ] along high - symmetry directions of 3d brillouin zone .
we used @xmath106 mev , and the ferromagnetic constant @xmath107 .
there are 4 different branches of the spin - wave spectrum , each of them is doubly degenerate .
furthermore , two of the branches have zero energy over a finite range of momenta ( see fig .
[ fig2 ] ( solid lines ) ) .
this so called zero modes emerge because the number of ferro- and antiferro - bonds connecting two neighboring antiferromagnetic chains is the same ( see fig .
[ fig1 ] ) , and the spins forming @xmath108chains can collectively rotate with no change in energy .
the existence of the zero modes is inconsistent with the observation of the magnetic ordering transition at @xmath109k .
the ordering requires that the zero mode be lifted .
the issue is what interactions are responsible for the lifting of spin degeneracy .
a natural first step would be to consider longer range interactions @xcite as these interactions generally remove the degeneracy ( it happens , e.g. , in a kagome antiferromagnet ) .
we show below that it is indeed the case , however , the energy of the relevant degeneracy - breaking mode is very small and can not explain the magnetic ordering temperature of @xmath109k . in fig.[fig2a ]
we show the interactions which include up to third neighbors .
there are 6 nearest neighbor interactions @xmath110
( @xmath111 along @xmath18 chain , and @xmath112 along @xmath22 and @xmath20 bonds ) , 12 second neighbor interactions @xmath113 , and 12 third neighbor interactions @xmath114
. quite often already inclusion of the second neighbor exchange lifts the degeneracy .
however , here second - neighbor interactions @xmath113 are frustrated and can not remove the degeneracy , and , therefore , zero modes in the spin - wave spectrum .
@xcite thus , one has to include third neighbor exchanges .
there are two inequivalent sets of third neighbors , one obtained by two nearest neighbor steps @xmath114 and the other through the empty space @xmath115 ( see fig.[fig2a ] ) .
when only @xmath31 hopping is considered , the exchange coupling through the empty space is zero , @xmath116 , and only @xmath114 interactions contribute .
these interactions are antiferromagnetic @xmath117 , and non - zero only if they connect sites along the direction corresponding to the symmetry of the orbital occupation , ( i.e. , for orbital occupation @xmath53 , the interaction is nonzero only along @xmath118 direction ) .
the third neighbor interaction is frustrated along @xmath18-chains , but it is small compared to nearest neighbor exchange along the chain , @xmath119 , and can not destroy antiferromagnetic ordering along the chain . along @xmath22 and @xmath20 directions , @xmath120 are not frustrated , and connect parallel antiferromagnetic chains located in second neighboring @xmath18 planes . the energy scale for @xmath121 is then @xmath122 .
this is a very small energy , only about one percent of the frustrated ferromagnetic inter - chain coupling @xmath123 .
this small interaction can not explain the magnetic ordering temperature of @xmath109 k. @xcite here we consider this interaction only qualitatively and assume the value of coupling constant @xmath124 .
one can see in fig .
[ fig2 ] ( dashed lines ) that the zero energy modes indeed become dispersive for @xmath125 .
we would like to note that experimentally it has been proven that all magnetic moments are aligned along the @xmath126-direction .
this experimental fact can not be explained in the framework of the se model , because it is isotropic in a spin space . in reality ,
vanadium spinels likely possess a single ion spin anisotropy which aligns spin along the @xmath126-direction .
the anisotropy does affect the magnetic excitation spectrum which in its presence acquires a gap . in this case , the zero mode will be lifted and the spin frustration will be removed . however , since the strength of magnetic anisotropy is not known experimentally at the moment , we can not estimate the magnitude of the anisotropy - induced gap and check whether or not this interaction alone can stabilize the ground state magnetic structure .
we believe that the role of the anisotropy deserves further experimental and theoretical investigation .
we now consider the ground state of the system in the presence of the so coupling and the anisotropy term ( eqs .
( [ anis ] ) and ( [ so ] ) ) . for any finite @xmath64
the so coupling prefers the orbital state with the unquenched effective orbital angular momentum @xmath127 . in such orbital state ,
one electron at each site occupies @xmath18 orbital due to the tetragonal distortion , while the second electron occupies the complex linear combination of @xmath22 and @xmath20 orbitals .
the effective @xmath128 should then be parallel to the spin magnetic moment in order to minimize the spin - orbit energy , i.e. a spin - up site will have @xmath129 while a spin - down site will have @xmath130 . as we discussed earlier , this type of orbital ordering is a coo state suggested first by tchernyshyov in ref . .
its energy @xmath131-\lambda$ ] is lower than the energy of the roo state for a wide range of parameters ( see the phase diagram in ref . ) .
the coo state is characterized by two strong afm bonds per tetrahedra , and its magnetic structure consists of afm chains in @xmath18-planes with the same interaction along the chain as in the roo state .
the strengths of the inter - chain coupling @xmath132 are also practically the same for the roo and the coo states @xcite , although @xmath133 $ ] is antiferromagnetic in the coo state , while it is ferromagnetic in the roo state . in this work
we assume @xmath134 in both orbital states . in the coo state
the preferred spin direction is fixed by anisotropy term ( [ anis ] ) via the so coupling to be along the @xmath126-axis , and , therefore , spins in @xmath18-chains can not rotate freely even for only nearest - neighbor exchange along the chains .
the long range order + + - - ... along diagonal directions in @xmath22 and @xmath20 planes can not be determined by local interactions , but for simplicity we do not include next neighbor hopping terms in this part of our calculations .
we now consider magnetic excitations in the coo state .
we follow the magnetic exciton model approach of refs . , which is the extension of the linear spin wave theory for systems with unquenched orbital angular moment .
we consider states with the effective total angular momentum @xmath135 .
often , energy levels with different @xmath136 are well separated in energy and both @xmath136 and its @xmath126-projection @xmath137 are good quantum numbers . however , in many transition metal oxides the strength of the spin - orbit coupling and the super - exchange interaction between localized @xmath138-electrons are comparable , and atomic energy levels with different values of @xmath136 can cross each other ; in such case only @xmath137 acts as a good quantum number .
we show below that magnetic excitations in vanadium spinels av@xmath0o@xmath1 can be understood as a propagation of excitations to states with a given @xmath137 through the crystal . to proceed , we rewrite the hamiltonian ( [ hamil ] ) as a sum of a single - ion hamiltonian @xmath139 and the term which describes the interaction between two different ions @xmath140 : @xmath141 where @xmath142 the molecular field part of the exchange interaction acting on site @xmath43 is given by @xmath143 .
@xmath144 is the number of @xmath145-th neighbors , @xmath146 is the corresponding exchange constant , and @xmath147 is the sublattice magnetization .
first , we diagonalize the single ion hamiltonian , @xmath139 , in the molecular field approximation .
it is convenient to express the eigenfunction @xmath148 for the states splitted by the spin - orbit interaction as linear combinations of the unperturbed eigenfunctions of @xmath149 and @xmath150 .
then @xmath139 can be represented as a block @xmath151 matrix in the subspace of @xmath152 as follows : @xmath153 @xmath154 @xmath155 where @xmath156 , @xmath157 , @xmath158 , @xmath159 and @xmath160 are represented in the basis @xmath161 , @xmath162 , ( @xmath163 , @xmath164 ) , ( @xmath165 , @xmath166 ) , and ( @xmath167 , @xmath168 , @xmath169 ) , respectively .
diagonalizing these matrices , we obtain eigenvalues and eigenvectors of @xmath139 as functions of the molecular field @xmath170 .
when the tetrahedra are flattened as observed in the experiments , i.e. when @xmath171 , for any strength of the molecular field @xmath170 the ground state of @xmath139 has @xmath172 ( the sign of @xmath137 depends on the sign of @xmath170 ) . for definiteness , we consider spin - down sublattice ( @xmath173 ) .
the resulting energy levels are presented in the fig.[fig3 ] .
the magnetic excitations of the local hamiltonian @xmath139 are transitions from the ground state to eight excited states of a single ion .
these excitations can be described by boson operators @xmath174 each pseudoboson describes the transition from the ground state @xmath175 to the excited state @xmath176 .
the single ion hamiltonian @xmath139 is diagonal in terms of pseudoboson operators : @xmath177 where @xmath178 is the energy difference between the excited state @xmath179 and the ground state of @xmath139 : @xmath180 we next introduce the representation for spin operators in terms of pseudobosons @xmath174 . the representation for spin @xmath181 on the sublattice @xmath182 can be written as : @xmath183 @xmath184 if the ground state of @xmath139 corresponds to @xmath157 , it follows from ( [ spmgeneral ] ) that only states with @xmath159 contribute to @xmath185 .
there are two such states , @xmath186 and @xmath187 . in the basis of @xmath188
they have the following structure : @xmath189 where we denote @xmath190 as the weight of the state with @xmath191 , and @xmath192 in the @xmath186 state .
the partial weight of orbital and spin contributions is determined by the competition between the spin - orbit coupling , the anisotropy energy and the molecular exchange field . in fig .
[ fig3a ] we present a dependence of @xmath190 on the molecular exchange field @xmath170 at different values of the tetragonal field @xmath12 , keeping the so coupling constant @xmath64 fixed . for all values of tetragonal field parameter @xmath190
shows similar field dependence : it increases with the increase of @xmath170 .
this happens because the molecular field @xmath170 , acting on the spin @xmath193 , is effectively created only by the nearest neighbors on the @xmath18-chain , which are antiferromagnetically aligned to @xmath193 .
therefore , when @xmath170 increases , the flipping of the spin costs more energy and the transitions with @xmath194 are suppressed .
the low lying excitations become more of an orbital character , which can be seen in the increase of the weight of the transitions with @xmath195 . as only two excited states of an ion are relevant to the spin - wave analysis , we consider the local excitations described by two pseudoboson operators @xmath196 and @xmath197 , which take an ion from the ground state to states @xmath186 and @xmath187 , respectively .
the explicit form of extended holstein - primakoff transformation for spin @xmath198 in terms @xmath199 and @xmath200 is given by : @xmath201 @xmath202 it is also useful to rewrite these expressions , ( [ szext ] ) and ( [ spmext ] ) , using the definition of @xmath190 ( for spin - down ) : @xmath203 the interactions between localized excitations are described by @xmath140 .
the excitation spectrum is obtained in a similar way as in the spin wave analysis for the roo state .
after diagonalization , the total hamiltonian @xmath204 can be written in the same form as eqs.([hdiag ] ) , however now the index @xmath101 runs from 1 to 16 and the modes have complex spin orbital character .
the details of calculation are given in appendix b. the numerically calculated magnetic excitation spectrum for the coo state is presented in fig.[fig4 ] . for comparison , we also plotted there the spin - wave dispersion for the roo state .
we used the following parameters : @xmath106 mev , @xmath205 , @xmath206 mev , and @xmath207 mev ( third neighbors interactions are not included here ) . as we see from fig.[fig4 ] , the excitation spectrum consists of 8 different branches , each of them is doubly degenerate . the excitations may be divided into two groups : four low lying branches with rather small gap @xmath208 mev , and four optical branches , with the gap @xmath209 mev . as @xmath210 , the lying branches are quasi - acoustic modes . as we see in fig.[fig4 ] , the lowest mode along the direction @xmath211 is the dispersionless mode .
this flat mode is a lifted `` zero energy mode '' of the roo state . here
the zero mode is lifted due to the combined effect of the anisotropy term and the so coupling .
the tetragonal distortion favors the non - zero value of the @xmath126-component of orbital angular momentum , which in turn , selects the local spin quantization axis in such a way , that the @xmath126 component of spin is non - zero .
we found that the gap , @xmath212 is mainly determined by the anisotropy term and only weakly depends on the so coupling ( see fig.[fig4a ] ) .
however , we note once again , that , in principle , the lifting of zero energy mode can appear also in roo state if single ion magnetic anisotropy is taken into account .
we caution that the size of these two different anisotropy gaps can be of the same order , and therefore , it will be rather difficult to distinguish between them . as a remark
, we would like to mention , that such lifted zero energy modes can be detected by an inelastic neutron scattering , and have been indeed recently observed in the frustrated kagome lattice antiferromagnet kfe@xmath213(oh)@xmath11(so@xmath1)@xmath0 @xcite .
there is another significant difference between the magnetic excitation spectrum for coo and roo states : as one can see from fig.[fig4 ] , the bandwidth of quasi - acoustic modes in the coo state is strongly reduced compared to the bandwidth of pure spin waves in the roo state .
this reduction is the effect of the mixing between orbital and pure spin excitations in the spectra of the coo state .
pure orbital excitations are non - dispersive , because they come form local interactions .
the presence of the orbital component in the spin - wave spectrum then obviously leads to the reduction of the bandwidth .
we believe that this large reduction of the bandwidth can be seen experimentally , in the neutron scattering measurements of the magnetic excitation spectrum .
such experimental results would discriminate between two types of orbital ordering , coo and roo .
unfortunately , at present these experiments are difficult to perform because to measure the full spin - wave spectrum one needs single crystals , which are still not available , as far as we know .
let us now discuss the optical branches .
these modes exist only in the coo state , and they arise from the hybridization between orbital angular momentum non - dispersive levels and dispersive spin branches due to the spin - orbit coupling . in fig.[fig3a ]
we show how the optical gap , @xmath214 , depends on the strength of the spin - orbit coupling .
one can see that at @xmath215 , @xmath214 is zero , but it rapidly increases with increasing @xmath64 . to summarize , the magnetic excitation spectrum for the coo state has two separated branches : quasi acoustic modes and optical modes .
both manifolds are gapped , but the gaps , @xmath212 and @xmath214 have different origins : @xmath212 is set by the anisotropy term , while @xmath214 is set by the spin - orbit coupling and is much larger .
these two gaps could be , in principle , determined by inelastic neutron or raman scattering even in powder samples of znv@xmath0o@xmath1 .
they are also easily distinguishable from the well - known haldane gap , @xmath216 , which is a characteristic feature of the antiferromagnetic @xmath2 chains .
@xmath216 , @xmath212 and @xmath214 have different temperature dependence : the optical gap , @xmath214 , does not depend on the temperature , as it is determined by the relativistic spin - orbit interaction ; the anisotropy gap , @xmath212 , is non - zero only below the temperature of the structural transition , @xmath217 ; and the haldane gap should disappear below @xmath14 , when a long range antiferromagnetic ordering emerges . at all temperatures ,
the optical gap , @xmath214 , is the largest , so we compare only the anisotropy and the haldane gaps . at the lowest temperatures , @xmath28 , the haldane gap vanishes , and the lowest mode will have a gap equal to an anisotropy gap @xmath212 . at intermediate temperatures , @xmath26 ,
the spin excitations are also gapped .
in this temperature range , the gap is the sum of the anisotropy gap , @xmath212 , and the haldane gap , @xmath216 .
the magnitude of the well - developed haldane gap is of the order of @xmath218 , and is compared to @xmath10 , hence it is very likely that the gap at @xmath26 will be larger then at low temperatures .
this behavior is exotic and , as far as we know , has not been yet observed in @xmath2 spin - chain systems . at high temperatures , @xmath219 ,
the sharp gap is washed out by thermal fluctuations , but the spectrum can still divided into acoustic and optical branches .
we presented in this paper a detail analysis of the magnetic excitations of vanadium spinels , whose low - temperature tetragonal phase can be modeled to a high accuracy by one - dimensional spin chains with weak inter - chain interaction . the formation of antiferromagnetic spin chains on the highly frustrated pyrochlore lattice is by itself non - trivial phenomena .
this can happen only because vanadium ions also possess an orbital degree of freedom , and the orbital modulation of the spin exchange partially lifts the geometrical degeneracy of the underlying lattice .
we considered two different ground states : _
i ) _ the one with the real orbital ordering , roo , and _
ii ) _ the one with the complex orbital ordering , coo .
we found that the excitation spectra in these two cases are qualitatively different .
the spectrum for the coo state consists of low lying quasi - acoustic modes with small anisotropy gap and optical branches with zone - center gap determined by the spin - orbit coupling .
the spectrum for the roo state has only quasi acoustic modes . within the superesxchange model considered in
the present study the spectrum is gapless , however , in reality we expect an anisotropy gap also for this state .
the bandwidth of the quasi - acoustic modes in the coo state is strongly reduced compared to the ones for the roo state , due to contributions from orbital @xmath220-modes .
because the spectra are so different , we argue that an effective way to determine experimentally the symmetry of orbital ordering in vanadium spinels is to measure their magnetic excitation spectrum . in the exciton model
, the base object is an enlarged set of pseudoboson operators @xmath226 , whose components @xmath227 and @xmath228 describe transitions to first and second excited levels .
these two vectors are analogous to @xmath229 for pure spin wave model . as a result ,
the matrix @xmath98 is enlarged and becomes a @xmath230 matrix @xmath231 .
its components are given by @xmath232 where @xmath233 here @xmath182 and @xmath234 runs from 1 to 8 .
the diagonal matrix elements of @xmath235 are zero ( we consider only the case @xmath236 ) , because the diagonal contribution is already included in @xmath178 . y. motome and h. tsunetsugu , prog .
. suppl . * 160 * , 203 ( 2005 ) .
we would like to note that the coupling between third neighbors arise not only due to the direct exchange , but also due to the various super - exchange paths through oxygen ions . in some pyrochlore compounds
this contribution can be rather significant , however , here , for simplicity , we did not take this coupling into account . | we study magnetic excitations in vanadium spinel oxides av@xmath0o@xmath1 ( a = zn , mg , cd ) using two models : first one is a superexchange model for vanadium @xmath2 spins , second one includes in addition spin - orbit coupling , and crystal anisotropy .
we show that the experimentally observed magnetic ordering can be obtained in both models , however the orbital ordering is different with and without spin - orbit coupling and crystal anisotropy . we demonstrate that this difference strongly affects the spin - wave excitation spectrum above the magnetically ordered state , and argue that the neutron measurement of such dispersion is a way to distinguish between the two possible orbital orderings in av@xmath0o@xmath1 . |
cosmic - rays ( crs ) are highly energetic particles ( with kinetic energies up to @xmath0 ev ) mainly composed by protons and nuclei with a small percentage of electrons ( 1@xmath1 ) .
since from their discovery cosmic - rays are one of the most debated issues of the high energy astrophysics . their origin is still a fundamental problem and is the subject of very intense research @xcite , ( for recent reviews , see * ? ? ?
* ; * ? ? ?
focusing on crs produced in our galaxy ( energies up to the `` knee '' , @xmath2 ev ) , strong shocks in supernova remnants ( snrs ) are considered the most probable cr sources ( e.g. , * ? ? ?
however , the final proof for the origin of crs up to the knee can only be obtained through two fundamental signatures .
the first one is the detection of a clear gamma - ray signature of @xmath3 decay in galactic sources ; the second one is the identification of sources emitting a photon spectrum up to pev energies .
both indications are quite difficult to obtain .
the `` pevatron '' sources are notoriously difficult to find ( for a review , see * ? ? ?
* ) , and the neutral pion decay signature is not easy to identify because of the possible contribution from co - spatial leptonic emission .
hadronic ( expected to produce the @xmath3 decay spectral signature ) and leptonic components can in principle be distinguished in the 50 - 200 mev energy band , where they are expected to show different behaviors .
+ over the last five years , agile and fermi gamma - ray satellites , together with ground telescopes operating in the tev energy range ( hess , veritas and magic ) , collected a great amount of data from snrs @xcite providing important information and challenging theoretical models .
most of the snrs detected in the @xmath4-ray band are interacting with a mc .
the presence of a high density target , indeed , enhances the possibility to detect @xmath4-ray emission from pp - interaction .
interestingly , most of the observed snrs show , apparently , a spectrum steeper than the one expected from linear and non - linear diffusive shock acceleration models ( dsa ) of index near 2 , and possible convex spectrum @xcite .
+ w44 is one of the most interesting snrs observed so far .
it is a middle - aged snr , which is bright at gamma - ray energies and quite close to us .
its gamma - ray spectral index is @xmath5 @xcite , in apparent contradiction with dsa models .
its environment is very interesting and complex , requiring a careful re - evaluation of theoretical models .
recently , an analysis of fermi - lat data confirmed these results @xcite . our new analysis of the agile data ,
comparing with the new fermi - lat data and considering also radio and new co data from vla and nanten2 , confirms that gamma - ray data can be fitted only with a hadronic model and that the spectral index is @xmath6 .
moreover , even if there is an ambiguity on the choice of a specific hadronic model , we obtain some important constraints on fundamental parameters , such as the magnetic field and the ism density @xcite .
thanks to a deep analysis of available multiwavelength data , we collected multiwavelength information for young and middle - aged snrs in order to compare their spectral behaviors with the w44 one and , in general , with the main theoretical model expectations .
+ in spite of the great amount of data from snrs emission , the understanding of cr acceleration and propagation processes is all but complete .
the snr w44 is a middle - aged ( @xmath720,000 yrs old ) snr located in the galactic plane ( @xmath8)= ( @xmath9 at a distance @xmath10 kpc @xcite .
multiwavelength observations revealed interesting features . in the radio band , w44 shows a quasi - elliptical shell ( * ? ? ? * and references therein ) ; the radio shell asymmetry is probably due to expansion in an inhomogeneous ism . in the northwest side of the remnant , which correlates with a peak of the radio emission
, there is bright [ sii ] emission characteristic of shock - excited radiative filaments @xcite . in the southeast side , instead , there is a molecular cloud ( mc ) complex embedded in the snr shell that interacts with the source @xcite .
the oh maser ( 1720 mhz ) emission detected in correspondence with the snr / mc region , confirm their interaction @xcite .
in @xcite the discovery of the radio pulsar psr b1853 + 01 is reported , which is located in the south part of the remnant and surrounded by a cometary - shaped pulsar wind nebula ( pwn ) @xcite .
this system , however , does not appear to be correlated with the detected gamma - ray emission .
the x - ray observations of w44 by the einstein observatory @xcite showed centrally peaked emission , which is later confirmed by chandra data @xcite .
+ the snr w44 is well studied also in the gamma - ray band . in @xcite ,
a gev morphology well correlated with the radio emission is showed , together with a steep photon spectrum ( index near 3 ) that , however , has a low - energy threshold of 200 mev , limiting the chance to identify a neutral pion signature .
the relatively large gamma - ray brightness of w44 and the good spectral capability of agile near 100 mev @xcite have stimulated a thorough investigation of this supernova remnant with the agile data .
the agile gamma - ray spectrum in the range of 50 mev to 10 gev confirms the high - energy steep slope up to 10 gev and , remarkably , identifies a spectral decrease below 200 mev for the first time , as expected from neutral pion decay ( * ? ? ? * hereafter g11 ) . in the analysis of g11 ,
both leptonic and hadronic models were considered in fitting both agile and fermi - lat data .
the best model was determined to be dominated by hadronic emission with a proton distribution of spectral index @xmath11 and a low - energy cut - off at @xmath12 gev .
the low - energy spectral behavior seen by agile was recently confirmed by the fermi - lat team that revisited the gamma - ray emission from w44 ( * ? ? ?
* hereafter a13 ) .
their best hadronic model with an _ assumed _ surrounding medium density @xmath13 @xmath14 is based on a smoothed broken power - law hadronic distribution with a break energy @xmath15 gev and indices @xmath16 for @xmath17 and @xmath18 for @xmath19 .
model parameters in a13 differ from those considered earlier in @xcite .
apparently , bremsstrahlung emission is not considered to be relevant in the hadronic modeling of a13 , even though this process could provide a non - negligible contribution to the gamma - ray emissivity in principle .
+ we present here a new analysis of agile data with a revised assessment of the w44 surrounding environment , which is based on new co data obtained from the nanten2 telescope @xcite .
+ the very important feature of the snr w44 spectrum , confirmed in every analysis , is its slope at gev energies : the index @xmath5 is substantially steeper than the range that is plausibly expected in linear and non - linear dsa models . in @xcite , this spectral feature
is explained by alfvn damping in the presence of a relatively large - density medium where acceleration occurs .
the w44 environment is quite challenging in its morphology and requires a reanalysis of its properties in the context of the crucial implications for the acceleration mechanism of crs .
we performed a global reassessment of the agile data on w44 , including new gamma - ray data obtained until june 2012 @xcite .
[ maps ] shows the w44 agile co maps in two velocity channels , 41 and 43 km / s , with radio ( vla ) and gamma - ray ( agile , 400 - 10000 mev ) .
gamma - ray emission appears to be mostly concentrated near a high - density region , the co peak at ( 34.7,-0.5 ) , indicating that most of the w44 gamma - ray emission is coincident with a site of snr / mc interaction ( for a more detailed description and image , see * ? ? ?
figure [ spectra ] shows the agile gamma - ray spectrum with the recently updated fermi - lat data from @xcite .
the agile spectrum is composed by six energy bins between 50 mev and 10 gev and our error - bars takes statistical errors into account . the measured flux of the source above 400 mev is @xmath20 ph @xmath21 s@xmath22 .
we notice the good agreement between the two spectra . especially important
is the confirmation of the drastic spectral decrement below 200 mev , a crucial feature that is discussed below .
both agile and fermi - lat spectra differ from the previously published spectra in g11 and @xcite , @xcite .
we model the radio , agile , and fermi - lat spectral data by hadronic and leptonic - only scenarios by considering the new nanten2 co data that provides a value for the ism density in the snr surroundings , @xmath23 @xcite .
this value of the average gaseous density that surrounds the gamma - ray emission is substantially larger than the one assumed in g11 and a13 ( @xmath24 ) .
since the agile gamma - ray emission is strongly correlated with one of the co peaks , we consider an average density @xmath25 @xmath14 @xmath26 in the following . in modeling the spectra
, we consider the most statistically significant fermi - lat data up to 50 gev .
we assume that the gamma - ray emission spectrum is due to the combined contribution of hadronic @xmath3 emission and leptonic bremsstrahlung emission by considering the proton component as the main one . for hadronic emission
, we use the formalism explained in @xcite that is a good approximation of the exact solution .
we consider a proton distribution in total energy @xmath27 rather than in kinetic energy @xmath28 , following @xcite and @xcite , but with @xmath29-function approximation for the cross section @xcite .. we fit the gamma - ray data by assuming different types of proton distributions in energy : a simple power - law with a high - energy cut - off ( * model h1 * ) , a smoothed broken power - law ( * model h2 * ) and a broken power - law ( * model h3 * ) @xcite . for leptons , we used a simple power - law with a high energy cut - off in all hadronic models .
we fix only the parameters for which we have solid observational evidence : the average medium density , @xmath30 , and the radio spectral index , @xmath31 ; all the other parameters , such as the normalization constants , @xmath32 and @xmath33 , and the cut - off and break energies , @xmath34 and @xmath35 , are free .
our best hadronic model is the model h3 with the following proton distribution : @xmath36 this is characterized by an index @xmath37 ( for @xmath17 ) , @xmath38 ( for @xmath39 ) , and an energy break @xmath40 gev . the leptonic contribution to this model
is given by a simple power - law for the electrons with @xmath31 , and @xmath41 gev ( see fig . [ hadronic ] ) .
this model provides a proton energy @xmath42 erg and requires an average magnetic field in the emission region , @xmath43 @xmath44 g .
+ the most important physical characteristics of the snr w44 can be summarized in the following way @xcite : * * neutral pion signature : * w44 is the first snr clearly showing the so - called `` pion bump '' that we expect at @xmath45 mev from @xmath3-decay photons .
the low - energy spectral index , @xmath46 could be affected by the fact that we estimate it in the energy space . in order to confirm ( or not ) the low - energy behavior we need to consider the snr w44 system in the momentum space ; * * high density of the surrounding environment : * we determined that the average density in the w44 shell is @xmath47 with @xmath48 @xmath14 , which corresponds with co peaks ( see medium panels in fig . [ maps ] ) .
this feature was also found in other middle - aged snrs , like w51c and ic443 @xcite and explains the high gamma - ray flux detected from these sources . *
* high magnetic field : * in w44 our best hadronic models imply a magnetic field @xmath49 @xmath44 g , which is lower than the post - shock magnetic field estimated from zeeman splitting in the oh masers by @xcite , and substantially higher than the equipartition magnetic field @xcite . in most of snrs , magnetic field estimations give values @xmath50
@xmath44 g that are much higher than the average diffuse galactic value [ e.g. , see @xcite for tycho , @xcite for w51c , and @xcite for ic443 ] .
this is hardly surprising since magnetic field compression due to the shock interaction with the ism leads to its amplification .
we need to then consider a non - linear scenario with a back - reaction of the accelerated particle at the shock @xcite.the large value for the magnetic field in w44 may be linked to the environment density value , @xmath51 given by nanten2 . for a lower density value , we notice that we can enhance the electron density and make plausible a lower magnetic field @xcite . * * steepness of the high energy index : * as in @xcite , g11 , and a13 , w44 shows a spectral index @xmath52 for energies above 1 gev , that is steeper than the values found in other middle - aged snrs .
alfvn damping in a dense environment @xcite is a mechanism for explaining this behavior , but other possibilities exist ( e.g. , * ? ? ?
* ; * ? ? ?
this is a point requiring deeper investigations in the future .
[ table_w44w28 ] [ cols="<,^,^,^,^,^,^,^,^ " , ] at the light of the snr w44 characteristics and of the snr / cr context , we want to present a comparison between this very important source and another important snr detected in the gamma - ray band , both at gev and tev energies , snr w28 @xcite .
even w28 is a middle - aged mixed morphology snr with dimensions of the shell radio very similar to the w44 ones . in spite of this , from table [ table_w44w28 ] and from fig .
[ fig_w44w28 ] , we can see that these two remnants have some very different features that lead to a different interpretation of their gamma - ray spectrum .
+ first of all , from the upper panel of fig
. [ fig_w44w28 ] , we can see that w44 gamma - ray emission has a very good correlation with its radio shell , differently by w28 where no correspondence there is between gamma - ray emission and radio shell
. observing gamma - ray / co emission correlation ( bottom panel of fig .
[ fig_w44w28 ] ) , in w44 the mc seems to be embedded in the remnant ; in w28 , instead , two different mcs are perfectly correlated with gev and tev peaks .
this is confirmed also by estimated average densities of the two snr shells ( see table [ table_w44w28 ] ) ; very small for w28 and of order of @xmath53 for w44 . moreover ,
no tev emission was detected from w44 , differently from w28 .
this could be explained by the fact that in the case of w44 , gamma - ray emission comes from both mc and the snr shell ; tev particles could be are already escaped by the remnant .
w28 is older than w44 and this implies that the most part of gev and tev cr particles are escaped from the remnant .
consequently , no emission is detected correlated with the shell but only with the two mcs @xcite .
+ this difference is fundamental for their spectral behavior interpretation .
we have seen that for w44 a simple linear dsa model fails because high energy spectral - index is @xmath54 when linear dsa model provides an index of @xmath55 .
for this reason , we need to consider all possible non - linear mechanisms that could explain the steepening of the spectrum . in the case of w28 , instead
, a simple linear dsa model can easily explain its spectral behavior . in @xcite
, we find that where is the peak of gev emission there is a minimum in the tev emission , and viceversa .
this behavior reflects also in the spectrum ; we can divided it into two components .
one from the east cloud , where there is the gev maximum , and the other from the west cloud complex , where is the tev maximum @xcite . from @xcite , considering the two clouds at different distances , we can explain gamma - ray data from w28 with the simple energy dependent diffusion , with @xmath56 and @xmath57 @xmath58s@xmath22 ( bohm diffusion regime ) .
+ in light of these considerations , it is clear that even in w28 the effect of some non - linear mechanisms affect the system ; consequently , considering only the diffusion mechanism in order to understand its characteristic seems a oversimplification .
however , the low average density of the snr shell , and the absence of mcs embedded within it exclude most of the non - linear mechanisms considered so far ( see section [ theoretical challenges ] ) .
now we extend the previous comparison at the most important young and middle - aged snrs emitting in the gamma - ray band .
the importance of young snrs is related to two fundamental issues .
first , their spectra are not affected by propagation effect as in the case of middle - aged snrs . consequently , analysis of their spectral behavior allows to understand injection spectral index and physical processes that could have an impact on acceleration mechanism . on the other hand ,
their young age enhances the chance to detect emission from particles with energies up to @xmath59 ev , called pevatrons , one of the direct signature for cr acceleration .
+ gamma - ray data collected by satellites and instruments , however , put us in front of a very challenging reality .
[ fig_snrspectra ] shows gamma - ray emission from cas a and tycho at gev and tev energies .
+ the first problem is that young snr fluxes are quite faint in the gamma - ray energy band .
this leads to the detection of a very low number of these objects ; consequently , it is very difficult have some reliable conclusions about their general behavior . from theoretical considerations ,
young snrs should to have a low energy flux at gev energies , but a high energy flux in the tev energy range .
even if the faintness of the detected gamma - ray flux is strongly related to the snr distance , there is another parameter that can affect it intrinsically ; this is the density value in the snr surroundings . both cas a and tycho snrs expand into a low average density medium @xcite , as well as the other young snrs , even if in all cases are detected some mcs .
if , on one side , this low average density can explain a low gamma - ray flux , on the other side , it is not so easy to explain the gamma - ray emission from these snrs with a ( most favorable ) hadronic model .
an explanation for this behavior is given in @xcite ; where the progenitor wind residual is considered as the cr target .
another explanation could be linked to proton trapping ; in a bohm - like diffusion regime the proton escape time is much greater than the time needed to cross the system lengths .
+ however , the most important challenge derived from experimental data is their spectral index , @xmath60-@xmath61 @xcite .
emission at the `` knee '' energies is expected by young snrs because their emission is not affected by propagation effects .
however , no sources were observed at @xmath62 tev .
consequently , steepening of young snr spectrum implies no detection of pevatrons ; in the case of cas a and rx j1713 - 3946 , the presence of high - energy cut - off enhances the system complexity .
+ differently by young snrs , middle - aged snrs have spectra that are influenced by propagation effects .
when we analyze their gamma - ray emission , we have to consider the modification of injection spectral index due to the diffusion . clearly , it is more difficult to detect the first phases of the acceleration process and we expect to be difficult to detect pevatrons because high energy particles are already escaped from these remnants .
detection at tev energies is possible only in the presence of a target not embedded in the remnant , like in the snr w28 .
+ in fig .
[ fig_snrspectra ] ( right panel ) , we collected all available gev and tev data of most of the middle - aged snrs studied so far
. the majority of these snrs have a gev flux quite high and are easily detectable .
the lower flux of w49b and w51c is probably due to their great distances from us with respect to the other remnants ) . in all cases we detect mc complexes interacting with remnants and magnetic fields
have large values .
+ puppis a seems to be a unique case .
it has a low magnetic field and no mc are detected in their surroundings @xcite .
its gev flux is low , even if its distance is not so large ( 2 kpc ) , probably due to the absence of a dense target .
moreover , its spectrum has an index @xmath63 , harder than all the other middle - aged snrs and also than young snrs . +
the middle - aged snrs gamma - ray spectral indices are in a range @xmath64 , and , in the oldest snrs , radio spectral index is harder than the one expected from modifications due to shock waves @xcite .
this implies an electron index , not only harder than @xmath65 , but also different from the proton one .
all these snrs seem to have similar surroundings and similar characteristics , such as high magnetic field , presence of mcs and so on . in spite of these facts ,
their gamma - ray spectral indices show that different physical mechanisms are at work .
the comparison between w44 and w28 was an example ( section [ w44 and w28 : comparison ] ) . + in order to explain snr spectral behavior , different from the theoretical expectations , several physical processes are considered so far : * * neutrals `` return flux '' * : there is a suppression of the mach number and compression ratio due to the formation of a shock precursor @xcite ; * * scattering center velocity * : there is the formation of a cr - induced precursor .
low - energy particles feel the lower compression factor at the subshock and the spectrum becomes steeper .
this effect disappears at @xmath66 few gev because high energy particles feel the whole precursor @xcite ; * * alfvn damping * : due to the presence of neutrals in the snr surroundings , it leads to a suppression of scattering center velocity in a certain energy range and , consequently , of the acceleration efficiency @xcite .
all these processes have solid physical reasons but most of them depend on poorly known parameters that , often , are considered separately . in a complex system like a snr
, we should consider all possible physical processes together , and their mutual interaction . only in this way we can have a correct picture of the system .
the snr w44 is a crucial source providing important information about the cr origin in our galaxy .
however , several characteristics of this snr , which have been deduced by a multifrequency approach ( gamma - ray spectral indices , large magnetic field ) , are challenging .
w44 is a relatively close and quite bright gamma - ray source .
therefore , an excellent characterization of its gamma - ray spectrum in the range 50 - 200 mev has been possible because of the good statistics achieved by agile and fermi - lat .
a re - analysis of the agile data from new and updated archives , revisiting radio and co data of w44 , shows the unlikeliness of leptonic - only models in their most natural form : electron distributions constrained by radio data , can not fit the broad - band w44 spectrum inside a 1-zone model . on the other hand , we find that both gamma - ray and radio data can be successfully modeled by different kinds of hadronic models ( h1 , h2 , and h3 ) .
+ the best one is a broken power - law with @xmath67 for @xmath17 and @xmath68 for @xmath39 . our results regarding the spectral properties of the accelerated proton / ion population by the w44 shock qualitatively agree with the results of @xcite .
source morphology resulted to be different from the previous one .
consequently , the interpretation of its surroundings and of the origin of the gamma - ray emission is different .
however , we confirmed the interaction between the remnant and a mc , probably embedded in the source ; most of the gamma - ray emission originates from this region .
this can be explained by the age of the snr .
w44 is a middle - aged snr and most of the high - energy particle diffused far away from the source and this is also the reason of the non detection of tev emission . in the mc
, high density reduces the diffusion and lower - energy particles are trapped inside it , emitting gev @xmath4-ray from @xmath3 decay that we can detect yet . independently from the hadronic model used , one big issue generated by fermi - lat and agile modeling is a photon spectral index , @xmath69 , steeper than the one provided by theoretical models and the steepest between all the other snr indices .
moreover , we find a high value of the magnetic field ( of order of @xmath53 @xmath44 g ) that appears to be strictly correlated with the surrounding medium density .
we see this feature in every snr known so far , in spite of different surroundings or spectral behavior .
placing the case of snr w44 in the general snr / cr context , with a direct comparison with the other snrs , in a direct way with snr w28 , we stress the necessity of a deeper knowledge of each snr and its surroundings .
259 fermi , e. , phys .
rev . 75 ( 1949 ) 1169 - 1174 ginzburg , v.l .
@xmath70 syrovatskii , s.i . ,
_ the origin of cosmic rays , pergamon _ , oxford , 1964 berezinskii , v.s . , bulanov , s.v . , dogiel , v.a .
_ , _ astrophysics of cosmic rays _ , north - holland , amsterdam , 1990 helder , e. a. , vink , j. , bykov , a.m. _ et al . _ 2012 , space science reviews , 173 ( 2012 ) 369 - 431 aharonian , f. 2013 , astroparticle physics , 43 , 71 - 80 vink , j. 2012 , a@xmath70a rev .
, 20 , 49 abdo , a.a . ,
ackermann , m. , ajello , m. , _ et al . _
2009 , apj , 706 , l1-l6 abdo , a.a . ,
ackermann , m. , ajello , m. , _ et al . _
2010 , apj , 710 , 92 - 97 abdo , a.a . ,
ackermann , m. , ajello , m. , _ et al . _
2010 , apj , 712 , 459 - 468 abdo , a.a . ,
ackermann , m. , ajello , m. , _ et al . _
2010 , aj , 722 , 1303 - 13011 abdo , a.a . , ackermann , m. , ajello , m. , _ et al . _
2010 , science , 327 , 1103 - 1106 abdo , a.a . ,
ackermann , m. , ajello , m. , _ et al . _ 2010 ,
apj , 718 , 348 - 356 abdo , a.a . , ackermann , m. , ajello , m. , _ et al . _
2011 , apj , 734 , 28 - 36 acciari , v.a . ,
aliu , e. , arlen , t. , _ et al . _ 2009 ,
apj , 698 , 133 - 137 tavani , m. , giuliani , a. , chen , a.w . , _
et al . _
2010 , apj , 710 , 151 - 155 acciari , v.a . ,
aliu , e. , arlen , t. , _ et al . _ 2010 ,
apj , 714 , 163 - 169 acciari , v.a . , aliu , e. , arlen , t. , _ et al . _ 2010 ,
apj , 730l , 720 - 725 aharonian , f. a. , akhperjanian , a. , barrio , j. _ et al . _ 2001 ,
a@xmath70a , 370 , 112 - 120 aharonian , f.a . ,
akhperjanian , a. g. , bazer - bachi , a. r. , _ et al . _ 2007 ,
a@xmath70a , 464 , 235 - 243 aharonian , f.a . ,
akhperjanian , a. g. , bazer - bachi , a. r. , _ et al . _ 2008 ,
a@xmath70a , 481 , 401 - 410 aleksic , j. , alvarez , e. a. , antonelli , l. a. , _ et al . _ 2012 ,
a@xmath70a , 541 , 13 - 23 giordano , f. , naumann - godo , m. and ballet , j. _ et al . _ 2012 ,
apj , 744 , 2 - 6 giuliani a. , tavani , m. , bulgarelli , a. _ et al . _ 2010 ,
a@xmath70a , 516 , 11 - 13 hewitt , j.w . ,
grondin , m.h .
, lemoine - goumard , m. _ et al . _ 2012 ,
apj , 759 , 89 - 98 katsuta , j. 2012 , apj , 752 , 135 - 146 lemoine - goumard , m. 2012 , a@xmath70a , 545 , 28 - 36 bell , a.r .
1978 , mnras , 182 , 147 - 156 malkov , m.a .
@xmath70 drury , o. 2001 , rep .
, 64 , 429 - 481 blasi , p. , gabici , s. @xmath70 vannoni , g. , 2005 , mnras , 361 , 907 - 918 giuliani , a. , cardillo , m. , tavani , m. , _ et al . _
2011 , apj , 742 , 30 - 34 ackermann , m. , ajello , m. and allafort , a. , _ et al . _
2013 , science , 339 , 807 - 811 cardillo , m. , tavani , m. , giuliani , a. , _ et al . _
2014 , a@xmath70a , 565 , 74 - 85 clark , d.h .
@xmath70 caswell , j.l 1976 , mnras , 174 , 267 - 305 wolszczan , a. , cordes , j. m. @xmath70 dewey , r. j. , 1991 , aj , 372 , l99-l102 castelletti , g. , dubner , g. , brogan , c. , _ et al . _ 2007 ,
a@xmath70a , 471 , 537 - 549 giacani e.b . , dubner , g. m. , kassim , n. e. , _ et al . _ 1997 , aj , 113 , 1379 - 1390 wootten , h.a .
1977 , apj , 216 , 440 - 445 rho , j. , petre , r. and schlegel , e. m. , _ et al . _
1994 , aj , 430 , 757 - 773 claussen , m. j. , frail , d. a. @xmath70 goss , w. m. 1997 , apj , 489 , 143 - 159 hoffman , m. , goss , w. m. , brogan , c. l. and claussen , m. j. 2005 , aj , 620 , 257 - 273 frail , d. a. , giacani , e. b. , goss , w. m. @xmath70 dubner , g. 1996 , apj , 464 , 165 - 169 watson , m. g. , willingale , r. , pye , j. p. , _ et al . _
1983 , iaus , 101 , 273 - 280 shelton , r.l . , kuntz , k.d
. @xmath70 petre , r. 2004 , apj , 611 , 906 - 618 tavani , m. , barbiellini , g , argan , a. , _ et al . _ 2009 ,
a@xmath70a , 502 , 995 - 1013 vercellone , s. , chen , a.w . , giuliani , a. , _ et al . _
2008 , apj , 676 , l13-l16 vercellone s. , chen , a.w . ,
vittorini , v. , _ et al . _ 2009 ,
apj , 690 , 1018 - 1030 malkov , m.a . , diamond p.h .
@xmath70 sagdeev , r.z .
2011 , nature com . , 2 , 194 - 198 yoshiike , s. 2013 , apj , 768 , 179 - 188 kelner , s. r. , aharonian , f. a. , @xmath70 bugayov , v. v. 2006 , phys .
, 74 , 034018 simpson , j.a .
1983 , ann.rev .
nucl . part .
, 33 , 323 - 381 dermer , c. d. 1986 , a@xmath70a , 157 , 223 - 229 aharonian , f. a. 2004 , _ very high energy cosmic gamma radiation : a crucial window on the extreme universe _ , ( world scientific publishing ) koo , b , heiles , c. , stanimirovi , s. @xmath70 troland , t. 2010 , aj , 140 , 262 - 265 castelletti , g. , dubner , g. , clarke , t. @xmath70 kassim , n. e. 2011 , a@xmath70a , 534 , 21 - 34 morlino , g. @xmath70 caprioli , d. 2012 , a@xmath70a , 538 , 81 - 95 bell , a.r .
2001 , mnras , 321 , 433 - 438 blasi , p.
@xmath70 amato , e. 2012a , jcap , 01 , 10 - 39 blasi , p.
@xmath70 amato , e. 2012b , jcap , 01 , 011 cardillo , m. , giuliani , a. and tavani , m. 2012 , aipc , 1505 ..
221c hwang , u. @xmath70 laming , j.m .
2012 , apj , 746 , 130 - 147 gomez , h.l . ,
clark , c. j. r. , nozawa , t. _ et al . _
2012 , mnras , 420 , 3557 - 3573 caprioli , d. 2011b , arxiv1111.0116c dubner , g. , loiseau , n. , rodrguez - pascual , p. , _ et al . _
2013 , a@xmath70a , 555 , 9 - 17 onic , d. 2013 , astrophysics and space science , 346 , 3 - 13 draine , b.t .
@xmath70 mckee , c.f .
1993 a@xmath70a , 31 , 373 - 432 blasi , p. 2012c , apj , 755 , 121 - 132 ohira , y. 2012 , apj , 758 , 97 - 105 kulsrud , r. @xmath70 pearce , w.p .
1969 , apj , 156 , 445 - 469 | until now , providing an experimental unambiguous proof of cosmic ray ( cr ) origin has been elusive . the supernova remnant ( snr ) study showed an increasingly complex scenario with a continuous elaboration of theoretical models .
the middle - aged supernova remnant ( snr ) w44 has recently attracted attention because of its relevance regarding the origin of galactic cosmic - rays .
the gamma - ray missions agile and fermi have established , for the first time for a snr , the spectral continuum below 200 mev which can be attributed to neutral pion emission .
our work is focused on a global re - assessment of all available data and models of particle acceleration in w44 and our analysis strengthens previous studies and observations of the w44 complex environment , providing new information for a more detailed modeling .
however , having determined the hadronic nature of the gamma - ray emission on firm ground , a number of theoretical challenges remains to be addressed in the context of cr acceleration in snrs .
acceleration of particles , astroparticle physics , shock waves , radiation mechanisms , supernova remnants , gamma - rays |
pump - probe experiments with femtosecond time resolution can record various nonequilibrium processes in solids directly in the time domain , including those induced by the coulomb interaction between electrons , or the scattering of electrons on defects and phonons . in these experiments
excitation of the sample and characterization of the excited state are accomplished by two distinct laser pulses ( pump and probe ) which hit the sample with controlled time delay , and either optical or photoemission spectroscopy may be used as probe technique .
the pump - probe setup has been used to investigate the dynamics of molecules,@xcite semiconductors,@xcite and metals@xcite for more than two decades .
more recently , such time - resolved experiments were also performed on several strongly correlated materials close to a phase transition , where many degrees of freedom contribute to the dynamics on very different time scales .
@xcite in such nonequilibrium solid - state experiments it is a major challenge to distinguish the electronic dynamics from other degrees of freedom .
this is crucial in particular for the mott metal - insulator transition,@xcite which is driven by the coulomb interaction between electrons moving in a crystal lattice .
an entirely new perspective on this phenomenon would open up if one could observe the transition as it happens in real time and , e.g. , monitor the formation of well - defined quasiparticles as the system goes from an insulating to a metallic state .
in fact , in several mott and charge - transfer insulators the transition to a metallic state can be induced by a laser pump pulse.@xcite so far these experiments have focused on the relaxation back to the insulating state , which involves coupling to degrees of freedom other than the valence band electrons , and is much slower than the buildup of the metallic state after the pump pulse . in the simplest case the observed relaxation is described by the two - temperature model,@xcite i.e. , as the cooling of a hot electron gas which is coupled to the colder lattice.@xcite the true dynamics of the electronic system has so far been observed only in simple metals , by looking at the thermalization of pump - excited electron distributions due to electron - electron scattering.@xcite in strongly correlated materials , thermalization is apparently much faster .
sufficient time resolution is now becoming available due to recent advances in femtosecond laser techniques,@xcite which have already allowed to investigate some solid - state systems even on the attosecond time scale.@xcite in equilibrium , electronic properties of correlated materials can be obtained directly from conventional photoemission spectroscopy with continuous light beams.@xcite by contrast , time - resolved measurements are likely to be restricted by the frequency - time uncertainty of the probe pulse : the energy @xmath0 of occupied states in the solid from which photoelectrons are released is determined from the kinetic energy of the photoelectrons , the work function of the solid , and the photon energy @xmath1 @xmath2 @xmath3 ; when the measurement pulse has finite duration @xmath4 , the latter is determined only up to an uncertainty @xmath5 @xmath6 @xmath7 . in a strongly correlated electron system
we would expect that typical relaxation times are directly related to the energy scales that appear in the spectrum , such as the bandwidth or the mott gap . in this case
all information on the initial energy @xmath0 is lost for pulses which are short enough to resolve the electronic dynamics .
the equilibrium interpretation of conventional photoemission data in terms of the electronic spectrum of the solid thus becomes meaningless in this limit .
a full theory for time - resolved photoemission spectroscopy ( trpes ) , which covers both the nonequilibrium effects of the electronic state and also the consequences of the frequency uncertainty of the pulse , was presented recently by freericks _
et al._.@xcite their approach extends existing theories of conventional photoemission spectroscopy to the case where the sample is not in equilibrium and measurement pulses have a finite time duration . the photoemission intensity as a function of the probe pulse delay time is related to electronic one - particle real - time green functions of the sample,@xcite which fully incorporate the nonequilibrium many - body dynamics after the pump pulse
this is in contrast to earlier green function approaches,@xcite which treat pump and probe on the same ( perturbative ) level .
the relation to real - time green functions allows to make direct contact to recent progress in nonequilibrium many - body theory , such as the extension of dynamical mean - field theory@xcite ( dmft ) to nonequilibrium .
@xcite dmft , which is exact in the limit of infinite dimensions,@xcite can provide insights into the real - time evolution of strongly correlated systems in a nonperturbative way . the new one - particle description of trpes given in ref
. leads to the question whether real - time green functions can be recovered fully from the time - dependent photoemission intensity , or whether parts of the electronic time evolution are not accessible by trpes at all .
freericks _ et al .
_ discussed the case when electronic _ equilibrium states _ are probed by pulses of finite time duration.@xcite this analysis covers experiments ( e.g. , those of ref . ) in which changes of external parameters such as the electronic temperature determine the dynamics of the electronic state , but the probe pulses are not short enough to resolve the thermalization of the electronic system in response to the pump pulse . for this case the electronic state is characterized by its frequency - dependent spectrum , and the photoemission intensity is given by this spectrum , broadened in accordance with the frequency - time uncertainty.@xcite such a broadening can hamper the determination of the electronic spectrum from photoemission data ; for the experiment of ref .
, however , it plays a minor role.@xcite by contrast , in this paper we investigate trpes with ultrashort pulses that do resolve the thermalization of the electrons after the pump pulse .
we consider systems with a purely hamiltonian time evolution involving only electronic degrees of freedom .
the electronic state is then no longer characterized only by a frequency - dependent spectrum , but rather by real - time green functions depending on two time variables .
we will show from the general theory of ref . that in this case the full time dependence on both time variables can not be recovered from the time - dependent photoemission intensity , no matter how the pulse length of the probe pulse is chosen . while time - resolved photoemission data can be predicted from calculated nonequilibrium green functions,@xcite
the converse procedure is thus impossible due to the frequency - time uncertainty relation .
we note that an analogous limitation is absent in time - resolved optical spectroscopy , where a two - time optical conductivity @xmath8 can be measured precisely by making the probe pulses sufficiently short . in nonequilibrium dmft , @xmath8 is directly related to momentum - averaged real - time green functions under certain conditions.@xcite below we employ the falicov - kimball model,@xcite which describes localized and mobile electrons on a lattice interacting via a local hubbard interaction , to study the relation between nonequilibrium green functions and time - resolved photoemission data in detail .
we consider an idealized setup in which the system is suddenly driven out of a metallic or insulating equilibrium state , and subsequently relaxes to a new phase due to the hamiltonian dynamics of the electrons .
this model situation was recently solved with nonequilibrium dmft.@xcite we then study hypothetical time - resolved photoemission experiments during this relaxation process , and find that some aspects of the formation of the metallic state are indeed obscured in the photoemission spectrum due to the frequency - time uncertainty .
on the other hand , the relaxation of an excited mott insulator leads to characteristic collapse - and - revival oscillations,@xcite which result in oscillating mid - gap weight in the time - resolved photoemission spectrum .
the above - mentioned uncertainty limitations notwithstanding , trpes with ultrashort pulses is well - suited to characterize nonequilibrium states of correlated electron systems . however , it will often be necessary to analyze in detail how the time evolution of the green function translates into the photoemission signal .
the outline of the paper is as follows . in sec .
[ sec : trpes ] we briefly outline the microscopic formulation of trpes derived in ref . and further discuss the role of the frequency - time uncertainty in this theory .
we then introduce the falicov - kimball model ( sec .
[ sec : fkm ] ) and discuss hypothetical time - resolved photoemission measurements on systems that relax to a metallic state ( sec . [ sec : i2 m ] ) and to an insulating state ( sec .
[ sec : m2i ] ) .
the discussion in sec .
[ sec : conclusion ] concludes the presentation .
in photoemission experiments with both temporal and angular resolution the sample is probed with a finite pulse of definite wave vector@xcite @xmath9 .
the detector collects the photoelectrons which are emitted in a certain direction @xmath10 , and it is sensitive to their kinetic energy @xmath11
@xmath2 @xmath12 , but not to their arrival time ( @xmath13 @xmath2 @xmath14 is the photoelectron momentum ) .
the time - resolved photoemission signal is thus proportional to the total number of electrons per solid angle @xmath15 and energy interval @xmath16 @xmath17 that are emitted in response to a pulse that hits the sample at time @xmath18.@xcite this definition includes only photoelectrons excited by the probe pulse and omits direct photoemission due to the pump pulse . in ref .
an expression for the photoemission signal ( [ arpes - intensity ] ) was derived , using only the so - called sudden approximation,@xcite which neglects the interaction of photoelectrons with the remaining sample .
the photoemission signal is then only related to matrix elements @xmath19 which couple bloch states with quasi - momentum @xmath20 in the solid and one - electron scattering states with asymptotic momentum @xmath13 via absorption of a photon with momentum @xmath9 , and to the real - time one - particle green function @xmath21.\ ] ] the latter incorporates the full nonequilibrium dynamics of the sample : @xmath22 @xmath2 @xmath23 are annihilation ( creation ) operators for electrons in the solid with momentum @xmath20 , whose propagation in time , with @xmath24 @xmath2 @xmath25 $ ] , includes all external fields except for the probe .
the initial state at some early time @xmath26 is usually given by the thermal ensemble at temperature @xmath27 , @xmath28 @xmath29
@xmath30 $ ] . the presence of the surface and the dependence of photoemission spectra on matrix elements can substantially complicate the comparison of theoretical and experimental data for specific materials . in order to reveal general aspects of trpes
we thus resort to further approximations that are commonly made in this context : ( i ) we assume that photoemission measures the bulk properties of the sample which are contained in the momentum - diagonal green function @xmath31 @xmath32 @xmath33 of the infinite and translationally invariant system , and ( ii ) , we take matrix elements to be constant but satisfying momentum conservation in the plane , @xmath34 @xmath32 @xmath35 . the time - resolved photoemission spectrum ( [ arpes - intensity ] )
is then given by@xcite @xmath36 where @xmath37 is the ( real ) pulse envelope function ( centered at @xmath38 @xmath2 @xmath39 ) and @xmath40 is the work function of the solid .
note that eqn .
( [ arpes-2d ] ) and ( [ arpes - simple ] ) become exact for perfectly layered ( two - dimensional ) structures , when @xmath20 denotes the two - dimensional momentum of the sample . in the following
we will discuss the momentum- and frequency - dependent expression ( [ arpes - simple ] ) .
for simplicity we will refer to eq .
( [ arpes - simple ] ) as the photoemission intensity ; observations made for this function presumably persist after summation over some part of the brillouin zone [ eq .
( [ arpes-2d ] ) ] .
( [ arpes - simple ] ) simplifies when the system is in equilibrium . in this case
green functions depend on the time difference only , and the fourier transform is given by@xcite @xmath41 where @xmath42 is the equilibrium spectral function,@xcite and @xmath43 @xmath2 @xmath44 is the fermi function for temperature @xmath27 .
the photoemission intensity then reduces to @xmath45 which is a convolution of the well - known expression @xmath46
@xmath29 @xmath47 for the intrinsic photocurrent in continuous beam experiments@xcite with the fourier transform @xmath48 @xmath2 @xmath49 of the pulse envelope . due to the frequency - time uncertainty of the pulse
, the frequency - dependent spectrum is thus smoothened on a scale @xmath50 @xmath51 @xmath52 when the pulse has a finite length @xmath4 , as discussed in refs . and .
here we study the case of an electronic system that is not in equilibrium .
the green function @xmath53 then contains important information both in the absolute time @xmath54 and in the time difference @xmath55 between addition and removal of an electron .
however , when the probe pulse extends only over a finite length @xmath56 , the product @xmath57 in eq .
( [ arpes - simple ] ) vanishes for all @xmath54 @xmath58 @xmath59 @xmath51 @xmath56 , and hence @xmath53 enters eq .
( [ arpes - simple ] ) only for @xmath55 @xmath60 @xmath56 .
it is therefore impossible to deduce @xmath53 from spectra that were recorded with pulses of length @xmath56 @xmath60 @xmath55 .
in other words , the time resolution ( in @xmath54 ) with which @xmath53 can be measured is limited by @xmath55 .
this also becomes clear when attempting to invert the convolution of @xmath61 in eq .
( [ arpes - simple ] ) : starting from the fourier transform @xmath62 and using , e.g. , gaussian pulses , @xmath63 we obtain @xmath64 while the integral in ( [ its3 ] ) apparently measures @xmath65 with a time resolution of @xmath56 , it is practically impossible to choose @xmath56 @xmath66 @xmath55 because then the result vanishes compared to any noise added to @xmath67 , due to the gaussian prefactor ( whose form is due to ( [ gauss ] ) but the suppression of the signal for @xmath56 @xmath66 @xmath55 is independent of the pulse shape ) .
we conclude that the nonequilibrium two - time green function can not be fully measured by means of trpes , and one must always carefully analyze how the time evolution of the green function translates into the photoemission signal for a given theoretical model .
this will be illustrated for the falicov - kimball model below .
in the remaining part of this paper we concentrate on one specific model for electronic dynamics in a single band , the falicov - kimball model.@xcite this lattice model describes itinerant ( @xmath68 ) and immobile ( @xmath69 ) electrons which interact via the local coulomb repulsion @xmath70 .
the hamiltonian is given by @xmath71 where @xmath72 are annihilation ( creation ) operators for the two species of fermions on lattice site @xmath73 , and @xmath74 @xmath2 @xmath75 is the corresponding density ( @xmath76 @xmath2 @xmath68 , @xmath69 ) .
hopping between sites @xmath73 and @xmath77 ( with amplitude @xmath78 ) is possible only for the mobile ( @xmath68 ) particles . the falicov - kimball model has been an important benchmark for the development of dmft in equilibrium , because the effective single - site problem for the mobile particles is quadratic and can be solved exactly.@xcite this model currently plays a similar role for nonequilibrium dmft,@xcite in particular since no appropriate real - time impurity solver is yet available for the hubbard model . in spite of its apparent simplicity
the falicov - kimball model has a rich equilibrium phase diagram containing metallic , insulating , and charge - ordered phases.@xcite in the following we consider only the homogeneous phase at half - filling for both particle species ( @xmath79 @xmath2 @xmath80 @xmath2 @xmath81 ) , which in equilibrium undergoes a metal - insulator transition at a critical interaction @xmath70 @xmath2 @xmath82 on the order of the bandwidth.@xcite this phase is studied in an idealized nonequilibrium situation , by preparing the system in thermal equilibrium for times @xmath54 @xmath60 @xmath39 , and changing the interaction parameter @xmath70 abruptly at @xmath54 @xmath2 @xmath39.@xcite of course within a more realistic description of the pump pulse the system would not be excited into a state that is an equilibrium state of any simple hamiltonian .
nonetheless , the _ interaction quench _ we study here allows us to discuss time - resolved photoemission signals for a situation in which the electronic dynamics drives the system between different phases .
since the exact dmft solution for the interaction quench is available@xcite we can directly relate green functions and photoemisson signals .
in particular , we will focus on two specific phenomena , namely ( i ) the formation of narrow quasiparticle resonances during the buildup of the metallic state ( sec .
[ sec : i2 m ] ) , and ( ii ) coherent collective oscillations after an excitation of the insulating phase ( sec . [ sec : m2i ] ) . in the following
we analyze the time - resolved photoemission signal of the mobile ( @xmath68 ) electrons using gaussian pulses ( [ gauss ] ) and omit the time - independent contribution of the immobile ( @xmath69 ) electrons . as in ref .
we assume a semielliptic density of states @xmath83 @xmath2 @xmath84 with half - bandwidth @xmath85 for the single - particle energies @xmath86 , which are the eigenvalues of the hopping matrix @xmath78 .
exact expressions for the real - time green functions @xmath31 of the mobile electrons are given in the appendix ( from now on we suppress the index @xmath68 of the mobile electrons ) .
we take @xmath87 @xmath2 @xmath88 as the unit of energy , so that the full bandwidth is @xmath89 and the critical interaction is given by @xmath82 @xmath2 @xmath85 = @xmath90.@xcite we will also set @xmath91 @xmath2 @xmath88 , setting the unit of time as @xmath92 .
for example , for @xmath87 @xmath2 1 ev we have @xmath92 @xmath2 0.66 fs .
[ eq . ( [ gkinf ] ) ] in the initial ( a ) and final ( b ) state for the quench from @xmath70 @xmath2 @xmath93 to @xmath70 @xmath2 @xmath94 ( @xmath95 @xmath2 @xmath96 @xmath2 @xmath81 ; temperature @xmath27 @xmath2 @xmath39 ) . due to the local self - energy in dmft
, @xmath97 depends on @xmath20 only via @xmath86 in the homogeneous phase .
( c ) density @xmath98 for momentum @xmath20 with @xmath86 @xmath2 @xmath99 [ thick red curve in ( a ) and ( b ) ] .
the horizontal line is at @xmath100 .
( d ) green function @xmath101 for the same @xmath86 .
differences between the green functions for @xmath54 @xmath2 @xmath39 and @xmath54 @xmath2 @xmath102 are best visible around @xmath55 @xmath2 @xmath39 ; their decay is almost identical . ] in this section we investigate the formation of a metallic state in real time , similar to pump - induced insulator - to - metal transitions on ultrashort time scales . @xcite in the falicov - kimball model , such a process takes place after a quench from an insulating state to the metallic parameter regime , which we consider now .
we prepare the initial state at @xmath70 @xmath2 @xmath93 and temperature @xmath27 @xmath2 @xmath39 , and change the interaction abruptly to @xmath70 @xmath2 @xmath94 at time @xmath54 @xmath2 @xmath39 .
subsequently the system relaxes to a new stationary state , in which green functions depend on time difference only.@xcite in the following we first discuss the real - time green functions for this process and then the corresponding time - resolved photoemission signal .
the difference between the initial and final state is evident from the momentum - diagonal green function @xmath31 of the mobile particles [ eq . ( [ gkk ] ) ] and its fourier transform @xmath103 in the limit @xmath54 @xmath2 @xmath104 , respectively : while @xmath105 has a broad maximum ( fig .
[ fig : gu3 - 0.5]a ) , a sharp peak in @xmath106 indicates that quasiparticle excitations have a long lifetime in the final state ( fig . [ fig : gu3 - 0.5]b ) .
note that for a quench in the falicov - kimball model the final state always retains memory on the initial configuration.@xcite this memory is contained in a nonuniversal occupation function @xmath107 which replaces the fermi function @xmath43 in eq .
( [ gleseq ] ) , i.e. , @xmath106 @xmath2 @xmath108 ( see appendix ) .
the development of the metallic state with its sharp quasiparticle - like resonances can be observed from the full time dependence of the momentum - diagonal green function @xmath97 . in particular
, we characterize the transition by means of ( i ) the total spectral weight , i.e. , the momentum occupation @xmath98 @xmath2 @xmath109 , and ( ii ) the decay of @xmath110 as a function of @xmath55 @xmath51 @xmath39 . from the latter one can read off the lifetime of a hole which is created at time @xmath54 in the nonequilibrium state .
the time evolution of these two quantities is similar for all @xmath20 ; it is shown for one representative value of @xmath20 in fig .
[ fig : gu3 - 0.5]c and [ fig : gu3 - 0.5]d ( namely @xmath86 @xmath2 @xmath99 , marked by the thick red line in fig . [ fig : gu3 - 0.5]a and [ fig : gu3 - 0.5]b ) : ( i ) relaxation of the momentum occupation @xmath98 takes place on a time scale on the order of the inverse bandwidth ; after this very short time interval the final value @xmath100 is almost reached , and a slower relaxation follows ( fig . [ fig : gu3 - 0.5]c ) .
a similar behavior was observed previously for the time dependence of the number of doubly occupied sites.@xcite ( ii ) as a function of @xmath55 , the green function @xmath110 decays slow compared to the inverse bandwidth ( fig .
[ fig : gu3 - 0.5]d ) . for @xmath54 @xmath2 @xmath102 ,
this is in accordance with the sharp peak in @xmath106 .
however , the fast decay is observed even for holes that are created at @xmath54 @xmath2 @xmath111 , when the system is still in the insulating state , indicating that the lifetime of hole excitations depends only weakly on the time of their creation .
the reason for this behavior is that in the falicov - kimball model scattering occurs only between mobile and immobile particles and is thus determined by the hamiltonian and the ( initial ) configuration of the immobile particles . if scattering occurred between two mobile electron species ( e.g. , as in the hubbard model ) , then we would expect the shape of the quasiparticle resonances to depend on the quantum state of the mobile particles as well , and therefore would expect it to change considerably during the relaxation process . ) ] ( a.u . ) for the same situation as the green functions in fig .
[ fig : gu3 - 0.5 ] ( @xmath86 @xmath2 @xmath99 ) , using gaussian probe envelopes ( [ gauss ] ) with @xmath56 @xmath2 @xmath112 ( a ) , @xmath56 @xmath2 @xmath90 ( b ) , and @xmath56 @xmath2 @xmath113 ( c ) . the vertical dashed lines are explained in the text . @xmath18 and @xmath56 are in units of @xmath92 @xmath2 @xmath88 . ] as discussed in the introduction , it is a central question whether the time - resolved photoemission spectrum @xmath114 [ eq . ( [ arpes - simple ] ) ] contains the same information as the green function @xmath97 . for the present case
, the answer is no : in the last subsection we saw that holes decay at a rate @xmath115 @xmath66 @xmath116 , which depends only weakly on the time @xmath54 when the hole is created , even for short times @xmath39 @xmath60 @xmath54 @xmath60 @xmath92 . to establish this behavior from the photoemission intensity , however , one would have to measure @xmath117 with time resolution better than @xmath92 , which is impossible according to the discussion at the end of sec .
[ sec : trpes ] . the effect of the frequency - time uncertainty can be seen in detail in the redistribution of spectral weight in the photoemission signal @xmath118 as a function of the probe time @xmath18 ( for @xmath86 @xmath2 @xmath99 , fig .
[ fig : pesu3 - 0.5 ] ) .
when the system is probed in a stationary state ( @xmath18 @xmath2 @xmath119 ) , the intensity is given by @xmath120 folded with the spectrum @xmath121
@xmath2 @xmath122 of the probe pulse [ cf .
( [ pes - stationary ] ) ] .
this broadening completely washes out the peak for short pulses ( @xmath56 @xmath2 @xmath113 in units of @xmath92 , fig .
[ fig : pesu3 - 0.5]c ) . on the other hand , pulses much longer than the inverse bandwidth
do not resolve the fast relaxation which is essentially complete for @xmath54 @xmath51 @xmath90 , but rather show a time dependence even for @xmath54 @xmath6 2 ( @xmath56 @xmath2 @xmath112 , fig .
[ fig : pesu3 - 0.5]a ) . for intermediate pulse length
neither time nor frequency dependence is resolved ( @xmath56 @xmath2 @xmath90 , fig .
[ fig : pesu3 - 0.5]b ) . for a quantitative analysis
we now consider the weight in the central region of the peak in the photoemission spectrum , @xmath123 in the present case we use @xmath124 @xmath2 @xmath125 and @xmath126 @xmath2 @xmath127 as indicated by the vertical dashed lines in fig .
[ fig : pesu3 - 0.5 ] ; they enclose twice the full width at half maximum of a lorentzian fit to the peak in @xmath128 .
we then define a time @xmath129 at which the relaxation of @xmath130 is essentially complete , @xmath131-w_\delta(-\infty ) } { w_\delta(\infty)-w_\delta(-\infty ) } = \eta , \end{aligned}\ ] ] with @xmath132 close to one ; we use @xmath132 @xmath2 0.95 . in fig .
[ fig : pestransfer]a we see first of all that @xmath129 is proportional to the pulse length for large @xmath56 , which is due to the fact that long pulses merely average the spectrum of the final and initial state . hence @xmath129 does not yield any information about intrinsic relaxation times for @xmath56 @xmath6 @xmath90 . on the other hand , for @xmath56 @xmath133 @xmath90 the ratio @xmath134 approaches the value @xmath135 that one would obtain by integrating over the whole spectrum instead of the peak region alone ( fig .
[ fig : pestransfer]b ) ; this is due to the insufficient energy resolution .
therefore only the relaxation time of the whole spectral width , i.e. , of the momentum occupation @xmath98 , can be determined reliably .
( [ taueta ] ) , @xmath132 @xmath2 @xmath136 for the spectral width in the central peak region of the photoemission spectrum at @xmath86 @xmath2 @xmath99 ( region between the vertical dashed lines in fig .
[ fig : pesu3 - 0.5 ] ) , plotted against the pulse length @xmath56 .
( b ) ratio @xmath134 .
the horizontal dashed line is at @xmath135 . ]
in this section we investigate the relaxation dynamics of a mott insulator in which a metallic state has been created by the pump pulse . in the falicov - kimball model this can be simulated in the strongly interacting regime ( @xmath70 @xmath2 @xmath112 ) by preparing the system in a metallic state ( @xmath70 @xmath2 @xmath88 , temperature @xmath27 @xmath2 @xmath39 ) at @xmath54 @xmath2 @xmath39 .
again we first discuss the real - time green functions for this situation and then the time - resolved photoemission signal that corresponds to them . in the present case the momentum - dependent green function evolves from a well - defined quasiparticle band at @xmath54 @xmath2 @xmath137 , which is cut off by the fermi function ( fig .
[ fig : gku1 - 10]a ) , to a gapped spectrum at @xmath54 @xmath2 @xmath102 ( fig .
[ fig : gku1 - 10]b ) .
note that in the final state spectral weight remains in the upper band because the system is strongly excited with respect to the insulating ground state at @xmath70 @xmath2 @xmath112 , and there is no coupling to an environment to which this excess energy could be passed during the relaxation process .
( [ gkinf ] ) ] in the initial ( a ) and final ( b ) state for the quench from @xmath70 @xmath2 @xmath88 to @xmath70 @xmath2 @xmath112 ( @xmath95 @xmath2 @xmath96 @xmath2 @xmath81 ; temperature @xmath27 @xmath2 @xmath39 ) .
( c ) density @xmath98 for momentum @xmath20 with @xmath86 @xmath2 @xmath88 [ thick red curve in ( a ) and ( b ) ] .
the vertical dashed lines are at multiples of the fundamental oscillation period @xmath138 .
the horizontal line is @xmath100 . ] as in the previous section we consider a representative momentum ( @xmath86 @xmath2 @xmath88 ) , chosen such that @xmath139 is small in the metallic state and increases after the quench ( fig .
[ fig : gku1 - 10]c ) .
again this relaxation takes place on a very short time scale ( on the order of the inverse bandwidth ) , but now @xmath139 passes through a series of damped oscillations with period @xmath138 before reaching its final value .
these oscillations are characteristic for the dynamics of a mott insulator which is dominated by a hubbard - type density interaction @xmath140 . in fact , if the hamiltonian were given only by this interaction term , then the time evolution operator @xmath141 would itself be @xmath138-periodic,@xcite and hence oscillations would occur in all nonlocal quantities . these so - called collapse - and - revival oscillations were first observed and described in experiments with ultracold atomic gases,@xcite where the hamiltonian of the system can be designed in a controlled way .
we will now discuss the fingerprint of these oscillations in the time - resolved photoemission spectrum . in . fig .
[ fig : pesu1 - 10 ] the angular - resolved photoemission spectrum @xmath114 is plotted for the same fixed momentum [ @xmath86 @xmath2 @xmath88 , using gaussian pulses ( [ gauss ] ) ] .
all features of the spectrum except for its total weight , which is proportional to @xmath98 , are washed out for short pulses ( @xmath56 @xmath2 @xmath142 , fig .
[ fig : pesu1 - 10]c ) , whereas long pulses show the formation of a gap , but can not resolve the oscillating nature of the state ( @xmath56 @xmath2 @xmath113 , fig .
[ fig : pesu1 - 10]a ) . for intermediate pulses , however , both the @xmath138-periodicity and the gap become visible ( @xmath56 @xmath2 @xmath143 , fig .
[ fig : pesu1 - 10]b ) .
@xmath2 @xmath88 to @xmath70 @xmath2 @xmath112 : photoemission signal ( a.u . )
( [ arpes - simple ] ) ] for @xmath86 @xmath2 @xmath88 , and gaussian probe envelopes ( [ gauss ] ) with @xmath56 @xmath2 @xmath113 ( a ) , @xmath56 @xmath2 @xmath143 ( b ) , and @xmath56 @xmath2 @xmath142 ( c )
. pulse lengths @xmath56 are in units of @xmath92 @xmath2 @xmath88 . ]
interestingly , the coherent oscillations are most pronounced in the center of the gap ( fig .
[ fig : pesu1 - 10]b ) .
this observation can be understood from the atomic limit of the hamiltonian ( [ fkm ] ) , i.e. , for @xmath78 @xmath2 0 . in the strongly interacting regime @xmath70
@xmath144 @xmath87 the atomic limit gives a good description of the transient behavior at short times @xmath54 @xmath133 @xmath92 . for the interaction term alone , @xmath145
@xmath2 @xmath146 @xmath58 @xmath147 the time evolution of annihilation operators is given by @xmath148 @xmath2 @xmath149 @xmath150 @xmath151 $ ] .
for @xmath54,@xmath59 @xmath51 @xmath39 , the green function then follows as @xmath152 , \end{gathered}\ ] ] with @xmath153 and @xmath154 is the expectation value in the ( arbitrary ) state at @xmath54 @xmath2 @xmath39 immediately after the pump . inserting this expression into eq .
( [ arpes - simple ] ) we find , for @xmath18 @xmath144 @xmath56 , @xmath155 .
\end{gathered}\ ] ] the first two terms are centered in the upper and lower hubbard bands at @xmath156 @xmath2 @xmath157 and @xmath156 @xmath2 @xmath158 and do not change with time . the third term , which oscillates with period @xmath138 , has its maximum where @xmath159 and @xmath160 overlap . for gaussian pulses ,
this is precisely the center of the gap because @xmath161 @xmath29 @xmath162 .
this discussion shows that on short time scales the observed time - dependent spectrum in the falicov - kimball model for large interactions resembled that of the atomic limit .
note that the initial state at @xmath54 @xmath2 @xmath39 determines only the weight of the three components , but not the frequency of the oscillations .
the oscillating midgap weight is thus a universal property of the mott insulator , which is largely independent of the excitation process . via this universal feature
it may eventually become possible to observe collapse - and - revival oscillations in trpes experiments on correlated materials .
in this paper we analyzed hypothetical time - resolved photoemission experiments on correlated electron systems that are not in equilibrium , building on the general theory of ref . .
we showed that the two - time green function , which characterizes the nonequilibrium state of the electrons , can not be fully measured with trpes , no matter how long or short the probe pulses are chosen .
for example , using dmft for the falicov - kimball model we found that in the buildup of the metallic state the green function of the transient state can not be determined from the photoemission signal . on the other hand ,
if an excited mott insulator is created by the pump pulse , its characteristic collapse - and - revival oscillations can nevertheless be inferred because they correspond to oscillating weight in the center of the mott gap in the time - dependent photoemission spectrum .
trpes is in some sense complementary to time - resolved optical spectroscopy , which measures the two - time optical conductivity @xmath8 . under certain conditions ,
the latter is obtained in dmft from a momentum - averaged product of two green functions that also enter the expression for the photoemission spectrum.@xcite the dependence of @xmath8 on both @xmath54 and @xmath59 can be measured precisely with sufficiently short probe pulses , unaffected by any minimum uncertainty , but unlike in photoemission spectroscopy there is no sensitivity toward specific momenta @xmath163 . in conclusion
, we showed that in spite of the frequency - time uncertainty of the probe pulse , trpes has the potential to discover fascinating details of the electronic thermalization process .
unlike for conventional photoemission on systems in equilibrium , however , the time - dependent photoemission signal does not yield the real - time green function directly , so that more detailed comparisons to theoretical predictions are needed .
we thank dieter vollhardt for valuable discussions .
m.e . acknowledges support by studienstiftung des deutschen volkes .
this work was supported in part by the sfb 484 of the deutsche forschungsgemeinschaft .
in this appendix , which is a direct extension of the work presented in ref . , we give the detailed derivation of the real - time green functions for the interaction quench in the falicov - kimball model , using dmft for nonequilibrium .
@xcite .
[ realtimeg ] @xmath164 \label{gret } \\ g^{{\text{a}}}_{{{{{{\bm k}}}}\sigma}}(t , t ' ) & = i\theta(t'-t ) \text{tr}[\rho_0\ { c^{{{\phantom{{{\dagger}}}\!}}}_{{{{{{\bm k}}}}\sigma}}(t),c^{{{\dagger}}}_{{{{{{\bm k}}}}\sigma}}(t')\ } ] \label{gadv } \\g^{<}_{{{{{{\bm k}}}}\sigma}}(t , t ' ) & = \;i \text{tr}[\rho_0 c^{{{\dagger}}}_{{{{{{\bm k}}}}\sigma}}(t')c^{{{\phantom{{{\dagger}}}\!}}}_{{{{{{\bm k}}}}\sigma}}(t ) ] \label{glesser } \end{aligned}\ ] ] respectively , where @xmath165 and @xmath28 are defined below eq .
( [ gkk ] ) .
dmft for nonequilibrium is based on the keldysh formalism,@xcite which yields the contour - ordered green function @xmath166 @xmath2 @xmath167 with time arguments on the contour @xmath168 that runs from @xmath26 to @xmath169 on the real axis , then from @xmath169 to @xmath26 , and finally to @xmath170 on the imaginary time axis ( @xmath171 : inverse temperature ) .
retarded , advanced , and lesser green functions ( [ realtimeg ] ) are obtained from the real - time components of the contour green function,@xcite [ contourrelations ] @xmath172 \\ g_{{{{{{\bm k}}}}\sigma}}^{{\text{a}}}(t , t ' ) & = [ g_{{{{{{\bm k}}}}\sigma}}^{11}(t , t')-g_{{{{{{\bm k}}}}\sigma}}^{21}(t , t ' ) ] \\
g_{{{{{{\bm k}}}}\sigma}}^{<}(t , t ' ) & = g_{{{{{{\bm k}}}}\sigma}}^{12}(t , t ' ) , \end{aligned}\ ] ] where superscripts refer to the two time arguments : @xmath88 , @xmath90 , and @xmath93 indicates whether a time argument is on the upper , lower , or vertical part of the contour , respectively .
these relations and also the symmetries [ dysonk ] @xmath177(t , t ' ) = { \delta^{\mathcal{c}}}(t , t ' ) , \end{aligned}\ ] ] where @xmath178 is the contour self - energy , and @xmath179 is the noninteracting green function , whose inverse @xmath180 \end{aligned}\ ] ] can be written as differential operator on the contour .
here we assumed a homogeneous state , and @xmath181 @xmath2 @xmath182 $ ] are the single - particle band energies .
we also introduced the convolution @xmath183
@xmath2 @xmath184 of two functions along the contour , the contour delta function @xmath185 [ defined by @xmath186 @xmath2 @xmath187 , and the contour derivative @xmath188.@xcite taking into account the boundary conditions ( [ boundaryconditions ] ) , the integro - differential equation ( [ dysonk ] ) has a unique solution for @xmath166 . in dmft
the self - energy is assumed to be local in space , and hence independent of @xmath20 for a homogeneous state .
this approximation is exact in the limit of infinite spatial dimensions,@xcite both for equilibrium and for the keldysh self - energy.@xcite the local self - energy @xmath189 and the local green function @xmath190 @xmath32 @xmath191 @xmath2 @xmath192 are then calculated from an auxiliary problem in which the degrees of freedom at a single lattice site @xmath73 are coupled to some unknown environment , which must be determined self - consistently . for the falicov - kimball model ,
the auxiliary problem is quadratic@xcite such that the equations of motion can be solved explicitly .
the dmft equations for the interaction quench in the falicov - kimball model were derived in ref . .
in the following we state these equations without derivation , and then give details of the solution . in particular we calculate the momentum - dependent green function @xmath193 @xmath32 @xmath194 of the mobile ( @xmath68 ) particles , which is needed for the photoemission intensity ( [ arpes - simple ] ) . from now on we consider only properties of the mobile particles , and omit the index @xmath68 . [ g - from - lambda ] @xmath195 of the local green functions @xmath196 and @xmath197 at sites with zero and one immobile ( @xmath69 ) particle , respectively , weighted with the average density @xmath198 @xmath2 @xmath199 of immobile particles . the functions @xmath196 and @xmath197 obey the equations of motion @xmath200 { q}(t , t ' ) - ( \lambda { \ast}{q})(t , t ' ) = { \delta^{\mathcal{c}}}\!(t , t ' ) \,,\label{eq : eqm - d } \\ & { } [ i\partial_{t}^{\mathcal{c}}\
! + \mu - u(t ) ] { r}(t , t ' ) - ( \lambda { \ast}{r})(t , t ' ) = { \delta^{\mathcal{c}}}\!(t , t ' ) , \label{eq : eqm - b } \end{aligned}\ ] ] and boundary conditions ( [ boundaryconditions ] ) . for
a quench the interaction is piecewise constant in time , @xmath201 @xmath2 @xmath202 @xmath150 @xmath203 .
the effective medium propagator @xmath204 must be determined self - consistently . for a semielliptic density of states of the mobile particles , which we adopt in the following
, the self - consistency cycle can be condensed into closed form,@xcite @xmath205 where @xmath85 is the half bandwidth of the density of states .
( [ g - from - lambda ] ) and ( [ sce ] ) form a complete set of equations for the local green function .
the local self - energy @xmath206 of the mobile particles is then obtained from the dyson equation of the local problem , @xmath207 @xmath58 @xmath208(t , t')$ ] @xmath2 @xmath209 . together with eq .
( [ g - from - lambda ] ) , this is easily transformed into @xmath210 finally the local self - energy @xmath211 is inserted into the lattice dyson equation ( [ dysonk ] ) , @xmath212(t , t ' ) = { \delta^{\mathcal{c}}}(t , t ' ) , \end{aligned}\ ] ] which yields the @xmath20-dependent green functions .
note that @xmath193 depends on momentum only via the single particle energy @xmath86 , because we assumed a homogeneous state . to solve the contour equations ( [ eq : eqm - d ] ) , ( [ eq : eqm - b ] ) ,
( [ eqm - sigma ] ) , and ( [ dysondmft ] ) we first rewrite them in terms of their retarded and lesser components , using the identities ( [ contourrelations ] ) . in effect , this means that contour derivatives @xmath213 @xmath2 @xmath214 are replaced by real - time derivatives,@xcite and convolutions @xmath213 @xmath2 @xmath216(t , t')$ ] of two contour green functions @xmath217 and @xmath218 are expressed in terms of their retarded , advanced , and lesser components according to the langreth rules@xcite the integral boundaries account for the fact that retarded ( advanced ) green functions @xmath220 vanish when @xmath54 @xmath60 @xmath59 ( @xmath54 @xmath51 @xmath59 ) .
furthermore , we shifted @xmath26 @xmath221 @xmath137 in the second equation , such that the convolution extends over the whole axis but contributions from the vertical part at @xmath222 can be dropped .
this step is discussed in further detail below .
the contour delta function on the right - hand - side of ( [ eq : eqm - d ] ) , ( [ eq : eqm - b ] ) , and ( [ dysondmft ] ) vanishes when the lesser component is taken , and it is replaced by the usual delta function @xmath223 for the retarded components . however , because any retarded function @xmath224 vanishes for @xmath54 @xmath60 @xmath59 , retarded equations of motion are only considered for @xmath54 @xmath51 @xmath59 , and the initial value at @xmath54 @xmath2 @xmath59 is determined by the weight of the delta function and the derivative operator .
in particular , we obtain @xmath225 from eqn .
( [ eq : eqm - d ] ) , ( [ eq : eqm - b ] ) , and ( [ dysondmft ] ) .
these conditions follow also directly from the anticommutation relation of creation and annihilation operators . for
the interaction quench we treat the equations of motion separately in the four regions where both @xmath54 and @xmath59 do not change sign ; we introduce additional subscripts @xmath150 and @xmath58 which indicate whether the time arguments are greater or less than zero , respectively . inserting ( [ langreth - deriv - ret ] ) and ( [ langreth - ret ] ) into eq .
( [ g - from - lambda ] ) yields a closed set of equations for @xmath226 and @xmath227 , [ eqm - ret ] @xmath228 \\ [ i\partial_t + \mu & ] { q}^{{\text{r}}}(t , t ' ) = \int_{t'}^{t } \!d{\bar t}\ , \lambda^{{\text{r}}}(t,{\bar t } ) { q}^{{\text{r}}}({\bar t},t ' ) \\ [ i\partial_t + \mu- u(t ) & ] { r}^{{\text{r}}}(t , t ' ) = \int_{t'}^{t } \!d{\bar t}\ , \lambda^{{\text{r}}}(t,{\bar t } ) { r}^{{\text{r}}}({\bar t},t ' ) , \end{aligned}\ ] ] which must be solved for @xmath54 @xmath51 @xmath59 , using the initial condition ( [ ic - ret ] ) .
the self - consistency equation ( [ sce ] ) was used in eq .
( [ sce - ret ] ) .
note that in eq .
( [ eqm - ret ] ) , green functions with both time arguments greater or less that zero , i.e. , the ( @xmath229 ) and ( @xmath230 ) components , do not mix with other components . because @xmath201 is constant for @xmath54 @xmath51 @xmath39 and @xmath54 @xmath60 @xmath39 , respectively , the solutions of ( [ eqm - ret ] ) are thus translationally invariant in time when both @xmath54 and @xmath59 have the same sign , and we make the ansatz for all contour functions @xmath217 @xmath2 @xmath232 , @xmath233 , @xmath234 , @xmath235 , @xmath236 , and @xmath206 ( with @xmath124 @xmath2 @xmath237 , @xmath238 , @xmath239 , @xmath240 , @xmath241 , and @xmath76 , respectively ) . using this ansatz in eq .
( [ eqm - ret ] ) we obtain a set of cubic equations , that can be solved analytically .
these cubic equations are well - known from the dmft solution of the falicov - kimball model in equilibrium.@xcite this is of course expected when both @xmath54 and @xmath59 @xmath60 @xmath39 , because before the quench the system indeed is in an equilibrium state . in a similar way , the retarded ( @xmath229 ) and ( @xmath230 ) components of @xmath206 and @xmath236 are obtained from eq .
( [ eqm - sigma ] ) and ( [ dysondmft ] ) , @xmath243^{-1}. \end{aligned}\ ] ] furthermore , advanced green functions are directly related to the retarded ones by symmetry ( [ contoursymmetry ] ) , so that we have the lesser green functions are translationally invariant in time only for both @xmath54 and @xmath59 @xmath60 @xmath39 [ ( @xmath230 ) component ] , when the system is still in equilibrium .
one then has@xcite where @xmath43 @xmath2 @xmath44 is the fermi function , and @xmath246 @xmath58 @xmath247 @xmath2 @xmath248 is proportional to the spectrum of the equilibrium green function .
mathematically this follows from the solutions of the equations of motion on the full contour , including the vertical part , and taking into account the antiperiodic boundary conditions . for the quench we use ( [ lessereq ] ) as _ initial condition _ for the lesser components
; only then can we then let @xmath26 @xmath221 @xmath137 , and disregard the vertical part in the langreth rule ( [ langreth - lesser ] ) . on the other hand
, we show below that in the limit where both @xmath54 and @xmath59 tend to @xmath102 ( but their difference is finite ) , the lesser ( @xmath229 ) components take a form very similar to ( [ lessereq ] ) , the function @xmath107 is common for all @xmath124 @xmath2 @xmath237 , @xmath238 , @xmath239 , @xmath76 , and @xmath241 .
one can in fact directly see from the equations of motion ( [ eq : eqm - d ] ) , ( [ eq : eqm - b ] ) , ( [ eqm - sigma ] ) , and ( [ dysondmft ] ) for the lesser component that _ if _ the stationary limit ( [ lesfitt ] ) exists , _ then _ green functions @xmath250 must have this common factor @xmath107 . to find this factor , however , the equations of motion must be solved , because it contains the entire information about the initial state . for the lesser components with mixed time arguments ( which holds for @xmath59 @xmath60 @xmath39 ) , and @xmath253 for the lesser green function with both time arguments after the quench .
in this subsection we derive explicit expressions for @xmath254 , @xmath255 , and @xmath256 ; the remaining are then obtained by symmetry ( [ contoursymmetry ] ) , using ( [ langreth - long ] ) and ( [ langreth - deriv ] ) one can rewrite eq . (
[ g - from - lambda ] ) for the various components and solve them using the self - consistency equation ( [ sce ] ) .
for instance , we obtain the ( @xmath264 ) component of the retarded green functions as @xmath265{r}^{{\text{r}}}_{{+}}(z){r}^{{\text{r}}}_{{-}}(-\eta ) , \\ { \tilde}{q}^{{\text{r}}}_{{{{+}}{{-}}}}(z,\eta ) & = [ { \tilde}\lambda^{{\text{r}}}_{{{{+}}{{-}}}}(z,\eta)+i]{q}^{{\text{r}}}_{{+}}(z){q}^{{\text{r}}}_{{-}}(-\eta ) , \end{aligned}\ ] ] where eq .
( [ eq : cubic ] ) was used once .
together with the self - consistency ( [ sce ] ) , this is a simple linear equation for @xmath266 . in a similar way
all components are determined successively : starting from the retarded ( @xmath264 ) and advanced ( @xmath267 ) components , the results enter the lesser ( @xmath264 ) and lesser ( @xmath267 ) components [ cf .
( [ langreth - long - less+- ] ) and ( [ langreth - deriv - less+- ] ) ] , which in turn enter the equations for the lesser ( @xmath229 ) component [ cf . eqn .
( [ langreth - long - less++ ] ) and ( [ langreth - deriv - less++ ] ) ] .
the procedure is repeated for eq .
( [ g - from - lambda ] ) , eq .
( [ eqm - sigma ] ) , and finally for the lattice dyson equation ( [ dysondmft ] ) , which yields the momentum - dependent green function @xmath193 .
[ kernelm ] @xmath268^{-1 } , \\ \kappa^x_\alpha & = { \tilde}{r}^x_\alpha { \tilde}{q}^x_\alpha / { \tilde}g^x_\alpha . \\
k^{xy}_{\alpha\beta } & = 1 + w_1w_0u_+u_-\kappa^x_\alpha \kappa^y_\beta m^{xy}_{\alpha\beta } , \end{aligned}\ ] ] with superscripts @xmath269,@xmath270 @xmath271 @xmath272 , and subscripts @xmath273 , @xmath171 @xmath271 @xmath274 , and we use the convention that the variables of a function @xmath275 is ( i ) @xmath276 when @xmath269 @xmath2 @xmath277 and @xmath273 @xmath2 @xmath150 , ( ii ) @xmath278 when @xmath269 @xmath2 @xmath279 and @xmath273 @xmath2 @xmath150 , and ( iii ) @xmath156 when @xmath273 @xmath2 @xmath58 . in terms of these expressions
the final result is to obtain the physical real - time green functions , we have to invert the double fourier transformations ( [ dft - ret ] ) , ( [ dft - les+- ] ) , and ( [ dft - les++ ] ) , using the final expressions ( [ kernelm ] ) through ( [ lesser++final ] ) . here
we give an explicit formula for the partially fourier - transformed lesser component @xmath286 ( @xmath217 @xmath2 @xmath232 , @xmath233 , and @xmath236 ) .
the singularity at @xmath287 @xmath2 @xmath39 in eq .
( [ lesser++general ] ) determines ( [ gles - wt ] ) in the limit @xmath54 @xmath221 @xmath102 , @xmath288 , \end{aligned}\ ] ] which is of the form discussed above [ cf .
( [ lesserfi ] ) ] , with @xmath107 @xmath2 @xmath289 .
the full result is given by where @xmath291 @xmath292 } { \omega+\eta } \\
\,\,+\ , f_a(\omega,\eta ) + \int \frac{d\omega'}{2\pi i } \frac{{\tilde}a^{<}_{{{{-}}{{+}}}}(\omega',\eta)}{\omega - i0-\omega'}\ , , \end{gathered}\ ] ] and the components for @xmath217 @xmath2 @xmath232 , @xmath233 , and @xmath236 were given in the previous subsection .
note that the first term on the right hand side of eq .
( [ lasteqn ] ) is regular at @xmath132 @xmath2 @xmath293 , because both @xmath294 and @xmath295 @xmath58 @xmath296 $ ] are then purely imaginary .
l. perfetti , p. a. loukakos , m. lisowski , u. bovensiepen , h. berger , s. biermann , p. s. cornaglia , a. georges and m. wolf , phys .
lett . * 97 * , 067402 ( 2006 ) ; l. perfetti , p. a. loukakos , m. lisowski , u. bovensiepen , m. wolf , h. berger , s. biermann , and a. georges , new .
j. phys .
* 10 * , 053019 ( 2007 ) .
a. l. cavalieri , n. mller , th .
uphues , v. s. yakovlev , a. baltuska , b. horvath , b. schmidt , l. blmel , r. holzwarth , s. hendel , m. drescher , u. kleineberg , p. m. echenique , r. kienberger , f. krausz , and u. heinzmann , nature * 449 * , 1029 ( 2007 ) a short pulse has necessarily finite extent in space , but its spread in wavenumber can nevertheless be neglected due to the steep dispersion of light : even when the frequency width of the pulse is on the order of a typical bandwidth ( @xmath297 @xmath2 @xmath298 ) , the spread in wavenumber , @xmath299 @xmath2 @xmath300 @xmath301 @xmath302 , which is small compared to the typical extent of the brillouin zone . | time - resolved photoemission experiments can reveal fascinating quantum dynamics of correlated electrons .
however , the thermalization of the electronic system is typically so fast that very short probe pulses are necessary to resolve the time evolution of the quantum state , and this leads to poor energy resolution due to the energy - time uncertainty relation .
although the photoemission intensity can be calculated from the nonequilibrium electronic green functions , the converse procedure is therefore difficult .
we analyze a hypothetical time - resolved photoemission experiment on a correlated electronic system , described by the falicov - kimball model in dynamical mean - field theory , which relaxes between metallic and insulating phases .
we find that the real - time green function which describes the transient behavior during the buildup of the metallic state can not be determined directly from the photoemission signal . on the other hand ,
the characteristic collapse - and - revival oscillations of an excited mott insulator can be observed as oscillating weight in the center of the mott gap in the time - dependent photoemission spectrum . |
experimental observation @xcite of the inclusive and exclusive radiative decays @xmath1 and @xmath2 stimulated the study of rare @xmath3 decays on a new footing .
these flavor changing neutral current ( fcnc ) @xmath4 transitions in the standard model ( sm ) do not occur at the tree level and appear only at the loop level .
therefore the study of these rare @xmath3-meson decays can provide a means of testing the detailed structure of the sm at the loop level .
these decays are also very useful for extracting the values of the cabibbo - kobayashi - maskawa ( ckm ) matrix elements @xcite , as well as for establishing new physics beyond the sm @xcite . currently , the main interest on rare @xmath3-meson decays is focused on decays for which the sm predicts large branching ratios and can be potentially measurable in the near future .
the rare and @xmath5 decays are such decays .
the experimental situation for these decays is very promising @xcite , with @xmath6 and hadron colliders focusing only on the observation of exclusive modes with @xmath7 and @xmath8 final states , respectively . at quark level
the process @xmath9 takes place via electromagnetic and z penguin and w box diagrams and are described by three independent wilson coefficients @xmath10 and @xmath11 .
investigations allow us to study different structures , described by the above mentioned wilson coefficients . in the sm ,
the measurement of the forward - backward asymmetry and invariant dilepton mass distribution in @xmath12 provide information on the short distance contributions dominated by the top quark loops and are essential in separating the short distance fcnc process from the contributing long distance effects @xcite and also are very sensitive to the contributions from new physics @xcite .
recently it has been emphasized by hewett @xcite that the longitudinal lepton polarization , which is another parity violating observable , is also an important asymmetry and that the lepton polarization in @xmath13 will be measurable with the high statistics available at the b - factories currently under construction .
in calculating the branching ratios and other observables in hadron level , i.e. for decay , we have the problem of computing the matrix element of the effective hamiltonian , @xmath14 , between the states @xmath3 and @xmath15 .
this problem is related to the non - perturbative sector of qcd .
these matrix elements , in the framework of different approaches such as chiral theory @xcite , three point qcd sum rules method @xcite , relativistic quark model by the light - front formalism @xcite , have been investigated .
the aim of this work is the calculation of these matrix elements in light cone qcd sum rules method and to study the lepton polarization asymmetry for the exclusive decays .
the effective hamiltonian for the @xmath9 decay , including qcd corrections [ 1214 ] can be written as _ eff = v_tb v^*_ts _ i=1 ^ 10 c_i ( ) o_i ( ) , which is evolved from the electroweak scale down to @xmath16 by the renormalization group equations . here
@xmath17 represent the relevant ckm matrix elements , and @xmath18 are a complete set of renormalized dimension 5 and 6 operators involving light fields which govern the @xmath4 transitions and @xmath19 are the wilson coefficients for the corresponding operators .
the explicit forms of @xmath19 and @xmath20 can be found in [ 1214 ] . for @xmath9 decay
, this effective hamiltonian leads to the matrix element = v_tb v^*_ts , where @xmath21 is the invariant dilepton mass , and @xmath22/2 $ ] are the projection operators . the coefficient @xmath23 , where the function @xmath24 contains the contributions from the one loop matrix element of the four - quark operators ,
can be found in [ 1214 ] .
note that the function @xmath25 contains both real and imaginary parts ( the imaginary part arises when the c - quark in the loop is on the mass shell ) .
the decay also receives large long distance contributions from the cascade process @xmath26 .
these contributions are taken into account by introducing a breit - wigner form of the resonance propagator and this procedure leads to an additional contribution to @xmath27 of the form @xcite - _
v= j/ , ^ . as we noted earlier ,
in order to calculate the branching ratios for the exclusive decays , the matrix elements @xmath28 and @xmath29 must be calculated .
these matrix elements can be parametrized in terms of the formfactors as follows ( see also @xcite ) : k ( q ) |s _ ( 1-_5 ) q b ( p+q ) & = & 2q_f^+(p^2)+p _ , + k ( q ) |s i _
p^ ( 1 + _ 5 ) q b ( p+q ) & = & , where @xmath30 and @xmath31 are the momentum of @xmath3 and @xmath15 and @xmath32 .
according to the qcd sum rules ideology , in order to calculate these formfactors , we start by considering the representation of a suitable correlator function in hadron and quark gluon languages . for this purpose , we consider the following matrix elements of the t product of two currents between the vacuum state and the @xmath15meson : _
^(1)(p , q ) & = & id^4x e^ipx
k(q ) t\{|s(x)_(1-_5)b(x ) |b(0)i_5q(0 ) } 0 , + _ ^(2)(p , q)&=&id^4x e^ipx k(q)t\ { |s ( x)i_p^(1+_5)b(x ) |b ( 0)i_5q(0 ) } 0 , where @xmath31 is @xmath15meson momentum and @xmath34 is the transfer momentum . the hadronic ( physical ) part of eqs.(5 ) and
( 6 ) is obtained by inserting a complete set of states including the @xmath3meson ground state , and higher states with @xmath3meson quantum number : + & & = k(q ) |s(x ) _ ( 1-_5)b(x ) b(p+q ) b(p+q ) |b(0 ) i_5 q(0 ) 0 + + & & + _
h k(q ) |s(x ) _
( 1-_5)b(x ) h(p+q ) h(p+q ) |b(0 ) i_5 q(0 ) 0 + & & = f_1(p^2,(p+q)^2 ) q_+ f_2(p^2,(p+q)^2 ) p _ , +
+ + + & & = k(q ) |s(x ) i _
p^(1+_5 ) b(x ) b(p+q ) b(p+q ) |b(0 ) i _ 5 q(0 ) 0 + + & & + _
h k(q ) |s(x ) i_p^(1+_5 ) b(x ) h(p+q ) h(p+q ) |b(0 ) i_5 q(0 ) 0 + & & = f_3((p+q)^2,p^2 ) , where @xmath35 .
then , for the invariant amplitudes @xmath36 , one can write a general dispersion relation in the @xmath3 meson momentum squared , @xmath37 , as : f_i(p^2,(p+q)^2))=_m_b^2^ , where the spectral densities are given by _
1(p^2,s ) & = & ( s - m_b^2)2f^+(p^2 ) + ^h_1(p^2,s ) , + + _ 2(p^2,s ) & = & ( s - m_b^2)+_2^h(p^2,s ) , + + _ 3 ( p^2,s ) & = & ( s - m_b^2 ) + _ 3^h(p^2,s ) .
the first terms in eqs.(10 ) , ( 11 ) and ( 12 ) represent the ground state @xmath3meson contribution and follow from eqs.(5 ) and ( 6 ) by inserting the matrix elements given in eqs.(3 ) , ( 4 ) and by replacing b |b i_5 q 0 = , in which we neglect the mass of the light quarks . in eqs.(10)-(12 ) , @xmath38 represent the spectral density of the higher resonances and of the continuum of states . in accordance with the qcd sum rules method we invoke the quark
hadron duality prescription and replace the spectral density @xmath39 by _ i^h(p^2,s)=f_i^qcd(p^2,s)(s - s_0 ) , where im@xmath40 is obtained from the imaginary part of the correlator functions calculated in qcd starting from some threshold @xmath41 . in order to suppress the higher states and continuum contributions , we follow the standard procedure in the qcd sum rules method and apply borel transformation @xmath42 on the variable @xmath37 to the dispersion integral to get , f_i(p^2,m^2)&=&b f_i(p^2,(p+q)^2 ) + & = & _ m_b^2^_i(p^2,s)e^-s / m^2ds .
using eqs.(10)-(12 ) and ( 14 ) we get f_1(p^2,m^2 ) & = & 2f^+(p^2 ) e^-m_b^2/m^2 + _
s_0^f_1^qcd(p^2,s ) e^-s / m^2ds , + + f_2(p^2,m^2 ) & = & e^-m_b^2/m^2 + + & & + _ s_0^f_2^qcd(p^2,s)e^-s / m^2ds , + + f_3(p^2,m^2 ) & = & e^-m_b^2/m^2 + _ s_0^ f_3^qcd(p^2,s)e^-s / m^2ds .
the main problem is then the calculation of the correlator functions ( 5 ) and ( 6 ) in qcd . after applying the borel transformation ,
the result can be written in the following form : f_i(p^2,m^2)=_m_b^2^ f_i^qcd(p^2,s)e^-s / m^2ds . equating the @xmath36 in eq.(19 ) to the corresponding @xmath36 in eqs.(16)-(18 ) , we arrive at the sum rules for the formfactors , which describe the @xmath43 transition : f^+(p^2 ) & = & _ m_b^2^s_0 f_1^qcd(p^2,s)e^-(s - m_b^2)/m^2ds , + + f^+(p^2)+f^-(p^2 ) & = & _ m_b^2^s_0 f_2^qcd(p^2,s)e^-(s - m_b^2)/m^2ds , + + f_t(p^2 ) & = & ( m_b+m_k)_m_b^2^s_0 f_3^qcd(p^2,s)e^-(s - m_b^2)/m^2ds
. one can calculate im@xmath40 , in the deep euclidean region , where both @xmath21 and @xmath37 are negative and large .
the leading contribution to the operator product expansion comes from the contraction of the @xmath44quark operators expressed in eqs.(5 ) and ( 6 ) to the free @xmath44quark propagator 0 t\{b(x ) |b(0 ) } 0 = -i e^-ikx . then we have _
^(1)(p , q ) & = & i e^i(p - k)x + + & & k(q ) 0 , + + _ ^(2)(p , q ) & = & - e^i(p - k)x p^ + + & & k(q ) 0 . note that , as can be seen from eqs.(23 ) and ( 24 ) , the problem is reduced to the calculation of the matrix elements of the gauge
invariant , nonlocal operators sandwiched in between the vacuum and the @xmath15 meson states .
these matrix elements define the @xmath15 meson light cone wave functions . following [ 1617 ]
, we define the @xmath15meson wave functions as : k(q ) |s(x ) _ ( 1-_5)q(0 ) 0 & = & iq_f_k_0 ^ 1 du e^iuqx- + & & - f_k x_- _ 0 ^ 1 du e^iuqxg_2(u ) . in eq.(25 ) @xmath45 is the leading twist two , @xmath46 and @xmath47 are the twist four @xmath15 meson wave functions , respectively .
the second matrix element of eq.(23 ) , can be split into two matrix elements using the identity @xmath48 , and the result can be evaluated using the twist three wave functions defined as [ 1619 ] : k(q ) |s(x ) i _ 5 q(0 ) 0 & = & _ 0 ^ 1 du e^iuqx_p(u ) , + + k(q ) |s(x ) _ ( 1+_5)q(0 ) 0 & = & i ( q_x_-q_x _ ) _ 0 ^ 1 du e^iuqx_(u ) .
the matrix elements in eq.(24 ) can be easily calculated using the identities , _ & = & - _ ^_5 + _ _ & = & i(_g_- _
^_5 , to express it in terms of the wave functions in eq.(25 ) .
substituting the matrix elements ( 2527 ) into the eqs.(23 ) and ( 24 ) and integrating over the variables @xmath49 and @xmath50 we get the following expressions for @xmath51 : f_1(p^2,(p+q)^2 ) & = & m_b f_k _ 0 ^ 1 + + & + & f_k _ k _ 0 ^ 1 , + + f_2(p^2,(p+q)^2 ) & = & m_b f_k _ 0 ^ 1 \ { } , + + f_3(p^2,(p+q)^2 ) & = & - f_k _ 0 ^ 1 \ { + . + + & + & .
m_b_k } , where @xmath52 there are also contributions to the above considered wave functions from multi - particle meson wave functions .
here we consider only the operator @xmath53 , which gives the main contribution and corresponds to the quark antiquark
gluon components in the kaon ( for more detail see @xcite ) . in this approximation
, the @xmath44quark propagator is defined as ( see @xcite ) : 0 t\ { b(x ) + & & _ 0 ^ 1 du . + + it is clear that when we take into account the second term in the right hand side of eq.(31 ) , new matrix elements appear . substituting eq.(31 ) into eqs.(5 ) and
( 6 ) , and using the identity _ _ _ & = & ( _ g_-_g _ ) + i(g_g_-g_g_)- + & -&__5- i _
g^__5 , one can express the resulting new matrix elements in terms of the three particle wave functions [ 1620 ] : k(q)|s(x ) g_s g_(ux)__5 q(0 ) 0 & = & if_3k + & & _
i_3k(_i ) e^iqx , + + k(q ) |s(x)__5 g_s g_(ux)q(0 ) 0 & = & f_k + & & _
i_(_i ) e^iqx + + & + & f_k q_x_-q_x _ _ i_(_i ) e^iqx , +
+ k(q ) |s(x)_g_s g_(ux)q(0 ) 0 & = & i f_k + & & _
i_(_i ) e^iq x + + & + & i f_k ( q_x_-q_x _ ) _
i_(_i ) e^iqx , where @xmath54 , @xmath55 , and @xmath56 .
here @xmath57 is a twist three wave function and the remaining functions @xmath58 , @xmath59 , @xmath60 and @xmath61 are all twist four wave functions .
when we substitute eqs.(32)(34 ) into eqs.(5 ) and ( 6 ) , perform integration over @xmath49 and @xmath50 , add ( 28 ) , ( 29 ) , ( 30 ) , apply the borel transformation over @xmath37 and equate the obtained results to ( 16 ) , ( 17 ) and ( 18 ) , we get the following sum rules for the @xmath43 transition formfactors : f^+(p^2)&=&e^ \ { _ ^1 exp(- ) . + & & .
f_3k _ 0 ^ 1 du _
i ( - ) exp ( - ) .
+ & & . } , + + + f^+(p^2)+f^-(p^2)&=&e^ \{_^1 exp ( - ) .
+ & + & . _ 0 ^ 1du_i ( - ) exp ( - ) .
+ & & . } , f_t(p^2)&= & e^ \ { _ ^1 du exp ( - ) .
+ & + & . _ 0 ^ 1 du _
i ( - ) exp(- ) .
+ & & . } , where @xmath62 .
the functions @xmath63 , @xmath64 in eqs.(35)(37 ) are defined in the following way @xmath65 note that the formfactor @xmath66 is investigated in @xcite and @xcite ( in @xcite , it is necessary to make the simple replacement @xmath67 ) , in light cone sum rules .
our results coincide with theirs if we let @xmath68 . at the end of this section
we calculate the differential decay rate with the longitudinal polarization of the final leptons and we obtain & = & \{m_b^2 ( 2 m_^2 + m_b^2 s ) ( a ^2 + c ^2 ) + + & + & 8 m_b^2 m_^2 + + & - & re ( a^ * c ) m_b^4 v } , where @xmath69 is the longitudinal polarization of the final lepton , @xmath70 and @xmath71 are its mass and velocity , respectively . in eq.(38 ) @xmath72 , @xmath73 and @xmath74 are defined as follows : & & a= 2c_9^efff^+- c_7 , + & & c= 2 c_10f^+ , + & & d= c_10(f^+ + f^- ) .for the dileptonic decays of the @xmath3 mesons , the longitudinal polarization asymmetry @xmath75 of the final state @xmath76 , is defined by @xmath77 where @xmath78 corresponds to the left ( right ) handed lepton . if in eq.(40 ) , we let @xmath79 , our results coincide with the results in @xcite and if @xmath80 , they coincide with the ones in @xcite .
the main input parameters in the sum rules ( 35)(37 ) are the kaon wave functions on the light cone . for kaon wave functions we use the results of ref .
@xcite : @xmath81 + 0.34\left[21(2u-1)^4- 14 ( 2u-1)^2 + 1\right]\right\}~ , \nnb \\ & & \varphi_p\simeq 1~,\nonumber\\ & & \varphi_\sigma\simeq 6u(1-u)~ , \nonumber\\ & & g_1(u)\simeq\frac{5}{2}\delta^2u^2(1-u)^2 ~ , \nonumber\\ & & g_2(u)\simeq \frac{10}{3}\delta^2u(1-u)(2u-1)~ , \nonumber\\ & & \varphi_{3k}(\alpha_i)\simeq 360 \alpha_1\alpha_2\alpha_3 ^ 2~,\nonumber\\ & & \varphi_{\perp}(\alpha_i)\simeq 10\delta^2 ( \alpha_1-\alpha_2)\alpha_3 ^ 2~,\nonumber\\ & & \varphi_{\parallel}(\alpha_i)\simeq 120\delta^2\epsilon(\alpha_1-\alpha_2 ) \alpha_1\alpha_2\alpha_3 ~ , \nonumber\\ & & \tilde\varphi_{\perp}(\alpha_i)\simeq 10\delta^2 \alpha_3 ^ 2(1-\alpha_3)~,\nonumber\\ & & \tilde\varphi_{\parallel}(\alpha_i)\simeq -40\delta^2\alpha_1\alpha_2\alpha_3~ , \nonumber\end{aligned}\ ] ] here @xmath82 at @xmath83 , which follows from the qcd sum rules analysis ( for more detail see @xcite ) , @xmath84 .
in order to estimate @xmath85 , we use the pcac relation for the pseudo goldstone bosons @xmath86 here we use @xmath87 and we assume @xmath88 , which follows from qcd sum rules for strange hadrons @xcite . as a result
we have @xmath89 for the values of the other parameters , we choose : @xmath90 , which is obtained from 2point qcd sum rules analysis @xcite , @xmath91 and @xmath92 , and @xmath93 . before giving numerical results on the formfactors
, we must first determine the region for the borel mass parameter @xmath94 , for which the sum rules yields reliable results .
the lower limit of this region is determined by the requirement that , the terms proportional to @xmath95 remain subdominant .
the upper limit of @xmath94 is determined by requiring the higher resonance and continuum contributions to be less than @xmath96 of the total result .
our numerical results show that both requirements are satisfied in the region @xmath97 and for the numerical analysis we use @xmath98 .
when @xmath21 approaches the region @xmath99 a breakdown of the stability is expected , similar to the @xmath100 case ( see @xcite ) . in fig.1 , we present the @xmath21 dependence of the formfactors @xmath66 , @xmath101 and @xmath102 . at zero momentum transfer ,
the qcd prediction for the formfactors are @xmath103 in fig.2a(b ) we present the @xmath21 dependence of the differential branching ratios for the @xmath104 and ( @xmath105 ) decay with and without long distance effects . in both cases summation over the final lepton polarization
is performed .
note that @xmath21 dependence of the differential branching ratio @xmath106 is analysed in @xcite and @xcite using the light front formalism and heavy meson chiral theory , respectively . in @xcite both short and long distance contributions
are considered while in @xcite only the short distance contributions are taken into account and in both works the obtained spectrum is fully symmetric while in the present work the spectrum we obtain seems to be slightly asymmetric as a result of a highly asymmetric resonance - type behaviour due to the nonperturbative contributions .
performing the integration over @xmath21 in eq.(38 ) and using the values of the life times @xmath107 s , @xmath108 s @xcite , for the branching ratios , including only the short distance contributions , we get : b(b_d k^0 ^+ ^- ) & = & ( 3.1 0.9 ) 10 ^ -7 , + b(b_d k^0 ^+ ^- ) & = & ( 1.7 0.4 ) 10 ^ -7 , + b(b_u k^+ ^+ ^- ) & = & ( 3.2 0.8 ) 10 ^ -7 , + b(b_u k^+ ^+ ^- ) & = & ( 1.770.40 ) 10 ^ -7 , where theoretical and experimental errors have been added quadratically .
in fig.3 we display the lepton longitudinal polarization asymmetry @xmath75 as a function of @xmath21 for the @xmath104 and @xmath105 decays , at @xmath109 , with and without the long distance contributions . from this figure
one can see that @xmath75 vanishes at the threshold due to the kinematical factor @xmath110 and the value of @xmath75 for the @xmath104 decay varies in the region ( @xmath111 ) and ( @xmath112 ) for the @xmath105 decay , when long distance effects are excluded .
in this work , we calculate the transition formfactors for the exclusive @xmath113(@xmath114 ) decay in the framework of the light cone qcd sum rules , and investigate the longitudinal polarization asymmetries of the muon and tau in this decay . from a comparison of our results with the traditional qcd sum rule predictions ( see @xcite ) , we observe that the behaviour of the formfactors are similar and the value of @xmath115 in both approaches coincides , while @xmath116 differs two times than that of @xcite at @xmath117 .
it is important to note that for a more refined analysis , it is necessary to take into account the su(3 ) breaking effects : the differences between @xmath118 , @xmath85 and @xmath119 , @xmath120 and the differences of pion and kaon wave functions .
these su(3 ) breaking terms can lead to differences between @xmath121 and @xmath33 formfactors .
but we expect that these effects can change the results about 1520@xmath122 and this lies at the accuracy level of the sum rules method .
few words about the possibility of the experimental observation of this decay are in order . experimentally , to observe an asymmetry @xmath75 of a decay with the branching ratio @xmath3 at the @xmath123 level
, the required number of events is @xmath124 ( see @xcite ) .
for example , to observe the @xmath8 lepton polarization at the exclusive channel @xmath125 at the @xmath126 level , one needs at least @xmath127 decays .
since in the future @xmath3-factories , it is expected that @xmath128 @xmath3-mesons would be created per year , it is possible to measure the longitudinal polarization asymmetry of the @xmath8 lepton .
* 1 . * the @xmath21 dependence of the formfactors @xmath129 and @xmath102 .
+ + * 2 .
* a ) invariant mass squared distribution of the lepton pair for the decay @xmath130 .
+ b ) the same as in a ) , but for the decay .
+ here and in all of the following figures the solid line corresponds to the short distance contributions only and the dashed line to the sum of both short and long distance contri- + butions .
+ + * 3 . * a ) the longitudinal polarization asymmetry @xmath75 for the decay .
+ b ) the same as in a ) , but for the decay .
+ c. anway - wiese , * cdf * collaboration , in : _ proc . of the _ @xmath132 _ meeting of the division of particle and fields of the american physical society _ , albuquerque , new mexico , 1994 , ed : s. seidel ( world scientific , singapore , 1995 ) .
a. ali , t. mannel and t. morozumi , _ phys .
* b273 * ( 1991 ) 505 ; + a. ali , g. f. giudice and t. mannel , _ z. phys .
_ * c67 * ( 1995 ) 417 ; + g. burdman , _ phys .
_ * d52 * ( 1995 ) 6400 ; + f. krger and l. m. sehgal , _ phys .
* b380 * ( 1996 ) 199 .
a. j. buras and m. mnz , _ phys .
rev . _ * d52 * ( 1995 ) 186 ; + m. ciuchini , e. franco , g. martinelli , l. reina and l. silvestrini , + _ phys
. lett . _ * b316 * ( 1993 ) 127 ; + m. ciuchini , e. franco , g. martinelli and l. reina , _ nucl .
_ * b415 * ( 1994 ) 403 ; + g. cella , g. curci , r. ricciardi and a. vicere , _ nucl .
phys . _ * b421 * ( 1994 ) 41 ; + _ ibid _ _ phys . lett . _
* b325 * ( 1994 ) 227 . | 0.7 cm we calculate the transition formfactors for the @xmath0 decay in the framework of the light cone qcd sum rules .
the invariant dilepton mass distribution and the final lepton longitudinal polarization asymmetry are investigated .
the comparison analysis of our results with traditional sum rules method predictions on the formfactors is performed .
16.3 true cm 23.0 true cm -0.1 true in 0.00 true in |
the x - ray properties of accreting black hole candidates ( bhcs ) lead naturally to classification into different spectral states ( e.g. van der klis 1994 ) . in their `` high / soft '' state ,
their x - ray spectra are dominated below 10 kev by a thermal blackbody - like component with temperature @xmath3 kev . in their `` low / hard '' state ,
this thermal component is absent , and bhcs have powerlaw x - ray spectra .
an `` intermediate '' state is also seen , in which the thermal component is weaker .
individual bhcs can exhibit all spectral states , but the fraction of time spent in the different states varies considerably between objects : for example lmc x-1 has only been observed in the high / soft state , while cyg x-1 spends the majority of its time in the low / hard state .
of all the persistent bhcs known , lmc x-3 is viewed through the smallest galactic column density . until recently lmc x-3 was thought to be permanently in a high / soft state , but observations with the _ rossi x - ray timing explorer _ ( _ rxte _ ) have shown that lmc x-3 changes to a low / hard state during periods of low count rate ( wilms et al .
the existence of state transitions in lmc x-3 has now been confirmed by further _ rxte _ observations ( boyd et al . 2000 ) and _ xmm - newton _ observations ( wu et al . 2001 ) .
wilms et al .
( 2001 ) argue that the long term variability and consequent state changes could be due to an accretion disk wind - driven limit cycle ( shields et al .
1986 ) . because of the small column density in the direction of lmc x-3 , it is the best persistent bhc in which to study the soft x - ray spectrum at a variety of spectral states .
the soft x - ray spectral region is particularly important for understanding the accretion and mass transfer processes , because winds from the accretion disk , or from the companion star will imprint the soft x - ray spectrum with emission and absorption lines and edges .
furthermore , the standard paradigm for the thermal component seen in the high and intermediate states of bhcs ( namely that it is the multi - temperature blackbody spectrum of an optically thick , gemoetrically thin , accretion disk ) has been poorly tested at soft x - ray energies .
this is because almost all instruments on the previous generation of x - ray observatories ( i.e. before the launches of _ chandra _ and _ xmm - newton _ ) had poor spectral resolution below 2 kev , leading to degeneracies between different spectral shapes and different amounts of absorption in spectral modelling .
in this paper we present soft x - ray spectra of lmc x-3 from the _ xmm - newton _ reflection grating spectrometer ( rgs ) .
the spectra , which have a resolution of @xmath470 m , cover the full range of observed spectral states in this source .
table [ tab : observations ] lists the _ xmm - newton _ observations of lmc x-3 that will be used in this analysis , and in fig .
[ fig : asm ] the observation times are marked on the _ rxte _ all - sky monitor lightcurve for lmc x-3 during 2000 .
lmc x-3 was also observed by _ xmm - newton _ in revolution 0028 on the 2nd february 2000 , but we have excluded this observation from the present analysis because we are unable to produce a satisfactory effective area calibration for rev .
0028 , in which the source is off - axis .
over the ten months spanned by these observations lmc x-3 varied by more than a factor of 1000 in rgs count rate .
lcccc date&xmm&instrument&exposure@xmath5&count rate@xmath5 + ( 2000)&orbit&&(ks)&(count s@xmath6 ) + 07 feb & 0030&rgs1 & 12.1&7.1 + 08 mar & 0045&rgs1 + 2&20.8&5.9 + 19 apr & 0066&rgs1 + 2&44.3&0.013 + 09 jun & 0092&rgs1 + 2&77.5&2.2 + 24 nov & 0176&rgs1 + 2&20.9&18.0 + & & & & + + the data were initially processed using the public release _ xmm - newton _ sas version 5.2 .
first and second order spectra from both rgs were extracted ( except for revolution 0030 in which only rgs1 was operating ) and response matrices were generated using the latest version of the sas task rgsrmfgen ( corresponding to sas version 5.3.3 ) . to correct for residual artifacts in the effective area calibration ( see den herder 2002 ) , the effective area of each response matrix
was then divided by the ratio of a powerlaw + galactic column fit to the _ xmm - newton _ rev 0084 rgs spectra of the continuum source mrk 421 ( brinkmann et al .
lmc x-3 was on - axis for all of the observations listed in table [ tab : observations ] , as was mrk 421 in revolution 0084 .
the data / model ratio of the mrk 421 rgs spectrum is shown in fig .
[ fig : corr_factor ] ; for most of the rgs wavelength range the effective area correction is only a few percent .
the statistical errors on effective area correction were added in quadrature to the statistical errors on the data .
finally , for each revolution , spectra and response matrices were resampled and coadded to produce a single spectrum per observation . to improve signal to noise , spectra from the february , march , june and november observations were grouped by a factor of 3 ,
resulting in spectra with @xmath7 channels @xmath8 m wide , well sampled with respect to the rgs resolution of @xmath9 m fwhm .
the flux level of the april spectrum is extremely low , neccessitating very heavy binning before a respectable signal to noise ratio is achieved ; we have grouped this spectrum by a factor of 64 .
the resultant spectra were then analysed using the xspec software package version 11.2.0 ( arnaud 1996 ) .
[ fig : allspec ] shows the flux calibrated spectra from all five observations . the four brightest spectra
are shown at full resolution in figs [ fig:0030spec ] - [ fig:0176spec ] .
the spectra appear to be continuum dominated without prominent absorption or emission lines ; this is particularly clear in the 24th november spectrum , which has the highest signal to noise ratio and shows no significant spectral features except for the interstellar o i edge at 23 .
the degeneracy between absorption and the intrinsic spectral shape has been a major source of uncertainty in attempts to model relatively low resolution soft x - ray spectra of bhcs ( e.g. haardt et al . 2001 ) . the excellent combination of sensitivity and resolution afforded by the _ xmm - newton _ rgs allows individual absorption edges and lines to be measured more or less independently of the underlying continuum shape , allowing a far more reliable determination of the amount of soft x - ray absorption .
cold material ( either interstellar or local to the source ) imprints a well known sequence of absorption edges on x - ray spectra . for cosmic abundances
, oxygen produces by far the strongest absorption edge in the rgs energy range , at 23 , and inspection of fig .
[ fig : allspec ] shows that this is consistently the strongest feature observed in the rgs spectra of lmc x-3 .
we have therefore used the depth of the 23 o i k edge to determine the cold absorbing column for each of the spectra . by studying a small spectral region ( 3 to either side of the edge )
we ensure that our column density determination is essentially independent of the overall continuum shape .
the confidence intervals for the position of the edge and its optical depth are shown in fig .
[ fig : pocontedge ] for all the observations excluding the very faint april observation .
we have also fit the four observations simultaneously with a single set of edge parameters but with the continuum slope and normalisation allowed to vary independently for each observation .
the best fit value for the simultaneous fit is shown as a cross on the confidence intervals in fig .
[ fig : pocontedge ] . the edge parameters derived from each of the observations independently are consistent with the best fit edge parameters from the simultaneous fit
; there is no significant change in the column density of cold material from one observation to the next .
the wavelength of the edge from the simultaneous fit is @xmath10 where the confidence interval is 95% .
this is consistent with the wavelength of the edge calculated by verner et al .
( 1996 ) , which is the edge wavelength used in the xspec ` tbabs ' model ( wilms , allen & mccray 2000 ) . to determine the equivalent h i column density we substituted the ` tbabs ' model for the edge and performed another simultaneous fit to the four spectra .
this results in @xmath11 @xmath1 , which is consistent with the h i column density in the direction of lmc x-3 as determined from radio observations ( @xmath12 @xmath1 ; nowak et al .
2001 ) .
although the column density of cold absorbing material is consistent with the interstellar medium of the galaxy , absorbing material associated with a wind from the accretion disk or the secondary star may be highly ionized by the intense x - ray and ultraviolet radiation .
absorption and emission lines from warm , ionized material have now been detected in the soft x - ray band from a number of agn and x - ray binaries ( e.g. kaastra et al .
2000 , brandt & schultz 2000 ) .
there are no obvious emission or absorption lines present in the rgs spectra of lmc x-3 ( figs [ fig:0030spec ] , [ fig:0045spec ] , [ fig:0092spec ] and [ fig:0176spec ] ) and so we have attempted a sensitive search for their presence .
the interstellar medium and the local - group intergalactic medium have imprinted absorption lines of o i at 23.52 and o vii at 21.60 on the spectrum of mrk 421 ( paerels et al .
2003 ) , which we have used to correct residual artifacts in the effective area calibration .
therefore we performed the absorption line search on spectra which have no effective area correction .
first , we calculated the transmission through the interstellar medium as a function of energy using ` tbabs ' with @xmath13 @xmath1 .
we then corrected the spectra for interstellar absorption by dividing the effective area of the response matrix by the transmission .
we then produced continuum spectra by smoothing our rgs spectra with a sliding box 30 pixels ( @xmath3 ) wide .
we rejected any points more than 2@xmath14 deviant from the resulting smooth curve ( i.e. potential emission / absorption features ) , and repeated the smoothing .
this smoothing / rejection process was iterated several times to ensure that even marginally significant discrete features do not influence its shape .
we then performed a grid search for emission and absorption lines where the grid was defined in terms of central wavelength , fwhm and equivalent width .
the grid values of central wavelength were chosen to oversample the wavelength bins of the data , fwhm was stepped between 0 and 1000 km s@xmath6 , with a stepsize of 200 km s@xmath6 ( well sampled at the rgs resolution ) , and the stepsize of the equivalent width was matched to the statistical uncertainty on the data .
at each point in the grid , we formed a model spectrum by adding a gaussian emission / absorption line to the smooth continuum spectrum , and we computed the goodness of fit of this model to the real ( unsmoothed ) spectrum using the poisson errors on the data points and the @xmath15 statistic . the significance of any features , and limits on the equivalent width were determined using @xmath16 .
since we search for lines in either emission or absorption , we expect @xmath16 to give an approximate measure of the significance of any lines ( see protassov et al .
the grid search was performed using our own software which propogates the model through the response matrix in an identical fashion to xspec .
the results of this line search are shown in fig .
[ fig : ewlimits ] .
there are @xmath41000 channels in each spectrum , and therefore we expect @xmath3 spurious feature at the @xmath17 level in each spectrum .
the spectra from the 7th february and 8th march each contain one @xmath18 significant absorption feature , at 36.18@xmath19 and 32.61@xmath20respectively .
these do not correspond to the wavelengths of strong transitions in abundant elements , and are therefore probably spurious . in the spectrum from 9th june
, there are three @xmath18 significant absorption features , at 17.86@xmath19 , 21.62@xmath21 and 37.38@xmath20 , and one @xmath22 significant feature at 23.51@xmath23 .
the features at 17.86 and 37.38 do not correspond to the wavelengths of any expected absorption line and are again probably spurious .
the 21.62 feature has an equivalent width of 19@xmath24m and is consistent with the @xmath25 transition in o vii , while the 23.51 feature is the @xmath25 transition of o i. in the spectrum from 24th november , the most significant feature ( @xmath26 ) is again the @xmath25 transition in o i. in addition , there are three @xmath18 and one @xmath27 significant absorption features in this spectrum .
two of the @xmath18 features , at 14.94@xmath28 and 31.30@xmath29 do not correspond to strong transitions in abundant elements and are therefore probably spurious .
the other @xmath18 line , at @xmath30 , corresponds to o viii ly@xmath31 , while the @xmath27 line is found at @xmath32 , and corresponds to o vii he-@xmath31 .
the 18.93and 21.58 have equivalent widths of @xmath33m and @xmath34mrespectively . the o vii lines detected on the 9th june and 24th november have similar equivalent widths of @xmath420m , which implies an o vii column density between 3@xmath35 @xmath1 and 5@xmath36 @xmath1 .
intervening material associated with the halo of our galaxy is likely to make a significant contribution to ( and may dominate ) this line , because galactic o vii absorption is seen with a similar equivalent width in the blazars observed during rgs calibration ( paerels et al .
it is notable that if we apply the mrk 421 effective area correction before performing the line search on lmc x-3 , the ovii absorption lines are not detected in any of the spectra .
the equivalent width of the o viii ly@xmath31 absorption observed on the 24th november implies an o viii column density of between 4@xmath35 @xmath1 and 1.4@xmath37 @xmath1 .
this absorption line is seen in two out of three rgs calibration blazars ( paerels et al . 2003 ) with a similar equivalent width , and so the line in lmc x-3 may or may not arise in the galactic halo . the observed wavelength is consistent with an outflow from lmc x-3 , but is not well enough constrained to rule out a galactic origin .
we can place very firm upper limits on the column density of an ionized wind by using the 3@xmath14 upper limits on the equivalent widths of @xmath25 absorption lines from o ii
o viii obtained from fig .
[ fig : ewlimits ] and assuming that there is no interstellar component to these lines .
the spectra from 9th june and 24th november both have sufficient signal to noise ratios to provide useful limits .
these limits are given in table [ tab : ewlimits ] , along with the limits on column density assuming a velocity width @xmath38 @xmath39 .
the inferred total column density for o ii
o viii is @xmath40 on the 9th june and @xmath41 on the 24th november .
.3@xmath14 upper limits to equivalent widths of o ii o viii @xmath25 absorption lines .
the upper limits to column density have been calculated assuming velocity width @xmath38 @xmath39 . [
cols="<,^,^,^,^,^ " , ] having established that the soft x - ray spectrum of lmc x-3 is essentially free of discrete emission and absorption features except for the interstellar absorption , we now consider the shape of the continuum .
for all the fits performed in this section we include a cold absorbing column fixed at our measured value of @xmath42 @xmath1 , and all the results are tabulated in table [ tab : fits ] .
we started by fitting a disk - blackbody model to our spectra . only the very faint spectrum from 19th april
is acceptably fitted by this model , which falls off too rapidly below 10 to fit the other spectra . at shorter wavelengths ,
bhc spectra typically require a powerlaw component in addition to ( or instead of ) a disk blackbody ( e.g. ebisawa et al . 1996 ) .
we therefore added a powerlaw to the model , with the photon index constrained to lie within the range of photon indices seen at higher energy in lmc x-3 : @xmath43 ( wilms et al .
2001 , wu et al .
this resulted in statistically acceptable fits to the february , march and april observations , but not to the higher signal to noise june and november spectra . in all cases
the best fit photon index converged to the minimum of the acceptable range , indicating that the powerlaw produces too many photons at long wavelength , where the data are already quite well reproduced by the disk - blackbody component .
therefore we substituted the powerlaw component for compton scattered emission from a hot corona , making use of the thcompdsptz / relrepr.html ] model ( ycki , done & smith 2001 ) .
this model upscatters photons to produce a power - law shape ; the seed photons for the compton scattered component were assumed to come from the disk blackbody , and the electron temperature of the scatterer was fixed to 100 kev ( the rgs energy range is too limited for the fit to be sensitive to this parameter ) .
this model produced better fits to all the spectra , and all of the fits are acceptable at the 3@xmath14 confidence level ; fig .
[ fig : diskbbcomp ] shows rgs spectra compared to the best fit model components .
lcccccc + & & & & & & + observation & @xmath44 & flux ( disk ) & - & - & @xmath45 & prob + & ( kev ) & ( @xmath46 erg s@xmath6 @xmath1 ) & & & & + & & & & & & + 07 feb & @xmath47 & @xmath48 & - & - & 1169/986 & @xmath49 + 08 mar & @xmath50 & @xmath51 & - & - & 1161/997 & @xmath52 + 19 apr & @xmath53 & @xmath54 & - & - & 15/16 & 0.53 + 09 jun & @xmath55 & @xmath56 & - & - & 2872/998 & 0 + 24 nov & @xmath57 & @xmath58 & - & - & 1344/1001 & @xmath59 + & & & & & & + + & & & & & & + observation & @xmath44 & flux ( disk ) & @xmath60 & flux ( powerlaw ) & @xmath45 & prob + & ( kev ) & ( @xmath46 erg s@xmath6 @xmath1 ) & & ( @xmath46 erg s@xmath6 @xmath1 ) & & + & & & & & & + 07 feb & @xmath61 & @xmath62 & @xmath63 & @xmath64 & 1045/984 & 0.085 + 08 mar & @xmath65 & @xmath66 & @xmath67 & @xmath68 & 954/995 & 0.82 + 19 apr & @xmath69 & @xmath70 & @xmath71 & @xmath72 & 13/14 & 0.54 + 09 jun & @xmath73 & @xmath74 & @xmath75 & @xmath76 & 1132/996 & @xmath77 + 24 nov & @xmath78 & @xmath79 & @xmath80 & @xmath81 & 1210/999 & @xmath82 + & & & & & & + + & & & & & & + observation & @xmath44 & flux ( disk ) & @xmath60 & flux ( comptonised ) & @xmath45 & prob + & ( kev ) & ( @xmath46 erg s@xmath6 @xmath1 ) & & ( @xmath46 erg s@xmath6 @xmath1 ) & & + & & & & & & + 07 feb & @xmath83 & @xmath84 & @xmath85 & @xmath86 & 1024/984 & 0.18 + 08 mar & @xmath87 & @xmath88 & @xmath89 & @xmath90 & 902/995 & 0.98 + 19 apr & @xmath91 & @xmath92 & @xmath71 & @xmath93 & 12/14 & 0.59 + 09 jun & @xmath94 & @xmath95 & @xmath96 & @xmath97 & 1079/996 & 0.034 + 24 nov & @xmath98 & @xmath99 & @xmath100 & @xmath101 & 1127/999 & @xmath102 + & & & & & & +
in the preliminary analysis of _ xmm - newton _ epic and rgs spectra , wu et al . ( 2001 ) argued that mass transfer mainly occurs in lmc x-3 by roche lobe overflow rather than via a strong stellar wind because the line of sight absorption is small ( @xmath103 @xmath1 ) . with the more detailed analysis presented here we can rule out accretion from a stellar wind with much higher confidence . for a continuous spherical wind leaving a body of mass @xmath104 and radius @xmath105 at the escape velocity
, the hydrogen column density @xmath106 is related to the mass - loss rate by : @xmath107 where @xmath108 is the mass of a proton .
this expression can be used to obtain a conservative upper limit to the wind mass - loss rate , because if we assumed a @xmath109-law velocity profile ( @xmath110^{\beta } , r > 1.04 r_{*}$ ] ) , as used to model the line - driven winds from hot stars ( pauldrach , puls & kudritzki 1986 ) , instead of a constant velocity wind , the mass - loss rate for a given column density would be lower than that given in equation ( [ eq : massloss ] ) .
in section [ sec : cold ] we obtained an upper limit of @xmath111 @xmath1 for the hydrogen column density intrinsic to lmc x-3 , based on the neutral o edge and assuming solar metallicity . if we instead assume a metal abundance of 0.4 @xmath112 solar as expected in the lmc ( caputo et al .
1999 ) , we find an upper limit of @xmath113 @xmath1 . we can also estimate the hydrogen column density associated with o ii
o viii using the limits obtained in section [ sec : ionized ] from the 24th november spectrum .
assuming 0.4 @xmath112 solar abundance , we obtain @xmath114 @xmath1 . summing these two upper limits to obtain the total hydrogen column density associated with o i
o viii , we find @xmath115 @xmath1 .
assuming that the companion star has mass @xmath116 and radius @xmath117 ( soria et al .
2001 ) and substituting into equation ( [ eq : massloss ] ) , we find that @xmath118 g s@xmath6 .
this is insufficient to power the x - ray emission except when lmc x-3 is in its very weakest state ( e.g. april 2000 ) .
however , we have so far neglected wind material in which o is completely ionized .
when the luminosity of the system is as high as @xmath119 erg s@xmath6 , the ionization parameter ( @xmath120 ) at the base of a homogeneous , constant - velocity wind from the secondary star is @xmath121 where @xmath122 is luminosity in erg s@xmath6 , @xmath123 is the ion density in @xmath124 , and @xmath125 is the binary separation , @xmath126 cm .
thus for mass - loss rates @xmath127 g s@xmath6 we expect @xmath128 erg cm s@xmath6 , implying that the majority of o in the wind is completely ionized ( kallman & mccray 1982 ) , and therefore undetectable by x - ray absorption line spectroscopy . for a wind with a @xmath109-law velocity profile
the wind is much denser near the base , but at a distance of @xmath129 from the stellar photosphere , when the wind has only achieved about a fifth of its terminal velocity ( assuming @xmath130 ) , @xmath131 reaches @xmath132 .
a stellar wind can not be driven by line - pressure at such a high ionization state , because the ions that provide most of the acceleration ( abbot 1982 ) will not be present ( mccray & hatchett 1975 ) . on the other hand
, we can rule out a mass - loss rate of @xmath133 g s@xmath6 , because the column density would be high enough , ( and @xmath131 low enough ) that absorption lines from o would be very apparent in the rgs spectra .
thus the lack of absorption lines in the rgs spectra implies that a line - driven stellar wind is not the primary source of accretion material for lmc x-3 except , possibly , when it is in a faint state . by comparing the variability properties of lmc x-3 to lmc x-1 , nowak et al .
( 2001 ) also conclude that lmc x-3 is not accreting from a stellar wind .
roche lobe overflow is therefore the most likely mode of mass transfer in lmc x-3 . in order to explain the state changes observed in lmc x-3 along with the long term @xmath134 day periodic variability , wilms et al .
( 2001 ) propose that the accretion rate could be controlled by an accretion disk wind - driven limit cycle as discussed by shields et al .
the optical photometric variability presented by brocksopp et al .
( 2001 ) is also consistent with this hypothesis . in this scenario , the central x - ray source in lmc x-3 drives a compton heated wind from the accretion disk that is sufficiently strong that it interrupts its flow of fuel .
for such an instability to take place , the mass loss in the wind must exceed the accretion rate by a considerable factor ( e.g. @xmath135 , shields et al .
1986 ) . at distances from the central source
greater than the outer radius of the accretion disk , the compton heated wind is approximately spherical ( begelman & mckee 1983 ) and so we can use equation [ eq : massloss ] to relate the mass loss rate of an accretion disk wind to the column density . taking the radius of the accretion disk to be @xmath126 cm , and the mass of the primary to be @xmath136 ( soria et al . 2001 ) , we find that an accretion disk wind capable of driving the long - term variability will have @xmath137 @xmath1 when the luminosity of the source is @xmath138 erg s@xmath6 .
if we assume a wind temperature of @xmath139 k ( begelman , mckee & shields 1983 ) , more than @xmath140 of o is in the form of o ix , and the column density of o viii can therefore be @xmath141 @xmath1 , consistent with the upper limits given in table [ tab : ewlimits ] .
thus we can not rule out that the periodic variability in lmc x-3 is driven by an accretion disk wind .
however the column density in any accretion disk wind must be small enough that that it is transparent to scattering : thermal emission from a wind with column density @xmath142 @xmath1 would dominate the x - ray luminosity of the system , giving rise to prominent k@xmath31 lines from fe xxv and fe xxvi .
this is not observed in lmc x-3 ( wu et al .
thus even if there is an accretion disk wind , it imparts no significant features or distortions on the x - ray spectrum of lmc x-3 .
the resolution and throughput of the rgs coupled with the lack of x - ray absorption in lmc x-3 provide a clearer picture of the soft x - ray continuum spectral shape than has ever been possible before . excluding the very faint april spectrum ,
the spectral fitting performed in section [ sec : model ] demonstrates that a single disk blackbody is unable to describe the 0.3 - 2 kev spectral shape even when lmc x-3 is in the high / soft state , mainly because the model falls off too fast at short wavelength . adding a powerlaw component results in
much improved ( though for the 24th november still unacceptable ) fits . however , the fitted powerlaw slopes consistently bottom out at the hardest allowed values in the fitting ( @xmath143 ) and are considerably harder than the powerlaw slopes found over the 0.3 - 12 kev energy range with _ xmm - newton _ epic ( wu et al . 2001 ) , even when the epic and rgs data were obtained simultaneously . a visual inspection of fig .
[ fig : allspec ] offers some insight into the reason for the discrepant powerlaw parameters : longward of 23 the spectra have almost the same shape , running approximately straight and parallel as expected from a disk blackbody with a varying inner temperature .
however , a powerlaw , even with a photon index as hard as @xmath144 , has a softer spectral shape than a disk blackbody longward of the peak .
thus for a powerlaw with @xmath145 to make a significant contribution at @xmath146 it would have to make a larger contribution at longer wavelength , and therefore the spectra should deviate significantly from the disk blackbody shape at @xmath147 .
therefore the straight parallel spectra at @xmath147 , and the low best - fit photon indices , imply that the component which is observed to have a powerlaw shape at @xmath148 ( i.e. at e @xmath149 kev ) does not continue with the same powerlaw shape through the soft x - ray band .
the powerlaw component is normally assumed to be the compton upscattered emission from a hot corona ( sunyaev & titarchuk 1980 ) .
the compton scattered spectrum is expected to deviate from a powerlaw shape close to the energy of the seed photons .
thus the absence of the powerlaw component in the soft x - ray is consistent with the seed photons for the comptonisation coming from the disk blackbody component .
this is predicted by successful models for black hole binary spectral states ( e.g. haardt & maraschi 1991 , haardt & maraschi 1993 , chakrabarti & titarchuk 1995 , haardt et al .
1997 ) , in which the temperature and optical depth of the corona depend on the amount of cooling via compton scattering off soft x - ray photons from the disk .
this result is borne out by the improved fits which are obtained when the powerlaw is exchanged for a comptonised disk blackbody .
however , even with this model , the photon index which describes the comptonised spectrum towards shorter wavelengths is lower than the photon indices obtained over a larger energy range using epic . furthermore , for the brightest spectra the soft x - ray flux from the comptonised component dominates the flux directly emerging from the disk .
this suggests some further complexity in the spectral shape at the short wavelength end of the spectra ( the disk blackbody and comptonised components have a similar spectral shape at the long wavelength end ) .
one possible explanation could be that in the soft state , the soft component in lmc x-3 falls off slower than a disk blackbody towards short wavelengths due to electon scattering near the surface of the disk .
such a broadening of the soft component spectral shape has been observed in the soft x - ray transients ( e.g. gs 2000 + 25 , gro 1655@xmath15040 ) studied by ycki et al .
( 2001 ) using _ ginga _ and _ rxte _ spectra .
however , the rgs does not extend to short enough wavelengths for a proper investigation of this , which we defer to a future study using the epic instruments which cover a wider energy range .
we have presented _ xmm - newton _ rgs spectra of lmc x-3 covering the full range of spectral states observed in this source ( more than 3 orders of magnitude in rgs count rate ) .
the spectra are completely dominated by continuum emission , and are devoid of notable spectral features except for the interstellar o i edge .
the depth of the edge is consistent between the different observations , and is consistent with the galactic interstellar column density of @xmath151 in the direction of lmc x-3 .
the upper limit to the edge depth implies @xmath152 associated with neutral o , intrinsic to lmc x-3 .
upper limits on the equivalent widths of @xmath25 resonant absorption lines from o ii
o viii imply @xmath153 for gas associated with partially ionized o. this implies that a line - driven stellar wind off the companion star can not supply a significant fraction of the accreting material except during the very low luminosity states .
therefore the majority of accretion in lmc x-3 probably takes place through roche lobe overflow .
a multi temperature disk blackbody is a good representation of the soft x - ray spectrum of lmc x-3 longward of the disk blackbody peak .
thus the powerlaw component which dominates at higher energies does not continue through the soft x - ray range with a powerlaw shape .
this suggests that the disk blackbody component supplies the seed photons which are compton upscattered to form the powerlaw at higher energies , as in conventional models for bhc spectral states .
this work is based on observations obtained with _ xmm - newton _ , an esa science mission with instruments and contributions directly funded by esa member states and the usa ( nasa ) .
this work makes use of results provided by the asm / rxte teams at mit and at the rxte sof and gof at nasa s gsfc .
den herder j.w . , 2003 , proceedings of `` high resolution x - ray spectroscopy with xmm - newton and chandra '' , ed . g. branduardi - raymont , available at http://www.mssl.ucl.ac.uk/@xmath4gbr/rgs_workshop/papers/ paerels f. , rasmussen a. , kahn s.m . , den herder j.w . ,
de vries c. , 2003 , proceedings of the workshop : `` xeus - studying the evolution of the hot universe '' , eds . g. hasinger , th .
boller , and a. parmar , mpe report 281 , 57 | we present soft x - ray spectroscopy of the black hole binary lmc x-3 from the _ xmm - newton _ reflection grating spectrometer .
the observations span the full range of spectral states seen in lmc x-3 .
the spectra are completely dominated by continuum emission , and the neutral absorbing column measured from the o i edge ( @xmath0 @xmath1 ) is consistent with the galactic interstellar column density towards lmc x-3 .
we find no evidence for variability of the neutral absorbing column .
we also constrain the ionized column density using the upper limits to the equivalent width of the o ii
o viii k - shell resonance lines : we find that the equivalent hydrogen column density of gas in which o is partially ionized is @xmath2 @xmath1 . from this upper limit
we can rule out a line driven stellar wind as the power source for the x - ray emission of lmc x-3except when it is faint . at wavelengths longward of the peak emission the spectral shape is well described by a multi - temperature disk - blackbody spectrum ; the powerlaw component which dominates at shorter wavelengths does not continue longward of the disk - blackbody peak .
this implies that the multi - temperature disk - blackbody component supplies the seed photons which are compton upscattered in the hot corona , consistent with the standard paradigm for black hole x - ray binary spectral states .
accretion , accretion disks black hole physics
galaxies : seyfert binaries : general stars : early type stars : winds |
numerical techniques and , in particular , computer simulations are by now firmly established as a third pillar of the scientific method , complementing the pillars of experiment ( or observation ) and mathematical theory , both of which were erected already at the birth of modern science in the scientific revolution . while in the early times @xcite simulation studies were not quite competitive with analytical calculations in condensed matter physics and quantum field theory ( usually based on perturbative and variational approaches ) nor were their outcomes adequate for comparison with experimental results ( usually due to limited length or time scales ) , this situation has dramatically changed over the past decades .
this very successful race to catch up was fueled by a combination of two factors : a continuous , sometimes revolutionary refinement of the numerical toolbox , for instance through the invention of cluster algorithms @xcite , reweighting @xcite or generalized - ensemble techniques @xcite in the field of monte carlo simulations , and the impressive increase in generally available computational power , which has followed an exponential form known as moore s law for the past forty years . at any time
, however , there has been no shortage of fascinating scientific problems whose computational study required resources at or beyond the cutting edge of the available technology .
this has led scientists to regularly use the latest commodity hardware as soon as it became available , but has also motivated the design and construction of a number of special purpose machines such as , e.g. , the cluster processor @xcite and janus @xcite for the simulation of spin systems or a series of initiatives for qcd calculations with its recent addition of the qpace architecture @xcite .
it has been true for the last couple of generations of graphics processing units ( gpus ) that their theoretical peak performance , in particular for single - precision floating point operations ( flop / s ) , significantly exceeds that of the corresponding x86 based cpus available at the same time ( as of this writing up to around 100 gflop / s in cpus vs. up to 5 tflop / s in gpus ) .
it is therefore natural that scientists and , increasingly , also programmers of commercial applications other than computer games , have recently started to investigate the possible value of gpu based computing for their purposes ; for scientific applications see , e.g. , refs .
apart from their mere peak performance , systems based on gpus or other highly parallel architectures such as the cell processor @xcite might outperform current cpu based systems also in terms of their efficiency , i.e. , in terms of flop / s per watt or flop / s per euro and thus might also contribute to the advancement of `` green computing . ''
the low prices and convenient over - the - counter availability of gpu devices clearly make for advantages as compared to custom - built special - purpose machines , for which many man - months or years need to be invested for the design , realization and programming of devices . the relative increase in peak performance of gpus versus cpus
is achieved at the expense of flexibility , however : while today s cpus will perform pretty well on most of a large variety of computer codes and , in particular , in a multi - tasking environment where on - the - fly optimizations such as branch prediction are essential , gpus are optimized for the highly vectorized and parallelized floating - point computations typical in computer graphics applications .
a one - to - one translation of a typical code written for cpus to gpus will therefore , most often , _ not _ run faster and , instead , algorithms and parallelization and vectorization schemes must be chosen very carefully to tailor for the gpu architecture in order to harvest any performance increases .
the efficiency of such calculations in terms of flop / s per _ human _ time crucially depends on the availability of easily accessible programming environments for the devices at hand .
while in view of a lack of such supporting schemes early attempts of general purpose gpu ( gpgpu ) calculations still needed to encapsulate the computational entities of interest in opengl primitives @xcite , the situation changed dramatically with the advent of device - specific intermediate gpgpu language extensions such as ati stream and nvidia cuda @xcite . in the future ,
one hopes to become independent of specific devices with general parallel programming environments such as opencl @xcite .
classical spin systems have turned out to be extremely versatile models for a host of phenomena in statistical , condensed matter and high - energy physics , with applications ranging from critical phenomena @xcite over surface physics @xcite to qcd @xcite . a rather significant effort ,
therefore , has been invested over the years into optimized implementations of monte carlo simulations of spin models .
they hence appear to be particularly suited for an attempt to fathom the potential gain from switching to the gpu architecture .
additionally , there are a plethora of questions relating to classical spin models which , decades of research notwithstanding , are still awaiting a combination of an increase in available computational resources , better algorithms and clever techniques of data analysis to find generally satisfactory answers .
this applies , in particular , to disordered systems such as random - field models and spin glasses @xcite .
as will be discussed below , due to their short - ranged interactions and the typically simple lattice geometries such models appear to be near ideal problems for gpu calculations .
this applies , in particular , to the inherently local single spin - flip algorithms discussed here . for studying the critical behavior of models without disorder
, cluster algorithms will outperform any optimized implementation of a local spin - flip dynamics already for intermediate lattice sizes ; the possibilities for porting such approaches to gpu will be discussed elsewhere @xcite . for disordered systems , on the other hand , efficient cluster algorithms are ( with few exceptions @xcite ) not known . for them , instead , local spin - flip simulations combined with parallel tempering @xcite moves are the state of the art . when comparing gpu and cpu performance for the case of general purpose computational applications , it has become customary to benchmark different implementations in terms of the relative speed - up ( or slow - down ) of the gpu code versus the cpu implementation @xcite .
while such numbers make for good selling points for the company producing the `` better '' type of computational device , it should be clear that speed - ups , being a relative measure , will vary to a large degree depending on how much effort is invested in the optimization of the codes for the different architectures . to avoid this problem ,
the main focus is put here on the _ absolute _ performance of different implementations , measured for the present application mostly in the wall - clock time for performing a single spin flip , citing speed - up factors only as an additional illustration of the relative performance .
if relative measures of performance are given , the question arises of what type of cpu code to compare to , in particular , since with the advent of multi - core processors and vector extensions such as sse , cpus also offer a significant potential for optimizations .
i decided here to use serial cpu code , reasonably optimized on the level of a high - level programming language and the use of good compilers , as i feel that this comes closest to what is most often being used in actual simulation studies .
as regards the possible effects of further cpu optimizations , see also ref .
@xcite which , however , unfortunately does not cite any measures of absolute performance .
simulations of the ferromagnetic ising model using implementations on gpu have been discussed before @xcite . compared to these implementations , the current approach with the double checkerboard decomposition and multi - hit updates to be discussed below offers significant advantages .
other applications , such as the simulation of ferromagnetic , short - range heisenberg models , the simulation of ising spin glasses with asynchronous multi - spin coding or parallel tempering for lattice spin systems are considered here for the first time .
for some very recent discussions of further spin models see also refs .
@xcite .
the rest of this article is organized as follows . in sec .
[ sec : architecture ] , i give some necessary background on the architecture of the nvidia gpus used in this study and its consequences for algorithm design .
section [ sec : ising ] discusses , in some detail , the chosen implementation of a single - spin flip metropolis simulation of the two - dimensional ( 2d ) ising model and its performance as well as a generalization to the three - dimensional ( 3d ) model .
section [ sec : heisen ] is devoted to generalizations of these considerations to continuous - spin systems , exemplified in the 2d heisenberg model .
in secs . [
sec : spinglass ] and [ sec : tempering ] , applications to simulations of spin - glass models and the use of parallel - tempering updates are discussed .
finally , sec .
[ sec : concl ] contains my conclusions .
as indicated above , there are significant differences in the general design of cpu and gpu units @xcite .
cpus have been optimized over the past decades to be `` smart '' devices for the rather unpredictable computational tasks encountered in general - purpose computing .
current intel cpus feature about 800 million transistors on one die .
a rather small fraction of them is used for the alus ( arithmetic logic units ) doing actual computations , whereas most of the available transistors are devoted to flow control ( such as out - of - order execution and branch prediction ) and cache memory .
this structure is very successful in increasing the performance of _ single_-threaded applications , which still make up the vast majority of general computer codes .
in contrast , about 80 % of the 1.4 billion transistors on a gpu die of the nvidia gt200 series ( now superseded by the fermi architecture ) are alus .
gpus do not try to be `` clever '' , but they are extremely efficient at doing the same computational steps on different bits of a large data - set in parallel .
this is what makes them interesting for applications in scientific computing .
figure [ fig : hardware ] shows a schematic representation of the general architecture of a current gpu .
the chip contains a number of multiprocessors each composed of a number of parallel processing units .
the systems based on the gt200 architecture used in this study feature 30 multiprocessors of 8 processors each ( versus 15 multiprocessors 32 cores for the gtx 480 fermi card ) .
the systems come with a hierarchy of memory layers with different characteristics .
each processor is equipped with a number of registers which can be accessed with essentially no latency .
the processors in a multiprocessor unit have read / write access to a small amount of shared memory ( @xmath1 kb in the gt200 architecture and @xmath2 kb for fermi cards ) which is on - chip and can be accessed with latencies around a hundred times smaller than those for global memory . the large device or global memory ( up to 6 gb in current tesla systems ) is connected with read / write access to all multiprocessors .
latency for global memory accesses is of the order of 400600 clock cycles ( as compared to , e.g. , one clock cycle for a multiply or add operation ) .
the additional constant and texture memory areas are cached and designed for read - only access from the processing units , in which case they operate very fast with small latencies .
the memory of the host computer can not be accessed directly from within calculations on the gpu , such that all relevant data need to be copied into and out of the gpu device before and after the calculation , respectively . since the processing units of each multiprocessor are designed to perform identical calculations on different parts of a data set ,
flow control for this simd ( single instruction multiple data ) type of parallel computations is rather simple .
it is clear that this type of architecture is near ideal for the type of calculations typical for computer graphics , namely rendering a large number of triangles in a 3d scene or the large number of pixels in a 2d projection in parallel .
the organization of processing units and memory outlined in fig .
[ fig : hardware ] translates into a combination of two types of parallelism : the processing units inside of each multiprocessor work synchronously on the same data set ( vectorization ) , whereas different multiprocessors work truly independent of each other ( parallelization ) .
the corresponding programming model implemented in the cuda framework @xcite is outlined schematically in fig .
[ fig : gridblock ] : computations on gpu are encapsulated in functions ( called `` kernels '' ) which are compiled to the gpu instruction set and downloaded to the device .
they are executed in a two - level hierarchic set of parallel instances ( `` execution configuration '' ) called a `` grid '' of thread `` blocks '' .
each block can be thought of as being executed on a single multiprocessor unit .
its threads ( up to 512 for the gt200 architecture and 1024 for fermi cards ) access the same bank of shared memory concurrently .
ideally , each thread should execute exactly the same instructions , that is , branching points in the code should be reduced to a minimum .
the blocks of a grid ( up to @xmath3 ) are scheduled independently of each other and can only communicate via global memory accesses .
the threads within a block can make use of cheap synchronization barriers and communicate via the use of shared ( or global ) memory , avoiding race conditions via atomic operations implemented directly in hardware . on the contrary
, the independent blocks of a grid can not effectively communicate within a single kernel call .
if synchronization between blocks is required , consecutive calls of the same kernel are required , since termination of a kernel call enforces all block computations and pending memory writes to complete .
since the latency of global memory accesses is huge as compared to the expense of elementary arithmetic operations , many computational tasks on gpu will be limited by memory accesses , i.e. , the bandwidth of the memory subsystem .
it is therefore crucial for achieving good performance to ( a ) increase the number of arithmetic operations per memory access , and ( b ) optimize memory accesses by using shared memory and choosing appropriate memory access patterns .
the latter requirement in particular includes the adherence to the specific _ alignment _ conditions for the different types of memory and clever use of the _ coalescence _ phenomenon , which means that accesses to adjacent memory locations issued by the different threads in a block under certain conditions can be merged into a single request , thus greatly improving memory bandwidth . due to most typical computations being limited by memory accesses ,
it is important to use an execution configuration with a total number of threads ( typically several @xmath0 ) much larger than the total number of processing units ( @xmath4 for the tesla c1060 and @xmath5 for the gtx 480 ) .
if a thread block issues an access , e.g. , to global memory , the gpu s scheduler will suspend it for the number of cycles it takes to complete the memory accesses and , instead , execute another block of threads which has already finished reading or writing its data .
the good performance of these devices thus rests on the idea of latency hiding and transparent scalability through flexible thread scheduling @xcite .
the layout of the gpu architecture outlined above implies guidelines for the efficient implementation of computer simulation codes . along these lines
, a code for single - spin flip metropolis simulations of the ferromagnetic ising model is developed .
we consider a ferromagnetic , nearest - neighbor , zero - field ising model with hamiltonian @xmath6 on square and simple cubic lattices of edge length @xmath7 , using periodic boundary conditions . in the single - spin flip metropolis simulation scheme for this model
each step consists of randomly selecting a spin @xmath8 and proposing a spin flip @xmath9 , which is accepted according to the metropolis criterion @xcite @xmath10,\ ] ] where @xmath11 corresponds to the energy change incurred by the flip and @xmath12 denotes the inverse temperature .
it is straightforward to show that this update is ergodic , i.e. , all states of the system can be reached in a finite number of spin - flip attempts with non - zero probability , and satisfies the detailed balance condition ensuring convergence to the equilibrium boltzmann distribution . in practice , one usually walks through the lattice in a sequential fashion instead of picking spins at random , which requires less pseudo - random numbers and , generically , leads to somewhat faster convergence .
this updating scheme , in fact , violates detailed balance , but is consistent with the global balance condition which is sufficient to ensure convergence @xcite .
this point is of some importance for the gpu implementation developed here , and will be further discussed below in section [ sec : balance ] . to leverage the effect of the massively parallel architecture of gpu devices for simulations of spin models , in most cases
domain decompositions where the system is divided into a large number of largely independent sub - units are the only approach with satisfactory scaling properties as the number of processors or the size of the system is increased . for the case of lattice systems , the simplest scheme amounts to a coloring of lattice sites with the minimally feasible number of colors . here , i focus on bipartite graphs such as the square and ( hyper-)cubic lattices where two colors suffice , resulting in a ( generalized ) checkerboard decomposition @xcite .
generalizations to other cases are trivial . for performing a single spin - flip metropolis simulation of the ising model on gpu .
each of the @xmath13 big tiles is assigned as a thread block to a multiprocessor , whose individual processors work on one of the two sub - lattices of all @xmath14 sites of the tile in parallel.,scaledwidth=60.0% ] in such a scheme , each site of one sub - lattice can be updated independently of all others ( assuming nearest - neighbor interactions only ) , such that all of them can be treated in parallel followed by an analogous procedure for the second sub - lattice . for an implementation on gpu , the checkerboard decomposition needs to be mapped onto the structure of grid blocks and threads .
although , conceptually , a single thread block suffices to update one sub - lattice in parallel , the limitation to 512 threads per block for the gt200 architecture ( resp .
1024 threads for fermi ) enforces an additional block structure for systems of more than 1024 resp .
2048 spins . for the square lattice , in ref .
@xcite stripes with a height of two lattice spacings were assigned to independent thread blocks while performing all spin updates in global memory .
this has several disadvantages , however : ( a ) shared memory with its much lower latency is not used at all , ( b ) determining the indices of the four neighboring spins of each site requires costly conditionals due to the periodic boundary conditions , and ( c ) the system size is limited to @xmath15 for gt200 ( @xmath16 for fermi ) . here , instead , i suggest a more versatile and efficient approach leveraging the inherent power of the memory hierarchy intrinsic to gpus . to this end
a twofold hierarchy of checkerboard decompositions is used .
figure [ fig : checker ] illustrates this for the case of a square lattice : on the first level , the lattice is divided into @xmath17 big tiles in a checkerboard fashion .
these are then updated by individual thread blocks , where the individual threads exploit a second - level checkerboard decomposition of the @xmath18 lattice sites inside each tile . the size of big tiles is thereby chosen such that the spin configuration on the tile ( plus a boundary layer of spins ) fits into shared memory , such that the spin updates can then be performed entirely inside of this much faster memory area .
on the block level , through the checkerboard arrangement all tiles of one ( `` even '' ) sub - lattice can be updated concurrently before updating the other ( `` odd '' ) half of tiles . inside of each block , again
all sites of the finer sub - lattices are independent of each other and are therefore updated concurrently by the threads of a thread block . in summary
, the updating procedure looks as follows : 1 .
the updating kernel is launched with @xmath19 thread blocks assigned to treat the _ even _
tiles of the coarse checkerboard .
2 . the @xmath20 threads of each thread block cooperatively load the spin configuration of their tile plus a boundary layer into shared memory .
3 . the threads of each block perform a metropolis update of each _ even _ lattice site in their tile in parallel . 4 .
all threads of a block wait for the others to finish at a barrier synchronization point .
the threads of each block perform a metropolis update of each _ odd _ lattice site in their tile in parallel .
the threads of each block are again synchronized .
a second kernel is launched working on the @xmath19 _ odd _ tiles of the coarse checkerboard in the same fashion as for the even tiles .
the cooperative loading of each tile into shared memory is organized in a fashion to ensure _ coalescence _ of global memory accesses @xcite ,
i.e. , subsequent threads in each block access consecutive global memory locations wherever possible .
while it turns out to be beneficial to load tiles into shared memory already for a single update per spin before treating the second coarse sub - lattice due to an avoidance of bank conflicts in global memory as well as the suppressed need to check for periodic wrap - arounds in determining lattice neighbors , the ratio of arithmetic calculations to global memory accesses is still not very favorable .
this changes , however , if a multi - hit technique is applied , where the spins on each of the two coarse sub - lattices are updated several times before returning to the other half of the tiles . as is discussed below in sec .
[ sec : balance ] , this works well , in general , and only close to criticality somewhat increased autocorrelation times are incurred . by design ,
monte carlo simulations of spin systems require a large amount of pseudo - random numbers .
depending on implementation details , for a system as simple as the ising model , the time required for generating random numbers can dominate the total computational cost .
hence , the efficient implementation of random - number generators ( rngs ) in a massively parallel environment is a crucial requirement for the efficient use of the gpu architecture .
speed is not everything , however , and it is well known that the statistical deficiencies that no _ pseudo _ rng can entirely avoid can have rather dramatic consequences in terms of highly significant deviations of simulation results from the correct answers @xcite .
hence , some effort must be invested in the choice and implementation of an appropriate rng . from the highly parallel setup described above
, it is clear that each of the @xmath21 concurrent threads must be able to generate its stream of random numbers to a large degree independently of all others , since the bottleneck in any configuration with a centralized random - number production would severely impede scaling of the code to a large number of processing units .
as each of these ( sub-)sequences of random numbers are used in the same simulation , one needs to ensure that ( a ) either each thread generates different members of the same global sequence of numbers or ( b ) the sequences generated by different threads are at least statistically uncorrelated .
the simplest choice of rng is given by the linear congruential generator ( lcg ) of the form @xmath22 with @xmath23 .
the authors of ref .
@xcite used @xmath24 and @xmath25 , originally suggested in ref .
the period of this generator is small with @xmath26 @xcite . in a simulation of a @xmath27 ising system , for instance ,
this period is exhausted already after 256 sweeps . on theoretical grounds , it is argued that one actually should not use more than @xmath28 numbers out of such a sequence @xcite , which would render the sequence of available numbers very short indeed .
additionally , simple lcgs are known to exhibit strong correlations which can be revealed by plotting @xmath29-tuples of successive ( normalized ) numbers as points in @xmath30 , where it is found that , already for rather small @xmath29 , the points are confined to a sequence of hyperplanes instead of being uniformly distributed .
the choice @xmath31 has the advantage that the whole calculation can be implemented entirely in 32-bit integer arithmetic since on most modern architectures ( including gpus ) integer overflows are equivalent to taking a modulo operation . for such power of two moduli @xmath32
, however , the period of the less significant bits is even shorter than that of the more significant bits , such that the period of the @xmath29th least significant bit is only @xmath33 .
an advantage for the parallel calculations performed here is that one can easily skip ahead in the sequence , observing that @xmath34 where @xmath35 therefore , choosing @xmath36 equal to the number @xmath37 of threads , all threads can generate numbers out of the same global sequence ( [ eq : lcg ] ) concurrently .
an alternative setup , that can not guarantee the desired independence of the sequences associated to individual rng instances , however , starts from randomized initial seeds for each generator , without using any skip - ahead @xcite . to improve on the drawback of a short period
, one might consider moving to a generator with larger @xmath32 , for instance @xmath38 , where the modulo operation again can be implemented by overflows , this time for @xmath39-bit integers , a data type which is also available in the cuda framework .
as multiplier i chose @xmath40 with provably relatively good properties @xcite , where an odd offset , here @xmath41 , needs to be chosen to reach the maximal period of @xmath42 .
as for the @xmath43-bit case , this generator can be used in the parallel framework yielding numbers from a common sequence via skip - ahead , or as independent generators with initial randomized seeds , leading to overlapping sub - sequences . and @xmath44 , @xmath45 , while `` lcg32 '' and `` lcg64 '' correspond to the recursion with @xmath46 and @xmath47 , respectively .
all thread blocks use @xmath48 threads . ] for high - precision calculations , one normally would not want to rely on a simple lcg
. a significant number of generators with much better statistical properties has been developed in the past , for a review see ref .
@xcite . for our purposes
, however , most of them have severe drawbacks in that they are either quite slow ( such as , for instance , generators that combine a number of elementary rngs or , instead , drop a certain fraction of random numbers generated as in ranlux @xcite ) , that they use a large state to operate on ( for instance @xmath49 32-bit words for the popular mersenne twister @xcite ) , or that it is rather hard to ensure that sequences generated independently are to a reasonable degree uncorrelated ( as for most high - quality generators ) .
while a state larger than a few words is usually not a problem for serial implementations , in the framework of gpu computing where fast local ( i.e. , shared ) memory is very limited , it is impossible to use several hundred words per thread of a thread block only for random number generation .
a reasonable compromise could be given by generators of the ( generalized ) lagged fibonacci type with recursion @xmath50 which operate on a buffer of length @xmath51 with @xmath52 and have good properties for sufficiently large lags @xmath53 and @xmath54 and choosing @xmath55 @xcite . if @xmath56 , the maximal period is @xmath57 .
for an implementation in single precision arithmetic , i.e. , @xmath58 , and @xmath59 ( see below ) , this results in a rather large period @xmath60 .
the recursion ( [ eq : lagged_fibonacci ] ) can be implemented directly in floating - point arithmetic yielding uniform random numbers @xmath61 $ ] via @xmath62 for good quality one needs @xmath63 , leading to relatively large storage requirements , but here the generation of @xmath54 random numbers can be vectorized by the threads of a thread block .
therefore , the required ring buffer of length @xmath51 words is shared by the @xmath20 threads of a block , and hence the storage requirement in only @xmath64 words per thread .
if one chooses @xmath54 only slightly larger than the number of threads per block , only a few words per thread are consumed .
a number of good choices for the lags @xmath53 and @xmath54 are collected in ref .
the best performance is achieved for unity multipliers @xmath65 , at only moderately reduced quality of the resulting random numbers . in this setup ,
different thread blocks use different sequences .
prescriptions for ensuring disjunct sequences are described in ref .
@xcite . in view of the astronomic period , however , it appears safe to just seed the ring buffers of the generators for different blocks with an independent rng . the lcg generator with @xmath66 and @xmath67 as well as the generalized fibonacci generator were implemented and benchmarked on a tesla c1060 gpu .
the lcgs have been realized with @xmath43- and @xmath39-bit integers , respectively , using the multipliers listed above . for the fibonacci generator , i chose @xmath65 and lags @xmath44 and @xmath45 ( suitable for simulations with tile size up to @xmath68 ) .
the generators were test with a variable number of blocks of @xmath48 threads each ( where @xmath48 is the maximum number of threads allowed on the c1060 device ) .
all generator state information is kept in shared memory .
the results are displayed in fig .
[ fig : rng ] . the characteristic zig - zag pattern results from the commensurability ( or not ) of the number of blocks @xmath69 with the number of @xmath70 multiprocessors on the device , such that the performance is best if @xmath71 , @xmath72 , @xmath73 , @xmath74 and worst if @xmath75 , since in the latter case @xmath76 of the multiprocessors are idle once the last batch of blocks has been scheduled for execution .
the @xmath43-bit lcg is found to take at best @xmath77 ns per random number , whereas the @xmath39-bit lcg and the lagged fibonacci generator are about @xmath78 and @xmath79 times slower , respectively .
lt9t2t7 . & & & & + exact & 1.106079207 & 0 & 0.8616983594 & 0 + + lcg32 & 1.1060788(15 ) & -0.26 & 0.83286(45 ) & -63.45 + lcg64 & 1.1060801(17 ) & 0.49 & 0.86102(60 ) & -1.14 + fibonacci , @xmath80 & 1.1060789(17 ) & -0.18 & 0.86132(59 ) & -0.64 + + lcg32 & 1.0944121(14 ) & -8259.05 & 0.80316(48 ) & -121.05 + lcg32 , random & 1.1060775(18 ) & -0.97 & 0.86175(56 ) & 0.09 + lcg64 & 1.1061058(19 ) & 13.72 & 0.86179(67 ) & 0.14 + lcg64 , random & 1.1060803(18 ) & 0.62 & 0.86215(63 ) & 0.71 + fibonacci , @xmath80 & 1.1060890(15 ) & 6.43 & 0.86099(66 ) & -1.09 + fibonacci , @xmath81 & 1.1060800(19 ) & 0.40 & 0.86084(53 ) & -1.64 + [ tab : rng ] the quality of the streams of random numbers generated in this way was tested with the metropolis update simulation of the 2d ising model using the double checkerboard update .
the resulting estimates for the internal energy and specific heat of a system of edge length @xmath82 and inverse temperature @xmath83 from simulations employing a number of different generators are collected in tab .
[ tab : rng ] . for comparison , the exact results calculated according to ref .
@xcite are shown as well .
for the sequential updates on cpu , the dominant correlations and short period of the @xmath43-bit lcg lead to highly significant deviations at the chosen rather high level of statistics ( @xmath84 updates per spin ) , which are , however , only visible in the specific heat data . already
the @xmath39-bit lcg and likewise the fibonacci generator result in data completely consistent with the exact values at the given level of statistical accuracy .
the double checkerboard simulations on gpu , on the other hand , appear to be much more sensitive to correlations in the rngs . if all threads sample from the same sequence of random numbers using skipping according to eq . , the @xmath43-bit lcg produces astronomic and the @xmath39-bit lcg still sizable deviations from the exact result , cf . the data collected in tab [ tab : rng ] .
somewhat surprisingly , no significant deviations are produced from either lcg when seeding the lcg of each thread randomly and independently with another rng . in some sense
, this setup appears to mimic the effects of a combined generator @xcite : although all of the sequences traversed by individual threads have significant overlap with each other , the random shifts between different sequences are sufficient to cancel the correlation effects .
the fibonacci generators of eq .
, on the other hand , do not suffer from such problems , although deviations are again larger than for the sequential implementation , and one needs to go to a lag as large as @xmath59 to get fully satisfactory results for the system at hand .
note that the observed differences between cpu and gpu results are entirely due to the different order of spin updates in the sequential and checkerboard setups , and have nothing to do with shortcomings of the gpu architecture .
in particular , all the generators ( and variants ) implemented here on gpu in connection with the corresponding simulation codes guarantee completely reproducible results , i.e. , the sequence of generated spin configurations only depends on the initial seeding of the rngs and not the order in which different blocks or threads are executed . for general practical applications , the use of the lagged fibonacci generator appears to be a reasonable compromise in that it is only rather insignificantly slower than the @xmath39-bit lcg , but produces random numbers of substantially better quality .
the setup with randomly seeded independent lcgs per thread also yields satisfactory results ( and can be very fast ) , but in the absence of a deeper understanding of how correlations are destroyed in this setup , it can not be entirely recommended for high - precision studies of simple , discrete models such as the ising ferromagnet considered here . for systems with continuous degrees of freedom or with additional quenched disorder ,
however , such correlation effects are presumably less of a problem .
it is worthwhile to note that the checkerboard update , like the sequential update of lattice sites usually applied in a scalar computing context , does not satisfy detailed balance ( see , e.g. , ref .
@xcite ) , @xmath85 on the level of single spin flips since the probability of selecting a spin twice for updating in direct succession vanishes if the lattice sites are treated in a deterministic order .
the global balance condition , @xmath86 on the other hand , which is sufficient to ensure convergence of the chain @xcite , stays satisfied at all times .
the same holds true if the checkerboard update is applied in a multi - hit fashion with each coarse sub - lattice being updated @xmath87 times before moving on to the other sub - lattice .
although detailed balance is not required for sampling the equilibrium distribution , it can be recovered , if desired , e.g. , to ensure that certain dynamical properties hold , on the level of compound moves such as lattice sweeps .
denote , for instance , an update of the even sub - lattice with `` a '' and an update of the odd sub - lattice with `` b '' and a measurement time by `` m '' .
then , a sequence of the form @xmath88 satisfies detailed balance with respect to compound updates and is hence more symmetric than the naive multi - hit implementation .
ising model as a function of inverse temperature @xmath89 .
the gpu calculations were performed with different numbers @xmath29 of multi - hit updates , but at a constant total number of updates . ]
while the multi - hit approach with @xmath90 is a perfectly valid updating scheme , it is clear that very close to the critical point it will be somewhat slower than the @xmath91 version at decorrelating the system as soon as the correlation length @xmath92 exceeds the size @xmath93 of the tiles . assuming that @xmath94 , it takes @xmath95 sweeps to decorrelate the configuration of the tile or , equivalently , for information traveling diffusively to cross the tile . for @xmath96
, one expects multi - hit updates to become inefficient .
figure [ fig : autocorr ] shows the resulting integrated autocorrelation times @xcite @xmath97 of the internal energy of a @xmath98 ising system as a function of inverse temperature @xmath89 and comparing the cpu calculation with gpu simulations with multi - hit updates ranging from @xmath91 up to @xmath99 , using tiles of @xmath100 spins .
note that , in the same way as sequential spin updates in a scalar simulation have different autocorrelation times than purely random updates , the checkerboard ( i.e. , gpu ) way of updating the spins yields slightly different autocorrelation times than the sequential update on cpu .
this effect , however , appears to be weak .
apart from the general expected rounded divergence of autocorrelation times in the vicinity of the critical coupling @xmath101 , fig .
[ fig : autocorr ] clearly reveals the relative increase of @xmath97 as @xmath29 is increased . as a result ,
in the vicinity of @xmath102 there are two competing effects : increasing @xmath29 reduces the overhead of loading tiles from global memory and thus decreases spin - flip times , whereas at the same time it increases autocorrelation times , thus reducing sampling efficiency . as a result , the autocorrelation time in physical units , @xmath103 where @xmath104 denotes the real time for a single spin flip , such that @xmath105 corresponds to the real time for generating an effectively uncorrelated spin configuration , close to @xmath102 has a minimum at intermediate values of @xmath29 .
this is illustrated in fig .
[ fig : autocorr2 ] , again for a @xmath98 system with tile size @xmath68 , where it is seen that , in the vicinity of @xmath102 , multi - hit updates with @xmath106 appear to be optimal . for the @xmath98 ising system and double checkerboard updates on gpu for different values of the multi - hit parameter @xmath29 .
minimal @xmath107 corresponds to a minimum time of generating a statistically independent spin configuration . ] using the double checkerboard decomposition loading tiles into shared memory and the rngs discussed above , an optimized code was first constructed for the case of the square - lattice ising model .
the sample code , which has undergone significant further optimizations as compared to the version discussed in ref .
@xcite , can be downloaded from the author s web site @xcite .
multi - hit updates of the square - lattice ising model as a function of linear system size @xmath7 .
the lines correspond to different block sizes @xmath93 . ]
it is a well - known trick valid for discrete models to tabulate the exponential occurring in the acceptance criterion ( [ eq : metropolis ] ) in advance and use a look - up instead of calculating the exponential directly .
it appears natural to store this array in constant memory which is fast through caching @xcite .
this memory , however , is optimized for the different threads of a block synchronously accessing the _ same _ memory element , which is not the typical case here since different spins usually have different local interaction energies . it is found more efficient , therefore , to store this table in texture memory which is fast with different threads accessing different memory locations . while for the cpu code it is beneficial to save arithmetic cycles by drawing random numbers and looking up the value of the boltzmann factor only when @xmath108 , this is not the case on gpu , where such an approach results in thread divergence requiring the corresponding branch to be visited twice . for the code discussed here , i used a single variable of the high - level language to represent a spin .
performance differences depending on whether int ( 32 bit ) , short int ( 16 bit ) or char ( 8 bit ) variables are used are only very moderate .
the results discussed below are for int variables .
a multi - spin coded version is discussed below in sec .
[ sec : spinglass ] . .
gpu data are for the tesla c1060 device apart from the lowest curve which is for a gtx 480 card .
all data are for tile size @xmath109 and an inverse temperature @xmath83 close to , but off the critical coupling .
there is a weak temperature dependence of spin - flip times due to differences in the number of write operations . ]
the optimum choice of the size @xmath93 of tiles depends on system size and is determined by the relative resource utilization on the considered gpu device .
the corresponding measured spin - flip times for @xmath99 shown in fig .
[ fig : blocks ] are easily understood from the parameters of the tesla c1060 device at hand : tiles should be large enough to provide a sufficient number of threads in each block .
the optimum load is seen when each multiprocessor is assigned with the maximum number of eight thread blocks . since each multiprocessor can accommodate a maximum allowed number of @xmath110 threads , this results in @xmath111 threads per block .
this number is reached for tiles of size @xmath112 spins ( remember that each thread updates two sites ) .
this choice turns out to be optimal for lattice sizes @xmath113 . for smaller lattices , on the other hand
, @xmath68 does not provide enough blocks to load all 30 multiprocessors of the device which requires @xmath114 or @xmath115 for @xmath116 and @xmath117 for @xmath118 .
since @xmath119 is the smallest tile size considered , this is also optimal for the smallest lattice sizes @xmath120 and @xmath121 . for smaller values of @xmath29 ,
the performance of global memory accesses comes into play , which favors larger tile sizes due to the resulting improvements in coalescence , such that @xmath68 turns out to be optimal for any @xmath122 there .
test simulations for the fermi architecture on a gtx 480 card lead pretty much to the same conclusions : there , the maximum number of resident threads per multiprocessor has been increased to 1536 , but this increase is too small to lift the optimum tile size to @xmath123 such that , again , @xmath124 is optimal apart from simulations on the smallest lattices . .
] figure [ fig : speedup ] shows the performance data for the final code as compared to the scalar cpu implementation . as a reference cpu
i used an intel core 2 quad q9650 at 3.0 ghz , which is one of the fastest available cpus for single - threaded applications .
the performance of the latter is essentially independent of system size with about @xmath125 ns per spin flip .
the maximum performance on gpu is reached only for system sizes @xmath126 with about @xmath127 ns per spin flip for @xmath128 and @xmath129 ns per spin flip for @xmath99 , resulting in speed - up factors ranging from @xmath70 to @xmath130 .
the fermi architecture results in an extra speed - up of about @xmath131 compared to the tesla c1060 device , leading to a maximum performance of @xmath132 ns per spin flip at @xmath99 or an overall speedup factor of @xmath133 ( the extra speedup for @xmath91 is even slightly larger , probably due to the reduced overhead of kernel calls for the fermi architecture ) . comparing to the previous implementation in ref .
@xcite , one notes that , probably due to inappropriate cache alignment , the ( single - spin coded ) cpu code used there is rather inefficient , such that , e.g. , for a @xmath134 system the code used here is about @xmath135 times faster than that of ref .
@xcite and @xmath136 times faster than the results quoted in ref .
@xcite , leading to a rather unfavorable comparison for the cpu based systems in these works .
their multi - spin coded cpu version is less than a factor of two more efficient than the single - spin coded implementation used here ( see below in sec . [
sec : spinglass ] for a multi - spin coded version of the program used here ) .
the maximum performance of the gpu implementation of ref .
@xcite with @xmath137 ns on the tesla c1060 is about @xmath138 times less than the present @xmath128 code and a factor of @xmath139 slower at @xmath99 .
the significant speedup at @xmath128 is due to the combined effect of the avoidance of bank conflicts in computing the local interaction energy , the fact that periodic boundaries do not need to be taken account with expensive modulo operations once the tile including its boundary layer has been loaded into shared memory , and the efficient implementation of the metropolis criterion using texture fetches .
the multi - spin coded gpu code ( shared memory version ) of ref .
@xcite is about of the same performance as our single - spin coded version at @xmath128 . in a practical application measurement cycles
are required on top of the updates , of course .
energy measurements were here realized with tracking the local energy changes at each spin flip and combining them in a final parallel reduction step after the update . at @xmath99 , corresponding to only one measurement per one hundred lattice sweeps , the overhead is small , of the order of 10% , see the data collected in fig .
[ fig : speedup2 ] . at @xmath128 ,
the overhead is about 20% . in view of the potential problems with the quality of random numbers discussed above in sec .
[ sec : rng ] , it is worthwhile to check the influence of the different speeds of rngs discussed above on the overall performance of the ising simulation code . in a simulation as simple as the metropolis update of the ising model , a substantial fraction of time is spent for generating random numbers .
therefore , it is no surprise that it is found that the simulation is slower by a factor of @xmath140 when using the lcg64 instead of the lcg32 generator and by a factor of @xmath141 when using the lagged fibonacci generator ( @xmath99 , @xmath142 ) . in view of the good results for
the randomly initialized lcg32 reported in sec .
[ sec : rng ] , it is probably acceptable to use it for the considered model in view of its good performance . for other applications , however , this issue would need to be revisited .
the generalization of the gpu code from two to three dimensions is straightforward , and only the part responsible for the correct bookkeeping in loading a tile into shared memory becomes more tedious to formulate .
the range of allowable tile sizes is somewhat more restricted by the limitation of the available gpus to 512 threads per block ( resp .
1024 for fermi ) . with the current setup ,
therefore , the only tile sizes that can be reasonably considered are @xmath143 and @xmath144 , if one restricts oneself to powers of two in system as well as in tile sizes .
more general and , perhaps , versatile , setups could be reached by relaxing the power - of - two condition or considering tiles with non - unit aspect ratios , which might lead to additional slight increases in performance . for the sake of simplicity , however , i here refrain from introducing such complications . the timing results for the 3d code are summarized in fig .
[ fig : speedup3d ] .
the performance of the cpu code is essentially independent of lattice size within the considered range of @xmath145 up to @xmath146 , arriving at around @xmath147 ns per spin flip on the reference cpu used here , which is slower than the 2d code by a bit more than the factor of @xmath148 expected merely from the increased number of nearest neighbors to be included in energy calculations . for the gpu code
, we find rather comparable performance from @xmath143 and @xmath144 with a slight advantage for the latter for lattice sizes exceeding @xmath149 .
the maximum performance on the tesla c1060 is around @xmath150 ns ( at @xmath151 ) , about by the expected factor of @xmath152 less than in two dimensions .
the fermi card is faster by another factor of @xmath153 , resulting in a peak performance of @xmath154 ns per spin flip . as a consequence ,
the observed maximal speed - up factors are @xmath155 for the c1060 and @xmath156 for the gtx 480 .
these speed - ups are very close to those observed in two dimensions , such that the relative efficiency of the cpu and gpu implementations are practically identical in two and three dimensions .
a natural generalization of the ising model is to systems with continuous degrees of freedom , i.e. , o(@xmath157 ) spin models with hamiltonian @xmath158 which are , most often , more realistic for the description of real magnetic materials . for implementations of simulation algorithms on gpu
, this offers the opportunity to test the relative performance of floating - point operations , an application for which gpus are highly optimized by design , since their original purpose in rendering graphics in technical terms to a large percentage reduces to manipulations of matrices of floating - point numbers . since for the original graphics purposes calculations in double precision
usually are not relevant , traditionally gpu devices have only offered single - precision arithmetic . although double precision calculations could be emulated in software , native support for double precision arithmetic was only introduced with the gt200 architecture , but it came with a massive performance penalty , double precision being around eight times slower than single precision . only with the recent fermi architecture ,
this drawback has been partially removed , double precision calculations there being only around a factor of two slower than single precision arithmetic .
sufficient precision is certainly important for numerical algorithms that have a potential of accumulating rounding and finite - precision errors such as , e.g. , molecular dynamics simulation codes . for monte carlo simulations , on the other hand ,
it is rather obvious that due to the random sampling rounding errors and numerical accuracy are much less of an issue , and it might be acceptable to represent the degrees of freedom in single precision and use double precision only for accumulated quantities in measurements .
implementations in single and in double precision are benchmarked here , also from the point - of - view of the resulting accuracy . and @xmath151 .
] in addition to the question of the number of digits , the floating - point implementation on cuda devices is not completely compliant with the ieee-754 floating - point standard which is otherwise practically universally accepted for current computing platforms . in particular ,
devices prior to the fermi series do no support the so - called denormalization feature ( for single - precision calculations ) that ensures sufficient accuracy for numbers close to zero .
furthermore , due to device - specific implementations , the accuracy of operations such as division and square root is sometimes less than the @xmath159 units in the last place ( ulp ) prescribed by the ieee standard . additionally , there are some less dramatic deviations such as the absence of a signaling of floating - point exceptions , the lack of configurable rounding modes and others .
the mentioned limitations are removed with the advent of fermi cards , which can be configured to almost completely adhere to the ieee standard .
regarding transcendental functions such as exponential , sine and logarithm , device - specific and highly performant , but less accurate , versions are provided in the cuda framework in addition to the standard c library implementations that are also available @xcite .
( note , however , that these highly performant implementations are only available in single - precision variants . ) apart from these questions of adherence to the floating - point standard and whether single - precision calculations are sufficient for the purpose at hand , one must keep in mind that parallel algorithms in themselves might lead to finite - precision effects which are different from those seen in serial algorithms .
if for instance , a binary - tree algorithm is used to perform a parallel reduction over an array of values to evaluate their sum ( as in the energy measurement code of the algorithms discussed here ) , the result will be slightly different to that found in a serial calculation since in finite precision addition is not commutative .
such effects can be reduced , however , with appropriate adaptations , for instance sorting the array by size prior to the reduction , such that values of similar magnitude are added first .
such considerations are not specific to gpus , of course , but need to be taken into account whenever parallel algorithms come into play . to study the effects of these particularities of the massively parallel gpu architecture on the performance and accuracy of a monte carlo simulation , a metropolis single - spin update simulation code was benchmarked for the case of the heisenberg model .
for the sake of simplicity , a new spin orientation is proposed uniformly at each update step , independent of the previous orientation of the spin . in practice , this leads to poor acceptance rates at low temperatures which could be improved by proposing small modifications instead of uniform re - orientations . for the benchmarks considered here , however , this question is irrelevant , since local reorientations lead to almost identical results for the relative performance of the different implementations considered .
in the standard implementation , drawing a random unit vector in three dimensions requires two random numbers and a total of five trigonometric function calls .
in addition , there are @xmath160 multiply - and - add operations for calculating the scalar products in determining the local energy change in an attempted flip for simulations on a @xmath161-dimensional ( hyper-)cubic lattice . compared to the ising system , the exponentials in the metropolis criterion ( [ eq : metropolis ] )
can not be tabulated in advance , leading to an extra special - function call per attempted update . as a result of these differences ,
an update step for the heisenberg model is significantly heavier with arithmetic operations than an ising update which , however , from the point - of - view of gpu computing is a rather favorable situation since this reduces the probability of memory bandwidth limitations .
( note , however , that heisenberg updates also require larger memory transfers since each spin variable requires three floating - point numbers . ) llt2t0 & & & + cpu ( intel q9650 ) & float or double & 183.16 & 1 + & float & 0.74 & 247 + tesla c1060 & float , fast_math & 0.30 & 608 + & double & 4.66 & 39 + & float & 0.50 & 366 + gtx 480 & float , fast_math & 0.18 & 1029 + & double & 1.94 & 94 + figure [ fig : heisen ] compares the resulting spin - flip times for various gpu and cpu implementations and a system size of @xmath27 spins .
the cpu implementation has exactly the same performance for single and double precision implementations , which is plausible since most arithmetic units in current cpus are double precision anyway .
for single - precision calculations , the use of the intrinsic , fast implementations of the special functions results in a factor of 2.5 speedup of the entire code , while the use of double variables for representing spins results in an about sixfold slow - down for the tesla c1060 and a fourfold slow - down for the fermi card .
overall speed - up factors are rather impressive , ranging from 39 for the double - precision version on the tesla c1060 and 1029 for the single - precision code with fast special functions on the fermi board , cf . the data collected in tab . [ tab : heisen ] .
it should be noted , of course , that reduced - precision implementations of special functions are , in principle , also possible on cpu architectures , such that the extra speed - up gained by the use of these functions is not specific to using gpus . to see whether numerical stability and finite precision are relevant for the calculations considered here , deviations from the required normalization @xmath162
have been monitored for the different implementations . as can be seen from the results presented in fig .
[ fig : stability_heisen ] , a drift of normalization is not an issue here . as expected , deviations of @xmath163 from unity are of the order of @xmath164 for single - precision implementations uniformly for cpu as well as gpu implementations .
maybe surprisingly , deviations are even somewhat smaller for the gpu codes than for the cpu implementation .
using the fast intrinsic implementations of special functions , on the other hand , reduces the accuracy by about one digit , such that deviation are of the order of @xmath165 there .
estimates of the magnetization , internal energy and specific heat extracted from the simulation runs also have been compared between the different implementations and , at the given level of statistical accuracy , no significant deviations have been observed .
( note that reductions in the measurement cycles have been implemented in double precision . )
it appears justified , therefore , to claim that for spin model simulations a representation of the spin variables in single precision , while still doing measurements in double precision , does not introduce systematic deviations of relevance as compared to statistical fluctuations .
the same seems to be valid even when using the fast intrinsic , reduced accuracy implementations of the special functions . in view of the dramatic speed - ups observed in particular for the `` float , fast_math '' implementation , cf.tab .
[ tab : heisen ] , this appears encouraging .
computationally particularly demanding problems in the field of spin systems occur for models with quenched disorder @xcite .
the nature of fluctuations between disorder realizations typically requires averages over at least several thousand disorder configurations to enable reliable results . for the case of spin glasses , where disorder is accompanied by the presence of competing interactions , each disorder configuration exhibits extremely slow dynamics close to or below the spin - glass transition . since no effective
cluster algorithms are available for spin glasses ( with the exception of systems in two dimensions @xcite ) , simulations are restricted to single - spin updates , and the combined effect of disorder average and slow relaxation results in rather excessive computational requirements .
it is not surprising , therefore , that a number of special - purpose computers have been constructed for such problems , the most recent being the janus machine based on fpgas @xcite .
independent disorder realizations of a @xmath112 ising spin glass system on the gt200 and fermi architectures as compared to the cpu implementation . ] the trivial parallelism inherent in the disorder average and the necessary restriction to local spin updates appears to render spin - glass problems ideal candidates to profit from the massively parallel architecture of gpus . to see to which degree such hopes are borne out in real implementations , i studied the short - range edwards - anderson ising spin glass model with hamiltonian @xmath166 where the exchange couplings @xmath167 are quenched random variables , here drawn from a symmetric bimodal distribution @xmath168 .
compared to the ferromagnetic ising model , the disordered version requires the values of the couplings @xmath167 to be taken into account when calculating the energy change of a spin flip such that an additional data structure for the couplings becomes necessary . for the tiled algorithm described in sec .
[ sec : checker ] above , the values of @xmath167 inside of each tile need to be loaded into shared memory in addition to the spin field .
furthermore , a number @xmath169 of completely independent instances of the spin system with different coupling values can be simulated in parallel .
these different instances are mapped to different thread blocks in the gpu architecture .
apart from these changes , the algorithm is identical to the implementation described for the ferromagnetic model . or @xmath170 bit variables as a function of the number of disorder realizations .
simulations are for a @xmath171 system , comparing cpu and gpu codes on various devices . ]
the spin - flip times resulting from this implementation are displayed in fig .
[ fig : spinglass ] as a function of the number @xmath169 of disorder realizations simulated simultaneously .
the reference cpu implementation performs at around @xmath172 ns per spin flip on the intel q9650
cpu used here , essentially independent of the number of realizations as those are merely worked on in sequence there .
the gpu code reaches a maximum performance of @xmath173 ns on the tesla c1060 and @xmath174 ns on the fermi card , respectively , corresponding to speed - ups of @xmath175 ( c1060 ) and @xmath176 ( gtx 480 ) .
these speed - ups are practically identical to the maximum speed - ups observed for the ferromagnetic model , whereas in absolute terms due to the extra calculations related to the non - uniform couplings the cpu and gpu implementations for the spin - glass model are about a factor of two slower than the ferromagnetic codes .
it is well known that storing the inherently one - bit information of the orientation of an ising spin in a ( typically ) 4-byte integer variable is a wasteful operation and packing several spin variables in each machine word can result in significant performance improvements @xcite .
while for the case of ferromagnetic models the only possible implementation needs to pack a number of spins from the same lattice into one word of @xmath177 bits , leading to a somewhat involved sequence of bit - wise operations becoming necessary in order to perform the metropolis updates of the coded spin in parallel @xcite ( `` synchronous '' multi - spin coding @xcite ) , the situation in the presence of quenched disorder implies that one can also code @xmath177 spins in the same spatial location of _ independent _ ising lattices corresponding to different disorder configurations in one word ( `` asynchronous '' multi - spin coding @xcite ) .
this setup allows for very efficient implementations using only a few bit - wise operations . depending on the type of integers used
, currently @xmath178 or @xmath170 spins can be represented in one variable . in parallel , 32 or 64 values of bimodal couplings of different realizations
are stored in another field of integers .
the resulting performance characteristics are shown in fig .
[ fig : spinglassmsc ] .
the cpu code performs at @xmath179 ns per spin flip for @xmath178 and a significantly better @xmath180 ns for @xmath170 , corresponding to a speedup of @xmath181 and @xmath182 as compared to the single - spin coded cpu version for @xmath178 and @xmath170 , respectively or @xmath170 spins , which appears justified in view of the statistical independence of the coupling realizations .
this setup was previously tested for the ising spin glass in three dimensions and no significant correlation effects were observed @xcite . ] .
the gpu implementation performs at around @xmath183 ps for 32-bit and @xmath184 ps for 64-bit on the tesla c1060 card . the rather disappointing result for
the @xmath170 version shows that the gt200 architecture is not optimized for @xmath39-bit operations . on the gtx 480 fermi card , on the other hand , the @xmath39-bit multi - spin coded implementation runs as fast as @xmath131 ps per spin flip as compared to @xmath185 ps for the @xmath43-bit version .
the resulting optimal speed - ups compared to the cpu code are hence around @xmath70 for the tesla c1060 and around @xmath186 for the gtx 480 . for the fermi card ,
the resulting code runs around @xmath187 times faster than the ( synchronous ) multi - spin coded simulation of the ferromagnetic model discussed in ref .
it is worthwhile to note that janus performs at around @xmath1 ps per spin flip on the 3d ising spin glass @xcite .
circumstances forbid a clear - cut comparison here , however , since the janus code uses synchronous multi - spin coding and a somewhat better random - number generator .
still , it appears fair to say that an appropriate gpu implementation seems to play in the same league as the special - purpose computer janus at a much smaller investment in capital and human resources .
a rather powerful technique for the simulation of systems with complex free energy landscapes , including spin glasses and bio - polymers , is given by the parallel tempering or replica exchange algorithm @xcite .
there , a number of identical copies of a system are simulated at a set of different , closely neighboring temperatures .
in addition to any chosen intra - replica update , such as the single spin flips according to the metropolis criterion , at fixed intervals an interchange of configurations ( typically ) simulated at neighboring temperatures is attempted and accepted with the generalized metropolis probability @xmath188,\ ] ] where @xmath189 and @xmath190 . with a properly chosen set of temperatures @xcite , this additional update allows for configurations with slow dynamics at low temperatures to successively diffuse to high temperatures , decorrelate there and finally return back to the low - temperature phase , likely arriving in a different valley of a complex free - energy landscape than where it started from .
$ ] with @xmath191 replicas with one exchange attempt per 100 sweeps of spin flips . ] through its inherent parallelism of intra - replica updates , this scheme appears to be well suited for the massively parallel architecture of current gpus .
it is tested here in a reference implementation for the ferromagnetic ising model along the lines of the previously discussed single spin - flip code .
the additional copies of the system are mapped to additional thread blocks running in parallel .
since the configurational energy is calculated online from the energy changes of single spin flips via a binary - tree reduction algorithm on tiles , it is automatically available after lattice sweeps and hence its determination does not incur any extra overhead . in the current implementation , replica exchange steps are performed on cpu since the computational effort is low and synchronization between thread blocks is required . as usual , instead of exchanging configurations between replicas only the corresponding temperatures are swapped .
the boltzmann factors according to ( [ eq : metropolis ] ) can still be tabulated , and are now implemented on gpu as fetches from a two - dimensional texture .
a successful replica exchange then requires an update of the texture which is easily possible in the current setup as the exchange moves are carried out on cpu .
the benchmark results of the parallel tempering simulation of the 2d ising model on gpu and cpu is shown in fig .
[ fig : parallel ] .
i chose to use @xmath191 replicas at equally spaced inverse temperatures in the interval @xmath192 $ ] . in a real - world application
one would probably want to use a more elaborate scheme for choosing these temperatures @xcite , but these questions do not have any bearing on the benchmark calculations performed here .
as is clearly visible , the presence of additional copies of the system leads to a much better resource utilization for smaller system sizes than that observed for the single - spin flip simulations .
hence , significant speed - ups are observed already for small systems .
the cpu code performs at a constant @xmath193 ns per combined spin - flip and replica exchange move ( mixed at a ratio of one exchange move per one hundred lattice sweeps ) .
the gpu code arrives at a maximum of around @xmath194 ns for the tesla c1060 and @xmath195 ns for the gtx 480 at @xmath196 .
the speed - up reaches up to @xmath197 ( c1060 ) resp .
@xmath198 ( gtx 480 ) at @xmath196 , but is already @xmath199 ( c1060 ) resp.@xmath200 ( gtx 480 ) for the @xmath118 system . as for disordered systems due to the severe slowing down usually only rather small systems can be studied , good performance of the code is crucial in this regime . increasing the number of exchange moves to one in ten lattice sweeps reduces the maximum performance of the gpu implementation somewhat to @xmath201 ns ( c1060 ) and @xmath202 ns ( gtx 480 ) , respectively .
current gpus have a significant potential for speeding up scientific calculations as compared to the more traditional cpu architecture . in particular
, this applies to the simulation of systems with local interactions and local updates , where the massive parallelism inherent to the gpu design can work efficiently thanks to appropriate domain decompositions .
the simulation of lattice spin systems appears to be a paradigmatic example of this class as the decomposition remains static and thus no significant communication overhead is incurred .
observing substantial speed - ups by factors often exceeding two orders of magnitude as for the case studies reported here requires a careful tailoring of implementations to the peculiarities of the considered architecture , however , in particular paying attention to the hierarchic organization of memories ( including more exotic ones such as texture memory ) , the avoidance of branch divergence and the choice of thread and block numbers commensurate with the capabilities of the cards employed . for achieving good performance ,
it is crucial to understand how these devices hide the significant latency of accessing those memories that do not reside on die through the interleaved scheduling of a number of execution threads significantly exceeding the number of available computing cores .
it is encouraging that employing such devices with the rather moderate coding effort mediated by high - level language extensions such as nvidia cuda or opencl updating times in the same ballpark as those of special purpose machines such as janus @xcite with a vastly higher development effort can be reached .
a regularly uttered objection against the systematic use of gpus for scientific computing criticizes them as being a possibly too special and exotic architecture with unknown future availability and architectural stability as compared to the traditional and extremely versatile x86 cpu design .
while there is certainly some truth to this argument , there can be no doubt about the fact that massive parallelism is not only the present state of the gpu architecture , but also the future of cpu based computing . as of this writing , intel cpus feature up to eight cores and amd chips up to twelve cores per die , the corresponding road - maps projecting even significantly more cores in the foreseeable future .
supercomputers will soon count their number of cores in the millions . due to this
development , serial computing will not remain a viable option for serious computational science much longer .
much effort will need to be invested in the years to come into solving scientific problems employing a _ massive _ number of parallel computational units . in this respect , gpu computing ,
apart from currently being more efficient for many problems in terms of flop / s per watt and per euro than cpu based solutions , is rather less exotic than pioneering .
an ideal application for gpu computing in the field of the simulation of spin systems appear to be disordered systems , where cluster algorithms are in general not efficient and a natural parallelism occurs from the quenched average over disorder , possibly combined with the parallel tempering algorithm . using asynchronous multi - spin coding for the ising spin glass , spin flip times down to @xmath73 ps can be achieved .
systems with continuous spins are particularly well suited for gpu deployment , since one finds a relatively stronger overhead of arithmetic operations over memory fetches as compared to systems with discrete spins there . for the heisenberg model , speed - ups up to a factor of @xmath0 can be obtained when making use of the highly optimized special function implementations in single precision .
while these examples of single - spin flip simulations look rather promising , it is clear that other well - established simulation algorithms for spin systems are less well suited for the parallelism of gpus , including the cluster algorithms in the vicinity of critical points , where spin clusters spanning the whole system need to be identified , or multi - canonical and wang - landau simulations , which require access to the current values of a global reaction coordinate ( such as , e.g. , the energy or magnetization ) for each single spin flip , thus effectively serializing all spin updates .
it is a challenge for future research to find implementations or modifications of such algorithms suitable for massively parallel computers @xcite .
i am indebted to t. yavorskii for a careful reading of the manuscript . support by the dfg through the emmy noether programme under contract no . we4425/1 - 1 and by the schwerpunkt fr rechnergesttzte forschungsmethoden in den naturwissenschaften ( srfn mainz ) is gratefully acknowledged .
52 natexlab#1#1[2]#2 , , , , , , ( ) . , , , ( ) .
, , , ( ) . , , , ( ) .
, , , ( ) . , , , , ( ) .
, , , , , , , , , , , , , , , , , , , , , ( ) . , , , , , , , , , , , , , , , , , , , , , , , , , ( ) . , , , , , , ( ) . , , , , , , ( ) . , , , , , ( ) . , , , , ( ) . , . ,
, , , ( ) . , , ( ) . , , , ( ) .
( ed . ) , , , , .
, , , , ( ) . , , , , ( ) . , , , , ( ) .
, , , ( ) . , , , , , , , , in : , , , , ( eds . ) , , volume of _ _ , , , , pp . . , , , , , , , ( ) . , , , , , . , , , , . , , , ( ) . , , , , ( ) . , , , ( ) . , , , , , , , edition , . , , , , edition , .
, , , , edition , .
, , , ( ) . , , in : , , pp . .
, , , ( ) . , , , , .
, , in : ( ed . ) , , , , , p. .
, , , ( ) . , , ( ) . .
, , , , ( ) . , , , , ( ) .
, , in : , , , , p. .
, , , , , ( ) .
, , , , ( ) . , , , ( ) . | graphics processing units ( gpus ) are recently being used to an increasing degree for general computational purposes .
this development is motivated by their theoretical peak performance , which significantly exceeds that of broadly available cpus . for practical purposes , however , it is far from clear how much of this theoretical performance can be realized in actual scientific applications .
as is discussed here for the case of studying classical spin models of statistical mechanics by monte carlo simulations , only an explicit tailoring of the involved algorithms to the specific architecture under consideration allows to harvest the computational power of gpu systems .
a number of examples , ranging from metropolis simulations of ferromagnetic ising models , over continuous heisenberg and disordered spin - glass systems to parallel - tempering simulations are discussed .
significant speed - ups by factors of up to @xmath0 compared to serial cpu code as well as previous gpu implementations are observed .
monte carlo simulations , graphics processing units , ising model , heisenberg model , spin glasses , parallel tempering |
since their discovery , carbon nanotubes have shown great promise for a wide variety of applications which utilize their unique electronic and mechanical properties.@xmath1@xcite@xmath2 $ ] for applications in which individual nanotubes act as the working element , such as nanotube field effect transistors ( fets)@xmath1@xcite@xmath2 $ ] or chemical sensors,@xmath1@xcite@xmath2 $ ] it is important to control the location and orientation of the nanotubes .
nanotubes can be prefabricated and then assembled into the desired geometry , or they can be fabricated in place using chemical vapor deposition ( cvd).@xmath1@xcite@xmath2 $ ] cvd is preferred since the growth can be widely tuned , both in yield and structure ( single- vs. multi - walled ) , by modifying the experimental conditions . while there are a vast number of cvd recipes available in the literature for both single- and multi - walled nanotube growth , most studies of the growth process have focused on the role of the carbon precursor ( e.g. co , ch@xmath3 , c@xmath4h@xmath4 ) and temperature as control parameters with less attention placed on the choice of catalyst and substrate.@xmath1@xcite@xmath2 $ ] optimization of the catalytic process requires an understanding of the catalyst chemistry throughout the growth process , including the initial chemical state of the catalyst before the introduction of the feedstock , as well as the catalytic decomposition of the feedstock in or on the catalyst particle.@xmath1@xcite@xmath2 $ ] only recently have studies focused on the catalyst - feedstock and catalyst - substrate interaction.@xmath1@xcite@xmath2 $ ] it is well established that the yield of single - walled nanotube growth on silicon substrates is dramatically reduced compared with growth on thick silicon dioxide layers , due to poisoning of the catalyst by the formation of a silicide.@xmath1@xcite@xmath2 $ ] intriguingly , the growth of multi - walled nanotubes on silicon substrates is regularly reported and seems less susceptible to catalyst poisoning,@xmath1@xcite@xmath2 $ ] yet the cause of this difference has not been addressed . for device applications it is desirable to use thin oxides to increase the gate capacitance and gate efficiency , and
thus it is important to understand how to minimize oxide thickness while still preventing catalyst poisoning .
though thick silicon dioxide layers have been used as diffusion barriers during nanotube growth , there has been no investigations to date that determine the the effectiveness of ultra - thin oxides .
here we demonstrate explicitly that catalyst diffusion through the ultra - thin silicon dioxide layer controls the formation of the non - catalytic silicide . using x - ray photoelectron spectroscopy ( xps ) , we study the interfacial reactions between the substrate and iron- or nickel - based nanotube catalysts during the initial temperature ramp portion of a cvd growth cycle . we show that ultra - thin oxide layers ( 4 nm or greater ) are sufficient to inhibit the silicide formation and permit high yield growth of single - walled carbon nanotubes . on thinner oxides or clean silicon ,
silicide formation begins by 600 @xmath0c ( ni ) or 800 @xmath0c ( fe ) and the catalyst is entirely consumed in the silicide at the growth temperature of single - walled nanotubes .
the silicide formation temperatures account for the difference in single- versus multi - walled nanotube growth because multi - walled nanotubes are grown at lower temperatures where some of the catalyst remains unreacted and active for catalysis .
interestingly , the silicides form while the silicon dioxide is still present on the substrate , indicating that diffusion of metal or silicon through the oxide is occurring .
we show that it is metal diffusion through the oxide , forming a metal silicide underneath the oxide , which dominates the interfacial reactions between the nanotube catalyst and the silicon substrate .
to understand the catalyst - substrate interfacial reactions , we first analyze xps spectra of iron nitrate catalyst that has been deposited on a thick ( 100 nm ) thermal oxide . on thick oxides
, silicides will not form and single - walled nanotube growth occurs with high yield ( figure [ fig1]a ) . .
( b ) fe 2p core level spectra of iron nitrate catalyst on 100 nm sio@xmath4 . as the sample
is annealed , there is no appreciable change in the oxidation state of the iron.,width=216 ] before annealing , the iron core level ( figure [ fig1]b ) exhibits broad , asymmetric peaks at 710.8 ev and 724.5 ev , corresponding to the fe 2p@xmath5 and 2p@xmath6 levels of iron oxide in a mixed fe@xmath7/fe@xmath8 oxidation state.@xmath1@xcite@xmath2 $ ] the extra intensity on the high binding energy side of the core levels is due to unresolved satellites that are characteristic of oxidized iron.@xmath1@xcite@xmath2 $ ] as the catalyst is heated on the thick oxide , there are no major changes in the chemistry of the iron oxide .
there is a small loss in iron intensity , due to either agglomeration of the catalyst on the surface ( thereby reducing the measured intensity due to the finite electron escape depth ) or desorption of the iron into the vacuum while annealing.@xmath1@xcite@xmath2 $ ] there is also a shift in the center of gravity of the fe 2p@xmath5 core level as the fe@xmath7 oxide ( feo ) transforms into the more stable fe@xmath8 oxide ( fe@xmath4o@xmath9)@xmath1@xcite@xmath2 $ ] and the satellite peaks at 715 ev and 730 ev become better defined .
even at 1000 @xmath0c , the catalyst remains as iron oxide because there is no reaction pathway to reduce it to pure iron .
the necessary reduction to metallic iron in a cvd reactor would come from flowing hydrogen or from feedstock decomposition products , enabling the catalysis of carbon nanotube growth . at the opposite extreme - that of hydrogen - terminated silicon on which single - walled nanotube growth is inhibited - surface reactions play a major role in the evolution of catalyst chemistry even at low temperatures . .
by 800 @xmath0c , the iron begins to alloy with the substrate , forming an iron silicide , seen in ( d ) by a 0.4 ev shift in the fe 2p@xmath5 core level and in ( c ) by the emergence of a plasmon loss peak.,width=216 ] while iron nitrate catalyst that has been deposited on a hydrogen terminated silicon substrate remains oxidized at room temperature ( figure [ fig2]b ) , the center of gravity of the fe 2p@xmath5 core level is shifted to 710.5 ev , indicating that the iron has been partially reduced as compared to the catalyst on the thick oxide .
whereas the si 2p spectrum of the thick oxide shows only a peak at 103.3 ev due to the sio@xmath4 layer , the spectrum of the hydrogen - terminated substrate ( figure [ fig2]a ) at room temperature resolves the si 2p atomic core levels ( 99.3 and 103.3 ev ) , as well as a small , broad peak on the high binding energy side of the si 2p peak , at around 102.7 ev , indicating the formation of a non - stoichiometric silicon suboxide.@xmath1@xcite@xmath2 $ ] this suboxide could have formed in air after the hf oxide strip but before introduction in the vacuum chamber , or could account for the small amount of reduction in the iron as compared with the catalyst on the thick oxide .
reports have indicated that a silicon dioxide is more stable than either feo or fe@xmath4o@xmath9 , which could lead to an exchange of oxygen from the iron oxide to the silicon substrate,@xmath1@xcite@xmath2 $ ] even at low temperature .
in contrast to the thick oxide , iron catalyst deposited on hydrogen terminated silicon begins to reduce to metallic iron when the catalyst by the time it is annealed to 550 @xmath0c .
this is seen in figure [ fig2]b by the decrease in fe 2p intensity at 710.5 ev , 724.1 ev , and the associated satellites of iron oxide , and by the appearance of peaks at 707.4 ev and 720.7 ev corresponding to the fe 2p core levels of fe@xmath10 .
the iron oxide is further reduced as the substrate is annealed , indicated by the reduced intensity of the iron oxide peaks ( figure [ fig2]b ) . by 800 @xmath0c
, the iron oxide has been fully reduced . as in the case of a thick oxide ,
the reduction in uhv must come from a surface reaction between the oxidized catalyst and the substrate .
indeed , there is an appreciable exchange of oxygen from the iron oxide to the silicon substrate , leading to a marked increase in intensity of the silicon oxide peak , accompanied by a shift in the peak position to higher binding energies ( figure [ fig2]a ) . at 550 @xmath0c the silicon oxide peak shifts to higher binding energy , and between 700 @xmath0c and 800 @xmath0c the peak reaches 103.3 ev which is characteristic of stoichiometric sio@xmath4 .
this indicates that the iron oxide catalyst is reduced by the silicon substrate leading to the formation of a thin sio@xmath4 layer .
after the iron has been fully reduced at 800 @xmath0c , the catalyst begins to form a silicide . at 800 @xmath0c ,
the fe 2p@xmath5 core level has shifted to higher binding energy by approximately 0.2 ev ( figure [ fig2]d ) , and a very broad peak begins to appear in the fe 2p spectrum centered around 729 ev ( figure [ fig2]c ) . in the silicon core levels
, a small shoulder appears at 100.5 ev on the high binding energy side of the si 2p peak ( figure [ fig2]a ) . by 900 @xmath0c
, the fe 2p@xmath5 peak has shifted further to 707.8 ev , giving a total binding energy increase of 0.4 ev ( figure [ fig2]d ) .
the 0.4 ev total shift in the fe 2p@xmath5 peak position , accompanied by the appearance of a peak @xmath1121 ev higher in binding energy , is indicative of the formation of an iron silicide , likely fesi@xmath4.@xmath1@xcite@xmath2 $ ] the peak at 729 ev is due to a plasmon loss , seen in figure 2c , and has been observed in electron energy loss spectroscopy@xmath1@xcite@xmath2 $ ] as well as xps@xmath1@xcite@xmath2 $ ] investigations of fe / si multilayers .
further , the intensity of the shoulder on the si 2p peak increases and the sio@xmath4 has desorbed entirely as evidenced by the loss of the @xmath11103.3 ev feature in the si 2p spectrum . though the shoulder in the si 2p spectrum at 100.5 ev seems to be associated with the formation of the silicide , the apparent binding energy of the shoulder is too high to be due to a fesi@xmath12 core level which would be expected to be shifted by only @xmath110.2 ev.@xmath1@xcite@xmath2 $ ] at present , the origin of this shoulder is not known .
annealing at 1000 @xmath0c ( not shown ) only leads to a decrease in the intensity of the iron peaks , but causes no noticeable changes in the peak positions , indicating that the silicide formation is complete at 900 @xmath0c .
catalyst poisoning by silicide formation is not unique to iron .
new also see that nickel thin film catalyst on clean silicon is incorporated into a silicide , only at a much lower temperature .
c , seen as an up - shift in the ni 2p core level . by 800 @xmath0c , the core level has shifted by 1.0 ev , indicating that all of the nickel has been incorporated into a silicide .
( d ) schematic reaction pathway from metal oxide catalyst to the formation of the non - catalytic silicide.,width=216 ] whereas iron catalyst is not fully reduced until 800 @xmath0c , the native nickel oxide has been fully reduced by 600 @xmath0c , and is accompanied by the growth of a sio@xmath4 layer ( figure [ fig3]a , b ) . by 600
@xmath0c some of the nickel has also been incorporated into a silicide , seen as a high binding energy shoulder on the ni 2p@xmath5 core level ( figure [ fig3]c ) . as
the substrate is further annealed , the ni 2p@xmath5 level shifts to higher binding energy , reaching a total shift of 1.0 ev by 800 @xmath0c where the entire nickel film has been incorporated into the silicide.@xmath1@xcite@xmath2 $ ] the chemical progression for both iron and nickel catalyst can be summarized as in figure 3d .
the initial metal or metal oxide is completely reduced by the silicon substrate leading to the growth of a silicon dioxide layer . after being reduced
, the metal can then react with the silicon substrate directly , leading to the formation of a silicide .
the initial silicide formation at @xmath11600 @xmath0c for nickel and @xmath11800 @xmath0c for iron explains the differing growth yields of single- and multi - walled nanotubes on clean silicon substrates .
since cvd of multi - walled nanotubes on iron occurs at relatively low temperatures , the metal remains unreacted and available for nanotube catalysis .
in contrast , the growth of single - walled nanotubes is performed at higher temperatures where all of the catalysts have been consumed by the silicide . in the case of iron presented in figure [ fig2]d ,
the fe 2p@xmath5 peak is only shifted by 0.2 ev at the multi - walled nanotube growth temperature of 800 @xmath0c .
this shift is approximately half of the binding energy shift that would correspond to a stoichiometric iron silicide , indicating that catalytic regions of pure iron are still present which permit high yield growth of multi - walled nanotubes . by 900
@xmath0c the iron silicide formation is complete , preventing nanotube catalysis and leading to poor growth yield .
since iron silicide reduces catalytic yield and forms when iron - based catalysts are used on bare silicon wafers but not on highly oxidized wafers , it is important to know how to maintain the iron in a catalytic state without requiring the use of a thick oxide . for electronic applications such as fets
, the use of thinner oxides gives higher gate efficiency due to the increased capacitance of the gate dielectric .
de los arcos and co - authors have shown that iron thin films maintain their catalytic state when they are placed on top of nitrides ( tin ) and oxides ( al@xmath4o@xmath9 and tio@xmath4),@xmath1@xcite@xmath2 $ ] but the use of silicon oxide would be more compatible with industrial silicon processing . in order to utilize silicon dioxide ,
it is important to determine what is the minimum thickness required to prevent the formation of a silicide . to determine this ,
iron catalyst is deposited onto 3 , 4 , and 8 nm oxides , and we use the measured binding energy of the reduced fe 2p@xmath5 core level around 707.3 ev to detect any silicide formation ( figs .
[ fig4]a - d ) .
if the silicon dioxide layer is thick enough to prevent silicide formation , the iron core level will remain at 707.3 ev , but if a silicide forms the peak will shift to higher binding energy , where 707.7 ev is the position of the fully formed iron silicide .
as seen above for the case of iron or nickel on hydrogen - terminated silicon , as the substrates are annealed some of the iron oxide is reduced by the formation of silicon dioxide , even for iron nitrate on an 8 nm oxide . after this reduction
, the metal catalyst could react further with the substrate to form the silicide . for catalyst deposited on the clean silicon ( figure [ fig4]a ) or a 3 nm oxide ( figure [ fig4]b )
, we see that the iron core level shifts to higher binding energy above 800 @xmath0c , indicating the nucleation of the silicide , followed by complete silicide formation at 900 @xmath0c .
in contrast , heating the catalyst to 900 @xmath0c on an 8 nm oxide ( figure [ fig4]d ) does not lead to silicide formation . in the case of the 4 nm silicon dioxide layer ( figure [ fig4]c )
, we see a small up - shift of 0.2 ev after the catalyst has been annealed at 900 @xmath0c , suggesting that some of the iron has been incorporated into a silicide , but that an appreciable amount of the iron remains in reduced form .
thus , a 4 nm oxide layer represents the minimum thickness that can inhibit the formation of the iron silicide during the growth of single - walled nanotubes at 900 @xmath0c .
this value is similar to a previous estimate of 5 - 6 nm , where the role of non - catalytic silicides was conjectured rather than directly shown.@xmath1@xcite@xmath2 $ ] having uncovered the minimum barrier oxide thickness in the case of iron catalyst , it is important to determine what mechanism sets this limit .
close observation of the xps spectra reveals a key signature of the mechanism of the interfacial reaction : the silicide is forming even though a silicon dioxide layer is present on the surface .
this indicates that either the silicon is diffusing up through the oxide to form the silicide , or that the metal is diffusing down .
the direction of diffusion is demonstrated most clearly in the case of a nickel thin film on a native oxide substrate ( figure[fig5 ] ) . and ( b ) si 2p core levels for a nickel film catalyst on native oxide .
the metallic nickel present at 600 @xmath0c has been incorporated into a silicide by 800 @xmath0c before the desorption of the sio@xmath4 layer . when the oxide layer is removed , the nickel signal increases indicating that there is a higher concentration of nickel beneath the oxide than on top due to diffusion of metal through the oxide at lower temperatures.,width=240 ] at low temperatures the fully reduced nickel film coexists with the thin sio@xmath4 layer , however by 800 @xmath0c the ni 2p@xmath5 core level is fully shifted due to the formation of the silicide , while the sio@xmath4 layer remains .
after annealing to 900 @xmath0c , to desorb the silicon dioxide layer , the spectra show an increase in intensity from nickel , while the binding energy remains the same .
this increase in intensity is clear evidence that the nickel silicide was originally beneath the sio@xmath4 , where its xps intensity was reduced due to the finite electron escape depth;@xmath1@xcite@xmath2 $ ] annealing the sample to remove the sio@xmath4 eliminates the scattering nickel photoelectrons , leading to an increase in the nickel intensity . for the silicide to form beneath the oxide
, the metal catalyst must diffuse through the oxide .
the diffusion mechanism further explains why single - walled nanotube growth is more inhibited than multi - walled nanotube growth since the diffusion kinetics are enhanced at the higher single - walled nanotube growth temperatures . given the diffusion mechanism , one can tailor the oxide thickness to maximize the gate coupling while maintaining the catalyst in an active state .
catalyst diffusion through the oxide is determined by the temperature dependent diffusion constant @xmath13 and the corresponding diffusion length @xmath14 , where @xmath15 is the activation energy for diffusion and @xmath16 is the diffusion time .
using the bulk diffusion constant for iron in electronic grade silicon dioxide ( @xmath17 @xmath18s@xmath19 , @xmath20)@xmath1@xcite@xmath2 $ ] one would only expect the iron to diffuse @xmath21 nm during the temperature ramp to 900 @xmath0c , roughly three times shorter than observed in xps .
the 4 nm diffusion length observed can be explained by a small reduction of @xmath22 in the activation energy , yielding a five - fold increase in the effective diffusion constant .
an increased diffusion constant is not surprising for such ultra - thin oxides since they have higher concentrations of pinhole defects as compared to thick oxides . indeed
, previous studies of metals on ultra - thin silicon oxides have shown that pinholes or defects in the oxide significantly enhances the diffusion process.@xmath1@xcite@xmath2 $ ] by determining the critical oxide thickness that prevents silicide formation for iron catalyst , we can extrapolate to determine the critical thickness for cobalt and nickel which are the other common single - walled nanotube catalysts . for cobalt ,
the bulk diffusion parameters are @xmath23@xmath18s@xmath19 and @xmath24,@xmath1@xcite@xmath2 $ ] leading to a diffusion length and critical oxide thickness of @xmath25 nm at a 900 @xmath0c growth temperature.@xmath1@xcite@xmath2 $ ] this bulk diffusion length is sufficiently thick that high quality oxide layers can be made to prevent defect - enhanced diffusion . for nickel , bulk diffusion ( @xmath26@xmath18s@xmath19
, @xmath27)@xmath1@xcite@xmath2 $ ] would predict a @xmath28 nm critical thickness , which is in the regime of the enhanced diffusion seen for iron .
if we assume a proportionate decrease in the activation energy for ni diffusion ( @xmath29 ) , the critical oxide thickness at the 900 @xmath0c growth temperature@xmath1@xcite@xmath2 $ ] becomes @xmath30 nm . by minimizing the concentration of defects in the oxide , one could reduce these critical oxide thicknesses further .
though the critical thickness could also be reduced by using a more effective barrier layer than silicon dioxide , the nanotube - gate coupling will also change due to changes in the insulator capacitance .
having determined the minimum oxide thickness to prevent catalyst diffusion through the oxide layer from xps , it is important to ensure that the conclusions are consistent with growth observations .
surface reactions observed in uhv may be inhibited when the process is repeated in a cvd reactor where reactant gases can influence catalyst chemistry .
since transition - metal silicides are more stable than transition - metal carbides at nanotube growth temperatures,@xmath1@xcite@xmath2 $ ] it is expected that the reactants will not be able to reverse the silicide formation and nanotube catalysis will be inhibited . to test these conclusions ,
growth was attempted on unannealed pieces of each of the thin oxides .
figures [ fig4]e - h show sem images of the results of this growth test . in the case of growth on hydrogen
terminated silicon and 3 nm oxides ( figs .
[ fig4]e , f ) , nanotubes are absent on the majority of the substrate .
the images shown depict one of the few locations where nanotubes did grow , and the reported densities represent an upper limit on the nanotube yield .
as is readily apparent , there is very little nanotube yield when growth was performed on hydrogen terminated silicon or a 3 nm oxide .
in contrast , and consistent with our xps results , yields on the 4 and 8 nm oxides ( figs .
[ fig4]g , h ) are significantly improved .
in summary , we have used x - ray photoelectron spectroscopy to monitor the formation of silicides that inhibit the growth of carbon nanotubes on silicon .
annealing metal catalysts on hydrogen - terminated silicon or ultra - thin oxide layers leads to the formation of a non - catalytic silicide around 800 @xmath0c ( fe ) and 600 @xmath0c ( ni ) . by 900 @xmath0c ,
all of the catalyst has been incorporated into the silicide . in order to prevent silicide formation , the oxide substrate must act as an effective barrier to the diffusion of metal catalyst at the high single - walled nanotube growth temperature . in the case of iron ,
our xps and growth experiments have shown that oxide layers that are 4 nm or greater are able to inhibit silicide formation at 900 @xmath0c , leading to a high nanotube growth yield . by limiting catalyst diffusion through the oxide layer ,
either with reduced growth temperatures or reduced diffusivity in the oxide , single - walled nanotubes can be grown with high yields on ultra - thin sio@xmath4 layers .
the substrates used in this study consist of highly doped silicon wafers capped with thermally grown sio@xmath4 ( 100 , 8 , 4 , or 3 nm ) , a native ( @xmath31 nm ) oxide , or hydrogen termination . all oxide thicknesses are measured using multi - wavelength ellipsometry and confirmed by direct measurements using atomic force microscopy . the oxide thickness uniformity across the substrates is better than 1 nm . to obtain a clean silicon wafer , a hydrofluoric acid etch
is used to remove the native oxide and hydrogen - terminate the silicon . immediately after hydrogen termination ,
the catalyst is deposited and the sample is introduced into the vacuum chamber within five minutes in order to limit the formation of a native oxide layer .
the quality of the hydrogen termination is confirmed by the low intensity of the oxidized silicon peak in the xps spectra .
although there are a large number of catalysts available , many require the incorporation of an insulating matrix , such as alumina , which can influence the chemistry of the catalyst .
for example , annealing a thin film of iron on an alumina buffer layer was shown to lead to an oxidation of the iron to fe@xmath4o@xmath9.@xmath1@xcite@xmath2 $ ] in order to avoid such interactions , we use catalysts that do not require a matrix : iron nitrate and thermally evaporated nickel thin films .
the iron nitrate catalyst is prepared by dissolving 50 mg iron ( iii ) nitrate nonahydrate ( fe(no@xmath9)@xmath329h@xmath4o ) in 100 ml isopropanol and stirring for two minutes.@xmath1@xcite@xmath2 $ ] the substrates are then soaked in the solution for one minute and the sample is dried with flowing nitrogen , forming a thin layer of catalyst over the entire substrate .
nickel catalyst is deposited onto the substrates by electron beam evaporation , resulting in a 7 film which decomposes at the growth temperature to form nanoparticles .
chemical vapor deposition tests are performed in an atmospheric pressure reactor .
the substrates are brought to the 900 @xmath0c growth temperature under argon flow , after which methane ( 400 sccm ) and hydrogen ( 20 sccm ) are used for the 10 minute nanotube growth .
the samples are then cooled to room temperature under argon before being removed from the reactor .
xps is performed in an ion - pumped ultra - high vacuum ( uhv ) chamber with a working pressure of better than 2x10@xmath33 torr , using monochromatic al@xmath34 ( 1486.6 ev ) x - rays at a resolution of @xmath35 ev .
sample heating is performed in uhv by passing current through the silicon substrate and the temperature is measured using an optical pyrometer .
after introduction into the uhv chamber , the samples are outgassed at @xmath11100 @xmath0c for 10 minutes to remove any residual contaminants . in order to mimic the kinetics of the temperature ramp used in a typical cvd growth , the sample is slowly heated from room temperature to the anneal temperature over 5 minutes . after reaching the desired temperature , substrate heating is stopped and the sample is allowed to cool to room temperature . during the anneal , the base pressure in the chamber increases , preventing the acquisition of spectra at the anneal temperature .
all energies are corrected to the si 2p@xmath5 core level at 99.3 ev for clean silicon and 103.3 ev for sio@xmath4@xmath1@xcite@xmath2 $ ] to correct for charging effects on the oxide substrates .
quantitative xps fitting is obtained by fitting raw data to voigt functions after a shirley background correction.@xmath1@xcite@xmath2 $ ] | the ability to integrate carbon nanotubes , especially single - walled carbon nanotubes , seamlessly onto silicon would expand the range of applications considerably .
though direct integration using chemical vapor deposition is the simplest method , the growth of single - walled carbon nanotubes on bare silicon and on ultra - thin oxides is greatly inhibited due to the formation of a non - catalytic silicide . using x - ray photoelectron spectroscopy
, we show that silicide formation occurs on ultra - thin oxides due to thermally activated metal diffusion through the oxide .
silicides affect the growth of single - walled nanotubes more than multi - walled nanotubes due to the increased kinetics at the higher single - walled nanotube growth temperature .
we demonstrate that nickel and iron catalysts , when deposited on clean silicon or ultra - thin silicon dioxide layers , begin to form silicides at relatively low temperatures , and that by 900 @xmath0c , all of the catalyst has been incorporated into the silicide , rendering it inactive for subsequent single - walled nanotube growth .
we further show that a 4 nm silicon dioxide layer is the minimum diffusion barrier thickness which allows for efficient single - walled nanotube growth . +
* keywords : * carbon nanotubes , catalysis , chemical vapor deposition , photoelectron spectroscopy + + * corresponding author : + m. a. eriksson + department of physics , university of wisconsin - madison + 1150 university ave
. + madison , wi 53706 + e - mail : [email protected] |
recently the atlas @xcite and cms @xcite collaborations reported event excesses , which may imply the higgs boson with mass of about @xmath0gev .
while it is difficult to explain the higgs mass in the minimal supersymmetric standard model ( mssm ) as long as the sparticle masses are around 1tev @xcite , such a higgs mass can be explained if the sparticles are heavier than multi - tev @xcite .
one of such scenarios is the anomaly - mediated susy breaking ( amsb ) model @xcite , where the sfermions and the gravitino are @xmath1@xmath2tev and the gaugino masses are @xmath1@xmath2gev , given by the amsb relation .
phenomenological aspects of this scenario have been discussed in refs .
@xcite , and it was shown that it is compatible with thermal leptogenesis @xcite , which requires the reheating temperature as high as @xmath3gev @xcite . is defined as @xmath4 where @xmath5 is the inflaton decay rate .
note that this definition of @xmath6 is smaller than that of ref .
@xcite by a factor @xmath7 . ]
another attractive scenario is that all sparticles are @xmath8tev in the gravity - mediated susy breaking .
the scenario alleviates the susy flavor / cp problems because of the heavy susy particles , while it explains the 125gev higgs boson for @xmath9 @xcite and the present dark matter abundance ( see ref .
@xcite for realization in the focus - point region @xcite ) .
however , the scenario suffers from the cosmological polonyi problem @xcite , since there must be a singlet susy breaking field , called the polonyi field , in order to generate sizable gaugino masses .
although the polonyi may decay before the big - bang nucleosynthesis ( bbn ) begins for the polonyi mass @xmath10tev , it releases a huge amount of entropy because it dominates the universe before the decay .
thus the leptogenesis scenario does not work in this setup .
an interesting solution to the polonyi problem was proposed long ago by linde @xcite .
it was pointed out that , if the polonyi field has a large hubble - induced mass , it follows a time - dependent potential minimum adiabatically and the resultant amplitude of coherent oscillations is exponentially suppressed .
recently , two of the present authors ( ft and tty ) noticed that there is an upper bound on the reheating temperature for the adiabatic solution to work @xcite and also showed that such a large hubble mass may be a consequence of the strong dynamics at the planck scale @xcite or the fundamental cut - off scale one order of magnitude lower than the planck scale @xcite .
more important , the present authors found that there are generally additional contributions to the polonyi abundance which depends on the inflation energy scale , and we showed that the polonyi problem is still solved or greatly relaxed in high - scale inflation models @xcite . in this solution , we do not need any additional mechanism to dilute the polonyi abundance .
therefore , it may revive the conventional polonyi model as a realistic susy breaking model , which is fully compatible with the current experiments and observations , including the 125gev higgs boson .
in this letter we study the adiabatic solution in detail , considering various production processes of the polonyi field as well as the thermal and non - thermal gravitino production .
in particular , we focus on a minimal model in which only a certain coupling of the inflaton to the polonyi field is enhanced .
we also consider explicit inflation models to see if there is an allowed parameter space where the polonyi and gravitino problems are solved .
first we briefly review the cosmological polonyi problem in the gravity mediation .
let us denote the polonyi field by @xmath11 , which makes a dominant contribution to the susy breaking .
its @xmath12-term is given by @xmath13 where @xmath14 is the gravitino mass and @xmath15 is the reduced planck scale .
it generally couples to the mssm superfields as @xmath16 where @xmath17 and @xmath18 collectively denote the matter and gauge superfields , respectively , and @xmath19 and @xmath20 are constants of order unity . here and hereafter ,
@xmath19 and @xmath20 are taken to be real and positive , for simplicity .
these couplings give masses of order @xmath14 to the susy particles , as @xmath21 note that @xmath11 must be a singlet field in order to give a sizable mass to the gauginos . the following term in the khler potential yields the sizable @xmath22 and @xmath23 terms @xcite , @xmath24 as @xmath25 and @xmath26 .
thus the framework naturally solves the @xmath27 problem .
if one takes the gravitino mass to be as large as 10tev , the susy flavor and cp problems are greatly relaxed and the cosmological gravitino problem is also ameliorated .
it also explains the 125gev higgs boson without tuning the @xmath28-parameter @xcite .
therefore the @xmath8tev susy is plausible from these phenomenological point of view .
however , the model suffers from the cosmological polonyi problem , which inevitably arises in the gravity - mediation scenario .
the polonyi abundance is estimated as @xmath29 where @xmath30 is the initial amplitude , which is in general of the order of @xmath15 . the reheating temperature @xmath6 is defined by @xmath31 where @xmath32 is the inflaton decay rate , and @xmath33 counts the relativistic degrees of freedom at the reheating .
here we have assumed that the potential for @xmath11 can be well approximated by a quadratic term for @xmath34 , and that the @xmath11 starts to oscillate before the reheating .
the polonyi abundance ( [ rhoz ] ) is so large that the @xmath11 dominates the energy density of the universe soon after the reheating , and causes cosmological problems .
the polonyi decays into gauge bosons through the interaction ( [ kw ] ) with the decay rate given by @xmath35 where @xmath36 is the polonyi mass at the zero temperature .
the decay into gauginos is suppressed by @xmath37 or @xmath38 , where @xmath39 denotes the gaugino mass , and as we will see later , the polonyi mass is considered to be slightly enhanced compared to @xmath39 or @xmath14 .
we parametrize it as @xmath40 with @xmath41 .
the interaction ( [ mu ] ) induces the decay of the polonyi into the higgs boson pair @xcite , @xmath42 while the decay into a higgsino pair is suppressed by a factor of @xmath38 .
the polonyi also decays into a pair of gravitinos if kinematically allowed @xcite .
the decay rate is given by @xmath43 for example , if the decay into gauge bosons is the dominant decay mode , the lifetime of the polonyi is given by @xmath44 if the decay into the gravitino pair is dominant , the lifetime is given by @xmath45 the lifetime must be ( much ) shorter than 1sec in order not to spoil the success of bbn @xcite .
even if it decays before bbn , it dilutes the pre - existing baryon asymmetry of the universe .
the dilution factor is roughly given by @xmath46 , where @xmath47 is the polonyi decay temperature .
the dilution factor is so large that thermal and non - thermal leptogenesis scenarios do not work . therefore some involved mechanisms to create the baryon asymmetry
is required if the polonyi problem is solved by increasing the polonyi mass . in the next section
we will consider another attractive solution to the polonyi problem in which there is no late - time entropy production . and
@xmath6 plane from thermally produced gravitinos and the polonyi coherent oscillation for @xmath4810tev ( upper panel ) and @xmath49tev ( lower panel ) with @xmath50 .
we set @xmath51 .
the dotted line shows the upper bound from the gravitino thermal production .
the gray band shows the upper bound on the reheating temperature from the polonyi coherent oscillation and thermal production .
the upper edge of the band corresponds to @xmath52 and lower one to @xmath53 . in the lower panel ,
the constraint comes from the lsp overproduction , hence all the constraints disappear if the r - parity is broken by a small amount .
, title="fig : " ] and @xmath6 plane from thermally produced gravitinos and the polonyi coherent oscillation for @xmath4810tev ( upper panel ) and @xmath49tev ( lower panel ) with @xmath50 .
we set @xmath51 .
the dotted line shows the upper bound from the gravitino thermal production .
the gray band shows the upper bound on the reheating temperature from the polonyi coherent oscillation and thermal production .
the upper edge of the band corresponds to @xmath52 and lower one to @xmath53 . in the lower panel ,
the constraint comes from the lsp overproduction , hence all the constraints disappear if the r - parity is broken by a small amount .
, title="fig : " ]
now we revisit the polonyi model in light of the recent developments in the suppression mechanism for the moduli abundance @xcite .
let us introduce the inflaton fields @xmath54 and @xmath55 , which have @xmath56-charges of @xmath57 and @xmath58 , respectively .
the inflaton superpotential has the form @xmath59 where @xmath60 is some function of @xmath55 .
the @xmath12-term of @xmath54 dominates the potential energy during inflation .
many known inflation models in supergravity fall into this category . the polonyi field in general couples to the inflaton fields as @xmath61 where @xmath62 and @xmath63 are taken to be real and positive .
the adiabatic suppression mechanism works if @xmath64 @xcite . however , the inflaton dynamics just after the inflation induces a non - negligible amount of the coherent oscillations of the polonyi field , which is estimated as @xcite @xmath65 where @xmath66 and @xmath67 is the hubble scale at the end of inflation .
this expression is valid for @xmath68 . for @xmath69
, there remains a contribution like ( [ co ] ) with @xmath63 replaced by @xmath70 .
this is much smaller than the naive estimate ( [ rhoz ] ) if @xmath71 , which is satisfied for the most known inflation models . from this expression
, we can see that the polonyi abundance is suppressed for @xmath72 and large inflation scale .
hereafter we take @xmath73 for simplicity .
now let us see how the present model is constrained from cosmological arguments .
first , gravitinos are effectively produced at the reheating , and its abundance is proportional to the reheating temperature . if the gravitino is heavier than the lightest susy particle ( lsp ) , it is unstable and decays emitting energetic particles .
such late gravitino decay changes the helium-4 abundance @xcite , and produces lsps non - thermally .
the polonyi causes similar effects : the polonyi decay may alter the standard bbn results and yield too many lsps .
if the polonyi decays into the gravitino , the subsequent gravitino decay also causes similar effects .
notice that the polonyi abundance is given by the sum of the coherent oscillation ( [ co ] ) and thermal production , the latter of which is comparable to the abundance of the ( transverse components of ) gravitino if @xmath74 .
[ fig:1 ] shows constraints on the @xmath67 and @xmath6 plane from thermally produced gravitinos and the polonyi coherent oscillations and thermal production for @xmath4810tev and @xmath50 ( upper panel ) and @xmath49tev and @xmath75 ( lower panel ) .
the choice of relatively small @xmath20 is motivated by the existence of the focus - point region , and the polonyi mainly decays into gravitinos in this case .
the dotted line shows the upper bound on @xmath6 from the thermal production of gravitinos .
the gray band shows the upper bound on @xmath6 from the polonyi coherent oscillation and thermal production , and the width of the band represents uncertainty of the polonyi abundance and couplings .
the upper edge of the band corresponds to @xmath76 , while the lower one to @xmath53 , where @xmath40 .
the polonyi mass is varied because it is strongly coupled with the inflaton @xmath54 @xmath77 , and the polonyi self interaction of the form @xmath78 with @xmath79 is expected in the khler potential . with the present parameter choice ,
the bino is the lsp of mass @xmath80gev ( upper panel ) and @xmath81gev ( lower panel ) .
the thermal relic abundance of the bino lsp is not taken into account in fig .
[ fig:1 ] , because it strongly depends on the mass spectrum .
for instance , if it has a sizable mixing with higgsino or wino , the thermal relic abundance can be smaller than the present dm abundance .
( in the latter case , we need to relax the gut relation on the gaugino mass . )
we note that , in the lower panel , the constraint comes from the lsp overproduction from the gravitino / polonyi decay , hence all the constraints disappear if the r - parity is broken by a small amount .
it is seen that the reheating temperature of @xmath82gev is allowed for @xmath83@xmath84gev .
it is important that we do not need any additional late - time entropy production for solving the polonyi problem .
thus the conventional polonyi model for the gravity - mediation for relatively heavy susy scale of @xmath8tev can be compatible with leptogenesis scenario once we assume the polonyi coupling to the inflaton is enhanced .
now let us see if the above solution works in some known inflation models in supergravity . in particular , we will show that there are consistent parameter regions where thermal @xcite or non - thermal @xcite leptogenesis scenario works , avoiding the polonyi and gravitino problems . first , let us consider the susy hybrid inflation model @xcite .
the superpotential is given by @xmath85 where @xmath86 .
the waterfall fields , @xmath55 and @xmath87 , can be identified with the higgs fields which break u(1)@xmath88 gauge symmetry .
this model , including the constant term @xmath89 , was analyzed in detail in refs .
we assume that the inflaton dominantly decays into the right - handed neutrinos @xmath90 @xmath91 through the interaction @xmath92 the inflaton decay rate into the right - handed ( s)neutrino pair is given by @xmath93 where we have taken into account a mixing between @xmath54 and @xmath55 ( and @xmath94 ) due to the constant term @xcite . here
we consider only the decay into the lightest right - handed neutrino . on the other hand
, the inflaton decays into a pair of gravitinos through the interaction in the khler potential @xcite , @xmath95 the decay rate into the gravitino pair is given by @xcite @xmath96 where the mixing between @xmath54 and @xmath55 ( and @xmath94 ) is taken into account @xcite .
notice that the same interaction induces the inflaton decay into the polonyi pair @xmath97 with the same decay rate .
since each polonyi field mainly decays into a pair of the gravitino , the gravitino abundance produced non - thermally by the inflaton decay is given by @xmath98 where the total decay rate is approximately given by @xmath99 .
this imposes severe constraints on the parameter space .
we have scanned parameters @xmath100 , which are rewritten in terms of @xmath67 and @xmath6 through the relation @xmath101 and @xmath102 .
we have also fixed @xmath103 : the non - thermal leptogenesis works for @xmath104gev in this case .
[ fig : hybrid ] shows constraints in the @xmath67 and @xmath6 plane for the hybrid inflation model with @xmath4810tev ( upper panel ) and @xmath49tev ( lower panel ) .
the red dashed line shows the lower bound on @xmath6 from the cosmic string .
the blue band shows the lower bound on @xmath6 from the non - thermal gravitinos for @xmath105 ( upper edge ) and @xmath106 ( lower edge ) .
the meanings of the gray band and the black dotted line are same as fig .
[ fig:1 ] : they set upper bounds on @xmath6 from the polonyi and thermal gravitino .
the density perturbation with a correct magnitude is generated on the solid line labels by wmap normalization " .
it is seen that there is a consistent parameter regions around @xmath107gev and @xmath108gev where the polonyi problem is solved within the framework of susy hybrid inflation model .
note that the constraints from the polonyi and gravitinos disappear if the r - parity is broken slightly for @xmath109tev , as already explained . and @xmath6 plane for the hybrid inflation model with @xmath4810tev ( upper panel ) and @xmath49tev ( lower panel ) .
the red dashed line shows the lower bound on @xmath6 from the cosmic string .
the blue band shows the lower bound on @xmath6 from the non - thermal gravitinos for @xmath105 ( upper edge ) and @xmath106 ( lower edge ) .
the meanings of gray band and the black dotted line are same as fig .
[ fig:1 ] : they set upper bounds on @xmath6 from the polonyi and thermal gravitino .
the density perturbation with a correct magnitude is generated on the solid line denoted by wmap normalization " .
, title="fig : " ] and @xmath6 plane for the hybrid inflation model with @xmath4810tev ( upper panel ) and @xmath49tev ( lower panel )
. the red dashed line shows the lower bound on @xmath6 from the cosmic string .
the blue band shows the lower bound on @xmath6 from the non - thermal gravitinos for @xmath105 ( upper edge ) and @xmath106 ( lower edge ) .
the meanings of gray band and the black dotted line are same as fig .
[ fig:1 ] : they set upper bounds on @xmath6 from the polonyi and thermal gravitino .
the density perturbation with a correct magnitude is generated on the solid line denoted by wmap normalization " .
, title="fig : " ] , but for the smooth - hybrid inflation model .
the blue band shows the lower bound on @xmath6 from the non - thermal gravitinos for @xmath105 ( upper edge ) and @xmath110 ( lower edge ) .
, title="fig : " ] , but for the smooth - hybrid inflation model .
the blue band shows the lower bound on @xmath6 from the non - thermal gravitinos for @xmath105 ( upper edge ) and @xmath110 ( lower edge ) .
, title="fig : " ] let us consider the smooth - hybrid inflation model @xcite where the inflaton superpotential is given by @xmath111 where @xmath112 is an integer .
the model has a discrete symmetry @xmath113 under which @xmath114 has a charge @xmath115 and @xmath54 has a zero charge .
this model has an advantage that it does not suffer from problematic topological defects formation since the @xmath55 and @xmath87 have nonzero vevs during inflation and topological defects are inflated away .
hereafter we consider the case of @xmath116 for simplicity .
results do not much affected by this choice .
the gravitino abundance is similarly estimated by eq .
( [ ygrav ] ) .
the inflaton can decay into ordinary particles through non - renormalizable interactions , for example , @xmath117 with cutoff parameter @xmath118 . the decay rate into the higgs boson and higgsino pair is given by @xmath119 where the mixing between @xmath54 and @xmath55 ( and @xmath94 ) is taken into account .
if the right - handed neutrino mass is not much smaller than the inflaton mass , the decay rate into them through the operator @xmath120 is comparable to the above expression .
we have scanned the parameters @xmath121 , in the range @xmath122 so that the effective theory ( [ smooth ] ) below the scale @xmath123 remains valid , which are rewritten in terms of @xmath67 and @xmath6 through the relation @xmath124 and @xmath102 .
we have fixed @xmath125gev .
[ fig : smooth ] shows constraints on the @xmath67 and @xmath6 plane for the hybrid inflation model with @xmath4810tev ( upper panel ) and @xmath49tev ( lower panel ) .
the blue band shows the lower bound on @xmath6 from the non - thermal gravitinos for @xmath105 ( upper edge ) and @xmath110 ( lower edge ) .
the meanings of the gray band and the black dotted line are same as fig .
[ fig:1 ] : they set upper bounds on @xmath6 from the polonyi and thermal gravitino .
the wmap normalization for the density perturbation is satisfied on the solid line .
the scalar spectral index @xmath126 also fits well with the wmap result : @xmath127 @xcite .
it is seen that there is a consistent parameter regions around @xmath128gev and @xmath129gev where the polonyi problem and the gravitino problem are solved within the framework of smooth - hybrid inflation model .
note again that the constraints from the polonyi and gravitinos disappear if r - parity is broken slightly for @xmath109tev , as already explained .
we have revisited the polonyi model for gravity mediation with a relatively high - scale susy breaking of @xmath8tev .
the higgs boson mass of around 125gev indicated by the recent lhc data is naturally explained in this framework , while constraints from flavor / cp violating processes are alleviated . the model , however , is plagued with the notorious cosmological polonyi problem .
we have shown that the polonyi problem is solved once we assume the relatively enhanced coupling of the polonyi to the inflaton .
we have also considered explicit inflation models ( hybrid and smooth hybrid inflation ) , and shown that there is a parameter space where thermal and/or non - thermal leptogenesis scenarios work successfully , avoiding the polonyi and gravitino problems .
thus , our solution revives the conventional polonyi model as a realistic susy breaking model .
this work was supported by the grant - in - aid for scientific research on innovative areas ( no .
21111006 ) [ kn and ft ] , scientific research ( a ) ( no .
22244030 [ kn and ft ] , 21244033 [ ft ] , 22244021 [ tty ] ) , and jsps grant - in - aid for young scientists ( b ) ( no .
21740160 ) [ ft ] .
this work was also supported by world premier international center initiative ( wpi program ) , mext , japan .
y. okada , m. yamaguchi and t. yanagida , prog .
* 85 * ( 1991 ) 1 ; j. r. ellis , g. ridolfi and f. zwirner , phys . lett .
b * 257 * ( 1991 ) 83 ; h. e. haber and r. hempfling , phys . rev .
* 66 * ( 1991 ) 1815 .
y. okada , m. yamaguchi and t. yanagida , phys .
b * 262 * ( 1991 ) 54 .
g. f. giudice and a. strumia , nucl .
b * 858 * , 63 ( 2012 ) [ arxiv:1108.6077 [ hep - ph ] ] .
m. ibe and t. t. yanagida , arxiv:1112.2462 [ hep - ph ] .
j. l. feng , k. t. matchev and d. sanford , arxiv:1112.3021 [ hep - ph ] . g. f. giudice , m. a. luty , h. murayama and r. rattazzi , jhep * 9812 * , 027 ( 1998 ) [ hep - ph/9810442 ] ; l. randall and r. sundrum , nucl .
b * 557 * , 79 ( 1999 ) [ hep - th/9810155 ] .
t. moroi and k. nakayama , arxiv:1112.3123 [ hep - ph ] .
m. ibe , s. matsumoto and t. t. yanagida , arxiv:1202.2253 [ hep - ph ] .
m. fukugita , t. yanagida , phys .
* b174 * , 45 ( 1986 ) ; for a review , see w. buchmuller , r. d. peccei , t. yanagida , ann .
nucl . part .
sci . * 55 * , 311 - 355 ( 2005 ) .
[ hep - ph/0502169 ] .
w. buchmuller , p. di bari and m. plumacher , nucl .
b * 665 * , 445 ( 2003 ) [ hep - ph/0302092 ] .
g. f. giudice , a. notari , m. raidal , a. riotto and a. strumia , nucl .
b * 685 * , 89 ( 2004 ) [ hep - ph/0310123 ] .
j. l. feng , k. t. matchev and t. moroi , phys .
lett . * 84 * , 2322 ( 2000 ) [ hep - ph/9908309 ] ; phys .
d * 61 * , 075005 ( 2000 ) [ hep - ph/9909334 ]
. g. d. coughlan , w. fischler , e. w. kolb , s. raby , g. g. ross , phys .
* b131 * , 59 ( 1983 ) ; j. r. ellis , d. v. nanopoulos , m. quiros , phys .
* b174 * , 176 ( 1986 ) ; a. s. goncharov , a. d. linde , m. i. vysotsky , phys .
* b147 * , 279 ( 1984 ) .
k. nakayama , f. takahashi and t. t. yanagida , phys .
d * 84 * , 123523 ( 2011 ) [ arxiv:1109.2073 [ hep - ph ] ] .
k. nakayama , f. takahashi and t. t. yanagida , arxiv:1112.0418 [ hep - ph ] .
a. d. linde , phys .
rev . * d53 * , 4129 - 4132 ( 1996 ) .
[ hep - th/9601083 ] .
g. f. giudice and a. masiero , phys .
b * 206 * , 480 ( 1988 ) .
m. endo and f. takahashi , phys .
d * 74 * , 063502 ( 2006 ) [ hep - ph/0606075 ] .
m. endo , k. hamaguchi , f. takahashi , phys .
* 96 * , 211301 ( 2006 ) .
[ hep - ph/0602061 ] ; s. nakamura , m. yamaguchi , phys .
lett . * b638 * , 389 - 395 ( 2006 ) .
[ hep - ph/0602081 ] .
m. kawasaki , k. kohri , t. moroi , phys .
* d71 * , 083502 ( 2005 ) ; m. kawasaki , k. kohri , t. moroi , a. yotsuyanagi , phys .
rev . * d78 * , 065011 ( 2008 ) .
m. bolz , a. brandenburg , w. buchmuller , nucl .
phys . * b606 * , 518 - 544 ( 2001 ) .
[ hep - ph/0012052 ] .
g. lazarides and q. shafi , phys .
b * 258 * , 305 ( 1991 ) .
k. kumekawa , t. moroi and t. yanagida , prog .
phys . * 92 * , 437 ( 1994 ) [ hep - ph/9405337 ] .
t. asaka , k. hamaguchi , m. kawasaki and t. yanagida , phys .
b * 464 * , 12 ( 1999 ) [ hep - ph/9906366 ] ; phys .
d * 61 * , 083512 ( 2000 ) [ hep - ph/9907559 ] .
e. j. copeland , a. r. liddle , d. h. lyth , e. d. stewart and d. wands , phys .
d * 49 * , 6410 ( 1994 ) [ astro - ph/9401011 ] .
g. r. dvali , q. shafi and r. k. schaefer , phys .
lett . * 73 * , 1886 ( 1994 ) [ hep - ph/9406319 ] .
a. d. linde and a. riotto , phys .
d * 56 * , 1841 ( 1997 ) [ arxiv : hep - ph/9703209 ] .
w. buchmuller , l. covi and d. delepine , phys . lett .
b * 491 * , 183 ( 2000 ) [ hep - ph/0006168 ] .
v. n. senoguz and q. shafi , phys .
d * 71 * , 043514 ( 2005 ) [ hep - ph/0412102 ] ; m. ur rehman , v. n. senoguz and q. shafi , phys . rev .
d * 75 * , 043522 ( 2007 ) [ hep - ph/0612023 ] .
k. nakayama , f. takahashi and t. t. yanagida , jcap * 1012 * , 010 ( 2010 ) [ arxiv:1007.5152 [ hep - ph ] ] .
m. kawasaki , f. takahashi and t. t. yanagida , phys .
b * 638 * , 8 ( 2006 ) [ hep - ph/0603265 ] ; phys .
d * 74 * , 043519 ( 2006 ) [ hep - ph/0605297 ] .
g. lazarides and c. panagiotakopoulos , phys .
d * 52 * , r559 ( 1995 ) [ hep - ph/9506325 ] .
e. komatsu _ et al .
_ [ wmap collaboration ] , astrophys .
j. suppl .
* 192 * , 18 ( 2011 ) [ arxiv:1001.4538 [ astro-ph.co ] ] . | recent indications of the 125gev higgs at the lhc can be explained in a relatively high - scale susy scenario where the sparticle masses are multi - tev as is realized in the focus - point region .
however , it suffers from the notorious cosmological polonyi problem .
we argue that the polonyi problem is solved and thermal or non - thermal leptogenesis scenario works successfully , if a certain polonyi coupling to the inflaton is enhanced by a factor of 10 - 100 .
# 1#2#3 ( ) ^#3 ut-12 - 06 + tu-901 + ipmu12 - 0027 1.0 cm |
the study of the hydrodynamics of colloidal suspensions of passive particles is an old yet still active subject in soft condensed matter physics and chemical engineering . in recent years
there has been a growing interest in suspensions of active colloids @xcite , which exhibit rich collective behaviors quite distinct from those of passive suspensions .
there is a growing number of computational methods for modeling active suspensions @xcite , which are typically built upon well - developed techniques for passive suspensions in steady stokes flow , i.e. , at zero reynolds number .
since active particles typically have metallic subcomponents , they are often significantly denser than the solvent and sediment toward the bottom wall , making it necessary to address confinement and implement non - periodic boundary conditions in any method aimed at simulating experimentally - relevant configurations .
furthermore , since collective motions seen in active suspensions involve large numbers of particles , and since hydrodynamic interactions among particles decay slowly like the inverse of the distance , it is crucial to develop methods that can capture long - ranged hydrodynamic effects , yet still scale to tens or hundreds of thousands of particles . for suspensions of passive particles the methods of brownian @xcite and stokesian dynamics @xcite have dominated in chemical engineering , and related techniques have been used in biochemical engineering @xcite .
these methods simulate the overdamped ( diffusive ) dynamics of the particles by using green s functions for steady stokes flow to capture the effect of the fluid .
while this sort of implicit solvent approach works very well in many situations , it has several notable technical difficulties : achieving near linear scaling for many - particle systems is technically challenging , handling non - trivial boundary conditions ( bounded systems ) is complicated and has to be done on a case - by - case basis @xcite , generalizations to non - spherical ( and in particular complex ) particle shapes is difficult , and including thermal fluctuations is non - trivial due to the need to obtain stochastic increments with the desired covariance . in this work
we develop relatively low - accuracy but flexible and simple rigid multiblob methods that address these difficulties .
our approach builds heavily on a number of existing techniques , combining elements from several distinct but related methods .
we extensively test the proposed methods and study their accuracy and performance on a number of test problems .
the continuum formulation of the stokes equations with suitable boundary conditions on the surfaces of a collection of rigid particles is well - known and summarized in more detail in appendix [ sec : continuumformulation ] . due to the linearity of the stokes equations
, there is an affine mapping from the applied forces @xmath0 and
torques @xmath1 and any specified _ apparent slip _ velocity due to active boundary layers @xmath2 to the resulting particle motion given by the linear velocities @xmath3 and the angular velocities @xmath4 .
specifically , @xmath5=\m{\mathcal{n}}\left[\begin{array}{c } \v f\\ \v{\tau } \end{array}\right]-\breve{\m{\mathcal{n}}}\breve{\v u},\label{eq : n_def_cont}\ ] ] where @xmath6 is the _ mobility matrix _ , and @xmath7 is an _ active mobility _ linear operator .
the _ mobility problem _ consists of computing the rigid - body motion given the applied forces and torques and apparent slip .
the inverse of this problem is the _ resistance problem _ , which computes the forces and torques on the body given a specified motion of the body and active slip .
solving the mobility problem is a key component of any numerical method for modeling the deterministic or fluctuating ( brownian ) dynamics of the particles . in this paper
we develop a _ mobility solver _ for suspensions of rigid particles immersed in viscous fluid , specifically , we develop novel preconditioners for iterative solvers for the unknown motions of the rigid bodies , given the applied external forces and torques as well as active apparent slip on the surface of the particles . as we discuss in more detail in the body of the paper , our formulation can readily solve the resistance problem ; however , our iterative solvers will prove to be more scalable for mobility computations ( which are of primary interest ) than for resistance computations .
key to the success of our iterative solvers is the idea that instead of eliminating variables using _ exact _ schur complements and solving a _ reduced _ system iteratively , as done in the majority of existing methods @xcite , one should instead iteratively solve an _ extended _ system that includes all of the variables .
this has the key advantage that the matrix - vector product becomes an efficient direct calculation , and the schur complement can be computed only _ approximately _ and used to construct an effective preconditioner .
like many other authors , we construct rigid bodies of essentially arbitrary shape as a collection of rigidly - connected collection of `` blobs '' or `` beads '' forming a composite object @xcite that we will refer to as a _ rigid multiblob_. the hydrodynamic interactions between blobs are represented using a rotne - prager tensor generalized to the desired domain geometry ( boundary conditions ) @xcite , specifically , we use the the rotne - prager - yamakawa ( rpy ) tensor @xcite for an unbounded domain , and the rotne - prager - blake ( rpb ) tensor @xcite for a half - space domain . in section
[ sec : rigidmultiblobs ] we describe how to obtain the hydrodynamic coupling between a large collection of rigid multiblobs by solving a large linear system for lagrange multipliers enforcing the rigidity .
a key contribution of our work is to develop an indefinite saddle - point preconditioner for iterative solution of the resulting linear system .
this preconditioner is based on a block - diagonal approximation of the blob - blob mobility matrix , in which all hydrodynamic interactions among distinct bodies ( more precisely , among blobs on distinct bodies ) are neglected .
the only system - specific component is the implementation of a fast matrix - vector multiplication routine , which in turn requires a scalable method for multiplying the rpy mobility matrix by a vector . for simple geometries such as an unbounded domain or particles sedimented next to a no - slip boundary , simple analytical formulas for the rpy tensor
are well - known @xcite , and can be used to construct an efficient matrix - vector multiplication routine , for example , using fast multipole methods ( fmms ) @xcite , or even direct summation on a gpu .
we numerically study the performance and accuracy of the rigid multiblob methods for suspensions in an unbounded domain in section [ sec : resultsunbounded ] , and study particles sedimented near a no - slip boundary in section [ sec : resultswall ] .
we find that resolving spherical particles with twelve blobs placed on the vertices of an icosahedron @xcite is notably more accurate than the fts ( force - torque - stresslet plus degenerate quadrupole ) truncation typically employed in stokesian dynamics simulations , provided that the effective hydrodynamic radius of the rigid multiblob is adjusted to account for the finite size of the blobs .
we also find that a small number of iterations of a krylov method are required to solve the required linear system , and importantly , the number of iterations is constant _ independent _ of the the number of rigid bodies , making it possible to develop a linear or near - linear scaling algorithm . for _ resistance problems _
, however , we observe a number of iterations growing at least as fast as the linear dimensions of the system .
this is consistent with similar studies of iterative solvers for stokesian dynamics by ichiki @xcite . for confined systems , however , even in the simplest case of a periodic system ,
the green s function for stokes flow and the associated rpy tensor is difficult to obtain in closed form , and when it is possible to write an analytical result , the resulting formulas are typically based on infinite series that are expensive to evaluate . for periodic systems
this is commonly addressed by using ewald summation @xcite based on the fast fourier transform ( fft ) @xcite ; the present state - of - the - art for stokes flow is the spectral ewald method @xcite , which has recently been used for stokesian dynamics simulations of periodic suspensions @xcite .
a key deficiency of most existing methods is that they rely critically on having triply periodic domains and the use of the fft .
generalizing these methods to non - periodic domains while keeping their linear scaling requires a large development effort and typically a new implementation for every different geometry @xcite .
furthermore , in a number of applications involving active particles @xcite , there is a surface slip ( e.g. , electrohydrodynamic or osmophoretic flow ) induced on the bottom boundary due to the gradients created by the particles , and this slip drives or at least strongly affects the motion of the particles .
accounting for this slip requires solving an additional equation such as a poisson or laplace equation for the electric potential or concentration of chemical fuel with nontrivial boundary conditions on the particle and wall surfaces .
the solution of this additional equation provides the slip boundary condition for the stokes equations , which must be solved to find the resulting fluid flow and active particle motion .
such nontrivial multi - physics coupling is quite hard to accomplish in existing methods .
to address these difficulties , in section [ sec : rigidibamr ] we develop a method for general cuboidal confined domains which does not require analytical green s functions .
this relies on an immersed boundary ( ib ) method for obtaining an approximation to the rpy tensor in confined geometries , as recently developed by some of us @xcite .
this technique has been combined with the concept of multiblob representation of rigid bodies in a follow - up work @xcite , but in this work stiff elastic springs were used to enforce the rigidity . by contrast , we ensure the rigidity of the multiblobs via lagrange multipliers which are solved concurrently with solving for the fluid pressure and velocity .
our key novel contribution is an effective preconditioner for the rigidly - constrained stokes problem in periodic and non - periodic domains , obtained by combining our recently - developed preconditioner for a rigid - body ib method @xcite with a block - diagonal preconditioner for the mobility subproblem . in the ib method developed in section [
sec : rigidibamr ] and studied numerically in section [ sec : resultsconfined ] , analytical green s functions are replaced by an `` on the fly '' computation carried out by a standard finite - volume fluid solver .
this stokes solver can readily handle nontrivial boundary conditions , for example , slip along the walls @xcite can easily be accounted for .
furthermore , suspensions at small but nonzero reynolds numbers can be handled with little extra work @xcite .
additionally , we avoid uncontrolled approximations relying on truncations of multipole expansions to a fixed order @xcite , and we can seamlessly handle arbitrary body shapes and deformation kinematics .
lastly , and importantly , in the spirit of fluctuating hydrodynamics @xcite , it is straightforward to generate the stochastic increments required to simulate the brownian motion of small rigid particles suspended in a fluid by including a fluctuating stress in the fluid equations , as we will discuss in more detail in future work ; here we focus on the deterministic mobility and resistance problems . at the same time
, our method also has some disadvantages compared to methods such as boundary integral or boundary element methods .
notably , it requires filling the domain with a dense uniform fluid grid , which is expensive at low densities .
it is also a low - order method that can not compute solutions as accurately as spectral boundary integral formulations .
we do believe , nevertheless , that the method developed here offers a good compromise between accuracy , efficiency , scalabilty , flexibility and extensibility , compared to other more specialized formulations .
we apply our methods to a number of test problems for which analytical solutions are known , but also study a few nontrivial problems that have not been properly addressed in the literature . in section [ sub : cylinderwall ]
we study the mobility of a cylinder of finite aspect ratio that is parallel to a no - slip boundary and compare to experimental measurements and asymptotic theory based on a slender - body approximation . in section
[ sub : activepair ] we study the formation of a stable rotating pair of active `` extensor '' or `` pusher '' nanorods next to a no - slip boundary , and confirm the direction of rotation observed in recent experiments @xcite . in section [ sub : boomerang ] we compute the effective diffusion coefficient of a boomerang - shaped colloid in a slit channel , and compare to recent experimental measurements @xcite .
in section [ sub : binarysedimentation ] we study the mean and variance of the sedimentation velocity in a binary suspension of spheres of size ratio two , and compare to recent stokesian dynamics simulations @xcite .
in this section we develop the rigid multiblob model of colloidal particles at zero reynolds number .
the kind of models we use here are not new , but we present the method in detail instead of relying on previous presentations , the most relevant of which are those of swan _ et al . _ @xcite .
this is in part to present the formulation in our notation , and in part to explain the differences with other closely - related methods .
our key novel contribution in this section is the preconditioned iterative solver described in section [ sub : preconditioner ] ; the performance and scaling of our mobility solver is studied numerically for unbounded domains in section [ sub : convergencefmm ] , and for particles confined near a single wall in section [ sub : convergencewall ] .
the modeling of suspensions of rigid spheres at small reynolds numbers is a well - developed field with a long history . a powerful class of methods
are related to brownian dynamics with hydrodynamic interactions ( bdhi ) @xcite and stokesian dynamics ( sd ) @xcite ( note that these terms are used differently in different communities ) .
the difference between these two ( as we define them here ) is that bdhi uses what we call a minimally - resolved model @xcite in which each colloid ( for colloidal suspensions ) or polymer bead ( for polymeric suspensions ) is only resolved at the monopole level , more precisely , at the rotne - prager level @xcite . by contrast , in sd the next level in a multipole expansion is taken into account and torques and stresslets are also accounted for .
it has been shown recently that yet one more order needs to be kept in the multipole expansion to model suspensions of active spheres @xcite , and a suitable galerkin truncation of the multipole hierarchy has been developed for active spheres in unbounded domains @xcite , as well as for active spheres confined near a no - slip boundary @xcite .
it is also possible to account for higher - order multipoles @xcite , leading to more complicated ( and computationally expensive ) but also more accurate models .
it has also been shown that multipole expansions converge very poorly for nearly touching spheres due to the divergence of the lubrication forces , and in most methods for dense colloidal suspensions of hard spheres pairwise lubrication corrections are added in a somewhat _ ad hoc _ manner ; we will refer to this approach as sd with lubrication . given the well - developed tools for modeling sphere suspensions , it is natural to leverage them when modeling suspensions of particles of more complex shapes .
here we describe a technique capable of , in principle , modeling passive rigid particles of arbitrary shape .
the method can also be used to model , without any extra effort , active particles with active slip layers , i.e. , particles which are phoretic ( e.g. , osmo - phoretic , electro - phoretic , chemo - phoretic , etc . ) due to an apparent slip at their surface . for the purposes of hydrodynamic calculations ,
we discretize rigid bodies by constructing them out of multiple rigidly - connected spherical `` blobs '' or beads of hydrodynamic radius @xmath8 .
these blobs can be thought of as hydrodynamically minimally - resolved spheres forming a rigid conglomerate that approximates the hydrodynamics of the actual rigid object being studied .
we prefer the word `` blob '' over `` sphere '' or `` point '' or `` monopole '' because blobs are not spheres as they do not have a well - defined surface like spheres do , they have a finite size associated with them ( the hydrodynamic blob radius @xmath8 ) unlike points , and they account for a degenerate quadrupole associated to the faxen corrections in addition to a force monopole . the word `` bead '' is also appropriate , but we prefer to reserve that for polymer models ( bead - spring or bead - link models ) .
examples of `` multiblob '' @xcite models of two types of colloidal particles are illustrated in fig .
[ fig : blobmodels ] . in the leftmost panel
, we show a minimally - resolved model of a rigid rod , with dimensions similar to active metallic `` nanorods '' used in recent experiments @xcite . in this minimally - resolved model the blobs , shown as spheres with radius equal to @xmath8 ,
are placed in a row along the axes of the cylinder .
such minimally - resolved models are particularly suited for cylinders of large but finite aspect ratio ; for very thin rods such as actin filaments boundary integral methods based on slender - body theory @xcite will be more effective . in the more resolved model illustrated in the second panel from the left , a hexagon of blobs is placed around the circumference of the cylinder to better resolve it . a yet more resolved model with a dodecagon of blobs around the cylinder circumference is shown in the third panel from the left . in the rightmost panel of fig .
[ fig : blobmodels ] we show a blob model of a colloidal boomerang with a square cross - section , as manufactured using lithography and studied in @xcite . similar `` bead '' or `` raspberry '' models appear in a number of studies of hydrodynamics of particle suspensions @xcite . in many studies ,
stiff elastic springs between the blobs are used to keep the structure rigid ; in some models the fluid or particle inertia is included also . here
, we keep the structures _ strictly rigid _ and refer to the resulting structures as _ rigid multiblob _ models .
such rigid multiblob models have been used in a number of prior studies @xcite , but we refer to @xcite for a detailed exposition . our primary focus in this section will be to develop algorithmic techniques that allow suspensions of tens or even hundreds of thousands of rigid multiblob particles to be simulated efficiently .
this is in many ways primarily an exercise in numerical linear algebra , but one that is _ necessary _ to make the rigid multiblob approach useful for simulating moderately dense suspensions .
a second goal , which will be realized in the results sections of this paper , will be to carefully assess the accuracy of rigid multiblob models as a function of their resolution ( number of blobs per body ) .
we now summarize the main equations used to solve the mobility and resistance problems for a collection of rigid multiblobs immersed in a viscous fluid .
we first discuss the hydrodynamic interaction between blobs , and then discuss the hydrodynamic interactions between rigid bodies . in the notation
used below , we will use the latin indices @xmath9 for individual blobs , and reserve latin indices @xmath10 for bodies .
we will denote with @xmath11 the set of blobs comprising body @xmath12 .
we will consider a suspension of @xmath13 rigid bodies with a chosen reference _
tracking point _ on body @xmath12 having position @xmath14 , and the orientation of body @xmath12 relative to a _ reference configuration _ represented by the quaternion @xmath15 @xcite . the linear velocity of ( the chosen tracking point on ) body @xmath12 will be denoted with @xmath16 , and its angular velocity will be denoted with @xmath17 .
the total force applied on body @xmath12 is @xmath18 , and the total torque is @xmath19 .
the composite configuration vector of position and orientation of body @xmath12 will be denoted with @xmath20 , the composite vector of linear and angular velocity will be denoted with @xmath21 , and the composite vector of forces and torques with @xmath22 .
the position of blob @xmath23 will be denoted with @xmath24 , and its velocity will be denoted with @xmath25 .
when not subscripted , vectors will refer to the composite vector formed by all bodies or all blobs on all bodies .
for example , @xmath26 will denote the linear and angular velocities of all bodies , and @xmath27 will denote the positions of all of the blobs .
we will use a superscript to denote portions of composite vectors for all blobs belonging to one body , for example , @xmath28 will denote the vector of positions of all blobs belonging to body @xmath12 .
the fact that the multiblob @xmath12 is rigid is expressed by the `` no - slip '' kinematic condition , @xmath29 this no - slip condition can be written for all bodies succinctly as @xmath30 where @xmath31 is a simple geometric matrix @xcite .
we will denote the apparent velocity of the fluid at point @xmath24 with @xmath32 .
for a _ passive blob _ , i.e. , a blob that represents a passive part of the rigid particle , the _ no - slip _ boundary condition requires that @xmath33 .
however , for _ active blobs _ an additional apparent slip of the fluid relative to the surface of the body can be imposed , resulting in a nonzero _ slip _ @xmath34 .
this kind of active propulsion is termed `` implicit swimming gait '' by swan and brady @xcite .
an `` explicit swimming gait '' @xcite can be taken into account without any modifications to the formulation or algorithm by simply replacing ( [ eq : noslip_rigid ] ) with @xmath35 that is , the only difference between `` slip '' and `` deformation '' is whether the blobs move relative to the rigid body frame dragging the fluid along , or stay fixed in the body frame while the fluid passes by them .
one can of course even combine the two and have the blobs move relative to the rigid body while also pushing flow , for example , this can be used to model an active filament where there is slip along the filament but the filament itself is moving . in the end ,
the only thing that matters to the formulation is the velocity difference @xmath36 in appendix [ app : permeable ] we explain how to model permeable ( porous ) bodies by making the apparent slip proportional to the fluid - blob force @xmath37 .
the fundamental problem tackled in this paper is the solution of the _ mobility problem _ , that is , the computation of the motion of the bodies given the applied forces and torques on the bodies and the slip velocity .
because of the linearity of the stokes equations and the boundary conditions , there exists an affine linear mapping @xmath38 where the _ body mobility matrix _ @xmath39 depends on the configuration and is the central object of the computation . the _ active mobility matrix _ @xmath40 is a discretization of the active mobility operator @xmath7 , and gives the active motion of force- and torque - free particles .
note that @xmath40 is related to , but different from , the propulsion matrix introduced in @xcite .
the propulsion matrix is essentially a finite - dimensional projection of the operator @xmath7 that only depends on the choice of basis functions used to express the surface slip velocity @xmath41 , and does not depend on the specific discretization of the body or quadrature rules , as does @xmath40 . in the remainder of this section
we develop a method for computing @xmath26 given @xmath42 and @xmath43 , i.e. , a method for computing the combined action of @xmath6 and @xmath40 , for large collections of non - overlapping rigid particles .
we will also briefly discuss the _ resistance problem _ , in which we are given the motion of the bodies as a specified kinematics , and seek the resulting drag forces and torques , which have the form @xmath44 where the _ body resistance matrix _ @xmath45 and @xmath46 is the _ active resistance matrix_. the blob - blob translational mobility matrix @xmath47 describes the hydrodynamic interactions between the @xmath48 blobs , accounting for the influence of the boundaries . specifically ,
if the blobs are free to move ( i.e. , not constrained rigidly ) with the fluid under the action of set of translational forces @xmath49 , the translational velocities of the blobs will be @xmath50 the mobility matrix @xmath47 is a block matrix of dimension @xmath51 , where @xmath52 is the dimensionality .
the @xmath53 block @xmath54 computes the velocity of blob @xmath55 given the force on blob @xmath56 , neglecting the presence of the other blobs in a _ pairwise _ approximation . to construct a suitable @xmath47 , we can think of blobs as spheres of hydrodynamic radius @xmath8 . for two well - separated spheres @xmath55 and @xmath56 of radius @xmath8 we have the far - field approximation * * @xcite @xmath57 where @xmath58 is the fluid viscosity and @xmath59 is the green s function for the steady stokes problem with unit viscosity , with the appropriate boundary conditions such as no - slip on the boundaries of the domain .
the differential operator @xmath60 is called the faxen operator @xcite .
note that the form of ( [ eq : mobilityfaxen ] ) guarantees that the mobility matrix is symmetric positive semidefinite ( spd ) by construction since @xmath59 is an spd kernel .
for a three dimensional unbounded domain with fluid at rest at infinity , the green s function is isotropic and given by the oseen tensor , @xmath61 using this expression in ( [ eq : mobilityfaxen ] ) yields the far - field component of the rotne - prager - yamakawa ( rpy ) tensor @xcite , commonly used in bdhi .
a correction needs to be introduced when particles are close to each other to ensure an spd mobility matrix @xcite , which can be derived by using an integral form of the rpy tensor valid even for overlapping particles @xcite , to give @xmath62 where @xmath63 , and @xmath64 the diagonal blocks of the mobility matrix , i.e. , the self - mobility can be obtained by setting @xmath65 to obtain @xmath66 , which matches the stokes solution for the drag on a translating sphere ; this is an important continuity property of the rpy tensor @xcite .
we will use the rpy tensor ( [ eq : rpytensor ] ) for simulations of rigid - particle suspensions in unbounded domains in section [ sec : resultsunbounded ] . in principle
, it is possible to generalize the rpy tensor to any flow geometry , i.e. , to any boundary conditions ( and imposed external flow ) @xcite , including periodic domains @xcite , as well as confined domains @xcite .
however , we are not aware of any tractable analytical expressions for the complete rpy tensor ( including near - field corrections ) even for the simplest confined geometry of particles near a single no - slip boundary . in the presence of a single no - slip wall , an analytic approximation to @xmath67
is given by swan and brady @xcite ( and re - derived later in @xcite ) as a generalization of the rotne - prager ( rp ) tensor @xcite to account for the no - slip boundary using blake s image construction @xcite .
as shown in ref .
@xcite , the corrections to the rotne - prager tensor ( [ eq : mobilityfaxen ] ) for particles that overlap each other but not the wall are independent of the boundary conditions , and are thus given by the standard rpy expressions ( [ eq : rpytensor ] ) for unbounded domains .
therefore , in section [ sec : resultswall ] we compute @xmath68 by adding to the rpy tensor ( [ eq : rpytensor ] ) wall corrections corresponding to the translation - translation part of the rotne - prager - blake mobility given by eqs . ( b1 ) and ( c2 ) in @xcite , ignoring the higher order torque and stresslet terms in the spirit of the minimally - resolved blob model .
the expressions derived by swan and brady @xcite assume that neither particle overlaps the wall and the resulting expressions are not guaranteed to lead to an spd @xmath68 if one or more blobs overlap the wall , as we discuss in more detail in the conclusions .
for more complicated geometries , such as a slit or a square ( duct ) channel , analytical computations of the green s function become quite complicated and tedious , and numerical computations typically require pre - tabulations @xcite . in section [ sec : resultsconfined ] we explain how a grid - based finite volume stokes solver can be used to obtain the action of the green s function and thus compute the action of the mobility matrix for confined domains , for essentially arbitrary combinations of periodic , free - slip , no - slip , or stress boundary conditions .
after discretizing the rigid bodies as rigid multiblobs , we can write down a system of equations that constrain the blobs to move rigidly in a straightforward manner . letting @xmath37 be a vector of forces ( lagrange multipliers ) that acts on each blob to enforce the rigidity of the body , we have the following linear system for @xmath37 , @xmath3 , and @xmath4 for all bodies @xmath12 , @xmath69 the first equation is the no - slip condition obtained by combining ( [ eq : w_mlambda ] ) and ( [ eq : noslip_rigid ] ) .
the second and third equations are the force and torque balance conditions for body @xmath12 .
note that the physical interpretation of @xmath37 is that of a total force on the portion of the surface of the body associated with a given blob .
if one wants to think of ( [ eq : rigidsystem ] ) as a regularized discretization of the first - kind integral equation ( [ eq : first_kind ] ) and obtain a pointwise value of the traction force _ density _ , one should divide @xmath70 by the surface area @xmath71 associated with blob @xmath56 , which plays the role of a quadrature weight @xcite ; we will discuss more sophisticated quadrature rules @xcite in the conclusions .
we can write the _ mobility problem _ ( [ eq : rigidsystem ] ) in compact matrix notation as a _ saddle - point _ linear system of equations for the rigidity forces @xmath37 and unknown motion @xmath26 , @xmath72\left[\begin{array}{c } \v{\lambda}\\ \v u \end{array}\right]=\left[\begin{array}{c } \slip\\ -\v f \end{array}\right].\label{eq : saddle_m}\ ] ] forming the schur complement by eliminating @xmath37 we get ( see also eq . ( 1 ) in @xcite or eq .
( 32 ) in @xcite ) @xmath73 where the body mobility matrix @xmath6 is @xmath74 and is evidently spd since @xmath47 is .
although written in this form using the inverse of @xmath47 , unlike in a number of prior works @xcite , we obtain @xmath26 by solving ( [ eq : saddle_m ] ) directly using an iterative solver , as we explain in more detail in section [ sub : preconditioner ] .
we note that one can compute a fluid velocity field @xmath75 from @xmath37 using a procedure we describe in appendix [ app : renderflow ] .
the _ resistance problem _ , on the other hand , consists of solving for @xmath37 in @xmath76 and then computing @xmath77 , giving @xmath78 at first glance , it appears that solving the resistance system ( [ eq : resistance_problem ] ) is easier than solving the saddle - point problem ( [ eq : saddle_m ] ) ; however , as we explain in more detail in section [ sub : convergencefmm ] , the mobility problem is significantly easier to solve using iterative methods than the resistance problem , consistent with similar observations in the context of stokesian dynamics @xcite . observe that the saddle - point formulation ( [ eq : saddle_m ] ) applies more broadly to _ mixed _
mobility / resistance problems , where some of the rigid body degrees of freedom are constrained but some are free @xcite .
an example is a suspension of spheres being rotated by a magnetic field at a specified angular velocity but free to move translationally , or a suspension of colloids fixed in space by strong laser tweezers but otherwise free to rotate , or even a hinged body that can only move in a partially - constrained manner . in cases such as these
we simply redefine @xmath26 to contain the free kinematic degrees of freedom and modify the definition of the kinematic matrix @xmath79 .
much of what we say below continues to apply , but with the caveat that the expected speed of convergence of iterative methods is expected to depend on the nature of the imposed constraints , as we discuss in section [ sub : convergencefmm ] .
note that the formula ( [ eq : bodymob ] ) is somewhat formal , and in practice all inverses should be replaced by pseudo - inverses .
for instance , in the limit when infinitely many blobs cover the surface of a body , the mobility matrix @xmath47 is not invertible since making @xmath37 perpendicular to the surface will not yield any flow because it will try to compress the ( fictitious ) incompressible fluid inside the body .
note that this nontrivial null space of the mobility poses no problem when using an iterative method to solve ( [ eq : saddle_m ] ) because the right hand side is in the proper range due to the imposition of the volume - preservation constraint ( [ eq : slip_solvability ] ) .
it is also possible that the matrix @xmath80 is not invertible .
a typical example for this is the minimally - resolved cylinder shown in the left - most panel of fig .
[ fig : blobmodels ] . because all of the forces @xmath37 are applied exactly on the semi - axes of the cylinder
, they can not exert a torque around the symmetry axes of the rod .
again , there is no problem with iterative solvers for ( [ eq : saddle_m ] ) if the applied force is in the appropriate range ( e.g. , one should not apply a torque around the semi - axes of a minimally - resolved cylinder ) . for a small number of blobs , the equation ( [ eq : saddle_m ] )
can be solved by direct inversion of @xmath47 , as done in most prior works . for large systems , which is the focus of our work ,
iterative methods are required . a standard approach used in the literature
is to eliminate one of the variables @xmath37 or @xmath26 . eliminating @xmath37 leads to the equation
@xmath81 which requires the action of @xmath82 , which must itself be obtained inside a nested iterative solver , increasing both the complexity and the cost of the method .
swan and wang @xcite have recently used the conjugate gradient method to solve ( [ eq : symmetric_1 ] ) , preconditioning using the block - diagonal matrix @xmath83 . an alternative is to write an equivalent system to ( [ eq : saddle_m ] ) , for an arbitrary constant @xmath84 , @xmath85\left[\begin{array}{c } \v{\lambda}\\ \v u \end{array}\right]=\left[\begin{array}{c } \slip\\ -\left(\v f+c\m{\mathcal{k}}^{t}\slip\right ) \end{array}\right],\label{eq : saddle_m_eq}\ ] ] from which we can easily eliminate @xmath26 to obtain an equation for @xmath37 only , in the form * @xmath86\v{\lambda}=\mathrm{rhs},\label{eq : symmetric_2}\ ] ] * where we omit the full expression for the right hand side for brevity .
the system ( [ eq : symmetric_2 ] ) can now be solved using ( preconditioned ) conjugate gradients , and only requires the inverse of the simpler matrix @xmath87 .
note that , although not presented in this way , this is the essence of the approach that is followed and recommended by swan _ et al_. @xcite ( see appendix @xcite and note that @xmath88 is denoted by @xmath89 in that paper ) ; they recommend computing the action of @xmath90 by an iterative method preconditioned by an incomplete cholesky factorization .
a similar approach is followed in boundary integral formulations ( which are usually formulated using a double layer density ) , where a continuum operator related to @xmath91 is computed and then discretized using a quadrature rule @xcite .
in contrast to the approaches taken by swan _
et al_. @xcite , we have found that numerically the best approach to solving for the unknown rigid - body motions of the particles is to solve the extended saddle - point problem ( [ eq : saddle_m ] ) for _ both _ @xmath26 and @xmath37 _ directly _ , using a preconditioned iterative krylov method .
in fact , as we will demonstrate in the results section of this paper , such an approach has computational complexity that is essentially linear in the number of blobs because the number of iterations required to solve ( [ eq : saddle_m ] ) is quite modest when an appropriate preconditioner , described below , is used .
this approach does not require computing ( the action of ) @xmath90 and leads to a very simple implementation .
a krylov solver for ( [ eq : saddle_m ] ) requires two components : 1 .
an efficient algorithm for performing the matrix - vector product , which in our case amounts to a fast method to multiply the dense but low - rank mobility matrix @xmath47 by a vector of blob forces @xmath37 .
2 . a suitable preconditioner , which is an approximate solver for ( [ eq : saddle_m ] ) . how to efficiently compute @xmath92 depends very much on the boundary conditions and thus the form of the green s function used to construct @xmath47 . for unbounded domains , in this work we use the fast multipole method ( fmm ) developed specifically for the rpy tensor in @xcite ; alternative kernel - independent fmms
could also be used , and have also been generalized to periodic domains @xcite .
the fmm method has an essentially linear computational cost of @xmath93 for a single matrix - vector multiplication . in the simulations presented here we use a fixed and rather tight relative tolerance for the fmm @xmath94 throughout the iterative solution process .
krylov methods , however , allow one to _ lower _ the accuracy of the matrix - vector product as the residual is reduced @xcite ; this has recently been used to lower the cost of fmm - based boundary integral methods @xcite
. we will explore such optimizations in future work . for rigid particles sedimented near a single no - slip wall
, we have implemented a graphics processing unit ( gpu ) based direct summation matrix - vector product based on the rotne - prager - blake tensor derived by swan and brady @xcite .
this has , asymptotically , a quadratic computational cost of @xmath95 ; however , the computation is trivially parallel so the multiplication is remarkably fast even for one million blobs because of the very large number of threads available on modern gpus
. gimbutas _ et al_. have recently developed an fmm method for the blake tensor by using a simple image construction ( image stokeslet plus a harmonic scalar correction ) and applying an infinite - space fmm method to the extended system of singularities @xcite .
however , this construction has not yet been generalized to the rotne - prager - blake tensor , and , furthermore , the fmm will not be more efficient than the direct product on gpus in practice unless a large number of blobs is considered . for fully confined domains , we will adopt an extended saddle - point formulation that will be described in section [ sec : resultsconfined ] . in this work
we demonstrate that a very efficient yet simple preconditioner for ( [ eq : saddle_m ] ) is obtained by neglecting hydrodynamic interactions between different bodies , that is , setting the elements of @xmath47 corresponding to pairs of blobs on _ distinct _ bodies to zero in the preconditioner .
this amounts to making a block - diagonal approximation of the mobility @xmath96 defined by only keeping the diagonal blocks corresponding to a single body interacting _ only _ with the boundaries of the domain , @xmath97 we will demonstrate here that the _ indefinite block - diagonal _ preconditioner , @xmath98,\label{eq : indef_block_p}\ ] ] is a very effective preconditioner for solving ( [ eq : saddle_m ] ) . applying the preconditioner ( [ eq : indef_block_p ] ) amounts to solving the linear system @xmath99\left[\begin{array}{c } \v{\lambda}\\ \v u \end{array}\right]=\left[\begin{array}{c } \slip\\ -\v f \end{array}\right],\label{eq : saddle_m_precon}\ ] ] which is quite easy to do since the approximate body mobility matrix ( schur complement ) , @xmath100 is itself a block - diagonal matrix where each block on the diagonal refers to a single body neglecting all hydrodynamic interactions with other bodies , @xmath101 computing @xmath102 requires a dense matrix inversion ( e.g. , cholesky factorization ) of the much smaller mobility matrix @xmath103 , whose size is @xmath104 , where @xmath105 is the number of blobs on body @xmath12 . in the case of an infinite domain
, the factorization of @xmath103 can be precomputed once at the beginning of a dynamic simulation and reused during the simulation due to the rotational and translational invariance of the rpy tensor ; one only needs to apply rotation matrices to the right - hand side and the result to convert between the original reference configuration of the body and the current configuration .
furthermore , particles of the same shape and size discretized with the same number of blobs as body @xmath12 can share a single factorization of @xmath103 and @xmath106 . in cases where @xmath103 depends in a nontrivial way on the position of the body , as for ( partially ) confined domains
, one needs to factorize @xmath103 for all bodies @xmath12 at every time step ; this factorization can still be reused during the iterative solve in each application of the preconditioner . because our preconditioner is indefinite , one can not use the preconditioned conjugate gradient ( pcg ) krylov method to solve ( [ eq : saddle_m ] ) without modification .
one of the most robust iterative methods , which we use in this work , is the generalized minimum residual method ( gmres ) .
the key advantage of gmres is that it is guaranteed to reduce the residual from iteration to iteration .
its main downside is that it requires storing a large number of intermediate vectors ( i.e. , the history of the iterates ) .
gmres also can stall , although this can be corrected to some extent by restarts .
an alternative to gmres is the ( stabilized ) bi - conjugate gradient ( bicg(stab ) ) method , which works for non - symmetric matrices as well . in our implementation
we have relied on the petsc library @xcite for iterative solvers ; this library makes it very easy to experiment with different iterative solvers .
the rigid multiblob method described in section [ sec : rigidmultiblobs ] requires a technique for multiplying the blob - blob mobility matrix with a vector .
therefore , this approach , like all other green s function based methods @xcite , is very geometry - specific and does not generalize easily to more complicated boundary conditions . to handle geometries for which there is no simple analytical expression for the green s function , such as slit or square channels , pre - tabulation of the green s function is necessary , and ensuring a positive semi - definite mobility matrix is in general difficult .
another difficulty with green s function based methods is that including a `` background '' flow is only simple when this flow can be computed easily analytically , such as simple shear flows .
but for more complicated geometries , such as poiseuille flow through a square channel , computing the base flow is itself not trivial or requires evaluating expensive infinite - series solutions .
an alternative approach is to use a traditional stokes solver to solve the fluid equations numerically @xcite .
this requires filling the domain with a grid , which can increase the number of degrees of freedom considerably over just discretizing the surface of the immersed bodies .
however , the number of fluid degrees of freedom can be held approximately constant as more bodies are included , so that the methods typically scale very well with the number of particles and are well - suited to dense particle suspensions .
previous work @xcite has shown how to use an immersed boundary ( ib ) method @xcite to obtain the action of the green s function in complex geometries . in this approach , spherical particles
are minimally resolved using only a single blob per particle . in subsequent work
this approach was extended to multiblob models @xcite , but the rigidity constraint was imposed only approximately using stiff springs , leading to numerical stiffness .
a class of related minimally - resolved methods based on the force coupling method ( fcm ) @xcite can include also torques and stresslets , as well as particle activity @xcite , but a number of these methods have relied strongly on periodic boundaries since they use the fast fourier transform ( fft ) to solve the ( fluctuating ) stokes equations . in recent work
@xcite , some of us have developed an ib method for rigid bodies .
this method applies to a broad range of reynolds numbers . in the case of zero reynolds number
it becomes equivalent to the rigid multiblob method presented in section [ sec : rigidmultiblobs ] , but with a blob - blob mobility that is computed by the fluid solver . in ref .
@xcite only rigid bodies with specified motion ( kinematics ) were considered ; here we extend the method to handle freely - moving rigid bodies in stokes flow .
we will present here the key ideas and focus on the new components necessary to solve for the unknown motion of the particles ; we refer the reader interested in more technical details to refs .
the key novel contribution of our work is the preconditioner described in section [ sub : preconditionining - algorithm ] ; the performance and scalability of our preconditioned iterative solvers is studied numerically in section [ sub : convergenceibamr ] .
to begin , we present a semi - continuum formulation where the relation to section [ sec : rigidmultiblobs ] is most obvious , and then we discuss the fully discrete formulation used in the actual implementation . in appendix
[ app : permeable ] we demonstrate how to handle permeable bodies using a small modification of the formulation .
numerical results obtained using the method described here are given in section [ sec : resultsconfined ] .
we consider here a semi - discrete model in which the rigid body has already been discretized using blobs but a continuum description is used for the fluid , that is , we consider a rigid multiblob model immersed in a continuum stokesian fluid . in the ib literature blobs are referred to as markers , and are often thought of as `` points '' or `` discrete delta functions '' .
we use the term `` blob , '' however , to connect to section [ sec : rigidmultiblobs ] and to emphasize that the blobs have a finite physical and hydrodynamic extent . in the ib method @xcite ( and also the force coupling method @xcite ) , the shape of the blob and its effective interaction with the fluid
is captured through a smooth kernel function @xmath107 that integrates to unity and whose support is localized in a region of size comparable to the blob radius @xmath8 . in our rigid multiblob
ib method , to obtain the fluid - blob interaction forces @xmath108 that constrain the unknown rigid motion of the @xmath48 blobs , we need to solve a constrained stokes problem @xcite for the fluid velocity field @xmath109 , the fluid pressure field @xmath110 , the blob constraint forces @xmath108 , and the unknown rigid - body motions @xmath111 and @xmath112 , @xmath113 note that here the velocity and pressure fields contain both the `` background '' and the `` perturbational '' contributions to the flow . in the first equation in ( [ eq : semi_continuum ] ) ,
the kernel function is used to transfer ( spread ) the force exerted on the blob to the fluid , and in the third equation the same kernel is used to average the fluid velocity in the region covered by the blob and constrain it to follow the imposed rigid body motion plus additional slip or body deformation .
the handling of the spreading of constraint forces and averaging of the fluid velocity near physical boundaries is discussed in appendix d in @xcite .
we have implicitly assumed that appropriate boundary conditions are specified for the fluid velocity and pressure .
notably , we will apply the above formulation to cases where periodic or no - slip boundary conditions are applied along the boundaries of a cubic prism ( recall that periodic boundaries are not actual physical boundaries ) .
this includes , for example , a slit channel , a square channel , or a cubical container .
it is also relatively straightforward to handle stress - based boundary conditions such as free - slip or pressure valves @xcite .
it is not difficult to show that ( [ eq : semi_continuum ] ) is equivalent to the system ( [ eq : rigidsystem ] ) with the mobility matrix between two blobs @xmath55 and @xmath56 identified with @xcite @xmath114 where we recall that @xmath59 is the green s function for the stokes problem with unit viscosity and the specified boundary conditions .
this expression can directly be compared to ( [ eq : mobilityfaxen ] ) after realizing that for a smooth velocity field @xcite , @xmath115\v v\left(\v r\right)\big|_{\v r=\v r_{i}}=\left(\m i+\frac{a_{f}^{2}}{6}\grad^{2}\right)\v v\left(\v r\right)\big|_{\v r=\v r_{i}},\ ] ] where we assumed a spherical blob , @xmath116 .
we have defined here the `` faxen '' radius of the blob @xmath117 through the second moment of the kernel function . in multipole expansion based methods , the self - mobility of a body is treated separately by solving the single - body problem exactly (
this is only possible for simple particle shapes ) .
however , in the type of approach followed here the self - mobility @xmath118 is also given by the same formula ( [ eq : greensmobility ] ) with @xmath119 and does not need to be treated separately .
in fact , the self - mobility of a particle in an unbounded three - dimensional domain _ defines _ the effective hydrodynamic radius @xmath8 of a blob , @xmath120 where the oseen tensor @xmath121 is given in ( [ eq : oseentensor ] ) . in general , @xmath122 , but for a suitable choice of the kernel one can accomplish @xmath123 ( for example , for a gaussian @xmath124 @xcite ) and thus accurately obtain the faxen correction for a rigid sphere @xcite . for an isotropic or tensor product kernel @xmath125 and an unbounded domain , the pairwise blob - blob mobility ( [ eq : greensmobility ] ) will take the form @xmath126
where @xmath63 , and hat denotes a unit vector .
the functions of distance @xmath127 and @xmath128 depend on the specific kernel ( and in the fully discrete setting on the spatial discretization of the stokes equations ) and will be different from those appearing in the rpy tensor ( [ eq : rpytensor ] ) . nevertheless , as we will show numerically in section [ sub : transinv ] , the functions @xmath129 and @xmath130 for our ib method are quite close in form to those appearing in the rpy tensor .
we note that the rpy tensor itself can be seen as a realization of ( [ eq : greensmobility ] ) with the kernel being a surface delta function over a sphere of radius @xmath8 @xcite .
we have demonstrated above that solving ( [ eq : semi_continuum ] ) is a way to apply the blob - blob mobility for a confined domain . in the method of regularized stokeslets @xcite the mobility
is obtained _ analytically _ by averaging the analytical green s function with a kernel or envelope function specifically chosen to make the resulting integrals analytical .
note however that in that method the kernel @xmath125 appears only once inside the integral in ( [ eq : greensmobility ] ) because only the force spreading is regularized but not the interpolation of the velocity ; this leads to non - symmetric mobility matrix inconsistent with the faxen formula ( [ eq : mobilityfaxen ] ) .
by contrast , our approach is guaranteed to lead to a symmetric positive semidefinite ( spd ) mobility matrix @xmath68 , which is crucial when including thermal fluctuations @xcite . to obtain a fully discrete formulation of the linear system ( [ eq : semi_continuum ] ) we need to spatially discretize the stokes equations on a grid .
the spatial discretization of the fluid equation used in this work uses a uniform cartesian grid with grid spacing @xmath131 , and is based on a second - order accurate staggered - grid finite volume ( equivalently , finite difference ) discretization , in which vector - valued quantities such as velocity , are represented on the faces of the cartesian grid cells , while scalar - valued quantities such as pressure are represented at the centers of the grid cells @xcite .
the viscous terms are discretized using a standard @xmath132-point laplacian ( in three dimensions ) , accounting for boundary conditions using ghost cell extrapolation @xcite . in the fully discrete formulation of the fluid - body coupling , we replace spatial integrals in the semi - continuum formulation ( [ eq : semi_continuum ] ) by sums over fluid grid points .
the regularized delta function kernel is discretized using a tensor product of one - dimensional immersed boundary kernels @xmath133 of compact support , following peskin @xcite .
to maximize translational and rotational invariance ( i.e. , improve grid - invariance ) we use the smooth ( three - times differentiable ) six - point kernel recently described by bao _
this kernel is more expensive than the traditional four - point kernel @xcite because it increases the support of the kernel to @xmath134 grid points in three dimensions ; however , this cost is justified because the new six - point kernel improves the translational invariance by orders of magnitude compared to other standard ib kernel functions @xcite .
the interaction between the fluid and the rigid body is mediated through two crucial operations .
the discrete velocity - interpolation operator @xmath135 averages velocities on the staggered grid in the neighborhood of blob @xmath55 via @xmath136 where the sum is taken over faces @xmath137 of the grid , @xmath138 indexes coordinate directions ( @xmath139 ) as a superscript , and @xmath140 is the position of the center of the grid face @xmath137 in the direction @xmath138 .
the discrete force - spreading operator @xmath141 spreads forces from the blobs to the faces of the staggered grid via @xmath142 where now the sum is over the blobs and @xmath143 is the volume of a grid cell .
these operators are adjoint with respect to a suitably - defined inner product , and the discrete matrices satisfy @xmath144 , which ensures conservation of energy @xcite .
extensions of the basic interpolation and spreading operators to account for the presence of physical boundary conditions are described in appendix d in @xcite .
we note that it is possible to change the effective hydrodynamic and faxen radii of a blob by changing the kernel @xmath125 .
such flexibility in the kernel can be accomplished without compromising the required kernel properties postulated by peskin @xcite by using shifted or _ split kernels _
@xcite , @xmath145+\phi_{a}\left[q_{\alpha}-\left(r_{k}\right)_{\alpha}+\frac{s}{2}\right]\right\ } , \ ] ] where @xmath146 denotes a shift that parametrizes the kernel . by varying @xmath146 in a certain range , for example , @xmath147
, one can smoothly increase the support of the kernel and thus increase the hydrodynamic radius of the blob by as much as a factor of two .
we do not use split kernels in this work but have found them to work as well as the unshifted kernels , while allowing increased flexibility in varying the grid spacing relative to the hydrodynamic radius of the particles . following spatial discretization
, we obtain a finite - dimensional linear system of equations for the discrete velocities and pressures and the blob and body degrees of freedom . for the resistance problem
, we obtain the following rigidly constrained discrete stokes system @xcite , @xmath148\left[\begin{array}{c } \v v\\ \pi\\ \v{\lambda } \end{array}\right]=\left[\begin{array}{c } \v g=\v 0\\ \v h=\v 0\\ \v w=-\slip \end{array}\right],\label{eq : constrained_stokes}\ ] ] where @xmath149 is the discrete ( vector ) gradient operator , @xmath150 is the discrete ( vector ) divergence operator , and @xmath151 where @xmath152 is a discrete ( vector ) laplacian ; these finite - difference operators take into account the specified boundary conditions @xcite . for impermeable bodies @xmath153 , which makes the linear system ( [ eq : constrained_stokes ] ) a nested saddle - point problem in both lagrange multipliers @xmath154 and @xmath37 .
as explained in appendix [ app : permeable ] , for permeable bodies @xmath155 is a diagonal matrix with @xmath156 for blob @xmath23 , where @xmath157 is the permeability of body @xmath12 and @xmath158 is a volume associated with blob @xmath55 .
the right - hand side could include any external fluid forcing terms , slip , inhomogeneous boundary conditions , etc .
the system ( [ eq : constrained_stokes ] ) can be made symmetric by excluding the volume weighting @xmath159 in the spreading operator ( [ eq : s_unweighted ] ) ; this makes @xmath37 have units of force density rather than total force .
this nested saddle - point structure continues if one considers impermeable rigid bodies that are free to move , leading to the _ _ discrete mobility problem _
_ @xmath160\left[\begin{array}{c } \v v\\ \pi\\ \v{\lambda}\\ \v u \end{array}\right]=\left[\begin{array}{c } \v g\\ \v h=\v 0\\ \v w=-\slip\\ \v z=\v f \end{array}\right].\label{eq : free_kinematics_stokes}\ ] ] after eliminating the velocity and pressure from this system , we obtain the saddle - point system ( [ eq : saddle_m ] ) with the identification of the mobility with its discrete approximation @xmath161 which is spd . here
@xmath162 is a discrete stokes solution operator , @xmath163 where we have assumed for now that @xmath164 is invertible ; see @xcite for the handling of periodic systems , for which the laplacian is not invertible . unlike for green s function based methods , we never explicitly compute or form @xmath162 or @xmath68 ; rather , we solve the stokes velocity - pressure subsystems iteratively using the preconditioners described in @xcite . in this section
we describe how to solve the system ( [ eq : free_kinematics_stokes ] ) using an iterative solver , as we have implemented in the immersed boundary adaptive mesh refinement software framework ( ibamr ) @xcite .
our codes are integrated into the public release of the ibamr library .
note that the matrix - vector product is a straightforward and inexpensive application of finite - difference stencils on the fluid grid and summations over blobs .
the key to an effective solver is the design of a good preconditioner , i.e. , a good approximate solver for ( [ eq : free_kinematics_stokes ] ) .
the basic idea is to combine a preconditioner for the stokes problem @xcite with the indefinite preconditioner ( [ eq : indef_block_p ] ) with a block - diagonal approximation of the mobility @xmath96 constructed based on empirical fits of the blob - blob mobility , as we know explain in detail .
a preconditioner for solving the resistance problem ( [ eq : constrained_stokes ] ) was developed by some of us in @xcite ; readers interested in additional details should refer to this work .
the preconditioner is based on approximating the blob - blob mobility with the functional form ( [ eq : m_tilde_ij ] ) , where the functions @xmath127 and @xmath128 are obtained by fitting numerical data for the blob - blob mobility in an _ unbounded _ system ( in practice , a large periodic system ) .
this involves two important approximations , the validity of which only affects the _ efficiency _ of the linear solver but does _ not _ affect the _ accuracy _ of the method since the krylov method will correct for the approximations .
the first approximation comes from the fact that the true blob - blob mobility for the immersed boundary method is not perfectly translationally and rotationally invariant , so that the form ( [ eq : m_tilde_ij ] ) does not hold exactly .
the second approximation is that the boundary conditions are not correctly taken into account when constructing the approximation of the mobility @xmath96 .
this approximation is crucial to the feasibility of our method and is much more severe , but , as we will demonstrate numerically in section [ sec : resultsconfined ] , the krylov solver converges in a reasonable number of iterations , correctly incorporating the boundary conditions in the solution .
the empirical fits of @xmath127 and @xmath128 are described in appendix a of @xcite , and code to evaluate the empirical fits is publicly available for a number of kernels constructed by peskin and coworkers ( three- , four- , and six - point ) at http://cims.nyu.edu/~donev/src/mobilityfunctions.c . as we show in section [ sub : transinv ] , these functions are quite similar to those appearing in the rpy tensor ( [ eq : rpytensor ] ) , and , in fact , it is possible to use the rpy functions @xmath165 and @xmath166 in the preconditioner , with a value of the effective hydrodynamic radius @xmath8 that depends on the choice of the kernel . nevertheless , somewhat better performance is achieved by using the empirical fits for @xmath127 and @xmath128 developed in @xcite . in @xcite , we considered general fluid - structure interaction problems over a range of reynolds numbers , and constructed @xmath96 as a dense matrix of size @xmath51 , which was then factorized using dense linear algebra .
this is infeasible for suspensions of many rigid bodies . in this work ,
we use the block - diagonal approximation ( [ eq : m_tilde_block_diag ] ) to the blob - blob mobility matrices , in which there is one block per rigid particle . once @xmath96 is constructed and its diagonal blocks factorized , the corresponding approximate body mobility matrix @xmath167 is easy to form , as discussed in more detail in section [ sub : preconditioner ] .
note that these matrices and their factorizations need to be constructed only once at the beginning of the simulation , and can be reused throughout the simulation . a key component of solving the constrained stokes problems ( [ eq : constrained_stokes ] ) or ( [ eq : free_kinematics_stokes ] ) is an iterative solver for the unconstrained discrete stokes sub - problem , @xmath168\left[\begin{array}{c } \v v\\ \pi \end{array}\right]=\left[\begin{array}{c } \v g\\ \v h \end{array}\right],\ ] ] for which a number of techniques have been developed in the finite - element context @xcite .
to solve this system , we can use gmres with a preconditioner @xmath169 that assumes periodic boundary conditions so that the various finite - difference operators commute @xcite .
specifically , the preconditioner for the stokes system that we use in this work is based on a projection preconditioner developed by griffith @xcite , @xmath170 where @xmath171 is the dimensionless pressure ( scalar ) laplacian , and @xmath172 and @xmath173 denote approximate solvers obtained by a _ single _
v - cycle of a geometric multigrid method , as performed using the _ hypre _ library @xcite in our ibamr implementation . in this paper
we will primarily report the options we have found to be best without listing all of the different combinations we have tried . for completeness , we note that we have tried the better - known lower and upper triangular preconditioners @xcite for the stokes problem . while these simpler preconditioners are better when solving pure stokes problems than the projection preconditioner ( [ p_stokes ] ) since they avoid the pressure multigrid application @xmath174 , we have found them to perform much worse in the context of suspensions of rigid bodies .
a possible explanation is that the projection preconditioner @xmath169 is the only one that is exact for periodic systems if exact subsolvers for the velocity and pressure subproblems are used .
observe that one application of @xmath169 is relatively inexpensive and involves only @xmath175 scalar multigrid v - cycles .
the number of iterations required for convergence depends strongly on the boundary conditions ; fast convergence is obtained within 10 - 20 iterations for periodic systems , but as many as a hundred gmres iterations may be required for highly confined systems @xcite .
we emphasize that the performance of this preconditioner is highly dependent on the details of the staggered geometric multigrid method , which is not highly optimized in the _ hypre _ library , especially for domains of high aspect ratios such as narrow slit channels .
for periodic boundary conditions , one can use ffts to solve the stokes problem , and this is likely to be more efficient than geometric multigrid especially because ffts have been highly optimized for common hardware architectures .
however , such an approach would require 3 scalar ffts for _ each _ iteration of the iterative solver for the constrained stokes problem ( [ eq : constrained_stokes ] ) or ( [ eq : free_kinematics_stokes ] ) , and this will in general be substantially more expensive than using only a few cycles of geometric multigrid as an _ approximate _ stokes solver . the use of an approximate stokes solver instead of an exact one is an important difference between implementing the rigid multiblob method for periodic systems using the spectral ewald method @xcite and our approach .
the product of the blob - blob mobility with a vector can be computed more accurately and faster using the spectral ewald method , in particular because one can adjust the cutoff for splitting the computation between real and fourier space arbitrarily , unlike in our method where the grid spacing is tied to the particle radius . however , for rigid multiblobs , one must solve the system ( [ eq : saddle_m ] ) , which requires potentially many matrix - vector products , i.e. , many ffts in the spectral ewald approach .
by contrast , in our method we solve the extended problem ( [ eq : free_kinematics_stokes ] ) , and only solve the stokes problems approximately using a few cycles of multigrid in each iteration .
this will require more iterations but each iteration can be substantially cheaper than performing three ffts each krylov iteration . for non - periodic systems , there is no equivalent of the spectral ewald method , but see @xcite for some steps in this direction .
our method computes the hydrodynamic interactions in a confined geometry `` on the fly '' without ever actually computing the action of the green s function exactly , rather , it is computed only approximately and the outer krylov solver corrects for any approximations made in the preconditioner .
we now have the necessary ingredients to compose a preconditioner for solving ( [ eq : free_kinematics_stokes ] ) , i.e. , to construct an approximate solver for this linear system .
each application of our preconditioner involves the following steps : 1 .
approximately solve the fluid sub - problem , @xmath168\left[\begin{array}{c } \tilde{\v v}\\ \tilde{\pi } \end{array}\right]=\left[\begin{array}{c } \v g\\ \v h \end{array}\right],\ ] ] using @xmath176 iterations of an iterative method with the preconditioner ( [ p_stokes ] ) .
2 . interpolate @xmath177 to get the relative slip at each of the blobs , @xmath178 , and rotate the corresponding component from the current frame to the reference frame of each body .
3 . approximately compute the unknown body kinematics @xmath26 : 1 . calculate @xmath179 and rotate the result back to the fixed frame of reference . here
@xmath96 is a block - diagonal approximation to the blob - blob mobility matrix in the reference frame , as described in section [ sub : approximateblobmob ] ; the factorization of the blocks of @xmath96 is performed once at the beginning of the simulation .
2 . calculate @xmath180 and transform ( rotate ) @xmath181 to the body frame of reference .
3 . compute @xmath182 and transform it back to the fixed frame of reference , where @xmath183 .
4 . calculate the updated relative slip velocity at each of the blobs , @xmath184 and transform ( rotate ) it to reference body frame .
5 . compute @xmath185 and transform @xmath37 back to the fixed frame of reference if necessary .
solve the corrected fluid subproblem to obtain the fluid velocity and pressure : @xmath168\left[\begin{array}{c } \v v\\ \pi \end{array}\right]=\left[\begin{array}{c } \v g+\m{\mathcal{s}}\v{\lambda}\\ \v h \end{array}\right],\ ] ] using @xmath186 iterations of an iterative method with the preconditioner ( [ p_stokes ] ) .
a few comments are in order .
the above preconditioner is not spd so the outer krylov solver should be a method such as gmres of bicgstab @xcite .
we prefer to use right - preconditioned krylov solvers because in this case the residual computed by the iterative solver is the true residual ( as opposed to the preconditioned residual for left preconditioning ) , and therefore termination criteria ensure that the original system was solved to the desired target tolerance .
we expect that the long - term recurrence gmres method will require a smaller number of iterations than the short - term recurrence used in bicgstab ( but note that each iteration of bicgstab requires _ two _ applications of the preconditioner ) .
however , observe that gmres can require substantially more memory since it requires storing a complete history of the iterative process .
floating - point numbers per grid cell , which can make the memory requirements of a gmres - based solver with a large restart frequency quite high for large grid sizes . ] this can be ameliorated by restarts at a cost of slowed convergence .
if the iterative solver used for the stokes solver in steps 1 and 6 is a nonlinear method ( most krylov methods are nonlinear ) , then the outer solver must be a flexible method such as fgmres .
this flexibility typically increases the memory requirements of the iterative method ( for example , it exactly doubles the number of stored intermediate vectors for fgmres versus gmres ) , and so an alternative is to use a linear method such as richardson s method . note that when a preconditioned krylov method is used for the stokes subsolver , one additional application of the preconditioner is required to convert the system to preconditioned form for both left and right preconditioning , making the total number of applications of the stokes preconditioner ( [ p_stokes ] ) be @xmath187 per krylov iteration .
by contrast , if richardson s method is used in the stokes subsolver , the number of preconditioner applications is @xmath188 . since in many practical cases
the cost is dominated by the multigrid cycles , this difference can be important in the overall performance of the preconditioner .
we will explore the performance of the preconditioner and the effect of the various choices in detail in section [ sub : convergenceibamr ] .
in this section we investigate the accuracy of rigid multiblob models of spheres as a function of the number of blobs .
we focus on spheres in an unbounded domain because of the availability of analytical results to compare to , and not because the rigid multiblob method is particularly good for suspensions of spheres , for which there already exist a number of well - developed multipole expansion approaches .
we also investigate the performance of the preconditioner developed in sec .
[ sub : preconditioner ] for solving ( [ eq : saddle_m ] ) , for suspensions of spheres in an unbounded domain ( e.g. , clusters of colloids formed in a gel ) . for unbounded domains ,
we compute the product of the blob - blob mobility matrix @xmath47 with a vector using the fast multipole method ( fmm ) developed specifically for the rpy tensor in @xcite ; this software makes four calls to the poisson fmm implemented in the fmmlib3d library ( http://www.cims.nyu.edu/cmcl/fmm3dlib/fmm3dlib.html ) per matrix - vector product . as we will demonstrate empirically , the asymptotic cost of the rigid - multiblob method scales as @xmath189 , where @xmath48 is the total number of blobs , with a coefficient that grows only weakly with density .
we note that in this paper we use relatively tight tolerances ( @xmath190 ) when computing the matrix - vector products solving the linear systems in order to test the robustness of the preconditioners ; in practical applications much lower tolerances ( @xmath191 ) would typically be employed , potentially lowering the overall computational effort considerably over what is reported here . in this work , each sphere
is discretized with @xmath192 blobs of hydrodynamic radius @xmath8 distributed on the surface of a sphere of _ geometric _ radius @xmath193 .
we discretize the surface of a sphere as a shell of blobs constructed by a recursive procedure suggested to us by charles peskin ( private communication ) ; the same procedure is used in @xcite .
we start with 12 blobs placed at the vertices of an icosahedron @xcite , which gives a uniform triangulation of a sphere by 20 triangular faces .
then , we place a new blob at the center of each edge and recursively subdivide each triangle into four smaller triangles , projecting the vertices back to the surface of the sphere along the way .
each subdivision approximately quadruples the number of vertices , with the @xmath137-th subdivision producing a model with @xmath194 blobs , leading to shells with 12 , 42 , 162 or 642 blobs , see fig . 2 in @xcite for an illustration .
in this section we study the optimal choice of @xmath8 for a given resolution ( number of blobs ) and @xmath193 .
an important concept that will be used heavily in the rest of this paper is that of an _ effective hydrodynamic radius @xmath195 _ of a blob model of a sphere ( more generally , effective hydrodynamic extent ) .
if we approach the rigid multiblob method from a boundary integral perspective , we would assign @xmath196 as the radius and treat the additional enlargement of the effective hydrodynamic radius as a numerical ( quadrature+regularization ) error .
this is more or less how results are presented in the recent work of swan and wang @xcite ( see for example their fig . 8) , making the accuracy appear low even in the far field for small number of blobs per sphere .
however , we instead think of a rigid multiblob as an effective _ model _ of a sphere , whose hydrodynamic response mimics that of an equivalent sphere
. a similar effect appears in lattice boltzmann simulations , with @xmath8 being related to the lattice spacing @xcite . to appreciate why it is imperative to use an effective radius , observe that even a single blob acts as an approximation of a sphere with radius @xmath197 .
similarly , one should not treat a line of blobs ( see left - most panel in fig .
[ fig : blobmodels ] ) as a zero - thickness object ( line ) ; rather , such a line of rigidly - connected blobs should be considered to model a rigid cylinder with finite thickness proportional to @xmath8 @xcite .
we compute the effective hydrodynamic radius of our blob models of spheres next .
[ [ subconvergencemomentseffective - hydrodynamic - radii - of - rigid - multiblob - spheres ] ] [ sub : convergencemoments]effective hydrodynamic radii of rigid multiblob spheres ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ in this section we consider an isolated rigid multiblob sphere in an unbounded domain , and compute its response to an applied force @xmath198 , an applied torque @xmath199 , and an applied linear shear flow with strain rate @xmath200 .
each of these defines an effective hydrodynamic radius by comparing to the analytical results for a sphere , therefore , each model of a sphere will have three distinct hydrodynamic radii .
the _ translational radius _ is measured from ( see also @xcite ) @xmath201 where @xmath202 is the resulting sphere linear velocity , the _ rotational radius _ is ( see also @xcite ) @xmath203 where @xmath204 is the resulting angular velocity , and the effective stresslet radius is @xmath205 here we compute the stresslet @xmath206 induced on the rigid multiblob under an applied shear by setting an apparent slip @xmath207 on blob @xmath55 , and then solving the mobility problem to compute the constraint ( rigidity ) forces @xmath37 . the stresslet @xmath206 is the symmetric traceless component of the * * first moment of the constraint forces @xmath208 in this work , we use @xmath209 as the effective hydrodynamic radius when comparing to theory .
this is because the translational mobility is controlled by the most long - ranged @xmath210 hydrodynamic interactions , and therefore the far - field response of a rigid multiblob is controlled by @xmath209 .
observe that since we only account for translation of the blobs , only @xmath209 is nonzero for a single blob , while @xmath211 and @xmath212 are zero .
therefore , the minimal model of a sphere that allows for nontrivial rotlet and stresslets is the icosahedral model ( 12 blobs ) @xcite .
since the rigid multiblob models are able to exert a stress on the fluid they can change the viscosity of a suspension @xcite , unlike the single - blob models , which do not resist shear . it is important to note that the rigid multiblob models of a sphere are _ not _ perfectly rotationally invariant , especially for low resolutions .
therefore , the rigid multiblobs may exhibit a small translational velocity even in the absence of an applied force , or they may exhibit a small rotation even in the absence of an applied torque .
in other words , the effective mobility matrix for a rigid multiblob model of a sphere can exhibit small off - diagonal components .
similarly , there will in general be small but nonzero components of the stresslet that would be identically zero for a perfect sphere .
in general , we find these spurious components to be very small even for the minimally resolved icosahedral rigid multiblob . a key parameter that we need to choose is how to relate the blob hydrodynamic radius @xmath8 with the typical spacing between the blobs .
since our multiblob models of spheres are regular the minimal spacing between markers @xmath146 is well - defined , and we expect that there will be some optimal ratio @xmath213 that will make the rigid multiblob represent a true rigid sphere as best as possible . in a number of prior
works the intuitive choice @xmath214 has been used , since this corresponds to the idea that the blobs act as a sphere of radius @xmath8 and we would like them to touch the other blobs .
however , as we explained above , it is not appropriate to think of blobs as spheres with a well - defined surface , and it is therefore important to study the optimal spacing more carefully .
.[tab : stresslets_3d](left ) effective translational , rotational and stresslet hydrodynamic radii for rigid multiblob models of a sphere , for two choices of the blob - blob relative spacing @xmath213 .
( right ) iterations to solve the mobility problem with tolerance @xmath215 for 4096 spheres discretized with 12 blobs , or for 512 spheres discretized with 42 blobs , arranged in a simple cubic lattice at different volume fractions @xmath216 , see sec .
[ sub : convergencefmm ] . [ cols="^,^,^,^,^,^,^ " , ] a ortega , d amors , and j garca de la torre .
prediction of hydrodynamic and other solution properties of rigid proteins from atomic - and residue - level models .
, 101(4):892898 , 2011 .
code available at http://leonardo.inf.um.es/macromol/programs/hydropro/hydropro.htm .
megan s davies wykes , jrmie palacci , takuji adachi , leif ristroph , xiao zhong , michael d ward , jun zhang , and michael j shelley .
dynamic self - assembly of microscale rotors and swimmers .
, 12(20):45844589 , 2016 .
tri t pham , ulf d schiller , j ravi prakash , and burkhard dnweg .
implicit and explicit solvent models for the simulation of a single polymer chain in solution : lattice boltzmann versus brownian dynamics .
, 131:164114 , 2009 .
simn poblete , adam wysocki , gerhard gompper , and roland g. winkler .
hydrodynamics of discrete - particle models of spherical colloids : a multiparticle collision dynamics simulation study .
, 90:033314 , 2014 .
david j smith .
a boundary element regularized stokeslet method applied to cilia - and flagella - driven flow .
in _ proceedings of the royal society of london a : mathematical , physical and engineering sciences _ , volume 465 , pages 36053626 . the royal society , 2009 .
simon k layton and lorena a barba .
inexact krylov iterations and relaxation strategies with fast - multipole boundary element method . , 2015 .
software available at https://github.com/barbagroup/fmm-bem-relaxed .
satish balay , william d. gropp , lois curfman mcinnes , and barry f. smith .
efficient management of parallelism in object oriented numerical software libraries . in e.
arge , a. m. bruaset , and h. p. langtangen , editors , _ modern software tools in scientific computing _ , pages 163202 . birkhuser press , 1997 .
software available at http://www.mcs.anl.gov/petsc .
yuanxun bao , jason kaye , and charles s. peskin .
a gaussian - like immersed - boundary kernel with three continuous derivatives and improved translational invariance .
, 316:139 144 , 2016 .
software available at https://github.com/stochastichydrotools/ibmethod .
r. falgout , j. jones , and u. yang .
the design and implementation of hypre , a library of parallel high performance preconditioners .
, pages 267294 , 2006 .
software available at http://www.llnl.gov/casc/hypre .
ayan chakrabarty , feng wang , chun - zhen fan , kai sun , and qi - huo wei .
high - precision tracking of brownian boomerang colloidal particles confined in quasi two dimensions .
, 29(47):1439614402 , 2013 .
pmid : 24171648 .
yu zhang , juan j de pablo , and michael d graham .
an immersed boundary method for brownian dynamics simulation of polymers in complex geometries : application to dna flowing through a nanoslit with embedded nanopits .
, 136:014901 , 2012 .
pat plunkett , jonathan hu , christopher siefert , and paul j atzberger .
spatially adaptive stochastic methods for fluid structure interactions subject to thermal fluctuations in domains with complex geometries .
, 277:121137 , 2014 . | we develop a rigid multiblob method for numerically solving the mobility problem for suspensions of passive and active rigid particles of complex shape in stokes flow in unconfined , partially confined , and fully confined geometries . as in a number of existing methods , we discretize rigid bodies using a collection of minimally - resolved spherical blobs constrained to move as a rigid body , to arrive at a potentially large linear system of equations for the unknown lagrange multipliers and rigid - body motions . here
we develop a block - diagonal preconditioner for this linear system and show that a standard krylov solver converges in a modest number of iterations that is essentially independent of the number of particles .
key to the efficiency of the method is a technique for fast computation of the product of the blob - blob mobility matrix and a vector . for unbounded suspensions , we rely on existing analytical expressions for the rotne - prager - yamakawa tensor combined with a fast multipole method ( fmm ) to obtain linear scaling in the number of particles . for suspensions sedimented against a single no - slip boundary ,
we use a direct summation on a graphical processing unit ( gpu ) , which gives quadratic asymptotic scaling with the number of particles . for fully confined domains , such as periodic suspensions or suspensions confined in slit and square channels ,
we extend a recently - developed rigid - body immersed boundary method [ `` an immersed boundary method for rigid bodies '' , b. kallemov , a. pal singh bhalla , b. e. griffith , and a. donev , communications in applied mathematics and computational science , 11 - 1 , 79 - 141 , 2016 ] to suspensions of freely - moving passive or active rigid particles at zero reynolds number .
we demonstrate that the iterative solver for the coupled fluid and rigid body equations converges in a bounded number of iterations regardless of the system size . in our approach
, each iteration only requires a few cycles of a geometric multigrid solver for the poisson equation , and an application of the block - diagonal preconditioner , leading to linear scaling with the number of particles .
we optimize a number of parameters in the iterative solvers and apply our method to a variety of benchmark problems to carefully assess the accuracy of the rigid multiblob approach as a function of the resolution .
we also model the dynamics of colloidal particles studied in recent experiments , such as passive boomerangs in a slit channel , as well as a pair of non - brownian active nanorods sedimented against a wall .
# 1 # 1 # 1 # 1#1 # 1#1 # 1#1 # 1|#1| # 1#1 # 1
s l |
the first beam test results for a new time - of - flight ( tof ) gaseous detector dielectric resistive plate chamber ( drpc ) @xcite were obtained in the slopes of r&d program for tof gaseous detectors intended to be used in the proposed nuclear experiment alice at lhc .
drpc was developed as a frontier detector between ppc and rpc .
the basic idea is to avoid large energy resolution in spark cases ( rpc feature ) , keeping high time resolution altogether with fast response provided by ppc @xcite .
a dielectric layer introduced between the electrodes is covered with a surface resistivity on the gas side which lets , in case of a breakdown , to decrease or limit the spark energy by charging an elementary plate condensor , which then discharges through the resistive surface .
the detector has shown stability of operation in the avalanche mode within a wide range of applied voltages .
time response and energy resolution in spark events may be adjusted by varying the main electric parameters of the detector : the dielectric layer capacity and the surface resistivity .
the basic detector construction used in the tests is shown in fig .
[ scheme ] .
a plate condensor has external dimensions @xmath4
mm@xmath3 , corresponding to the required dimensions of a single detecting cell in the alice tof system , and is several milimeters thick .
one of the electrodes is made of a conducting material , the other has a resistance - dielectric - metal structure .
the gas gap has been chosen to be 0.6 mm wide basing on the ppc experience as this value provides high mip registration efficiency altogether with a rather good time resolution . the gap is formed by means of ceramic spacers .
accuracy requirements are similar to those for the traditional ppc : the electrodes and spacers are kept flat and parallel within the several micrometers of accuracy .
the layer electrode is the crucial part of the detector . the dielectric thickness @xmath5 is determined from the condition @xmath6 , where @xmath7 is the gas gap width , @xmath8 and @xmath9 are the dielectric and gas dielectric constants correspondently .
we used 1 mm thick al@xmath10o@xmath11 with @xmath12 .
the dielectric was covered with a thin layer of resistant material sic or tic with the volume resistivity @xmath13 @xmath14 .
the outer side of this electrode was metallized through evaporation , and actually appeared to be the second condensor electrode producing the electric field responsible for the gas amplification . in the tests
double - gap detectors formed of two identical single gaps were investigated .
[ options ] shows other possible options of the detector assembly .
drpc has been studied as a detector for high precision and efficient time measurements on ps beams at itep and cern .
@xmath15-beams with 27 gev/@xmath16 momenta , 10@xmath310@xmath17 s@xmath18 intensities , and several @xmath1 cross - section at the detector position were usually used . trigger and start counters , made of scintillators of different sizes and modifications , and adapted for time measurements , were positioned on the same beam line .
the trigger consisted of 6 to 10 counters , including those situated on precise movable platforms , which determined the space position of the beam ahead and behind the target , important for efficiency measurements , as well as the crossing point between the beam and the detector plane .
the time measurement scheme included 2 start scintillation counters ( each with 2 pmts ) , it was always possible to measure their time resolution continiously during the data taking .
a typical resolution of the start system equaled to 150 ps .
the signal from the detector was used as the _ stop _ in the scheme .
the resolution of start counters was quadratically subtracted from the time of particles flight along a fixed base . in this way the detector resolution was measured . all trigger and timing electronics were assembled in the nim and camac standards .
specially designed fast and low - noise electronics , consisting of a preamplifier , a main amplifier and a discriminator , were used for registration of signals from the gaseous detector .
the scheme comes as a development of ideas previously described in @xcite and lies beyond the slopes of this paper .
various gas mixtures , based on either dme or different types of freons with normal quality , were used as working gases in the detector . a simple gas system allowed to mix gases at rather small fluxes of about several litres per hour .
no special control of quality and other gas parameters was undertaken .
the mixture dme + 5%cf@xmath11br was most thoroughly studied as a working gas for drpc .
the best results were obtained with c@xmath10h@xmath10f@xmath19 + 15%dme and c@xmath10h@xmath10f@xmath19 + 5%isobutane + 10%sf@xmath20 .
[ sevent ] represents a typical result of the drpc efficiency and time resolution measurements .
[ cols="^,^ " , ] all streamer signals have same shapes with fast rise times and large amplitudes .
such working regime seems to be very attractive and must be additionally investigated .
the cost of a large tof system built of many drpc cells may be estimated supposed that the single channel price is about $ 4050 , including the cost of development of new front - end and readout electronics , prototype construction , etc .
this price is more than an order of magnitude less , than that which might be required for tof based on scintillation counter technique . | principles of operation , construction and first test results of a dielectric resistive plate chamber ( drpc ) are described .
the detector has shown stability of operation in the avalanche mode of gas amplfication within a wide range of applied voltages .
double - gap drpcs have demonstrated the mip registration efficiency of 97% and the time resolution of 180200 ps .
no changes in drpc operation have been observed with test beam intensities up to 10@xmath0 hz/@xmath1 . _
@xmath2institute for theoretical and experimental physics ( itep ) , + b. cheremushkinskaya 25 , moscow , 117218 , russia .
+ @xmath3center of perspective technologies and apparatus ( cpta ) , + preobrazhenskaya pl . 6/8 , moscow , 107076 , russia . |
stars are surprisingly stable for most of their evolution . the exceptions provide windows into important transitions in stellar properties . the helium flash , for example , marks the temporal transition from shell hydrogen burning to core helium burning .
interiors theory also predicts changes in structure as a function of mass .
a classic example is the prediction that progressively lower mass models should have progressively deeper surface convection zones , culminating in a transition to a fully convective state at around a third of a solar mass .
the global behavior of models near this transition is traditionally thought to be smooth .
we have uncovered a novel instability for models just above the fully convective boundary that has , to the best of our knowledge , never before been noted for these objects .
non - equilibrium @xmath1 burning in a marginally stable and small radiative core leads to a steady increase in @xmath1 and energy generation , and the development of a small convective core separated from the convective envelope by a radiative buffer zone . above a critical threshold
a transition to a fully convective state occurs .
convective mixing lowers the core @xmath1 abundance , quenching the nuclear reactions and causing the core to become radiative again .
the sudden drop in energy generation leads to a sudden drop in radius , followed by a gradual recovery as @xmath1 production in the core resumes .
this phenomenon repeats in a series of pulses , with an amplitude which is steadily damped as the envelope @xmath1 is enriched and the drop in core @xmath1 becomes smaller . for non - interacting stars this will apply to a narrow range of masses just above the fully convective boundary .
the secondary stars of cataclysmic variables , however , will all pass through this mass range , and the onset of the instability occurs at the upper end of the cv period gap .
the astrophysical consequences may therefore be surprisingly broad , and the instability can be triggered for mass - loss rates of interest in the cv context . in this paper
we describe the physics of the instability and present some preliminary assessment of its effects .
we describe our stellar models in section [ sec : models ] , discuss the details of the physical processes responsible for the instability in section [ sec : physics ] , observational implications of the instability in section [ sec : observations ] , and close with our conclusions in section [ sec : conclusion ] .
we utilize the yale rotating stellar evolution code ( yrec ) @xcite to produce a grid of stellar models with a standard set of input physics .
we create a dense grid near the fully convective boundary to investigate the mass range over which this instability operates .
models range in mass from @xmath2 and are spaced every @xmath3 .
models generally contain @xmath4 shells and take @xmath5 timesteps reach to 14 gyr , at which point the evolution is truncated .
all models include helium and heavy element diffusion following the procedure of @xcite , atmosphere and boundary conditions of @xcite , nuclear reaction rates of @xcite with weak screening @xcite , and employ a mixing length theory of convection @xcite .
opacities are from the opacity project ( op ) @xcite for a @xcite solar mixture , supplemented with the low temperature opacities of @xcite .
we utilize the 2006 opal equation of state @xcite and the @xcite eos for temperature and density combinations outside of the opal tables .
convective instability is determined using the schwarzschild criterion .
we impose an initial @xmath1 mass fraction of @xmath6 , for a @xcite mixture with the @xmath7 isotopic ratio from @xcite on a series of pre - main sequence starting models . because @xmath8 chain burning is by far the most significant source of @xmath1 , the exact value of this initial abundance is of relatively little importance .
we find that low mass objects at the fully convective boundary undergo non - equilibrium @xmath1 burning during the early stages of their main - sequence lifetimes which gives rise to periodic fully convective episodes .
we discuss the nature of the instability , the conditions under which it operates , and compare our results obtained with yrec to those obtained with the mesa stellar evolution code .
@xmath1 is produced in low mass stars via proton capture on deuterium ; the former can be either primordial or produced from the @xmath8 chain .
the @xmath8 reaction itself has a low cross section but a weak temperature dependence , so @xmath1 can be generated even in relatively cool environments given sufficient time .
@xmath1 is destroyed by both @xmath1 + @xmath1 ( @xmath8 i ) and @xmath1 + @xmath9 reactions ( @xmath8 ii and @xmath8 iii ) ; the former is far more important at low temperatures . at high temperatures equilibrium
is achieved with a @xmath1 abundance that decreases as the temperature increases , and net destruction of @xmath1 can occur . however , at low temperatures the rate of production exceeds the rate of destruction and @xmath1 is produced ( @xcite ; @xcite ; see also @xcite for a discussion ) .
one therefore expects a @xmath1 abundance peak in the outer layers of higher mass stars ( see figure 1 in * ? ? ?
* for an example ) . for lower mass stars the peak shifts towards the center of the star and the development of a deep convective envelope homogenizes the outer layers . for sufficiently cool central temperatures
@xmath1 is never in equilibrium .
therefore the nuclear energy generation from @xmath1 + @xmath1 burning will depend sensitively on the integrated @xmath1 production over the lifetime of the star . if there is a radiative core it is possible to have a local enhancement in the @xmath1 abundance which is much larger than the average for a fully convective star .
this special circumstance can therefore create a small convective core , even in stars far too cool to support the cno burning responsible for convective cores on the upper main sequence .
the growth in central @xmath1 will cause this core to grow in mass . for a sufficiently small radiative core
it is therefore possible to induce a `` convective kissing '' instability where the central and surface convective regions merge .
to understand the nature of the this instability , we introduce figure [ fig1 ] , which contains plots of the location of the base of the envelope convection zone , core @xmath1 fraction ( @xmath10 ) , luminosity , and radius of several stellar models of different masses . fully convective episodes ( when @xmath11 ) are in phase with sudden drops in @xmath10 in the core . after an episode
the fraction of @xmath1 in the core slowly recovers , along with the luminosity from @xmath1 burning reactions , whose rates are inflated by the presence of the elevated @xmath1 abundances . the total luminosity and radius of the star both increase on nuclear timescales in response to the additional burning .
the response of the core and envelope convection zones is shown in figure [ fig2 ] , which plots the temperature gradients of a single @xmath12 model at different stages in the convective cycles . at times
early in the cycle ( ( a ) , bottom panel of figure [ fig2 ] ) , the core and envelope convection zones are separated significantly in radius . as time passes and the mean @xmath1 of the entire star increases with each mixing episode , the increased @xmath1 burning inflates the radiative temperature gradient , resulting in a decrease in the separation between the envelope and core convection zones , as seen at times ( b ) and ( c ) . by time ( d )
the additional luminosity is sufficient to make the convection zones meet , at which point the star is fully convective , and mixes the @xmath1 in the core throughout the entire model on a convective overturn timescale . in the case of a @xmath13 model , the luminosity from the reactions @xmath14 and @xmath15 in the convective core represents between @xmath16 and @xmath17 of the total model luminosity , depending on the location within the cycle .
the dilution of the @xmath1 in the core results in decreased luminosity and thermal support , and the model radius and luminosity decrease on a thermal timescale .
the sudden loss of substantial @xmath1 burning is somewhat counter - balanced by energy release due to gravitational collapse , which results in modest , few percent changes in the _ total _ luminosity over the course of these episodes .
the structure returns to a convective core , radiative buffer zone , and convective envelope , while the @xmath1 again begins to accumulate in the core .
the middle panel of figure [ fig2 ] shows the temperature gradient at the timestep immediately following the fully convective episodes for several cycles .
each successive cycle begins with higher @xmath1 abundances throughout the model , since the products of previous cycles are mixed throughout the star . because the total @xmath1 abundance is higher at the onset than in previous cycles , the boundaries of the convective core and envelope begin the cycle more closely spaced in radius , and the duration of the cycles decreases .
cycles continue until the @xmath1 abundance throughout the star is high enough that the model is fully convective for the rest of its ms lifetime .
figure [ fig1 ] demonstrates the response of the convection zone , central @xmath1 abundance , radius and luminosity for models of different masses over the course of many convective cycles .
more massive models tend to have cycles of longer duration and greater amplitude , because they are further from the fully convective boundary , and thus require more @xmath1 burning to cause them to be convectively unstable . likewise , lower mass models have shorter cycles with smaller amplitude variations in core @xmath1 fraction , luminosity , and radius . in our models
( at solar composition @xmath18 , with a solar calibrated mixing length parameter @xmath19 ) , the instability occurs over the mass range @xmath20 . for masses @xmath21
the star undergoes a brief period in which it has a radiative zone between a convective core and convective envelope without any episodic mixing , but is then fully convective on the remainder of the ms .
models with masses @xmath22 are fully convective ( in that they _ never _ have a radiative core ) throughout their ms lifetimes .
this fully convective boundary is somewhat lower than the typical values quoted in the literature primarily due to our definition : we consider the fully convective boundary to be the division between models that host a radiative core at some point in their lives , but that may be fully convective for the vast majority of their lifetimes , versus models that never have even a transient radiative zone . to verify that this instability is visible in other stellar evolution codes apart from yrec
, we made use of the publically available mesa ( * ? ? ?
* , http://mesa.sourceforge.net/ ) stellar evolution code .
we ran solar calibrated models with z = 0.016498 , y = 0.26275 , mixing length parameter @xmath23 and @xmath24 @xcite at 4.57 gyr @xcite and the default , out - of - the - box physics and numerical parameters released with mesa version 3372 , with the exception of the model timesteps , which we forced to be of order @xmath25 years or smaller .
we found the same qualitative behavior with @xmath1 at the convective boundary in the mesa models , with some quantitative differences .
the fully convective boundary in these models occurs for masses @xmath26 , and the mass range over which the convective episodes are present is @xmath27 ( versus @xmath28 for the yrec models ) . we expect that this difference is due to the detailed differences between the input physics in the two different codes .
although the timescales differ somewhat , the behavior is qualitatively the same , which is evidence in favor of a physical phenomenon rather than numerical artifact . a mesa model for @xmath29
is shown in figure [ fig1 ] among other yrec models for comparison .
the exact masses at which the star is unstable to these convective episodes is also model dependent within yrec .
for example , models run with @xcite atmospheres rather than allard atmospheres become unstable at @xmath30 rather than @xmath31 .
however , in both cases the instability occurs over a mass range of @xmath32 , consistent with a simple shift in the location of the fully convective boundary .
models run at non - solar metallicities also display unstable behavior over slightly different mass ranges .
for example , a @xmath33 model with @xmath34 } = -0.4 $ ] ( referenced to solar metallicity ) displays instabilities with the longer timescales and larger amplitudes of a more massive @xmath35 solar metallicity object .
physical processes such as convective overshoot have not been considered here , and will undoubtedly affect the exact location in mass of the instability boundaries .
there are two distinct cases in which overshooting may affect the instability ; the first is simply by shifting the effective location of the boundaries of the convection zones , resulting in a shift in the masses at which the instability occurs .
the second effect is more subtle , and is related to the interplay and exchange of material between two closely separated convection zones due to overshooting .
the importance of overshooting in a system with closely separated convection zones has been considered in the case of a stars , which are thought to have two thin surface convection zones , one at the hydrogen ionization zone , and another at the deeper second ionization of he .
these are formally separated by a radiative buffer , although it is unclear whether they are truly distinct from one another when overshooting is considered , as @xcite suggest . in the particular case
we consider here , overshooting would serve to connect the core and envelope convection zones either chemically or physically before they formally touch according to the schwarzschild criterion .
yrec treats overshooting as `` overmixing '' only , and does not adjust the temperature gradients in the overshoot regions , and so is incapable of capturing the full behavior of the system when the czs are in this close configuration .
one could imagine that the dilution of the core @xmath1 due to @xmath1 poor material mixed in through overshooting would lead to an overall drop in the @xmath1 luminosity in the core , a global decrease in the radiative temperature gradient , and therefore a shrinking of the convective core .
likewise , the mixing of @xmath1 rich material into the convective envelope might lead to increased burning , and cause the envelope convection zone to deepen slightly .
essentially , if the production timescale for @xmath1 in the core is much shorter than the timescale for the mixing of @xmath1 poor material into the central cz , then the effects of overshooting should be minimal .
if the mixing timescale is comparable or shorter than the production timescale for @xmath1 , then it is possible that the effects of overshooting could prevent these fully convective episodes from occurring at all .
however , if the temperature gradients in the overshoot regions are very close to adiabatic , the boundaries of the the two czs are actually closer in the case of overshooting than they would otherwise be .
further work on this interaction is therefore of interest .
we offer , however , that the convective velocities in these stars should be low , and the stellar material strongly stratified .
the extent of convective envelope overshooting in the sun inferred from helioseismology is less than 0.05 pressure scale heights @xcite , and both studies of cluster cmds @xcite and asteroseismology @xcite suggest that the typical extent of core overshooting in intermediate mass stars is 0.1 - 0.2 pressure scale heights . there has been very little observational or theoretical investigation of the amount of overshooting in stars of a few tenths of a solar mass , but we might argue , given what information we have about other masses , that the degree of overshooting will be minimal . likewise , we have neglected to include the effects of magnetic fields , which could be significant for very active m - dwarfs , and could potentially stabilize regions of the model against convection ( as in * ? ? ?
to investigate whether the luminosity and radius changes associated with the instability are observable , we focus on open clusters , in which we have a single , coeval , chemically homogeneous sample of stars . we chose a simple initial mass function , with @xmath36 , where @xmath37 is a normalization constant , and @xmath38 @xcite .
we sample the imf in @xmath3 bins , and determine our model luminosities and effective temperatures at fixed time for each mass .
we add noise to the model hr diagram by assigning luminosities and effective temperatures randomly sampled from a normal distribution centered on the theoretical values for each stellar mass .
we mimic the presence of binaries by randomly selecting half of the stars in the theoretical open cluster , and assigning a companion of mass @xmath39 where @xmath40 is a random number between 0 and 1 drawn from a uniform distribution , motivated by the distribution of binary mass ratios in @xcite .
we assign the binaries the luminosity weighted mean effective temperature of the two components , and add the two luminosities
. we do not account for the effects of activity .
the presence of the instability produces some weak structure along the single - star ms , which will be challenging to detect in any open cluster survey . however , a few stars lie below the ms , where , with the addition of proper motion or rv membership information , they might be recognized as odd cluster members .
detection of such sources would rely on there being an open cluster with enough members , of an interesting age and metallicity such that this instability is visible .
the presence of this instability could also manifest itself as a scatter in radius as a function of age in m dwarf wide binary systems in which the stellar parameters are well known .
given a sample of stars of the same mass and composition with a variety of ages , one should observe more scatter in the radii of the younger objects in comparison to the older objects .
the effect will only be visible if one can disentangle the contributions from stellar activity and observational uncertainty , both of which are current challenges .
it is important to note , however , that this is potentially an additional source of scatter in m - dwarf radii which is not related to activity or environment , but rather nuclear burning processes .
cataclysmic variables are short period binaries in which a white dwarf accretes mass through roche lobe overflow from a relatively unevolved donor star .
cvs are observed over a wide range of binary periods , ranging from @xmath41 hrs to just over 1 hour .
the primary means of angular momentum loss are postulated to be magnetic braking associated with the stellar wind of the donor star , and losses through gravitational radiation , which operates most efficiently for the closest orbital separations .
one of the most notable features in the cv period distribution is the so - called `` period gap '' ( see * ? ? ?
* ; * ? ? ?
* ) in which we observe a paucity of systems with periods of 2 - 3 hrs .
this instability is interesting in the context of cv systems . given sufficient time and mass - loss rates
, all cv donor stars will pass through the mass range over which these convective episodes occur in single - star models .
the response of the donor star radius to mass - loss is a critical element of the cv picture ; sudden changes could truncate the mass - loss by bringing the system out of roche lobe contact , or conversely result in temporarily inflated accretion rates .
although a detailed discussion of the manner in which this instability interacts with the process of regulated mass - loss through roche - lobe overflow is beyond the scope of this article , we do consider the response of single star models to constant mass - loss rates . under the right conditions
, this instability induces a significant , complicated response in the stellar radius , at an extremely interesting mass range in cv systems .
we apply constant mass - loss rates between @xmath42 to models of @xmath43 and a single mass loss rate of @xmath44 to models over a range of masses ( @xmath45 ) , all of which have been evolved to 1 gyr prior to beginning mass - loss .
this exploration of of mass - loss rates is presented in figure [ fig3 ] . in general ,
relatively unevolved models of the lowest masses ( @xmath46 ) coupled with low mass - loss rates ( @xmath47 ) show the largest radius `` glitches '' due to convective kissing .
whether or not strong radius glitches appear in these models depends sensitively on a combination of the mass - loss rate , initial mass , and evolutionary state of the model . at fixed initial mass ,
the instability does not manifest itself in models until the mass - loss timescale is greater than the thermal timescale , @xmath48 .
for high mass loss rates , the mass loss timescale is always smaller than the thermal timescale , the star is out of thermal balance , and no instability is present , even over the mass ranges where we expect it to occur . for lower mass - loss rates ,
the instability is present and the glitch in radius becomes visible ( see figure [ fig3 ] ) .
the initial mass of the model sets the @xmath1 profile and abundance , which in turn determines if the @xmath1 abundance in the core is sufficient upon arrival at @xmath49 to become unstable .
low mass models have @xmath1 profiles peaked closer to the central region , and in general the absolute central @xmath1 abundance increases with decreasing mass for models of the same initial compositions .
higher mass models , which are more effective at burning @xmath1 , have @xmath1 profiles that peak at larger radii , and generally have more @xmath1 depleted cores . finally , because models of different masses have different nuclear equilibrium timescales for @xmath1 , in general , the later the mass - loss begins , the less @xmath1 is present in the core of the model .
these three effects conspire to make the radius glitches most visible in relatively unevolved , low mass models with low mass - loss rates .
we note a small radius glitch occurs even in models with low mass - loss rates but insufficient @xmath1 to induce the instability , which occurs as @xmath1 is homogenized throughout the star as the model crosses the fully convective boundary ; we believe this is the same feature noted in @xcite and @xcite .
apart from the question of whether this instability affects the degree of mass loss through physically altering the radius of the donor star , one should also note that the composition of any donated material is influenced by the presence of the instability .
both @xcite and @xcite suggest that for @xmath1 mass fractions @xmath50 the ignition of novae may have some sensitivity to the @xmath1 abundance , and that the burning of accreted @xmath1 on the primary can lead white dwarfs that are more luminous than their average accretion rates suggest . in the case of a 0.4 @xmath51 model with a constant mass loss rate of @xmath52 ,
the surface @xmath1 abundance increases by a factor of 2 over the course of the first convective episode , from @xmath53 to @xmath54 .
this increase in the @xmath1 abundance of accreted material would appear at masses above the fully convective boundary , where it would otherwise not exist .
we have negelected to include rotational mixing in our models , which could be non - negligible in the case of such close binary systems in which the secondary may be relatively rapidly rotating due to tidal locking , which could help to drive meridional flows @xcite and result in mixing . furthermore , mixing may also be driven by tidal distortions of the potential due to its close companion @xcite . if the timescale for mixing is shorter than the timescale in between convective kissing episodes , the mixing of material could serve to eliminate episodic mixing events . with the inclusion of tidal deformation , the estimates of the minimum timescale for mixing are of order the timescale between convective episodes for systems at 3 hrs ( see * ? ? ? * fig . 9 for timescales ) . the initial feature in the curve in figure [ fig3 ] , as the @xmath1 in the core
is first mixed throughout the envelope would remain even in the case of strong mixing , but the subsequent variations in the radius would be damped .
therefore , the precise shape of these curves may be modified in the case in which rotational mixing is important , and our neglect of mixing is an important caveat to this discussion .
we have discovered a novel instability in low mass , @xmath55 , stellar models in which a buildup of @xmath1 in the core results in brief convective episodes within the first few gyrs on the main sequence .
these events are accompanied by changes in the radius and luminosity of the star .
we have verified that the same instability appears to occur in the mesa code as well , which suggests that it is not a merely numerical artifact .
stars undergoing these convective episodes may be visible as both single stars or as members of binary systems .
we have demonstrated that the radius of the donor star in a cv system could have a sizable response to the onset of the convective kissing instability , given donor stars in which the mass - loss rates , initial stellar mass , and evolutionary state ensures that there is sufficient @xmath1 available in the core by the time the stellar mass reaches and interesting range .
the complicated structure of the glitches , which includes both increases and decreases in radius over the course of the instability , could interact in interesting ways with the process of cv mass - loss , either by severing contact , or by leading to temporarily inflated mass - loss rates , provided they are not suppressed by rotational mixing processes .
we suggest that this instability be investigated in the context of full cv evolutionary models , where the interaction between the mass - loss rate and the convective kissing instability could be better quantified .
we would like to thank lars bildsten for useful discussion , and to acknowledge the kitp staff of ucsb for their warm hospitality during the research program `` asteroseismology in the space age '' .
we also thank ben shappee and andy gould for useful discussions .
this research was supported in part by the national science foundation of the united states under grant no .
nsf phy0551164 , nsf graduate research fellowship grant rf#743796 ( jvs ) , and nasa grant nnx11ae04 g ( mhp ) . | we report on the discovery of an instability in low mass stars just above the threshold ( @xmath0 ) where they are expected to be fully convective on the main sequence .
non - equilibrium @xmath1 burning creates a convective core , which is separated from a deep convective envelope by a small radiative zone .
the steady increase in central @xmath1 causes the core to grow until it touches the surface convection zone , which triggers fully convective episodes in what we call the `` convective kissing instability '' .
these episodes lower the central abundance and cause the star to return to a state in which is has a separate convective core and envelope .
these periodic events eventually cease when the @xmath1 abundance throughout the star is sufficiently high that the star is fully convective , and remains so for the rest of its main sequence lifetime .
the episodes correspond to few percent changes in radius and luminosity , over myr to gyr timescales .
we discuss the physics of the instability , as well as prospects for detecting its signatures in open clusters and wide binaries .
secondary stars in cataclysmic variables ( cvs ) will pass through this mass range , and this instability could be related to the observed paucity of such systems for periods between two and three hours .
we demonstrate that the instability can be generated for cv secondaries with mass - loss rates of interest for such systems , and discuss potential implications . |
contributions from the meson cloud of the nucleon to deep - inelastic scattering ( dis ) were first discussed in the early 1970s by drell , levy and yan @xcite and sullivan @xcite .
it was of little interest , however , until in the early 1980s a possible enhancement of the pion cloud of a bound nucleon was suggested as an explanation of the nuclear emc effect @xcite .
while that issue is still of interest , the modern importance of the pion cloud in high - energy scattering originates in the work of thomas @xcite , who showed ( within a particular chiral quark model ) that the pion cloud contribution provided a natural explanation for the observed excess of non - strange over strange sea quarks in the nucleon .
this calculation also predicted an excess of @xmath2 over @xmath3 quarks in the nucleon sea , which was confirmed by the violation of the gottfried sum rule observed in the nmc experiment at cern @xcite , and the @xmath4 ratio measured in the drell - yan reaction in @xmath5 and @xmath6 scattering by the e866/nusea collaboration at fermilab @xcite .
although this excess of @xmath2 over @xmath3 , as well as the @xmath7 asymmetry @xcite , was motivated by the physics of spontaneous chiral symmetry breaking in qcd , the early calculations were based on models whose connections to qcd were not manifest .
it was later realized @xcite , however , that because of their origin in chiral loops , these contributions had a nonanalytic dependence on quark mass that could only be generated by a goldstone mechanism , which placed these effects on a more rigorous theoretical footing .
in fact , it is a model - independent consequence of spontaneous chiral symmetry breaking in qcd that @xmath8 and @xmath9 are nonzero .
the only open question is how large these asymmetries actually are .
this issue has become even more important in the last few years because of the widespread interest in the so - called five - quark components of baryon wave functions @xcite .
in particular , there is some evidence that the roper resonance may not be a simple excitation of a single valence quark in the nucleon but rather involves a large five - quark contribution @xcite .
this naturally leads to questions about such components in the wave function of the nucleon , and there have been suggestions that these may be sizable . even though the pentaquark now appears to be defunct , legitimate questions about five - quark components in baryon spectroscopy
remain , especially within models such as the chiral quark soliton model @xcite . in the context of low energy tests of the standard model
this issue is also very topical , with a potential asymmetry between the @xmath10 and @xmath11 parton distributions potentially yielding a large correction to the value of @xmath12 derived from the nutev measurement @xcite .
model - dependent estimates of the five - quark component of the nucleon wave function involving three light quarks plus @xmath13 pairs were recently used to investigate this correction @xcite . in all of these examples , the contribution from the meson cloud of the nucleon has been a crucial factor in reconciling the physics results . indeed , because of the model - independent constraints imposed through chiral symmetry on the nonanalytic behavior @xcite , the presence of meson cloud contributions to the various observables is firmly established , justifying the extensive theoretical attention that has been paid to this issue over the past decade .
( we should note , however , that in chiral perturbation theory observables in general receive contributions from pions as well as local short - distance operators , or counter - terms , and it is possible under renormalization to move strength between them @xcite ; the nonanalytic contributions , on the other hand , remain model independent . )
a further motivation relates to the fact that lattice qcd is now providing considerable information on the moments of parton distribution functions @xcite , but at larger quark masses than occur in nature .
the extrapolation of those moments as a function of quark mass to the physical point is critically dependent on knowing the correct nonanalytic behavior @xcite , and that in turn is uniquely determined by the pion cloud of the nucleon . given the phenomenological importance of the meson cloud , it is unfortunate that the literature contains a number of sometimes contradictory results for meson cloud contributions to dis . in particular , the majority of calculations , from the original work of drell _ et al .
_ @xcite and sullivan @xcite , to the early studies of the nonanalytic behavior @xcite , and essentially all model analyses of pion cloud effects in dis in between @xcite , have used pseudoscalar ( ps ) coupling , which is , by itself , inconsistent with chiral symmetry @xcite .
the restoration of chiral symmetry can be achieved through the addition of a scalar `` @xmath14 '' field ; however , determining its practical consequences for dis is problematic because of uncertainties in identifying the nature of low - mass scalars in meson spectroscopy @xcite .
more recently , the matrix elements of nonsinglet twist-2 operators were computed by chen and ji @xcite and arndt and savage @xcite using lowest order , heavy baryon chiral perturbation theory . by construction this theory uses pseudovector ( pv ) @xmath15 couplings , which are manifestly invariant under chiral transformations @xcite .
the twist-2 matrix elements are related through the operator product expansion to moments of parton distributions measured in inclusive dis . at lowest order in the low - energy expansion ( minimum number of derivatives of the pion field @xmath16 ) ,
the effective chiral lagrangian for the interaction of pions and nucleons , consistent with chiral symmetry , can be written as @xcite @xmath17 where @xmath18 is the nucleon field , @xmath19 mev is the pion decay constant , and @xmath20 is the nucleon axial vector charge .
the first term in the lagrangian ( [ eq : lpin ] ) gives rise to the well - known `` rainbow '' diagram in which a pion is emitted and reabsorbed by the nucleon at different space - time points .
the second is the so - called weinberg - tomozawa term @xcite , in which two pion fields couple to the nucleon at the same point , and gives the leading contribution to s - wave pion
nucleon scattering @xcite .
it also generates the pion tadpole or bubble diagrams , which in the presence of external fields generally give non - vanishing contributions to nucleon matrix elements .
this term does not appear in the ps theory , which leads to some of the differences between the moments of twist-2 parton distributions from pion loops computed in the ps @xcite and pv @xcite theories .
in this paper we present a detailed analysis of the light - cone momentum distributions of pions , and the corresponding recoil baryons , inside a physical nucleon within the chirally symmetric effective field theory defined by the lagrangian ( [ eq : lpin ] ) .
we further contrast the results with those obtained in the ps theory , conventionally defined by the lagrangian @xmath21 where the @xmath1 coupling constant @xmath22 is related to @xmath23 and @xmath24 by the goldberger - treiman relation @xcite , @xmath25 with @xmath26 the nucleon mass . unlike the earlier chiral effective theory calculations which only computed the light - cone distributions of pions @xcite or considered the nonanalytic behavior of their moments to lowest order in the pion mass @xmath27 @xcite , we compute the complete set of diagrams relevant for dis from nucleons dressed by pions resulting from the lagrangians ( [ eq : lpin ] ) and ( [ eq : lps ] ) , without taking the heavy baryon limit .
in particular , we demonstrate explicitly the consistency of the computed distribution functions with electromagnetic gauge invariance .
this requires consideration of the kroll - ruderman terms @xcite , which although entering at higher orders in @xmath27 , are nonetheless essential for ensuring conservation of charge to all orders in @xmath27 @xcite .
formally , the light - cone distribution functions can be defined by considering the vertex renormalization constant @xmath28 for the physical @xmath29 vertex .
if the contributions from the pion cloud are not large , then @xmath30 and one has @xmath31 . the light - cone distributions @xmath32 associated with a particular contribution @xmath33 can then be defined as @xcite @xmath34 where @xmath35 is the momentum fraction of the physical nucleon carried by the pion , with @xmath36 and @xmath37 the four - momenta of the pion and physical nucleon , respectively .
note that in this work we define the light - front `` @xmath38 '' and `` @xmath39 '' components of a four - vector @xmath40 as @xmath41 .
it will be most natural to perform the calculation in light - front coordinates ; however , as demonstrated in ref .
@xcite for the nucleon self - energy , the results for the model - independent , long - distance physics associated with the pion cloud are reproduced in any formalism , be it instant form , covariant , or light - front . because the tadpole or bubble diagrams in the pv theory involve pions emitted and absorbed at the same point , these can only contribute at zero light - cone momentum fraction , @xmath42 . a novel feature of the pv theory , however , is the appearance of @xmath0-function contributions arising also from the rainbow diagram at @xmath42 .
inclusion of these singular terms is in fact vital for preserving gauge invariance in the pv theory . in the following
we will elucidate the origin of the @xmath0-function terms and discuss their possible physical implications .
the rest of this paper is organized as follows .
we derive in sec .
[ sec : fypi ] the light - cone momentum distribution functions for pions in a physical nucleon , including both the pion rainbow and bubble contributions .
the analogous nucleon momentum distributions are discussed in sec .
[ sec : fyn ] , where the kroll - ruderman couplings required by gauge invariance are introduced .
the complete set of pion and nucleon distributions allows us for the first time to explicitly verify the symmetry relations between them .
the application of the results to inclusive dis is described in sec .
[ sec : sumrules ] within the convolution approach , where we discuss the effect of the pion loop corrections , including the @xmath0-function contributions , on various sum rules . finally , we summarize our findings and their implications in sec . [
sec : conc ] .
assuming isospin symmetry , the contributions from pion loops to nucleon matrix elements cancel for isoscalar combinations . in this work
we therefore consider only isovector light - cone momentum distributions , @xmath43 .
the distributions are evaluated by taking the `` @xmath38 '' components of the operators .
we begin by considering the contributions from diagrams involving a direct coupling to the pion , as illustrated in fig .
[ fig : pi ] . the isovector light - cone momentum distribution function for the pion rainbow diagram in fig .
[ fig : pi](a ) is given by @xmath44 where the pion and nucleon propagators are given by @xmath45 respectively . in eq .
( [ eq : fypi_def ] ) and throughout this work the spinors @xmath46 are normalized such that @xmath47 .
( note that for the pion coupling diagrams , the isovector combination @xmath48 is equivalent to the notation `` @xmath49 '' @xcite . ) using the dirac equation , the distribution function in eq .
( [ eq : fypi_def ] ) can be decomposed into several terms , @xmath50 2y\ , \delta\left(y - { k^+\over p^+}\right ) .
\label{eq : fypi_d}\end{aligned}\ ] ] to evaluate the expression in eq .
( [ eq : fypi_d ] ) , we proceed to first integrate over the @xmath51 component of the pion momentum . without loss of generality
, we work in a frame where @xmath52 . for the first term in ( [ eq : fypi_d ] ) we use cauchy s integral formula and take the single nucleon pole by closing the contour in the upper half - plane , @xmath53 note that the arc contribution at infinity vanishes for this term .
a straightforward calculation then yields the contribution which we associate with the on - shell part of the nucleon propagator , @xmath54 ^ 2 } , \label{eq : fyon}\end{aligned}\ ] ] where for convenience we have factored out all isospin factors .
this term corresponds to the usual pion distribution function for the `` sullivan process '' @xcite , which is computed from a ps @xmath15 coupling . performing the @xmath55 integration with an ultraviolet cut - off @xmath56 gives @xmath57 + { \cal o}\left ( \frac{m^2}{\lambda^2 } \right ) , \nonumber\\ & & \label{eq : fyon_final}\end{aligned}\ ] ] where only the leading term in @xmath56 has been kept . in the second and third terms in eq .
( [ eq : fypi_d ] ) , the cancellation of the nucleon propagators @xmath58 leaves tadpole - like contributions involving pion propagators only .
this will have important consequences for the structure and interpretation of the distribution functions .
integrating the second term in ( [ eq : fypi_d ] ) proportional to @xmath59 over @xmath51 gives a result which is zero everywhere except at @xmath60 , @xmath61 after multiplication by @xmath62 , this term therefore vanishes .
evaluation of the third term in eq .
( [ eq : fypi_d ] ) proportional to @xmath63 requires particular care .
since @xmath64 , integration over @xmath51 involves terms such as in ( [ eq : intdpi ] ) which vanish , as well as terms of the type @xmath65 . to evaluate the latter
, one can make use of the identity @xmath66 to write @xmath67 where @xmath68 is a mass parameter .
this gives rise to a @xmath0-function contribution ( again defined without isospin factors ) @xmath69 with the total pion distribution function then given by @xmath70 when integrated over @xmath55 from 0 to @xmath56 , the function @xmath71 becomes @xmath72 where we have isolated the leading nonanalytic term in @xmath73 , whose coefficient is independent of the short - distance physics parametrized by @xmath56 .
thus the pion rainbow diagram has contributions from the regular @xmath74 region as well as the ultrasoft @xmath75 point . in the ps formulation of pion nucleon interactions , where the pion light - cone momentum distribution is given by just the first term in eq .
( [ eq : fypi_d ] ) , the @xmath76 term in ( [ eq : fypi ] ) does not arise , and the total ps pion distribution is given by @xcite @xmath77 in addition to the pion rainbow diagram in fig .
[ fig : pi](a ) , there is also a contribution from the term in the lagrangian ( [ eq : lpin ] ) involving two pion fields , which gives rise to the bubble diagram in fig .
[ fig : pi](b ) . a straightforward calculation of the resulting contribution to the pion light - cone momentum distribution gives @xmath78 a calculation similar to that for the @xmath71 term above gives the final pion bubble contribution as @xmath79 where for convenience we have defined the distribution @xmath80 without the isospin factor 2 .
integrating over @xmath55 from 0 to @xmath56 then yields a result for @xmath80 similar to that for the @xmath71 term in eq .
( [ eq : fydelta ] ) , @xmath81 finally , combining the results in this section , the total isovector pion distribution in the nucleon , including both the pion rainbow and bubble diagrams in fig .
[ fig : pi ] , can be written as @xmath82 \nonumber\\ & & + \ { ( 1-g_a^2 ) \over ( 4\pi f_\pi)^2 } m_\pi^2 \log m_\pi^2\ \delta(y)\ + \
\textrm{terms involving $ \lambda$ } , \label{eq : fypi_final}\end{aligned}\ ] ] where we have explicitly separated the long - distance contributions from the short - distance effects involving the cut - off @xmath56 . in particular , the total term proportional to the @xmath0-function is found to vanish in the limit @xmath83 , as one would have in a ps pion
nucleon theory @xcite .
note that a similar behavior for the @xmath84 term is observed for the pion cloud contribution to the isovector form factor @xcite .
for the nucleon rainbow diagram in fig .
[ fig : n](a ) , the distribution function for the neutron is @xmath85 that of the proton , so that for the isovector distribution one has @xmath86 . specifically , the isovector nucleon distribution function is given by @xmath87 which , using the dirac equation , can be written as @xmath88 \delta\left(y - { k^+\over p^+}\right ) .
\label{eq : fyn_def}\end{aligned}\ ] ] the first term ( @xmath89 in eq .
( [ eq : fyn_def ] ) corresponds to the on - shell nucleon contribution , and is proportional to the function @xmath90 in eq .
( [ eq : fyon ] ) . the second term
( @xmath91 arises from the off - shell components of the nucleon propagator , while the third term involving the single pion propagator ( @xmath92 ) contributes only at @xmath42 , and is proportional to @xmath71 .
the total isovector nucleon light - cone distribution function arising from the nucleon rainbow diagram is then given by @xmath93 where @xmath94 with the @xmath55 integration taken up to the cut - off scale @xmath56 .
for the pseudoscalar model the nucleon light - cone momentum distribution is again just given by the on - shell term in eq .
( [ eq : fyn ] ) , @xmath95 this result was also obtained in the infinite momentum frame calculation of drell , levy and yan @xcite .
on the other hand , the pv theory contains , in addition to the off - shell and @xmath0-function pieces , the contribution from the operator insertion at the @xmath96 vertex , fig .
[ fig : n](b ) . the distribution function associated with this diagram
can be written @xmath97 performing the @xmath51 integration and again using the relation in eq .
( [ eq : dpi_id ] ) then gives @xmath98 comparing the pion bubble and tadpole contributions , eqs .
( [ eq : fypitad ] ) and ( [ eq : fyntad ] ) , one finds that their sum vanishes , @xmath99 these contributions themselves therefore have no net effect on the sum of the pion and nucleon light - cone distributions , and hence on the total nucleon charge , as will be discussed in sec .
[ sec : sumrules ] below . finally , the light - cone momentum distribution associated with the kroll - ruderman diagrams in fig .
[ fig : n](c ) , which arises from the derivative coupling in the pv theory , is given by @xmath100 u(p)\ , { i \over d_\pi}\ , \delta\left(y - { k^+\over p^+}\right ) \\ & = & 4i \left ( { g_a \over 2 f_\pi } \right)^2 \int\!\!{d^4k \over ( 2\pi)^4 } \left [ { 4 m^2 y \over d_\pi d_n}\ + \ { 2 \over d_\pi } \right]\ , \delta\left(y - { k^+\over p^+}\right ) .
\label{eq : fykr_def}\end{aligned}\ ] ] after @xmath51 integration , one then obtains @xmath101 one observes then that the sum of the pion [ eq . ( [ eq : fypi ] ) ] and kr [ eq . ( [ eq : fykr ] ) ] distributions is proportional to the nucleon light - cone distribution ( eq . ( [ eq : fyn ] ) ) , @xmath102 which , as will be discussed in the next section , is a necessary condition for the gauge invariance of the theory .
the effect of pion loops on observables such as parton distribution functions can be computed by considering dis from a nucleon viewed in the infinite momentum frame @xcite . here
the total contribution to the nucleon isovector quark distribution @xmath103 from the pionic corrections in figs .
[ fig : pi ] and [ fig : n ] can be written in convolution form ( for a review see _
e.g. _ ref .
@xcite ) , @xmath104 where @xmath105 is the isovector quark distribution in the bare nucleon state @xmath106 , and the sum over @xmath33 includes the nucleon and pion rainbow , bubble , tadpole and kr contributions .
the symbol @xmath107 denotes convolution ; for the pion rainbow contribution , for instance , one has @xmath108 , where @xmath109 is the isovector quark distribution in the pion , and similarly for the other terms . while the shapes of the bare quark distribution functions , and especially for the kr and tadpole terms , are not known _ a priori _ , charge conservation is nevertheless ensured as long as the distributions are normalized according to @xmath110 for all @xmath111 . the wave function renormalization factor @xmath112 in eq .
( [ eq : qniv ] ) can be evaluated from the nucleon self - energy @xmath113 by @xcite , @xmath114 where the derivative is taken at the nucleon pole , @xmath115 . for weak pion fields ,
explicit evaluation gives @xmath116 where the factor 3 arises from the sum of the proton and neutron intermediate state contributions , for either a proton or neutron initial state .
in fact , eq .
( [ eq : z2 ] ) is required by the ward - takahashi identity which relates the wave function and vertex renormalization factors by @xmath117 , where @xmath118 is the lowest moment of the nucleon distribution function @xmath119 for a _ proton _ initial state .
integrating the convolution expressions in eq .
( [ eq : qniv ] ) over all @xmath120 and using the relation in eq .
( [ eq : cons ] ) one obtains , for a nonsinglet quark distribution @xmath121 normalized to unity , the quark number sum rule , @xmath122 with the contributions from the @xmath123 and @xmath124 terms canceling as in eq .
( [ eq : tadsumrule ] ) . from the expression for @xmath112 in eq .
( [ eq : z2 ] ) one can verify that the baryon number of the nucleon is not modified by pion loops .
inclusion of the @xmath42 contributions is vital for formal sum rules to be satisfied ; if these were to be evaluated for @xmath125 only , the @xmath84 terms would be missed .
indeed , as was illustrated in ref .
@xcite in the case of the bukhardt cottingham sum rule for the twist-3 polarized parton distribution @xmath126 ,
if such @xmath127 terms are present then sum rules based on the formal operator product expansion appear to be violated when evaluated for @xmath128 only . the same was observed for the corresponding sum rule for the chirally odd twist-3 parton distribution @xmath129 . in the context of dispersion relations ,
@xmath127 terms reflect the presence of subtractions . a deep - inelastic scattering experiment to test such sum rules
would thus find them to be violated , since @xmath130 is never measured . in the convolution calculation
the @xmath84 terms would thus imply a @xmath127 contribution to the quark distribution @xmath121 , which would lead to the violation of both the gottfried @xcite and adler sum rules @xcite , if these are applied to nonzero @xmath120 only .
the appearance of the @xmath127 terms in the isovector quark distributions is possibly an artifact of taking chiral perturbation theory too literally in this context .
for example , although a bubble diagram in the chiral effective theory inevitably leads to an @xmath10-independent contribution to the forward compton amplitude and hence a subtraction in the corresponding dispersion relation it may no longer be justified to assume a point - like @xmath15 four - point coupling that gives rise to the bubble when @xmath131 .
while a @xmath127 contribution formally appears in the chiral effective theory framework , this contribution will presumably soften into a peak near @xmath130 in qcd .
nevertheless , even if the adler sum rule is in the end satisfied in qcd , the presence of @xmath127 terms in the effective theory indicates that in order to confirm this sum rule it is essential to include the small-@xmath120 regime .
of course , none of these considerations applies to the case of the ps pion - nucleon coupling , which contains nonzero contributions only at @xmath74 .
in this paper we have presented a detailed derivation of the isovector light - cone momentum distributions of pions in the nucleon for both the pseudovector and pseudoscalar @xmath15 interactions . in the pv theory the direct coupling to the pion
gives rise to a pion rainbow distribution , @xmath132 , and a distribution from the pion bubble , @xmath124 , with the latter involving emission and absorption of a pion at the same point .
consequently , the bubble contribution to the momentum distribution exists only at zero light - cone momentum , @xmath42 .
we found , however , that the pion rainbow diagram also has a @xmath0-function component , on top of the regular terms at @xmath125 .
the @xmath84 contributions arise because the pv @xmath1 vertex is proportional to the momentum of the pion , which leads to a more singular integrand in the @xmath133 limit than in the case of the momentum - independent ps coupling .
in addition to the pion distribution , we have also computed the isovector light - cone momentum distribution of the corresponding recoil nucleon dressed by the pion , including contributions from the rainbow diagram with nucleon coupling , @xmath134 , and a tadpole coupling at the @xmath135 vertex , @xmath123 , as well as the kroll - ruderman terms @xmath136 which arise at the same order in the chiral expansion of the pv lagrangian .
the latter are in fact essential for preserving the gauge invariance of the pv theory .
the nucleon rainbow distribution @xmath134 contains pieces that are associated with the off - shell components of the nucleon propagator , in addition to the on - shell components , and a @xmath0-function term analogous to that in the pion light - cone distribution @xmath132 .
in contrast , the nucleon distribution in the ps model only contains the on - shell contributions .
the kr distribution @xmath136 has contributions from the off - shell components of the nucleon and from the singular @xmath0-function terms .
the singular nucleon tadpole distribution is related to the pion tadpole by @xmath137 for all @xmath62 .
combined , the sum of the pion rainbow and kr distributions in the pv theory is shown to be proportional to the nucleon rainbow distribution , @xmath138 which is a necessary condition for ensuring gauge invariance . within the convolution model
, we have demonstrated that this in fact gives the correct number sum rule : the sum of all the pion cloud contributions does not alter the valence quark number of the nucleon , although it does of course affect the shape .
this is also true in the ps formulation .
our results pave the way for future phenomenological applications of pion cloud models that are manifestly consistent with the chiral symmetry properties of qcd .
most calculations in the literature about the effects of the pion cloud in deep - inelastic scattering have been made using the ps @xmath15 coupling @xcite , for which chiral symmetry is not manifest without the addition of a scalar field .
in fact , the absence of the @xmath0-function contributions in the ps version of the rainbow diagrams leads to the coefficient of the leading nonanalytic term in the chiral expansion of the quark distribution function moments that differs by a factor 4/3 from the pv result @xcite .
a detailed study of the phenomenological consequences of the distributions derived here in the pv theory will be the subject of an upcoming study @xcite .
this work was supported by the doe contract no .
de - ac05 - 06or23177 , under which jefferson science associates , llc operates jefferson lab , doe contract no .
de - fg02 - 03er41260 , the doe science undergraduate laboratory internship ( suli ) program , and the australian research council through an australian laureate fellowship and the arc centre of excellence in particle physics at the terascale . | we compute the light - cone momentum distributions of pions in the nucleon in chiral effective theory using both pseudovector and pseudoscalar pion nucleon couplings . for the pseudovector coupling
we identify @xmath0-function contributions associated with end - point singularities arising from the pion nucleon rainbow diagrams , as well as from pion bubble and tadpole diagrams which are not present in the pseudoscalar model .
gauge invariance is demonstrated , to all orders in the pion mass , with the inclusion of kroll - ruderman couplings involving operator insertions at the @xmath1 vertex .
the results pave the way for phenomenological applications of pion cloud models that are manifestly consistent with the chiral symmetry properties of qcd . |
in situ observations in the solar wind , in planetary magnetosheaths , in the heliosheaths and in other weakly collisional , generally turbulent astrophysical plasmas show isolated or wave - trains of compressible pressure - balanced structures @xcite .
many of these structures are thought to be generated by the mirror instability driven by the perpendicular particle temperature anisotropy @xcite .
the mirror instability has peculiar features .
it generates nonpropagating modes ( at least in a plasma without differential streaming ) and near threshold the unstable modes appears on fluid scales , i.e. , on large scales with respect to the particle characteristic scales .
on the other hand , the instability is resonant , the resonant particles ( with nearly zero parallel velocities with respect to the ambient magnetic field ) have a strong influence on the instability growth rate @xcite .
the nonlinear properties of the mirror instability are not well understood .
@xcite proposed a nonlinear model for this instability near threshold , based on a reductive perturbative expansion of the vlasov - maxwell equations .
this model extends the mirror dispersion relation by including the dominant nonlinear coupling whose effect is to reinforce the mirror instability . in this approach both the linear and nonlinear properties are strongly sensitive to the details of the proton distribution function @xcite . in the perturbative nonlinear model
the particle distribution is , however , fixed . for bi - maxwellian particle distribution functions
the nonlinear model predicts formation of magnetic depressions / holes at the nonlinear stage of the instability . on the other hand ,
direct numerical simulations typically show generation of magnetic enhancements / humps @xcite .
this behavior is in agreement with expectations based on the energy minimization argument in the simplified framework of usual anisotropic magnetohydrodynamics @xcite .
@xcite attempted to combine the reductive perturbative expansion approach with the quasilinear approximation @xcite .
this combined model leads to a fast deformation of the proton distribution function that modifies the ( sign of the ) nonlinear term , and , consequently , magnetic humps are generated in agreement with fully self - consistent simulations .
the quasilinear approximation is however questionable in the case of coherent structures and , moreover , one expects particle trapping to be important at the nonlinear level of the mirror instability @xcite . in situ observations in the terrestrial magnetosheath
show that mirror magnetic humps are typically observed in the mirror - unstable plasma whereas in the mirror - stable plasma magnetic holes are more probable @xcite . the two - dimensional hybrid expanding box simulation of a homogeneous plasma system ( with the magnetic field in the simulation box , without turbulent fluctuations but with an expansion that drives the perpendicular temperature anisotropy ( cf .
* ; * ? ? ?
* ) ) of @xcite predicts that in a high - beta plasma the mirror modes are dominant , and , as the expansion pushes the system to lower betas , the system becomes dominated by the proton cyclotron waves whereas the mirror mode structure continuously disappear .
the mirror modes , however , survive relatively far the stable region where they are ( linearly ) damped ; in the unstable region the mirror modes have the form of magnetic humps and on the way to the stable region they transform to magnetic holes similar to the observations in the terrestrial magnetosheath @xcite .
the transition from humps to holes is not understood but is in agreement with the energetic arguments .
the mirror instability is usually investigated in homogeneous or weakly inhomogeneous plasmas @xcite . behavior of this instability is a strongly turbulent ( and strongly inhomogeneous ) plasma such as the solar wind one is an open problem . as in the case of the oblique fire hose instability
@xcite one expects that the mirror instability coexists with plasma turbulence if it is fast enough to compete with the turbulent cascade . in this paper
we investigate properties of the mirror instability in 2-d expanding box simulation where we include important turbulent plasma motion .
the paper is organized as follows : section [ code ] describes the numerical code , section [ simul ] presents the simulation results , and in section [ discussion ] we discuss the obtained results .
in this paper we test the relationship between proton kinetic instabilities and plasma turbulence in the solar wind using a hybrid expanding box model that allows to studying self - consistently physical processes at ion scales . in the hybrid expanding box model a constant solar wind radial velocity @xmath0 is assumed .
the radial distance @xmath1 is then @xmath2 where @xmath3 is the initial position and @xmath4 is the initial value of the characteristic expansion time @xmath5 .
transverse scales ( with respect to the radial direction ) of a small portion of plasma , co - moving with the solar wind velocity , increase @xmath6 .
the expanding box uses these co - moving coordinates , approximating the spherical coordinates by the cartesian ones @xcite .
the model uses the hybrid approximation where electrons are considered as a massless , charge neutralizing fluid and ions are described by a particle - in - cell model @xcite .
here we use the two - dimensional ( 2-d ) version of the code , fields and moments are defined on a 2-d @xmath7@xmath8 grid @xmath9 ; periodic boundary conditions are assumed .
the spatial resolution is @xmath10 where @xmath11 is the initial proton inertial length ( @xmath12 : the initial alfvn velocity , @xmath13 : the initial proton gyrofrequency ) .
there are @xmath14 macroparticles per cell for protons which are advanced with a time step @xmath15 while the magnetic field is advanced with a smaller time step @xmath16 .
the initial ambient magnetic field is directed along the @xmath17 direction , perpendicular to the simulation plane ( that includes the radial direction @xmath18 ) , @xmath19 , and we impose a continuous expansion in @xmath7 and @xmath17 directions with the initial expansion time @xmath20 . due to the expansion with the strictly transverse magnetic field the ambient density and the magnitude of the ambient magnetic field decrease as @xmath21
while @xmath22 ( the proton inertial length @xmath23 increases @xmath6 , the ratio between the transverse sizes and @xmath23 remains constant ; the proton gyrofrequency @xmath24 decreases as @xmath25 ) .
a small resistivity @xmath26 is used to avoid accumulation of cascading energy at grid scales ; we set @xmath27 ( @xmath28 being the magnetic permittivity of vacuum ) .
the simulation is initialized with an isotropic 2-d spectrum of modes with random phases , linear alfvn polarization ( @xmath29 ) , and vanishing correlation between magnetic and velocity fluctuations .
these modes are in the range @xmath30 and have a flat one - dimensional ( 1-d ) ( omnidirectional ) power spectrum with rms fluctuations @xmath31 .
we set initially the parallel proton beta @xmath32 and the system is characterized by a perpendicular temperature anisotropy @xmath33 ; for these parameters the plasma system is already unstable with respect to the mirror instability , however , the geometrical constraints and the presence of relatively strong fluctuations inhibit the growth of mirror modes .
electrons are assumed to be isotropic and isothermal with @xmath34 at @xmath35 .
figure [ evol ] shows the evolution of different quantities in the simulation as functions of time : the fluctuating magnetic field ( panel a , solid line ) perpendicular @xmath36 and ( a , dashed ) parallel @xmath37 with respect to @xmath38 ; ( b ) the average squared parallel current @xmath39 ; the ( c , solid ) parallel @xmath40 and ( c , dashed ) perpendicular @xmath41 proton temperatures ( the @xmath42 and @xmath43 directions are here with respect to the local magnetic field ; the dotted lines on panel c denote the corresponding cgl predictions ) ; ( d , solid ) the nonlinear eddy turnover time @xmath44 at @xmath45 ( d , dotted ) the expansion time @xmath46 , and ( d , dashed ) the linear time @xmath47 for the mirror instability ; here @xmath48 is the maximum growth rate of the mirror instability in the corresponding homogeneous plasma with bi - maxwellian protons .
figure [ evol ] gives an overview of the simulation : initially , the parallel current fluctuations are generated , @xmath39 ( normalized to @xmath49 ) reaches a maximum at @xmath50 indicating the presence of a well - developed turbulent cascade @xcite .
after that the system is dominated by a decaying turbulence , @xmath51 decreases . during the initial phase , the compressible component @xmath37 is also generated , and then decays ; at later times , however , @xmath52 stagnates and even increases .
this indicates generation of compressible fluctuations .
figure [ evol]c shows that the parallel and perpendicular temperatures roughly follow the double adiabatic predictions .
figure [ evol]d presents a comparison between characteristic timescales ; the longest timescale is the expansion time , the system is ( on average ) unstable with respect to the mirror instability but the nonlinear eddy turnover time @xmath44 at @xmath45 is faster than the mirror linear time but at later times the two timescales become comparable
. this may be favorable for generation of compressible mirror modes that may be responsible for the increasing compressible magnetic fluctuations . and ( dashed ) parallel @xmath37 with respect to @xmath38 ;
( b ) the average squared parallel current @xmath39 ; ( c ) the parallel @xmath40 ( solid line ) and perpendicular @xmath41 ( dashed line ) proton temperatures ( the @xmath42 and @xmath43 directions are here with respect to the local magnetic field ; the dotted lines denote the corresponding cgl predictions ) ; ( d ) ( solid ) the nonlinear eddy turnover time @xmath44 at @xmath45 ( dotted ) the expansion time @xmath46 , and ( dashed ) the linear time @xmath53 for the mirror instability .
[ evol],width=302 ] figure [ spec ] presents 1-d power spectral density ( psd ) of the ( left ) perpendicular @xmath54 , ( middle ) parallel @xmath55 , and total ( right ) @xmath56 as functions of @xmath57 at @xmath35 ( dotted line ) , @xmath58 ( blue ) @xmath59 ( green ) , @xmath60 ( red ) , and @xmath61 ( black solid ) .
the perpendicular component @xmath54 has roughly a kolmogorov - like slope on large scales that steepens on the sub - ion scales .
the slopes of @xmath54 , @xmath55 , and @xmath56 in the sub - ion range ( below the transition / break ) are quite similar , about @xmath62 . however , the range where the spectra are power - law like is quite narrow , at smaller scales @xmath63 they flatten ( especially at later times ) indicating a bottleneck problem possibly connected with the numerical noise .
the amplitude of the @xmath54 spectra decreases with time owing to the cascade and the expansion .
the amplitude of the compressible ( @xmath55 ) spectrum also initially decreases but , at later times , the level of fluctuations on large scales increases ( see figure [ evol]a ) .
the incompressible @xmath54 fluctuations dominate the total spectrum but , at later times , the compressible component @xmath55 becomes important on relatively large scales and importantly contributes to the total power spectra of @xmath56 .
figure [ specun ] presents 1-d psd of the proton velocity field @xmath64 ( left ) and @xmath65 ( middle ) , and the proton ( number ) density ( right ) as functions of @xmath57 at @xmath35 ( dotted line ) , @xmath58 ( blue ) @xmath59 ( green ) , @xmath60 ( red ) , and @xmath61 ( black solid ) .
the power spectra of @xmath64 exhibit an exponential - like behavior from large to sub - ion scales ( alternatively , this may be a smooth transition between two power - law like dependencies but it s hard to distinguish between the two for the relatively short range of wave vectors ) ; for @xmath66 the spectra are dominated by the numerical noise due to the finite number of particles per cell @xcite .
the power spectra of @xmath65 are relatively flat at large scales and below about @xmath67 they exhibit rather power - law like properties ( with a slope about @xmath68 ) and again , for @xmath66 the spectra are dominated by the numerical noise .
the amplitudes of @xmath64 and @xmath65 decreases with time but at later times they are roughly constant .
the power spectra of @xmath69 at early times have properties of two power laws ( but the ranges of wave vectors are too short to be sure ) with a relatively thin transition , and their amplitude decreases in the time . at later times
the amplitude of density fluctuations increases and has a property of a wide spectral peak with the maximum around @xmath70 . as in the case of the velocity
fluctuations the density spectra are dominated by the numerical noise for @xmath66 .
the generation of compressible fluctuations affects some ratios used to analyze properties of turbulent fluctuations .
figure [ rats ] shows ( left ) the ratio between perpendicular electric and magnetic fluctuations , ( middle ) the ratio between squared amplitudes of the parallel and the total magnetic fluctuations , and ( right ) the ratio between squared amplitudes of density and perpendicular magnetic fluctuations as functions of @xmath57 at different times .
the simulation results show that the transverse fluctuations are not strongly affected by the presence of the compressible fluctuations ( cf .
* ; * ? ? ?
* ) whereas , unsurprisingly , the compressible ratios @xmath71 and @xmath72 are strongly affected ( cf .
* ; * ? ? ?
* ) .
the spatial properties of the magnetic fluctuations are shown in figure [ bb ] where @xmath36 and @xmath73 are displayed as functions of @xmath7 and @xmath8 at different times : ( top ) @xmath59 , ( middle ) @xmath60 , and ( bottom ) @xmath61 .
only a part of the simulation box is shown ; note that the radial , @xmath8 size of the simulation box normalized to @xmath74 decreases in time as @xmath75 ( see the animation corresponding to figure [ bb ] ) .
the slow expansion introduces an anisotropy with respect to the radial direction @xcite ; this happens mainly on the large scales , the turbulent characteristic time scales on the scales resolved in the present simulation are much faster then the expansion time so that no clear anisotropy is observed in our simulation ( cf .
figure [ bb ] shows the turbulent field of magnetic islands / vortices in @xmath36 and formation of localized magnetic enhancements / humps in the compressible magnetic component @xmath73 that are evident at later times but weak signatures of these structures are already seen at @xmath59 .
the compressible structures are likely the expected mirror mode structures , a more detailed analysis indicates that these structures are standing in the local plasma frame ( they are moving with the turbulent plasma flow , see the animation corresponding to figure [ bb ] ) . and ( right )
@xmath73 as functions of @xmath7 and @xmath8 for ( top ) @xmath59 , ( middle ) @xmath60 , and ( bottom ) @xmath61 .
the solid lines show selected ( projected ) magnetic field lines .
only a part of the simulation box is shown .
[ bb],width=302 ] a detailed view of the spatial structure is displayed in figure [ profb ] showing 1-d cuts of @xmath76 , @xmath77 , @xmath73 ( all normalized to @xmath78 ) , and @xmath69 ( normalized to @xmath79 ) as functions of @xmath7 at @xmath80 and @xmath61 ( see figure [ bb ] , bottom ) . @xmath76 and @xmath77 components of the fluctuating magnetic field have a complex structure , at around @xmath81 the cut passes a center of a relatively large magnetic vortex . close to this center
the compressible component @xmath73 forms a magnetic hump with a strong amplitude @xmath82 .
the magnetic enhancement is compensated by a density decrease ; the magnetic hump / density hole structure is roughly at pressure balance .
, @xmath77 , @xmath73 ( all normalized to @xmath78 ) , and @xmath69 ( normalized to @xmath79 ) as functions of @xmath7 at @xmath80 for @xmath61 .
( see figure [ bb ] , bottom ) .
[ profb],width=302 ] the pressure - balanced magnetic humps are characterized by an anti - correlation between the magnetic field component @xmath83 and the proton density and the distribution of @xmath83 values has a skewed distribution with a positive skewness @xcite .
figure [ corskew ] shows the evolution of the correlation between @xmath84 and @xmath83 and the skewness of @xmath83 , @xmath85 , calculated over the whole box , as functions of time .
initially in the simulation , @xmath83 and @xmath84 are correlated indicating fast mode - like properties . @xmath83 and
@xmath84 rapidly become anti - correlated suggesting a slow mode - like behavior ; a similar evolution is seen in standard 2-d hybrid simulations of @xcite . at later times
the anti - correlation is strengthened by development of mirror structures .
the skewness @xmath85 starts around zero , then steadily increases , and for @xmath86 saturates around @xmath87 .
it is interesting to note that a similar analysis applied to the results from standard 2-d hybrid simulations of @xcite shows that the skewness @xmath85 is negative in a well - developed turbulent cascade starting from an isotropic proton distribution for a wide range of betas . and the proton number density @xmath88 and ( bottom ) of the skewness of @xmath83 , @xmath85 , as functions of time .
[ corskew],width=302 ] figure [ bean ] shows the simulation results @xmath89 at different times , ( top ) @xmath59 , ( middle ) @xmath60 , and ( bottom ) @xmath61 .
the shades of blue display the distribution of the local ( grid ) values whereas the solid circles indicate the values averaged over the simulation box .
the empty circle denotes the initial condition and the solid line gives the evolution of the averaged values . during the evolution , a large spread of local values develops in the space @xmath89 ( cf .
* ; * ? ? ?
the expansion drives the system towards more unstable situations but the development of mirror modes reduces the anisotropy and tends to stabilize the system . a small subset of local values in the space
@xmath89 have smaller values of the maximum growth rates ( compared to the average value ) ; the places where the mirror instability is weakened appear in the vicinity of the mirror structures .
the reduction of the local proton temperature anisotropy is mainly governed by the enhanced magnetic field that leads to higher proton perpendicular temperatures , the magnetic moment of protons is varying only weakly ( cf .
the variation of the magnetic field is not , however , the only saturation mechanism .
figure [ vdf ] shows the proton velocity distribution @xmath90 ( averaged over the simulation box ) as a function of parallel and perpendicular velocities @xmath91 and @xmath92 ( with respect to the local magnetic field ) at different times : ( top ) @xmath59 , ( middle ) @xmath60 , and ( bottom ) @xmath61 . the averaged proton distribution function exhibit a clear flattening around @xmath93 , i.e. , @xmath94 for @xmath95
this is compatible with the quasi - linear diffusion of protons through the landau ( transit time ) resonance @xcite . at different times .
the empty circles give the initial condition whereas the solid circles denote the average values and the solid lines show their evolution .
the dashed contours show the maximum growth rate @xmath48 ( in units of @xmath24 ) of the mirror instability as a function of @xmath96 and @xmath97 in the corresponding plasma with bi - maxwellian protons .
[ bean],width=302 ] both the linear and nonlinear properties of the mirror instability in the nonlinear reductive perturbative model @xcite are sensitive to the details of the proton distribution function in the resonant region @xmath98 .
the flattening observed in the simulation likely modifies the nonlinear properties and leads to generation of magnetic humps ( instead of holes that are expected for bi - maxwellian velocity distribution functions
@xcite ) in agreement with the simulation results . as a function of parallel and perpendicular velocities @xmath91 and @xmath92 ( with respect to the local magnetic field ) for ( top ) @xmath59 , ( middle ) @xmath60 , and ( bottom ) @xmath61 . [ vdf],width=302 ]
the presented 2-d hybrid simulation of plasma turbulence with the expansion forcing demonstrates that mirror instability may coexist with the fully developed ( strong ) turbulence and generate compressible , nonpropagating pressure - balanced magnetic structures with amplitudes comparable to or even greater than those of the ambient turbulent fluctuations .
the compressible component of the magnetic field @xmath55 becomes at later times important on scales comparable to and larger than the typical proton scales and this affects the total power spectra of @xmath56 around the transition between large mhd and sub - ion scales ( cf .
* ) , as well as different compressibility ratios .
the mirror structures reduce locally the temperature anisotropy through a fluid mechanism ( generation of enhanced magnetic field that increases the proton perpendicular temperature ) and through a quasilinear - like proton scattering via the landau ( transit time ) resonance .
the role of particle trapping in the present case is unclear @xcite .
the dominant , compressible magnetic component of the mirror structures only weakly interacts with the incompressible turbulent alfvnic fluctuations through the main fluid nonlinearities . on the other hand , the minor , transverse component of mirror fluctuations
have a vortex like properties around the compressible structures @xcite and likely couples directly to the turbulent plasma motions ( the present simulation indicates presence of such vortical structures ) .
further work is necessary to understand the interaction between turbulence and the mirror instability . in this paper
we drive the temperature anisotropy by the expansion with the transverse magnetic field .
we expect a similar evolution for other driving forces that generate the perpendicular proton / ion temperature anisotropy @xcite . our work is relevant mainly for high beta plasmas where the mirror instability is dominant ; for low and moderate beta plasmas the ion cyclotron instability is prevalent @xcite .
the nonlinear competition between these instabilities is a nontrivial problem even in a homogeneous system and generally requires fully three - dimensional ( 3-d ) simulations @xcite . in the present case ,
both turbulence and the 2-d geometry constraints strongly affect the dynamics of the mirror instability . in the 2-d simulation box
we have only limited access to oblique modes ; however , the mirror instability appear at strongly oblique angles with respect to the ambient magnetic field ( in a homogeneous plasma system ) near threshold @xcite .
the 2-d geometry is likely a smaller problem for the mirror instability compared to the oblique fire hose ( cf .
* ) but , in any case , 3-d simulations are needed to investigate the interplay between turbulence and instabilities .
we expect that 3-d simulations with turbulent fluctuations will exhibit an evolution similar to the 2-d simulation of @xcite modified by turbulence ; this will be subject of future work . despite the limitations ,
the present simulation results confirm that the mirror instability is a viable mechanism that can generate magnetic pressure - balanced structures in turbulent astrophysical plasmas .
the authors wish to acknowledge valuable idea exchanges with t. tullio .
. acknowledges grant 15 - 10057s of the czech science foundation .
is funded by fondazione cassa di risparmio di firenze , through the project `` giovani ricercatori protagonisti '' .
l.m . was funded by the uk stfc grant st / n000692/1 .
the ( reduced ) simulation data are available at the virtual mission laboratory portal ( http://vilma.asu.cas.cz ) developed within the european commission s 7^th^ framework programme under the grant agreement # 284515 ( project-shock.eu ) . | the relationship between a decaying strong turbulence and the mirror instability in a slowly expanding plasma is investigated using two - dimensional hybrid expanding box simulations .
we impose an initial ambient magnetic field perpendicular to the simulation box , and we start with a spectrum of large - scale , linearly - polarized , random - phase alfvnic fluctuations which have energy equipartition between kinetic and magnetic fluctuations and vanishing correlation between the two fields .
a turbulent cascade rapidly develops , magnetic field fluctuations exhibit a kolmogorov - like power - law spectrum at large scales and a steeper spectrum at sub - ion scales .
the imposed expansion ( taking a strictly transverse ambient magnetic field ) leads to generation of an important perpendicular proton temperature anisotropy that eventually drives the mirror instability .
this instability generates large - amplitude , nonpropagating , compressible , pressure - balanced magnetic structures in a form of magnetic enhancements / humps that reduce the perpendicular temperature anisotropy . |
the cosmic microwave background ( cmb ) provides us with perhaps the clearest window on the very early universe . based just on our current understanding ,
the impact of the next generation of high resolution cmb experiments on theoretical cosmology is guaranteed to be enormous , and full implications of the new data have yet to be determined .
the subject of this paper is the distinctive `` secondary doppler peaks '' or `` sakharov oscillations@xcite '' in the angular power spectra of the microwave anisotropies calculated for inflationary models .
i discuss how these features reflect the very specific properties of `` passive perturbations '' of which the inflationary models are a subset .
the very different nature of the defect based ( or `` active '' ) models suppresses the tendency to generate sakharov oscillations , although a different mechanism _ can _ produce sakharov oscillations in certain active models .
section 2 outlines the basic ingredients of a of cmb anisotropy calculation , emphasizing the differences between active and passive models .
section 3 spells out how the sakharov oscillations appear in passive models .
section 4 describes the basic properties of active models make it hard to produce oscillations in the angular power spectrum , but also points out how some degree of oscillation is still possible . at the end of section 4 i comment of the degree to which this issue has been addressed quantitatively in the literature . concluding comments appear in section 5 .
much of this paper is based on work with my collaborators p. ferreira , j magueijo , and d. coulson , as reported in @xcite .
most models of structure formation consider perturbation which originate at an extremely early time ( eg the gut era or even the planck era ) and which have very small amplitudes ( of order @xmath1 ) until well into the matter era .
perturbations of inflationary origin start as short wavelength quantum fluctuations which evolve ( during the inflationary period ) into classical perturbations on scales of astronomical interest .
defect based models undergo a phase transition ( typically at around gut temperatures , eg @xmath2 ) forming defects which generate inhomogeneities on all scales .
for all these models , once the inflationary period and/or phase transition is over , the universe enters an epoch where all the matter components obey linear equations except for the defects ( if they are present ) .
this `` standard big bang '' epoch can be divided into three distinct periods .
the first of these is the `` tight coupling '' period where radiation and baryonic matter are tightly coupled and behave as a single perfect fluid .
when the optical depth grows sufficiently the coupling becomes imperfect and the `` damping period '' is entered .
finally there is the `` free streaming '' period , where the cmb photons only interact with the other matter via gravity . while the second and third periods can have a significant impact on the overall shape of the angular power spectrum , all the physics which produces the sakharov oscillations takes place in the tight coupling regime , which is the focus of the rest of this paper .
working in in synchronous gauge , and following the conventions and definitions in references @xcite , the fourier space perturbation equations are : @xmath3 \\ \dot{\delta}_c & = & 4 \pi g { a \over \dot a}\left(\tau_{00 } - \theta_{00}\right ) -{\dot a \over a}\left({3\over 2 } \omega_c + 2\left [ 1 + r \right ] \omega_r \right)\delta_c \nonumber \\ & & \mbox{}- { \dot a \over a}2\left[1 + r\right]\omega_r s\\ \label{eq:3}\ddot s & = & - { \dot r \over 1 + r}\dot s - c_s^2k^2\left(s + \delta_c\right ) \label{eq : pert}\end{aligned}\ ] ] here @xmath4 is the pseudo - stress tensor , @xmath5 , @xmath6 is the defect stress energy , @xmath7 is the cosmic scale factor , @xmath8 is newton s constant , @xmath9 is the density contrast and @xmath10 is the mean energy density over critical density of species @xmath11 ( @xmath12 for relativistic matter , @xmath13 for cold matter , @xmath14 for baryonic matter ) , @xmath15 , @xmath16 , @xmath17 and @xmath18 are the mean densities in baryonic and relativistic matter respectively , @xmath19 is the speed of sound and @xmath20 is the comoving wavenumber .
the dot denotes the conformal time derivative @xmath21 . in the inflationary case
there are no defects and @xmath22 . with suitable initial conditions
these linear equations completely describe the evolution of the perturbations . in the defect case @xmath23 , and certain components
are required as input ( a matter mainly of numerical convenience)@xcite .
] of @xmath24 are required as input .
cosmic defects are `` stiff '' , which means @xmath24 can be viewed as an external source for these equations .
the additional equations from which one determines @xmath24 are highly non - linear , although the solutions tend to have certain scaling properties which allow @xmath24 to be modelled using a variety of techniques ( see for example @xcite ) .
quite generically , for wavelengths larger than the hubble radius ( @xmath25 ) , eqns [ 1 - 3 ] have one decaying and one growing solution .
the growing solution reflects the gravitational instability , and , as required of any system which conserves phase space volume , there is a corresponding decaying solution .
for example , in the radiation dominated epoch , the two long wavelength solutions for the radiation perturbation @xmath26 are @xmath27 and @xmath28 . over time , @xmath29 grows compared with a comoving wavelength so in the standard big bang epoch a given mode starts with wavelength @xmath30 but eventually crosses into the @xmath31 regime . in the period of tight coupling modes with @xmath31
undergo oscillatory behavior since the radiation pressure stabilizes the fluid against gravitational collapse .
this process of first undergoing unstable behavior which eventually converts to oscillatory behavior is the key to the formation of sakharov oscillations .
i will now illustrate the process with a simple toy model .
the simplest example of growing / decaying behavior is given by the upside down harmonic oscillator ( @xmath32 ) which has the general solution @xmath33 ) .
figure [ fig : squeeze ] ( left panel ) shows the trajectories in phase space for this system .
as the `` particle '' rolls down the inverted parabolic potential both @xmath34 and @xmath35 increase arbitrarily and _ any _ starting point in phase space evolves arbitrarily close to the @xmath36 axis .
any initial region in phase space will become squeezed against this axis and elongated along it , as illustrated by the circle in the figure .
this `` squeezing '' process is generic to any equation which has one growing and one decaying solution , although the simple shape of the squeezed region is specific to models with linear equations .
the right - side - up harmonic oscillator ( @xmath37 ) serves to illustrate the oscillating regime ( when the wavelength is smaller than @xmath29 ) .
the right hand panel in fig [ fig : squeeze ] shows the phase space trajectories for the right - side - up harmonic oscillator .
the familiar oscillatory behavior describes circles in the @xmath38 plane ( in polar coordinates the angle corresponds to phase of the oscillation ) .
the linearity ensures that the period of rotation is the same on all trajectories , thus preserving the shapes of any initial region as it rotates around .
the physical system in question ( e.g. @xmath26 ) undergoes first squeezing and then oscillatory behavior . during the unstable period
the initial phase space distribution ( dictated , in the case of an inflationary scenario , by the quantum zero - point fluctuations ) is squeezed by many orders of magnitude .
the distribution which enters the oscillatory period is thus highly squeezed ( much more so than the oblong shape depicted in the right panel of fig [ fig : squeeze ] ) .
the end result it that the temporal phase of oscillation is rigorously dictated by the period of squeezing which went before .
this effect is illustrated more directly in fig [ fig : phfo ] , where different solutions for @xmath26 are shown .
representative solutions are shown from across the entire ensemble , the growing solution domination guarantees that each one goes through zero at essentially the same time .
figure [ fig : acfm2 ] shows the ensemble averaged values of @xmath39 at a fixed time as a function of wavenumber ( the power spectrum ) .
the zeros correspond to modes which have been caught at the `` zero - point '' of their oscillations .
the phase focusing across the entire ensemble guarantees that there will always be some wavenumbers where the power is zero . it is these zeros which are at the root of the oscillations in the angular power spectrum .
an important point is that these zeros are an absolutely fundamental feature in any passive theory .
no amount of tinkering with the details can counteract that fact that an extended period of liner evolution will lead to growing mode domination , which in turn fixes the phase of the oscillations in the tight coupling era .
if one were to require oscillation in a passive model to be out of phase from the prescribed value , one would imply domination by the decaying mode outside the hubble radius in other words a universe which is not at all robertson - walker on scales greater than @xmath29 .
as described in section 2 , the active case is very different from the passive case , due to the presence of what is effectively a source term in eqns [ 1 - 3 ] .
one consequence is that the whole notion of the ensemble average is changed . in the passive case any model with gaussian initial conditions
can be solved by solving eqns [ 1 - 3 ] with the initial values for all quantities given by their initial rms values .
the properties of linearly evolved gaussian distributions guarantee that this solution will always give the rms values at any time .
thus the entire ensemble is represented by one solution . in the active case
this is not in general possible .
one has to average over an ensemble of possible source histories , which is a much more involved calculation . in @xcite
we `` square '' the evolution equations to write the power spectrum as convolution of two - point functions of the sources , but there the added complexity requires the use of the full _
un_equal time correlation functions . in general
, the source term will `` drive '' the other matter components , and temporal phase coherence will be only as strong as it is within the ensemble of source terms .
an illustration of this appears in fig [ fig : acfm3 ] . in many active models
the sources are sufficiently decoherent that no oscillations appear in the power spectrum ( see for example the dashed curve in fig [ fig : acfm2 ] ) .
the source evolution is a highly non - linear process , so from the point view of a single wavenumber the source may be viewed as a `` random '' force term . at first glance it may seem impossible for such random force term could induce _ any _ temporal coherence , but here is how temporal coherence can occur : the source term only plays a significant role in eqns [ 1 - 3 ] for a _ finite _ period of time .
this is somewhat apparent in fig [ fig : acfm3 ] .
( the y - axis of fig [ fig : acfm3 ] shows a quantity specially chosen to indicate the significance of the source in eqns [ 1 - 3 ] . ) in the limit where this period of significance is short compared to the natural oscillation time of @xmath26 ( and happens at the same time across the entire ensemble ) the ensemble of source histories _ can _ be phase coherent .
the tendency for a given active model to produce oscillations in the power spectrum depends on the how `` sharply peaked '' the significance of the source term is in time .
_ note added:_it is also the case that on scales larger than @xmath29 there are `` squeezing '' mechanisms at work , even for active perturbations .
( the gravitational instability is , after all , still present . ) in @xcite a green function method is developed which clearly illustrates how the active case involves a competition between squeezing effects , which tend to produce oscillations in the power spectrum , and the randomizing effect of the nonlinear source evolution . in extreme cases , where the `` randomizing '' effects are minimized
it is even possible to have active models with mimic an inflationary signal@xcite .
so far , we are just beginning to learn the degree of coherence which is present in the most popular active models .
much of the work makes use of the `` high coherence limit '' in which a single solution to eqns [ 1 - 3 ] is meant to represent the rms value .
this only makes physical sense for the `` sharply peaked '' sources discussed in the previous subsection@xcite , but allows one to use code designed for passive perturbations with only minor changes .
the question of coherence has been most aggressively pursued in the cosmic string case@xcite , and every indication is that cosmic strings give a highly decoherent ensemble
. however , there will still be room for some degree of skepticism until the production of gravity waves and small loops is incorporated in some realistic way ( @xcite , see also the methods of@xcite ) .
most of the work on cosmic texture models has not dealt quantitatively with the question of evidence for coherence in the microwave sky although it is pretty clear that textures should have more coherence than the cosmic strings . in @xcite
the `` high coherence '' limit was used for calculating the microwave sky , but numerical simulations were used to illustrate some coherent behaviour in the tight coupling era .
in other calculations the high coherence limit has been used for convenience , and these papers have simply not claimed to treat the question of coherence .
recent simulations by turok @xcite are the sort which can in principle address the coherence question for a large number of active models , but the author is not willing to claim conclusive results until a large dynamic range is achieved .
the sakharov oscillations ( or secondary doppler peaks ) in the angular power spectrum of cmb anisotropies signify a high degree of coherence in the primordial perturbations .
these oscillations are a certain prediction of all passive models ( which includes all inflation based models ) and can not be adjusted away . as such , this prediction represents probably the most clear - cut opportunity to falsify _ all _ scenarios in which the perturbations have an inflationary origin . on the other hand the observation of substantial sakharov oscillations in the data would have an enormous impact on the active models , ruling out all be a very special subset of these .
i conclude that the search for sakharov oscillations in the cmb sky represents an opportunity to gain very deep insights into the origin of the primordial perturbations .
every effort should be made to make sure that the experiments are able to achieve conclusive results on this matter@xcite .
990.4 cm see page 280 of ya .
b. zeldovich and i. d. novikov , _ `` the structure and evolution of the universe '' _ , publ .
univ . of chicago press .
a. albrecht , d. coulson , p. ferreira , and j. magueijo , _ phys .
lett . _ * 76 * 1413 - 1416 ( 1996 ) .
j.magueijo , a. albrecht , d. coulson , p. ferreira , _ phys .
lett _ , * 76 * 2617 ( 1996 ) , j.magueijo , a. albrecht , p. ferreira , d. coulson , _ phys .
, in press , ( 1996 ) see also the contributions of p. ferreira and j. magueijo in this volume .
s. veeraraghavan and a. stebbins , _
_ * 365 * 37 - 65 ( 1990 ) .
u. pen , d.n .
spergel and n. turok , _ phys .
d _ , * 49 * 692 - 729 ( 1994 ) .
r. durrer , a. gangui and m. sakellariadou , _ phys .
_ , * 76 * 579 ( 1996 ) .
r. crittenden and n. turok , _ phys .
_ , * 75 * 2642 ( 1995 ) .
n. turok , damtp - preprint-96 - 44 , astro - ph/9604173 ( 1996 ) , n. turok , damtp - preprint-96 - 69 , astro - ph/9607109 ( 1996 ) .
w. hu and m. white , ias preprint astro - ph/9603019 ( 1996 ) w. hu .
d. spergel , and m. white ias preprint astro - ph/9605193 ( 1996 ) a. albrecht , p. ferreira , m. joyce , and t. prokopec , phys .
d * 50 * 4807 ( 1994 ) .
see also the contribution of m. hindmarsh in this volume .
a. albrecht and j. robinson , in preparation ( 1996 ) n. turok , damtp p 96 - 59 , astro - ph/9606087 ( 1996 ) see for example albrecht and wandelt imperial / tp/95 - 96/23 , astro - ph/9602040 , j. magueijo and m. hobson , damtp-96 - 13 , astro - ph/9602023 , m. hobson and j. mageijo damtp 96 - 01 , astro - ph/96030 . | 0.4 cm i discuss the origin of the `` sakharov oscillations '' ( or `` secondary doppler peaks '' ) in standard angular power spectra of the cosmic microwave background anisotropies calculated for inflationary models .
i argue that these oscillations appear because perturbations from inflation have a set of properties which makes them `` passive perturbations '' .
all passive perturbations undergo a period of linear `` squeezing '' resulting in a dramatic degree of ( classical ) phase coherence of pressure waves in the photon - baryon fluid .
this phase coherence eventually is reflected in oscillatory features in the angular power spectrum of the temperature anisotropies observed today .
perturbations from cosmic defects are `` active perturbations '' which have sharply contrasting properties .
active perturbations are highly non - linear and the degree of phase coherence in a given model reflects the interplay between competing effects . a large class of active models
have _ non_-oscillatory angular power spectra , and only an extremely exotic class has the same degree of coherence as is found in all passive models .
i discuss the significance of the search for these oscillations ( which transcends the testing of any given model ) and take a critical look at the degree to which the question of coherence has been treated so far in the literature .
imperial college , london , uk . |
the research of citation networks has drawn increasing attention and been applied to many fields@xcite .
it can help scientists find useful academic papers@xcite , help inventors find interesting patents@xcite , or help judges discover relevant past judgements@xcite .
the scientific citation networks considered in this paper are directed graphs , in which nodes represent papers , while edges stand for the citation relationships between them .
since new papers can only cite the published papers@xcite , these graphs are acyclic .
degree distribution is a fundamental research object of citation networks , and a series of models have been proposed to illustrate it .
the price model appears to be the first to discuss about cumulative advantage in the context of citation networks and their in - degree distributions@xcite .
the idea lies in that the rate at which a paper gets new citations should be proportional to the citations that it already has@xcite .
this can lead to a power - law distribution according to the price model@xcite .
a copy mechanism in which a new node attaches to a randomly selected target node as well as all its ancestors has been proposed by krapivsky et al@xcite . based on their viewpoints ,
an author may be familiar with a few primary references and may simply copy the secondary references from the primary ones .
this rule also leads to a power - law distribution .
in addition , the cumulative advantage is also known as the preferential attachment in other literatures@xcite .
jeong et al@xcite have measured the rate at which nodes acquire links on four kinds of real networks , and found that it depends on the node s degree .
their results offer direct quantitative support for the presence of preferential attachment .
moreover , an investigation has been conducted by eom et al@xcite on the microscopic mechanism for the evolution of citation networks by raising a linear preferential attachment with time - dependent initial attractiveness .
the model reproduces the tails of the in - degree distributions of citation networks and the phenomenon called burst " : the citations received by papers increase rapidly in the early years since publication .
the above - mentioned models have studied the tail of the in - degree distribution only , while the two - mechanism model proposed by george et al@xcite characterizes the properties of the overall in - degree distributions .
with respect to the research of networks from the real world ( e.g. citation networks ) , using random geometric graph has become a hot topic in recent years .
xie et al@xcitedefine the academic influence scope as a geometric area and present an influence mechanism , which means that an existing paper will be cited by a new paper located in its influence zone .
based on this mechanism , they further propose the concentric circles model ( cc model ) , which can fit the power - law tails of the in - degree distributions of the citation networks with the exponentially growing nodes .
nevertheless , the forepart of the in - degree distribution and the out - degree distribution can not be well fitted by this model . in reality , node - increment in many current citation networks enjoys a linear growth , e.g. cit - hepph , cit - hepth@xcite ( fig [ fig1 ] ) and pnas including articles published during 2000 - 2015 ( shown in our later study ) .
therefore , a model with linearly growing node - increment is proposed .
the edges in the model are still linked according to the influence mechanism , whereas they are revised in that the influence scopes of papers are determined by their topics and ages ( the time that has passed since publication ) .
different from the previous models that only focus on the tails of in - degree distributions , the improved model can well predict the overall in - degree distributions of the empirical data well . in consideration of the citations among different disciplines in real citation networks , a mechanism that is referred to as the interdisciplinary citation mechanism is proposed . under appropriate parameters
, these mechanisms can reproduce a range of properties of citation networks , including the power - law tail of the out - degree distribution , giant component and clear community structure .
meanwhile , some other properties can also be obtained like the relationship between in - degree and local clustering coefficient as well as in- and out - assortativity .
these results show that our model can be used as a medium to study the intrinsic mechanism of citation networks .
-2.25in0 in .*the first two networks extract from arxiv which cover paper from january 1993 to april 2003 ( 124 months ) in high energy physics theory and in high energy physics phenomenology@xcite . *
the last network is generated according to the generating process of the model , where parameters are @xmath0 , @xmath1 , @xmath2 , @xmath3 , @xmath4 , @xmath5 , @xmath6 , @xmath7 , @xmath8 , @xmath9 , @xmath10 . in the header of the table ,
cc , ac , ac - in , ac - out , pg and mo denote the clustering coefficient , the assortative coefficient , the in - assortative coefficient , the out - assortative coefficient , the node proportion of giant component and modularity , respectively . [ cols="<,^,^,^,^,^,^,^,^",options="header " , ] [ table1 ] the structure of this paper is as follows . the model is described in section 2 .
the degree distributions , clustering and assortativity are analyzed in section 3 to section 5 , and finally the conclusion is provided in the last section .
-2.25in0 in . * panels ( a , b ) show the trends for the papers of cit - hepth and cit - hepph . panel ( c ) shows the trend for the modeled network .
they are fitted by linear functions .
the coefficient of determination @xmath11 is used to measure the goodness of fits .
, title="fig:",width=576,height=144 ]
-2.25in0 in is bigger than other s . the edges in the model
are linked according to the influence mechanism and the interdisciplinary citation mechanism : node @xmath12 belongs to the zone of node @xmath13 , then @xmath12 cites @xmath13 ; node @xmath14 is an interdisciplinary paper " , so it could cite node @xmath15 , even though @xmath14 does not belong to the influence zone of @xmath15 .
, title="fig:",width=249,height=220 ] since many journals and databases publish papers monthly or yearly and papers in the same issue can not cite each other normally , models like the price model or the copy model that publish one paper at each time step do not consider the growing trends of papers .
xie et al@xcite pay attention to the citation networks in which the annual numbers of papers grows exponentially , such as the citation network collected by tang et al@xcite for papers ( which are published in the period from 1936 - 01 - 01 to 2013 - 09 - 29 ) in dblp dataset .
however , in some real citation networks ( e.g. cit - hepph and cit - hepth ) , the monthly or annual numbers of papers published grow linearly ( fig [ fig1 ] ) .
for purpose of the evolution and features of these networks , a geometric graph model , in which the node - increment in specific time unit experiences a linear growth , is proposed here . in our model , some spatial coordinates
are given to the nodes to represent the research contents of papers ( the differences of research contents are illustrated by the geometric distances between nodes ) .
besides , a simple spacetime , ( 2 + 1)-dimensional minkowski spacetime of two spatial dimensions , along with one temporal dimension is considered in this model , so that the time characteristics of the nodes in citation networks can be modeled .
the nodes in the model are uniformly and randomly sprinkled onto a cluster of concentric circles ( the centers of which are on the time axis ) .
in addition , the nodes on different circles are generated in different time units , while those in the same circle represent the papers published in the same issue .
the number of nodes in a circle is a linearly increasing function of the circle s temporal coordinate . in the spacetime ,
nodes are identified by their locations @xmath16 , where @xmath17 is the generated time of the node , @xmath18 is the radius of the circle born at time @xmath17 , and @xmath19 is the angular coordinate . considering that the radius @xmath18 and the time @xmath17 are @xmath20-to-@xmath20 correspondence , each node is identified by its location only with time coordinate @xmath17 and angular coordinate @xmath19 .
the edges in the model are linked according to the influence mechanism and the interdisciplinary citation mechanism , which are displayed in fig [ dia ] .
the influence zone of node @xmath13 contains node @xmath12 , and thus a directed edge is drawn from node @xmath12 to node @xmath13 under a given probability . as node @xmath14 is an
interdisciplinary paper " , it could cite node @xmath15 , even though the influence zone of node @xmath15 does not contain @xmath21 . supposing that a modeled network has @xmath22 papers ( @xmath23 ) published in the @xmath24th unit of time @xmath25 , including some interdisciplinary papers , the generating process of the model is listed as follows . 1 .
generate a new circle @xmath26 with radius @xmath27 ( @xmath28 ) centered at point @xmath29 at each time @xmath30 , sprinkle @xmath31 nodes ( papers ) on it randomly and uniformly , and fix nodes with their coordinates , e.g. node @xmath32 with @xmath33 .
+ 2 . for each node with coordinate @xmath34 ,
the influence zone ( academic influence scope ) of the node is defined as an interval of angular coordinate with center @xmath19 and arc - length @xmath35 , where @xmath36 is used to tune the exponent of power - law tail of in - degree distribution , and @xmath37 is used to make the in - degree distribution of papers published in each time unit have a power - law tail .
+ 3 . for node @xmath32 and node @xmath38 ,
the coordinates of which are @xmath33 and @xmath39 respectively , if the distance of angular coordinates @xmath40 and @xmath41 , a directed edge is drawn from @xmath38 to @xmath32 under a probability @xmath42 .
select @xmath43 percent nodes as interdisciplinary papers to continually cite a number of existing papers randomly to make the reference lengths ( out - degrees ) of those papers to be random variables drawn from a power - law distribution @xmath44 @xmath45 .
the function @xmath37 in step 2 is a staircase function of @xmath19 @xmath46$}\\ \beta_{0}+\lambda , & \textrm{$\theta\in[\theta_{1},\theta_{2}]$}\\ \vdots\\ \beta_{0}+(s-1)\lambda , & \textrm{$\theta\in[\theta_{s-1},2\pi]$ } , \end{array}\right.\ ] ] where @xmath47 , @xmath48 , @xmath49 , and @xmath50 , ... , [ \theta_{s-1},\theta_{s}]$ ] are a specific partition of @xmath51 $ ] satisfying @xmath52 , @xmath53 , @xmath54 , @xmath55 , @xmath56 , and the aging of the papers influences is ignored here due to the short time span of the empirical data ( around ten years ) ( table [ table1 ] ) . in this paper ,
the model is developed to fit cit - hepth and cit - hepph ( table 1 ) .
the evolutionary trends of the monthly numbers of papers in this two networks are sufficiently fitted by linear functions ( figs [ fig1]a , [ fig1]b ) . to make the modeled time span ( around ten years ) and the modeled size of nodes match with the empirical data
, parameters are properly selected and listed in the end of table [ table1 ] . especially , the unit of the parameter @xmath17 is set as @xmath57 month ( fig [ fig1]c ) , while the rise rate of node - increment @xmath58 and the number of circles @xmath59 are set to be @xmath60 and @xmath61 , respectively .
-2.25in0 in for the foreparts of the in- and out - degree distributions , and the power - law distribution @xmath62 for the tails of the in - degree distributions in panels ( a , b , c ) ( fitted by the method in ref ( @xcite ) ) .
the root mean squared error ( rmse ) and coefficient of determination @xmath11 are used to measure the goodness of fits.,title="fig:",width=576,height=288 ]
the out - degree distributions of the empirical data ( table [ table1 ] ) take the form of fat tails and curves in the forepart ( figs [ fig3]e,[fig3]f ) .
the curves in the forepart of the out - degree distributions can be well fitted by the generalized poisson distribution . in reality , the behavior that paper @xmath38 cites paper @xmath32 is influenced by the number of the citations@xcite and the popularity of paper @xmath63 author . at the same time
, it can be viewed as a low - probability event ( the reference length of paper @xmath38 is very small compared with the large number of papers ) .
these settings are suitable for the use of the generalized poisson distribution@xcite .
now the formulas of the forepart and tail of the out - degree distribution of the modeled network ( table [ table1 ] ) are derived to show how our model generates the similar curve and fat tail ( fig [ fig3]d ) .
the edges in the model are linked according to the influence mechanism and the interdisciplinary citation mechanism .
firstly , the non - interdisciplinary paper @xmath32 with coordinate @xmath33 is considered . for prior node @xmath38 ,
its coordinate is @xmath39 , where @xmath64 . if @xmath65 , node @xmath32 is located in the influence zone of node @xmath38 .
when @xmath66 is small enough , @xmath67 , because @xmath68 is a staircase function .
then the expected out - degree of node @xmath32 is as follows : @xmath69 which is an increasing function of the temporal coordinate @xmath70 .
when @xmath70 is large enough , @xmath71 , indicating that the reference length of the paper denoted by node @xmath32 is approximately a constant .
this is in accordance with the actual situation that the reference length of papers can not grow infinitely .
since the process of sprinkling nodes follows the poisson point process , the actual out - degree of node @xmath32 is not exactly equal to the expected out - degree .
therefore , in order to obtain the correct out - degree distribution , it is necessary to average the poisson distribution@xcite , @xmath72 which is the probability that node @xmath32 has out - degree @xmath73 , with the temporal density @xmath74 . in this model , @xmath75 so the out - degree distribution is @xmath76 it is a mixture poisson distribution similar to that of the empirical data .
moreover , the curve in the forepart of the modeled in - degree distribution can be well fitted by the generalized poisson distribution ( fig [ fig3]d ) .
the interdisciplinary papers make the tail of the modeled out - degree distribution fat ( step 4 ) ( fig [ fig3]d ) .
thus , in combination with the non interdisciplinary papers , the out - degree distribution is @xmath77 where @xmath78 denotes the proportion of interdisciplinary papers , and @xmath79 refers to the power - law distribution defined in step 4 .
the in - degree distributions of the empirical data have been investigated with the result showing that the curves in the forepart of the in - degree distributions can be well fitted by the generalized poisson distribution ( figs [ fig3]b,[fig3]c ) .
actually , the citations of one paper are affected by the new papers of its authors , and the probability of one paper receiving citations ( be selected from plenty of papers ) is small and not equal to that of other papers .
these are the conditions in which the generalized poisson distribution can be applied@xcite . besides ,
the in - degree distributions of the empirical data have a fat tail ( figs [ fig3]b,[fig3]c ) which can be interpreted as a consequence of the cumulative advantage@xcite . in this model
, this phenomenon is caused by the highly cited papers with large influence zones .
now , an expression of the forepart and tail of the modeled in - degree distribution is derived to show how our model generates the similar curve and fat tail(fig [ fig3]a ) . for the modeled paper @xmath32 , it can receive citations from the papers located inside or outside of its influence zone .
therefore , the expected in - degree of paper @xmath32 with coordinate @xmath33 is @xmath80 when @xmath70 is small , the first item in formula ( [ eq6 ] ) is larger than the second item , so @xmath81 averaging the poisson distribution , the in - degree distribution in the large in - degree region is @xmath82 where @xmath83 , @xmath84 . the laplace approximation and the stirling s approximation
are used in this approximation .
it can be proven that the integral term of @xmath85 is approximately independent of @xmath73 .
the derivation process is as follows : @xmath86 when the in - degree @xmath73 is large enough , which is satisfied by the small @xmath17 ( formula ( [ eq7 ] ) ) , the integration is approximately equal to a constant . in this way ,
the modeled in - degree distribution in the large-@xmath73 has a power - law tail with exponent @xmath87 .
when @xmath70 is large , the time derivative of the influence zone of paper @xmath32 is considered , @xmath88it means that the influence zone of paper @xmath32 in this model is approximately a constant when @xmath70 is large .
hence , it is assumed that @xmath89 ( d is a constant ) .
the expected in - degree of paper @xmath32 is @xmath90 and thus the in - degree distribution in the small in - degree region is @xmath91 it indicates that the in - degree distribution of the modeled network in the small in - degree region is a mixture poisson distribution similar to that of the empirical data .
also , the in - degree distribution of the modeled network in the small in - degree region can be well fitted by the generalized poisson distribution ( fig [ fig3]d ) .
the local clustering coefficient is equal to the probability that two vertices , both neighbors of the third vertex , will be the neighbors of one another@xcite .
also , it is found that the highly cited papers often have low local clustering coefficients in empirical data ( figs [ fig4]b,[fig4]c ) . due to the short time span of the empirical data ( only ten years ) ( table [ table1 ] ) ,
the highly cited papers get lots of citations from the new published papers with few citations . in this paper ,
the highly cited papers are considered and the formula of the relation between the highly cited papers and their local clustering coefficients is derived to show how well our model fits the tail of the local clustering coefficient .
-2.25in0 in suppose @xmath32 is a highly cited paper and @xmath70 is small enough .
paper @xmath38 and paper @xmath92 are the new published papers , which are the neighbors of paper @xmath32 . if @xmath38 has coordinate @xmath39 , a reasonable assumption is made that the overlap of the influence zones of @xmath32 and @xmath38 in circle @xmath93 ( @xmath94 is the current time ) is approximately @xmath66 because of the small @xmath70 and large @xmath95 .
particularly , if the connection probability @xmath42 equals to 1 , the probability that paper @xmath92 is the common neighbor of paper @xmath32 and paper @xmath38 is approximately equal to @xmath96 .
thus , for the general connection probability @xmath42 , the conditional probability @xmath97 . the effect of the interdisciplinary papers is ignored here owing to the low probability that paper @xmath92 is an interdisciplinary paper and connects to paper @xmath32 and paper @xmath38 simultaneously .
summing over the possible values of @xmath95 , it can be found that @xmath98 where @xmath99 denotes the number of the papers in the influence zone of paper @xmath32 at time @xmath95 .
since paper @xmath32 is a highly cited paper , the papers citing @xmath32 dominate the neighbors of @xmath32 and the effect of papers cited by @xmath32 can be ignored .
moreover , the expected in - degree of the highly cited paper @xmath32 is @xmath100 . by substituting it into eq ( [ eq13 ] ) , we get @xmath101 which is inversely proportional to the in - degree @xmath73 of paper @xmath32 .
thus , the local clustering coefficient of the highly cited paper @xmath32 in this model is also small . to show the similarity more clearly
, the range of @xmath73 is divided into equal small intervals and @xmath102 for each interval is also averaged to reduce the noises caused by random factors ( fig [ fig4 ] ) .
it can be seen that the highly cited papers tend to cite the same highly cited papers in the empirical data , which means that they are in - assortative ( table [ table1 ] ) .
moreover , it is intuitive that researchers are often wild about tracing back to hot topics .
if the topics of papers have great research value , numerous researchers will focus on them and publish a large number of papers that will cite each other . as a result , these papers become highly cited as well . the empirical data are also out - assortative ( table [ table1 ] ) , which refers to the tendency of papers to cite other papers with similar out - degrees to themselves . actually , the researchers often put emphasis on the new published papers that have novel contents .
-2.25in0 in -2.25in0 in our model also has these two properties . to show the performance of this model ,
the formulas of the scaling relations between the in- and out - degree of a node and the mean in- and out - degree of the neighbors pointing to and pointed at by the node are derived .
the relations are denoted by @xmath103 and @xmath104 , respectively . for node @xmath32 , the coordinate of which
is @xmath33 , the in - degrees of all nodes pointing to it are averaged , and found @xmath105 in this formula , some approximations are made , and they hold for small @xmath70 , meaning that formula ( [ eq15 ] ) can only fit the tail of the scaling relation .
therefore , substituting the in - degree @xmath73 of node @xmath32 born early into formula ( [ eq15 ] ) , we get @xmath106 where @xmath107 and @xmath108 are constants . formula ( [ eq16 ] ) is an increasing function , suggesting that the model is in - assortative .
in addition , when in - degree @xmath73 is large enough , @xmath109 is approximately equal to a constant ( fig [ fig5]a ) .
it is close to the actual situation that a hot topic will be out of fashion .
then @xmath110 is considered .
if node @xmath32 is not an interdisciplinary paper , we could get @xmath111 substituting the expected out - degree of node @xmath32 into formula ( [ eq17 ] ) , we get @xmath112 where @xmath113 is a constant . formula ( [ eq18 ] ) is an increasing function about out - degree @xmath73 , whereas @xmath110 will not satisfy the result given by formula ( [ eq18 ] ) ( fig [ fig6]a ) if the out - degree of node @xmath32 is large , which shows that most nodes with large out - degrees represent interdisciplinary papers ( fig [ fig3]d ) . if node @xmath32 is an interdisciplinary paper , the nodes that are pointed at by it are also interdisciplinary papers due to the out - assortativity of the model .
so @xmath114 where @xmath115 is a constant .
it indicates that the average out - degree of nodes pointed at by node @xmath32 fluctuates around a constant ( fig [ fig6]a ) .
meanwhile , the model is out - assortative , as the number of interdisciplinary papers in the model is small .
a model of scientific citation networks with linearly growing node - increment is proposed , in which the influence mechanism and the interdisciplinary citation mechanism are involved . under appropriate parameters , the formula of the modeled network s in - degree
distribution is derived , and it shows a similar behavior to the empirical data in the small in - degree region and a power - law tail in the large in - degree region .
different from most previous models that just study the forepart of the out - degree distribution of the empirical data , this model also captures the fat tails .
the model can also predict some other typical statistical features like clustering , in- and out - assortativity , giant component and clear community structure .
for example , it vividly characterizes the academic influence power of papers by geometric zones , and interprets the power - law tails of citation networks in - degree distributions by the papers inhomogeneous influence power .
therefore , it is believed that this model is a suitable geometric tool to study the citation networks .
however , some shortcomings still need to be overcome in future work : how to design a mechanism to characterize the citations of the interdisciplinary papers rather than randomly and uniformly select the existing papers ; and how to model the out - degree distribution better .
the authors would like to thank pengyuan zhang , zonglin xie and han zhang for helpful discussions and dan zhuge for proofreading this paper .
conceived and designed the experiments : ql .
performed the experiments : ql .
analyzed the data : ql zx ed . contributed reagents / materials / analysis tools : ql zx ed . wrote the paper : ql zx ed jl .
10 brooks ta ( 1986 ) evidence of complex citer motivations .
j am soc inf sci technol 37(1 ) : 34 - 36 . radicchi f , fortunato s , vespignani a , citation networks ( 2012 ) in : scharnhorst a , brner k , besselaar pvd editors . models of science dynamics .
233 - 257 .
radicchi f , castellano c ( 2015 ) understanding the scientific enterprise : citation analysis , data and modeling . in social phenomena .
springer pp .
135 - 151 .
harzing aw ( 2010 ) the publish or perish book .
melbourne : tarma software research .
abbas a , zhang l , khan s u ( 2014 ) a literature review on the state - of - the - art in patent analysis .
world patent information 37 : 3 - 13 .
martin pw ( 2007 ) introduction to basic legal citation .
legal information institute .
newman me ( 2003 ) the structure and function of complex networks .
siam review 45(2 ) : 167 - 256 .
price dj de solla ( 1965 ) networks of scientific papers .
science 149(3683 ) : 510 - 515 .
cad , c a d ( 1976 ) a generai theory of bibiiometric and other cumulative advantage processes .
j am soc inf sci technol : 293 .
peterson gj , steve press s , dill ka ( 2010 ) nonuniversal power law scaling in the probability distribution of scientific citations .
proc natl acad
sci usa 107 : 16023 - 16027 .
krapivsky pl , redner s ( 2005 ) network growth by copying .
phys rev e 71 : 036118 .
barabsi al , albert r ( 1999 ) emergence of scaling in random networks .
science 286(5439 ) : 509 - 512 .
newman me ( 2001 ) clustering and preferential attachment in growing networks .
phys rev e 64(2 ) : 025102 .
jeong h , nda z , balabsi al ( 2001 ) measuring pre - ferential attachment for evolving network .
arxiv preprint cond - mat/0104131 .
eom yh , fortunato s ( 2011 ) characterizing and modeling citation dynamics .
xie z , ouyang zz , zhang py , yi dy , kong dx ( 2015 ) modeling the citation network by network cosmology .
plos one 10 : e0120687 .
leskovec j , kleinberg j , faloutsos c ( 2007 ) graph evolution : densification and shrinking diameters .
acm tkdd 1(1 ) : 2 .
doi : 10.1145/1217299.1217301 gehrke j , ginsparg p , kleinberg j ( 2003 ) overview of the 2003 kdd cup .
sigkdd explorations 5 : 149 - 151 .
doi : 10.1145/980972.980992 .
tang j , zhang j , jin rm , yang z , cai kk , zhang l , et al .
( 2011 ) topic level expertise search over heterogeneous networks .
mach learn j 82 : 211 - 237 .
doi : 10.1007/s10994 - 010 - 5212 - 9 .
clauset a , shalizi cr , newman mej ( 2009 ) power - law distributions in emprical data .
siam rev 51 : 661 - 703 .
tuenter hj ( 2000 ) on the generalized poisson distribution .
statistica neerlandica 54(3 ) : 374 - 376
. krioukov d , kitsak m , sinkovits rs , rideout d , meyer d , bogum ( 2012 ) network cosmology .
sci rep 2 : 793 .
krapivsky pl , redner s , leyvraz f ( 2000 ) connectivity of growing random networks .
phys rev l 85(21 ) : 4629 .
dorogovtsev sn , mendes jff , samukhin an ( 2000 ) structure of growing networks with preferential linking .
phys rev l 85(21 ) : 4633 .
watts dj , strogatz sh ( 1998 ) collective dynamics of ` small - world ' networks .
nature 393(6684 ) : 440 - 442 . | due to the fact that the numbers of annually published papers have witnessed a linear growth in some citation networks , a geometric model is thus proposed to predict some statistical features of those networks , in which the academic influence scopes of the papers are denoted through specific geometric areas related to time and space . in the model , nodes ( papers )
are uniformly and randomly sprinkled onto a cluster of circles of the minkowski space whose centers are on the time axis .
edges ( citations ) are linked according to an influence mechanism which indicates that an existing paper will be cited by a new paper located in its influence zone .
considering the citations among papers in different disciplines , an interdisciplinary citation mechanism is added to the model in which some papers with a small probability of being chosen will cite some existing papers randomly and uniformly .
different from most existing models that only study the power - law tail of the in - degree distribution , this model also characterizes the overall in - degree distribution .
moreover , it presents the description of some other important statistical characteristics of real networks , such as in- and out - assortativity , giant component and clear community structure .
therefore , it is reasonable to believe that a good example is provided in the paper to study real networks by geometric graphs . |
the _ global action construction of k - theory , introduced by a.bak @xcite , associates to a ring @xmath0 an algebraic object , namely the global action @xmath1 which constitutes the algebraic analogue of the standard topological construction .
the underlying set of the global action @xmath1 consists of the points of the general linear group of @xmath0 and the action consists of the virtual triangular subgroups of the general linear group acting on the general linear group by left multiplication . _
this new approach introduced by bak has the advantage that the solutions of algebraic problems can be followed algebraically step by step .
the notion of global action gives algebraic objects such as groups , structures which allow one to develop homotopy theory similarly to the classical way , by defining paths and deformations of morphisms . in @xcite ,
the second named author developed an axiomatic homotopy theory for categories with natural cylinders , which can be applied to global actions .
recently , a. bak , r. brown , e.g. minian and t. porter generalized ideas and constructions of global actions and introduced _ groupoid atlases _ @xcite . as it was pointed out in @xcite , there were many advantages in formulating the concept of global action in terms of groupoids instead of group actions , so that it becomes part of a wider notion , namely the concept of a groupoid atlas .
this was done using the well know transition of group actions to groupoids .
a groupoid atlas can be regarded as an _ algebraic manifold _
, where the local groupoids play the role of the charts .
in this paper we define and investigate homotopy invariants of global actions and groupoid atlases .
we study the strong fundamental groupoid , the weak and strong nerves , classifying spaces and homology theory of groupoid atlases .
we also relate all these new invariants to classical constructions in topological spaces , simplicial complexes and simplicial sets and obtain this way new combinatorial formulations of classical and non classical results in terms of groupoid atlases .
the rest of the paper is organized as follows . in section [ s2 ]
we recall the basic definitions , examples and results on global actions and groupoid atlases .
nothing is very new in this section with the exception of a couple of new examples .
one of these examples appears naturally when the global action @xmath2 ( defined in @xcite ) acts on a @xmath3- set @xmath4 .
this induces a new and interesting global action . in the particular case that the general linear group acts on the quadratic forms , this construction is related to hermitian k - theory .
it is important to remark that the model for the line @xmath5 that we introduce here is not exactly the same one that is used in @xcite or @xcite .
both models are isomorphic at the weak level , but the original notion of the line becomes quite rigid when one works with morphisms that also preserve information of the local actions .
this change is essential when defining the strong fundamental group of a groupoid atlas .
we explain below the reasons of this change .
we shall also prove that both models are _ equivalent groupoid atlases .
moreover , our model is the _ regularization of the original one . _ _ in section [ s3 ] we introduce some of the fundamental concepts of this work , such as the notion of _ equivalence between maps of groupoid atlases and the notions of _ irreducible and _ regular atlases .
_ _ _ in section [ s4 ] we study the strong fundamental group of a groupoid atlas .
we use first a _ geometric approach and later we prove that it can also be defined and computed with a more algebraic approach , related to the vertex group of the colimit groupoid of the atlas .
we relate the strong fundamental groupoid with the weak one ( introduced in @xcite ) and we also show how to compute the fundamental groupoid of a topological space using groupoid atlases , via the van kampen theorem .
new formulations of classical and non classical results on the fundamental group of an open covering are also discussed .
_ the last section of the paper is devoted to simplicial methods . to each groupoid atlas
we associate a simplicial set , which we call the _ ( strong ) nerve _ for historical reasons and also because it generalizes in some sense the nerve of a groupoid .
we also introduce a weak version of the nerve ( compare with @xcite ) .
we obtain this way another definition for the fundamental group of a groupoid atlas .
this independently defined simplicial version is proved to coincide with the other two defined in the previous section .
we finish the paper with an introduction to the homology theory of groupoid atlases .
we compute some easy but clarifying examples .
the article ends with a result which relates the local homology groups with the homology of the whole atlas .
before recalling the basic definitions of _ global actions _ , let us begin with an easy example , which was already exhibited in @xcite .
let @xmath3 be a group and let @xmath6 be a family of subgroups of @xmath3 , acting on @xmath3 by left multiplication .
this will induce a global action , denoted by @xmath2 .
if @xmath7 consists of a single group @xmath8 , then this global action is simply the set of orbits of the action , but when @xmath7 consists of more than one subgroup of @xmath3 , then the different orbits of the actions interact .
we will see later that this interaction is crucial from the homotopical point of view .
for example , consider the case @xmath9 , where @xmath10 denotes the dihedral group of order @xmath11 .
take @xmath12 , @xmath13 with @xmath14 a rotation of order @xmath15 and @xmath16 with @xmath17 a symmetry .
the actions of each @xmath18 divides @xmath3 in orbits which determine the following covering of @xmath3 .
we denote by @xmath19 this covering .
its nerve @xmath20 is the simplicial complex corresponding to the following diagram @xmath21 \ar@{-}[d ] \ar@{-}[dr]\\ h_b & h_b.r & h_b.r^2\\ & h_a.s \ar@{-}[ul ] \ar@{-}[u ] \ar@{-}[ur ] } \ ] ] in section 3 we shall study the strong fundamental group of global actions . as we can guess , in this case the fundamental group will be isomorphic to @xmath22 .
a _ global action @xmath0 consists of a set @xmath23 together with a family of group actions @xmath24 on subsets @xmath25 .
these actions are related by certain morphisms which glue them together coherently .
more exactly , _
@xmath26 where 1 .
@xmath23 is a set , 2 .
@xmath27 is an index set equipped with a reflexive relation , 3 . for each @xmath28
, there is a subset @xmath29 and a _ local group @xmath30 acting on @xmath31 and _ 4 .
if @xmath32 , there exists a homomorphism @xmath33 such that @xmath34 for every @xmath35 in @xmath36 , @xmath37 .
+ moreover , @xmath38 and @xmath39 , whenever these compositions have sense .
note that @xmath40 , i.e. the @xmath41-orbits of each element in the intersection are included in the intersection .
we call @xmath0 a _ single domain global action _ if @xmath42 for each @xmath41 .
if @xmath43 , we say that @xmath0 is _
covered_. the global action @xmath44 of above is an example of a single domain global action .
in this case , @xmath45 , the local actions @xmath46 are the subgroups of the family acting by left multiplication , and any two indices @xmath41 and @xmath47 satisfy @xmath48 if and only if @xmath49 . the associated group homomorphisms @xmath50 are the inclusions
. another interesting example of a single domain global action is the _ general linear global action _
@xmath51 , where @xmath52 is an associative ring with unit .
this example was already studied in @xcite and @xcite , but we recall it here briefly , since it is one of the motivating examples for this theory .
the homotopy groups of this global action coincide with the k - theory groups of the ring @xmath52 .
let @xmath53 .
a subset @xmath41 of @xmath54 is called _ closed _ if every time that it contains the pairs @xmath55 and @xmath56 , then the pair @xmath57 is also in @xmath41 .
consider the poset @xmath58 partially ordered by inclusion .
let @xmath59 be the subgroup of @xmath51 generated by the matrices @xmath60 where @xmath61 is the matrix containing @xmath62 in the diagonal , @xmath14 in the position @xmath63 and @xmath64 elsewhere .
it is not difficult to verify that a matrix @xmath0 belongs to @xmath65 if and only if @xmath66 for @xmath67 and @xmath68 if @xmath69 . for @xmath70
, we denote @xmath71 the inclusion @xmath72 .
now let @xmath73 .
the subgroup @xmath65 acts on @xmath51 by left multiplication .
note that the general linear global action is a particular example of the actions @xmath2 introduced before .
if @xmath2 is considered as an atlas with discrete index set ( see @xcite ) , this is not longer true since in the general linear global action situation we are making some identifications between local actions .
let @xmath44 and let @xmath74 be a @xmath3-set .
the global action @xmath75 is an extension of the action of @xmath3 on @xmath4 and it is defined as follows .
take @xmath76 for each @xmath77 , with the action induced by the action of @xmath3 . as a particular case ,
consider the general lineal group @xmath78 acting over the quadratic forms in @xmath79 by base change , i.e. @xmath80 this action restricts well to the subset @xmath4 of positive defined quadratic forms .
if @xmath0 is the matrix of @xmath81 , then @xmath82 is the matrix of @xmath83 .
consider again our first example @xmath9 , where @xmath10 denotes the dihedral group of order @xmath11 , and take the element @xmath84 .
when different elements of the local groups @xmath85 and @xmath86 act consecutively on it , we obtain a path like the following @xmath87 in general , the various actions of the local groups @xmath30 interact on the intersections of the local sets @xmath31 and this induces a global dynamics in @xmath0 .
the elements of @xmath23 move along @xmath23 through the actions of the different local groups .
this idea suggests definitions for paths and loops in @xmath0 and therefore notions for path connectedness and simply connectedness .
let @xmath88 be a global action and let @xmath89 .
@xmath90-frame _ is a finite subset @xmath91 such that for each @xmath92 there exists @xmath93 with @xmath94 .
a ( weak ) _ path _ is a finite sequence @xmath95 such that for each @xmath92 the set @xmath96 is a _ local frame _ ,
i.e. an @xmath90-frame for some index @xmath41
. given two elements @xmath97 of @xmath23 , we say that they are in the same _ connected component _ of @xmath0 if there exists a path joining both points . as usual , we denoted by @xmath98 the set of ( path ) components of @xmath0 .
let us compute @xmath98 in the examples of above .
consider first the case @xmath99 .
if @xmath35 and @xmath100 are in the same component , then there is a finite sequence of matrices @xmath101 such that @xmath102 which implies that they determine the same class in the quotient @xmath103 . here
@xmath104 denotes the subgroup of elementary matrices .
since every elementary matrix can be factored as a finite products of @xmath105 , we obtain that @xmath106 for more details , see @xcite . in the case
@xmath44 , a similar argument shows that @xmath107 are in the same component if and only if there exists a sequence @xmath108 such that @xmath109 .
if we denote @xmath110 the subgroup of @xmath3 generated by all @xmath111 , then we obtain @xmath112 recall that a _ weak morphism _ @xmath113 is a set theoretic function @xmath114 which preserves local frames .
a path is a particular example of a weak morphism between global actions .
more exactly , let @xmath5 be the global action with underlying set @xmath115 , with the index set @xmath116 the family of subset of @xmath117 of the form @xmath118 and @xmath119 , and whose local actions are the free and transitive actions of the trivial group and @xmath120 respectively .
a path in @xmath0 is simply a morphism @xmath121 that _ stabilizes _ in both directions ( i.e. @xmath122 such that @xmath123 for @xmath124 ) .
if a weak morphism @xmath125 does not stabilize , we call it a _ weak curve_. a path with the same initial and final point is a _
weak loop_. as we pointed out in the introduction , the model for the line @xmath5 that we introduce here is not the same one used in @xcite or @xcite .
both models are isomorphic at the weak level , but the original notion becomes very rigid when one works with morphisms that preserve the information of the local actions .
recall that the product of global actions is defined as follows . given two global actions @xmath0 and @xmath126 , the _ product _
@xmath127 is the global action with underlying set @xmath128 and index set @xmath129 , equipped with the product relation .
the local action of @xmath127 indexed by @xmath130 is the product action between @xmath131 and @xmath132 .
the product @xmath127 , defined as above , satisfies the universal property of the categorical product in the category of global actions .
a homotopy between paths @xmath133 and @xmath134 is defined as a weak morphism @xmath135 for which there exist integers @xmath136 such that @xmath137 and @xmath138 , and that stabilizes in an appropriate sense ( see @xcite ) . in particular , the local frames of the product @xmath139 are the subsets @xmath140 satisfying @xmath141 for some @xmath142 . [ weak pi1 ]
let @xmath133 and @xmath134 be loops based on @xmath143 .
a _ homotopy _ between @xmath133 and @xmath134 is a function @xmath144 such that * for all @xmath145 , the sets @xmath146 are local frames of @xmath0 , and * there exist @xmath147 such that @xmath148 , @xmath149 and @xmath150 we can thus define the weak fundamental group of a global action @xmath23 with base point @xmath151 as follows .
the _ weak fundamental group _
@xmath152 is the set of homotopy classes of loops at @xmath35 .
the multiplication is defined via concatenation of paths .
let us compute @xmath152 in the case @xmath44 .
without loss of generality , we may suppose @xmath153 the neutral element of @xmath3 .
we denote with @xmath154 the amalgamated product of @xmath18 , i.e. the colimit in the category of groups over the diagram @xmath155 .
the weak fundamental group of @xmath2 can be computed as the kernel of the canonical map @xmath156 .
the group @xmath157 is isomorphic to the fundamental group of the nerve of the cover of @xmath3 by @xmath8-coclasses , with @xmath158 ( see below ) .
let @xmath159 and write @xmath160 for the image of @xmath161 .
let @xmath162 be the nerve of the cover of @xmath163 by @xmath160-coclasses .
the canonical group homomorphism @xmath164 induces the universal covering @xmath165 between the nerves ( see @xcite,@xcite ) .
given @xmath160 a vertex of @xmath162 , a deck transformation @xmath166 is determined by @xmath167 , which could be any element of @xmath168 , the fiber over @xmath8 .
multiplication by @xmath169 gives a simplicial map on @xmath162 for @xmath170 , hence it gives a deck transformation for @xmath171 .
we conclude that the group of deck transformations is exactly @xmath172 .
* here , by a _ covering _ of simplicial complexes , we mean a simplicial map with the unique lifting property of simplices .
observe that a simplicial map @xmath173 is a covering if and only if @xmath174 is a covering of topological spaces . at the weak level ,
a global action is the same as a set equipped with a cover .
let @xmath0 be a global action , @xmath23 the underlying set and @xmath175 the covering determined by the local orbits .
let @xmath176 be the _ vietoris complex _ of @xmath175 , whose simplices are the finite subsets of @xmath23 that are included in some element of @xmath175 .
let @xmath177 be the _ nerve _ of this covering , whose simplices are the finite subsets of @xmath175 with non trivial intersection . since local frames in @xmath0 are just simplices in @xmath178 , we have @xmath179 and @xmath180 dowker s theorem , proved in @xcite , relates both simplicial complexes @xmath181 and @xmath178 . in general , one has the following result .
[ ser debil no sirve ] the functor @xmath182 is an equivalence between the category of ( covered ) global actions with weak morphism and the category of simplicial complex .
an inverse can be obtained by giving to a simplicial complex @xmath183 a global action @xmath184 with the simplices of @xmath183 as indices and whose underlying set is the set of vertices of @xmath183 . here
@xmath185 means the group of bijections of the set @xmath4 .
there exists a stronger notion of morphisms of global actions , which originally were called regular in @xmath186 .
a _ regular morphism _
@xmath113 is a triple @xmath187 satisfying 1 .
@xmath188 is a relation preserving map , 2 .
@xmath189 is a group morphism such that if @xmath48 the diagram @xmath190 \ar[d ] & g_{\phi_f(\a ) } \ar[d]\\ g\subb \ar[r ] & g_{\phi_f(\b)}}\ ] ] commutes , 3 .
@xmath191 is a set function such that @xmath192 and 4 . for each @xmath28
, @xmath193 is a morphism of actions , i.e. @xmath194 . in order to generalize the constructions of above , we must be very careful about what a strong path or a strong homotopy is .
regular morphisms , that a priori allow us to work with curves , paths and loops in a strong sense , are very restrictive .
to give an idea , let us consider a global action such that all its local groups are finite of odd order . because a regular morphism contains group morphisms as part of its data , there are no regular morphisms @xmath121 which are not constant .
the regular maps @xmath121 do not measure , in general , the dynamics of @xmath0 .
* important note .
* in this paper , the word * regular will mean a different concept ( see [ regu ] ) . since regular morphisms of global actions ( in the sense of @xcite ) are not used in this article , there should be no confusion . * giving a group @xmath3 acting on a set @xmath4 , one can associate to the group action @xmath74 a groupoid @xmath195 .
the objects of @xmath196 are the elements of @xmath4 and for any @xmath197 , the arrows from @xmath35 to @xmath100 are the pairs @xmath198 with @xmath199 such that @xmath200 .
composition is defined in the obvious way . applying this construction to each local orbit of a global action @xmath0
, we obtain a groupoid atlas .
the concept of a groupoid atlas was introduced in @xmath201 . a _ groupoid atlas _ @xmath202 consists of a set @xmath23 , an index set @xmath203 equipped with a reflexive relation , and for each @xmath204 , a groupoid @xmath205 such that 1 .
@xmath206 , 2 . if @xmath207 in @xmath203 , @xmath36 is union of components of @xmath205 , and there is a functor @xmath208 which restricts to the identity in objects .
the structural functor @xmath50 is the identity when @xmath209 , and it commutes with the compositions .
the groupoids @xmath205 are called the _ local groupoids _ of @xmath0 .
let @xmath196 be a groupoid and let @xmath210 .
we denote by @xmath211 the full subgroupoid of @xmath196 on @xmath4 .
as we mentioned above , every global action induces a groupoid atlas .
in particular , we endow @xmath2 and @xmath51 with a structure of global action . although not every groupoid atlas comes from a global action .
for some examples , see @xcite .
any groupoid @xmath196 can be viewed as a groupoid atlas with trivial index set .
this groupoid atlas is denoted by @xmath212 .
this construction induces a fully faithful functor into any of the categories of groupoid atlases that we discuss later .
[ sc example ] let @xmath183 be a simplicial complex .
we define a groupoid atlas @xmath213 as follows : @xmath214 , the set of vertices of @xmath183 ; @xmath215 , the set of simplices of @xmath183 ordered by inclusion ; for each simplex @xmath17 , the local groupoid @xmath216 is the simply connected groupoid ( tree ) with object set @xmath17 ; the structural functors @xmath217 are the inclusions . as a particular case ,
n - sphere _ is defined to be the groupoid atlas @xmath218)$ ] , where @xmath219 $ ] is the simplicial complex with vertices @xmath220 and simplices all the nonempty proper subsets of @xmath221 .
this functorial construction @xmath222 associates to every simplicial complex @xmath183 a groupoid atlas @xmath223 satisfying some extra properties that we will discuss later .
for example , @xmath223 is irreducible , regular and infimum .
also , this construction induces a fully faithful functor into the category @xmath224 $ ] defined in the next section . this way one might view groupoid atlases as generalized simplicial complexes , for which there are many others local models than the ( homotopy trivial ) simplices .
another interesting examples of groupoid atlases arise from the fundamental groupoids of topological spaces .
[ topological example ] let @xmath4 be a topological space and @xmath19 an open cover of @xmath4 .
the groupoid atlas @xmath225 is defined as follows .
the underlying set is @xmath4 and the index set @xmath226 is the poset @xmath227 . for each @xmath228 ,
the local groupoid @xmath229 is the fundamental groupoid @xmath230 and the morphisms @xmath231 are induced by the inclusions @xmath232 . in section [ s4 ]
we relate both the weak and the strong version of the fundamental group of @xmath225 with @xmath233 . between groupoid atlases
are , as one might suppose , functions between the underlying sets which preserve local frames . in this context , a _ local frame _ is a finite subset of some connected component of a local groupoid .
this category , with groupoid atlases as objects and weak morphism as arrows , is canonically equivalent to those of global actions and simplicial complexes .
we call it the _ weak category _ of groupoid atlases .
the notion of ( strong ) morphism of groupoid atlases is not as rigid as the notion of regular morphism of global actions .
a _ morphism _ @xmath113 between groupoid atlases is a triple @xmath234 satisfying 1 .
@xmath191 is a set - theoretic function , 2 .
@xmath188 is a function which preserves the relation @xmath235 , 3 .
@xmath236 is a ( generalized ) natural transformation of groupoid diagrams over the function @xmath237 , which restricts to @xmath238 in the objects .
+ in other words , for each @xmath41 a functor @xmath239 is given in such a way that @xmath240 and , if @xmath48 , the diagram @xmath241 \ar[d ] & \g_{\phi_f(\a ) } \ar[d]\\ \g\subb \ar[r ] & \g_{\phi_f(\b)}}\ ] ] commutes .
we denote by @xmath242 the category of groupoid atlases with ( strong ) morphisms .
the atlas @xmath0 is _ covered _ if every element of @xmath23 is an object of some local groupoid .
all the groupoid atlases that we consider are assumed to be covered .
note that , if @xmath0 is covered and @xmath113 is any morphism , then the function @xmath238 is locally determined by the values of the functors @xmath243 in objects .
however , @xmath238 must be part of the data . to illustrate this ,
consider the following example .
suppose @xmath0 is the groupoid atlas with @xmath244 , @xmath245 with the discrete order and @xmath246 and let @xmath126 be any groupoid atlas
. suppose also that we are given a map @xmath188 and a generalized natural transformation @xmath247 .
note that the map @xmath188 picks two indices @xmath248 and @xmath249 of @xmath250 , and @xmath247 determines two objects @xmath35 and @xmath100 of @xmath251 and @xmath252 , respectively .
if we take @xmath253 , then @xmath254 does not come from a morphism of atlases .
under certain condition on the atlas @xmath0 , a relation preserving map @xmath188 and a natural family of functors @xmath255 do determine a map @xmath256 .
we introduce the concept of a _ good atlas , which solves this problem and plays a fundamental role in the next sections .
this concept is weaker than ( but intimately related with ) the notion of _ infimum _ , introduced in @xmath186 and @xmath201 ( cf .
[ infimum ] ) .
_ [ def of good ] let @xmath0 be a groupoid atlas .
we say that @xmath0 is _ good _ if for every @xmath143 the set @xmath257 has an initial element , i.e. there exists @xmath258 such that @xmath259 for all @xmath260 .
suppose @xmath0 is a good atlas .
given a relation preserving map @xmath261 and @xmath247 a natural family of functors over @xmath237 , the various local functions @xmath262 agree in the intersections because @xmath263 must be @xmath264 for all @xmath41 , and therefore there is a well defined function @xmath191 .
in this case , a morphism @xmath113 can be regarded as a pair @xmath254 .
in order to define the right notion of strong fundamental group we need to change first a little bit our notion of morphism .
many maps @xmath256 seem to carry the same information , and they just differ in the indices . for example , if @xmath0 is such that @xmath265 with @xmath266 a subgroupoid of @xmath267 and @xmath268 is the inclusion , this atlas does not contain more information than the one coming from @xmath267 .
moreover , given @xmath151 , it is necessary to identify the various identities of this element viewed as an object of the different local groupoids . also , we should understand @xmath269 as an _ extension of the _ movement @xmath169 . _ _
this induces the notion of equivalence between morphisms .
given @xmath270 , we say that @xmath271 is a _ corestriction _ of @xmath272 if @xmath273 for all @xmath28 and all the diagrams @xmath274_{\g_f(\a ) } \ar[dr]^{\g_{f'}(\a ) } \\
\g_{\phi_f(\a ) } \ar[rr]_{\phi_{\phi_f(\a)}^{\phi_{f'}(\a ) } } & & \g_{\phi_{f'}(\a)}}\ ] ] are commutative .
we write @xmath275 .
the morphisms @xmath270 are _ equivalent _ if they are in the same class of the equivalence relation generated by the corestrictions .
we write @xmath276 .
it is easy to see that @xmath276 implies @xmath277 .
note that the relation @xmath278 is transitive .
moreover , if @xmath275 , then @xmath279 and @xmath280 for any @xmath281 .
it follows that , if @xmath270 and @xmath282 are such that @xmath276 and @xmath283 , then @xmath284 .
therefore , we can define a new category @xmath224 $ ] , with groupoid atlases as objects and the classes of ( strong ) morphisms as maps .
we denote the class of @xmath271 by @xmath285 $ ] .
the notion of equivalence between morphisms loosen the dependence of @xmath0 from the index set @xmath203 .
let us go back to the example at the beginning of this section .
the information of the atlas @xmath0 seems to lie just in @xmath267 . if we call @xmath286 the atlas with this only local groupoid , the canonical arrows @xmath113 and @xmath287 satisfy @xmath288 and @xmath289 , so they are isomorphic in @xmath224 $ ] .
we say that @xmath113 is an _ equivalence _ if there exists @xmath287 such that @xmath288 and @xmath290 , i.e. @xmath285 $ ] is an isomorphism in @xmath224 $ ] .
given atlases @xmath0 and @xmath126 , we say that they are _ equivalent _ if there is an equivalence between them .
recall that the set @xmath203 of a groupoid atlas @xmath291 is equipped with a reflexive relation . if @xmath292 are such that @xmath48 , by definition @xmath36 can be written as union of connected components of the groupoid @xmath205 .
if in addition @xmath205 is connected , it follows that @xmath293 and the structural functor @xmath208 can be regarded simply as a functor between @xmath205 and @xmath294 .
so , when each local groupoid is connected , a groupoid atlas is nothing but a particular diagram of groupoids .
we will say that a groupoid atlas @xmath0 is _ irreducible _ if @xmath205 is a connected groupoid for every index @xmath41 in @xmath203
. given any atlas @xmath0 , there is a natural way to associate to it an irreducible atlas @xmath295 , considering each component of any local groupoid as an individual local groupoid .
the underlying set of @xmath295 is the same of @xmath0 , so @xmath296 .
the index set of @xmath295 must repeat each index of @xmath0 as many times as many components of the corresponding groupoid .
explicitly , @xmath297 and the relation is induced from @xmath203 : @xmath298 finally , the local groupoid @xmath299 is the component of @xmath205 with object set @xmath4 , and if @xmath300 , the morphism @xmath301 equals @xmath50 .
like in any irreducible atlas , the structural maps of @xmath295 are defined over the whole local groupoid .
this construction is functorial , since every map @xmath302 induces a new map @xmath303 given by @xmath304 , @xmath305 where @xmath306 is the component of @xmath307 that contains @xmath308 , and @xmath309 is the restriction of the functor @xmath239 .
the canonical map @xmath310 that forgets the second coordinate of the indices ( @xmath311 ) and such that @xmath312 is the inclusion of each component , is natural in @xmath0 .
it satisfies also the following universal property , whose proof is straightforward .
let @xmath287 a map of groupoid atlases , and suppose that @xmath126 is irreducible .
then , there exists a unique map @xmath313 such that @xmath314 .
it is clear that @xmath315 when the atlas @xmath0 is already irreducible .
the functor @xmath92 is a right adjoint for the inclusion of the full subcategory of irreducible atlases into the category of atlases .
the map @xmath316 is a weak isomorphism , but it is not an isomorphism or an equivalence in general .
however , the groupoid atlases @xmath0 and @xmath317 are intimately related : they share all the invariants that we will study , such as the nerve , the homology groups and even the fundamental group .
given an irreducible atlas @xmath0 and indices @xmath318 and @xmath319 in @xmath203 such that @xmath320 , then @xmath321 and the composition @xmath322 is defined and it agrees with @xmath323 if @xmath324 in @xmath226 .
therefore , in an irreducible atlas we can assume that the relation @xmath325 is transitive without loss of generality .
[ up to equivalence ] given an irreducible atlas @xmath0 , the relation @xmath325 defined in its set of indices @xmath203 is a partial order `` up to equivalence '' , i.e. there exists an atlas @xmath126 whose set of indices is partially ordered by @xmath325 , and an equivalence @xmath256 .
recall that the relation @xmath325 defined in @xmath203 is always reflexive , and since @xmath0 is irreducible , it can also be taken transitive .
it only remains to make @xmath325 antisymmetric .
the atlas @xmath126 will be obtained from @xmath0 by deleting some local groupoids .
we say that two indices @xmath292 are _ paired _ if @xmath48 and @xmath326 .
clearly , been paired is an equivalence relation .
we denote by @xmath327 $ ] the paired class of @xmath41 .
if @xmath41 and @xmath47 are paired indices , then @xmath328 and the functor @xmath50 must be an isomorphism .
let @xmath329 be a selector function that assigns to each paired class an element of itself .
we define @xmath126 as follows , 1 .
@xmath330 , 2 .
@xmath331\ |\ \a\in\phi_a}$ ] , with the relation induced by the inclusion @xmath332 , 3 .
@xmath333 is the restriction of @xmath334 to @xmath250 .
the map @xmath335 is the canonical inclusion .
the inverse , @xmath287 is given by 1 .
@xmath336 , 2 .
@xmath337 is the map @xmath338 $ ] . if @xmath48 , because @xmath41 and @xmath339 $ ] are paired , @xmath47 and @xmath340 $ ] are paired and @xmath325 is transitive , we have @xmath339\ls c[\b]$ ] , 3 .
the functor @xmath341}$ ] equals @xmath342}$ ] the structural functor of @xmath0 .
the square @xmath343^{\phi\suba^{c[\a ] } } \ar[d]_{\phiab } & \g_{c[\a ] } \ar[d]^{\phi_{c[\a]}^{c[\b ] } } \\ \g\subb \ar[r]_{\phi\subb^{c[\b ] } } & \g_{c[\b ] } } \ ] ] is commutative because of the naturality of @xmath334 .
the composition @xmath344 is the identity of @xmath126 , and @xmath345 is a corestriction of @xmath346 .
it follows that @xmath0 and @xmath126 are equivalent .
the next definition , which is motivated by the last proposition , gives another approach to irreducible groupoid atlases .
the proof of the equivalence between both definitions is omitted .
[ alternative definition ] given a partially ordered set @xmath226 , an _ irreducible groupoid atlas _
@xmath0 with index set @xmath226 is a diagram in the category of groupoids @xmath347 such that for each @xmath41 the groupoid @xmath348 is connected and if @xmath48 then the induced functor @xmath50 , is an inclusion on the objects . with this definition ,
the underlying set @xmath23 is the union @xmath349 .
a morphism @xmath113 is a pair @xmath254 , with @xmath261 a map of posets and @xmath350 a natural transformation between @xmath334 and @xmath351 that induces a function @xmath238 .
@xmath352^{\phi_f } \ar[dr]_{\g_a } & & \phi_b \ar[dl]^{\g_b } \\ & \gpd}\ ] ] in this context , the corestrictions arise naturally .
given @xmath270 , the relation @xmath275 is equivalent to the existence of a natural transformation @xmath353 such that @xmath354 equals the composition @xmath355 , @xmath356^{\phi_{f ' } } \ar@/_.6pc/[rr]_{\phi_f } \ar[dr]_{\g_a } & \uparrow & \phi_b \ar[dl]^{\g_b } \\ & \gpd}\ ] ] we may think a regular atlas as a _
well pointed _ atlas .
our model for the line is regular , and this is very convenient because of the existence of sections for the projection @xmath357 ( they are necessary for the homotopies ) .
[ regu ] an atlas @xmath0 is called _ regular _ if it is good and @xmath358 for every @xmath143 , where @xmath359 is the minimum of @xmath360 ( cf .
[ def of good ] ) . given @xmath143 ,
the index @xmath359 in the definition of regular atlas is not unique , but two of such indices must be mutually related by @xmath325 , and therefore the uniqueness can be assumed up to equivalence ( proposition [ up to equivalence ] ) when @xmath0 is irreducible .
there is a regularization functor @xmath361 defined as follows .
@xmath362 and @xmath363 the local groupoids @xmath364 are singletons and the @xmath365 are as before . the line @xmath5 is a regular groupoid atlas . as we pointed out above , it is the regularization of the one used in @xcite and @xcite . sometimes one need to replace a groupoid atlas @xmath0 by its regularization @xmath366 , in order to work with paths , homotopies and morphisms in a right way .
the following proposition asserts that , when @xmath0 is good , this replacement does not change the equivalence class of the atlas .
then , every good atlas can be supposed to be regular up to equivalence .
[ a sim ra ] let @xmath0 be a good atlas .
the canonical inclusion @xmath367 is an equivalence .
let @xmath271 be the inclusion @xmath367 , we define @xmath368 by sending @xmath369 to the initial element of @xmath360 .
it is easy to see that @xmath169 is a morphism , @xmath370 and @xmath371 ( in fact , @xmath372 ) .
we will say that a morphism @xmath113 between regular atlases is _ regular _ if @xmath373 for all @xmath41 .
the following three results will be used to characterize the ( strong ) fundamental group of groupoid atlases .
let @xmath113 be a morphism between regular atlases .
then there exists @xmath374 regular such that @xmath276 .
let us define @xmath272 and prove that it is equivalent to @xmath271 .
take @xmath375 equals @xmath238 . for each @xmath41
there is at least one @xmath143 such that @xmath376 is the initial element of @xmath360 .
define @xmath377 as the function that sends the initial element of @xmath360 to the initial element of @xmath378 , and the other indices @xmath41 to @xmath379 .
the family of functors @xmath380 is defined as follows : if there is an @xmath35 such that @xmath376 , @xmath381 is uniquely determinated by @xmath238 ; otherwise , @xmath381 is just @xmath382 .
naturality follows from the naturality of @xmath350 , and since @xmath383 the result follows .
[ r adj inc ] let @xmath113 .
if @xmath126 is good , then there exists @xmath384 that extends @xmath271 .
moreover , if @xmath385 is another extension , then @xmath283 .
it is sufficient to extend @xmath237 to the indices @xmath386 of @xmath387 that are not in @xmath203 , and this can be done sending each @xmath359 to the minimum of @xmath388 .
the morphism @xmath169 obtained this way is a corestriction of any other extension .
when @xmath0 is good and @xmath389 , the map @xmath368 is exactly the one constructed in [ a sim ra ] .
proposition [ r adj inc ] proves that the regularization functor is right adjoint to the inclusion of the full subcategory of @xmath224 $ ] formed by the good atlases .
[ regular proposition ] let @xmath390 with @xmath0 regular and @xmath126 good . if @xmath391 and @xmath392 for all @xmath28 except perhaps for the indices @xmath393 , then @xmath394
. follows from the previous result .
along this section , we work with good groupoid atlases . from the notion of equivalence
, we will obtain a definition for points and arrows of an atlas @xmath0 in a global sense , i.e. independent of the local groupoids .
we denote with @xmath395 the singleton groupoid ( a single object and a single arrow ) , and with @xmath396 the 2-point simply connected groupoid .
recall that @xmath397 is the functor that maps each groupoid @xmath196 into the atlas whose unique local groupoid is @xmath196 .
a morphism @xmath398 is a pair @xmath399 , where @xmath400 and @xmath401 are the images of the unique index and the unique object respectively .
a same element @xmath143 gives rise to many morphisms @xmath398 , one for each @xmath41 such that @xmath401 . on the other hand ,
if @xmath48 , the two corresponding morphisms are equivalent .
a _ global point _ in @xmath0 is the equivalence class of an arrow @xmath398 .
@xmath402}(a(*),a)\ ] ] the canonical function @xmath403 that maps the class of @xmath404 in @xmath405 is bijective when the atlas @xmath0 is good . in fact , it is surjective because @xmath0 is covered and injective because each @xmath360 has initial element .
as we pointed out before , we think @xmath406 as an extension of the _ movement _ @xmath407 .
let us consider the case of @xmath2 , with the family @xmath7 closed under ( finite ) intersections .
the arrows of a local groupoid @xmath408 come from the action of the subgroup @xmath8 in @xmath3 by left multiplication . recall that @xmath409 is the arrow with source @xmath169 and target @xmath410 .
these arrows are identified with the corresponding elements of @xmath3 , without considering which @xmath8 does actually contain them : if @xmath411 belongs to @xmath8 and @xmath412 , since @xmath7 is closed under intersections we might think @xmath413 and @xmath414 as extensions of @xmath415 .
thus , the arrows of @xmath0 are @xmath416 for a general atlas @xmath0 , we propose the following definition . [ def de global arrow ] a _ global arrow _
$ ] of @xmath0 is the class of some local arrow @xmath407 by the relation generated by @xmath418 .
equivalently , a global arrow is the equivalence class of a morphism @xmath419 .
@xmath420}(a(\i),a)\ ] ] since @xmath283 implies @xmath421 , there are well defined _ source _ and _ target _ of a global arrow @xmath417 $ ] .
they are @xmath422 and @xmath423 , respectively .
let @xmath0 be a good atlas , and @xmath143 .
if @xmath424 , then @xmath425 because we can reach one from the other by functors @xmath71 .
we denote this global arrow as @xmath426 .
observe that @xmath417=id_x$ ] does not imply that @xmath169 is the identity of @xmath35 in some local groupoid , because the structural functors @xmath50 may not be faithful . finally , recall that the _ standard @xmath427-simplex _ @xmath428 $ ] in the category of groupoids is the simply connected groupoid with object set @xmath220 .
in particular , we have @xmath429 $ ] and @xmath430 $ ] .
these classical definitions give us models for the simplices in @xmath242 , composing with @xmath431 . as a generalization of what we did with points and arrows ,
they will be used in section [ s5 ] to define the nerve of a groupoid atlas . in this subsection
we will interpret a curve @xmath121 as a sequence of global arrows .
this leads to a very nice result : the fundamental group of a groupoid atlas equals the fundamental group of the colimit groupoid over the set of indices . in the category
@xmath242 , we consider the inclusions @xmath432 given by @xmath433 and @xmath434 a morphism @xmath435 gives rise to a sequence @xmath436 defining @xmath437 , @xmath438
. given @xmath435 , in the triple @xmath439 the function @xmath440 is determinated by the others , and any natural family of functors @xmath441 induces a map @xmath435 since @xmath5 is regular .
moreover , the local groupoids of @xmath5 of the form @xmath442 are singletons so each functor @xmath443 must be trivial , and each functor @xmath444 is given by an arrow in @xmath445(the image of the arrow @xmath446 ) .
thus , a morphism @xmath447 determines and is determined by the following data .
* @xmath448 where @xmath449 , and * @xmath450 , where @xmath451 satisfies @xmath452 and @xmath453 .
a map @xmath121 gives rise to a curve with a _ framing _ @xmath47 in the vocabulary of @xcite .
here , @xmath47 is the function @xmath454 .
conversely , one can get a map @xmath121 from a weak curve @xmath455 with a framing @xmath47 , specifying local arrows @xmath456 which make @xmath455 a curve .
a _ curve _ @xmath457 in @xmath0 is an equivalence class of a morphism @xmath435 .
if @xmath450 is a sequence of local arrows such that @xmath452 , we can construct a morphism @xmath435 taking for each @xmath427 an index @xmath458 , which exists when @xmath0 is good . here
@xmath459 is the index of the groupoid that contains the local arrow @xmath460 .
since @xmath5 is good , by proposition [ regular proposition ] , the class of this @xmath447 does not depend on the choice of the @xmath461 .
hence , we have established the following ( partial ) correspondence : @xmath462\text { curve}\\ ( [ \l_n])_n\subset \aa & \leftarrow \l : l\to a\end{aligned}\ ] ] if @xmath463 then @xmath464=[\l]\circ[i_n]=[\l']\circ[i_n]=[\l'_n]$ ] for all @xmath427 .
we investigate now under which hypothesis the converse is also true , namely : @xmath465=[\l']$ ] .
let @xmath466 be such that @xmath467 and @xmath468 .
then , @xmath463 . for each @xmath427 ,
let @xmath469 be less or equal than @xmath470 and @xmath471 .
let @xmath472 be given for the sequences @xmath473 and @xmath450 .
it is easy to check that @xmath474 and @xmath475 .
let @xmath466 .
suppose that there exists a finite subset @xmath476 with @xmath477 for @xmath478 and @xmath479 for @xmath480 , then @xmath463 .
it suffices to construct a finite sequence @xmath481 such that @xmath482 and @xmath483 are in the conditions of the proposition .
we say that @xmath435 _ stabilizes _ to the left ( resp . to the right ) if there are @xmath143 and @xmath28 such that @xmath484 for small ( resp .
big ) enough values of @xmath427 .
[ equiv entre curvas ] if @xmath466 stabilize in both directions and @xmath485 for every @xmath427 , then @xmath463 .
let @xmath486 be a stabilization pair for @xmath447 and @xmath487 .
we have @xmath488 for @xmath489 and some @xmath490 in @xmath203 , and @xmath491 for @xmath492 and some @xmath493 .
the map @xmath447 gives an infinit sequence @xmath494 that not necessary stablize in any direction .
however , we may suppose that @xmath495 for @xmath489 and @xmath496 for @xmath492 without change the equivalence class of @xmath447 ( actually , @xmath447 is a corestriction of a map satisfying this ) .
similarly , let @xmath497 and @xmath498 be the indices where @xmath487 stabilizes .
choose @xmath499 in @xmath203 such that @xmath500 , @xmath501 .
define @xmath502 as @xmath503 define @xmath504 by @xmath505 denote @xmath506 the morphism induced by the sequences @xmath507 and @xmath508 . clearly , @xmath509 .
analogously , define @xmath510 such that @xmath511 . by the propostion of above and since @xmath512 for @xmath513 or @xmath514 and @xmath515 for @xmath516 , it follows that @xmath517 . therefore , @xmath518 .
if @xmath457 is a curve , we say that @xmath457 _ stabilizes _ to the left ( resp . to the right ) if there exists @xmath435 such that @xmath519=l$ ] and @xmath447 stabilizes to the left ( resp . to the right ) .
a _ path _ in @xmath0 is a curve which stabilizes in both directions .
note that if @xmath457 is a path in @xmath0 , then there exists @xmath435 , @xmath520 , @xmath521 and @xmath522 such that @xmath523 $ ] , @xmath524 for @xmath525 and @xmath526 for @xmath527 .
consider again the correspondence between curves and sequences of global arrows .
a path @xmath523 $ ] is associated to the sequence @xmath528)_n\subset \aa$ ] which stabilizes in identities at both sides .
conversely , given a sequence @xmath529 such that @xmath530 , @xmath531 for @xmath513 and @xmath532 for @xmath514 , one can take local arrows @xmath450 such that @xmath464=g_n$ ] and construct the curve associated to this sequence . note that this curve does not have to be a path .
this can be solved choosing @xmath533 such that @xmath534 , @xmath535 and taking @xmath488 and @xmath491 for @xmath427 small or big enough .
this new curve is a path , and is uniquely determinated by proposition [ equiv entre curvas ] .
thus we have obtained the following result .
the constructions of above are mutually inverse .
they define a bijection between the set of paths of an atlas @xmath0 and the sequences @xmath536 that stablizes on identities and satisfies @xmath530 for all @xmath427 .
this correspondence preserves sources and targets , and concatenations of paths up to homotopy .
this leads to a simplicial computation of the fundamental group .
in fact , we will prove in the next section that the fundamental group of an atlas @xmath0 is the fundamental group of its _ nerve _ @xmath537 . in this subsection
we introduce the ( strong ) fundamental group of a good groupoid atlas @xmath0 with a fixed base point .
later we will generalize this construction to any groupoid atlas .
the definitions and results introduced here admit formulations in terms of the fundamental groupoid .
recall that @xmath5 is regular .
its points can be represented by the inclusions @xmath538 the projections @xmath539 admit sections @xmath540 respectively .
we will simply write @xmath541 instead of @xmath542 , the inclusion of @xmath5 in the @xmath427-th row
. a _ homotopy _ between two paths @xmath457 and @xmath543 is a map @xmath135 such that there exist integers @xmath544 and @xmath545 satisfying * @xmath546 with @xmath519=l$ ] , * @xmath547 with @xmath548=l'$ ] , * there are @xmath522 such that for all local arrow @xmath271 of @xmath139 , @xmath549 if the first coordinate of @xmath550 is less than @xmath551 and @xmath552 if the first coordinate of @xmath553 is greater than @xmath545 . it is clear that homotopy of paths is an equivalence relation .
[ def de pi1 ] let @xmath0 be a good atlas and @xmath143 .
the fundamental group @xmath554 is the set of homotopy classes of loops at @xmath35 , with the operation induced by the concatenation of paths .
[ pi1 y sim ] if @xmath555 is an equivalence , then @xmath556 .
it is a consequence of the definition of paths and homotopies .
the fundamental group is actually a well defined invariant in the category @xmath224 $ ] . for an alternative proof ,
see section [ s5 ] .
* important remark : * we extend the definition of the fundamental group to non necessarily good atlases , taking the fundamental group of their regularizations . by the last proposition ,
this new definition agree with the original one on good atlases .
a homotopy @xmath557 can be decomposed into more elementary homotopies , considering the various ways to reach the upper - right point from the lower - left one in a lattice ( say @xmath558\times[0,n]\subset x_{l\times l}$ ] ) , that gives rise to a sequence of paths @xmath559 where two consecutive paths just differ in a single square .
note that all these paths become loops when @xmath8 is applied .
these elementary homotopies relate a path @xmath560,\dots,[g_{m_1}])$ ] , viewed as a sequence of global arrows , with another @xmath561,\dots,[h_{m_1}])$ ] , where @xmath562 for some @xmath92 and some @xmath41 , and @xmath563 for @xmath564 . inserting the loop @xmath560,\dots,[g_i\circ g_{i-1}],\dots,[g_{m_1}])$ ] between them
, we can observe that the homotopies in @xmath0 define the same equivalence relation that the one used to represent the morphisms in a colimit on the category of categories ( cf .
@xcite,@xcite ) .
this proves the following result .
[ fundamental group ] let @xmath0 be an irreducible groupoid atlas and let @xmath143 . consider @xmath565 the colimit over the diagram @xmath203 .
then the fundamental group @xmath554 equals the fundamental group of the groupoid @xmath566 .
the hypotesis of irreducible on @xmath0 is necessary to make @xmath567 a groupoid diagram . otherwise , the functors @xmath50 are partially defined and the colimit does not have sense .
last theorem shows another way to define the group @xmath568 when @xmath0 is irreducible .
when @xmath0 is not irreducible , the group @xmath569 defined in this way is isomorphic to the fundamental group defined as before .
this approach will be study in section [ s5 ] . [ topological example 2 ] we return to example [ topological example ] . by definition , the local groupoid @xmath570 is connected when the open subset @xmath571 is path connected .
so , the atlas @xmath225 is irreducible if and only if @xmath571 is path connected for all @xmath228 .
if in addition the family @xmath19 is closed under finite intersections , then @xmath572 as a consequence of the last theorem and ( the groupoid formulation ) of van kampen s theorem ( cf .
@xcite , @xcite ) .
the fundamental group @xmath554 appears as a natural invariant for groupoid atlases in many ways : from paths and homotopies of paths ( [ def de pi1 ] ) , by simplicial calculation ( [ pi1 de na ] ) and even as a purely algebraic invariant ( [ fundamental group ] , [ colimit groupoid ] ) . in some special cases
coincides with the weaker one @xmath152 .
the group @xmath152 describes some of the algebra of the global nontrivial loops , while @xmath554 detects also the local ones .
it is clear that these groups are not equal in general .
the following examples are the simplest cases where the differences between these two groups are noticed . in a proposition of below we
will prove that these are essentially the unique kinds of differences they could have .
given a groupoid atlas @xmath0 and @xmath143 , recall that the group @xmath152 is the fundamental group of the simplicial complex @xmath178 .
we denote this complex by @xmath573 .
let @xmath196 be a groupoid and let @xmath35 be an object of @xmath196 .
the groups @xmath574 and @xmath575 depends only of the component of @xmath196 that contains @xmath35 , so @xmath196 can be supposed connected . the vietoris complex @xmath576 is the simplex spanned by @xmath577 ,
so it is contractible .
hence , @xmath578 . on the other hand , @xmath579 , which is not trivial in general .
let @xmath0 be a groupoid atlas with discrete set of indices @xmath580 and such that @xmath581 and the local groupoids @xmath205 and @xmath294 are simply connected .
every finite subset of @xmath23 is a local frame , so as in the last example , @xmath573 is contractible and @xmath582 for any @xmath35 .
let @xmath583 be the unique arrow in @xmath584 with source @xmath100 and target @xmath585 .
the ( strong ) loop whose associated sequence of global arrows is @xmath586,[y\xto{\b}x])\ ] ] is a non trivial element of @xmath554 for any @xmath587 , @xmath588 ( in fact , @xmath554 is the free group with one generator @xmath589,[y\xto{\b}x])$ ] for each @xmath100 ) .
therefore , the groups @xmath590 and @xmath591 are distinct if @xmath0 has other points than @xmath35 .
there is a canonical group morphism @xmath592 which sends a loop @xmath457 to the weak loop @xmath440 , where @xmath447 is a representative of @xmath457 that stabilizes in both directions .
the map @xmath593 can also be defined using the simplicial information : mapping a path of global arrows @xmath594 to the path @xmath595 of edges of @xmath573 .
this map is defined up to homotopy .
note that @xmath593 is the group morphism induced by the simplicial map @xmath593 of [ p en nervios ] .
the last examples show that @xmath593 is not an isomorphism in general , but it is so in many interesting cases .
consider for example the groupoid atlas @xmath44 , with @xmath3 a group and @xmath7 a family of subgroups closed under finite intersections .
let @xmath35 be an element of @xmath3 .
in order to compute the group @xmath554 , note that @xmath596 , as one can easily check from definition [ def de global arrow ] and the paragraph above it . then , we have the isomorphism @xmath597 since ( the groupoid form of ) @xmath593 maps bijectively global arrows into edges and elementary homotopy triangles into three - elements local frames .
this result is not true if the family @xmath7 is not closed under intersections .
the general linear global action is a particular case of an @xmath2 where @xmath7 is closed under intersections . to see that , recall that for @xmath99 the index set @xmath203 consists of the closed subsets of @xmath598 partially ordered by inclusion .
if @xmath41 and @xmath47 are in @xmath203 , it is clear that @xmath599 is also closed , and hence it is in @xmath203 . the subgroup @xmath600 of the linear group equals @xmath601 .
therefore , the natural map between both fundamental groups of @xmath51 is an isomorphism .
the homotopy of @xmath2 is locally trivial , in the sense that the local groupoids of @xmath0 are simply connected groupoids .
this leads us to a more general family of atlases in which the map relating both groups @xmath152 and @xmath554 is an isomorphism .
[ infimum ] recall the definition of infimum from @xcite , @xcite .
a groupoid atlas @xmath0 is _ infimum _ if for every local frame @xmath17 the set @xmath602 has an initial element .
note that an infimum groupoid atlas is , in particular , a good atlas .
this condition , when all local groupoids are simply connected , implies that both groups @xmath554 and @xmath152 are isomorphic by @xmath593 . [ pi1a equals pi1wa ] if @xmath0 is an infimum groupoid atlas such that every local groupoid @xmath205 of @xmath0 is simply connected , then the canonical map @xmath603 is an isomorphism .
given a two elements local frame @xmath604 , since @xmath605 has an initial element , there is a unique global arrow with source @xmath100 and target @xmath585 .
then , @xmath593 maps bijectively the paths of global arrows into the paths of edges .
the 2-simplices of @xmath573 arise from the three elements local frames . under this bijection , any three edges which are the faces of a 2-simplex in @xmath573 correspond to a homotopy triangle .
we propose a stronger version of the definition of infimum atlases ( the definition of above is the original introduced in @xcite ) that seems to fit better from the strong point of view .
we might say that @xmath0 is _ infimum _ if @xmath606 has an initial element for every simplex @xmath17 of @xmath537 .
when @xmath0 is such that all its local groupoids are simply connected , the original and the strong definition of infimum agree , since , in this case , a simplex is essentially determined by its underlying local frame .
theorem [ pi1a equals pi1wa ] remains true when all the local groupoids are simply connected and @xmath0 satisfies the following condition : the sets @xmath606 are filtered for all local frame @xmath17 .
note that this condition is weaker than the infimum condition , but it is sufficient to prove the result .
we finish this section considering the atlas @xmath607 , where @xmath4 is a topological space and @xmath19 is an open cover of @xmath4 .
the atlas @xmath0 is infimum if @xmath19 is closed under intersections , and we have seen in [ topological example 2 ] that @xmath0 is irreducible if @xmath571 is path connected for all @xmath228 .
when these conditions hold , the strong fundamental group of @xmath0 equals @xmath233 ( cf .
[ fundamental group ] ) and the weak one is , by dowker s theorem , the fundamental group of the nerve of the cover . by last proposition ,
these groups are isomorphic when each @xmath571 is simply connected . by the remark of above , it is sufficient for @xmath19 to be closed under finite intersections .
thus , we obtain as a corollary of [ pi1a equals pi1wa ] an alternative proof of the following result . let @xmath4 be a topological space and @xmath19 an open cover by simply connected open subsets .
if @xmath19 is closed under finite intersections , then the group @xmath608 equals the fundamental group of the nerve of the cover @xmath20 .
let @xmath609 the category of simplicial sets and simplicial morphisms .
we introduce two functorial constructions @xmath610 for the _ nerve of a groupoid atlas @xmath0 . the first one , @xmath611 , is based on the cover by components of local groupoids .
the second one , @xmath537 , preserves more information about the local groupoids . using these constructions
, we can define homology theory of groupoid atlases and also a weak and a strong version of the classifying space of a groupoid atlas .
_ the _ weak nerve _ @xmath613 of a groupoid atlas is the simplicial set whose @xmath427-simplices are the sequences of @xmath614 elements of the same component of a local groupoid , @xmath615 equipped with usual faces and degeneracies : the face @xmath616 erases the @xmath92-th element and the degeneracy @xmath617 repeats the @xmath618-th . given a weak map of groupoid atlases @xmath113 , we have @xmath619 defined by @xmath620 .
it is well defined since @xmath271 preserves local frames .
note that this construction is functorial : @xmath621 and @xmath622 .
consider @xmath623)$ ] the @xmath427-simplex in the category @xmath242 . since it has only one index , the entire set @xmath624)}=\set{0, ... ,n}$ ] is a local frame .
a weak map @xmath623)\to a$ ] is a function @xmath625 such that its image is a local frame .
thus , a simplex @xmath626 can be viewed as a weak map @xmath623)\to a$ ] .
faces and degeneracies are , under this correspondence , compositions with the canonical inclusions @xmath627)\to a(\delta[n])$ ] and projections @xmath628)\to a(\delta[n])$ ] , respectively . hence , @xmath629),a).\ ] ] the weak nerve of @xmath0 is the simplicial set that naturally arises from the complex @xmath573 ( cf .
the geometric realization @xmath630 has the homotopy type of the polyhedron induced by @xmath573 , so it has the same fundamental group and the same homology groups .
the weak nerve is constructed from the covering of @xmath23 by the components of each @xmath205 .
it has no more information about the groupoid structure than that .
the strong version of the nerve appears naturally when one looks for a construction that preserves the local information .
let @xmath0 be a regular atlas .
the _ nerve _
@xmath537 is the simplicial set whose @xmath427-simplices are @xmath632 where we identify @xmath633 with @xmath634 , its image through the structure functors @xmath50 .
the face and degeneracy maps are defined as it is usual for nerves of categories : a face composes two arrows , a degeneracy inserts an identity ( cf .
they pass to the quotient since @xmath50 are functors .
note that the @xmath64-simplices of an atlas @xmath0 are the points and the @xmath62-simplices are the global arrows . in general , the set @xmath637 is a quotient of the disjoint union of the @xmath638-simplices of the nerves of the local groupoids @xmath639 .
given @xmath113 a map of groupoid atlases , it determines a map between the nerves @xmath640 by the formula @xmath641=[(\g_f(\a)(g_1), ... ,\g_f(\a)(g_k))]\ ] ] for arrows @xmath642 in @xmath205 .
this definition does not depend on the representative of the class @xmath643\in na_k$ ] .
the simplex @xmath643 $ ] is mapped by @xmath645 and @xmath646 into @xmath647 $ ] and @xmath648 $ ] , respectively . since @xmath275 , we have @xmath649 and @xmath650 .
@xmath274_{\g_f(\a ) } \ar[dr]^{\g_{f'}(\a ) } \\ \g_{\phi_f(\a ) } \ar[rr]_{\phi_{\phi_f(\a)}^{\phi_{f'}(\a ) } } & & \g_{\phi_{f'}(\a)}}\ ] ] we conclude that @xmath651 .
recall that the @xmath638-simplices of the nerve @xmath654 of a small category @xmath655 can be presented as @xmath656 and when @xmath655 is a groupoid , by the universal property of localization , we have @xmath657,c).\ ] ] it is clear that @xmath658),a)$ ] equals the disjoint union of the @xmath638-simplices of @xmath636 , since a map @xmath659)\to a$ ] picks an index @xmath41 of @xmath203 and gives a functor @xmath660\to \g\suba$ ] .
note that two maps @xmath659)\to a$ ] are equivalent if and only if the simplices that they define are identified in @xmath537 .
thus , @xmath661),a)/\sim\ = hom_{[\ga]}(a(\delta[k]),a).\ ] ] therefore , as it happens with the weak nerve , the functor @xmath662\to \ss$ ] is an example of _ singular functor _ in the vocabulary of @xcite .
it means that there exists a functor @xmath663 $ ] from the category of finite ordinal numbers such that @xmath664}(\theta(\set{0< ... <k}),a)$ ] , and the faces and degeneracies are given by composition with @xmath665 , where @xmath35 is an elementary injection or surjection in @xmath666 . in this case , the functor @xmath667 is @xmath668)$ ] .
[ na is not kan ] the nerve of a groupoid atlas @xmath537 is not in general a kan complex .
it might happen that two 1-simplices can not be extended to another @xmath62-simplex : take local groupoids @xmath205 and @xmath294 and let @xmath97 be objects of @xmath205 and @xmath294 , respectively .
let @xmath585 be an object shared by these two groupoids .
suppose that @xmath669 is not a local frame .
if @xmath670 and @xmath671 are arrows of @xmath205 and @xmath294 respectively , then there is not a 2-simplex @xmath17 in @xmath537 such that @xmath672 and @xmath673 .
if the atlas @xmath0 is irreducible , the set @xmath637 is the colimit over @xmath203 of the sets @xmath639 , since it is the largest quotient of the disjoint union that makes the diagrams @xmath678^{\phiab}\ar[rd ] & & ( n\g\subb)_k\ar[dl]\\ & ( \coprod ( n\g\suba)_k)/\sim}\ ] ] commutative . since limits and colimits in @xmath609
can be computed coordinatewise , we have @xmath679 when @xmath0 is not irreducible , there is not a well defined diagram in the category of sets since the functions @xmath680 are partially defined .
[ na equals nia ] the nerve @xmath537 of an atlas @xmath0 is equal to the nerve @xmath681 of the irreducible atlas @xmath317 , since both atlases have the same simplices by the universal property of @xmath317 .
then , the nerve of an arbitrary atlas can be computed as the nerve of an irreducible atlas . in the rest of this section we will assume , without loss of generality , that @xmath0 is irreducible .
we have seen that the fundamental group of @xmath0 is the set of paths of global arrows modulo the relations generated by @xmath683 .
note that @xmath281 and @xmath410 are the three faces of the 2-simplex @xmath684 , so the group @xmath554 is the set of paths of 1-simplices that start and end in @xmath35 modulo the simplicial homotopies , i.e. @xmath554 equals @xmath682 . in theorem
[ fundamental group ] we introduced another algebraic object related to @xmath0 , namely the _ colimit groupoid _ over the diagram @xmath334 , which we denoted @xmath687 .
@xmath688 let @xmath689 be a map in @xmath242 .
the family @xmath690 commutes with the structural functors @xmath71 , hence induces a functor @xmath691 by the universal property of @xmath687 . with this definition on arrows
, @xmath692 becomes a left adjoint for the functor @xmath397 .
@xmath693 the functor @xmath196 factors through @xmath224 $ ] , since a map and a corestriction of it give rise to the same family @xmath690 .
it is easy to see that @xmath694 .
recall that an arrow of @xmath687 is a path of colimit arrows modulo the smallest equivalence class that contains the local identities and compositions . here , a _ colimit arrow _
@xmath695 means an element of @xmath696 . in other words , it is an arrow of the colimit graph of the underlying graphs of @xmath205 .
let @xmath140 be the union @xmath702 . with the faces and degeneracies of @xmath700
, @xmath140 results a simplicial set . recall that an @xmath427-simplex @xmath17 of @xmath537 is the class of a functor @xmath703\xto{s } \g\suba\ ] ] for some @xmath41 , under the quotient map that identifies @xmath17 with @xmath704 for all @xmath48 .
we assign to @xmath17 the composition @xmath705\to \g(a)$ ] .
this factors through the quotient , so there is a function @xmath706 .
since this function preserves faces and degeneracies , it gives rise to a simplicial map @xmath707 .
it is surjective by definition of @xmath140 .
it is not difficult to prove that it is also injective . like the geometric realization of any simplicial set ,
@xmath708 is a cw - complex with a cell for each non - degenerated simplex of @xmath537 . as a consequence of our results on the nerve of a groupoid atlas
, we obtain the following propositions . given a simplicial complex @xmath183 , its associated groupoid atlas @xmath213 is infimum and all its local groupoids are simply connected .
then , by proposition [ na equals nia ] its weak nerve equals its strong nerve .
it follows that @xmath712 . since @xmath713 is homotopy equivalent to the polyhedron @xmath183
, we conclude that any _ interesting homotopy type _ can be obtained as the classifying space of a groupoid atlas .
an algebraic loop in @xmath0 induces a topological one in @xmath708 , since @xmath714 and a map of groupoid atlases @xmath121 which stabilizes can be restricted to a closed real interval .
this assignation preserves homotopy classes and induces an isomorphism @xmath715 , which is just the composition @xmath716 .
consider @xmath726 .
suppose that there exists a section @xmath727 for @xmath593 .
then @xmath728 , and @xmath729 , arrows of @xmath267 and @xmath723 , respectively .
since @xmath730 , @xmath731 and since @xmath92 commutes with the face maps then we have @xmath732 which is a contradiction .
[ def de homologia ] let @xmath0 be a groupoid atlas and let @xmath52 be a commutative ring .
the _ homology of @xmath0 with coefficients in @xmath52 _ is the homology of the associated simplicial set @xmath537 .
@xmath735 explicitly , for each @xmath427 we put @xmath736 $ ] , the free @xmath52-module with basis @xmath737 , and the boundary map @xmath738 is defined , as usual , by @xmath739 in the basis elements .
we will simply denote @xmath740 and @xmath741 when @xmath742 . 1 .
two equivalent atlases have the same homology groups , since their nerves are isomorphics ( cf .
[ iso ] ) .
2 . similarly , the homologies of @xmath0 and @xmath317 are the same .
thus , we can restrict our attention to the homology of irreducible atlases .
3 . if @xmath203 has a single element @xmath41 , the homology groups of @xmath0 are the homology groups of the groupoid @xmath205 , which are the direct sum of the homologies of the vertex groups of each component of @xmath205 .
if @xmath743 , the group @xmath744 is free with one generator for each component of @xmath0 .
thus , we have an isomorphism @xmath745.\ ] ] 5 . when @xmath742 , the groups @xmath746 and @xmath747 are related by a quotient map @xmath748 with kernel the commutator of the fundamental group ( the first one coincides with @xmath749 and the second one with @xmath750 ) .
thus , the classical equation @xmath751\ ] ] remains true in this context .
recall that a simplicial set and its geometrical realization share all their homology groups .
we can compute the homology of the atlas @xmath0 of [ rombitos ] from its classifying space , since it has the homotopy type of @xmath752 .
@xmath753 given @xmath0 a groupoid atlas , note that the chain complex @xmath754 is the colimit of the complexes @xmath755 : the free functor @xmath756 $ ] is left adjoint to the forgetful functor , hence it preserves colimits .
@xmath757=\colim{}{r[n\g\suba]}=\colim{}{c_*(\g\suba , r)}\ ] ] [ a , h y bh ] in the case of @xmath2 , the chain complex @xmath758 coincides with the chain complex @xmath759 , defined in @xcite as the colimit of the @xmath760$]-complexes induced by the non normalized homogeneous bar resolutions @xmath761 of the groups @xmath18 .
the homology groups of @xmath0 are the homology groups of the nerve of the cover of @xmath3 by the @xmath8-orbits , @xmath158 .
the nerve @xmath537 is a simplicial subset of @xmath700
. they may differ ( [ na neq nga ] ) as the functor @xmath762 does not preserve colimits .
when @xmath0 is a connected atlas , @xmath763 equals the homology groups of its fundamental group .
we deduce that @xmath764 in general , since @xmath708 can reach the homotopy type of any cw - complex , and the fundamental group of a cw - complex does not determine all its homology groups .
[ [ relation - between - ha - and - the - local - homology - groups - hgsuba ] ] relation between @xmath765 and the local homology groups @xmath766 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ consider the map @xmath767 .
it induces an epimorphism of chain complexes @xmath768 , with kernel @xmath769 .
the short exact sequence @xmath770 induces a long exact sequence relating the homologies of the complex @xmath769 , the local groupoids @xmath205 and the groupoid atlas @xmath0 .
@xmath771 in some cases , the complex @xmath769 can be written as a direct sum of some of the complexes @xmath772 . in these cases ,
the homology of @xmath0 can be computed from the homology of the local groupoids .
let @xmath0 be the _
1-sphere _ , defined in [ sc example ] .
the set @xmath23 is @xmath773 , the set of indices @xmath203 consists of the proper non empty subsets of @xmath23 , ordered by inclusion , and for every @xmath17 , the local groupoid @xmath216 is the tree over @xmath17 .
since it is infimum and all its local groupoids are simply connected , by ( [ na equals nwa ] ) we have @xmath774 for each @xmath17 , the nerve of the local groupoid @xmath216 is @xmath775 .
then , there is a short exact sequence of complexes @xmath776 where @xmath777 since the groupoids @xmath216 are simply connected , the non trivial part of the long exact sequence relating the homology groups is , in this case , @xmath778 the group @xmath779 is isomorphic to @xmath117 because @xmath0 is connected , and so @xmath780 : it is free of rank 1 since @xmath781 and the euler characteristic of the last complex is 0 .
all the local groupoids are , in this example , simply connected .
then , by the long exact sequence , the homology groups @xmath741 should be trivial for @xmath783 .
this does not happen , for example , with the @xmath427-sphere ( @xmath783 ) , whose local groupoids are also simply connected .
therefore , for many examples we can not find any expression of @xmath769 as a sum of complexes @xmath772 .
in the example of above , we use the usual presentation of the colimit @xmath758 as the cokernel of the map @xmath784 which is defined on the basis elements by @xmath785 . here
@xmath786 denotes the element of the factor @xmath92 given by the simplex @xmath17 .
in general @xmath618 is not a monomorphism , then one can not identify @xmath769 with the sum @xmath787 .
let @xmath789 be the map defined in the basis elements by @xmath790 . here
@xmath791 denotes the class of the simplex @xmath17 in @xmath537 .
let @xmath792 be an element of @xmath793 .
we have @xmath794 then , for all @xmath795 we have @xmath796 and @xmath797 , where @xmath798 is the boundary map of the simplicial chain complex of @xmath606 . this way , a non trivial element of @xmath793 gives a cycle in some @xmath606 .
the converse is clear .
we expose a simple aplication of last result .
let @xmath0 be a groupoid atlas such that @xmath203 is discrete ( @xmath801 ) . recall that , since @xmath0 is not good , @xmath802 by definition , and note that @xmath803 is discrete for @xmath17 of dimension @xmath804 and is a star for @xmath17 a point .
thus , we have the procedure of above can be emulated in other contexts , considering a cofinal subset of local groupoids @xmath809 to make the canonical map @xmath810 an epimorphism ( when @xmath203 is finite , then @xmath140 could be the set of maximal objects ) , and then construct @xmath811 as the sum of the chain complexes associated to certain local groupoids . | a global action is the algebraic analogue of a topological manifold .
this construction was introduced in first place by a. bak as a combinatorial approach to k - theory and the concept was later generalized by bak , brown , minian and porter to the notion of groupoid atlas . in this paper
we define and investigate homotopy invariants of global actions and groupoid atlases , such as the strong fundamental groupoid , the weak and strong nerves , classifying spaces and homology groups .
we relate all these new invariants to classical constructions in topological spaces , simplicial complexes and simplicial sets .
this way we obtain new combinatorial formulations of classical and non classical results in terms of groupoid atlases .
_ 2000 mathematics subject classification .
19d99 , 20l05 , 18g55 , 55u35 .
_ _ key words and phrases . global actions , groupoid atlases , simplicial objects , homology , k - theory . |
perturbative qcd corrections are well - known for being the dominant contributions to the radiative inclusive penguin decay @xcite .
this perturbative dominance was recently reassured by a dedicated analysis @xcite in which non - perturbative corrections to the inclusive decay mode @xmath1 have been estimated to be well below @xmath2 . within a global effort , a perturbative qcd calculation to the next - to - next - to - leading - logarithmic ( nnll ) level within the standard model ( sm ) has been performed and has led to the first nnll prediction of the @xmath3 branching fraction @xcite . using the photon energy cut @xmath4 gev ,
the branching ratio reads @xmath5 this result is based on various highly - nontrivial perturbative calculations @xcite .
the combined experimental data according to the heavy flavor averaging group ( hfag ) @xcite leads to @xmath6 where the first error is combined statistical and systematic , and the second is due to the extrapolation in the photon energy .
thus , the sm prediction and the experimental average are consistent at the @xmath7 level .
this is just one example among the impressive confirmation of the sm in all experiments in flavour physics during the last decade @xcite , including the first generation of the @xmath8 factories at kek ( belle experiment at the kekb @xmath9 collider ) @xcite and at slac ( babar experiment at the pep - ii @xmath9 collider ) @xcite , and the tevatron @xmath8 physics programs ( cdf @xcite and d0 @xcite experiments ) .
also the first results of the lhcb experiment @xcite are in full agreement with the simple ckm theory of the sm .
this feature is somehow unexpected because in principle flavour changing neutral current ( fcnc ) processes like @xmath3 offer high sensitivity to new physics ( np ) .
additional contributions to the decay rate , in which sm particles in the loops are replaced by new particles such as the supersymmetric charginos or gluinos are not suppressed by the loop factor @xmath10 relative to the sm contribution .
thus , fcnc decays provide information about the sm and its extensions via virtual effects to scales presently not accessible otherwise .
this approach is complementary to the direct production of new particles at collider experiments .
the experimental fact that none of the dedicated flavour experiments has observed any unambiguous sign of new physics yet , in particular no @xmath11 np effects in any fcnc process , implies the famous flavour problem , namely why fcnc processes are suppressed .
it has to be solved in any viable new physics model .
the hypothesis of minimal flavour violation ( mfv ) @xcite , i.e. that the np model has no flavour structures beyond the yukawa couplings , solves the problem formally . however , new flavour structures beyond the yukawa couplings are still compatible with the present data @xcite because the flavour sector has been tested only at the @xmath2 level in the @xmath12 transitions .
today supersymmetric models are often given priority in our search for np beyond the sm .
this is primarily suggested by theoretical arguments related to the well - known hierarchy problem .
supersymmetry eliminates the sensitivity for the highest scale in the theory and , thus , stabilizes the low energy theory .
there are other features in supersymmetric theories which are promising like the unification of the gauge couplings and the existence of a dark matter candidate .
supersymmetry also represents the unique extension of poincare symmetry .
the precise mechanism of the necessary supersymmetry breaking is unknown .
a reasonable approach to this problem is the inclusion of the most general soft breaking term consistent with the sm gauge symmetries in the so - called unconstrained minimal supersymmetric standard model ( mssm ) .
this leads to a proliferation of free parameters in the theory .
the decay @xmath1 is sensitive to the mechanism of supersymmetry breaking because , in the limit of exact supersymmetry , the decay rate would be just zero : @xmath13 this follows from an argument first given by ferrara and remiddi in 1974 @xcite . in that work the absence of the anomalous magnetic moment in a supersymmetric abelian gauge theory was shown . in the mssm
there are new sources of fcnc transitions . besides the ckm - induced contributions , which are brought about by a charged higgs or a chargino
, there are generic supersymmetric contributions that arise from flavour mixing in the squark mass matrices in case they are not aligned to the ones in the quark sector .
then the gluino contribution enhanced by an extra factor @xmath14 instead of @xmath15 significantly contributes to the decay rate .
thus , the general structure of the mssm does not explain the suppression of fcnc processes , which is observed in experiments ; the gauge symmetry within the supersymmetric framework does not protect the observed strong suppression of the fcnc transitions .
this is the crucial point of the well - known supersymmetric flavour problem .
parameter bounds on np from flavour physics is a model - dependent issue .
the present data on @xmath3 implies a very stringent bound for example on the inverse compactification radius of the minimal universal extra dimension model ( macd ) ( @xmath16 at @xmath17 cl ) @xcite .
the bound is much stronger than the ones derived from other measurements .
moreover , there is a bound induced by @xmath1 on the charged higgs mass in the two higgs - doublet model ( ii ) : @xmath18 at @xmath17 cl @xcite .
it is based on a nll qcd calculation within this model presented in refs .
the latter bound is not valid in general two - higgs doublet models , especially in supersymmetric models .
however , the two - higgs - doublet model ( ii ) is a good approximation for gauge - mediated supersymmetric models with large @xmath19 , where the charged higgs contribution dominates the other supersymmetric contributions . simplifying assumptions about the parameters often introduce model - dependent correlations between different observables .
thus , flavour physics will also help in discriminating between the various models that will be proposed by then .
in view of this , it is important to calculate the rate of the rare @xmath8 decays , with theoretical uncertainties as reduced as possible and general enough for generic supersymmetric models .
the rare decay @xmath1 has already carved out large regions in the space of free parameters of most of the supersymmetric models .
once more precise data from the super @xmath8 factories are available , this decay will undoubtedly gain even more efficiency in selecting the viable regions of the parameter space in the various classes of models .
constraints based on nontrivial qcd calculations within various supersymmetric extensions are heavily analyzed in the literature , see for example the refs . .
finally , model - independent analyses in the effective field theory approach without @xcite and with the assumption of minimal flavour violation @xcite also show the strong constraining power of the @xmath20 branching fraction .
while in the sm , the rate for @xmath0 is known up to nnll in qcd , also within supersymmetric theories higher order calculations have been pushed forward in recent years . at the ll level
there are several contributions to the decay amplitude : besides the contributions solely induced by flavour mixing in the quark sector with a @xmath21 boson or a charged higgs boson and a top quark in the loop , there is also a chargino contribution with an up - type squark which can be induced by the ckm matrix .
if we consider also generic new sources of flavour violation induced by a disalignement of quarks and squarks , there are additional contributions from a chargino , gluino and also neutralino .
the first complete analysis of the decay rate of @xmath1 has been presented in ref .
@xcite .
it is highly desirable to analyse these non - standard contributions with nll precision : besides the large uncertainties in the ll predictions , the step from the ll to the nll precision is also necessary in order to check the validity of the perturbative approach in the model under consideration .
moreover , it was already shown in specific np scenarios that bounds on the parameter space of non - standard models are very sensitive to nll contributions .
the mfv hypothesis is a formal model - independent solution to the np flavour problem .
it assumes that the flavour and the cp symmetry are broken as in the sm .
thus , it requires that all flavour- and cp - violating interactions be linked to the known structure of yukawa couplings .
a renormalization - group - invariant definition of mfv based on a symmetry principle is given in @xcite ; this is mandatory for a consistent effective field theoretical analysis of np effects .
the mfv hypothesis is an important benchmark . because any measurement which is inconsistent with the general constraints and relations induced by the mfv hypothesis @xcite indicates the existence of new flavour structures .
this hypothesis can also be used within the mssm .
it can be implemented by assuming that the squark and quark mass matrices can be simultaneously diagonalized ( alignement ) . in this case
there are no flavour - changing interactions induced by the gluino at the tree level . the first nll calculation of the inclusive decay @xmath1 in the mssm with the mfv hypothesis includes the gluon corrections to the charged higgs and the chargino contribution @xcite , see figure [ figure1 ] .
in particular the possibility of destructive interference of the chargino and the charged higgs contribution is studied .
the analysis is done under the mfv assumption that the only source of flavour violation at the electroweak scale is that of the sm , encoded in the ckm matrix .
other flavour - changing interactions were suppressed by assuming the gluino being heavy .
it is found that , in this specific supersymmetric scenario , bounds on the parameter space are rather sensitive to nll contributions and they lead to a significant reduction of the stop - chargino mass region , where the supersymmetric contribution has a large destructive interference with the charged - higgs boson contribution @xcite .
there are also further analyses within the mfv hypothesis which try to include only the potentially large contributions beyond the leading order which are enhanced by large @xmath19 factors or by large logarithms of the form @xmath22 where the masses of the supersymmmetric particles are assumed to be significantly larger than the @xmath21-boson mass @xcite .
a practically complete mfv analysis has been presented in ref . @xcite . to ll precison
this calculation includes the one - loop diagrams containing a @xmath21 boson and up - type quark , or a charged higgs boson and an up - type quark , or a chargino and an up - type squark ( see figure [ figure2 ] ) .
neutralino and gluino exchange diagrams are neglected under the mfv assumption . to nll precision the gluonic two - loop corrections to the sm and charged higgs loops
are included , also two - loop diagrams with a gluino together with a higgs or w boson , and finally two - loop diagrams with a chargino together with a gluon or a gluino or a quartic squark coupling .
as already shown in ref.@xcite , the two - loop gluonic corrections to the chargino loops are not uv finite : in order to obtain a finite result one has to combine them with the chargino - gluino diagrams
. however , a mfv analysis should take into account the fact that the simultaneous diagonalization of the quark and squark mass matrices can be imposed at one scale @xmath23 .
the renormalization group evolution of the mssm parameters then leads to a disalignement between the squark and quark mass matrices at scales different from @xmath23 @xcite .
so if the mfv condition is imposed at a scale much larger than the superparticle mass scale @xmath24 , very large logarithms of @xmath25 occur in the wilson coefficients .
then the soft susy - breaking mass parameters which are assumed to be flavour - diagonal at the scale @xmath23 must be evolved down to @xmath24 with the help of the appropriate renormalization group equations ( rge ) , thus , generating some flavour violation in the squark mass matrices which gets absorbed in the couplings of the squark mass eigenstates with the gluinos and charginos . in ref .
@xcite , it is argued , that the effects of the rge - induced flavour mixing is relatively small and , therefore , are only included to ll order , in the one - loop diagrams with gluinos and down - type squarks and in the one - loop diagrams with charginos and up - type squarks .
there is a public computer code for this mfv calculation available which includes all contributions discussed above @xcite . beyond minimal flavour violation , the most important role
is played by the non - diagonal gluino - quark - squark vertex due to the large strong coupling which comes with this vertex .
as discussed above , this flavour non - diagonal vertex is induced by squark - mixing to the extent as it is misaligned with quark mixing .
it represents a new flavour structure beyond the sm yukawa couplings . to understand these new sources of flavour violation that may be present in supersymmetric models in addition to those enclosed in the ckm matrix
, one has to consider the contributions to the squark mass matrices @xmath26
\left(m_{\,f,\,lr}^{2}\right)^{\dagger } + f_{f\,rl } & \
\ m^2_{\,f,\,rr } + f_{f\,rr } + d_{f\,rr } \end{array } \right ) \
, , \label{squarku}\ ] ] where @xmath27 stands for up- or down - type squarks . in the super - ckm basis , where the quark mass matrices are diagonal and
the squarks are rotated in parallel to their superpartners , the @xmath28 terms from the superpotential and the @xmath29 terms from the gauge sector turn out to be diagonal @xmath30 submatrices of the @xmath31 mass matrices @xmath32 .
this is in general not true for the additional terms @xmath33 with @xmath34 , originating from the soft supersymmetric breaking potential . because the squark - quark - gluino coupling is flavour - diagonal in the super - ckm basis , the gluino vertex in the mass eigenstate basis is non - diagonal in flavour - space due to the off - diagonal elements of the soft terms @xmath35 , @xmath36 , @xmath37 .
a complete ll analysis of the corresponding gluino contribution to the inclusive decay rate of @xmath1 has been presented in ref .
the sensitivity of the bounds on the down squark mass matrix to radiative qcd ll corrections is systematically analysed , including the sm and the gluino contributions . in ref .
@xcite the interplay between the various sources of flavour violation and the interference effects of sm , gluino , chargino , neutralino and charged higgs boson contributions is studied .
the bounds on simple combinations of elements of the soft part of the squark mass matrices are found to be , in general , one order of magnitude weaker than the bound on the single off - diagonal element , which was derived in previous work by neglecting any kind of interference effects .
some effects beyond ll precision like large @xmath38 effects are estimated in ref .
@xcite in analogy to the mfv analyses of refs .
@xcite .
recently , the complete nll corrections to the wilson coefficients ( at the matching scale @xmath39 ) of the various versions of magnetic and chromomagnetic operators which are induced by a squark - gluino loop have been calculated @xcite . in this analysis
all the appearing heavy particles ( which are the gluino , the squarks and the top quark ) are simultaneously integrated out at the high scale .
there are two classes of two - loop diagrams which have to be considered : diagrams with one gluino and a virtual gluon and diagrams with two gluinos ( see figure [ figure3 ] ) or with one gluino and a squark - loop .
the former have been presented already in ref .
@xcite and now confirmed , while the latter have been calculated for the first time @xcite . besides these nll contributions due to the gluino vertex , there are of course more nll corrections with non - minimal flavour violation ; they involve electroweak ( gaugino and higgsino ) vertices
. however , such contributions are in general suppressed compared to the ones related to the gluino .
there are two types of such contributions at the nll level : first , there are electroweak corrections to the non - minimal ll gluino contribution ( in which the electroweak vertex is flavour - diagonal or mfv - like ) which are naturally suppressed due to the smaller coupling constants and due to the ckm hierarchy .
second , there is also non - minimal flavour violation via squark - mixing in the electroweak vertices possible . but
such contributions are already suppressed at the ll level compared to the gluino contribution due to the smaller coupling constant , apart from the chargino contributions in specific parts of the parameter space in which for example the trilinear coupling @xmath40 is very large .
these features do not change of course when gluon- and gluino - induced nll corrections are added to such ll contributions .
still , the leading chirally enhanced corrections can be easily calculated by inserting the effective feynman rules of @xcite into the results of @xcite .
summing up , the complete nll corrections induced by the gluino vertex given in ref .
@xcite represent the dominant contribution beyond mfv at this order in most parts of the mssm parameter space .
they are complementary to the mfv contributions at the nll level which are given in ref .
the results are presented also in public computer code @xcite .
th thanks the organizers of the conference for the interesting and valuable meeting and the cern theory group for its hospitality during his regular visits to cern where part of this work was written .
+ m. misiak and m. steinhauser , nucl .
b * 764 * ( 2007 ) 62 [ arxiv : hep - ph/0609241 ] .
d. asner _
[ heavy flavor averaging group ] , arxiv:1010.1589 [ hep - ex ] + and online update at http://www.slac.stanford.edu/xorg/hfag ( 2010 ) u. haisch and a. weiler , phys .
d * 76 * , 034014 ( 2007 ) [ arxiv : hep - ph/0703064 ] . m. ciuchini , g. degrassi , p. gambino and g. f. giudice , nucl .
b * 527 * ( 1998 ) 21 [ arxiv : hep - ph/9710335 ] .
f. borzumati and c. greub , phys .
d * 58 * , 074004 ( 1998 ) [ arxiv : hep - ph/9802391 ] .
g. degrassi , p. gambino and g. f. giudice , jhep * 0012 * , 009 ( 2000 ) [ arxiv : hep - ph/0009337 ] .
m. s. carena , d. garcia , u. nierste and c. e. m. wagner , phys .
b * 499 * , 141 ( 2001 ) [ arxiv : hep - ph/0010003 ] .
f. borzumati , c. greub , t. hurth and d. wyler , phys .
d * 62 * , 075005 ( 2000 ) [ arxiv : hep - ph/9911245 ] .
t. besmer , c. greub and t. hurth , nucl .
b * 609 * , 359 ( 2001 ) [ arxiv : hep - ph/0105292 ] .
m. ciuchini , e. franco , a. masiero and l. silvestrini , phys .
d * 67 * , 075016 ( 2003 ) [ erratum - ibid .
d * 68 * , 079901 ( 2003 ) ] [ arxiv : hep - ph/0212397 ] .
m. ciuchini , a. masiero , l. silvestrini , s. k. vempati and o. vives , phys .
. lett . * 92 * , 071801 ( 2004 ) [ arxiv : hep - ph/0307191 ] .
m. ciuchini , a. masiero , p. paradisi , l. silvestrini , s. k. vempati and o. vives , nucl .
b * 783 * , 112 ( 2007 ) [ arxiv : hep - ph/0702144 ] .
w. altmannshofer , d. guadagnoli , s. raby and d. m. straub , phys .
b * 668 * , 385 ( 2008 ) [ arxiv:0801.4363 [ hep - ph ] ] . | we give a brief overview about perturbative corrections to the inclusive decay mode @xmath0 in supersymmetric models . |
much of the recent interest in theories with extra dimensions stems from the fact that they provide a possible solution to the hierarchy problem of the standard model . in the case of flat extra dimensions
@xcite , the fundamental planck scale could take values of order of the weak scale .
the gravity interactions at long distances are governed by the four dimensional planck scale , @xmath0 , which is related to the fundamental planck scale @xmath1 by a factor proportional to the volume of the extra dimensional space .
therefore , for large values of the compactification scale and/or large number of extra dimensions , one can reconcile the observed value of the planck scale with a fundamental scale @xmath2(tev ) .
since this mechanism demands an a priori unexplained large extra - dimensional volume , it could be argued that rather than providing a solution to the hierarchy problem , theories with flat extra dimensions allow a reformulation of the problem . on the other hand , flat extra dimensions
would only be possible in the presence of a tension - less brane , confining all sm fields . under these conditions , however , it is somewhat more natural to think about a brane with finite tension , and an induced curvature in the extra dimensional scenarios .
non - vanishing curvature provides a distortion of the metric , which allows new alternatives to address the hierarchy problem .
for instance , one can find simple solutions to the case of branes with finite tension by assuming that the local four dimensional metric is affected by an exponential warp factor , which depends linearly on the position of the brane .
in this simple famework , one can get a solution to the hierarchy problem without assuming any unnatural large factor @xcite . indeed , assuming all fundamental mass scales to be of the same order , the ratio of the physical higgs vacuum expectation value to the observable four dimensional planck scale is exponentially suppressed , with an exponent that depends on the warp factor @xmath3 times the position @xmath4 of the higgs brane in the extra dimensions .
taking @xmath5 provides a good solution to the hierarchy problem .
it is important to emphasize that , as in any other solution , new physics appears at the tev scale .
this new physics includes graviton kaluza klein states and a graviscalar , also called the radion , with effective interactions suppressed by a scale of order of the tev , @xmath6 .
it is natural to assume that not only the higgs field , but all matter fields are confined to the brane at @xmath7 , which we shall call the infrared brane .
the gauge fields , however , may propagate in the extra dimensions .
one motivation for this is the fact that , from the field theoretical point of view , it is more difficult to localize gauge fields than to localize fermions .
however , there is a phenomenological obstacle to this realization : the kaluza klein excitations of the gauge bosons couple with a strength @xmath8 times the zero mode coupling , an order of magnitude larger for values of @xmath9 necessary to provide a solution of the hierarchy problem . the exchange of these strongly coupled higher kk states can be used to set a lower bound on the mass of the first kk excitation that is about 20 tev @xcite .
precision electroweak observables present additional challenges , since these observables are affected at the tree - level @xcite .
the modification of these observables is related to the shape of the extra dimensional wave functions in the presence of a localized symmety breaking vev .
the presence of the localized scalar vev induces a repulsion of the gauge fields from the brane , leading to a breakdown of the usual linear relation between the gauge boson mass with its coupling and the higgs vev . in order to recover acceptable phenomenological predictions ,
the bound on the first kk excitation has to be pushed to even larger values than the ones obtained from the kk boson exchange . in this article
we analyze how the results for the gauge field interactions described above is modified by the presence of local gauge field kinetic terms .
such terms are naturally expected to be present in any realistic theory , and indeed will be induced radiatively by the localized higgs and fermions on the ir brane even if the underlying dynamics is such that they vanish at any particular scale @xcite .
the presence of a local brane kinetic term on the infrared brane implies that at sufficiently high energies , the gauge interactions on the brane should be four dimensional and renders the brane opaque to gauge fields of short wavelength along the extra dimension .
this can only be possible if there is an important modification of the couplings of the gauge kk modes to matter localized on the brane @xcite .
moreover , for sufficiently large local gauge kinetic terms , the physics on the brane should be dominated by the four dimensional behavior , and therefore it should be possible to relax these bounds on the mass of the lightest gauge boson kk mass .
the plan of this work is as follows . in section [ sec :
prop ] , we shall give the expression for the gauge field propagator with brane kinetic terms .
we shall analyze its behavior in different momentum regimes and for different values of the local gauge kinetic terms . in section [ sec : kk ] , we proceed with the kaluza - klein decomposition . after reviewing the situation in the case of transparent branes , in the absence of local gauge kinetic terms
, we analyze the case of one or two opaque branes .
we also provide a comparison of the results obtained with the kk decomposition with the ones derived from the behavior of the five dimensional gauge field propagator . in section [ sec : higgs ] , we discuss the effects induced by the presence of a localized higgs vev .
we discuss both the case of vanishing local gauge kinetic terms and the effects associated with the presence of these local terms . in section [ sec : pheno ] we apply our results to the case where the electroweak gauge bosons propagate in the bulk , and analyze the phenomenological consequences in this scenario .
in particular , we show that the presence of the brane - localized gauge kinetic terms opens the possibility for the first kk excitations to be within reach of the lhc .
we reserve section [ sec : concl ] for the conclusions .
in this section we derive the tree - level propagator for a gauge field described by the action [ action ] s = - d^4xdy , where @xmath10 is the determinant of the metric , and capital latin letters refer to the full 5d coordinates , @xmath11 , whereas lower case greek letters refer only to the four uncompactified dimensions , @xmath12 . in the above we
have rescaled the bulk kinetic term by absorbing @xmath13 into @xmath14 , so that @xmath14 has mass dimension one , the canonical dimension for a gauge boson propagating in four dimensions
. then @xmath15 has dimensions of mass and the coefficients of the local brane terms , @xmath16 , have dimensions of length.$ ] , with the branes located at the endpoints .
consequently , each delta function contributes a factor of 1/2 when performing the @xmath17 integration .
] @xmath18 is the usual field - strength functional of the gauge fields , ^a_m n & = & _ m a^a_n - _
n a^a_m + f^abc a^b_m a^c_n , for a non - abelian yang - mills theory , with the final term omitted in the abelian case . for simplicity of notation
, we denote the extra dimensional coordinate @xmath19 as @xmath17 .
the full set of five dimensional coordinates are denoted by capital letters , i.e. @xmath20 .
we assume a background metric defined by the line element , [ lineelement ] ds^2 = e^-2 _ dx^dx^+ dy^2 , with @xmath21 , @xmath22 orbifold such that @xmath23 is even and @xmath24 is odd under reflection in @xmath17 .
it is then always possible to choose a gauge in which @xmath25 , corresponding to a unitary gauge
. this will be sufficient for our purposes , but it is also interesting to exhibit the full gauge - dependence in an @xmath26 gauge . in an @xmath26 gauge both the four dimensional vector field @xmath27 and the scalar @xmath24 propagate .
the gauge choice is defined such that terms which mix the two are zero .
this is enforced by the gauge - fixing term , - ^2 .
we could work out the ghost interactions corresponding to this choice of gauge - fixing , but will not need them for our ( tree - level ) analyses .
the propagators @xmath28 and @xmath29 , satisfy \
{ p^ + ^ _ y [ e^-2 _ y ] + 2 r_i ( y - y_i ) p^ + ^^ } g_(x , x ) & = & ^_(x - x ) + \ { ^2 + _ y^2 e^-2ky } g_55(x , x ) & = & ( x - x ) , where @xmath30 and @xmath31 indicates the sum over both the uv and ir branes . due to the translational invariance of eq .
( [ lineelement ] ) along the four noncompact dimensions , it is convenient to work in mixed position and momentum space , defined by @xmath32 .
the momentum defined in this way is a conserved quantity .
however , it is important to keep in mind that it is not necessarily the physical momentum of the propagating field .
rather , from eq .
( [ lineelement ] ) it can be inferred that the momentum that an observer would measure when standing at @xmath17 is @xmath33 . in the following
, we will be working with the `` coordinate '' momentum @xmath34 , as is standard in the literature , with the understanding that @xmath34 is the momentum as measured by uv observers . keeping the fifth coordinate explicit will also make the locality properties in @xmath17 manifest .
we choose to work in euclidean space , which will simplify the analysis later .
then the @xmath35 propagator can be written as g_(p;y , y ) = ( _ - ) g_p(y , y ) + g_p^(y , y ) , where @xmath36 and @xmath37 satisfy \ { _ y [ e^-2 _ y ] - p^2 [ 1 + 2 r_i ( y - y_i ) ] } g_p(y , y ) = ( y - y ) , + \ { _ y [ e^-2 _ y ] - } g^_p(y , y ) = ( y - y ) .
the @xmath38-functions impose the following boundary conditions at @xmath39 : [ propbound0 ] _ y = 0 & = & 0 + [ propboundl ] _ y = & = & 0 , whereas @xmath40 must vanish at the boundaries . at @xmath41 , both @xmath42 and @xmath43
must be continuous and satisfy e^-2 = g_5 ^ 2 , where @xmath44 .
the solutions can be written as [ propagator ] g_p(y , y ) = where @xmath45 , @xmath46 are modified bessel functions of order @xmath47 , @xmath48 are the smallest ( largest ) of @xmath17 , @xmath49 and [ coefficients ] a & = & i_0 ( ) - p r_uv i_1 ( ) + b & = & k_0 ( ) + p r_uv k_1 ( ) + c & = & i_0(e^k l ) + p e^k l r_ir i_1(e^k l ) + d & = & k_0(e^k l ) - p e^k
l r_ir k_1(e^k l ) .
the solution for @xmath43 takes the same form with @xmath50 and @xmath51 in eq .
( [ coefficients ] ) .
the @xmath52 propagator is of the same form as @xmath43 , but with second order bessel functions and multiplied by @xmath53 . from these results we see how the transition into the unitary gauge , @xmath54 ( or @xmath55 ) proceeds ; @xmath56 and @xmath57 . in this gauge ,
all of the physics is contained in eqs .
( [ propagator ] ) and ( [ coefficients ] ) . to understand the physics contained in eq .
( [ propagator ] ) it is instructive to analyze it in various limits . we define the following auxiliary function b(y_>,y_<,r ) = \ { + p r } , which , as the notation indicates , will be used only when @xmath58 . in this case
, @xmath59 has a very simple behavior when the condition @xmath60 is satisfied : @xmath61 when @xmath62 is ( extremely ) close to @xmath63 , and @xmath64 otherwise .
specializing to the case of interest for addressing the hierarchy problem , where @xmath65 , we can write : \i ) for @xmath66 , [ veryhighepropagator ] g_p(y , y ) & ~ & - e^k(y+y ) e^- ( e^k y _ > - e^k y _ < ) b(y_<,0,r_uv ) b(l , y _ > , e^k l r_ir ) .
\ii ) for @xmath67 , [ highepropagator ] g_p(y , y ) & ~ & - e^k(y+y ) e^- ( e^k y _ > - e^k y _ < ) b(l , y _
> , e^k l r_ir ) .
\iii ) for @xmath68 , [ nonlocalpropagator ] g_p(y , y ) & ~ & - e^k y _
e^- e^k y _
> b(l , y _ > , e^k l r_ir ) .
\iv ) for @xmath69 , [ lowepropagator ] g_p(y , y ) ~- .
\v ) for @xmath70 , [ verylowepropagator ] g_p(y , y ) ~- . several remarks are in order .
consider first the properties of the propagator when the kinetic terms vanish .
if we are doing physics at a fixed location along the warped dimension ( say , on one of the branes ) , it follows from the previous asymptotic forms that the propagator @xmath71 changes from a four dimensional ( @xmath72 ) to a five dimensional ( @xmath73 ) behavior at a scale @xmath74 .
the onset of the higher dimensional scaling indicates that the gauge theory ceases to be predictive when the external 4d momenta are much larger than @xmath75 , and that it should be cut off at a scale @xmath76 , for some constant @xmath77 .
thus , a @xmath17-dependent cutoff on @xmath34 appears naturally in this language .- dependence of the cutoff is simply a consequence of general covariance and the fact that ads is homogeneous .
the statement that the highest energy modes described by the effective theory have @xmath78 translates in position space into minimum _ coordinate _ wavelengths of size @xmath79 .
that is to say , the effective theory describes _ proper _
wavelengths larger than @xmath80 , independent of the position in the bulk . by a similar reasoning
, one can see that the powers of the warp factors @xmath81 in eqs .
( [ veryhighepropagator])([verylowepropagator ] ) are also dictated by general covariance . ] for the consistency of this picture , it is important to note that , at energy scales above @xmath75 , the observables localized at @xmath17 are only sensitive to the physics of nearby points @xmath49 , such that @xmath82 .
furthermore , case iii ) shows that the contributions from far away points @xmath83 are exponentially suppressed for all momenta above the lower cutoff @xmath84 .
thus , the physics associated with some energy scale at @xmath17 is effectively shielded from contributions that would be outside the region of validity of the effective theory .
it is remarkable that the gauge field tree - level propagator with endpoints at @xmath17 exhibits a purely four - dimensional scaling at all scales below @xmath75 .
this property together with the shielding discussed in the previous paragraph are an important ingredient in the understanding of the four - dimensional running " of the gauge couplings that has been discussed recently @xcite .
in fact , if one defines the gauge coupling at @xmath17 at a scale @xmath34 via the two point gauge correlator with endpoints at @xmath17 and external momentum @xmath34 , as advocated in @xcite , one can understand from the above discussion that , provided @xmath85 , all the propagators that are relevant in a loop calculation will exhibit a four dimensional scaling .
therefore , such momentum integrals will depend logarithmically on @xmath34 , with no power - law sensitivity . from the behavior of the five dimensional propagator @xmath86 , eqs .
( [ veryhighepropagator])([verylowepropagator ] ) , we can also quickly understand some of the effects induced by the presence of the local terms . considering again coincident points @xmath41
, we see that for @xmath87 , there are three different regions . away from the boundaries we find , as before , the five dimensional scaling @xmath88 , and we conclude that the brane kinetic terms have an exponentially small effect ( that does not show up at this order ) , as expected from locality . on the branes , however , we find @xmath89 and similarly for @xmath90 with @xmath91 in place of @xmath92 and @xmath93 in place of @xmath94 .
this shows that , for brane observers , a four dimensional behavior is recovered for @xmath95 where @xmath96 on the uv brane and @xmath97 on the ir brane . note that the ir crossover distance has been appropriately red - shifted as is generally expected for dimensionful parameters in the randall - sundrum setup .
we postpone a further analysis of the physics behind the behavior of the propagator when @xmath92 and @xmath98 are nonzero to the next section . here
we only note that , from a theoretical point of view , it is perfectly consistent to take the coefficients @xmath92 , @xmath98 to be large in units of the fundamental length @xmath99 , as long as we include their effects exactly as has been done here .
as we have seen , the previous exact propagator does not change the divergence structure of the 5d theory , except on the branes , where it actually softens it .
in particular , it is consistent to assume that all other couplings are small so that perturbation theory is valid , even if the brane kinetic terms are larger than their `` natural '' scale @xcite .
nevertheless , when the gauge fields are identified with the carriers of the standard model interactions , one can derive some constraints on the size of the brane kinetic terms from the fact that the observed standard model gauge couplings are of order one .
for simplicity , and because it is the case that will interest us in the phenomenological applications of section 5 , let us consider the case where @xmath100 . from eq .
( [ verylowepropagator ] ) , we see that the observed zero mode gauge coupling , @xmath101 , is given by [ 4dgaugecoupling ] g_0 ^ 2 = = ( 1 + ) ^-1 . in the limit
that @xmath102 , this gives @xmath103 .
this relation does not directly impose a bound on @xmath104 which , as will be shown in the following sections is the most relevant quantity .
however , if one assumes that the gauge observables at any given position @xmath17 are perturbative below the local compactification scale " @xmath105 , as would be desirable from the point of view of gauge coupling unification , one has @xmath106 , a four - dimensional loop factor ( this can be inferred from eq .
( [ lowepropagator ] ) ) . under this assumption
, it follows that @xmath107 .
recalling now that a solution to the hierarchy problem requires @xmath108 , we conclude that @xmath109 , which can be of order twenty for the electroweak interactions .
from the point of view of an observer at low energies , the relevant description of the phenomenology is in terms of the kaluza - klein modes .
bulk gauge fields in warped backgrounds with transparent branes were considered in @xcite . in order to elucidate the effects of the opaque branes ,
we begin with a brief review of some of the features of the transparent brane case .
we begin with the linearized 5d equation of motion for a bulk gauge field . from the action , eq .
( [ action ] ) , with @xmath110 , we derive the equation for the @xmath111 component of @xmath112 , _
^a^- ^ ( _ a^ ) - 2 k e^-2 k y ( _ y a^ ) + e^-2 k y ( _ y^2 a^ ) & = & 0 . in order to determine the kk spectrum , we expand @xmath113 in wave functions , a^(x ) & = & _ n f_n(y ) a_n^(x^ ) , and require the @xmath114 to obey the 4d equation of motion for a free massive gauge field , _ ^a_n^- ^ ( _ a^_n ) - m_n^2 a^_n & = & 0 .
this requires the wave functions to obey the equation , f_n(y ) & = & 0 .
the solutions to this equation are bessel functions , f_n(y ) & = & n_m e^k |y| \ { j_1 ( e^k|y| ) + b y_1 ( e^k|y| ) } , [ eq : wavefunction ] where @xmath115 is an overall normalization factor , determined by requiring each kk mode to have canonically normalized kinetic terms , [ orto ] _
0^l dy f_n(y ) f_m(y ) & = & _ nm , + _ 0^l dy e^-2ky f^_n ( y ) f^_m ( y ) & = & m_n^2 _ nm . and @xmath116 is a function of @xmath117 , both of which are determined by boundary conditions below . in order to find the masses ,
we impose boundary conditions that the first derivative of the wave functions be continuous at the @xmath118 and @xmath119 boundaries .
this requires , b^0 = - & , & b^l = - . requiring both boundary conditions be satisfied simultaneously , @xmath120 , provides a transcendental equation for the quantized masses .
this equation may be solved numerically , resulting in masses , in units of @xmath121 , of @xmath122 , and approximately @xmath123 , @xmath124 , @xmath125,
@xcite , where the difference between two subsequent mass eigenvalues tends asymptotically to @xmath126 .
the coupling of the @xmath127-@xmath128 kk gauge boson mass eigenstate to matter on the brane at @xmath7 is equal to @xmath129 times the appropriate charge of the matter field . for large values of @xmath130 ,
the kk modes couple universally ( up to a sign ) to fermions located on the ir brane at @xmath131 , with a coupling @xmath132 larger than the zero mode s coupling ( for @xmath133 , the coupling is therefore approximately @xmath134 times larger than the zero mode coupling ) .
the kk modes couple to fields on the uv brane at @xmath135 with couplings that vary with the kk mode number , and are typically suppressed compared to the zero mode coupling .
this behavior may be seen in fig .
1 , for @xmath136 .
it is interesting to compare these results with the behavior of the five - dimensional propagator for two points located on either the ir or the uv brane . for @xmath137
, that is for momenta much larger than the mass of the lightest kk mode , the five dimensional propagator with endpoints on the ir brane reads [ gpll ] g_p(l ,
l ) & = & - e^k l + & = & - ( ) . using the result that , for @xmath138 , one can approximate @xmath139 \simeq ( \pi/2 p^2 ) ( p/\tilde{m})$ ] , eq .
( [ gpll ] ) has a clear interpretation .
the five dimensional behavior of the propagator is given by the sum over kk modes with a constant coupling @xmath140 to the ir brane .
since the zero mode coupling is given by @xmath141 , the relation between these couplings is the one described above . on the uv brane ,
the propagator reaches a five dimensional behavior only at momenta much larger than @xmath3 . indeed , from eq .
( [ veryhighepropagator ] ) one obtains [ gp00 ] g_p(0,0 ) & = & - . the comparison between eqs .
( [ gpll ] ) and ( [ gp00 ] ) suggests that those kk modes with masses larger than @xmath3 couple with a constant coupling @xmath142 to the uv brane .
observe that these couplings are exponentially suppressed with respect to the couplings on the ir brane .
a numerical evaluation of the couplings of asymptotically large mass eigenstates for @xmath143 confirms that this is indeed the case .
now we allow for opacity on both the ir and the uv branes , allowing @xmath144 , @xmath145 in eq .
( [ action ] ) . in this case , the orthonormality conditions for the kk decomposition become , [ ortor ] _ 0^l dy f_n(y ) f_m(y ) & = & _ nm , + _ 0^l dy e^-2ky f^_n ( y ) f^_m ( y ) & = & m_n^2 _ nm . eq .
( [ ortor ] ) reduces to eq .
( [ orto ] ) for @xmath146 .
these conditions may be simultaneously solved , as before , by imposing the condition that the 4d gauge fields @xmath147 are on the mass shell .
the resulting wave functions satisfy , [ equationfn ] f_n(y ) & = & 0 .
in fact , the introduction of opaque branes does not affect the bulk solution for the kk modes , as written in eq .
( [ eq : wavefunction ] ) , but instead modifies the boundary conditions at @xmath118 and @xmath7 to reflect the discontinuity in the slope of the wave function .
thus , the new solutions have the same form , but with different @xmath116 s , [ boundary0 ] b^0 & = & - , + [ boundaryl ] b^l & = & - , [ eq : eigenmass ] indicating different admixture of the bessel functions @xmath148 and @xmath149 in the solutions . once again
, we determine the quantized masses by imposing the conditions on both boundaries , @xmath150 . while there is no analytic way to solve for the masses , solutions may easily be obtained numerically .
once the eigenmasses ( and therefore @xmath116 ) have been found , we normalize the wave functions and determine the coupling to either brane - localized or bulk fields . the normalization condition to determine @xmath151 in the presence of brane kinetic terms , eq .
( [ ortor ] ) , may be expressed as , = 1 , where the explicit factors of two multiplying the @xmath152 brane terms that were introduced in eq .
( [ action ] ) cancel as a result of our having considered only the physical space @xmath153 and not the reflection , @xmath154 . to examine the couplings of the kk tower to various types of fields , either confined to branes or living in the bulk , consider some representative interaction terms in the 5d theory , _ int & = & _ 0^l dy \ { ( y - y _ ) .
+ ( 1 + 2 r_uv ( y ) + 2 r_ir ( y - l ) ) .
+ & & . + ( 1 + 2 r_uv ( y ) + 2 r_ir ( y - l ) ) } , where we have taken fermions canonically normalized and confined to a brane located at @xmath155 .
the first term represents the gauge boson - fermion coupling and the latter two terms are the interactions among the bulk gauge fields for a non - abelian theory . in order to derive the effective interactions between various kk modes ,
one inserts the kk decomposition into this equation . from our normalization convention for the @xmath156
, it follows that the coupling of brane fields localized at @xmath157 to the @xmath127th kk mode is [ eq : branecoupling ] g_n = f_n ( y _ ) .
equation ( [ equationfn ] ) always has a solution with zero mass and constant wave function .
the constant wave function insures that the zero mode couples universally to all charged brane matter regardless of where it is localized , as required by the unbroken gauge invariance of the zero mode .
the normalization of the zero mode in terms of @xmath13 , @xmath98 , @xmath92 , and @xmath4 determines the zero mode gauge coupling in terms of the fundamental parameters , [ zeromodecoupling ] g_0 & = & f_0(y ) = . since it is the coupling of the lightest mode which we identify at low energies with our four dimensional gauge interaction , we choose to present the couplings of the higher kk modes relative to the zero mode coupling .
couplings of the higher kk modes of bulk fields are model - dependent , being given by integrals of products of several of the wave functions .
for the three- and four - point gauge field vertices , we have , g_nml & = & _ 0^l dy ( 1 + 2 r_uv ( y ) + 2 r_ir ( y - l ) ) f_n(y ) f_m(y ) f_l(y ) + g_nmlk & = & _ 0^l dy ( 1 + 2 r_uv ( y ) + 2 r_ir ( y - l ) ) f_n(y ) f_m(y ) f_l(y ) f_k(y ) between the @xmath158-@xmath159-@xmath160 and @xmath158-@xmath159-@xmath160-@xmath161 modes , respectively . for simplicity of notation
we have suppressed the vector and color indices , but these are readily restored . for the vertices involving the zero mode , the results are much simpler , since the zero mode wave function is independent of @xmath17 . in the three - point vertex , we find that the zero mode has constant coupling @xmath162 with all other kk mode pairs . in the four - point vertex , setting @xmath163 results in a vertex factor @xmath164 . together ,
these results demonstrate the fact that the zero mode gauge field s couplings take a universal form as dictated by its unbroken gauge invariance , resulting in the same coupling to both bulk and brane fields .
these results are the same as in the flat brane scenario studied in @xcite .
to begin with , we consider the case where @xmath144 but @xmath100 .
we fix @xmath165 as a typical value which generates approximately the correct hierarchy between the planck and weak scales . in figure
[ fig : masses ] , we present the masses of the four lowest modes as a function of @xmath166 . as the brane kinetic term increases , the kk mode masses decrease , each one approaching an asymptotic value .
the features are qualitatively similar to the flat space case @xcite .
also plotted in figure [ fig : masses ] are the couplings of each mode to fields localized in either the uv or the ir branes . in agreement with the intuition that the localized kinetic term expels the higher kk modes from the brane
, we see that the modes decouple from the ir brane , with the heavier modes decoupling more quickly .
( in contrast , the couplings to the uv brane fields increase with larger @xmath98 . )
this behavior can also be understood from the point of view of quantum mechanics . for sufficiently large spatial momenta ,
the particle loses information about the extension of the extra dimensions and the physics on the brane becomes determined by the value of the local gauge coupling .
the larger the value of @xmath167 and the larger the spatial momentum , the more striking this phenomenon must be .
in the large @xmath167 limit , the particle propagator @xmath168 should recover its four dimensional behavior , and its interactions must be equivalent to the one of a particle with gauge coupling equal to the local brane one .
this is only possible if the higher kk modes , which dominate the large momentum behavior of the propagator , decouple from the brane , a behavior that is shown in fig .
[ fig : masses ] .
some analytical understanding of the behavior of couplings and masses is provided by the analysis of the five dimensional propagator for @xmath169 and @xmath170 . for @xmath171 the propagator ,
( [ veryhighepropagator ] ) , at @xmath172 reduces to g_p(l , l ) - .
therefore , for values of @xmath173 , one recovers a four dimensional behavior in this energy regime , @xmath174 this expression can be compared with the one obtained for momenta @xmath175 from eq .
( [ verylowepropagator ] ) , [ one4d ] g_p(l ,
l ) ~- , which describes the zero mode with coupling @xmath176 . the simplest interpretation of the transition from eq . ( [ two4d ] ) to eq . ( [ one4d ] )
is that there is a single massive mode with mass of order @xmath177 , that couples to the brane with strength [ onebraneg1 ] g_1 = g_0 , while the other modes decouple .
, as in fig . 1 , @xmath178 as @xmath179 .
] we also see that when @xmath102 then @xmath180 , and therefore , in the large brane kinetic term limit , only the zero mode couples to the brane with coupling @xmath181 .
this result agrees with the numerical results displayed in fig .
[ fig : masses ] .
= 10.0cm=10.0 cm = 10.0cm=10.0 cm the physics in the two opaque brane scenario is somewhat different from the one brane case . in analogy with the case of flat extra dimensions
@xcite for values of @xmath98 and @xmath92 large compared to both @xmath4 and @xmath182 , the physics at each brane must be determined by the local couplings and , in the asymptotic limit of @xmath183 an observer on either the uv or ir brane must be insensitive to the presence of the extra dimensions , including the other brane .
this can only be true if , in this limit , two massless modes appear , and appropriate linear combinations of them couple to one of the branes with a strength given by the local coupling , while decoupling from the other brane and vice versa . since for any finite value of @xmath98 and @xmath92 there is only one zero mode
, what should happen is that the mass of the first kk mode tends to zero as the local brane terms increase , while keeping a relevant coupling to the ir and uv branes .
this behavior was already apparent from the form of the five dimensional propagator for large values of the local brane kinetic terms .
consider the propagator , @xmath86 for @xmath70 , eq .
( [ verylowepropagator ] ) , for @xmath172 : [ gpllruv ] g_p(l , l ) & ~ & - + & = & - - .
the first term is the contribution of the massless zero mode , with coupling @xmath101 as defined in eq .
( [ zeromodecoupling ] ) . the second term in eq .
( [ gpllruv ] ) describes an additional mode of mass @xmath184 , where [ gamma ] = .
provided that @xmath185 , the propagator , eq . ( [ gpllruv ] ) , shows that there is a light mode ( in addition to the zero mode ) , whose coupling is given by [ twobraneg1 ] g_1 = g_0 , which generalizes eq .
( [ onebraneg1 ] ) for nonzero @xmath92 .
we see that , when @xmath186 , the light mode couples with the same strength as the zero mode to fields localized on the ir brane .
notice also that , similarly to the case of vanishing @xmath92 , when @xmath187 , the light mode decouples from the ir brane
. in the numerical analysis , for simplicity , we consider the case @xmath188 . in figure
[ fig : masses2 ] we show the couplings to brane fields localized on the ir and uv branes .
the first mode mass goes to zero and its coupling becomes equal ( and opposite in sign in the case of uv brane fields ) to the zero mode .
this agrees with the limit @xmath189 in which bulk propagation switches off and we are left with two brane gauge theories which do not interact with each other .
the higher modes decouple from both branes .
note that the couplings to the uv brane grow slightly as @xmath167 increases , and then begin to fall off again as the higher modes decouple .
= 10.0cm=10.0 cm = 10.0cm=10.0 cm
in order to describe the electroweak theory and address the hierarchy problem , the higgs responsible for electroweak symmetry - breaking ( ewsb ) must be localized on the ir brane with action , ) .
the factor of two in front of the @xmath38-function has the same origin as for the localized terms in eq .
( [ action ] ) . ] - d^4x dy 2 ( y - l ) \ { ( d_h)^d^h + ( |h|^2 - v^2 ) ^2 } , where @xmath190 is the usual gauge - covariant derivative . in the low energy
effective theory ( after rescaling the higgs kinetic term to canonical normalization ) this results in , -d^4x \{^ ( d_h)^d_h + ( |h|^2 - v^2 e^-2k ) ^2 } , with the ewsb vev red - shifted to @xmath191 .
as mentioned above , a natural solution to the hierarchy problem is obtained for values of @xmath192 . in 5d language and working in 4d unitary gauge , the localized vev results in a gauge boson mass which is itself localized on the ir brane , [ localmass ] - d^4x dy 2(y - l ) ^2
^ a_a _ .
it is therefore interesting to analyze the effects on the gauge field propagation induced by the presence of a higgs vev on the ir brane .
in fact , the localized higgs vev bears a certain resemblance to the local gauge kinetic terms , and it can be analyzed using similar techniques .
for the purposes of this section , we consider a simple gauge group . we discuss the subtleties of the @xmath193 theory in section [ sec : pheno ] .
the localized mass only affects the boundary condition to be imposed at @xmath194 . comparing eqs .
( [ action ] ) and ( [ localmass ] ) we see that , for example , eq .
( [ propboundl ] ) is modified to _
y = = 0 .
therefore , we can obtain @xmath36 from eqs .
( [ propagator ] ) and ( [ coefficients ] ) by making the replacement [ repl ] r_ir r_ir + .
the conditions for a diagonal kk decomposition may be expressed as [ ortorv ] _ 0^l
dy f_n(y ) f_m(y ) & = & _ nm , + _ 0^l dy e^-2ky \ { f^_n ( y ) f^_m ( y ) + 2 g_5 ^ 2 v^2 ( y - l ) f_n(y ) f_m(y ) } & = & m_n^2 _ nm . once again the kk mode wave functions can be found by requiring each mode to satisfy the free field equation of motion for a massive vector field .
the kk masses are determined as before , by imposing @xmath195 , where now the coefficient @xmath196 in eq .
( [ eq : wavefunction ] ) , determined by the boundary condition at @xmath197 , can be simply obtained from eq .
( [ boundaryl ] ) by the replacement ( [ repl ] ) with @xmath198 : b^l & = & - .
we will be interested in the case that @xmath92 vanishes ( or is small ) . in this case , @xmath199 is given by b^0 = - .
let us start with the case @xmath136 .
a first intuitive phenomenon that occurs is that , due to energy considerations , the presence of the vev tries to induce a repulsion of the gauge field from the brane location .
this is balanced by the cost in energy associated with a non - vanishing @xmath17-derivative . in a kk decomposition
, this produces mixing between the whole tower of kk modes , resulting in a deformation of the five dimensional wave functions , including the zero mode , which is no longer absolutely flat .
when @xmath200 , the cost associated with the deformed wave function overshadows the localized mass induced by the higgs , and the zero mode remains approximately flat with mass given by @xmath201 .
= 10.0cm=10.0 cm = 10.0cm=10.0 cm however , when @xmath202 , the cost associated with the derivative is small compared to that associated with the brane mass . in that case
the light kk modes are expelled from the brane .
the repulsion of the gauge field from the ir brane has as an immediate consequence that all the kk mode gauge couplings , as well as the zero mode one , tend to small values . for large values of @xmath203 , the ratio of the kk mode couplings to
the zero mode tends to a constant that becomes larger than the value @xmath204 obtained for @xmath205 .
the second effect is that , due to the decoupling of the gauge field from the ir brane , the masses of the zero mode ( and light kk modes for sufficiently large values of @xmath203 ) are no longer governed by the higgs vev .
this behavior is apparent in figure [ fig : massesr0v ] where one sees that for @xmath206 the mass of the zero mode increases linearly with @xmath207 , with value @xmath208 . for @xmath209 ,
the mass of the lightest mode becomes more and more insensitive to @xmath207 as the mode bends away from the brane .
also shown in figure [ fig : massesr0v ] are the couplings of the four lowest modes to the ir and uv branes relative to the zero mode coupling in the absence of a localized higgs vev , @xmath210 .
we see the behavior inferred above : for @xmath200 the coupling is close to @xmath211 , while for @xmath212 the kk modes are repelled from the brane .
this results in the higher kk modes more and more strongly coupled to the ir brane compared to the zero mode . in this case , the theory becomes strongly coupled more quickly than it would have in the absence of the vev , and the kk modes have a dramatic effect on the phenomenology .
the situation in the case of a local gauge kinetic term on the ir brane is somewhat different .
this is due to the fact that at sufficiently large momenta the physics should be dominated by the local brane term and therefore for large values of @xmath98 compared to @xmath4 and @xmath182 there should be a mode with coupling to the ir brane given by the local gauge coupling and mass given approximately by this coupling times the vev of the higgs field .
all other modes should decouple from the ir brane , and their masses need not be correlated with @xmath207 .
we can make these observations more concrete by studying the propagator with endpoints on the ir brane .
its limiting forms in the high energy and low energy regimes are , respectively , [ largeponebrane ] g_p(l , l ) & ~ & - p k e^-k l + [ smallponebrane ] g_p(l , l ) & ~ & - p k e^-k l .
we see that for sufficiently large momenta , @xmath213 , the propagator eq .
( [ largeponebrane ] ) reduces to @xmath214 , where @xmath215 .
thus , when @xmath216 , the propagator describes a single four dimensional state of mass @xmath217 ( up to corrections of order @xmath218 ) ) .
it is also clear that the energy at which this behavior sets in is lower for larger @xmath219 .
in fact , when @xmath220 ( @xmath221 ) the low energy propagator eq .
( [ smallponebrane ] ) takes exactly the previous form , so that the full propagator reduces to the propagator of a single four dimensional state with mass @xmath217 and coupling @xmath222 . the above described behavior is illustrated in fig .
[ fig : massesrv ] from a kk point of view , where we plot the masses and couplings ( now relative to @xmath223 ) of the four lightest modes for two different values of @xmath98 , as a function of @xmath203 .
the mass spectrum exhibits a level - crossing phenomenon , more pronounced for @xmath224 than for @xmath225 , where each mode successively `` takes a turn '' feeling the higgs effect . for small @xmath207 , it is the zero mode whose mass grows linearly with @xmath207 , whereas at @xmath226 the zero mode mass becomes insensitive to @xmath207 , and it is the first kk mode mass which grows linearly .
this pattern repeats ; as each mode becomes insensitive to @xmath207 , the one above it becomes sensitive and grows linearly with @xmath207 . for large @xmath203
, the transition points occur in intervals of approximately @xmath227 , with effective coupling approximately @xmath228 , apart from the zero mode whose coupling is always given by @xmath229 . for the phenomenological applications to be discussed in section 5
, we will be interested mostly in the case where @xmath230 .
it is clear from eq .
( [ smallponebrane ] ) that , for @xmath231 , there is a single state with coupling and mass given by @xmath232 and @xmath233 , up to corrections of order @xmath234
. it will be useful to have the next order corrections to the above parameters .
we find [ m0firstorder ] m_0 & = & , + [ g0firstorder ] g_0 & = & , where [ eta ] = .
we observe that when @xmath220 then @xmath235 , which shows explicitly that , provided @xmath234 is small , in the large @xmath98 limit the standard relations among @xmath236 , @xmath237 and @xmath238 are recovered .
= 10.0cm=10.0 cm = 10.0cm=10.0 cm for the sake of completeness we now consider the case where both @xmath92 and @xmath98 are turned on .
the ir brane propagator has the following limits for arbitrary values of @xmath92 , @xmath98 and @xmath239 : [ largep ] g_p(l , l ) & ~ & - p k e^-k l + [ smallp ] g_p(l , l ) & ~ & - g_5 ^ 2 ^-1 p k e^-k l . we see that , as required by locality , when @xmath171 , the ir propagator depends only on @xmath98 , not on @xmath92 which is localized far away .
thus , the high energy properties are the same as when @xmath170 , as discussed in the previous subsection : for sufficiently large momenta , ir observers see a single four dimensional state of mass @xmath217 and coupling @xmath222 .
the low energy limit is more complicated and depends on the relative size of the various localized parameters .
first , inspecting the coefficient of the @xmath240 term in eq .
( [ smallp ] ) , we see that when @xmath241 there is a qualitative difference depending on the size of @xmath242 as defined in eq .
( [ gamma ] ) . when @xmath243 , as is the case when either brane localized kinetic term vanishes , the propagator is just [ zeromodepropagator ] g_p(l , l ) ~- , where @xmath244 was defined in eq .
( [ zeromodecoupling ] ) .
the physics is qualitatively similar to the case where @xmath100 .
the low energy physics is considerably richer when both localized kinetic terms are present and @xmath185 . in this case
we have to be more careful about the size of the higgs vev relative to other scales .
in fact , compared to the analysis of subsection [ subsec : twobranes ] , the introduction of the higgs vev introduces two new relevant scales : @xmath245 and @xmath246 , where @xmath247 and @xmath101 is given in eq .
( [ zeromodecoupling ] ) .
these two scales can be very different if there is a hierarchy between @xmath92 and @xmath98 . for simplicity
, we will analyze here the case where @xmath248 .
we find three regions according to the size of @xmath249 . when @xmath250 there are no light states in the theory and we get an effective contact term interaction @xmath251 .
when @xmath252 we find g_p(l ,
l ) ~\ { cc - & g_ir p k e^-kl + - & p g_ir , . which shows that , below the compactification scale @xmath177 , there is a single state with mass @xmath245 , and with coupling @xmath222 . finally , when @xmath253 , we conclude from g_p(l , l ) ~\ { cc - & k e^-kl p k e^-kl + - & p k e^-kl . . that there are two light states below @xmath177
the lightest one has mass @xmath254 and coupling @xmath101 .
the heavier one has a mass of order @xmath255 and coupling @xmath256 as given in eq .
( [ twobraneg1 ] ) .
we now turn to the electroweak theory @xmath193 , and realistic phenomenology .
we assume that the higgs and all fermions are confined to the ir brane , but the gauge fields are allowed to propagate in the bulk . for simplicity , we treat the uv brane as transparent : @xmath170 .
after ewsb , the relevant terms in the 5d lagrangian are thus , ^5_ew & = & \ { - w_mn w^mn ( 1 + 2 r_2 ( y - l ) ) - b_mn b^mn ( 1 + 2 r_1 ( y - l ) ) . + & & . - v^2 ( y - l ) } , where the @xmath257 are the three @xmath258 gauge bosons ( with bulk coupling @xmath13 ) , @xmath259 is the @xmath260 gauge boson ( with bulk coupling @xmath261 ) , and @xmath262 and @xmath263 are the respective field strength tensors . @xmath264 and @xmath265 are the ir brane gauge kinetic terms for the @xmath260 and @xmath258 gauge groups . in ref .
@xcite , which addressed the localized vev effects , but treated the brane as transparent , the situation was analyzed in the limit @xmath200 ( for which the symmetry - breaking effects of section [ sec : higgs ] can be treated as a perturbation ) by introducing the gauge rotations which diagonalize the masses in terms of the bulk couplings , @xmath266 .
then , in the basis , [ eq : gaugerotation ] w^3_= c^2 z_+ a _ & , & b_= -s^2 z_+ a_the kk towers for the photon ( @xmath112 ) and @xmath267 decouple from each other . for @xmath268 ,
this choice is not particularly convenient , because although it does decouple the kk tower of the photon and @xmath267 with respect to the bulk terms , they remain mixed together by the brane terms .
thus we consider the simplified case , @xmath269 , for which the same rotation diagonalizes both terms . while there is no reason why @xmath269 should hold , and
in fact since they renormalize differently , they will certainly differ at different energy scales , it does simply illustrate the effects of the brane kinetic terms on the phenomenological picture . when @xmath270 , the field redefinitions of eq .
( [ eq : gaugerotation ] ) diagonalize the lagrangian , resulting in , [ eq : lew5 ] l^5_ew & = & \ { - w^+_mn w_-^mn - f_mn f^mn . + & & . - z_mn z^mn - 2 v^2 ( y - l ) ( w^+_m w_-^m + z_m z^m ) } , where we have introduced @xmath271 , the 5d photon coupling
. we can now make contact with the results of section [ sec : higgs ] .
clearly the zero mode of the photon is associated with the gauge interactions we see at low energies .
the weak gauge bosons are somewhat more subtle . in section
[ sec : higgs ] we saw that for @xmath272 the first few kk modes couple a factor of order 10 ( or more for large @xmath203 ) more strongly to the ir brane than the zero mode , and have approximately equal spacing .
this suggests that for a viable phenomenological picture , it is the zero modes that should be associated with the weak gauge bosons observed in experiments , with the higher kk modes suitably heavy such that they evade current experimental limits .
however , for large @xmath167 , there is only ever one gauge boson with relevant gauge coupling to the fermions
. provided the lighter modes are sufficiently weakly coupled , they could have escaped detection up until now . thus ,
for @xmath273 we have a choice as to which mode plays the role of the observed weak bosons , at least for some range of parameters . for simplicity
, we will restrict our attention to the simpler case in which we associate the known weak interactions with the zero modes of the @xmath274 and @xmath275 fields , and leave the more exotic case in which we may actually be identifying the higher kk modes as the mediators of the weak force for future work . in the standard model
the physics predominantly depends on three parameters of the electroweak theory , @xmath276 , @xmath277 , and @xmath278 .
the accuracy of the data is such that loop - level effects are important , introducing a strong dependence on the top mass ( @xmath279 ) , and relevant dependence on the strong coupling at the @xmath267 pole ( @xmath280 ) and higgs mass ( @xmath281 ) .
the usual procedure is to use the three most precisely measured quantities , the fermi constant from muon decay ( @xmath278 ) , the mass of the @xmath267 boson ( @xmath282 ) , and the electromagnetic coupling at the @xmath267-pole ( @xmath283 ) , to fix the three tree - level parameters , and then to combine the sum of the precision data in a fit to either @xmath281 ( and @xmath279 and @xmath280 ) , or , allowing for nonstandard corrections to the weak boson self - energies ( oblique corrections ) , to fit the oblique parameters @xmath284 , @xmath285 , and @xmath286 @xcite .
this procedure is not completely appropriate whenever there are important non - oblique corrections which directly modify the fermion couplings , or which modify the self - interactions of the gauge bosons beyond what is contained in the oblique parameters .
however , it is still sensible whenever the oblique corrections capture the dominant new physics contributions . in the specific case of the rs model with bulk gauge bosons ,
one immediately encounters a problem related to non - oblique corrections .
the muon decay now proceeds through the entire tower of kk modes of the @xmath287 .
this was handled in refs .
@xcite by introducing an additional fit parameter @xmath288 which measures the amount of contribution to @xmath278 from the higher kk modes relative to the zero mode .
however , as we will see below , we will find it more convenient to avoid introducing @xmath288 , and instead define effective parameters , @xmath284 , @xmath285 , and @xmath286 , which in practice take into account all relevant oblique and non - oblique corrections necessary to describe the z - pole precision electroweak observables and the w mass . at tree level ,
our model point is specified in terms of the lagrangian in eq .
( [ eq : lew5 ] ) by the six quantities @xmath289 , @xmath290 , @xmath207 , @xmath167 , @xmath4 , and @xmath3
. @xmath3 can be taken to be the 5d planck scale , and thus may be thought of as setting the overall dimensionful scale .
we must identify the 4d quantities @xmath276 , @xmath277 , and effective 4d higgs vev @xmath236 from the measurements of @xmath283 , @xmath282 , and @xmath278 .
this will determine three of our six parameters in terms of the other three .
we represent this freedom by treating @xmath167 and @xmath4 as the free parameters , and use the input data to specify @xmath289 , @xmath290 , and @xmath207 in terms of them . applying the results of section [ sec : higgs ]
, we derive the effective lagrangian for the zero modes , - w^+ _ w_-^ - z _ z^ - f _ f^ - m_w^2 w^+_w_-^- z_z^ + + f_w ( ^t_+ w_^+ + ^t_- w_^- ) + f_z ^(t_3 - s^2 q ) z_+ f_a ^q a _
, where @xmath291 , @xmath292 and @xmath293 are the @xmath287 , @xmath267 and photon zero mode wave functions evaluated at the ir brane , @xmath7 , @xmath294 stands for the brane - localized fermions , @xmath295 , @xmath296 are the relevant weak isospin matrices and @xmath297 is the electric charge .
note also that here and below , 4d indices are raised and lowered by the canonically normalized lorentz metric , @xmath298 .
the quantities @xmath299 and @xmath300 are determined numerically for a given choice of @xmath207 , @xmath289 , @xmath290 , @xmath167 , @xmath4 , and @xmath3 , as in fig .
[ fig : massesrv ] . note that this implies that these quantities have no direct relation to those which one
would have expected in the sm . since the photon experiences no symmetry breaking , the zero mode wave function is flat , and thus we can identify the 4d electromagnetic coupling @xmath276 with , e f_a = .
this allows us to reproduce the electromagnetic coupling @xmath301 .
the muon decay constant @xmath278 should be identified with the ( full ) @xmath287 boson at zero momentum transfer .
thus , it implicitly contains the effects of the entire kk tower . from eq .
( [ smallponebrane ] ) we see that at zero momentum transfer , the tree - level gauge boson propagator reduces to the very simple form , [ gfermi ] - g(p^2 = 0;l , l ) & = & + _ n 0 = = 4 g _ , which implies that fixing @xmath278 to its experimentally measured value of @xmath302 ge@xmath303 determines @xmath304 gev .
this result for the propagator is exact at @xmath305 and includes the sum of the contributions of all the kk modes of the charged weak gauge bosons .
the final quantity we would like to use to determine the input parameters is the @xmath267 mass .
having already determined @xmath289 and @xmath207 ( as functions of @xmath167 and @xmath4 ) , we can accomplish this by adjusting @xmath290 to the value which produces the correct @xmath300 .
we find this value numerically as in section [ sec : higgs ] , adjusting the quantity @xmath306 until the mass of the lowest mode is the @xmath267 boson mass measured in experiments , @xmath307 .
the advantage of matching directly to @xmath283 , @xmath282 , and @xmath278 is that , with appropriate definitions , we can carry over the established machinery for electroweak fits , and use the usual bounds on @xmath284 , @xmath285 , and @xmath286 to directly constrain the rs model , without refitting the data ( see below ) .
it remains to determine the effect on precision observables as a function of @xmath167 and @xmath4 , which allows us to determine the region of @xmath167 and @xmath4 consistent with experimental data .
since the gauge boson interactions with the fermions are affected in a universal way , most of the effects may be captured by effective @xmath284 , @xmath285 and @xmath286 parameters .
the primary remaining non - oblique effect is the coupling strengths of the self - interactions of the zero mode @xmath287 s and @xmath267 .
these depend on integrals over @xmath17 of three or four gauge boson wave functions .
thus , while they are determined from the @xmath308 , they are not related to them in a simple way . however , as these interactions are only mildly constrained by lep-2 data @xcite , the resulting constraints are very weak
. we will find it convenient to rescale the zero mode wave functions , f_a & = & _ a + f_z & = & _ z = _ z + f_w & = & _ w = g _ w for which the deviation from the tree - level sm predictions appears as a deviation of @xmath309 from unity .
note that given our matching prescription we have @xmath310 .
we have defined @xmath10 and @xmath311 with respect to @xmath290 and @xmath312 , the sine and cosine of the gauge rotation angle defined above eq .
( [ eq : gaugerotation ] ) .
this angle is different from the weak mixing angle implicit from our choice of input parameters , s_0 ^ 2 c_0 ^ 2 & = & . in order to make contact with the oblique parameters , we further rescale the fermion interactions to unity , resulting in the self - energy part of the effective lagrangian taking the form , & & - w^+ _ w_-^ - z _ z^ - f _ f^ + & & - w^+_w_-^- z_z^ .
[ eq : weakzeromodes ] the tree - level contributions to @xmath284 , @xmath285 , and @xmath286 can be simply determined by matching the effective lagrangian for the zero modes , eq .
( [ eq : weakzeromodes ] ) to a generic effective lagrangian including a parameterization of the oblique corrections , [ eq : leweff ] l & = & - ( 1 - ^_ww ) w^+ _ w_-^ - ( 1 - ^_zz ) z _ z^ + & & - ( 1 - ^ _ ) f _ f^ - ^_z f^ z _ - ( ^2 + _ ww(0 ) ) w^+_w_-^ + & & - ( ^2 + _ zz(0 ) ) z_z^ + & & + ( ^t_+ w_^+ + ^t_- w_^- ) + ^(t_3 - s^2 q ) z_+ ^q a _ , where @xmath313 plays the role of the higgs vev extracted from @xmath278 , eq . ( [ gfermi ] )
we have followed ref .
@xcite and written these in standard self - energy notation , despite the fact that the 5d contributions are in fact tree level . the ads / cft correspondance suggests that the 5d theory is dual to a 4d theory of walking technicolor @xcite .
thus , the tree level 5d contributions correspond to the loop - level corrections in the dual theory @xcite . matching eqs .
( [ eq : weakzeromodes ] ) and ( [ eq : leweff ] ) , we obtain , & = & ^_11 = ( 1 - ) + & = & ^_33 = ( 1 - ) + ^_3q & = & ^_qq = 0 + & = & _ 11(0 ) = - ^2 + & = & _ 33(0 ) = - ^2 , where the @xmath314 refer to the fact that these are only the tree - level self - energy contributions from the extra - dimensional effects
. the full @xmath315 will also include both loop - level sm contributions , and also loop - level extra dimensional ones .
the usual oblique parameters @xmath284 , @xmath285 , and @xmath286 are defined in terms of the self - energies by , s & & 16 ( ^_33 - ^_3q ) + t & & ( _ 11(0 ) - _ 33(0 ) ) + u & & 16 ( ^_11 - ^_33 ) , where @xmath282 is the @xmath267 mass from the sm relations in terms of @xmath276 , @xmath316 and @xmath236 .
given our matching conditions , @xmath317 .
thus , we derive the tree level 5d contributions to the usual @xmath284 , @xmath285 , and @xmath286 , [ eqs ] & = & ( 1 - ) + [ eqt ] & = & ( - ) + [ equ ] & = & where once again , the barred quantities indicate that these are only the contributions from the tree level extra - dimensional effects . at the end of subsection 4.2
, we found approximate expressions for the mass and coupling to ir fields of the zero mode gauge bosons , that are valid when @xmath318 ( see eqs .
( [ m0firstorder ] ) and ( [ g0firstorder ] ) ) , where @xmath13 stands for the appropriate gauge coupling ( @xmath13 for the charged @xmath287 bosons and @xmath319 for the @xmath267 )
. it will be useful to have analytical expression for @xmath284 , @xmath285 and @xmath286 to the same order . in our approach
we determine the sine of the rotation angle between the gauge and mass eigenbasis , @xmath290 , by requiring that the @xmath267 mass is reproduced .
this angle is in general different from the standard model value , but when @xmath320 we expect them to differ by order @xmath321 . using eq .
( [ m0firstorder ] ) with the general @xmath322 of section 4 replaced by @xmath323 we have m_z & = & ( 1 - + ) , where @xmath324 was defined in eq .
( [ eta ] ) and @xmath325 .
the above relation gives [ deltasine ] s = s_0 ( 1 - + ) , where = .
we can also write @xmath326 , @xmath327 and @xmath328 as : [ mwapprox ] m_w & = & = , + [ fzapprox ] _ z ( ) & = & 1 - 2 + ( ^2 ) , + [ fwapprox ] _ w ( ) & = & 1 - 2 c_0 ^ 2 + ( ^2 ) . plugging these expressions in eqs .
( [ eqs])-([equ ] ) , we find & = & - + ( ^2 ) , + & = & - + ( ^2 ) , + & = & ( ^2 ) . due to the presence of non - oblique corrections associated with heavy kk modes , the @xmath284 , @xmath285 and @xmath286 parameters defined above are not sufficient to describe the precision electroweak observables .
however , as observed in ref .
@xcite , the z - pole precision observables may be properly described by the introduction of an effective parameter , @xmath329 , which includes the non - oblique corrections to the weak mixing angle , [ newdeltasine ] s^2-s_0 ^ 2 = ( s - s^2 c^2 t_eff ) , where @xmath329 is given by t_eff = t + t and t = - = - _ n 0 ( ) ( ) = - ( - 1 ) = - + ( ^2 ) , where @xmath330 is the non - oblique contribution to the muon decay constant due to the exchange of kk modes . with this definition of @xmath329
, we can recover eq .
( [ deltasine ] ) , using eqs .
( [ mwapprox ] ) and ( [ fwapprox ] ) .
note that in the rs theory , the bare parameter @xmath290 coincides with the kennedy - lynn running coupling @xmath331 @xcite .
the non - oblique corrections to @xmath278 affect also the expression of @xmath332 , but the relative coefficient between the oblique and non - oblique corrections encoded in @xmath329 is different from the one appearing in @xmath333 .
therefore , to properly parameterize @xmath334 in terms of effective parameters @xmath335 , @xmath329 and @xmath336 , [ newmw ] - c_0 ^ 2 = ( - s_eff + c^2 t_eff + u_eff ) , one can introduce u_eff = u - 4 s^2 t .
observe that the above expressions , eqs .
( [ newdeltasine ] ) and ( [ newmw ] ) reproduce eqs .
( [ deltasine ] ) and ( [ mwapprox ] ) , respectively . in terms of the effective parameters
@xmath337 , @xmath329 and @xmath336 differs from the one presented in the appendix of ref .
@xcite by the additional contribution from @xmath336 . ]
the above parameterization serves to describe all z - pole observables as well as the @xmath287 mass . following ref .
@xcite , we shall use a fit to these observables to place limits upon the free parameters of our theory . in terms of the expansion in @xmath321 ,
we find that the tree - level five - dimensional contributions , including the non - oblique corrections to @xmath278 , can be given by [ seff ] _ eff & = & - + - 366 , + [ teff ] _ eff & = & - + - 258 , + [ ueff ] _ eff & = & + 183 , we see that @xmath338 and @xmath339 are negative while @xmath340 is positive ( and not small ) . before closing this section ,
let us stress that , following ref .
@xcite , we are ignoring loop - level contributions to the precision electroweak observables , including those induced by a potentially light radion .
the computations of ref .
@xcite indicate that , unless there is strong mixing between the radion and the higgs field @xcite , these corrections are small and thus may be typically neglected .
we have also checked that for the parameters that give a good description to the data , the effect of the tree - level neutral kk mode interactions in the z - pole observables is unobservably small . the sm with a light higgs , with mass of about 120 gev , provides an excellent description of the precision electroweak observables measured at the lep , tevatron and sld colliders . while @xmath341 plays a key role in the determination of the above quoted higgs mass range
, there is a discrepancy of more than about 3@xmath342 s between the value of the weak mixing angle extracted from the lepton asymmetries and the one extracted from the hadron asymmetries .
a good fit to the hadron asymmetries , which are dominated by the forward - backward asymmetry of the @xmath116-quark , @xmath343 , tends to require much larger values of the higgs mass . on the other hand ,
a good fit to the lepton asymmetries leads to the preference of a higgs mass value of about 50 gev @xcite , well below the present direct search limit .
in fact , the internal consistency of the precision electroweak data is dramatically improved by the exclusion of the hadron asymmetries : while the fit to the combined data from these colliders has a confidence of about @xmath344 @xcite , it increases to more than 30@xmath345 once the hadronic asymmetries are excluded .
this suggests that , in order to improve the consistency of the precision electroweak data either one must postulate an error in the experimental determination of @xmath343 and invoke new physics to raise the higgs mass above the direct search limits @xcite , or one must introduce new physics that directly modifies the couplings of the bottom quark in order to restore the consistency of the measurements @xcite . the rs model does not discriminate between the leptons and quarks , and thus this second option is denied to us .
below , we will consider the fits to precision electroweak parameters @xmath284 , @xmath285 , and @xmath286 both with and without @xmath343 . while we do not advocate ignoring @xmath343 in general , it is interesting to see that the rs model with bulk gauge bosons can in fact lead to a remarkable improvement of the description of the electroweak observables once the hadron asymmetries are ignored , for a higgs mass @xmath281 well above the experimental limits .
we use the @xmath284,@xmath285 , and @xmath286 fits of @xcite , which , for @xmath346 gev yield ( removing @xmath343 from the fit ) , [ eq : stulep ] s & = & -0.14 0.12 + t & = & -0.08 0.13 + u & = & 0.20 0.14 , or ( keeping @xmath343 in the fit ) , [ eq : stuhad ] s & = & 0.00 0.11 + t & = & -0.03 0.13 + u & = & 0.27 0.14 .
we compute the full @xmath337 , @xmath329 , and @xmath336 as the sum of the extra dimensional contributions , eqs .
( [ seff])([ueff ] ) , and also the contributions from the higgs @xcite , s_h & & ( ) + t_h & & - ( ) + u_h & & 0 , where @xmath347 is a reference higgs mass , chosen for the fits of eqs .
( [ eq : stulep ] ) and ( [ eq : stuhad ] ) to be 113 gev .
we scan the parameter space of @xmath4 and @xmath167 for two fixed values of @xmath348 and @xmath349 gev , and identify the regions of @xmath167 and @xmath4 which are consistent with eq .
( [ eq : stulep ] ) or ( [ eq : stuhad ] ) at @xmath350 , @xmath351 , and @xmath352 .
we define consistency at the @xmath353 level to mean that all three oblique parameters predicted by the rs model are within @xmath354 error intervals of the central fitted values .
we fix the top quark mass to its experimental mean value , @xmath346 gev .
the one - sigma variations of the top quark mass lead to corrections of about @xmath355 , @xmath356 and @xmath357 , which are smaller than the one @xmath342 errors on @xmath284 , @xmath285 and @xmath286 obtained from the fit to the data and do not affect our results in a relevant way .
when comparing how well rs fits the data compared to the sm , it is important to remember that the sm fails this analysis for any region of parameters at @xmath350 , and agrees at roughly @xmath351 for higgs masses above the lep limit and below about 200 gev .
furthermore , for the same range of higgs mass parameters , removing @xmath343 does not improve the agreement between the experimental data and the standard model predictions .
= 10.0cm=10.0 cm = 10.0cm=10.0 cm the results for @xmath348 gev are shown in fig .
[ fig : fitmh115 ] , plotted in the plane of @xmath167 and @xmath358 , the mass of the first kk mode of the photon .
generally , the masses of the first kk modes of the @xmath287 and @xmath267 bosons will be somewhat larger , because of the electroweak symmetry - breaking effects . in the limit @xmath359 , we find agreement with the limit obtained in ref .
@xcite , @xmath360 tev . as @xmath167 increases , there are two different effects . for small non - zero @xmath167 ,
the overall fit to the data remains roughly the same as the @xmath361 case ( in other words , the properties of the zero modes do not change much ) , but the properties of the first kk mode and its coupling to fermions are quite radically affected , with both dropping .
, as demanded by the fit to the data , the properties of the kk mode masses and couplings can be read from fig . 1 .
] this results in the kk photon being much lighter than one would have assumed for zero @xmath167 , and , for moderate values of @xmath167 , its coupling is comparable to that of the zero mode photon . for @xmath362 and @xmath358 on the order of 3.5 tev , the fit to the entire data set is roughly as good as the sm itself , and the fit to the data without the hadron asymmetries is actually consistent within @xmath350 . indeed , for the large region of parameters denoted by light squares in fig .
[ fig : fitmh115 ] , we obtain an improvement of the fit to the data with respect to the standard model , with @xmath358 of order of a few tev . in fig .
[ fig : fitmh200 ] we show the same analysis , but with the higgs mass fixed to @xmath363 gev .
we see some small variation of the allowed regions , but by relatively small amounts , indicating that the rs fit does not prefer a light higgs , but fits about equally well to higgs masses as large as 200 gev .
the analysis for larger @xmath281 is straight - forward , and shows somewhat worse agreement than for @xmath364 gev .
= 10.0cm=10.0 cm = 10.0cm=10.0 cm our results have very important implications for the rs model with gauge fields in the bulk .
the preferred parameter space , for moderate @xmath365 , has a first kk photon with fermionic couplings of the order of the zero mode couplings , and a mass of a few tev , most likely within reach of the lhc experiments .
the rs model is an interesting construction , with a novel solution to the hierarchy problem . while the simplest versions have gauge fields confined to the ir brane , there are interesting motivations to allow them to propagate into the extra dimension , including grand unification and a relation to a class of walking technicolor theories through the ads / cft correspondance . in that case ,
the lowest kk modes of the gauge bosons are usually very strongly coupled to fermions on the ir brane , with the lowest mass given by a factor of a few times @xmath366 , at the tev scale .
precision electroweak observables are challenging to fit in this framework , owing to corrections to the zero mode @xmath287 and @xmath267 boson wave functions induced by the presence of the higgs boson in the infra - red brane . in the absence of local gauge kinetic terms ,
consistency of the fit to the precision electroweak observables at the 2@xmath342 level , as defined in the text , demands masses of the kk modes of the gauge bosons larger than about 27 tev , beyond the reach of the next generation of collider experiments .
however , the inclusion of brane kinetic terms is helpful in lowering this bound . for moderately large values of @xmath167
, 2-@xmath342 consistency may be obtained for kk gauge boson masses of the order of a few tev . furthermore ,
if the hadronic asymmetries are excluded from the fit , the bulk gauge bosons may lead to a consistent fit to the data at the 1 @xmath342 level for a lightest kk mode of the order of a few tev and higgs masses in the range 115 - 200 gev .
these light kk modes have couplings of order of the zero mode coupling , opening new exciting possibilities at the lhc .
+ * acknowledgements * + + the authors are grateful for conversations with g. burdman .
work at anl is supported in part by the us doe , div . of hep , contract w-31 - 109-eng-38 .
fermilab is operated by universities research association inc . under contract
de - ac02 - 76ch02000 with the doe .
+ * note added : * related work on this subject appeared at a similar time in ref .
we have explored the case @xmath367 proposed in this reference and we found that the bounds from @xmath337 , @xmath329 , and @xmath336 are more constraining than those based on the kk mode contributions to the muon decay constant alone .
n. arkani - hamed , s. dimopoulos and g. r. dvali , phys .
b * 429 * , 263 ( 1998 ) [ arxiv : hep - ph/9803315 ] ; i. antoniadis , n. arkani - hamed , s. dimopoulos and g. r. dvali , phys .
b * 436 * , 257 ( 1998 ) [ arxiv : hep - ph/9804398 ] .
l. randall and r. sundrum , phys .
lett . * 83 * , 3370 ( 1999 ) [ arxiv : hep - ph/9905221 ] .
h. davoudiasl , j. l. hewett and t. g. rizzo , phys .
b * 473 * , 43 ( 2000 ) [ arxiv : hep - ph/9911262 ] . c. csaki , j. erlich and j. terning , phys
. rev .
d * 66 * , 064021 ( 2002 ) [ arxiv : hep - ph/0203034 ] .
g. burdman , phys .
d * 66 * , 076003 ( 2002 ) [ arxiv : hep - ph/0205329 ] . s. j. huber and q. shafi , phys .
d * 63 * , 045010 ( 2001 ) [ arxiv : hep - ph/0005286 ] ; s. j. huber , c. a. lee and q. shafi , phys .
b * 531 * , 112 ( 2002 ) [ arxiv : hep - ph/0111465 ] .
g. r. dvali , g. gabadadze and m. a. shifman , phys .
b * 497 * , 271 ( 2001 ) [ arxiv : hep - th/0010071 ] .
h. georgi , a. k. grant and g. hailu , phys .
b * 506 * , 207 ( 2001 ) [ arxiv : hep - ph/0012379 ] .
m. carena , t. m. p. tait and c. e. m. wagner , acta physica polonica , b33 , 2355 ( 2002 ) arxiv : hep - ph/0207056 .
m. carena , a. delgado , j. lykken , s. pokorski , m. quiros and c. e. wagner , nucl .
b * 609 * , 499 ( 2001 ) [ arxiv : hep - ph/0102172 ] .
g. r. dvali , g. gabadadze , m. kolanovic and f. nitti , phys .
d * 64 * , 084004 ( 2001 ) [ arxiv : hep - ph/0102216 ] . l. randall and m. d. schwartz , jhep * 0111 * , 003 ( 2001 ) [ arxiv : hep - th/0108114 ] ; phys .
lett . * 88 * , 081801 ( 2002 ) [ arxiv : hep - th/0108115 ] ; k. w. choi , h. d. kim and y. w. kim , arxiv : hep - ph/0202257 ; k. agashe , a. delgado and r. sundrum , nucl .
b * 643 * , 172 ( 2002 ) [ arxiv : hep - ph/0206099 ] ; a. lewandowski , m. j. may and r. sundrum , arxiv : hep - th/0209050 . k. agashe , a. delgado and r. sundrum , arxiv : hep - ph/0212028
. w. d. goldberger and i. z. rothstein , arxiv : hep - th/0204160 ; arxiv : hep - th/0208060 .
e. pontn and e. poppitz , jhep * 0106 * , 019 ( 2001 ) [ arxiv : hep - ph/0105021 ] .
a. pomarol , phys .
b * 486 * , 153 ( 2000 ) [ arxiv : hep - ph/9911294 ] .
m. e. peskin and t. takeuchi , phys .
d * 46 * , 381 ( 1992 ) .
t. g. rizzo and j. d. wells , phys .
d * 61 * , 016007 ( 2000 ) [ arxiv : hep - ph/9906234 ] .
n. arkani - hamed , m. porrati and l. randall , jhep * 0108 * , 017 ( 2001 ) [ arxiv : hep - th/0012148 ] .
d. c. kennedy and b. w. lynn , nucl .
b * 322 * , 1 ( 1989 ) .
g. altarelli , f. caravaglios , g. f. giudice , p. gambino and g. ridolfi , jhep * 0106 * , 018 ( 2001 ) [ arxiv : hep - ph/0106029 ] . c. csaki , m. l. graesser and g. d. kribs , phys .
d * 63 * , 065002 ( 2001 ) [ arxiv : hep - th/0008151 ] .
g. f. giudice , r. rattazzi and j. d. wells , nucl .
b * 595 * , 250 ( 2001 ) [ arxiv : hep - ph/0002178 ] .
m. s. chanowitz , phys .
lett . * 87 * , 231802 ( 2001 ) [ arxiv : hep - ph/0104024 ] .
d. choudhury , t. m. p. tait and c. e. m. wagner , phys .
d * 65 * , 053002 ( 2002 ) [ arxiv : hep - ph/0109097 ] .
p. langacker , arxiv : hep - ph/0211065 ; lep electroweak working group , http://lephiggs.web.cern.ch / lepewwg/. | we examine localized kinetic terms for gauge fields which can propagate into compact , warped extra dimensions .
we show that these terms can have a relevant impact on the values of the kaluza - klein ( kk ) gauge field masses , wave functions , and couplings to brane and bulk matter .
the resulting phenomenological implications are discussed . in particular , we show that the presence of opaque branes , with non - vanishing brane - localized gauge kinetic terms , allow much lower values of the lightest kk mode than in the case of transparent branes .
moreover , we show that if the large discrepancies among the different determinations of the weak mixing angle would be solved in favor of the value obtained from the lepton asymmetries , bulk electroweak gauge fields in warped - extra dimensions may lead to an improvement of the agreement of the fit to the electroweak precision data for a higgs mass of the order of the weak scale and a mass of the first kk gauge boson excitation of a few tev , most likely within reach of the lhc . |
uplink co - ordinated multi - point ( comp ) is a promising technique for increasing the capacity of 4 g networks @xcite .
the uplink is gaining increasing attention due to the dramatic increase of user - generated data in the form of photos , videos and file - sharing . in practice ,
sharing of uplink received signals across cells is limited by backhaul bandwidth .
in addition , the receiver _ aperture _ , or number of signals from antennas at neighboring cell sites that can be processed at a particular cell , may be limited due to hardware constraints .
several approaches for signal sharing and combining have been proposed for uplink comp .
a performance analysis of different combining methods with different backhaul bandwidth requirements is presented in @xcite . to reduce
the amount of sharing , dynamic clustering of cells has been proposed , e.g. , in @xcite .
in addition to limiting the amount of information that can be shared across cells , the limited backhaul bandwidth also introduces latency , which is addressed in @xcite .
previous work on comp has generally assumed that the set of cells that share information is fixed _ a priori _ by the topology and the bandwidth of the backhaul links . in this paper
we relax this assumption and consider the problem of _ optimizing _ _ sets _ of _ helper _ cells that pass along their uplink signals to other cells .
( a cell can both share its signals as a helper cell and receive signals , or help from other cells . )
we account for architectural constraints that limit the set of potential helper cells ( which differs from cell to cell ) , backhaul constraints that limit the number of cells to which a helper cell can send its received signals ( _ egress _ constraint ) , and hardware constraints at the cell site , which may limit the number of incoming signals the cell can combine ( _ ingress _ constraint ) .
related work in @xcite has considered the problem of mmse receiver estimation under compression and backhaul constraints .
we focus here instead on a simpler sharing formulation that introduces explicit constraints on egress bandwidth , to arrive at a convex weighted sum rate formulation with guaranteed convergence . imposing these explicit egress bandwidth and ingress aperture constraints
captures important architectural limitations in centralized or distributed co - operative networks .
our problem is then to maximize a sum rate objective subject to these architectural constraints .
since the cell sites are assumed to have multiple antennas , we refer to this problem as _ multi - antenna aperture selection ( maas ) _ for joint reception ( jr)-comp ( see also @xcite ) .
the optimization of sets of helper cells is an integer program . assuming max - ratio combining of received signals and relaxing the integer constraints
, the optimization problem becomes convex .
we present a distributed algorithm in which each helper cell announces an _ egress price _ , indicating the demand for its signals to help other cells , and each assisted cell computes an _ ingress price _ , indicating the potential improvement from adding a helper cell .
the prices are used to compute an ordering of users / cells , which is then used to allocate the helper antennas across assisted cells .
this assignment is iterated with updates for the prices .
we refer to this algorithm as _ liquidmaas _ due to its ability to flexibly allocate help based on network load conditions , and show that it converges to the optimal allocation of helper cells across the network .
numerical results are presented that show that the algorithm converges in relatively few iterations even for a system with a large number of cells .
furthermore , the gains relative to an _ a priori _ fixed allocation of helper cells can be substantial .
consider a network of @xmath0 cells , each with multiple antennas .
each cell @xmath1 serves a set of users @xmath2 .
the uplink signal from each user @xmath3 is typically strongest at its serving cell ; however , the signal could also be received with significant strength at other cells depending on the user location and network topology .
we assume the antenna - combined signal for a particular user @xmath3 can be shared with other cells , and is , of course , provided to that user s serving cell for comp combining .
figure ( [ cell - topology ] ) shows an example of a network with uplink data sharing .
there are four cells with seven users , where @xmath4 , @xmath5 , @xmath6 , and @xmath7 .
let @xmath8 denote the set of cells with data that can be requested by the cell serving user @xmath9 ( _ ingress _ neighborhood ) .
that typically corresponds to the set of cells where the user s sinr is above a minimum threshold value ( usually -10 db ) .
that is , @xmath10 , where @xmath11 is the sinr at cell @xmath12 for a user @xmath9 in cell @xmath1 ( i.e. , @xmath13 ) , and @xmath14 is the serving cell for user @xmath9 .
note that @xmath8 may not be the actual set of helper cells for user @xmath9 due to backhaul and aperture constraints . also shown in
figure ( [ cell - topology ] ) are the possible sharing variables along with the inter - connect .
note that cells 1 and 3 , and cells 2 and 4 do not exchange information .
the sharing variable @xmath15 indicates whether or not cell @xmath12 shares data for user @xmath9 in cell @xmath1 over the backhaul link @xmath16 .
for the example shown in figure ( [ cell - topology ] ) , the local neighborhood sets for the different users are @xmath17 the maximum receiver aperture size , or number of helper cells , for user @xmath9 is denoted as @xmath18 , and is the number of cells in the neighborhood @xmath19 . for a cell @xmath12 , its _ egress _ neighborhood to cell @xmath1 is denoted by @xmath20 , and is the set of users in cell @xmath1 it can potentially help , ignoring backhaul and aperture constraints .
that is : @xmath21 for the example in figure ( [ cell - topology ] ) , @xmath22 .
the egress neighborhoods are : @xmath23 note that if there is no possibility of sharing between two cells , then the corresponding egress set is null , _
i.e. , _ @xmath24 .
the notation is summarized in table [ table ] .
.variable definitions [ cols="^,^",options="header " , ] [ table2 ] users were allocated bandwidth equally based on the number of users connected to a given cell .
a user s sinr to all cells in the network was calculated based on this bandwidth allocation , and open - loop fractional power control @xcite was assumed .
the update for @xmath25 and @xmath26 used a step - size of @xmath27 .
four algorithms were evaluated : _ ( 1 ) no comp , ( 2 ) maas without egress constraints , ( 3 ) maas with a randomized egress bandwidth control mechanism , and ( 4 ) liquidmaas_. in case _
( 2 ) _ , each cell requests help based on their aperture limit , and helper cells grant any requested help . in _ ( 3 ) _ , helper cells randomly grant requests for help until their egress constraint limit is reached .
the ingress aperture constraint limit was set to @xmath28 , so that each user being gets a maximum of @xmath29 helper cells data .
the egress bandwidth limit @xmath30 was varied to see the effect on performance and convergence .
the top part of figure ( [ fig2 ] ) shows the convergence behavior of all cells egress bandwidth demand ( for @xmath31 ) , from which we observe that the algorithm converges in @xmath32 iterations .
the convergence time is dependent on the step - size for updating the prices .
it was observed that if the aperture and egress limits are similar , the step - size should be reduced to enable smoother convergence .
the bottom part of figure ( [ fig2 ] ) shows the distribution of weighted sum rate ( wsr ) gain obtained by the various comp approaches .
this shows that liquidmaas gives substantial gain over no comp and also over the randomized egress bandwidth control strategy .
figure ( [ fig3 ] ) shows the distribution of the egress bandwidth and ingress apertures at convergence .
we observe that the liquidmaas algorithm maintains tight control over the egress bandwidth , and adapts the ingress aperture to ensure that the constraints are met .
finally , figure ( [ fig4 ] ) shows the average wsr gain ( over no comp ) as a function of the egress limit ( @xmath30 ) .
liquidmaas gives significant gains across the range of egress limits compared to the randomized strategy , and converges to the unconstrained maas solution for larger values of @xmath30 .
.,height=359 ] .,height=359 ]
in this paper , we presented an approach for joint algorithm and architecture optimization for co - operative communication networks .
we considered the problem of uplink joint reception comp with backhaul bandwidth constraints , and showed that the resulting network utility maximization problem is convex . a distributed algorithm to solve this problem
was presented , which consisted of local nodes computing their desired helper requests , followed by helper cells computing and updating an egress price for their bandwidth . at convergence ,
these iterations give a set of connections between helper and recipient cells , such that egress and ingress constraints are met at all cells , and overall network utility is maximized .
simulation results for a 57-cell system were presented to illustrate the efficacy of the approach , and show that the distributed dual decomposition - based gradient algorithm converges within tens of iterations .
the authors acknowledge useful discussions with r. agrawal , m. r. raghavendra , p. rasky , and c. schmidt .
k. m. karakayli , g. j. foschini , and r. a. valenzuela , `` network coordination for spectrally efficient communications in cellular systems , '' _ ieee wireless commun .
_ , vol . 13 , no .
4 , pp . 5661 , aug 2006 .
c. liao , h. huang , `` performance evaluation of intra - site coordinated multi - point transmission with inter - cell phase information in 3gpp lte - a , '' _ proc .
ieee wireless personal multimedia communications ( wpmc ) _ , pp .
241 - 245 , 2012 .
a. papadogiannis , d. gesbert , and e. hardouin , `` a dynamic clustering approach in wireless networks with multi - cell cooperative processing , '' _ proc .
ieee international conference on communications , ( icc ) _ , pp .
4033 - 4037 , may 2008 .
y. zhou and w. yu , optimized beamforming and backhaul compression for uplink mimo cloud radio access networks , " _ ieee journal on selected areas in communications , special issue on 5 g wireless communication systems , _ vol .
32 , no . 6 , pp.1295 - 1307 , june 2014 .
k. zeineddine , m. l. honig , s. nagaraj and p. j. fleming , `` antenna selection for uplink comp in dense small - cell clusters , '' _ proc .
ieee signal processing advances in wireless communications ( spawc ) _ , pp . 81 - 85 , 2013 | we address the problem of uplink co - operative reception with constraints on both backhaul bandwidth and the receiver aperture , or number of antenna signals that can be processed .
the problem is cast as a network utility ( weighted sum rate ) maximization subject to computational complexity and architectural bandwidth sharing constraints .
we show that a relaxed version of the problem is convex , and can be solved via a dual - decomposition .
the proposed solution is distributed in that each cell broadcasts a set of _ demand prices _ based on the data sharing requests they receive .
given the demand prices , the algorithm determines an antenna / cell ordering and antenna - selection for each scheduled user in a cell .
this algorithm , referred to as _
liquidmaas _ , iterates between the preceding two steps .
simulations of realistic network scenarios show that the algorithm exhibits fast convergence even for systems with large number of cells . |
this paper is part of a larger project which develops various aspects of de rham and hodge cohomology theory for the natural class of ` iterated edge ' metrics on smoothly stratified spaces . in the course of this , we develop new analytic tools adapted to this singular geometry .
+ stratified spaces arise naturally even when one is primarily interested in smooth objects . for example
, whitney showed that if @xmath1 is the set of common zeros of a finite family of polynomials , then the singular part of @xmath2 @xmath3 is again algebraic and of dimension strictly less than that of @xmath4 this gives a natural filtration @xmath5 by algebraic subvarieties , or alternately a decomposition of @xmath1 into a union of analytic manifolds : @xmath6 where these different ` strata ' fit together in a precise way .
there has been extensive investigation of this notion of stratified spaces .
this class includes every analytic variety ( whitney ) , subanalytic set ( verdier ) , as well as the generalization of this last class to the class of definable sets in an arbitrary order - minimal structure ( loi ) .
stratified spaces also include polyhedra , orbit spaces of many group actions on manifolds , and mapping cylinders of maps between manifolds .
these spaces , the definition of which is recalled below , thus arise as ubiquitously as smooth manifolds .
nonetheless , many basic aspects of geometric analysis on these spaces have not yet been developed in the literature , and this paper should be regarded as one step toward this goal . a fundamental problem is to understand the de rham and hodge theory of these spaces .
the seminal work on this was carried out by cheeger in the early 1980 s @xcite .
if @xmath1 is a compact stratified pseudomanifold and @xmath7 its top - dimensional regular part , and if @xmath8 is an ` iterated incomplete edge ' ( which we abbreviate as @xmath9 ) metric on @xmath7 , then there are two natural de rham complexes of @xmath0 forms : the first uses the minimal extension of the exterior derivative , with domain @xmath10 while the second uses the maximal extension of @xmath11 , with domain @xmath12 both complexes have finite dimensional cohomology and cheeger ( * ? ? ?
* theorem 6.1 ) showed that the latter is dual to the topological lower middle perversity intersection homology group @xmath13 of goresky - macpherson .
it follows that the cohomology of the minimal domain @xmath0 complex is the dual of the upper middle perversity groups @xmath14 there is a special class of stratified spaces , satisfying the witt condition , on which these two cohomologies coincide .
even more strongly , for suitably chosen metrics , the domains @xmath15 and @xmath16 coincide . in a previous paper @xcite we have given a new treatment of the de rham / hodge theory of an @xmath9 metric on witt spaces ; this was then applied to the novikov conjecture on these spaces . in the present paper , we extend that theory to handle general non - witt spaces . for any metric on a non - witt space , @xmath17 , and one key part of our task is to define geometrically appropriate domains for the exterior derivative that lie between these two extremes .
this is done following ideas of cheeger , published by him in the case of isolated conic singularities @xcite , and here extended to this general stratified setting .
a space @xmath1 is non - witt if and only if there is a stratum @xmath18 with link at @xmath19 @xmath20 of even dimension such that @xmath21 if this is the case , these spaces form a vector bundle over @xmath18 with a natural flat connection . at every stratum where the witt condition fails ,
we choose a smooth subbundle , with @xmath22 which is parallel with respect to the flat connection , and so that whenever two strata have intersecting closures , these choices are compatible . such a choice will be called a _ mezzoperversity .
_ we work with ` suitably scaled , rigid ' @xmath9 metrics .
the rigidity condition states that the metric has an ` exact conic ' structure on the conic fibers of the normal neighborhoods of each stratum . as shown in (
* propositions 3.1 - 3.2 ) , there is no loss of generality in assuming this .
the other condition of being suitably scaled is also easy to arrange , and requires simply that the restriction of the metric to the link of each of these conic fibrations be multiplied by a small enough factor so as to exclude spurious small eigenvalues of the induced operators on these links , and hence also excludes unnecessary indicial roots .
we say more about this in [ sec : indicialderham ] , see in particular lemma [ lem : suitablescaling ] .
let @xmath23 be a stratified pseudomanifold with a ( suitably scaled ) @xmath24-metric .
each mezzoperversity @xmath25 on @xmath1 determines a domain for the exterior derivative and for the de rham operator @xmath26 we refer to this specification by saying that we are imposing _
cheeger ideal boundary conditions . _
the latter makes @xmath27 a self - adjoint operator with compact resolvent and induces a strong kodaira decomposition on differential forms , the former induces a fredholm complex whose cohomology @xmath28 satisfies a hodge theorem and is independent of the choice of @xmath9-metric . in a companion @xcite to this paper , we show that this cohomology is invariant under stratified homotopy equivalences
. in proving this theorem , there are two main analytic difficulties to overcome .
the first is the definition of the domains .
if @xmath18 is a stratum of depth one , then its link @xmath29 is a closed manifold , and we establish below that any element @xmath30 has a distributional asymptotic expansion in terms of @xmath31 a ` boundary defining function ' for @xmath32 @xmath33 if @xmath34 is not even , or more generally , if the witt condition holds , then @xmath35 otherwise these forms constitute the cauchy data , on which the ideal boundary conditions are ( algebraic ) restrictions .
we show that once we have imposed suitable boundary conditions on the first @xmath36 singular strata ( in terms of increasing depth ) , then there is an expansion of the form at the contiguous stratum of depth @xmath37 . the second analytic difficulty , to the resolution of which the bulk of the paper is devoted , involves the construction of a parametrix for @xmath38 with cheeger ideal boundary conditions . on a space with a simple edge , i.e. only one singular stratum
, @xmath39 is an ` elliptic edge operator ' in the sense of @xcite , and there is a complete pseudodifferential theory to study it @xcite , at least when @xmath34 is even so that there is no indicial root exactly at the @xmath0 cutoff .
the strategy here is to construct a parametrix for @xmath38 acting on the weighted space @xmath40 and then modify it to obtain a parametrix for @xmath41 with an error term taking values in @xmath42 using this parametrix , we show that the domain @xmath43 is compactly included in @xmath44 there is now a substantial amount of literature concerning related problems on spaces with simple edge singularities .
we mention in particular , beyond @xcite and @xcite , the ongoing work of krainer and mendoza @xcite and the recent work of cheeger and dai @xcite .
the oeuvre of schulze @xcite handles certain problems , though less general than considered here , for stratified spaces .
+ every mezzoperversity @xmath45 has a dual mezzoperversity @xmath46 such that the intersection pairing of differential forms restricts to a non - degenerate pairing between the de rham cohomology of @xmath45 and that of @xmath47 a mezzoperversity which coincides with its dual is called a _ self - dual mezzoperversity_. it is not hard to see that there are topological obstructions , for example involving the signatures of the links , to the existence of a self - dual mezzoperversity on a stratified pseudomanifold .
we shall call a compact stratified space which carries a self - dual mezzoperversity a _ cheeger space_.
let @xmath23 be a stratified pseudomanifold with a ( suitably scaled ) @xmath24-metric . if @xmath25 is a self - dual mezzoperversity on @xmath2 then the domains @xmath48 are invariant under the action of the hodge star and the cohomology @xmath28 satisfies poincar duality .
moreover , its signature is the index of the analytic signature operator , i.e. the de rham operator @xmath38 with the involution induced by the hodge star , with boundary conditions and the domain induced by @xmath49 [ thm2 ] in a companion paper @xcite , we apply the constructions of this paper to define higher signatures on a cheeger space and then prove the novikov conjecture for cheeger spaces which have fundamental groups satisfying the strong novikov conjecture .
we also prove there that the signature of theorem [ thm2 ] above is independent of the choice of self - dual mezzoperversity and is also a bordism invariant .
these properties of the signature can also be deduced by comparing the analytic approach developed in the present paper with a purely topological definition of self - dual mezzoperversities and poincar duality developed earlier by banagl .
the fact that these two approaches lead to equivalent signatures is proved in @xcite , joint with banagl .
+ notice that at each stratum there are two canonical choices of mezzoperversity , the subbundles of @xmath50 of full rank or of rank zero .
it follows from the work of cheeger , and is shown directly in @xcite , that these choices correspond to the upper - middle and lower - middle perversities of goresky - macpherson at this stratum .
this companion paper also shows that any choice of self - dual sheaf in the sense of @xcite yields a mezzoperversity , satisfying the self duality studied in
[ sec : signlag ] .
in particular , from @xcite we see that the reductive borel - serre compactification of a hilbert modular surface is an example of a space carrying interesting non - trivial mezzoperversities .
+ * acknowledgements . *
was partly supported by nsf grant dms-1104533 and an ihes visiting position and thanks sapienza universit di roma , stanford , and institut de mathmatiques de jussieu for their hospitality and support .
acknowledges support by nsf grant dms-1105050 .
thanks the _ projet algbres doprateurs _ of _ institut de mathmatiques de jussieu _ for hospitality during several short visits and a two months long visit in the spring of 2013 ; financial support was provided by universite paris 7 , istituto nazionale di alta matematica ( indam ) and cnrs ( through the bilateral project noncommutative geometry " ) and ministero delluniversit e della ricerca scientifica ( through the project spazi di moduli e teoria di lie " ) .
thanks sapienza universita di roma for hospitality during several week - long visits ; financial support was provided again by indam and cnrs through the bilateral project noncommutative geometry " the authors are happy to thank richard melrose , jeff cheeger , michel hilsum , thomas krainer , gerardo mendoza , francesco bei and markus banagl for many useful and interesting discussions .
in this section we give an overview of the types of operators and boundary conditions to which our methods apply . to help motivate and better
explain all of this , we describe it in the now well - known simple edge setting , i.e. where @xmath7 has only one singular stratum , then explain the analogous constructions in the next most complicated case , the depth two setting , and finally turn to spaces with singular strata of arbitrary depth .
as we will see , much of the new complexity already occurs in the depth two case ; the extension to higher depth singularities involves an induction argument , but otherwise differs mostly in the notation .
although we present all of this for a class of differential operators of arbitrary order satisfying a few structural hypotheses , our main interest is in the de rham and signature operators associated to an incomplete iterated edge metric , and so we use these operators as the key examples throughout , and describe the types of boundary problems in full detail only for these operators .
let @xmath1 be a compact stratified space , with singular strata @xmath51 ordered by increasing depth .
we also denote by @xmath52 the open dense stratum of top dimension , the so - called regular part of @xmath1 ; the other strata @xmath53 are thus regarded as higher codimension boundaries of @xmath1 .
we refer the reader to @xcite for a careful discussion of this class of spaces , and for the accompanying description of a resolution process which associates to any such stratified space @xmath1 a compact manifold with corners @xmath54 with iterated fibration structure on its boundary faces , obtained by blowing up the strata @xmath53 in order of decreasing depth . for notational convenience , we assume for the rest of this paper that each stratum is connected , and even more strongly that there is a single maximal chain of strata @xmath55 where @xmath56 for all @xmath57
. both of these assumptions are simple to remove and are adopted here to make the already complicated notation as simple as possible . without yet specifying the specific degeneracies of the operators we shall consider later ( i.e. the behavior of their coefficients near the singular strata ) , we first recall some generalities about closed extensions of unbounded operators on hilbert spaces .
let @xmath58 be an elliptic operator on @xmath1 acting between sections of two bundles @xmath59 and @xmath60 , and fix a smooth measure on the regular part of @xmath1 and hermitian metrics on the fibres of @xmath59 and @xmath60 .
we may then regard @xmath58 as an unbounded operator acting on @xmath61 sections of @xmath59 .
its graph , as a subspace of @xmath62 , is not closed .
there are two canonical closed extensions of this subspace , both of which are graphs over dense domains @xmath63 .
these are the minimal and maximal extensions : @xmath64 is the domain corresponding to the graph closure of @xmath58 on the core domain @xmath65 , while @xmath66 consists of all sections @xmath67 with @xmath68 .
the imposition of boundary condition consists of a choice of an intermediate domain @xmath69 lying between @xmath64 and @xmath66 .
the goal is to choose such a domain so that the corresponding opertor @xmath70 is not only closed but also fredholm .
it is very difficult to analyze @xmath66 for operators on a stratified space directly .
instead , we proceed inductively .
if @xmath1 has depth @xmath36 , assume that ` good ' boundary conditions have already been chosen at the strata @xmath71 of depth less than @xmath36 .
we study the ` partially minimal ' and ` partially maximal ' domains obtained by imposing these boundary conditions at the lower depth strata and applying the general definitions above at the stratum @xmath72 of depth @xmath36 .
functions ( or sections ) in this partially maximal domain which satisfy @xmath73 ( or @xmath74 for @xmath75 ) have partial asymptotic expansions at @xmath72 , the coefficients of which can be regarded as the cauchy data of @xmath76 .
a choice of good boundary conditions at @xmath72 consists of imposing appropriate conditions on these cauchy data ; in this paper we shall only impose local algebraic conditions on these cauchy data , but more generally one could impose nonlocal ( e.g. pseudodifferential ) conditions . a refined analysis of more general boundary value problem of this type on spaces with simple edge singularities ( i.e. on depth @xmath77 spaces )
is carried out in @xcite , and in a series of articles by krainer and mendoza , @xcite .
as noted earlier , we shall make a set of structural assumptions on the operator @xmath58 , and for the sake of exposition , we introduce these assumptions gradually , once we have provided motivation and suitable notation .
our first of these is the [ ass : localizable ] suppose that @xmath78 , i.e. @xmath79 and @xmath80 . if @xmath81 is a point in a singular stratum of @xmath2 then there is a smooth function @xmath82 supported in a distinguished neighborhood of @xmath81 , with @xmath83 in a smaller neighborhood of @xmath84 such that @xmath85 [ assump1 ] we define the notion of distinguished neighbourhoods later .
we showed in @xcite that assumption [ assump1 ] holds for the de rham operator .
the same argument extends easily to show that it also holds for any dirac - type operator associated to an @xmath9 metric , as defined below .
as promised , we now embark on a much more detailed discussion , initially on stratified spaces @xmath1 with only one singular stratum @xmath86 we call these simple edge spaces .
one of the basic features of a stratified space is its thom - mather data ( see @xcite ) , which in this simple edge setting consists of a tubular neighborhood @xmath87 of @xmath18 in @xmath1 , equpped with a fibration over @xmath18 , the typical fiber of which is a cone over a smooth closed manifold @xmath29 .
the ` resolution ' @xmath88 of @xmath1 is obtained by radially blowing up each of these conic fibres at its vertex ; thus @xmath88 is a smooth manifold with boundary , and the thom - mather data of @xmath1 is converted to a fibration of @xmath89 over @xmath18 with fiber @xmath90 @xmath91 there is a canonical identification of the interior of @xmath88 and @xmath52 , and we denote these common smooth spaces by @xmath7 : thus @xmath92 now choose any smooth boundary defining function @xmath93 for @xmath89 .
a complete edge metric on @xmath7 is , by definition , a metric which near @xmath18 takes the form @xmath94 where @xmath95 is metric on @xmath18 and @xmath96 a family of two - tensors that restricts to a metric on each fiber of @xmath97 we always assume that all of these terms are smooth for @xmath98 .
we have buried some technicalities about the precise form of the asymptotics we allow , and shall return to these details when they become relevant below .
complete edge metrics may be regarded as the basic structural regulators of the geometry of simple edge spaces , and many of the auxiliary quantities defined below are most naturally phrased in terms of these metrics .
however , the focus of this paper is on _ incomplete _ edge metrics ; by definition @xmath8 is an incomplete edge metric if @xmath99 is a complete edge metric , see @xcite .
the main reference for all of the definitions and constructions in this simple edge case is the paper @xcite .
a vector field @xmath100 on @xmath88 has pointwise length which is uniformly bounded with respect to a complete edge metric if the restriction of @xmath100 to @xmath89 is tangent to the fibers of @xmath97 the space of all such @xmath100 is called the space of ` edge vector fields ' ; these are sections of the naturally defined ` edge tangent bundle ' , @xmath101 .
this bundle naturally includes in the ordinary tangent bundle , @xmath102 , and this inclusion is an isomorphism over @xmath7 . over @xmath89 ,
however , @xmath103 is a bundle of rank @xmath104 a differential operator @xmath58 on @xmath7 is an ` _ edge differential operator _ ' if it can be written as a locally finite sum of products of edge vector fields .
this means , in local coordinates near the boundary , that it has the form @xmath105 for some @xmath106 . here , as before , @xmath93 is a boundary defining function , while @xmath107 are coordinates along @xmath32 and @xmath108 are coordinates along @xmath29 .
the coefficients @xmath109 are assumed to be smooth functions in these variables ; if @xmath58 acts between sections of smooth bundles @xmath59 and @xmath60 over @xmath88 , then the @xmath109 are smooth sections of @xmath110 .
the space of these operators of degree at most @xmath111 is denoted @xmath112 .
there is an invariantly defined _ edge symbol _ of @xmath58 , @xmath113 which is naturally defined as a section of the pullback bundle @xmath114 over the edge cotangent bundle @xmath115 , which is the dual of @xmath116 .
this edge symbol is part of a short exact sequence @xmath117 an edge differential operator @xmath58 is said to be ( edge-)elliptic if @xmath118 is invertible away from the zero section of @xmath119 . for brevity
, we always say ellipticity rather than edge ellipticity ; the latter is always implied .
next , the wedge operators ( also known as incomplete edge operators ) of order @xmath111 are defined as the class @xmath120 thus @xmath58 is in @xmath121 if , in the same type of local coordinates , @xmath122 we say that any such @xmath58 is elliptic if the associated edge operator @xmath123 is edge elliptic . as we described in general earlier , if @xmath58 is a wedge operator , then @xmath124 and if @xmath7 is endowed with an ( incomplete ) edge metric and @xmath125 @xmath60 with hermitian metrics , there are two canonical closed extensions : the _ minimal extension _ , with domain @xmath126 where by definition , if @xmath127 then @xmath128 , and the _ maximal extension _ , with domain @xmath129 here @xmath130 is defined distributionally . we can perform the same constructions for the associated edge operator @xmath131 , and if @xmath132 is elliptic then it is a consequence of the results in @xcite that @xmath133 this is the edge sobolev space of order @xmath111 .
there is some relationship between these spaces and the domains of the wedge operator @xmath58 , namely @xmath134 but in general , these subspaces may all be different . to prove this chain of inclusions , note that clearly @xmath78 implies @xmath135 ;
on the other hand , since @xmath136 then given any @xmath137 , we can find @xmath138 which converges to @xmath139 in @xmath140 and such that @xmath141 .
this shows that @xmath142 a key difference from the closed ( nonsingular ) case is that an elliptic edge operator is not automatically fredholm , and one of the main points of the theory is to determine the extra conditions which need to be imposed to obtain a fredholm problem .
these conditions are phrased in terms of a secondary model for @xmath58 beyond its principal symbol ; this is a family of operators known collectively as the ` normal operator , ' @xmath143 here @xmath144 is the inward - pointing half of the bundle @xmath145 and @xmath146 is the natural projection .
the restriction to the fiber over a point @xmath147 is denoted @xmath148 where , in an abuse of notation , @xmath149 denotes the restriction of the bundle @xmath150 to @xmath151 and similarly @xmath152 in local coordinates , after identifying @xmath153 with @xmath154 @xmath155 where @xmath156 and @xmath76 are coordinates on @xmath157 and @xmath158 respectively . in @xcite
it is shown that an elliptic edge differential operator @xmath58 induces a fredholm operator on @xmath159 precisely when @xmath160 is invertible for every @xmath147 .
the method for proving this result relies on a detailed analysis of the structure of the schwartz kernel of a parametrix for @xmath58 .
this parametrix is constructed using both @xmath161 , via a modification of the standard symbol calculus parametrix method , and @xmath162 .
the inverse of this family of model problems is integrated into the parametrix using a fundamentally geometric construction .
specifically , we make the ansatz that the schwartz kernel of the parametrix lifts to a polyhomogeneous ( or at least conormal ) distribution on the ` edge double space ' , @xmath163.\ ] ] this space is obtained from @xmath164 by radially blowing up the fiber diagonal of @xmath165 i.e. , replacing the fiber diagonal with its spherical inward - pointing normal bundle , see @xcite .
the boundary hypersurface produced by this blow - up is called the _ edge front face _ and denoted @xmath166 a pseudodifferential edge operator @xmath167 , by definition , is a pseudodifferential operator on @xmath7 , the schwartz kernel @xmath168 of which is a distribution on @xmath164 , but which lifts to @xmath169 as a polyhomogeneous distribution .
we typically place restrictions on , or at least keep track of , the exponents in the expansions of this lift of @xmath168 at the various boundary faces . in particular
, this lift depends smoothly on the normal variable to @xmath170 , and its restriction to this face is , by definition , the normal operator of @xmath171 .
the face @xmath170 is the total space of a fibration over @xmath18 with each fiber naturally identified with a projective compactification of @xmath172 , and @xmath171 should be regarded as a family of operators @xmath173 on this space .
there are analogues of the various symbolic rules , including that @xmath174 specializing to the case where @xmath167 is an elliptic differential edge operator with invertible normal operator , one should therefore expect to construct a parametrix @xmath175 for @xmath167 by choosing @xmath175 so that @xmath176 and @xmath177 .
this is indeed the case , and is the basis for the parametrix construction in @xcite .
thus pseudodifferential operators include both differential edge operators and parametrices of elliptic edge operators with invertible normal operator .
we carry out a somewhat stripped down version of this construction here on stratified spaces , see [ sec : invnormal ] . in studying the normal operator of @xmath178 ,
one uses properties of a simpler model operator .
this is the _
indicial family _ ,
@xmath179 given in local coordinates by @xmath180 it is also defined by the equation @xmath181 for any fixed @xmath19 the indicial family is invertible away from a discrete set @xmath182 , the elements of which are called the _ indicial roots _ of @xmath183 at @xmath184 there is a slightly more refined object , @xmath185 we also set @xmath186 returning now to the incomplete edge operator @xmath187 , observe that @xmath188 in other words , @xmath189 decays faster than expected , compared to a generic element of @xmath190 .
the paper @xcite restricts attention only to those elliptic operators @xmath183 which satisfy the strong hypothesis that the set @xmath191 is independent of @xmath192 . using this ,
a basic result in @xcite , recorded in lemma [ lem : expansionexistence ] below , states that using , @xmath76 has a partial expansion : @xmath193 this sum is over the subset @xmath194 of triples @xmath195 for which both @xmath196 . moreover , each map @xmath197 is continuous from @xmath198 , endowed with the graph topology , to @xmath199 let @xmath200 be the largest logarithmic power accompanying a given @xmath201 .
it turns out that only the subset @xmath202 of these asymptotic coefficients are relevant in the sense that if these _ leading coefficients _ vanish , then all coefficients @xmath203 with @xmath204 also vanish .
it is also true that every leading coefficient @xmath202 is in the null space of @xmath205 the paper @xcite studies elliptic edge operators @xmath58 for which only the indicial roots @xmath206 lying in a particular strip @xmath207 ( where @xmath208 and @xmath209 are in a range determined by mapping properties of the normal operators @xmath210 ) may vary . the orders of the indicial roots are allowed to change , however . in the present paper we work with the intermediate [ ass : cstindicialroots ] the set @xmath194 is independent of @xmath192 .
this strip turns out to be the same as the one considered in @xcite .
quite a lot can still be done even without assumption [ ass : cstindicialroots ] .
for example , in a series of papers @xcite and further ongoing work , krainer and mendoza study elliptic edge operators assuming only that @xmath211 does not intersect the boundaries of this same strip , but allowing the indicial data in this range to vary .
under assumption [ ass : cstindicialroots ] , it is straightforward to regard the leading asymptotic coefficients @xmath202 as distributional sections of a finite dimensional smooth vector bundle over @xmath32 called the trace bundle @xmath212 one of the key results by krainer and mendoza is that one can still make sense of this trace bundle even when the indicial roots vary . in any case , we henceforth restrict attention to operators for which the elements of @xmath191 in this strip are constant .
we show in lemma [ lem : suitablescaling ] that the de rham operator for a ( suitably scaled ) metric satisfies this assumption , and do not discuss the more general situation further .
the starting point for the formulation of boundary conditions is the cauchy data map @xmath213 we now claim that @xmath214 to prove this , first note that by ( a simplified version of ) the proof of theorem [ thm : witlessdecay ] , if @xmath210 ( @xmath215 ) is injective on @xmath216 for some sufficiently small @xmath217 and all @xmath147 , and if @xmath218 has all @xmath219 ( for @xmath204 ) , then @xmath220 . however , we already know that @xmath221 so an approximation argument , cf .
* proposition 5.11 ) , completes the argument .
this fact now shows that the projection @xmath222 can be identified with the cauchy data map , and hence every closed domain for @xmath58 can be realized as @xmath223 where @xmath175 is an arbitrary closed linear operator acting on the coefficients @xmath224 thus @xmath58 has a unique closed extension , or equivalently @xmath225 , if and only if @xmath194 is empty . a * local ideal boundary condition * for @xmath58 is a differential operator @xmath226 where @xmath227 is some auxiliary bundle , with associated domain @xmath228 for the de rham operator @xmath229 of an incomplete edge metric ( suitably scaled , see lemma [ lem : suitablescaling ] below ) , @xmath230 is nonempty if and only if @xmath1 does _ not _ satisfy the _ witt _ condition . in fact , from corollary [ cor : derhamexp ] , @xmath230 contains a single element @xmath231 , and the corresponding coefficient @xmath232 ( we drop the other indices ) is a distributional section of the bundle @xmath233 here @xmath234 is the vertical hodge bundle with respect to the induced metric and @xmath235 thus when @xmath1 is witt , there is no need to impose boundary conditions , while if @xmath1 is not witt , then a local ideal boundary condition is a differential operator @xmath236 we shall be primarily interested in very special local ideal boundary condition associated to arbitrary sub - bundles @xmath237 indeed , for any such @xmath238 , if @xmath239 is the orthogonal complement with respect to the induced bundle metric , and @xmath240 is orthogonal projection onto @xmath241 then we define the * cheeger ideal boundary condition * associated to @xmath238 by @xmath242 the bundle @xmath243 has a natural flat structure , ( * ? ? ?
* proposition 10.1 ) , ( * ? ? ?
* proposition 3.4 ) , ( * ? ? ?
* proposition 14 ) .
this is a general fact about the vertical cohomology of a fibration : if @xmath244 is a smooth fibration endowed with a connection @xmath245 then there is a splitting of the differential forms on @xmath246 as @xmath247 where @xmath248 and @xmath81 are the horizontal and vertical degrees , respectively .
the exterior differential @xmath249 decomposes as @xmath250 the first term , @xmath251 , is the vertical de rham operator on the fibers , the second , @xmath252 , is the sum of a lift of @xmath253 and a tensorial term built from the second fundamental form of the fibration , and the third , @xmath254 , is purely tensorial and involves the curvature of the fibration .
the identity @xmath255 implies various identities among these terms , one of which implies that @xmath252 induces a flat connection on the vertical de rham cohomology bundle @xmath256 and , in the presence of a riemannian submersion metric , @xmath252 projects to a connection on the vertical hodge cohomology bundle @xmath257 for a simple edge space @xmath1 , the boundary fibration @xmath258 extends to small values of @xmath31 @xmath259 & x \cap \ { x = c \ } = { \partial}_c x \ar[r]^-{\phi_c } & y. } \ ] ] it is not hard to calculate how each of the individual terms in the decomposition of @xmath11 , acting on @xmath260 forms , scale with @xmath261 : @xmath262 as before , @xmath263 induces flat connections @xmath264 we shall call a choice of subbundle @xmath238 in a ( hodge ) * mezzoperversity * or * flat structure * if it is parallel with respect to @xmath265 .
we show in theorem [ thm : mainhodgethm ] below that given any mezzoperversity @xmath241 the closed realization of the de rham operator with cheeger ideal boundary conditions , @xmath266 is a self - adjoint fredholm operator on @xmath0 with compact resolvent .
now suppose that the stratification of the pseudomanifold @xmath1 has two strata , one contained in the closure of the other : @xmath267 thus the strata of @xmath1 are @xmath268 we denote by @xmath269 the link of @xmath270 and @xmath271 the link of @xmath272 , both in @xmath1 .
notice that @xmath273 is a stratified space in its own right ; in fact , it is a simple edge space with singular stratum @xmath272 , and we denote by @xmath246 its link in @xmath273 .
moreover , the link @xmath271 is also a simple edge space with singular stratum @xmath246 and such that the link of @xmath246 in @xmath271 is @xmath269 .
note finally that @xmath272 , @xmath269 and @xmath246 are all smooth closed manifolds .
this figure is a schematic representation of a stratified space with two strata : * the stratum @xmath272 is the base circle in blue ; * the stratum @xmath270 is the union of the large red punctured cones @xmath274 , as @xmath248 varies within @xmath272 ; the link of a point @xmath81 in @xmath270 is the closed manifold @xmath275 denoted by the small green dotted circles ; * the closure @xmath273 is the union of the large red cones @xmath276 as @xmath248 varies within @xmath272 ; the link of @xmath273 at @xmath277 @xmath278 is denoted by the large bold green circle ; * the link of @xmath1 at @xmath277 @xmath279 is denoted in bold green in the picture and consists of the cones @xmath280 as @xmath81 varies in @xmath281 ; it is a depth one stratified space with singular stratum @xmath281 notice that given @xmath81 in @xmath270 as in the picture , its link @xmath282 is also equal to the link of @xmath81 viewed as a point in the singular stratum @xmath283 of @xmath284 this exhibits the well known slogan the link of a link is a link . " as explained in @xcite , it is convenient to resolve @xmath1 to a smooth manifold with corners , @xmath88 with iterated fibration structure on its boundary .
this resolution process is explained in great detail in 2 of that paper . in this depth two case , @xmath88 is a manifold with corners of codimension two obtained by radially blowing up first @xmath272 and then the lift of @xmath270 : @xmath285.\ ] ] this space has two boundary hypersurfaces , one corresponding to @xmath270 and the other to @xmath272 , and each of these boundary faces fibers over the resolution of the corresponding stratum , @xmath286 with fiber @xmath269 over @xmath287 and @xmath288 the resolution of @xmath289 over @xmath290 these fibrations are compatible at the corner @xmath291 in the sense that they fit into a commutative diagram @xmath292 & & & & m \ar@{-}[ld ] \\ z^1 \ar@{-}[r ] & h^1 \cap h^2 \ar[rr]^-{\phi_1 } \ar[rd]^-{\phi_2 } & & { \partial}{\widetilde{y}}^1 \ar[ld ] & \\ & & y^2 & & } \xymatrix { \\ \text { which induces } z^1 \ar@{-}[r ] & { \partial}{\widetilde{z}}^2 \ar[r]^-{\phi_1 } & m } \ ] ] a riemannian metric @xmath8 on @xmath7 is an _ iterated incomplete edge ( or @xmath9 ) metric _ if near @xmath270 it is an incomplete edge metric as defined above in [ sec : simpleedgebc ] , while near points of @xmath272 it has the form @xmath293 where @xmath93 is a boundary defining function for @xmath294 @xmath295 is a riemannian metric on @xmath272 and @xmath296 restricts to an incomplete edge metric on each fiber of @xmath297 as in the simple edge case , this can be interpreted as a metric on the iterated incomplete edge cotangent bundle @xmath298 over all of @xmath299 see @xcite .
if @xmath59 and @xmath60 are vector bundles over @xmath2 then , following @xcite , a differential operator @xmath58 of order @xmath111 acting between sections of @xmath59 and @xmath60 is an _ iie operator _ or _ iterated wedge operator _
if * in local coordinates near @xmath300 @xmath301 where @xmath302 is a bdf for @xmath303 @xmath304 are local coordinates along @xmath270 and @xmath305 are local coordinates along @xmath306 thus , restricted to a subset of @xmath307 @xmath58 is an @xmath9 operator as in
[ sec : simpleedgebc ] . * in local coordinates near @xmath308 @xmath309 where here @xmath93 is a bdf for @xmath310 @xmath311 are local coordinates along @xmath310 and @xmath312 are @xmath313-vertical vector fields such that each @xmath314 is an incomplete edge operator on the fibers of @xmath297 we start by analyzing elements of @xmath198 at points in @xmath315 each point @xmath81 in this stratum has a distinguished neighborhood @xmath316 in which @xmath59 and @xmath60 are trivial , and which is identified with @xmath317 . here
@xmath302 is a bdf for @xmath270 and @xmath318 is the unit ball in @xmath319 this neighborhood is an ( open ) simple edge space and the restriction of @xmath58 to @xmath316 is an element of @xmath320 this restricted operator may then be analyzed as described above ; in particular , we impose _ assumption [ ass : cstindicialroots ] _ there .
elements in @xmath321 have distributional partial asymptotic expansions as @xmath322 , and we can impose local ideal boundary conditions @xmath323 for @xmath58 over @xmath318 ; thus @xmath324 for some auxiliary bundle @xmath325 over @xmath315 we define @xmath326 the key point is to analyze elements of @xmath327 near @xmath328 thus , near any @xmath329 choose a distinguished neighborhood @xmath316 as before .
thus , @xmath59 and @xmath60 are trivial on this neighbourhood and @xmath330 where @xmath93 is a bdf for @xmath272 and @xmath331 is the unit ball in @xmath332 in each such neighborhood , @xmath333 we prove below that elements of @xmath327 with support in @xmath316 have partial asymptotic expansions at @xmath334 with coefficients in the null space of @xmath335 if @xmath58 is as in in this distinguished neighborhood , then the ` partially completed operator ' has the form @xmath336 it is natural to think of this operator @xmath183 near @xmath272 as an edge differential operator ` with values in the space of wedge differential operators on @xmath271 ' .
thus we can proceed essentially just as in the simple edge case , via model operators , with the proviso that we must keep track of the boundary conditions along @xmath270 .
first define the family of normal operators of @xmath183 along @xmath272 by @xmath337 this an element of @xmath338 for each @xmath329 .
there is a stratification of @xmath339 ; the face @xmath340 , which is just @xmath341 resolves the depth two stratum , and has link ( or fiber ) @xmath342 while the resolution of the depth one stratum is @xmath343 , or equivalently , @xmath344 , which has link @xmath345 an @xmath9 metric on @xmath1 and hermitian metrics on @xmath59 and @xmath60 induce corresponding objects on @xmath346 . at an interior points
@xmath347 , @xmath210 is an incomplete edge operator . by the same analysis
as described earlier , see also lemma [ lem : expansionexistence ] below , elements of @xmath348 have distributional expansions as @xmath349 , @xmath350 with error term @xmath351 the indicial family of @xmath210 is @xmath352 , so @xmath353 and @xmath354 .
the coefficients are distributional sections of the trace bundle , @xmath355 unwinding the definitions , this is identified with the pull - back and restriction of @xmath356 the cauchy data map assigns to @xmath357 the coefficients in its expansion , @xmath358 now consider the differential operator @xmath323 on @xmath270 which determines the boundary conditions there . recalling that @xmath273 is a simple edge space with singular stratum @xmath272 and link @xmath246 , and that @xmath323 is an edge operator , we analyze its normal operator at points of @xmath272 , @xmath359 this defines a domain at the level of normal operators consistent with @xmath327 : @xmath360 the partially completed operator @xmath183 has indicial family at point @xmath329 , given in local coordinates as @xmath361 this has domain @xmath362 here @xmath363 is the cauchy data map at @xmath364 , which is defined as before , recalling that @xmath271 is a simple edge space with @xmath9 metric .
we now impose the [ fredholm indicial family ] [ assindfamily ] for each @xmath365 the family of indicial operators @xmath366 is fredholm on @xmath367 and there is some @xmath368 for which this operator is invertible .
we prove in theorem [ thm : mainhodgethm ] that this holds for the de rham operator .
since this is a holomorphic family of fredholm operators , this assumption provides the extra piece of information needed to apply the analytic fredholm theorem , which implies that the family of inverses @xmath369 is meromorphic .
hence for each @xmath81 , its poles are discrete .
this meromorphic family determines the two sets @xmath370 { \mathrm{spec}}_b(p;b^1,q ) & = \ { ( \zeta , p ) \in { \mathbb{c}}\times { \mathbb{n}}_0 : ( i_q(p;\eta ) , { \mathcal{d}}_{\max , i_q(b^1;\eta)}(i_q(p;\eta))^{-1 } ) \\ & \qquad \qquad \qquad \qquad \qquad \text { has a pole of order $ p$ or greater at } \eta=\zeta \}. \end{split}\ ] ] consider the set @xmath371 where @xmath372 consists of triples @xmath373 in @xmath374 such that @xmath375 if @xmath376 remains disjoint from all other indicial roots of @xmath58 as @xmath81 varies , then by ( * ? ? ? * theorems 3.2 and 6.3 ) there is a smooth vector bundle @xmath377 the smooth sections of which are the functions in .
we impose again the _ assumption [ ass : cstindicialroots ] _ that the set @xmath376 is independent of @xmath378 lemma [ lem : expansionexistence ] below asserts that @xmath379 where this sum is a distributional section of @xmath380 and @xmath381 we again identify sections of @xmath380 with the ( finite ) set of coefficients @xmath203 , and then define the cauchy data map of @xmath58 at @xmath272 by @xmath382 we can now define local ideal boundary conditions at @xmath272 , @xmath383 where @xmath384 is an auxiliary bundle .
the corresponding domain for @xmath58 is @xmath385 we refer to the pair @xmath386 as * local ideal boundary conditions * for @xmath58 .
we show later that the operator @xmath387 is closed for any choice of local ideal boundary conditions .
we now specialize this discussion to where @xmath58 is the de rham operator of an @xmath9 metric ( scaled as in lemma [ lem : suitablescaling ] ) . if @xmath1 is witt , there is no need to impose boundary conditions . if either @xmath270 or @xmath272 , but not both , does not satisfy the witt condition , then we can proceed just as in [ sec : simpleedgebc ] .
hence we turn immediately to the case where the witt condition fails at both @xmath270 and @xmath272 .
the leading coefficient of an element of the maximal domain is a distributional section of the bundle @xmath388 each summand here has a flat connection .
suppose that @xmath389 is a flat sub - bundle of @xmath390 with orthogonal projection @xmath391 .
then we can impose the cheeger ideal boundary condition @xmath392 at the next stratum @xmath310 the leading coefficient of an element in @xmath393 is a distributional section of the bundle @xmath394 the typical fiber in either of these summands is @xmath395 it is not immediate that these spaces vary smoothly with @xmath81 , but this follows from theorem [ thm : selfdual ] , which identifies these hodge cohomology spaces with de rham cohomology spaces ; the fact that they are metric - independent is theorem [ thm : metricindep ] .
this also follows by combining theorem [ thm : mainhodgethm ] , which shows that _ assumption [ assindfamily ] _ holds , with @xcite .
the bundle @xmath396 has a natural flat connection , obtained as in [ sec : simpleedgebc ] from the exterior derivative on @xmath397 .
since the vertical cohomology is now an @xmath0-cohomology , we discuss this more closely .
choose a small collar neighborhood @xmath398_x$ ] of @xmath397 in @xmath88 and extend the boundary fibration @xmath399_x \right ) } - { \mathrm{coll}}(h^2 ) { \xrightarrow{\phantom{x } { \mathrm{coll}}(\phi ) \phantom{x } } } y^2.\ ] ] using the metric on @xmath1 , we obtain an orthogonal splitting of the tangent bundle @xmath400_x \cong { { } ^{{\operatorname{iie}}}}ty^2 \oplus { { } ^{{\operatorname{iie}}}}tz^2 \oplus t[0,{\varepsilon}]_x.\ ] ] this induces a decomposition of @xmath11 on differential forms supported in @xmath401 : @xmath402 where @xmath403 is independent of @xmath404 the graph closure of @xmath11 with core domain @xmath405 has domain @xmath406 see also [ sec : l2coho ] for an alternate description .
we also consider @xmath407 this is the space of forms with support in @xmath401 projecting down to a compact subset of @xmath310 and which annihilate @xmath408 . as in @xcite
there is a natural surjective quotient map @xmath409 which defines a differential operator ( for any @xmath106 ) by @xmath410 & { { \mathcal{c}}^{\infty}_c } ( y^2 ; \lambda^{p+1}({{}^{{\operatorname{iie}}}}t^*y^2 ) \otimes { \operatorname{h}}_{w^1}^\ell ( h^2/y^2 ) ) \\
\omega = \psi(e ) \ar@{|->}[r ] & \psi(d^{1,0}_{{\mathrm{coll}}(h^2 ) } e ) } \ ] ] it follows from @xmath411 that this is well - defined independently of the choice of @xmath412 lifting @xmath413 and that @xmath414 the leibnitz rule satisfied by @xmath415 induces a leibnitz rule for @xmath416 so we see that @xmath417 is a flat connection on the bundle @xmath418 we have defined the action of @xmath417 on compactly supported sections , but one can now extend it to arbitrary smooth sections . alternately , it is easy to see that the map @xmath419 extends to a surjective quotient map @xmath420 ( where @xmath421 refers to the annihilator of @xmath408 ) and so the definition of @xmath417 above extends naturaly to this domain . there is a similar decomposition of @xmath422 on forms supported in @xmath423 @xmath424 and finally , a natural identification @xmath425 this allows us to define a flat connection @xmath265 on the bundle of vertical harmonic forms @xmath426 namely , @xmath427 where @xmath428 is the orthogonal projection onto @xmath429 the hodge map induces a bundle isomorphism @xmath430 which also identifies @xmath265 with @xmath417 ( ( * ? ? ?
* proposition 3.14 ) ) .
+ a local ideal boundary condition at @xmath272 is a differential operator acting on sections of the bundle .
we associate to each flat sub - bundle @xmath431 the cheeger ideal boundary condition @xmath432 the pair @xmath433 constitute a ( hodge ) * mezzoperversity*. we show in theorem [ thm : mainhodgethm ] that for any such mezzoperversity the de rham operator with cheeger ideal boundary conditions @xmath434 is a self - adjoint fredholm operator on @xmath0 with compact resolvent . now consider a space @xmath1 which has singular strata up to depth @xmath37 .
one reason for our elaboration of the depth two case is that this discussion extends to the higher depth almost unchanged .
order the strata @xmath435 so that @xmath436 , @xmath437
. thus , at one end , @xmath270 is a stratum of least depth and its link is a smooth closed manifold , while @xmath72 is a smooth closed manifold which has link a stratified pseudomanifold with depth @xmath438 .
we assume inductively that we have defined local ideal boundary conditions for all spaces with depth less than or equal to @xmath36 .
we now describe how to extend this definition to @xmath439 . as before ,
we let @xmath93 denote a boundary defining function for the hypersurface corresponding to this stratum in the resolution @xmath88 , and work with the partially completed operator @xmath215 .
we use now the notation of @xcite for iterated edge and iterated incomplete edge operators .
let @xmath440 be an elliptic @xmath9 operator acting on sections of a vector bundle @xmath441 the local ideal boundary conditions at @xmath442 are differential operators @xmath443 where each @xmath444 is an auxiliary bundle over @xmath53 .
we always impose assumptions [ ass : localizable ] , [ ass : cstindicialroots ] and [ assindfamily ] at each stratum .
recall also that @xmath445 @xmath446 are interiors of manifolds with corners with iterated fibration structures and the boundary condition operators are iterated edge operators with respect to this structure .
fix @xmath447 and choose a ( punctured ) neighborhood @xmath448 .
the indicial and normal operators for the elements of @xmath449 define local ideal ideal boundary conditions @xmath450 for @xmath451 and @xmath452 for @xmath453 . assuming that @xmath454 satisfies assumption [ assindfamily ] , lemma [ lem : expansionexistence ]
then shows that elements of @xmath455 have partial distributional asymptotic expansions at @xmath456 @xmath457 where @xmath458 and , with @xmath459 as above , we have @xmath460 if the elements of @xmath461 are isolated from the rest of @xmath462 , then ( * ? ? ?
* theorems 3.2 and 6.3 ) gives a smooth vector bundle @xmath463 of which the sum in is a distributional section . imposing _
assumption [ ass : cstindicialroots ] _ as well , we define the cauchy data map of @xmath58 at @xmath439 , @xmath464 ^ -{{\mathcal{c}}_{y^{k+1 } } = { \mathcal{c}}_{y^{k+1}}(l;{\mathbf{b } } ) } & { { \mathcal{c}}^{-\infty}}{\left ( y^{k+1 } ; { \mathcal{tr}}_{y^{k+1}}(l;{\mathbf{b } } ) \right ) } \\ u \ar@{|->}[r ] & ( u_{\zeta_j , k , p } ) } \ ] ] local ideal boundary conditions at @xmath439 are defined by a differential operator @xmath465 with @xmath466 an auxiliary bundle .
we complete the inductive step by setting @xmath467 the collection of differential operators @xmath468 constitutes a * local ideal boundary condition * for @xmath469 theorem [ thm : closeddomains ] shows that for any such choice of @xmath470 , @xmath471 is a closed operator .
+ specialize again to the de rham operator for a scaled @xmath9 metric .
generalizing the discussion in [ sec : twoedgebc ] we make the [ def : flatsystem ] let @xmath472 be a stratified pseudomanifold and @xmath55 an ordering of the strata of @xmath1 with increasing depth .
a ( hodge ) * mezzoperversity * is a collection of bundles @xmath473 with @xmath474 here @xmath475 is the bundle with fiber @xmath476 as before , that these spaces form a bundle follows from ( * ? ? ?
* theorems 3.2 and 6.3 ) once we establish the appropriate versions of _ assumption [ assindfamily ] _ , but also follow directly from theorem [ thm : selfdual ] where these hodge cohomology groups are identified with de rham cohomology groups .
the discussion of the flat connection in
[ sec : twoedgebc ] extends to the general case essentially unchanged .
every mezzoperversity determines cheeger ideal boundary conditions for the de rham operator on @xmath477 @xmath478 and we shall prove that this is a closed , self - adjoint fredholm operator on @xmath479 with compact resolvent .
we work as usual with a stratified space @xmath1 with singular strata @xmath55 ordered with increasing depth , an @xmath9 metric @xmath8 on @xmath480 , and an elliptic @xmath9 differential operator @xmath481 . for simplicity
we assume that @xmath58 acts on sections of a single bundle @xmath59 .
assume finally that local ideal boundary conditions @xmath482 have been chosen at the first @xmath36 strata , and that @xmath58 satisfies assumptions [ ass : localizable ] , [ ass : cstindicialroots ] and [ assindfamily ] at @xmath483 we show now that elements in @xmath455 have distributional asymptotic expansions at @xmath439 .
we also derive the ` integration by parts trick ' , lemma [ lem : intbyparts ] , which is a key tool in the proof that the de rham operator is fredholm . + for @xmath484 , let @xmath485 be a distinguished neighborhood with fixed trivialization @xmath486 .
the mellin transform is a good tool to study asymptotic expansions .
we fix the measure @xmath487 on @xmath316 , and set @xmath488 so that @xmath489 as before , we work with the partially completed operator @xmath215 . [
lem : expansionexistence ] if @xmath490 and @xmath491 then , as @xmath492 @xmath493 with @xmath461 from and with @xmath494 and @xmath495 the coefficients @xmath496 and the remainder @xmath497 all depend continuously on @xmath490 ( with respect to the graph norm ) .
the leading coefficients @xmath498 are the images of @xmath76 under the finite rank projectors which are the leading coefficients in the laurent expansion of the indicial family at @xmath499 so can be interpreted as sections on @xmath500 of @xmath59 twisted by the kernel of the indicial operator of @xmath501 @xmath502 if any leading coefficient @xmath503 vanishes , then so do the subsequent coefficients @xmath504 for the same indicial root .
following @xcite , a variant of the standard symbol calculus produces a `` small calculus '' parametrix @xmath167 for @xmath505 , where @xmath506 is a product of boundary defining function for each of the hypersurface boundaries of @xmath88 .
this has the following properties : if @xmath507 , then @xmath508 for any @xmath509 ; furthermore , if @xmath510 and @xmath511 , then @xmath512 is smooth in @xmath7 , and @xmath513 for all @xmath514 .
writing @xmath515 , then these properties show that the term @xmath516 can be included into the error term @xmath497 .
hence we may assume for the remainder of the proof that @xmath517 to simplify notation we also replace @xmath518 and @xmath519 by @xmath520 and @xmath29 for the duration of this section .
since @xmath521 , its mellin transform in @xmath93 , @xmath522 , takes values in @xmath523 this extends to a holomorphic function in the region @xmath524 and satisfies the ideal boundary conditions for @xmath525 induced by @xmath470 . writing @xmath526 so that @xmath527 , and using assumption [ assindfamily ] , we have that @xmath528 is holomorphic in the region @xmath529 with values in @xmath530 , even though the right hand side is _ a priori _ a meromorphic function of @xmath368 ( because of the meromorphy of the inverse of the indicial operator ) .
notice that @xmath531 , and hence @xmath532 is a holomorphic function in the half - plane @xmath533 with values in @xmath534 .
reasoning as in @xcite , and recalling that @xmath535 , we obtain that for every @xmath536 $ ] , @xmath537 is holomorphic in @xmath538 with values in @xmath539 .
altogether then , implies that @xmath540 extends to a meromorphic function from @xmath538 to @xmath539 for any @xmath536\cap [ 0,\ell]$ ] .
moreover , the poles all come from the inverse of the indicial operator and hence the leading term at each pole is the image of @xmath76 under the projector appearing as the leading coefficient in the laurent expansion of @xmath541 at that pole . similarly , for @xmath542 , @xmath537 is meromorphic in @xmath543 to @xmath544 ; the poles here arise from the poles in the extension of @xmath545 this yields , in turn , an extension of @xmath540 to the larger region @xmath543 with values in @xmath546 for all @xmath536\cap [ 0,\ell-1]$ ] .
the new poles correspond to those of @xmath547 , i.e. are the indicial roots of @xmath183 with imaginary part lying between @xmath548 and @xmath549 and also to the poles of @xmath550 , which are located at points @xmath551 with imaginary part between @xmath552 and @xmath553 these latter poles only occur if the residue at the indicial root between @xmath552 and @xmath554 which will correspond to the leading coefficient of that root in the expansion , did not vanish .
continuing in this fashion we see that @xmath522 is meromorphic in @xmath533 with values in @xmath555 with an expansion as above .
this identity even holds in @xmath556 and taking inverse mellin transform along @xmath557 yields the element @xmath558 the existence of this expansion has some simple but useful consequences .
for example , it is often convenient to work with the subspace of @xmath559 consisting of elements for which this expansions holds in a strong sense and has smooth coefficients , and fortunately this turns out to be a dense subspace : [ lem : regulardomain ] the domain @xmath455 is the graph closure of @xmath560 if @xmath561 is sufficiently small , then @xmath562 indeed the asymptotic expansion is valid in @xmath140 and the right hand side of is in @xmath563 for some @xmath564 any @xmath78 has a distributional asymptotic expansion at @xmath315 assume that @xmath76 is supported in a distinguished neighborhood @xmath565 around @xmath566 .
choose a standard family of mollifiers @xmath567 on @xmath568 , which we regard as functions on @xmath316 .
then @xmath569 still has support in @xmath316 when @xmath217 is small .
since @xmath570 is independent of @xmath31 @xmath571 hence @xmath572 and @xmath573 the proof of lemma [ lem : expansionexistence ] shows that @xmath574 has an expansion at @xmath270 and that the coefficients are smooth in @xmath568 . the same argument
can then be applied at the strata of increasingly higher depth .
[ thm : closeddomains ] let @xmath575 be local ideal boundary conditions for @xmath58 .
then @xmath576 is a closed operator on @xmath577 suppose that @xmath578 is a sequence in @xmath455 which is cauchy in the graph norm .
since @xmath579 is closed , the limit @xmath580 lies in @xmath198 and @xmath581 in @xmath0 so it remains to show that @xmath76 satisfies the ideal boundary conditions , i.e. that @xmath582 we may assume that the statement is proven at all strata @xmath53 , @xmath583 .
it follows that @xmath76 has an expansion at @xmath439 . since the coefficients @xmath584 depend continuously on @xmath76 , they are the limits of the corresponding coefficients of @xmath578 , and hence belong to @xmath585
this proves that @xmath490 .
the ` weak ' asymptotic expansions we have been discussing here are not always as useful as one might hope because of the low regularity of their coefficient functions .
however , additional information about @xmath58 , such as the injectivity of its normal operators , can sometimes provide more information .
one way to carry this out is the following integration by parts identity .
note that , in contrast to the analogous theorem in @xcite , we make no assumption of self - adjointness here .
[ lem : intbyparts ] let @xmath490 be supported in a distinguished neighborhood @xmath316 around a point @xmath484 , and choose @xmath586 so that there are no terms in the expansion with exponent having real part less than or equal to @xmath587 ; in other words , all nonvanishing terms in this expansion are @xmath588 .
the spaces @xmath589 and @xmath590 are dual with respect to the inner product on @xmath591 .
let us denote the transpose of @xmath183 with respect to this pairing by @xmath592 then for any @xmath593 which satisfies @xmath594 , we have @xmath595 however tempting it may be to think so , this identity is not a direct consequence of the definition of @xmath596 .
it suffices to prove this equality for @xmath597 in the dense subspace @xmath598 for notational convenience , conjugating @xmath183 by @xmath599 , we reduce to the case that @xmath600 , and hence our hypotheses become @xmath601 the natural pairing between @xmath602 and @xmath603 is @xmath604 . by parseval s formula for the mellin transform we obtain first that @xmath605 shifting the contour , this equals @xmath606 where the integrand is now interpreted as a pairing between @xmath607 and we are using our assumption of extra regularity for @xmath608 this integral can be rewritten using integration by parts as @xmath609 and another application of parseval s formula identifies this with @xmath610 as required .
this integration by parts identity makes it possible to transfer an equality of operators from one dense domain to another . to illustrate this ,
suppose that @xmath611 is an unbounded operator defined on either of the two domains @xmath612 .
suppose furthermore that @xmath167 is injective with closed range on @xmath613 , but has nontrivial null space in @xmath614 .
( for example , this can happen on spaces with isolated conic spaces between the minimal and maximal domains . )
then @xmath615 has a bounded left inverse @xmath616 , so that @xmath617 on @xmath618 although both sides of this equation make sense for elements in @xmath619 this is no longer an identity since it fails on @xmath620 the next result shows how lemma [ lem : intbyparts ] may be applied in such a situation .
[ lem : indopconclusion ] let @xmath621 be a dense subspace of @xmath0 with @xmath622 fix @xmath623 let @xmath624 , and then define @xmath625 to be the space consisting of those elements in @xmath621 which lie in @xmath0 with no term in their expansion having exponent with real part less than or equal to @xmath587 , i.e. @xmath626 suppose that @xmath627 is a bounded operator from @xmath628 to itself
. then there is at most one extension of the operator @xmath629 such that simultaneously @xmath630 are bounded . restating the conclusion ,
suppose that @xmath631 is any operator whose restriction @xmath632 to @xmath633 has image in @xmath634 @xmath635 our assertion is that if @xmath636 vanishes on @xmath637 then @xmath638 vanishes on @xmath639 we must first understand the relation between the transposes , @xmath640 observe that @xmath641 { \text { then } } { \left\langle q^t w , v \right\rangle } = { \left\langle w , qv \right\rangle } = { \left\langle w , { \mathscr{q}}v \right\rangle } = { \left\langle { \mathscr{q}}^t w , v \right\rangle}. \end{gathered}\ ] ] since @xmath642 is dense in @xmath479 we conclude that @xmath643 in particular , for any @xmath644 we have @xmath645 thus whenever @xmath646 we can apply lemma [ lem : intbyparts ] to show that @xmath647 since @xmath648 is dense in @xmath634 we conclude finally that @xmath649 for all such @xmath650 this lemma is our main tool for proving that sections of some domain decay faster than expected at the boundary . for example , if @xmath183 is essentially injective on @xmath633 with generalized inverse @xmath651 so that @xmath629 is the projection onto the finite dimensional null space of @xmath183 on @xmath652 then @xmath653 in particular , in this case @xmath654 we now _ specialize to first order @xmath9 operators _ , and describe a structure theorem for the generalized inverse of the normal operator and apply this to the de rham operator in the next section .
we shall use some extensions and adaptations to the present setting of the calculus of pseudodifferential edge operators , @xmath655 , from @xcite and already mentioned in 1.1 in the setting of depth @xmath77 spaces .
we refer to that paper for details on this calculus , and shall follow its language and notation freely
. + let @xmath656 be elliptic , with local ideal boundary conditions @xmath482 at the first @xmath36-strata of @xmath1 .
we work near a point @xmath484 with distinguished neighborhood @xmath316 , and denote by @xmath29 the link of @xmath1 at @xmath81 and fix a boundary defining function @xmath93 for @xmath439 ; we also write @xmath657 and @xmath658 . briefly , the idea is to analyze @xmath659 by treating it as an ( incomplete ) edge operator , making allowances for the fact that the fibre @xmath29 is stratified of depth @xmath36 . by virtue of the inductive structure of the argument , the operator induced by @xmath58 on @xmath29 , together with the boundary conditions above , is fredholm and hence can be treated essentially as when @xmath29 is closed and nonsingular .
beyond assumptions [ ass : localizable ] , [ ass : cstindicialroots ] and [ assindfamily ] , this requires the additional [ compact domain ] [ asscmpdomain ] the inverse @xmath660 when it exists , is a compact operator on @xmath661 we denote the space of compact operators on @xmath662 by @xmath663 this assumption is verified for the de rham operator in theorem [ thm : mainhodgethm ] .
+ as we have shown , the boundary conditions @xmath470 induce a domain @xmath664 for @xmath210 on @xmath665 taking the fourier transform in @xmath666 , writing the dual variable as @xmath667 , and rescaling by @xmath668 ( which reflects the dilation invariance of @xmath183 in @xmath669 ) we can transform @xmath210 to a family of ` bessel - type ' operators parametrized by points in the cosphere bundle over @xmath456 @xmath670 here @xmath671 is the variable in the cosphere bundle . each @xmath672 can be regarded near @xmath673 as an elliptic @xmath674-operator on @xmath675 depending smoothly on @xmath676 , and inherits the domain @xmath677 by passing to the bessel reductions of the boundary operators , @xmath678 . however , as already discussed , there is an important proviso : in the standard @xmath674-theory , the link @xmath29 is assumed to be a smooth compact manifold , while here it is a stratified space of depth @xmath36 . by the assumptions above , and by induction , the indicial family , which is a family of @xmath9 operators on @xmath29 with ideal boundary conditions , has a discrete set of poles and a good regularity theory for elements in its nullspace , and these two facts suffice to proceed with the main constructions of the @xmath674-calculus . however
, beyond this behavior near @xmath673 , the family @xmath672 has a ` bessel structure ' as @xmath679 , and this plays an important role too .
we now analyze @xmath680 following @xcite closely , cf . also @xcite . in the following , use boundary defining functions @xmath681 for @xmath673 and @xmath682 for @xmath683 .
suppose that no indicial root of @xmath672 has real part equal to @xmath684 then , for any @xmath685 , @xmath686 is fredholm .
[ lemfred ] it suffices to construct a right parametrix for @xmath672 since a left parametrix is obtained as the dual of the right parametrix for @xmath687 . to do this
, we construct a parametrix near @xmath673 and another @xmath688 and then patch these together .
near @xmath689 and for any @xmath690 , let @xmath691 ( equivalently , @xmath692 ) on @xmath693 this is obtained by restricting the inverse of the indicial family to the line @xmath694 and taking the inverse mellin transform by integrating along this line . that this is a parametrix for @xmath695 near @xmath673 follows easily from assumption [ asscmpdomain ] .
indeed , @xmath696 and this is a compact operator on functions supported near @xmath673 . as for its behavior
when @xmath697 is large , conjugate @xmath695 by the fourier transform in @xmath697 . letting @xmath587 be the variable dual to @xmath698 we have @xmath699 where @xmath700 is the ` partial principal symbol ' of @xmath701 as explained in ( * ? ? ?
* lemma 5.5 ) , the ellipticity of @xmath183 implies that @xmath702 where @xmath703 is a cut - off function , induces a bounded map @xmath704 patching together @xmath705 and @xmath706 , we obtain an operator @xmath707 which is bounded as in , and such that @xmath708 is compact . dualizing a right parametrix for @xmath709
gives a left parametrix .
the reader may have noticed that the specific value of the weight parameter @xmath552 at @xmath710 plays no real role , and indeed , it follows from this construction that elements of the nullspace of @xmath672 which are tempered as @xmath679 necessarily decrease rapidly : @xmath711 let us now describe the finer structure of @xmath705 , the portion of the parametrix acting on functions supported near @xmath673 .
the indicial family @xmath712 takes values in @xmath713 and by assumption its inverse is a meromorphic function with values in @xmath714 the compact operators on @xmath715 taking the inverse mellin transform , the resulting inverse of the indicial operator is an element @xmath716 and satisfies @xmath717 let us recall here that @xmath718 is the space of @xmath674-pseudodifferential operators @xmath167 of order @xmath719 and with index family @xmath720 describing the exponents in the asymptotic expansions of the schwarz kernel of @xmath167 at the various boundary components of the @xmath674-stretched product @xmath721 .
the index sets @xmath722 are determined by the indicial roots of @xmath183 and the weight @xmath723 and are given by @xmath724 the error term @xmath725 contains an extra factor of @xmath697 , which means that it vanishes to first order on the front face and also at the left face .
( this latter vanishing is achieved by using a careful extension off of the front face , as in ( * ? ? ?
* proof of proposition 5.43 ) ) .
since @xmath705 is obtained by integrating the inverse of the indicial family along a vertical line in the complex plane , it is not surprising that the indicial roots determine the index sets .
since this inverse is meromorphic , changing the contour of integration to a parallel vertical line would produce another operator which differs from the first by the ( finite ) sum of residues at the poles between these two lines .
these residues are the coefficients of the expansion of @xmath726 at the left and right faces , see ( * ? ? ?
* lemma 5.16 ) .
in particular , if @xmath727 is a simple pole of @xmath728 with @xmath729 then the coefficient of @xmath730 of @xmath726 at the left face is the orthogonal projection onto @xmath731 ( of course if @xmath732 this is the coefficient of @xmath733 of the expansion at the right face . )
this is consistent with the fact that @xmath705 maps into @xmath677 and elements in this domain have partial asymptotic expansions with coefficients in the null space of the indicial family at the corresponding indicial root . + before proceeding ,
let us note that tensoring with @xmath734 does not interact with the constructions of the @xmath674-calculus .
we now improve @xmath705 to a finer parametrix , i.e. one with an even smaller remainder term .
the first step in this is to solve away the expansion of the error @xmath725 at the left face .
this is done as in ( * ? ? ?
* lemma 5.44 ) , using @xmath728 again , and in particular staying within the @xmath674-calculus twisted by @xmath735 this leads to the next parametrix @xmath736 which satisfies @xmath737 for some index set @xmath738 the second step is to form the asymptotic sum of the neumann series @xmath739 this is possible because the kernel of @xmath740 vanishes to order one at the front face and to infinite order at the left face . writing this asymptotic sum as @xmath741
, then @xmath742 and @xmath743 is the residual term @xmath744 for some index set @xmath745 . by construction , @xmath746 ,
where @xmath747 for index sets satisfying @xmath748 and @xmath749 by the same sort of duality argument , we can also construct a parametrix @xmath750 on the left for @xmath701 having constructed these two parametrices , then lemma [ lemfred ] implies abstractly that @xmath672 has a generalized inverse @xmath751 . a further argument , ( * ? ? ? * ( 4.25 ) ) , gives the identity @xmath752 which allows us to understand this generalized inverse as an element in the twisted @xmath674-calculus .
in particular , if there is no indicial root in @xmath753 with real part @xmath754 , ( to understand this shift , recall that @xmath755 ) , then we obtain finally that @xmath756 the index sets here are ( * ? ? ? * ( 4.22 ) ) @xmath757 so far we have been working at a fixed @xmath758 however , we can now apply the analysis in ( * ? ? ? *
( 5.11 - 19 ) ) verbatim .
the generalized inverses of the bessel operators @xmath695 can be reassembled into the generalized inverse of @xmath210 : @xmath759 the change from the parameter @xmath667 to the variable @xmath76 means that we should interpret this not as a @xmath674-operator on @xmath760 but as an element of the @xmath761-calculus of @xcite on @xmath762 indeed , some analysis of this formula leads to the [ prop : invnormalop ] if there is no element of @xmath753 with real part equal to @xmath763 and if @xmath764 is either injective or surjective on @xmath765 , then there is an operator @xmath766 with image in @xmath767 and such that @xmath768 are the orthogonal projections onto the kernel and cokernel of @xmath664 . in particular , for any such @xmath723 the operator @xmath210 ( which implicitly depends on @xmath422 ) has closed range and is either injective or surjective .
the integral kernels of these operators depend smoothly on @xmath769 an advantage of this explicit description for this integral kernel is that we can read off the mapping properties on weighted @xmath0 spaces . to state this ,
define the ` real part ' of an index set , e.g. @xmath770 , by @xmath771 from ( * ? ? ? * theorem 3.25 ) , if @xmath772 then @xmath773 extends to a bounded map @xmath774 if and only if @xmath775 equivalently , in terms of the volume form of an @xmath9-metric , @xmath776 & \text { if and only if } \\[0.3ex ] { \operatorname{re}}g_{10 } & > b'-\tfrac{f+1}2 , \quad { \operatorname{re}}g_{01 } > \tfrac{f+1}2-a ' , \quad { \operatorname{re}}g_{11 } \geq b'-a ' , \end{split}\ ] ] which again uses @xmath777 + since it will be of particular importance below , let us spell out these mapping properties more carefully when the index sets @xmath778 @xmath779 and @xmath780 are as in . define @xmath781 \eta^- = ( \delta - \tfrac{f+1}2 ) - \sup \
{ { \operatorname{re}}\zeta : \zeta \in { \mathrm{spec}}_b(p;{\mathbf{b}},q ) , \ { \operatorname{re}}\zeta < \delta - \tfrac{f+1}2 \}. \end{gathered}\ ] ] then @xmath782 defines a bounded map @xmath783 provided @xmath784 to conclude this section , let us say a bit more about the meaning of `` reassembling '' the integral kernels of the @xmath785 as @xmath147 varies .
let @xmath786 be a distinguished neighborhood of @xmath787 where @xmath568 is the unit ( @xmath520-dimensional ) ball in @xmath788 , and fix a trivialization @xmath789 we define the * zero double space * of @xmath316 by radially blowing up the fiber diagonal of the boundary : @xmath790 } \times z_q^2.\ ] ] this space has a natural blow - down map @xmath791 and we denote by @xmath792 @xmath793 the compositions of this blow - down with the projections onto the left or right factor of @xmath794 we refer to @xmath795 , @xmath796 and @xmath797 as the * left * , * right * and @xmath798 boundary faces , denoted @xmath799 , @xmath800 and @xmath801 , respectively .
choose coordinates @xmath802 on @xmath803 .
then , we define projective coordinates on @xmath804 away from the right face @xmath805 by @xmath806 in these coordinates , @xmath807 defines the left face while @xmath808 is a defining function for the front face .
we can use these coordinates to express the lifts from the left factor of functions and edge vector fields : @xmath809 \beta_l^*(x{\partial}_x ) = s{\partial}_s , \quad \beta_l^*(x{\partial}_y ) = s{\partial}_u , \quad \beta_l^*({\partial}_z ) = { \partial}_z . \end{gathered}\ ] ] the front face is naturally the total space of a fibration with base @xmath810 and fiber a quarter - sphere ; in these coordinates , this is the map @xmath811 ( with @xmath805 since it is at the front face ) . using @xmath812
, we can extend this fibration to a neighborhood of the front face by @xmath813 now suppose that @xmath814 is the integral kernel of @xmath815 from , which is the generalized inverse of @xmath816 .
we say that @xmath817 is the * amalgamation * of these integral kernels ; here @xmath82 a smooth cut - off function on @xmath818 which equals one near the front face . by construction , @xmath819 is compatible with the extended fibration and @xmath820 moreover , the conormal singularity in each @xmath821 at @xmath822 gives rise to a conormal singularity of the same order for @xmath819 at @xmath823 which is the ( interior ) lift of the diagonal in @xmath824 just as for the @xmath674-calculus , see ( * ? ? ?
* proof of proposition 5.43 ) , the advantage of this extension is seen in the behavior of the index set at the left face when we compose with @xmath825 the restriction of the lift of @xmath183 to the fibres composed with the restriction of @xmath627 there is a bounded operator , and we have that @xmath826 .
consider just the expansion at the left face : since @xmath627 has index set @xmath827 there , we would expect @xmath828 to have index set @xmath829 but instead , because of the action of the indicial operator of @xmath183 , it actually has index set @xmath830 at that face . essentially by definition ,
@xmath831 acts on the leading terms of an asymptotic expansion ( see especially ( * ? ? ?
* ( a.9 ) ) )
. however , in the expansion of @xmath832 at the left face , all coefficients of the leading terms in @xmath833 are in the null space of @xmath831 , and hence are annihilated . if , in local coordinates , @xmath834 then in the projective coordinates on @xmath835 @xmath836 } { \mathcal{k}}_{g_{y'+x'u}}(s , u , z , z').\ ] ] clearly , the expansion of @xmath837 at the left face is given by @xmath838 applied to the terms in the expansion of @xmath839 at that face . in particular , we have been careful to match the base points of the indicial family with the extension of the schwartz kernel of @xmath840 so we do indeed obtain the improvement of the index set .
this means that we can omit the leading terms in @xmath841 in the index set of @xmath842 at the left face , and hence take @xmath843 for the index set of @xmath842 there .
finally , note that the amalgamated integral kernels form a @xmath844-twisted @xmath761-pseudodifferential operator .
the indicial roots may vary with the base point @xmath845 , and we refer the reader to @xcite for a detailed treatment of this , see also @xcite . for our purposes , it is enough to have uniform bounds on the index sets .
for example , if @xmath846 is obtained from amalgamating operators with index set @xmath847 as above , and if we assume that @xmath667 from is uniformly bounded below by @xmath848 then @xmath167 defines a continuous map @xmath849
in this section we assume that @xmath1 , with singular strata @xmath55 , ordered by increasing depth , is endowed with a _ rigid _
@xmath9 metric @xmath8 , and we study the associated de rham operator @xmath850 we first show that any choice of cheeger ideal boundary conditions yields a self - adjoint extension .
we then show inductively that assumptions [ ass : cstindicialroots ] , [ assindfamily ] and [ asscmpdomain ] all hold in this case .
this is done through the analysis of its indicial and normal operators .
our first task is to compute the indicial roots of @xmath229 ; in particular , we simplify and sharpen the computation from @xcite and extend it to non - witt spaces .
assume that we have chosen local ideal boundary conditions @xmath482 at the first @xmath36 strata , and work near @xmath769 the link of @xmath1 at @xmath81 is denoted @xmath29 and @xmath93 is a boundary defining function for @xmath483 assume inductively that @xmath851 has compact resolvent , and hence discrete spectrum , as an operator on @xmath852 assume too that there is a strong kodaira decomposition on @xmath29 .
this will be justified in [ sec : l2coho ] . + it is not hard to deduce from and that @xmath853 and we want to understand the indicial roots of this operator .
first , @xmath854 preserves @xmath855 and in this subspace the indicial roots are the values of @xmath368 for which @xmath856 is not invertible .
thus @xmath857 to study the indicial roots on the orthogonal complement of this nullspace , we compute the indicial roots of a closely related operator . conjugating @xmath858 by @xmath859
gives @xmath860 then @xmath861 the two summands here commute , so the next step is to compute the spectrum of the second summand on the right ; this operator is in fact the square of @xmath862 we assume inductively that @xmath863 and hence @xmath864 is self - adjoint , with compact resolvent , and this means that the spectrum of @xmath865 is non - negative and discrete .
thus , fix any @xmath866 write @xmath867 , and decompose @xmath868 , and similarly for @xmath674 , according to the kodaira decomposition on @xmath29 .
( we have already handled the case where @xmath561 and @xmath674 are harmonic . )
then we must solve @xmath869 start with @xmath870 : first apply @xmath871 to both sides of the first equation to get @xmath872}a_d = 0 .
\end{gathered}\ ] ] this last operator factors as @xmath873 we can remove the absolute value from @xmath874 since @xmath875 is symmetric about @xmath761 ; restricting to an eigenspace of @xmath876 with eigenvalue @xmath877 , we find that @xmath878 next consider @xmath879 : applying @xmath880 to both sides of the second equation above gives @xmath881 and hence @xmath882 altogether we conclude that @xmath883 where all of the @xmath884 signs can be chosen independently .
+ on the @xmath885 eigenspace of @xmath886 @xmath887 this means that the indicial roots of @xmath888 are contained in @xmath889 } \ { -k , k - f : 0 \in { \mathrm{spec}}(\delta^z{\big\rvert_{k-\text{forms } } } ) \ } \\
\cup \bigcup_{k \in [ 0,f ] } { \left\lbrace -\tfrac f2\pm \tfrac12 \pm \sqrt { ( k-\tfrac f2 \pm \tfrac12)^2 + \mu } : \mu \in { \mathrm{spec}}(\delta^z{\big\rvert_{k-\text{forms}}})\setminus \ { 0 \ } \right\rbrace}. \end{gathered}\ ] ] this proves the [ lem : suitablescaling ] let @xmath23 be a stratified pseudomanifold with @xmath9 metric , with all notation as before , and @xmath890 a set of cheeger ideal boundary conditions corresponding to a mezzoperversity at the first @xmath36 strata .
let @xmath891 be the dimension of the link of @xmath1 at @xmath439 .
* if @xmath891 is even and the induced hodge laplacians on the links @xmath892 of @xmath1 at @xmath447 satisfy @xmath893 \subseteq \ { 0 \ } \quad { \text { whenever } } |j-\tfrac f2|\leq 1,\ ] ] then @xmath894 \subseteq \ { -\tfrac f2 \}\ ] ] * if @xmath891 is odd and the induced hodge laplacians on the links @xmath892 of @xmath1 at @xmath447 satisfy @xmath895 then @xmath894 \subseteq \ { -\tfrac f2 \pm \tfrac12 \}\ ] ] in particular , any @xmath9 metric @xmath8 can be replaced by another one where the metric on the link at @xmath439 is scaled by a large constant so that these condition as well as _
assumption _ [ ass : cstindicialroots ] are all true .
we say that @xmath8 is * suitably scaled * if the hodge laplacians of the links satisfy the hypotheses of lemma [ lem : suitablescaling ] .
we are abusing notation slightly by regarding this as a property of the metric , since it also depends on the ideal boundary conditions @xmath470 .
[ cor : derhamexp ] let @xmath472 be a stratified pseudomanifold with cheeger ideal boundary conditions @xmath470 at the first @xmath36 strata , and a suitably scaled metric . if @xmath896 is supported in a distinguished neighborhood @xmath897 , @xmath787 then as in lemma [ lem : expansionexistence ] , @xmath898 u_{f/2 } \in h^{-1/2}({\mathbb{b}}^{h } ; \ker i(x\eth_{{\operatorname{dr } } } ; { \mathbf{b } } , -f/2 ) ) { \text { and } } { \widetilde{u } } \in x l^2 ( c(z_q ) , h^{-1}({\mathbb{b}}^h)\otimes \lambda^*({{}^{{\operatorname{iie}}}}t^*x ) ) .\end{gathered}\ ] ] if @xmath899 is odd or @xmath900 then @xmath901 ; in all other cases , @xmath902 to put this into perspective , recall that @xmath903 ; thus since @xmath904 for every @xmath905 the expansion here is of the form studied in lemma [ lem : expansionexistence ] .
we now turn to a consideration of the normal operator of @xmath229 at a point @xmath484 .
the link of @xmath1 at @xmath81 is denoted @xmath29 and @xmath93 is a boundary defining function for @xmath483 we assume that the metric is suitably scaled , and that there is kodaira decomposition for forms on @xmath29 as in
[ sec : l2coho ] .
the strategy is to construct a parametrix for @xmath906 near @xmath439 using the inverses of its normal operators @xmath907 however , we must consider these model operators acting on weighted @xmath0 spaces since they are not invertible on @xmath0 itself . for any @xmath908 define @xmath909 since we will be working with @xmath9-differential forms on @xmath910 it simplifes notation to write @xmath911 the normal operator at @xmath81 acts on sections of @xmath912 pulled - back to @xmath316 and restricted to the fiber over @xmath81 , which can be identified with @xmath913 ( the graded tensor product of forms on @xmath914 and @xmath9 forms on @xmath29 ) .
thus we also write @xmath915 in @xcite we have shown that the two assumptions @xmath916 imply that @xmath917 is injective for all @xmath918 _ now consider the situation where ( @xmath561 ) holds but ( @xmath674 ) does not . _
thus we assume that @xmath919 is an even integer and , letting @xmath920 be the domain of @xmath229 induced by the boundary conditions @xmath921 that @xmath922 because of the inductive hypothesis that the de rham operator on the fibers is fredholm and self - adjoint , and that the de rham complex on the fibers satisfies a strong kodaira decomposition and a hodge theorem , we can use the computation in ( * ? ? ?
* lemma 5.5 ) here .
in particular , with @xmath923 this shows that elements in @xmath924 arising from the fact that b ) no longer holds are necessarily of the form @xmath925 with @xmath926 from , on forms of middle vertical degree , the equation @xmath927 reduces to @xmath928 setting @xmath929 , then these become @xmath930 \eth^{{\mathbb{r}}^b}_{{\operatorname{dr}}}\beta ' = { \partial}_s \alpha ' .
\end{cases}\ ] ] since @xmath931 and @xmath932 commute , implies that @xmath933 .
taking fourier transform in @xmath934 ( with dual variable @xmath667 ) , we find that @xmath935 however , a solution lying in any @xmath765 must have @xmath936 .
moreover , @xmath937 so the @xmath0 condition becomes @xmath938 given @xmath939 then determines @xmath940 , so altogether @xmath941 where @xmath942 .
since the relation between @xmath943 and @xmath944 in @xmath945 is symmetric , we have @xmath946 where @xmath947 is the involution @xmath948 [ prop : nullnormal ] the operator @xmath949 is injective for @xmath950 and has closed range if @xmath951 , @xmath952 .
moreover if @xmath953 , then @xmath954 observe that the space is an infinite dimensional space and hence the normal operator is fredholm if and only if it is invertible .
closedness of the range follows from proposition [ eq : gnqp ] .
the other statements follow directly from the computation . since , by induction , the de rham operator with cheeger ideal boundary conditions coming from a mezzoperversity is self - adjoint , see theorem [ thm : selfdual ] , the following holds by duality .
[ cor : normalopsurj ] the normal operator with domain @xmath955 , defined by duality from @xmath956 is surjective when @xmath957 this description of the null space yields a direct sum decomposition of the trace bundle of the de rham operator at @xmath86 assume now that @xmath958 is a sub - bundle of @xmath959 with @xmath960 the fibration with typical fiber @xmath961 [ lem : decomp ] there is a direct sum decomposition @xmath962 where @xmath963 note that @xmath964 and so @xmath965 thus we only need to show that @xmath966 contains the left hand side of .
decompose any @xmath967 as @xmath968 according to the splitting @xmath969 .
let @xmath970 and notice that @xmath971 we can now decompose an arbitrary pair @xmath972 as @xmath973 define @xmath974 then by lemma [ lem : expansionexistence ] , elements in @xmath975 have a partial asymptotic expansion at @xmath976 the same indicial root computation of the de rham operator for a suitably scaled @xmath9 metric implies that this partial asymptotic expansion has only one term , and this has exponent @xmath977 , and coefficient in .
denote the corresponding cauchy data map by @xmath978 @xmath979 from lemma [ lem : decomp ] , there is a ( not necessarily orthogonal ) projection into the null space @xmath980 indeed , if @xmath981 is the natural projection , and @xmath982 then we can set @xmath983 and it follows from that @xmath984 maps into @xmath985 [ lem : phieps ] the operator @xmath984 is a twisted @xmath761-pseudodifferential operator @xmath986 with index set @xmath987 using the notation .
we compute the integral kernel of @xmath988 by corollary [ cor : normalopsurj ] , @xmath989 is surjective , hence proposition [ prop : invnormalop ] can be applied with @xmath990 first the integral kernel of @xmath991 can be obtained from the generalized inverse @xmath992 of @xmath993 notice that @xmath994 so it suffices to focus on those @xmath995 orthogonal to @xmath996 thus @xmath997 we are looking for the coefficient of @xmath859 in the expansion of the left hand side . since @xmath998 and@xmath999 this coefficient comes from the expansion of @xmath1000 at the left face , @xmath1001 of @xmath1002 indeed , there is a term @xmath1003 in the expansion of @xmath1000 at the left face , and @xmath1004 is given by the action of @xmath1005 on @xmath1006 it is convenient to write this coefficient in terms of the generalized inverse of the bessel - type operator @xmath701 indeed , from ( cf .
* ( 5.20 ) ) ) we have @xmath1007 where @xmath1008 is the coefficient in the corresponding coefficient in the asymptotic expansion of @xmath1009 at the left face . thus for all @xmath995 orthogonal to @xmath1010 we have @xmath1011 and @xmath1012 is given by the action of @xmath1013 on @xmath1006 so to see that @xmath984 acts as a @xmath844-twisted @xmath761-pseudodifferential operator , we need only note that @xmath1014 is such an operator . directly from this expression we see that @xmath1015 on the other hand , the restriction of the kernel of @xmath751 to the left face has index set @xmath1016 at the front face and @xmath1017 at the right face , and so we can take these as the index sets for @xmath1018 finally , for general @xmath1019 we have @xmath1020 and we know that @xmath1021 has index sets @xmath770 at the left face and @xmath1022 at the right face .
continuing as above , fix @xmath1023 and the boundary operators @xmath482 at the first @xmath36 singular strata of @xmath4 assume that @xmath439 is * non - witt * for @xmath1024 in the sense that @xmath1025 even and @xmath1026 as explained in [ sec : arbitrarydepthbvp ] the union of these null spaces form a flat vector bundle over @xmath439 .
we now choose a flat sub - bundle @xmath1027 and let @xmath1028 be the cheeger ideal boundary condition associated to @xmath1029 in this section we show that any @xmath1030 supported in a distinguished neighborhood of @xmath447 is in @xmath1031 + fix a distinguished neighborhood @xmath1032 of @xmath484 , a trivialization @xmath1033 and flat trivializations @xmath1034 we start by modifying the generalized inverse @xmath1035 of the normal operator @xmath854 with domain @xmath1036 ( this is surjective by corollary [ cor : normalopsurj ] ) using the projector @xmath984 from lemma [ lem : phieps ] .
recall that @xmath1037 in we defined a domain for @xmath975 using the first @xmath36 boundary conditions @xmath470 .
we refine this to @xmath1038 where @xmath1039 is as in .
now define @xmath1040 the premultiplication by @xmath156 serves to mediate between @xmath854 and @xmath1041 [ prop : other ] for any choice of cheeger ideal boundary conditions @xmath1042 the operator @xmath1043 satisfies @xmath1044 hence @xmath1045 is invertible with inverse @xmath1046 moreover , the composition @xmath1047 is a bounded operator on @xmath1048 the operator @xmath1035 satisfies @xmath1049 where @xmath1050 is the orthogonal projection onto the null space of @xmath854 with domain @xmath1051 this null space coincides with that of @xmath1052 so @xmath1053 moreover , we know that @xmath1054 is the identity on @xmath1055 and hence @xmath1056 this shows @xmath1057 and so @xmath1058 putting these observations together , we have shown @xmath1059 and so @xmath1060 is the generalized inverse of @xmath1061 + next consider the operator .
since @xmath1062 we have @xmath1063 this shows that @xmath1064 makes sense , so directly from the definition of @xmath1043 @xmath1065 moreover , @xmath1066 since @xmath1067 directly from the definition of @xmath1039 we have @xmath1068 on @xmath1069 so indeed @xmath1070 is invertible with inverse @xmath1046 + to prove that @xmath1047 is a bounded operator , first consider @xmath1071 using that @xmath1072}g_q.\ ] ] since @xmath1073 , the composition of the second summand above with @xmath1035 is bounded .
it is clear from the definition of @xmath984 that its image consists of sections of the vertical hodge bundle and so @xmath1074 thus @xmath1075 note that this means that both @xmath1076 and @xmath1077 are bounded operators .
let us point out some further properties of @xmath1046 we know from proposition [ prop : invnormalop ] that @xmath1078 with @xmath780 given by . from lemma [ lem : phieps ]
, @xmath1079 thus altogether , @xmath1080 using ( * ? ? ?
* theorem 3.15 ) , the composition formula for @xmath761-pseudodifferential operators , the index sets are given by @xmath1081 ( the operator @xmath1082 lies in a similar space , but with @xmath1083 replaced by @xmath1084 because of the factor @xmath156 on the right in the definition of @xmath1046 ) note that since @xmath1085 the indicial operator of @xmath906 must annihilate all of the terms in the expansion of @xmath1086 at the left face ( see the discussion above ) .
+ we can read off the mapping properties of @xmath1087 from .
indeed , in terms of @xmath1088 and @xmath1089 from , we have @xmath1090 and hence @xmath1087 defines a bounded map @xmath1091 provided that @xmath1092 using lemma [ lem : suitablescaling ] we can be more explicit : @xmath1093 the corresponding values for @xmath1043 are @xmath1094 @xmath1095 we now restrict to the distinguished neighborhood @xmath316 . following the procedure in [ sec : invnormal ] , we amalgamate the family of operators @xmath1087 to an operator @xmath1096 on @xmath1 supported in this neighborhood .
thus @xmath1096 is defined through its integral kernel @xmath1097 where @xmath82 is a smooth cut - off function equal to one in a neighborhood @xmath1098 we define @xmath1099 similarly , @xmath1100 as in ( * ? ? ?
* ( 3.5 ) ) , @xmath1099 acts on a section @xmath891 by @xmath1101 for the appropriate measure @xmath1102 notice two features here : first , since @xmath1103 @xmath1104 and second , @xmath1105 from we see that @xmath1106 so @xmath1096 satisfies properties analogous to those of @xmath1107 while @xmath1099 satisfy similar mapping properties with @xmath1108 as long as @xmath1109 since @xmath1110 vanishes to first order at both the front face and the left face .
( note that though the index sets vary with @xmath81 , the bounds on their real parts , hence the mapping properties , discussed above are uniform in @xmath184 ) let us show that @xmath1111 from the way @xmath1096 acts , we see that @xmath1112 where @xmath1113 is the vertical de rham operator for the metric at @xmath1114 since the metric @xmath8 is rigid ( in fact , this argument works for a slightly larger class of metrics ) , we have @xmath1115 . by , this composes with @xmath1116 to a bounded operator on @xmath44 we have shown that @xmath1096 satisfies the assumption preceeding , so if @xmath1117 is the collection of index sets for @xmath1118 , then @xmath1119 these , in turn , imply that @xmath1120 is bounded so long as @xmath1121 in particular , we have shown that @xmath1122 is bounded , and hence @xmath1123 is a bounded operator on @xmath42 essentially the same reasoning shows that @xmath1124 is also a bounded operator on @xmath42 + we have established that the image of @xmath1099 lies in @xmath1125 in fact , since @xmath1126 we have the the operator @xmath1099 maps @xmath1127 into @xmath1128 if @xmath1129 then @xmath1130 so @xmath1131 has a distributional partial asymptotic expansion with leading term @xmath1132 consistent with this , the integral kernel of @xmath1043 has an asymptotic expansion at the left face with leading term @xmath1133 the coefficient @xmath1134 is obtained by applying @xmath1135 to @xmath1136 it follows that @xmath1135 takes values in @xmath1137 similar reasoning can be applied to @xmath1138 from the discussion above , @xmath1099 maps into @xmath1125 by , the expansion of @xmath1139 at a singular stratum of @xmath892 is a parametrized version of the expansion at this stratum of the individual @xmath1140 since the latter take values in the space determined by the cheeger ideal boundary conditions @xmath921 so is the former .
thus @xmath1099 maps into @xmath1141 finally , consider the expansion as @xmath1142 by construction , the asymptotic expansion of @xmath1143 at the left face is the amalgam of the asymptotic expansions of the individual @xmath1144 , so its leading term at the left face is @xmath1145 in particular , sections in the image of @xmath1099 have a partial asymptotic expansion as @xmath1146 with leading term a distributional section of @xmath1147 thus @xmath1099 maps into @xmath1148 define @xmath1149 , where @xmath1150 is a cut - off function on a neighborhood of @xmath18 smaller than the one to which we extended @xmath1099 . from the mapping properties of @xmath1151 if @xmath1152 then @xmath1153 are bounded , so we can apply lemma [ lem : indopconclusion ] .
[ thm : witlessdecay ] if @xmath472 is a stratified space with singular strata @xmath55 , ordered with increasing depth and rigid @xmath9 metric @xmath8 .
suppose that @xmath1154 are cheeger ideal boundary conditions at the first @xmath37 strata , and @xmath439 is non - witt ( so holds there ) .
then @xmath1155 we may assume that @xmath76 is supported in a distinguished neighborhood @xmath316 of @xmath769 first suppose that @xmath1156 referring to lemma [ lem : indopconclusion ] for notation , identify domains to which we can apply that lemma .
set @xmath1157 from corollary [ cor : derhamexp ] , @xmath1158 and directly from its definition we have @xmath1159 now @xmath1160 is a bounded operator on @xmath1127 so the lemma guarantees that there is at most one extension of @xmath1161 from @xmath633 to @xmath1162 thus knowing that @xmath1163 and that both sides extend to @xmath1164 we must also have @xmath1165 moreover @xmath1166 for any @xmath1167 so we just need to show that @xmath1168 has extra vanishing . + to discuss the normal operator of @xmath707 we recall that @xmath1169 the normal operator of @xmath1170 at the point @xmath1171 is the operator @xmath1172 from lemma [ lem : phieps ] .
let @xmath1173 be the amalgamation of the family @xmath1174 as in , so that we have @xmath1175 this operator satisfies with @xmath1109 so in particular @xmath1176 finally , since @xmath1177 if @xmath1178 is a section of @xmath238 over @xmath439 then clearly @xmath1179 this establishes the extra decay when @xmath1180 it was necessary to restrict to these @xmath217 because we needed to take @xmath1181 in .
but we can now repeat the argument starting with any @xmath217 , which allows us to take @xmath1182 in . by inspection
, we can thus take any @xmath1183 iterating this @xmath1184 times , we establish decay for any @xmath1185 , and hence since @xmath1184 is arbitrary , for any @xmath1186 .
this proves .
now return to the situation at the beginning of this section .
namely assume that @xmath482 are local ideal boundary conditions for the de rham operator of a rigid @xmath9 metric on @xmath4 we say that @xmath439 is a * witt stratum * for @xmath1187 either if @xmath891 is odd , or else if @xmath891 is even and @xmath1188 the argument in @xcite now shows that @xmath1189 is invertible for all @xmath918 we prove here that any @xmath896 ( supported in the distinguished neighborhood @xmath316 ) must lie in @xmath1190 the construction in this case is simpler than the one above for a non - witt stratum .
fix @xmath918 since @xmath917 is invertible , the construction in [ sec : invnormal ] produces an inverse @xmath1191 arguing as in proposition [ prop : other ] , we see that @xmath1060 is an inverse for @xmath1192 with domain @xmath1193 and that @xmath1194 is a bounded operator on @xmath1195 continuing on , define the integral kernel @xmath1196 of @xmath1197 by amalgamating the family of operators @xmath1035 .
this lies in @xmath1198 ; the index sets are such that @xmath1199 is bounded , and @xmath1200 extends from @xmath1127 to a bounded operator on @xmath44 we can thus apply the integration by parts trick as before , and so obtain the analogue of theorem [ thm : witlessdecay ] .
[ thm : wittydecay ] let @xmath23 and @xmath470 be as in theorem [ thm : witlessdecay ] , but suppose that holds at @xmath439 , so this is a witt stratum .
then @xmath1201 we have shown how to construct a left parametrix at all singular strata .
this construction on spaces of depth @xmath37 relies on inductive information about the de rham operator on singular spaces of depth less than or equal to @xmath36 , and in particular that this operator with cheeger ideal boundary conditions from a mezzoperversity is self - adjoint .
we now use these parametrices to prove that this de rham operator is essentially injective , i.e. , has closed range and finite dimensional null space .
then , once we have proved theorem [ thm : selfdual ] , which asserts that the de rham operator on spaces of depth @xmath37 is self - adjoint , then this essential injectivity proves that it must actually be fredholm .
[ thm : mainhodgethm ] consider @xmath472 as before , where @xmath8 is suitably scaled , and @xmath575 a set of cheeger ideal boundary conditions .
then @xmath229 with domain @xmath1202 is closed , essentially injective and has compact resolvent . indeed if @xmath506 is a total boundary defining function on @xmath299 the resolution of @xmath2 then @xmath1203 in particular , the de rham operator satisfies _ assumptions _ [ assindfamily ] and [ asscmpdomain ] at depth @xmath37 .
assume that this result is true for all spaces of depth less than or equal to @xmath36 .
if @xmath1204 and @xmath82 is an indicator function supported in the distinguished neighborhood @xmath316 of @xmath484 , then @xmath1205 , and using the inductive hypothesis and theorems [ thm : witlessdecay ] and [ thm : wittydecay ] , @xmath1206 since @xmath81 is arbitrary , @xmath76 itself lies in this space . since this space is compactly included in @xmath1207 this shows that @xmath1208 is essentially injective and has compact resolvent .
let @xmath472 be a stratified pseudomanifold endowed with a rigid @xmath9 metric .
in this section we relate the domains of the de rham operator to domains of the exterior derivative .
useful consequences include the self - adjointness of the de rham operator , a kodaira decomposition for @xmath0 differential forms , and an identification of the hodge cohomology associated to a mezzoperversity with the associated de rham cohomology .
+ the exterior derivative @xmath11 and its formal adjoint @xmath422 have two canonical closed extensions from the smooth forms of compact support on @xmath1209 namely , @xmath1210 in the first case , ` cauchy ' means cauchy in @xmath479 while in the second @xmath1211 is computed distributionally .
we define in the same way @xmath1212 and @xmath1213 it is clear from these definitions that @xmath1214 and similarly , with @xmath64 and @xmath66 exchanged .
+ to identify the adjoints of domains ( for @xmath1215 @xmath422 or @xmath229 ) intermediate between @xmath64 and @xmath1216 introduce the boundary pairings @xmath1217 where @xmath1218 is the @xmath0 pairing .
these are closely related : in fact , @xmath1219 can be expressed in terms of @xmath1220 as follows .
let @xmath1221 note that since @xmath1222 we have @xmath1223 similarly , if @xmath1224 denotes the orthogonal projection of @xmath995 onto @xmath1225 we have @xmath1226 thus it is always true that @xmath1227 moreover , since @xmath1228 we have @xmath1229 and similarly @xmath1230
so we also have @xmath1231 thus altogether , @xmath1232 let us start by considering the setting of @xmath1 with a single singular stratum @xmath86 for any @xmath1233 denote the leading term in its expansion from lemma [ lem : expansionexistence ] and corollary [ cor : derhamexp ] by @xmath1234 recall from lemma [ lem : regulardomain ] the core subdomain @xmath1235 defined by requiring that @xmath1236 and @xmath1237 are smooth instead of just distributional sections .
[ lem : stokesthmedge ] if @xmath1238 and one of them is in @xmath1235 then from stokes theorem @xmath1239 where on @xmath18 we have the pairing between smooth forms and currents with coefficients in the vertical hodge bundle , induced by the @xmath0-pairing .
we always have @xmath1240 if both @xmath1241 are in @xmath1235 then @xmath1242 and so @xmath1243 if only one of them , say @xmath76 is in @xmath1244 we can find @xmath1245 such that @xmath1246 in the graph norm . by continuity of the coefficients of the asymptotic expansion in lemma [ lem : expansionexistence ]
, we have @xmath1247 as distributions , so @xmath1248 the pairing @xmath1220 is analogous .
this lemma shows the usefulness of having elements in @xmath1244 so we define analogous domains for @xmath11 and @xmath1249 the domain @xmath1250 is a core domain for @xmath16 by the same argument that showed that @xmath1235 is a core domain for @xmath1251 similarly we define @xmath1252 and point out that this is a core domain for @xmath1213 [ lem : prescribedforms ] let @xmath316 be a distinguished neighborhood of @xmath147 over which we have trivializations @xmath1253 where we recall that @xmath1254 refers to the @xmath0 harmonic forms on @xmath892 of degree equal to half of the dimension of @xmath1255 for differential forms @xmath1256 supported in @xmath316 we have @xmath1257 moreover , given @xmath1258 * there exists @xmath1259 such that @xmath1260 @xmath1261 * there exists @xmath1262 such that @xmath1263 @xmath1264 * for any @xmath1265 compactly supported in @xmath316 with @xmath1266 @xmath1267 there exists @xmath1268 such that @xmath1269 let @xmath703 be a smooth function identically equal to one on the support of @xmath1270 and vanishing outside a compact subset of @xmath794 + let @xmath1271 and let @xmath1272 since @xmath76 and @xmath995 are smooth and in @xmath1273 we clearly have @xmath1274 also @xmath1275 and @xmath1276 + now we can identify @xmath1277 and we have just shown that @xmath1278 can be prescribed arbitrarily , so @xmath1279 precisely when @xmath1280 similarly , from @xmath1281 we can characterize @xmath1212 as required .
moreover we can now recognize that @xmath1282 so we have established ( @xmath1283 ) and ( @xmath1284 ) .
+ for ( @xmath1285 ) , let @xmath1286 in @xmath1287 we know from , that we can write @xmath1288 and we know from lemma [ lem : expansionexistence ] that @xmath1289 is in the null space of @xmath1290 it follows from the kodaira decomposition on @xmath892 and the explicit expressions for @xmath11 and @xmath422 that @xmath1291 since @xmath1292 are both closed domains for @xmath1293 they both contain @xmath1294 and hence so does their intersection @xmath1295 on the other hand , in the pairing @xmath1296 this lemma shows that we can prescribe @xmath1297 and @xmath1298 arbitrarily and so @xmath1299 in particular , if @xmath1265 then we have @xmath1300 @xmath1095 another useful consequence of this lemma is that whenever @xmath1301 @xmath1302 indeed , we have @xmath1303 for any @xmath1304 and from the lemma we can prescribe @xmath1305 arbitrarily .
this shows that @xmath1306 and a similar argument shows @xmath1307 + returning to , we see that whenever @xmath1308 @xmath1309 and either @xmath1310 or @xmath1224 is in @xmath1244 @xmath1311 and so @xmath1312 let us see that this is a reflection of the flatness of the bundle of vertical harmonic forms over @xmath86 [ lem : alphanabla ] for @xmath1308 @xmath1313 @xmath1314 where @xmath265 is the flat connection on the bundle of vertical harmonic forms discussed in [ sec : simpleedgebc ] .
any @xmath1315 has a partial asymptotic expansion @xmath1316 if @xmath76 is polyhomogeneous then it has a full asymptotic expansion @xmath1317 for some index set @xmath1318 from the form of the partial asymptotic expansion above , we must have @xmath1319 consider a differential form @xmath76 such that @xmath1320 applying @xmath11 to the expansion of @xmath76 we have @xmath1321 on the other hand , we have @xmath1322 with @xmath1323 comparing coefficients we must have @xmath1324 for some form @xmath1325 from the kodaira decomposition , @xmath1326 we see that projecting @xmath1327 onto the harmonic forms comes down to subtracting an exact form , say @xmath1328 thus we have @xmath1329 however , note that the left hand side is harmonic while the right hand side is exact .
another appeal to the kodaira decomposition shows that both sides must vanish , so we have shown @xmath1330 similar reasoning shows that whenever @xmath1331 we have @xmath1332 now for any @xmath1308 let us take @xmath995 such that @xmath1333 and consider the boundary pairings @xmath1334 on the one hand , we can write @xmath1335 while on the other we have @xmath1336 since from lemma [ lem : prescribedforms ] @xmath1305 can be arbitrarily prescribed , we conclude that @xmath1337 similarly we see that @xmath1338 thus in , @xmath1339 is in the closure of the image of @xmath1340 while @xmath1341 is in the closure of the image of @xmath1342 since the connection is flat these images are orthogonal , which explains . + next
consider cheeger ideal boundary conditions , @xmath921 corresponding to the flat bundle @xmath1343 we define @xmath1344 and then set @xmath1345 and @xmath1346 equal to the closures in the respective graph norms .
the discussion above allows us to compute the adjoints of these domains and then their closures .
first , @xmath1347 is equivalent to @xmath1348 for all @xmath1349 i.e. , @xmath1350 from lemma [ lem : prescribedforms ] , we can prescribe @xmath1351 arbitrarily among smooth forms on @xmath18 with coefficients in @xmath1029 thus @xmath1352 this domain has a core given by @xmath1353 it is easy to see that a form @xmath1354 will be in @xmath1355 precisely when @xmath1356 this pairing is equal to @xmath1357 and from lemma [ lem : prescribedforms ] we can prescribe @xmath1278 arbitrarily , so we conclude that @xmath1358 a similar computation lets us identify @xmath1346 and we find @xmath1359 + since @xmath1360 is a closed operator , @xmath1361 is a closed domain for @xmath850 notice that since @xmath1362 is in @xmath1363 we can use observation to replace @xmath1351 and @xmath1364 by @xmath1297 and @xmath1365 thus directly from its description we have @xmath1366 since this domain clearly contains @xmath1367 and the only closed domain for @xmath229 containing @xmath1368 is @xmath1369 we have the first part of the following proposition : [ prop : derhamcohosimpleedge ] let @xmath1 be a stratified pseudomanifold with a single singular stratum and @xmath470 a cheeger ideal boundary condition corresponding to a mezzoperversity at @xmath86 we have @xmath1370 so in particular @xmath1208 is self - adjoint .
together with theorem [ thm : mainhodgethm ] , this shows that it is also fredholm .
+ the differential forms in @xmath1345 together with @xmath11 form a hilbert complex , whose cohomology we denote @xmath1371 since @xmath1208 is fredholm , this is a fredholm complex and we have @xmath1372 moreover , the domain @xmath1345 and the cohomology @xmath1373 are independent of the choice of metric ( within suitably scaled rigid @xmath9 metrics ) .
the discussion above the proposition shows that @xmath1374 we know that @xmath1360 is a closed operator , so it will form a hilbert complex iff @xmath1375 which , from the definition of @xmath1376 is equivalent to @xmath1377 which follows from the flatness of @xmath1029 the well - known general theory of hilbert complexes of brning - lesch @xcite implies that this complex is a fredholm complex satisfying a kodaira decomposition and a hodge theorem . + for metric independence , let us start with @xmath1378 compactly supported in a distinguished neighborhood @xmath316 of @xmath1379 if we denote @xmath1380 then we know that @xmath1381 and that @xmath1382 so from , @xmath1383 now if @xmath1384 is another suitably scaled rigid @xmath9 metrics on @xmath477 then it is quasi - isometric to @xmath8 since they both define metrics on @xmath298 over all of @xmath88 which is compact .
thus the sets @xmath1385 and @xmath1386 coincide and the @xmath0-norms are equivalent . directly from the definitions we see that @xmath1387 the form @xmath1388 is in @xmath1389 with leading tangential term @xmath1390 hence it is in @xmath1391 and so from the description above we have @xmath1392 finally , the graph closure of @xmath1393 with respect to @xmath8 is the same as the graph closure with respect to @xmath1394 so we conclude that @xmath1395 now consider a general stratified space @xmath1 , with strata @xmath55 and cheeger ideal boundary conditions with mezzoperversity , @xmath470 .
recall that for any @xmath1354 we denote @xmath1310 the orthogonal projection onto @xmath1396 and that @xmath1397 similarly given @xmath1398 we denote the orthogonal projection onto @xmath1399 by @xmath1224 and this is an element of @xmath1251 define @xmath1344 and then set @xmath1400 and @xmath1401 equal to the closures in the respective graph norms . in this section
we will show that @xmath1402 so that the de rham cohomology and hodge cohomologies with cheeger ideal boundary conditions coincide . in the next section
we will show that surprisingly the de rham cohomology is actually independent of the choice of metric ( among suitably scaled , rigid @xmath9 metrics ) .
+ for @xmath1403 we will denote the leading term at the stratum @xmath72 by @xmath1404 we will make strong use of the boundary pairing @xmath1219 from the previous section .
we start with @xmath1347 which is equivalent to @xmath1309 and @xmath1405 for all @xmath1406 we know that @xmath1407 and so it has a distributional asymptotic expansion at @xmath315 choosing @xmath1378 supported in a distinguished neighborhood of a point @xmath566 we see that @xmath1408 and as before we can choose @xmath1409 arbitrarily so we must have @xmath1410 we have a similar domain for @xmath11 @xmath1411 and since @xmath1412 is a closed operator extending @xmath1413 we must have @xmath1414 thus we have shown that to determine @xmath1415 we can restrict the pairing @xmath1219 to @xmath1416 [ lem : defmindomb1 ] in terms of @xmath1417 the domains @xmath1418 are , respectively , equal to @xmath1419 and @xmath1420 notice that @xmath1421 is a closed domain for @xmath422 since it is the graph closure of @xmath1422 if @xmath1423 and @xmath1424 then @xmath1425 so @xmath1426 on the other hand , if @xmath1256 is such that @xmath1427 then restricting to those @xmath995 supported in a distinguished neighborhood of @xmath566 we have a continuous map @xmath1428 letting the @xmath0-norm of @xmath995 go to zero while keeping @xmath1305 fixed shows that we must have @xmath1429 and hence that @xmath1430 thus we have shown that @xmath1431 and so @xmath1432 is contained in the closure of @xmath1433 which is @xmath1434 putting these together , @xmath1435 and reversing the roles of @xmath11 and @xmath422 yields @xmath1436 since @xmath1437 is a closed domain we have @xmath1438 given @xmath1439 let us denote its projection onto the second summand by @xmath1440 notice that , for any @xmath1441 we have @xmath1442 and @xmath1443 so @xmath1444 and in particular @xmath1445 has an asymptotic expansion at @xmath315 since @xmath1445 differs from @xmath1224 by an element of @xmath1446 we see that @xmath1447 and since @xmath1448 we have , using the analogue of lemma [ lem : alphanabla ] and that @xmath389 is a flat bundle , @xmath1449 thus altogether we have shown the important fact that @xmath1450 which shows that @xmath1451 has a distributional asymptotic expansion at @xmath328 let us similarly use the decomposition @xmath1452 to define @xmath1453 to be the projection of @xmath1354 onto the second summand .
we have @xmath1454 the pairing @xmath1455 restricted to @xmath1456 satisfies @xmath1457 if @xmath1378 is supported in a distinguished neighborhood of @xmath329 then , just as in lemma [ lem : stokesthmedge ] , stokes theorem shows that @xmath1458 and , as in lemma [ lem : prescribedforms ] , we may pick @xmath1459 arbitrarily among the smooth sections of @xmath1460 so we conclude that @xmath1461 we have a similar domain for @xmath1215 @xmath1462 and reasoning as above shows that @xmath1463 @xmath1095 we now continue inductively in this way to define @xmath1464 for all @xmath1465 and show that @xmath1466 once this is done for the first @xmath1467 strata , then the argument above shows @xmath1468 since at this stage we have @xmath1347 if and only if @xmath1348 for all @xmath1378 supported in a distinguished neighborhood of a point @xmath1469 this final domain has a useful core domain @xmath1470 computations analogous to those carried out above show that @xmath1471 and that @xmath1472 finally , directly from the definition of @xmath1202 we can conclude that @xmath1473 to see that the final two intersections are superfluous we can proceed as we did on spaces of depth one .
indeed , the same argument we gave there shows that @xmath1474 and iterating we find @xmath1475 now since the latter space is a closed domain for @xmath229 it must contain the closure of the former , namely @xmath1476 [ thm : selfdual ] if @xmath472 is a stratified pseudomanifold with a suitably scaled @xmath9 metric and @xmath1477 flat bundles forming a mezzoperversity with associated cheeger ideal boundary conditions @xmath921 then @xmath1360 forms a hilbert complex with dual complex @xmath1478 and associated de rham operator @xmath1479 in particular , using theorem [ thm : mainhodgethm ] , this operator is fredholm and so @xmath1360 is a fredholm complex with corresponding kodaira decomposition and hodge theorem : @xmath1480 let @xmath472 be a stratified pseudomanifold with a rigid , suitably scaled @xmath9 metric , with hodge mezzoperversity @xmath1481 as in definition [ def : flatsystem ] , with associated cheeger ideal boundary conditions @xmath1482 recall that the definition of the mezzoperversity is that @xmath1483 we now want to show that the de rham cohomology is independent of the choice of appropriate @xmath9 metric . to state
this properly , we need to rephrase the boundary conditions slightly since the hodge cohomology will not be independent of the choice of metric .
let @xmath472 be a stratified pseudomanifold , as above .
a * de rham mezzoperversity * consists of a list of bundles @xmath1484 with @xmath1485 given a metric @xmath8 , a hodge mezzoperversity @xmath1486 is equivalent to a de rham mezzoperversity ; we denote this by @xmath1487,$ ] since we have shown that the hodge cohomology groups coincide with the de rham cohomology groups . the following result , using the notation from [ sec : graldepthderham ] , allows us to identify domains defined using different metrics .
[ lem : coefisexact ] let @xmath1488 be a stratified pseudomanifold with a suitably scaled rigid @xmath9 metric and hodge mezzoperversity .
let @xmath93 be a boundary defining function for @xmath456 @xmath29 the link of @xmath1 at @xmath439 and @xmath1489 if @xmath1490 has an expansion at @xmath439 with leading term @xmath1491 then @xmath1492 consider the pairing @xmath1493 given as usual by @xmath1494 since @xmath1495 is orthogonal to the image of @xmath723 @xmath1496 we have @xmath1497 if @xmath1498 is supported in a distinguished neighborhood of @xmath484 then we have @xmath1499 ( note that since @xmath1236 is not a harmonic form , this pairing does not descend to @xmath483 ) as this vanishes for all @xmath995 and we can prescribe @xmath1500 arbitrarily among @xmath9-forms on @xmath18 with coefficients in the harmonic forms on @xmath29 satisfying the boundary conditions @xmath1501 we can conclude that @xmath1236 is a closed form perpendicular to harmonic forms , i.e. , exact : @xmath1502 since @xmath1503 is closed .
[ thm : metricindep ] let @xmath1 be a stratified pseudomanifold as above .
let @xmath8 and @xmath1384 are suitably scaled rigid @xmath9 metrics on @xmath1 with hodge mezzoperversities @xmath1486 and @xmath1504 respectively , so that @xmath1486 and @xmath1505 are both equivalent to the same de rham mezzoperversity , @xmath1487 $ ] . then @xmath1506 thus the de rham cohomology groups @xmath1507}({\widehat{x}})$ ] are independent of the choice of suitably scaled rigid @xmath9 metric .
it suffices to prove that @xmath1508 as this implies that the de rham cohomologies are isomorphic and hence so are the hodge cohomologies .
+ if @xmath1 is a space of depth one , then the result was proven in proposition [ prop : derhamcohosimpleedge ] .
we now inductively carry out the same argument in the general case .
since @xmath8 and @xmath1384 are quasi - isometric , we have @xmath1509 given @xmath1354 its projection @xmath1310 from will be different for @xmath8 or @xmath1394 so we denote them by @xmath1510 and @xmath1511 respectively .
we know that @xmath1512 and @xmath1513 are both in @xmath1514 hence so is @xmath1515 since both @xmath1510 and @xmath1511 have asymptotic expansions at @xmath270 of the sort required in lemma [ lem : coefisexact ] , we can conclude that @xmath1516 and so we have the crucial fact @xmath1517 = [ \alpha(u_{\delta_{g ' } } ) ] { \text { in } } { \operatorname{h}}^{{\operatorname{mid}}}(h^1/y^1).\ ] ] note that the hodge theorem allows us to write @xmath1518 \text { is a current on $ y^1 $ with coefficients in $ [ w^1]$ } \}\ ] ] and so our observation implies @xmath1519 then from lemma [ lem : defmindomb1 ] it follows that @xmath1520 inductively the same argument shows that once we know @xmath1521 then we see that @xmath1522 and from lemma [ lem : coefisexact ] that @xmath1523 = [ \alpha(u_{\delta_{g ' } , ( b^1 , \ldots , b^k ) } ) ] { \text { in } } { \operatorname{h}}^{{\operatorname{mid}}}_{b^1 , \ldots , b^k}(h^{k+1}/y^{k+1})\ ] ] for all @xmath1524 this in turn shows that @xmath1525 completing the induction and the proof .
on any smooth , oriented , closed manifold @xmath1526 there is a natural intersection pairing @xmath1527 the signature of this pairing is by definition the signature of the manifold . note that by stokes theorem @xmath1528 and so if one of @xmath1270 @xmath597 is closed and the other exact then @xmath1529 thus @xmath1170 descends to a map @xmath1530 poincar duality on @xmath246 is the assertion that this quadratic form is non - degenerate .
if we endow @xmath246 with a riemannian metric @xmath8 and denote the hodge star by @xmath1531 then non - degeneracy is immediate from @xmath1532 the generalized poincar duality of goresky - macpherson @xcite is , for every stratified pseudomanifold , a non - degenerate map @xmath1533 induced by the usual intersection product , but involving two dual perversities @xmath1534 and
@xmath1535 in @xcite goresky - macpherson adopted a sheaf hypercohomology approach and showed that this generalized poincar duality can be deduced from verdier duality . for witt spaces , the intersection cohomology groups with ` lower middle ' @xmath1536 and ` upper middle ' @xmath1537 perversities coincide , so the generalized poincar duality statement involves only a single sequence of groups .
cheeger @xcite showed that poincar duality could be realized as the same integration map @xmath1170 on @xmath0 differential forms for an @xmath9 metric . for non - witt spaces with conic singularities , he also indicated @xcite how to incorporate ideal boundary conditions to refine poincar duality .
the sheaf approach has been extended by banagl @xcite to a refined poincar duality using verdier duality . in this section
we generalize cheeger s analytic approach to generalized poincar duality results to non - witt spaces . given a mezzoperversity @xmath1538 we introduce a dual mezzoperversity @xmath1539 and show that there is a non - degenerate intersection pairing @xmath1540 realizing a refined generalized poincar duality . in a companion paper @xcite we will show that this is consistent with banagl s sheaf theoretic approach .
recall from the end of [ sec : simpleedgebc ] the construction of the natural flat connection on the vertical cohomology of a fibration of closed manifolds .
if we assume that @xmath246 is the total space of a fibration @xmath1541 with all spaces closed and oriented , then we have a natural push - forward @xmath1542 given by integration over the fibers .
this factors as @xmath1543 and satisfies @xmath1544 we can use @xmath1545 to define the family intersection pairing @xmath1546 by using the surjection @xmath1547 and then setting @xmath1548 notice that this pairing is non - degenerate and from that @xmath1549 in particular this shows that whenever @xmath1550 is a flat sub - bundle then so is @xmath1551 the @xmath1552-orthogonal complement of @xmath1029 + at the end of [ sec : twoedgebc ] we discussed how one can adapt the construction of the connection @xmath417 to the setting of @xmath0 cohomology with ideal boundary conditions .
that discussion extends to this context and shows that @xmath1553 notice that we can rewrite @xmath1554 in terms of the vertical hodge star @xmath1555 and the vertical @xmath0-inner product , namely @xmath1556 so @xmath1557 @xmath1095 [ prop : dualmezzo ] let @xmath1558 be a stratified pseudomanifold with a suitably scaled @xmath9-metric and cheeger ideal boundary conditions @xmath470 corresponding to a hodge mezzoperversity @xmath1559 and for each @xmath1560 let @xmath1561 be the @xmath1562-orthogonal complement of @xmath1563 then @xmath1564 is a hodge mezzoperversity and if we denote the corresponding cheeger ideal boundary conditions by @xmath1565 then the hodge star @xmath1566 defines bounded involutions @xmath1567 the hodge star of @xmath8 defines a unitary involution @xmath1568 that for any @xmath1569 satisfies @xmath1570 for any subbundle @xmath238 of the vertical harmonic forms of @xmath1571 there is an induced map @xmath1572 by identifying these forms with forms on @xmath1573 in the kernel of the vertical de rham operator . denoting @xmath1574 the @xmath1554-orthogonal complement of @xmath241 this map
is @xmath1575 now if @xmath1 has a single singular stratum , this is enough to prove the proposition . indeed , clearly @xmath1539 is a mezzoperversity and @xmath1576 if and only if @xmath1577 which suffices since @xmath1202 and @xmath1345 are the graph closures of @xmath1578 if @xmath1 has two singular strata and they are both non - witt , then @xmath1579 is a flat subbundle of @xmath1580 applying the proposition to the typical link at a point in @xmath310 @xmath289 a stratified space with a single singular stratum , we see that @xmath1581 is a flat subbundle of @xmath1582 thus @xmath1539 is a mezzoperversity and the same argument as above shows that @xmath1576 if and only if @xmath1577 and hence holds .
the case of more strata is proven inductively in the same way .
let @xmath1558 be a stratified pseudomanifold with a suitably scaled @xmath9-metric and cheeger ideal boundary conditions @xmath470 corresponding to a hodge mezzoperversity @xmath1583 the mezzoperversity @xmath1539 of proposition [ prop : dualmezzo ] is known as the * dual hodge mezzoperversity * to @xmath1583 the de rham mezzoperversity @xmath1584 $ ] is the * dual de rham mezzoperversity * of @xmath1487.$ ] note that the dual de rham mezzoperversity is independent of the choice of metric .
let @xmath1558 be an oriented stratified pseudomanifold with a suitably scaled @xmath9-metric and cheeger ideal boundary conditions @xmath470 corresponding to a hodge mezzoperversity @xmath1486 with duals @xmath1565 and @xmath1585 the intersection pairing @xmath1586 is non - degenerate and descends to a non - degenerate ` generalized poincar duality ' pairing @xmath1587 at the level of differential forms the statement follows easily from @xmath1588 with both @xmath667 and @xmath1589 in @xmath1590 to see that the pairing descends to cohomology , note that if both @xmath667 and @xmath597 are regular then by stokes theorem for manifolds with corners @xmath1591 and each of these summands vanishes since @xmath1592 by continuity in the graph norm , it follows that @xmath1593 for all pairs in @xmath1594 and hence @xmath1595 vanishes if one of the forms is closed and the other exact .
the generalized poincar duality of goresky - macpherson is most interesting when the dual perversities @xmath1534 and @xmath1596 coincide , i.e. , on witt spaces . similarly our generalized poincar duality is most interesting when the dual mezzoperversities @xmath1486 and @xmath1539 coincide .
we say that a de rham mezzoperversity is * self - dual * if it coincides with its dual mezzoperversity .
if @xmath1 is a stratified pseudomanifold that carries a self - dual mezzoperversity we say that @xmath1 is a * cheeger space .
* not all stratified pseudomanifolds are cheeger spaces .
indeed , @xmath1597 is isomorphic to @xmath1598 precisely when the vertical hodge star sends @xmath1599 to its orthogonal complement with respect to the induced vertical metrics .
this implies that the vertical family of signature operators on @xmath1600 has vanishing families index .
in particular the individual links @xmath1601 must have vanishing index .
+ if @xmath1 has only conic singularities , then it is a cheeger space if and only if the links all have vanishing signature . on a cheeger space
we can refine the de rham operator to the signature operator by restricting to self - dual forms .
if @xmath1 is even - dimensional , the hodge star induces a natural involution on the differential forms on @xmath477 @xmath1602 that extends to @xmath0 @xmath9 forms @xmath1603 the @xmath1604 @xmath719 eigenspaces are known as the self - dual and anti - self - dual forms and denoted by @xmath1605 the signature operator @xmath1606 is the de rham operator @xmath38 as an unbounded operator between self - dual and anti - self - dual forms .
its formal adjoint is @xmath1607 the de rham operator @xmath38 as an unbounded operator between anti - self - dual and self - dual forms .
if @xmath1486 is a self - dual mezzoperversity and @xmath470 are the associated cheeger ideal boundary conditions , we define @xmath1608 let @xmath1488 be a stratified space endowed with a suitably scaled @xmath9 metric and cheeger ideal boundary conditions @xmath470 corresponding to a self - dual mezzoperversity @xmath1583 the signature operator @xmath1609 is closed and fredholm .
its adjoint is @xmath1610 and its fredholm index is the signature of the generalized poincar duality quadratic form @xmath1170 from .
in order to relate the cohomology of the complex @xmath1360 to the hypercohomology of a complex of sheaves , as we do in the companion paper @xcite , we must compute the cohomology of a distinguished neighborhood with respect to ideal boundary conditions .
so assume that @xmath1611 is a distinguished neighborhood , with trivialized vertical cohomology with boundary conditions , ( consistent with the flat structure ) , @xmath1612 , and flat subbundle @xmath1613 ; we compute @xmath1614 a mezzoperversity for @xmath1 induces a mezzoperversity on each link @xmath1615 and hence ideal boundary conditions @xmath1616 on @xmath1617 for every @xmath1618 the cohomology of the resulting complex is given by the expected formula : the proof follows the usual computation , as in @xcite and @xcite .
note that cheeger defines certain operators on the maximal domain of the exterior derivative ; we point out that , using notation as in @xcite , @xmath1622\leq f/2 } { \longrightarrow}{\mathcal{d}}_{\max , { \mathbf{b}}(z)}(d^{c(z)})_{[j]\leq f/2 } \\ w\in { \mathcal{d}}_{\max , { \mathbf{b}}(z)}(d^{c(z ) } ) \implies kw = \pi_*(w ) { \text { or } } \pi_*(\chi w ) \in { \mathcal{d}}_{{\mathbf{b}}(z)}(d^{c(z ) } ) \end{gathered}\ ] ] hence cheeger s proof carries over directly .
finally , observe that one can reduce the cohomology @xmath1623 to @xmath1624 this requires a bit of care since , even though @xmath1625 is contractible , the boundary conditions vary with @xmath1618 however the natural filtration by vertical degree with respect to @xmath1626 shows that the cohomology @xmath1623 can be computed by first passing to @xmath1627 ( indeed this is the @xmath1628 term in the serre spectral sequence ) . since this is now a trivial local system , we can use the contractibility of @xmath1625 to see that this reduces to @xmath1624 one can proceed alternately as in ( * ? ? ?
* corollary 2.15 ) . in any case
, this gives @xmath1629 , _ on the hodge theory of riemannian pseudomanifolds _ , geometry of the laplace operator ( proc .
pure math .
hawaii , honolulu , hawaii , 1979 ) , proc .
pure math .
, xxxvi , amer . math . soc . , providence , r.i .
, 1980 , pp .
91146 .
jeff cheeger and xianzhe dai , _
@xmath0-cohomology of spaces with nonisolated conical singularities and nonmultiplicativity of the signature _ , riemannian topology and geometric structures on manifolds , progr .
271 , birkhuser boston , boston , ma , 2009 , pp .
124 .
schulze , _ the iterative structure of the corner calculus _
, pseudo - differential operators : analysis , applications and computations , oper . theory adv .
213 , birkhuser / springer basel ag , basel , 2011 , pp . 79103 . | we extend the study of the de rham operator with ideal boundary conditions from the case of isolated conic singularities , as analyzed by cheeger , to the case of arbitrary stratified pseudomanifolds .
we introduce a class of ideal boundary operators and the notion of mezzoperversity , which intermediates between the standard lower and upper middle perversities in intersection theory , as interpreted in this de rham setting , and show that the de rham operator with these boundary conditions is fredholm and has compact resolvent .
we also prove an isomorphism between the resulting hodge and @xmath0 de rham cohomology groups , and that these are independent of the choice of iterated edge metric . on spaces which admit ideal boundary conditions of this type which are also self - dual , which we call `
cheeger spaces ' , we show that these hodge / de rham cohomology groups satisfy poincar duality . |
coupled map lattices ( cml ) are dynamical systems with discrete space and time and continuous state variables .
their most common form in applications is the diffusively coupled model @xmath0\ ] ] where @xmath1 is a discrete time step , @xmath2 is a lattice point with a periodic boundary condition , @xmath3 the local map , @xmath4 the coupling and @xmath5 a continuous parameter .
they prove particularly useful in the investigation of chaotic dynamics of spatially extended systems ( spatiotemporal chaos ) , but have also been employed in chemical , biological and engineering modelling , as well as pattern formation . the discrete structure in space and time
can either be the approximation of a less tractable continuous system ( as in turbulence ) , or may arise naturally in applications ( as in population dynamics of discrete colonies of organisms ) .
refs . @xcite
give a good overview of early work to 1998 including the above applications ; recent developments include advances in the mathematics of the relevant banach spaces and transfer operators @xcite as well as ever widening applications ; examples include ecology @xcite , neurons @xcite , traffic flow @xcite cryptography @xcite and field theory @xcite .
from the point of view of understanding the dynamics of coupled chaotic maps , the most common cml studied are where the local dynamics @xmath6 is quadratic ( the `` logistic '' map ) ; this class of one dimensional maps is now reasonably well understood .
the simplest example of a quadratic map is the second degree tchebyscheff polynomial @xmath7 which is well known to be conjugated to a piecewise linear map with complete binary symbolic dynamics , and hence exactly solvable . in this paper
we consider diffusively coupled tchebyscheff maps ( see sec .
2.2 for details ) of degree 2 and 3 .
these have been previously used by beck @xcite in some interesting but unconfirmed ideas concerning dynamical generation of noise for stochastic quantisation leading to predictions of parameters in the standard model of particle physics .
one of the authors @xcite subsequently showed that four of beck s six cml models exhibit stable synchronised time periodic states at arbitrarily small values of the coupling parameter @xmath5 , albeit with small basins of attraction .
the motivation for considering these models derives not only from possible applications , but also theoretical simplicity in that the uncoupled state is fully chaotic and exactly solvable , the small coupling limit has interesting structure , and the equations ( being low degree polynomials ) allow many exact calculations . here
we apply classical periodic orbit theory @xcite to a lattice of two coupled tchebyscheff maps .
this theory allows very precise computations of long time properties ( averages , lyapunov exponents , dimensions , decay of correlations , etc . ) of chaotic systems , but has been mostly restricted to low dimensional systems in the past .
this paper contains the first steps towards developing a high dimensional periodic orbit theory , making use of the properties of tchebyscheff maps to locate a large number of periodic orbits .
if the beck theory turns out to be true , the insight gained regarding these cml will be invaluable in developing periodic orbit based computation methods which are potentially far more precise than direct simulation , and which are required in particle physics . more generally , a periodic orbit theory of cml will allow a far greater precision and insight into the many spatiotemporal systems discussed at the beginning than is currently available using direct methods . in order to achieve our immediate goal of computing beck s observables on a two site lattice ,
we first need to look for periodic orbits of ( [ main ] ) by using appropriate numerical methods and study some of their properties , as described in the following sections .
the locations , the stability and the number of periodic orbits of such maps depend on the value of @xmath5 and current results do not allow prediction of these changes in detail . as we discuss in section 2 ,
unexpected phenomena take place as soon as the coupling parameter @xmath5 is increased from 0 , for example a few periodic points suddenly move off the region @xmath8 $ ] , which we consider for our analysis . as @xmath5
is increased from , periodic points drift on the domain @xmath8 $ ] , moving into the complex plane and then reappearing on the real axis as @xmath5 is increased further .
the stability of periodic orbits is also heavily affected by the value of the coupling and stable and/or almost stable orbits appear for certain @xmath5 s .
such behaviour has important consequences for the convergence of cycle expansions @xcite , as discussed in section 3 .
in fact both the rate and the kind of convergence of any average we compute on the lattice strongly depends on the value of the coupling parameter .
still in the same section we address the applications of equations ( [ main ] ) to the theory of beck . as pointed out by one of the authors @xcite , the system ( [ main ] ) does not show ergodicity for all values of @xmath5 , therefore direct simulation is not always guaranteed to be reliable in computing averages over the lattice .
that is one reason why we investigate the efficiency of averaging over the periodic points . when it is possible to achieve a significant precision for the zeta function ( see sec .
3.1 ) , we have computed the average of some string potentials defined in sec . 3.2 , and then compared the results with direct simulation on the same lattice . conclusions and outlook are presented in section 4 .
the first problem is to find periodic solutions to the map ( [ main ] ) up to the highest possible period .
a number of numerical methods can be employed for that purpose . in our case
the quickest one seems to be the inverse iteration method , as follows : let @xmath9 be a map whose cycles are all expanding , then it follows that its inverse @xmath10 has only stable periodic orbits , which can easily be found by iteration .
a different symbol sequence ( where symbols are associated with branches of the multivalued inverse ) is to be used for every orbit @xcite .
unfortunately this method can not be applied a priori , unless we are certain that the map has only expanding cycles .
it turns out that the system ( [ main ] ) has got a few periodic orbits whose jacobian has one stable and one unstable eigenvalue ( see sec 2.3 ) .
this makes it impossible for inverse iteration to find them , since we can not guess initial conditions that lie right on the unstable manifold .
a different approach to the problem is that to find the zeros of the equation @xmath11 , where @xmath1 is the period in question .
the multipoint newton s method @xcite is believed to be one of the best algorithms for that purpose .
it has the advantage of converging relatively fast to the solution , provided that sufficiently accurate initial conditions are given .
a cycle of period @xmath1 is a root of the vector function @xmath12 : @xmath13 the iteration we need in order to approach the roots has the form @xmath14^{-1 } f(x)\ ] ] therefore , after a certain number of iterations we expect to find one root .
of course the method has to be performed every time with different initial conditions , until we are sure all the solutions have been found .
the issue concerning the choice of initial conditions is a delicate one , because in principle it is difficult to guess the initial points on a 2-dimensional plane accurately .
nevertheless , it is useful to remember that we are dealing with weakly coupled maps ( @xmath15 ) .
for that reason we expect the periodic points to be relatively close to those of the uncoupled system ( @xmath16 ) , that can be found analytically ( see sec .
2.2 ) and hence exactly .
( 200,200 ) ( 0,150 ) ( 180,150 ) the spatially uncoupled dynamics of tchebyscheff maps is given by @xmath17 where @xmath18 is the @xmath19th tchebyscheff polynomial , defined in general as @xmath20 of which we consider the first two , namely @xmath21 @xmath22 the general ( coupled ) tchebyscheff maps have the form @xmath23 where @xmath5 is the string coupling parameter , @xmath24 can be @xmath25 and @xmath26 can take on the values 1 and 0 , determining respectively advanced coupling ( @xmath27 , `` a '' ) or backward coupling ( @xmath28 , `` b '' ) .
we have then six models @xmath29 , @xmath30 , @xmath31 , @xmath32 , @xmath33 and @xmath34 .
the other two @xmath35 maps are equivalent to these by a trivial symmetry . the model ( [ main ] ) , becomes @xmath36 on a two site lattice . in this section we consider the uncoupled case @xmath16 , in which the system ( [ sys1],[sys2 ] ) reduces to @xmath37 this is useful as there is a close correspondence between orbits at @xmath16 and orbits with @xmath38 .
this correspondence is used to find orbits in the coupled dynamics using the uncoupled orbits as starting points for the multipoint newton algorithm .
the correspondence is not exact the @xmath39 degree polynomial equation giving the periodic condition always has this number of ( possibly degenerate ) solutions , however for @xmath38 some of these solutions are not in the range @xmath40 $ ] or are complex , and hence not relevant to the dynamics .
this is discussed further in the next section .
we begin first with some definitions .
a _ periodic point _ of length @xmath41 is a solution of @xmath42 ; as discussed above , there are at most @xmath39 of them .
a _ periodic orbit _ is the orbit @xmath43 of a periodic point .
prime orbit _ is a periodic orbit which is not the repeat of a shorter orbit , that is , @xmath44 only if @xmath45 is a multiple of @xmath41 .
@xmath46 is a periodic point of length @xmath41 of the uncoupled system if and only if @xmath47 and @xmath48 are both periodic points of length @xmath41 of the map @xmath18 .
it is well known @xcite that the @xmath18 maps are conjugate to piecewise linear maps with complete @xmath19-ary symbolic dynamics , by the transformation @xmath49 .
this leads to exact expressions for all the periodic points of length @xmath41 , namely @xmath50 $ ] with @xmath51 .
from these we enumerate all the prime orbits of the uncoupled two dimensional dynamics , eliminating cyclic permutations which correspond to points on the same orbit , and repeats of shorter orbits . as an example we consider @xmath52 .
let symbol 0 ( 1 ) denote @xmath53 $ ] ( respectively @xmath54 $ ] ) ; in general there are @xmath19 symbols . in the symbolic representation of periodic points , the spatial variable ( @xmath2 in eq .
( 1 ) ) will increase to the right , and the temporal variable ( @xmath1 in eq .
( 1 ) ) will increase downwards .
the periodic points of the one dimensional map are @xmath55 leading to 16 periodic points in the two dimensional map , classified as six prime orbits @xmath56 their cyclic permutations @xmath57 and four repeated period one orbits @xmath58 until now we have not considered the symmetry of the system .
this is particularly important for cycle expansions @xcite , for which a correct treatment of symmetry significantly accelerates convergence of the expansions .
an uncoupled lattice has symmetry group @xmath59 where @xmath60 is the number of lattice sites ( equal to two in this paper ) , @xmath61 is the symmetric group , of order @xmath62 , corresponding to permutations of the lattice points , and @xmath63 is the symmetry group of the dynamics .
a symmetrically coupled lattice with @xmath64 has symmetry group @xmath65 , the dihedral group , of order @xmath66 , corresponding to spatial cyclic permutations and reflections of the lattice sites . in the case
@xmath67 the cyclic shift and reflection are equivalent , so the group is @xmath68 , the cyclic group .
remnants of the internal symmetry group @xmath63 may or may not survive , depending on the nature of the coupling . for @xmath69
the coupling breaks the permutation symmetry , reducing the order of the symmetry group ( ignoring internal symmetries ) , from @xmath62 to @xmath66 @xcite .
each symmetry operation permutes periodic points .
if the image of a periodic point lies in the same periodic orbit , that orbit is called symmetric , and is invariant under a cyclic subgroup of symmetry operations .
if it does not , the new orbit is equivalent to the first under the symmetry , and shares many of the same properties , in particular stability and averages invariant under the symmetry .
it is also possible for a periodic point to be a fixed point of the symmetry transformation , in which case its orbit is called a boundary orbit . in the presence of symmetry , the best strategy is desymmetrisation , in which all images of a phase space point are considered equivalent , and so may be described by a representative point on a subset of the original phase space called the fundamental domain .
equivalent orbits in the full dynamics correspond to a single orbit in the desymmetrised dynamics , and symmetric orbits in the full dynamics correspond to repeats of shorter orbits in the desymmetrised dynamics .
thus the enumeration of prime orbits up to a given period in the desymmetrised dynamics contains much more information than the enumeration up to the same period in the original dynamics , hence the accelerated convergence of cycle expansions .
we continue with the previous example , considering the @xmath70 maps , which have no internal symmetry .
thus the symmetry group is the cyclic group @xmath68 , with the nontrivial element corresponding to @xmath71 .
the fundamental domain can be taken as @xmath72 , from which the full domain is obtained by taking symmetry operations .
the boundary satisfies @xmath73 .
now the prime orbits of period two in the desymmetrised dynamics are classified as follows .
there are two asymmetric orbits of period two in the original dynamics , @xmath74 one boundary orbit of original period two @xmath75 and three symmetric orbits of original period four @xmath76 note that the total number ( six ) is the same , but the information content is higher , since all the original period two orbits are included as period one or two , and also some orbits of original period four . in the case
@xmath35 , the symmetry group of the interacting case is @xmath77 , since there is an internal symmetry @xmath78 , which in the coupled map survives as @xmath79 .
this is in addition to the permutation @xmath71 .
thus the fundamental domain can be taken as @xmath80 .
orbits may now be classified as asymmetric , symmetric under @xmath71 , symmetric under @xmath79 , symmetric under @xmath81 , boundary with @xmath73 , boundary with @xmath82 and ( if it occurs ) boundary with @xmath83 . in this section
we illustrate the behaviour of the @xmath18 map periodic solutions when the coupling parameter @xmath5 is in the range @xmath84 $ ] .
in particular , we describe two phenomena , that is the way and stability and number of cycles change with @xmath5 .
stability is expressed by the _ _ real parts _ _ eigenvalue , but we always mean its real part . ]
@xmath85,@xmath86 of the eigenvalues of the product of the jacobian matrices @xmath87 of each point of one cycle , where @xmath88 and @xmath89 are the set of equation ( [ sys1],[sys2 ] ) .
@xmath90 is said to be stable when @xmath91 .
when @xmath16 , the systems is uncoupled and the @xmath18 maps have no stable cycles . since the eigenvalues are continuous functions of @xmath5 , we expect no stable eigenvalue to show up for very small @xmath5 at fixed orbit period . as for stability , we analyse the trend of the maximum and minimum eigenvalues found for each period . in general , we consider the product @xmath92 so that in the 2-d case it is either @xmath93 or just the unstable @xmath90 .
stable orbits are not used for cycle expansions , hence they are not considered for this analysis .
the graphs in figure [ f : trend1 ] show that both the maximum and the minimum eigenvalues increase hyperbolically ( i.e. exponentially ) with respect to the period @xmath41 of the cycles when @xmath16 and still up to @xmath94 , but the trend of the minimum @xmath90 s changes as long as @xmath5 grows , so that their increase gets slower and slower .
besides , as we will explain later in this section , the few cycles with @xmath95 move off the fundamental domain when @xmath96 , so that the maximum eigenvalues take the form @xmath97 . coupled maps where the maximum @xmath98 is @xmath99 when @xmath16 and @xmath100 or lower otherwise .
] when @xmath101 , the minimum eigenvalues are perhaps best fitted by a power law , rather than an exponential .
such a behaviour is common to all the coupled @xmath102 and @xmath103 maps .
figure [ f : trend1 ] also shows that the trend of the minimum @xmath90 s also gets less and less smooth , as @xmath5 increases .
this phenomenon , together with the loss of hyperbolicity for some @xmath104 , dramatically slows down the convergence of cycle expansions , as we discuss in the next section .
( 500,500 ) ( 50,500 ) ( 50,330 ) ( 50,160 ) figure [ f : trend2 ] shows the trend of the eigenvalues in the @xmath32 map when @xmath101 .
( 150,150 ) ( 50,150 ) as we can see , it is very difficult to make a good fit for the trend of the minimum @xmath98 s in this case , because all the coupled @xmath102 ( and @xmath103 ) maps have got almost stable cycles for such @xmath5 s , including some at high period .
it is interesting to compare this behaviour with what happens with intermittent @xmath105 maps @xcite .
the other interesting issue we analyse is whether and why the number of cycles of a coupled map for a fixed period varies as a function of the coupling parameter @xmath5 .
numerical simulation shows that few periodic orbits of the uncoupled map ( [ uncoupled ] ) move off the fundamental domain when the system is coupled ( @xmath38 ) .
moreover , the number of @xmath106 cycles continues to decrease as @xmath5 grows , because some periodic solutions of the set ( [ sys1],[sys2 ] ) become complex or degenerate for certain values of @xmath5 , so that most of the missing orbits have just disappeared into the complex plane .
data representing the number of periodic points @xmath107 with @xmath108 $ ] have been fitted with the curve @xmath109 where @xmath110 and @xmath111 are parameters that depend on the model employed , as shown in table [ t : tab ] .
.parameters @xmath111 and @xmath110 of the fitting curve ( [ h ] ) with @xmath112 $ ] , uncertainty of about 2 on the last digit [ cols="^,^,^",options="header " , ] as we can see , @xmath113 in all the models , which is theoretically plausible since the dynamics and bifurcations are controlled by the critical point of the tchebyscheff map , which is quadratic .
that implies that perturbations of order @xmath5 have effects of order @xmath114 , hence the exponent of @xmath115 .
@xmath107 is related to the _ topological entropy _ @xmath116 of the system by @xmath117}{n_p}\ ] ] so that @xmath118\ ] ] which decreases with a power law as @xmath5 increases .
figure [ f : sinking ] shows good agreement of the @xmath116 fitting function with the data for very weak coupling ( @xmath119 ) while a slight deviation for larger @xmath5 .
this deviation is not unexpected , since this form for the topological entropy is only an approximate expression in the small coupling limit .
of course a detailed investigation could reveal that the topological entropy is not smooth in fine detail , however this form is remarkably accurate for practical purposes .
( 200,200 ) ( 50,200 )
( 200,200 ) ( 70,225 ) ( 60,120 ) @xmath48 ( 190,0 ) @xmath47 given a quantity @xmath120 , we want to compute its average on the attractor of the dynamics ( [ sys1],[sys2 ] ) , such as the one depicted in figure [ f : attractor ] . assuming that such attractor is ergodic for any value of the coupling parameter we consider , direct simulation is the quickest method to use for this purpose , but its precision is limited by statistics .
that is one reason why we use cycle expansions , which are potentially far more precise .
according to periodic orbit theory @xcite , the average of @xmath121 can be found by computing the dynamical zeta function @xmath122 where @xmath123 @xmath6 is the map in question , @xmath124 is a parameter , @xmath125 the eigenvalue of the cycle @xmath126 and @xmath1 is the orbit period .
the uncertain digit in the average is obtained by comparing results from the computation of ( [ zeta ] ) up to the two largest periods ( for example 10 and 11 ) .
it is possible to check the rate of convergence of the zeta function by setting @xmath127 and @xmath128 and computing @xmath129 , which should approach zero .
this way we can have an idea of how effective our cycle expansion is and whether longer cycles are needed .
figure [ f : zeta ] shows the convergence of @xmath129 versus the truncation period of the expansion .
( 220,220 ) ( 0,250 ) ( 170,250 ) ( 0,120 ) ( 170,120 ) as we can see , the convergence is monotonic and exponential when the coupling parameter is relatively small ( say @xmath130 ) , while it becomes nonmonotonic and increasingly slower for larger values of @xmath5 .
the accuracy of @xmath131 is particularly affected towards @xmath101 , where longer cycles or stability ordering may be needed .
ordering by stability means to expand the zeta function keeping all terms whose period and stability are less than a given cutoff ( @xmath132 ) .
this scheme is usually preferred to length ordering in case of intermittency or bad symbolic dynamics grammar @xcite , but it hasnt been attempted here since it would need much a higher period of truncation for the cycle expansion and , as a consequence , impractical cpu times .
a full resolution would require understanding which orbits are close to stability , which is beyond the scope of this paper .
anyway the behaviour shown in figure [ f : zeta ] traces back to the trend of the minimum eigenvalues of the cycles ( sec .
2.3 ) , since we can see from equation ( [ zeta ] ) that the smaller @xmath98 s are more relevant in the computation of the zeta function .
that means that if their trend is exponential with respect to increasing @xmath5 s , the @xmath133 s are smaller and smaller as @xmath129 approaches zero .
otherwise the convergence gets more complicated .
the expansion for the @xmath103 maps has been truncated at period @xmath134 , yet the precision obtained is comparable with what we have got in the @xmath102 maps .
that is because @xmath135 , meaning a similar number of periodic points and hence more or less the same number of terms in the two expansions .
we now want to use cycle expansions to evaluate the average of an observable over the two - site lattice ( [ sys1],[sys2 ] ) and compare the outcomes with what we get with direct simulation on the same lattice . according to the theory of beck @xcite , an interaction potential @xmath136 and self - interacting potentials
@xmath137 can be defined for the chaotic strings described by equation ( [ main ] ) .
zeros of ( [ int ] ) and minima of ( [ self2],[self3 ] ) may represent parameters of the standard model of particle physics such as coupling constants and masses of elementary particles . since both the potentials defined above depend on @xmath5 in a nontrivial way , we can only find zeros and minima numerically by averaging both @xmath138 and @xmath139 over the lattice ( [ main ] ) .
beck has obtained his results by running a long trajectory on a lattice with many sites , we are going to see what happens instead by computing the zeta function ( [ zeta ] ) on the two - site lattice ( [ sys1],[sys2 ] ) . ( 220,220 ) ( 0,250 ) ( 170,250 ) ( 0,120 ) ( 170,120 ) figure [ f : averages ] shows our results ( achieved by computing ( [ zeta ] ) up to two different orbit periods ) compared with direct simulation of the potentials @xmath140.a similar behaviour has been observed with the interaction potential @xmath139 .
five main features emerge . 1 . in most points cycle expansion
coincides with direct simulation .
cycle expansion is far offset on some points , which can be explained with the presence of almost stable cycles ( @xmath141 ) of high periods whose terms in eq .
( [ zeta ] ) have too much weight in the expansion .
3 . there are points where both cycle expansion and direct simulation are offset .
that is probably caused by one or more stable cycles or chaotic attractors , whose basins of attraction alter the phase space making it no longer ergodic .
the precision of the expansion visibly decreases as @xmath5 approaches @xmath142 , as we expect from the observations made in the previous sections on the trend of the minimum eigenvalues and the behaviour of @xmath129 .
figure [ f : screwed ] shows an offset at some point between cycle expansion and direct simulation in the model @xmath31 performed with different starting values from those used in figure [ f : averages ] .
that proves a lack of ergodicity in some regions of the interval @xmath84 $ ] of @xmath5 and indicates that the unstable periodic orbits are mostly unaffected by the presence of a stable orbit ( or other ergodic component such as an elliptic region ) . hence the cycles form a transient chaotic set more or less covering the attractor ( fig .
[ f : attractor ] ) for similar values of the parameter . in this sense , the periodic orbit expansion results are more stable than direct simulation with respect to variation of the parameter , and may be preferred for this reason .
( 140,140 ) ( 70,170 )
periodic orbits have provided a unique window with which to view properties of weakly coupled cml .
the periodic orbits in the uncoupled limit provide a reliable means of finding periodic orbits of the weakly coupled system .
the symmetry of the coupled system has been exploited to improve convergence of the cycle expansion .
cycle expansion techniques work as expected , well near the uncoupled , more chaotic limit , but with slower convergence at stronger coupling .
the computation of averages using cycle expansions leads to important information at a glance concerning the number and nature of nearly or fully stable states coexisting with the chaotic web of unstable periodic orbits .
the observed power law dependence of the topological entropy with parameter requires a more detailed investigation of the density and nature of bifurcations in this system to explain the value of the power @xmath143 , the extent of its apparent universality and corrections for finite values of the parameter . for a more rigorous approach to the topological entropy of a different two site model ,
see @xcite .
this work has clarified a number of issues and challenges regarding a more general periodic orbit approach to cml . for stronger coupling ( hence weaker chaos ) stability ordering of cycle expansions @xcite
is likely to be more effective than the length ordering employed here ; this requires an understanding of the symbolic dynamics of the most stable cycles , which may be obtained in the study of the bifurcations mentioned above .
larger lattices are of course an important aspect , however the rapid decay of spatial correlations for weakly coupled cml make this less important than it might seem .
the fact , as observed here , that small windows of stability at weak coupling affect the cycle expansions at only a few parameter values , gives great confidence since their presence is expected to be even less evident in lattices of larger size , in which typical orbits have many unstable directions .
we acknowledge financial support of the nuffield foundation ( grant nal/00353/g ) and helpful discussions with p. cvitanovi .
dl would also like to thank g. servizi and v. smith for having made available computing facilities in bologna and bristol respectively , as well as n. chevalier , r. morris and a. shall for technical support .
papers special issue : chaos * 2 * , issue 3 ( july 1992 ) .
k. kaneko ( ed . ) `` theory and applications of coupled map lattices '' ( wiley , chichester , 1993 ) . special issue : physica d * 103 * ( april 1997 ) .
t. bohr , m. h. jensen , g. paladin and a. vulpiani , dynamical systems approach to turbulence ( cambridge university press , 1998 ) . v.
baladi and h. h. rugh , commun .
phys . * 220 * , 561 ( 2001 ) .
bardet , probab .
theory relat .
fields * 124 * , 151 ( 2002 ) .
a. satake and y. iwasa , ecology * 83 * , 993 ( 2002 ) .
f. a. labra , n. a. lagos , p. a. marquet ecol
* 6 * , 197 ( 2003 ) .
p. c. tobin and o. n. bjornstad , j. anim .
ecol . * 72 * , 460 ( 2003 ) .
a. o. dudkin and v. i. sbitnev , biofiz .
* 45 * , 119 ( 2000 ) .
r. m. c. de almeida and m. a. p. idiart phys .
e , * 65 * , 061908 ( 2002 ) .
t. nagatani , physica a * 289 * , 267 ( 2001 ) .
r. tenny , l. s. tsimring , l. larson and h. d. i. abarbanel , phys .
lett . * 90 * , 047903 ( 2003 ) .
wang et al , commun .
theor . phys . *
40 * , 57 ( 2003 ) . c. beck , physica d * 171 * , 72 ( 2002 ) . c. beck , `` spatio - temporal chaos and vacuum fluctuations in quantized fields '' ( world scientific , 2002 ) . c. p. dettmann , physica d * 172 * , 88 ( 2002 ) .
r. artuso , e. aurell and p. cvitanovi , nonlinearity * 3 * , ( 325 ( 1990 ) ; * 3 * , 361 ( 1990 ) .
p. cvitanovi et al . , `` chaos : classical and quantum '' , chaosbook.org ( niels bohr institute , copenhagen 2003 ) .
b. lauritzen , phys .
a * 43 * , 603 ( 1991 ) .
j. m. robbins , phys .
a * 40 * , 2128 ( 1989 ) .
dettmann and p. cvitanovi , phys .
e * 56 * , 6687 ( 1997 ) .
v. afraimovich and b. fernandez , nonlinearity * 13 * 973 ( 2000 ) . | coupled map lattices have been widely used as models in several fields of physics , such as chaotic strings , turbulence , and phase transitions , as well as in other disciplines , such as biology ( ecology , evolution ) and information processing .
this paper investigates properties of periodic orbits in two coupled tchebyscheff maps .
then zeta function cycle expansions are used to compute dynamical averages appearing in beck s theory of chaotic strings .
the results show close agreement with direct simulation for most values of the coupling parameter , and yield information about the system complementary to that of direct simulation . |
the normal resting shape of the human red blood cell ( rbc or erythrocyte ) is a flattened biconcave disc ( discocyte ) about 8 @xmath1 m in diameter .
treatment of erythrocytes in vitro with a variety of amphipathic agents is known to transform this shape systematically and reversibly into various other shapes such as echinocytes ( crenated shapes ) and stomatocytes ( cup - like shapes ) , which are further subdivided into classes labelled i , ii , iii ( brecher and bessis , 1972 ; bessis , 1973 ; chailley et al .
1973 ; mohandas and feo , 1975 ) .
in particular , the echinocyte iii is a more - or - less spherical body covered evenly with 25 - 50 rounded protuberances , which we shall call spicules .
these shape transformations from the discocyte to the stomatocyte on one side and from the discocyte to the echinocyte on the other have been studied experimentally for more than 50 years .
understanding these shapes and shape transformations is a classic problem in cell biology ; over the past three decades it has also attracted the attention of physicists .
what makes this problem so intriguing is that the structure of the red cell is remarkably simple . to a good approximation
it is simply a bag of fluid , a concentrated solution of hemoglobin surrounded by a thin macroscopically homogeneous membrane .
thus , the diverse resting shapes which it exhibits can be thought of as ( locally ) stable mechanical equilibrium shapes of the membrane .
this membrane is composite ( steck , 1989 ; alberts et al . , 1994 ) . on the outside
is the plasma membrane , a self - assembled fluid bilayer with a thickness of about 4 nm , composed of a complex mix of phospholipids , cholesterol , and dissolved proteins .
lipids such as phosphatidylcholine , sphingomyelin , and glycophospholipids , which are neutral at physiological ph , are common in the outer leaflet , while phosphatidylserine and phosphatidylethanolamine predominate in the inner leaflet ( gennis , 1989 ; alberts et al . , 1994 ) .
due to the negative charge of phosphatidylserine , located in the inner leaflet , there is a significant difference in charge between the two leaflets of the bilayer . inside the plasma membrane but linked to it by protein anchors is a thin , two - dimensionally cross - linked protein cytoskeleton , the membrane skeleton ( steck , 1989 ; bennett , 1990 ) .
the skeleton is an hexagonally linked net , each unit of which is a filamentous spectrin tetramer with an extended length of about 200 nm . the spectrin tetramer is negatively charged at physiological ph .
thermal fluctuations reduce the projected distance between the vertices of the net to about 76 nm , while the offset between the plasma membrane and the spectrin network is 3050 nm ( byers and branton , 1985 ; liu et al . ,
functionally , the plasma membrane serves as an osmotic barrier , passing water with relative ease but controlling , via a system of pumps and channels , the flow of ions and larger solute molecules .
the membrane skeleton has the role of supporting and toughening the plasma membrane , which would otherwise disintegrate in circulatory shear flow . while the rbc membrane is certainly heterogeneous at the length scale of individual lipid molecules , lipid patches or the basic spectrin tetramer , on scales longer than 100 nm it is reasonably homogeneous in its properties , and it may make sense to treat it as a mechanical continuum . in this spirit and following early work by canham ( 1970 ) , helfrich ( helfrich , 1973 ; deuling and helfrich , 1976 ) , and others , we imagine replacing the ( fluid - phase ) plasma membrane by an ideal uniform two - dimensional surface characterized by a bending rigidity @xmath2 , taken for the rbc to be ( waugh and bauserman , 1995 ; strey et al . , 1995 ) @xmath3 j ( roughly 50 @xmath4 , where @xmath5 is boltzmann s constant ) , a spontaneous curvature @xmath6 , which recognizes the outside - inside asymmetry of the leaflet composition , and an area - difference - elasticity ( ade ) term ( helfrich , 1973 ; evans , 1974 ; svetina et al . , 1985 ; boi et al . , 1992 ; wiese et al . , 1992 ;
miao et al . , 1994 ) , which reflects the fact that a difference in relaxed area , @xmath7 , between the outer and inner leaflets produces a `` bilayer couple '' ( sheetz and singer , 1974 ) tending to bend the membrane away from the larger - area side ( like a bimetallic strip ) .
we assign to the modulus @xmath8 associated with the ade term a value of ( waugh and bauserman , 1995 ) @xmath9 j. these terms taken together constitute what is called the ade model of the plasma membrane .
in addition , we represent the rbc membrane skeleton as a two - dimensional elastic continuum characterized by a bulk stretching modulus @xmath10 and a shear modulus @xmath1 with @xmath11 j / m@xmath12 and @xmath13 ( see discussion in sec .
v ) . these two mechanical subsystems , constrained to the same mathematical surface , will constitute our model of the rbc shape at length scales above 100 nm ( see sec .
ii for full details ) .
the basic hypothesis of the mechanical approach is that the observed rbc shapes must be shapes which minimize the energy subject to appropriate constraints on volume and area , i.e. , that all the observed rbc shapes must emerge as local energy minima of this model and that the observed shape transformations must come about as a response of the minimum - energy shapes to changes in the `` control parameters '' which characterize the model .
there are relatively few such control parameters : in addition to the known area and volume of the rbc , we include in the list the three mechanical moduli noted above , the spontaneous curvature @xmath6 , the relaxed area difference @xmath7 between the bilayer leaflets , and parameters describing the effective relaxed size and shape of the membrane skeleton .
if all these control parameters were known or easily measurable , we could predict the stable rbc shapes using continuum mechanics .
the fact that they are not restricts us , as we shall see , to somewhat more generic predictions .
this program has already been carried out , albeit in somewhat restricted form , with respect to the discocyte and stomatocyte shapes . the shapes and shape transitions of laboratory - prepared fluid - phase phospholipid vesicles ( lacking any skeleton ) have been successfully described by the ade model ( miao et al . , 1994
; dbereiner et al . , 1997 ) .
in particular , plausible discocytic and stomatocytic shapes do occur as energy - minimizing shapes of the ade model .
furthermore , the bilayer - couple mechanism ( sheetz and singer , 1974 ) accounts qualitatively for many of the chemically - induced discocyte - stomatocyte transitions observed in the laboratory , in that stomatocytogenic agents tend to segregate to the inner leaf of the bilayer , thus making @xmath7 negative and favoring the invaginated stomatocyte shape ( see more below ) .
there has been discussion in the literature of the way in which introducing the elastic energy of the membrane skeleton modifies or improves the detailed prediction for the shape of the normal resting discocyte ( zarda et al . , 1977 ; evans and skalak , 1980 ) . however , most treatments in the literature tend to regard the cytoskeletal elastic energy as providing only perturbative modification of the basic shapes which emerge from the pure - bending - energy model
. we shall argue that this point of view fails completely in describing echinocyte shapes .
until recently , it was a major difficulty for a fully mechanical picture of the rbc shapes that echinocytic shapes have not been found in the catalog of minimum - energy shapes of the ade model ; however , several authors have now pointed out what we believe is the correct resolution of this problem ( waugh , 1996 ; igli , 1997 ; igli et al .
, 1998a , b ; wortis , 1998 ) : a pure - lipid vesicle with a sufficiently positive spontaneous curvature ( or , equivalently , sufficiently large positive @xmath7 ) adopts a vesiculated or budded shape , in which one or more quasi - spherical buds appears on the outside of the main body of the vesicle , attached to it via a narrow neck or necks ( see fig .
such necks are allowed as low - energy structures for pure - lipid vesicles , because the bending energy depends on the sum of the two principal curvatures ( see sec .
ii ) , so that large curvatures of opposite sign ( characteristic of necks ) can still lead to a small mean curvature and , thus , to low energy ( fourcade et al . ,
1994 ) . but and this is the crucial point
small necks involve large elastic strains in the membrane skeleton and are , thus , inhibited by elastic energy cost .
it is not surprising that they are not seen for undamaged rbc s .
what our calculations show is that the formation of echinocytes is driven by positive @xmath6 ( equivalently , positive @xmath7 ) and that the spiculated shapes seen in experiments can arise from a competition between the bending energy of the plasma membrane , which by itself would promote budding , and the stretching and shear elastic energies of the membrane skeleton , which prevent it . to make this more precise
, we note that the ratio of the bending modulus to the shear modulus ( or the stretching modulus ) defines an elastic length scale , @xmath14 which yields @xmath15
@xmath1 m . we shall see below that this number sets the scale of spicule size and , thus , fixes the spicule density of the fully developed echinocyte ( echinocyte iii ) .
the success of this approach strengthens the hypothesis that rbc shapes can be understood on the basis of continuum mechanics ; however , in themselves , such calculations can not answer the question , `` why does a particular shape occur under given experimental conditions ? ''
such a question has two parts .
why do the control parameters have the values they do ? and ,
how does this set of control parameters lead to the observed shape ?
the second part is the proper domain of continuum mechanics .
the first part is primarily biological or biochemical in content and revolves around the mechanisms which control the lipid composition of the leaflets of the plasma membrane , local electrostatic effects , the composition and environment of the spectrin network and its coupling to the plasma membrane , and any other determinants of @xmath6 , @xmath16 , the elastic moduli , and other control parameters .
hopefully , this separation of the question is useful .
the plan of this paper is as follows . in the remainder of this section
, we briefly review some history of echinocyte shapes and shape calculations .
section ii presents the details of the continuum - mechanics model .
section iii discusses the theory required to solve the model numerically for echinocyte shapes , including a full account of the scaling argument suggested above .
section iv presents our results for the predicted echinocyte shapes and spicule density .
section v contains discussion of our results and further speculations .
the discocyte - to - echinocyte transformation was precisely identified by ponder ( 1948 , 1955 ) and has since been studied extensively by many other authors , as reviewed , for example , by bessis ( 1973 ) , lange et al .
( 1982 ) , and steck ( 1989 ) .
although this transformation can be driven in many ways , the final spiculated shapes produced are apparently the same ( smith et al . , 1982 ; mohandas and feo , 1975 ) .
this quasi - universal behavior is evidence for the kind of mechanism suggested above , in which quite diverse biochemical processes may set ( a few ) important control parameters to the same or similar values . in the same spirit
, we note that rbc ghosts ( formed by hemolysis ) behave quite similarly to intact red cells ( lange et al . , 1982 ) .
the simplest picture of this type consistent with some ( but not necessarily all ) experimental findings is the so - called bilayer - couple hypothesis of sheetz and singer ( 1974 ) , which focuses attention on the control parameter @xmath7 .
thus , adding exogenous phospholipids to the exterior solution is observed to promote echinocyte formation ( ferrell et al .
, 1985 ; christiansson et al . , 1985 ) .
this is explained in the bilayer - couple picture by arguing that such addition is expected to produce an area increase in the outer leaflet only , since phospholipids do not readily flip - flop from one leaflet to the other .
this increases the area difference @xmath7 and is expected to promote echinocyte formation .
similarly , cholesterol does equilibrate across the bilayer but apparently prefers the outer leaflet , presumably because of the particular lipids there present .
thus , adding cholesterol will tend to increase @xmath7 and promote echinocytosis , while depleting cholesterol will tend to decrease @xmath7 and to promote stomatocytosis ( lange and slayton , 1982 ) .
more generally , anionic amphipaths all tend to produce echinocytes , while cationic amphipaths tend to produce stomatocytes ( deuticke , 1968 ; weed and chailley , 1973 ; mohandas and feo , 1975 ; smith et al . , 1982 ) .
the bilayer - couple explanation of these observations is that , when incorporated into the plasma membrane , the cationic compounds segregate preferentially to the inner leaflet and the anionic compounds , to the outer leaflet because of the predominantly negative charge of the inner - leaflet lipids .
outer - leaflet segregation increases @xmath7 , thus promoting echinocytosis ; conversely , inner - leaflet segregation promotes stomatocytosis .
time - dependent effects have even been seen , where a molecule initially intercalates into the outer leaflet , causing echinocytosis , but then slowly migrates to the inner leaflet , following the electrostatic gradient , causing return to the discocyte and subsequent stomatocytosis ( isomaa et al . , 1987 ) .
a similar hypothesis is made to explain the observed tendency of ghosts to become echinocytes at high salt concentrations but stomatocytes at low salt concentrations ( lange et al . , 1982 ) .
the argument relies on electrostatics .
cationic species neutralize the negative lipid charges on the inner leaflet and salt decreases the debye length , more effectively screening such charges .
both effects lead to a contraction of the inner leaflet and consequent increase in @xmath7 , promoting crenation .
other observations can be linked to the bilayer - couple effect but somewhat less directly .
thus , it is found that hemolyzed echinocytes produce discocytic ghosts ( lange et al . , 1982 ) and that electroporation suppresses shape changes ( schwarz et al .
, 1999a , b ) .
to explain these effects , it is argued that hemolysis and electroporation both involve perforation of the membrane and resulting in contact via the pore surface between the lipids in the inner and outer leaflets .
this provides a mechanism for lipid transport which may reduce the magnitude of @xmath7 by partial symmetrization of the lipid composition of the leaflets , thus ( equivalently ) reducing @xmath6 .
similarly , it is known that atp depletion drives discocytes to crenate ( nakao et al . , 1960 , 1961 , 1962 ;
backman , 1986 ) .
the hypothesis here is that atp is required in some way to maintain the lipid asymmetry of the leaflets ( related to @xmath6 ) , perhaps for the operation of atp - driven translocases ( steck , 1989 ) .
some effects are not readily explained by the bilayer - couple mechanism .
weed and chailley ( 1972 ) , and gedde et al . ( 1995 , 1997a , 1997b , 1999 ) report that rbc shape can be controlled experimentally by varying the external ph , with high ph promoting echinocyte formation and low ph , stomatocyte formation ( the effect of proximity to a glass surface in promoting echinocytosis ( furchgott and ponder , 1940 ) is probably related to this effect ) .
the mechanism for this effect does not seem to be well established . some authors argue for a mechanism involving membrane proteins .
band 3 has been specifically implicated ( gimsa and ried , 1995 ; jay , 1996 ; gimsa , 1998 ; hgerstrand et al . , 2000 ) .
for example , it is known that high ph induces dissociation of ankyrin and band 3 , which plays a role in the linkage of the membrane skeleton to the bilayer ( low et al . , 1991 ) .
indeed , it was shown that in vitro the membrane skeleton expands at high ph and contracts at low ph ( elgsaeter et al . , 1986 ; stokke et al . ,
it has recently been proposed that ph change may induce conformational transformation in band 3 protein , leading to a change in @xmath7 .
similarly , wong ( 1994 , 1999 ) has proposed that the folding of the spectrin in the cytoskeletal net is controlled by a conformational change of the band 3 protein .
such mechanisms might shift the emphasis of shape determination from the dominantly lipid - related control parameters , @xmath6 and @xmath7 , to the elastic parameters and the relaxed shape of the membrane skeleton .
we shall comment briefly on these issues in sec .
v. numerous shape calculations have previously been done using variants of the model we describe in sec .
ii . early calculations ( fung and tong , 1968 ; zarda et al . ,
1977 ; evans and skalak , 1980 ) omitted the ade term ( @xmath17 ) , set @xmath18 , and concentrated entirely on the shape of the resting discocyte and the modifications caused by osmotic swelling .
they solved numerically the full mechanical shape equations .
landman ( 1984 ) treated a viscoelastic shell surrounding a viscous droplet and modeled the formation of protrusions as sudden local addition of shell material .
waugh ( 1996 ) studied a full ade model with local shear elasticity ( but treating the compressibility modulus as infinite ) and studied the instability of a flat membrane to formation of a local `` bump '' ( axisymmetric spicule ) for positive values of @xmath6 or @xmath7 or both .
he assumed a specific spicule - shape parametrization , focussed on a single spicule , calculated its energy , and in this way was able to discuss generically the conditions under which spicule formation might become energetically favorable .
most recently , igli and collaborators ( igli , 1997 ; igli et al . , 1998a , b ) have used the same model as waugh ( 1996 ) but parametrize the spicule more simply , using a hemispherical cap on a cylindrical body and joining this to a spherical surface via a base shaped like a toroidal section .
they correctly identify the emergence of echinocyte shapes as a competition between the bending energy and the skeletal elasticity .
they model the echinocyte as a collection of @xmath19 spicules attached to a spherical body and then determine the radius of the sphere , the parameters of the spicules , and the preferred number of spicules by minimizing the mechanical energy subject to constraints on cell area and volume . in this way , they are able to estimate the expected number of spicules and their shape as a function of the relaxed area difference @xmath7 .
the present paper builds on the work of waugh and igli : we extend the model to include the area modulus @xmath20 of the membrane skeleton .
we calculate spicule shape for the first time from the euler - lagrange equations ( these calculations are closely related to the earlier work of zarda et al .
( 1977 ) ) .
although we continue with igli et al .
( 1998a , b ) to treat the echinocyte as a sphere decorated with spicules , we deal seriously ( although still approximately ) with the mechanical boundary conditions which must be met where the spicule joins the sphere .
the central assumption of our work is that the red blood cell assumes a shape that minimizes its overall membrane energy subject to the appropriate constraints .
the rbc membrane is a composite of the plasma membrane and the cytoskeletal network ; correspondingly , we take the membrane energy to be a sum of two terms , @xmath21 the bending energy , @xmath22 associated with the bilayer and an elastic energy of stretching and shear , @xmath23 associated with the skeleton . in doing this , we treat the plasma membrane as incompressible .
an estimate of its stretching modulus is @xmath24 j / m@xmath12 , comparable to that of a pure - phospholipid bilayer and some four orders of magnitude larger than that of the skeleton .
correspondingly , we ignore any bending rigidity associated with the isolated membrane skeleton ( the dimensional estimate @xmath25(thickness)@xmath12 leaves it two orders of magnitude smaller than that of the bilayer ) .
eq . [ bendingenergy ] for the plasma membrane s contribution to the overall membrane energy is the now - standard ade hamiltonian ( svetina et al . , 1985 ; boi et al . , 1992 ; wiese et al . , 1992 ; miao et al . ,
the first term was originally proposed by helfrich ( helfrich , 1973 ; deuling and helfrich , 1976 ) . @xmath26 and @xmath27 are the two local principal curvatures , @xmath28 is the spontaneous curvature , and the integral is over the membrane surface .
the two leaflets of a closed bilayer of fixed interleaflet separation @xmath29 are required by geometry to differ in area by an amount , @xmath30 in calculations , we shall take @xmath31 corresponding to the distance between the neutral surfaces of the leaflets . the second term in eq . [ bendingenergy ]
is the so - called area - difference - elasticity energy and represents the cost in stretching ( or compressional ) energy of the individual leaflets necessary to force the change from the relaxed area difference @xmath7 so as to conform to this geometric requirement .
this effect occurs because a strong hydrophobic barrier prevents lipids in one leaflet from `` flip - flopping '' passively to the other on the time scales of mechanical shape changes .
@xmath32 is the area of the plasma membrane .
the modulus @xmath33 associated with the ade term is generically of the same scale as @xmath2 .
note , finally , that we can substitute eq . [ areadifference ] to rewrite eq .
[ bendingenergy ] , @xmath34 + { \rm const . } , \label{bendingenergy1}\end{aligned}\ ] ] where @xmath35 @xmath36 , and the constant term is shape independent .
this shows that , in determining minimal shapes at given values of the control parameters , @xmath6 and @xmath7 do not enter independently but only in the form of an effective spontaneous curvature @xmath37 or , equivalently , an effective relaxed area difference , @xmath38 which we shall often quote in the dimensionless , reduced form @xmath39 note for future reference that increasing @xmath40 by @xmath41 is equivalent to increasing @xmath6 by @xmath42m@xmath43 .
[ elasticenergy ] measures the elastic - energy cost of the spectrin network .
it depends both on the relaxed shape of the membrane skeleton and on the way it is actually distributed over the membrane surface .
( the notion of a relaxed shape is , of course , somewhat nominal , since removing the skeleton from the plasma membrane which can certainly be done ( sheetz , 1979 ) would radically change its local biochemical environment , in a way which would modify its shape and elastic constants ) . in this redistribution ,
each element of the network will be stretched or compressed .
the quantities @xmath44 and @xmath45 are the local principal extension ratios of each membrane element ( evans and skalak , 1980 ) . @xmath10 and @xmath1 are the ( two - dimensional ) moduli for stretch and shear , respectively , as introduced in sec .
i , and the integrals are over the undeformed shape . in writing eq .
[ elasticenergy ] , we are assuming , as appears to be typical , that the membrane skeleton does not plastically deform in the course of the experimental shape transformation being described .
note that eq .
[ elasticenergy ] makes a particular choice of terms in the elasticity beyond those quadratic in the weak - deformation parameters , @xmath46 . for the large elastic strains which are present at narrow necks and for small spicules , these nonlinearities do play a role . as far as we are aware , rbc
elasticity has not been measured well enough to produce a clear preference for a particular form of the nonlinearities .
various authors have used eq .
[ elasticenergy ] for the elastic energy in other contexts with apparent success ; however , alternative forms of the elasticity have also been proposed for dealing with problems involving large local deformations ( evans and skalak , 1980 ; discher et al . , 1994 ) .
we shall make some additional comments in sec .
v. one attractive feature of eq .
[ elasticenergy ] is that the effect of changing the size of the relaxed membrane skeleton by a pure , uniform dilation is particularly simple .
suppose that the skeleton is decreased in linear scale by a factor @xmath47 .
this means that undeformed areas are reduced by a factor @xmath48 , while the extension ratios @xmath49 are each increased by a factor @xmath47 .
note that the integrand of the shear term in eq .
[ elasticenergy ] is invariant under this change but the factor @xmath50 in the stretching integrand increases by @xmath51 .
only the quadratic part of the stretching term influences the membrane shape because @xmath52 is just the membrane area a , which is fixed by the incompressibility of the plasma membrane .
it follows that the effect of decreasing the size of the skeleton is completely equivalent to keeping the size fixed but , instead , changing the elastic moduli according to @xmath53 i.e. , this `` prestress '' in the membrane skeleton is completely equivalent to no prestress , a harder stretching modulus , and a softer shear modulus . it is not known with certainty what `` prestress '' exists in the typical rbc skeleton .
computer simulation has suggested that the relaxed skeleton may be as much as 1020% smaller in area than the plasma membrane ( boal , 1994 ) . on the other hand , the experiment by svoboda et al .
( 1992 ) shows that isolated skeletons are expanded . in our calculations
, we shall assume zero net prestress ( relaxed skeletal area equal to rbc area , i.e. , @xmath54 ) .
any deviation from this would result in modified , effective elastic constants according to eqs .
[ cytoscaling0 ] and [ cytoscaling ] . in our calculations
, we will identify shapes which are minima of the ( free ) energy defined by eqs .
[ totalenergy][elasticenergy ] subject to constraints of fixed membrane area @xmath32 ( bilayer incompressibility ) and volume @xmath55 ( since the rbc volume is typically set by osmotic balance ) .
constrained minimization is achieved variationally by introducing the functional , @xmath56 where @xmath57 and @xmath58 are lagrange multipliers used to enforce the surface - area and volume constraints .
@xmath58 is the pressure difference across the bilayer , while @xmath57 has the dimension of a surface tension .
making @xmath59 stationary with respect to variations of membrane shape and cytoskeletal distribution leads to a set of coupled euler - lagrange equations .
these equations can then be solved to give the shapes of mechanical equilibrium . in principle , such shapes are expected to correspond to observed equilibrium shapes at temperature @xmath60 only ; however , because the energy scale @xmath2 is large on the scale of @xmath61 , thermal fluctuations are generally negligible and play an important role only for `` soft '' modes , especially near instabilities ( wortis et al . , 1997 ) .
the euler - lagrange equations are in general nonlinear and can have multiple solutions .
the lowest - energy solution for given @xmath32 and @xmath55 is automatically stable to small fluctuations ; higher - energy solutions must be tested for stability .
all local - minimum solutions are candidates for stable observable shapes , except in exceptional cases ( near instability ) where energy barriers become comparable to @xmath61 .
the stable shapes produced by this process depend on the control parameters : the geometrical parameters , @xmath32 and @xmath55 ; the determinants of curvature , @xmath6 and @xmath7 , which only enter in the combination ( [ c0bar ] ) or , equivalently , ( [ deltaa0bar])-([deltaa0 ] ) ; the moduli , @xmath62 , @xmath63 , @xmath10 , and @xmath1 ; plus , finally , any parameters required to characterize the relaxed shape of the membrane skeleton .
the area and volume of a typical rbc we take as ( bessis , 1973 ) @xmath64m@xmath12 and @xmath65m@xmath66 .
the moduli , given in sec .
i , are less well determined by experiment ( we shall have more to say on this point in sec .
this leaves as unknown control parameters the curvature variables and the cytoskeletal parameters .
the euler - lagrange variational equations derived from ( [ phi ] ) are not numerically tractable except in the case of axisymmetry , which clearly does not apply to echinocytic shapes . in this paper
, we aim to treat only fully developed echinocytic shapes ( echinocyte iii ) .
for this case , individual spicules are identical and axisymmetric in shape to a good approximation and , in addition , the central body which they decorate is approximately spherical with radius @xmath67 .
the observed distribution of spicules is rather regular , and we shall approximate the local spicule packing as a triangular array ( except for special numbers of spicules , there must , of course , be some defects ) , which will look increasingly like an hexagonal close - packed structure , as the number of spicules , @xmath68 , becomes large .
the base of each individual spicule , where it meets the sphere tangentially , is a circular contour @xmath69 of radius @xmath70 .
if @xmath71 is the angle subtended by @xmath69 at the center of the sphere , then @xmath72 because of the close - packed structure , the circles @xmath69 meet tangentially .
we explain in sec .
iii how to derive appropriate boundary conditions where the spicules meet the sphere . solving
the axisymmetric euler - lagrange equations then determines a spicule shape , including an individual spicule volume @xmath73 and area @xmath74 . in terms of these variables , the overall area and volume of the rbc
are taken to be @xmath75 and @xmath76 with the spicule number @xmath77 in writing these relations , we have assumed that 10% of the spherical surface is not covered by the circular spicule bases ( close packing on a flat surface would give 9.3% ; curvature effects and packing defects both increase this number ) .
the spicule volume @xmath73 is calculated with respect to a plane through @xmath69 , and eq .
[ volume ] neglects curvature corrections .
these approximations are good when the number of spicules is large , as it will turn out to be for the echinocyte shapes .
in our treatment , spicules are assumed to be axisymmetric , and individual spicules are joined to the central body along the contour @xmath69 , as illustrated in fig .
thus , our calculation involves finding a family of energy - minimizing axisymmetric spicule shapes and selecting from that family those shapes consistent with appropriate mechanical boundary conditions applied along @xmath69 .
parametrization of the axisymmetric spicule shape is illustrated in fig .
2 . the variables @xmath78 and @xmath79 measure distances along and perpendicular to the symmetry axis , respectively , while @xmath80 measures the arclength from the pole .
the function @xmath81 determines the shape .
@xmath82 is the angle between the local normal and the symmetry axis ; @xmath83 , @xmath84 are the principal curvatures , @xmath85 in calculating the spicule shape , we shall assume that the relaxed cytoskeleton is locally flat , a good approximation as long as the number of spicules is large , so that the spicule size is small compared to @xmath67 .
thus , in forming each spicule , a flat circular patch of relaxed cytoskeleton must be elastically deformed to fit the spicule contour .
we assume that this deformation is axisymmetric , so that the center of the patch remains at the apex of the spicule , and each point of the patch at relaxed radius @xmath86 maps to a point at arclength @xmath80 and radius @xmath79 on the spicule contour .
the principal extension ratios can then be written , @xmath87 when these expressions for @xmath44 and @xmath45 are inserted into eq .
[ elasticenergy ] we obtain an explicit form for the elastic energy in terms of the functions @xmath88 and @xmath89 , @xmath90 using eqs . [ totalenergy ] , [ bendingenergy ] , [ curvatures ] , and [ elasticenergy1 ] , we can implement the stationarity condition for the free - energy functional ( [ phi ] ) with respect to variations of membrane shape and cytoskeletal strain .
this leads to a set of five coupled first - order differential equations , which are written down explicitly in the appendix .
note that the ade term in eq .
[ bendingenergy ] enters only through the appearance of an effective spontaneous curvature @xmath91 which must be determined self - consistently via eq .
[ areadifference ] .
although these euler - lagrange equations look complicated , they can be related to simple mechanical force - balance conditions in a way that makes their content entirely transparent ( mukhopadhyay and wortis , in preparation ) .
we have used these equations to calculate ( approximate ) spicule shapes .
we now turn to the boundary condition where the spicules meet the sphere ( see fig .
along each element of @xmath69 , the spicule membrane exerts a tension @xmath92 in the plane of the membrane and perpendicular to @xmath69 , and a tension @xmath93 normal to the plane of the membrane .
there is no tension in the third ( `` shear '' ) direction because of the membrane fluidity .
in addition , the membrane skeleton exerts an independent tension which must be in - plane , since the skeleton lacks bending rigidity . in general
, these tensions would be different at different points of @xmath69 ; however , in the axisymmetric approximation , the tensions are uniform around @xmath69 and the skeletal tension is directed radially . to maintain mechanical equilibrium
, these tensions must balance where the spicules meet at point a. note that a is a point of bilateral symmetry between the two adjacent spicules , so that in - plane tensions , which act in opposite directions , always balance by symmetry . on the other hand , the normal tensions @xmath93 must vanish individually , since they act in the same direction .
the normal tension is related to the isotropic bending moment per unit length , @xmath94 by ( evans and skalak , 1980 ; mukhopadhyay and wortis , in preparation ) @xmath95 .
thus , an appropriate boundary condition along @xmath69 is @xmath96 this is the boundary condition which we shall apply in our calculations .
of course , this boundary condition is approximate , since the real spicule is only approximately axisymmetric . indeed , with equal logic
, we could argue by symmetry that @xmath93 should vanish at point b , where three adjacent spicules meet at a radius from the spicule axis some 15% larger than @xmath70 .
we shall use this observation in sec .
iv to get a rough measure of the error introduced by the approximation ( [ boundarycondition ] ) .
we now outline the algorithm used to solve numerically the euler - lagrange equations for stationary spiculated shapes .
a `` shooting method '' was employed to integrate the euler - lagrange equations from the apex to the edge of the spicule , the contour @xmath69 along which the normal tension @xmath93 vanishes .
the set of solutions can be characterized by five parameters .
three of them are the global parameters : the pressure difference @xmath58 , the lagrange multiplier @xmath57 , and the effective spontaneous curvature @xmath97 , eq .
[ effectivecurvature ] .
the remaining two parameters correspond to initial values of variables in the integration procedure , i.e. , the curvature ( @xmath98 ) and the stretching ratios ( @xmath99 ) of the cytoskeleton at the pole . integrating the euler - lagrange equations to a point where eq .
[ boundarycondition ] is satisfied determines @xmath70 and @xmath71 and , therefore , @xmath67 and @xmath19 from eqs .
[ geometry ] and [ spiculenumber ] .
the initial conditions may be adjusted iteratively to fit four parameters , @xmath32 ( eq . [ area ] ) , @xmath55 ( eq . [ volume ] ) , @xmath100 ( eq . [ deltaa0bar ] ) , and the cytoskeletal stretching ratio , @xmath101 , which measures the ratio of the area of the calculated spicule to that of the corresponding relaxed cytoskeletal patch ( roughly the same as the ratio of the area of the full echinocyte to that of the full relaxed membrane skeleton , since the corrections in eq .
[ area ] for the spherical segments are small ) .
we are left , finally , with a one - parameter family of mechanical - equilibrium solutions , which we take to be labelled by the spicule number @xmath19 .
the predicted equilibrium configuration is the solution @xmath102 which minimizes the overall energy .
of course , for the exact solution , @xmath19 is a discrete variable : probably , there exists , for given initial conditions , a set of distinct branches of solutions characterized by different @xmath19 s and different arrangements of spicules on the surface , as discussed more fully in sec .
v. before proceeding to solutions , we wish to identify the important energy and length scales of the problem .
because we are working at a temperature which is effectively low , one overall energy scale drops out of the mechanical problem .
we take this scale to be set by the bending modulus @xmath2 ( @xmath2 and @xmath8 are similar in magnitude ) .
there are three length scales : the first is the overall scale of the rbc , which we denote @xmath103 . the second , we may take to be @xmath104 from eq .
[ c0bar ] ( which is equivalent to using a length based on @xmath100 , eq .
[ deltaa0bar ] ) ; alternatively , we can take the second length to be @xmath105 , where @xmath97 is the effective spontaneous curvature from eq .
[ effectivecurvature ] ( we assume that @xmath106 and @xmath97 are comparable in magnitude ) . the third length is the elastic length scale @xmath0 , eq .
[ lambda ] .
in general , there may be other length scales associated with the relaxed cytoskeletal shape ; however , we shall assume that any such lengths are comparable to @xmath103 .
for the rbc , @xmath107 .
the length scale @xmath104 is a control parameter whose magnitude can be tuned across the range defined by @xmath103 and @xmath0 , resulting in the observed rbc shape classes .
the size of the rbc sets the total area and volume in the problem .
note that the rbc has a relatively smaller volume for a given surface area than a sphere .
if a sphere with the same surface area has a volume @xmath108 , then the ratio of rbc volume to that of the equivalent sphere defines a reduced volume , @xmath109 . for the rbc , @xmath110 , significantly less than unity , which allows the rbc to assume its sequence of distinct nonspherical shapes . in its usual discocytic phase
, the red blood cell is expected to have a magnitude of @xmath104 that is of the order of the red blood cell dimensions or larger .
tuning @xmath37 to sufficiently large positive values generates the sequence of evaginated echinocytic shapes ; tuning it to sufficiently large negative values generates the sequence of invaginated stomatocytic shapes .
the significance of the elastic length scale @xmath0 is that it determines how strong the effect of elasticity is for structures with some characteristic length , say @xmath69 . to see this , we imagine rescaling the problem with the energy scale @xmath111 and the length scale @xmath69 in such a way as to make all the parameters appearing in the euler - lagrange equations dimensionless . in this way , we arrive at the rescaled parameters , @xmath112 equivalent rescaling of the variables , @xmath113 , etc . , leaves the form of the euler - lagrange equations unchanged , although the parameter values ( eq . [ rescaling ] ) vary with @xmath69 .
for example , suppose @xmath69 is the echinocyte spicule radius . if @xmath114 , then @xmath115 is greater than unity , and the elastic terms from eq .
[ totalenergy ] dominate the bending - energy terms ; conversely , if @xmath116 , then @xmath115 is smaller than one , and the bending - energy terms dominate .
thus , generically , elasticity is strong at length scales larger than @xmath0 and weak at length scales smaller than @xmath0 .
this statement may seem surprising to those who think of the discocyte as effectively shaped by bending energy .
the key point here is the plasticity of the membrane skeleton , albeit on time scales much longer than the induced shape changes we are discussing : if the relaxed membrane skeleton closely matches the shape preferred by bending energy alone , then ( of course ) elastic effects are small .
this is apparently the case for the resting discocyte ( but definitely not for the echinocyte ! ) .
a similar scaling argument gives insight into the characteristic spicule size in the echinocyte region .
suppose we let @xmath117 in eq .
[ rescaling ] . at this length scale , the bending and elastic terms ( @xmath118 , @xmath119 , and @xmath115 ) are comparable and the only remaining parameters are @xmath120 , @xmath121 , and @xmath122 ( or , equivalently , @xmath123 ) . if , in addition , @xmath120 and @xmath121 are small , then the euler - lagrange equations are equivalent to those for the unconstrained minimization of the combined bending - plus - elastic problem ( [ totalenergy ] ) .
but , this is a problem which we understand : when @xmath122 is small in magnitude , then the membrane remains flat ( at the length scale of @xmath0 ) ; on the other hand , when @xmath122 is large in magnitude , then bending energy ( which dominates the shape at scales below @xmath0 ) will favor budded shapes on the scale of @xmath124 , either invaginated or evaginated depending on the sign of @xmath125 . between these limits , when @xmath0 and @xmath126 are comparable , we expect spicules ( for positive @xmath125 ) scaled by @xmath126 , tending towards smoother shapes when @xmath0 is somewhat smaller than @xmath126 and towards sharper , more columnar shapes when @xmath0 is somewhat larger than @xmath126 .
note that , as long as we rescale to the length @xmath0 , the expected spicule shape is predicted to be independent of cell size @xmath103 and to depend only on the scaled value of @xmath125 and the ratio of the two elastic constants , @xmath1 and @xmath10 .
the above argument depends on the condition that the scaled values of @xmath58 and @xmath57 be small at the elastic length scale . for smooth and flaccid shapes like the discocyte ( and
when the cytoskeleton is not significantly stretched or compressed ) the pressure difference @xmath58 is generated by the bending energy , so @xmath120 is of order unity on the scale of the cell size @xmath103 , and similarly for @xmath121 .
it then follows from eq .
[ rescaling ] that the condition is well satisfied for the discocyte and other nearby shapes ( as we shall see , the initial echinocyte shapes fall into this class ) .
of course , if the system is pushed too hard , then this condition breaks down and other length scales enter the spicule - shape problem . in summary ,
the shape of a spicule arises as a result of the interplay between the two length scales , @xmath127 and @xmath126 .
when @xmath126 is positive and much larger than @xmath127 , the effect of elasticity is strong and we can expect spicules that are at most gentle bumps .
it is only when @xmath126 becomes of the order @xmath0 or smaller , that we may expect sharp spicules .
thus , if we look at the series of shape transformations , from discocytes through discocytes with gentle bumps to sharply spiculated structures , the sharp spicules will first arise with a radius comparable to @xmath0 .
we shall verify this scenario in the next section by explicit solution for the stable shapes .
in this section we report echinocyte shapes calculated according to the program outlined in sec .
iii and using the elastic moduli ( @xmath2 , @xmath128 , @xmath1 , and @xmath10 ) given in sec .
i and the rbc parameters ( @xmath32 and @xmath55 ) given in sec .
the cytoskeletal stretching ratio @xmath129 is not known to good accuracy for the rbc . in this section , we take @xmath130 ; however , we have tested values in the range 0.71.2 , and it is one of our results that variation within this range produces no appreciable changes in the number of spicules or their shape . the spontaneous curvature and relaxed area difference are not known and presumably vary somewhat over a typical rbc population ; however , they enter the mechanics problem only in the combination @xmath131 , eq .
[ deltaa0 ] , which we take as our principal control parameter .
figure 3 displays the equilibrium spicule shapes that we obtain for a sequence of values of @xmath131 .
figures 4 and 5 plot the corresponding calculated number of spicules @xmath19 which appear on the fully developed echinocyte iii at equilibrium and the corresponding values of @xmath97 as functions of @xmath132 . how well do the calculated shapes and spicule dimensions agree with experimental observations ( bessis , 1973 ; brecher and bessis , 1972 ; chailley et al . , 1973 ) .
as shown in fig .
3 , the calculated shapes fall into three distinct categories . for @xmath19 less than about 20 , corresponding to @xmath131 appreciably smaller than @xmath133 ,
the spicule shapes are broad and rather hat - shaped , rather different from those seen in experiment .
as @xmath131 and @xmath97 increase , the number of spicules increases and the individual spicules become smaller , just as predicted by the scaling argument at the end of sec .
iii . by the time @xmath131 is at or above @xmath133 , for @xmath19 in the range 3060 , the spicule shapes are columnar with rounded tops , in good general agreement with the kinds of shapes seen in experiment . beyond @xmath134 ,
the spicules become narrow - necked , a shape not typically seen in experiment but illustrating the dominance of the bending energy , as discussed in sec .
iii . in the `` good , '' central region ,
the spicule dimensions and numbers are very comparable to those seen in experiment . a typical example ( @xmath135 ) is shown in fig . 6(a ) .
for this case , we find @xmath136 , a radius of the central , spherical body of @xmath137 , and an aspect ratio , height over width at base ( which also corresponds approximately to the distance between adjacent spicules ) , of about 0.8 , all of which are in reasonable agreement with observation .
the only significant discrepancy is that the spicule height is close to 1.0 @xmath1 m , somewhat larger than the range 0.60.8 @xmath1 m seen in experiment , as is the aspect ratio .
we shall comment on possible reasons for this discrepancy in sec .
v. we emphasize that these shapes are a prediction of the model , not an assumption as they are in other recent work ( igli , 1997 ; igli et al .
, 1998a , b ) .
the only other spicule - shape predictions are those given by waugh ( 1996 ) , whose calculation suggests cross sections that are less steep at the sides and more pointed at the apex than what is observed . whether this is a consequence of the variational form assumed or of the neglect of stretching elasticity in the model is not clear .
the calculations of igli et al . , based on a postulated spicule shape and assuming skeletal incompressibility ( @xmath138 ) , produce somewhat higher values of the aspect ratio . at a given reduced area difference @xmath131
, they find a spicule number which is larger than ours by almost a factor of two .
we will return to some of these issues in sec .
v. before discussing further what we shall argue are the non - biological regions of hat - like and narrow - necked shapes , we must review the full echinocyte sequence ( brecher and bessis , 1972 ; bessis , 1973 ; chailley et al . , 1973 ; mohandas and feo , 1975 ) .
experimental observations of rbc s which are initially smooth , axisymmetric discocytes show that echinocytosis occurs in three stages . the first stage ( echinocyte i )
corresponds to irregular disc - like shapes with a central dimple . in the second stage ( echinocyte ii )
the irregularities evolve into more or less well defined spicules and the central dimple disappears , resulting in an oblately elliptical body decorated with spicules . in the third stage ( echinocyte iii ) the central body becomes approximately spherical and the spicules become somewhat smaller and more numerous .
when pushed beyond the echinocyte iii , some kind of disconnection between the plasma membrane and the membrane skeleton occurs ( see below ) and plasma membrane is shed from the spicule tips into microscopic exovesicles . the remaining spherical shape with pointed protuberances
is called a spheroechinocyte ( bessis , 1973 ) .
our calculation explicitly assumes that the central body is spherical and is , therefore , appropriate only for the late stage - iii echinocyte .
what we believe happens is that , for @xmath139 , the branch of echinocyte iii shapes becomes energetically unstable to a branch of echinocyte ii shapes .
there is no internal way that our calculation can show that this crossover takes place ; however , there are two pieces of evidence we find convincing .
the first is a simple estimate : a separate calculation of the discocyte branch shows that its shape ( and @xmath140 , for example ) is roughly independent of @xmath132 , allowing us to estimate its energy as a function of @xmath141 .
the calculated energy of our echinocyte iii branch is lower than the discocyte branch but increases as @xmath132 decreases .
the crossing point of these two branches at @xmath142 is , thus , an approximate lower limit to the transition out of the echinocyte iii phase .
the actual transition out of the echinocyte iii branch must presumably take place at a somewhat higher value of @xmath132 .
the second piece of evidence is that , in separate work which we shall report elsewhere ( lim , wortis , boal , and mukhopadhyay , 2002 ) , we have done numerical simulations of rbc shapes using the same set of numerical parameters chosen here , and we find just such a transition at @xmath143 from the echinocyte iii to a branch of echinocyte ii shapes . at the opposite end of the range of `` good '' spicule shapes
, our calculation fails again and , again , for reasons that can not be assessed internally . here
the size of the predicted spicule set by the scale of @xmath144 becomes comparable to the length of the elementary spectrin tetramer of the membrane skeleton ( liu et al . , 1987 ) .
at this length scale , the continuum picture breaks down for the membrane skeleton ( although it remains valid for the lipid component ) .
what presumably happens is that a point of instability is reached , where lipid flow occurs and the plasma membrane buds outwards in regions between the cytoskeletal anchors .
such buds , lacking cytoskeletal support , are known to be fragile .
they presumably break off , leading to microvesiculation , membrane loss , and the observed spheroechinocytosis .
naturally , this vesiculation is expected to happen first at the spicule tips , where the elastic network is maximally expanded and the distance between anchors is largest . in nature , the plasma membrane and its cytoskeleton are bound together by protein anchors ; in our model this binding is replaced by the condition that the bilayer and the elastic network co - exist on the same mathematical surface .
this requirement means that there is a local force per unit area or pressure @xmath145 with which the skeleton pulls inward ( @xmath146 ) or pushes outward ( @xmath147 ) on the membrane surface ( or , equivalently , with which the membrane pulls outward or pushes inward , respectively , on the skeleton ) . the expression for this pressure is ( mukhopadhyay and wortis , in preparation ) @xmath148 \nonumber \\ & & + c_{p}\left[k_{\alpha}\left({r \over s}{dr \over ds } - 1\right ) - { \mu \over 2}\left(\left(s\over r\right)^{2 } -\left(ds \over dr\right)^{2}\right)\right ] \label{pressure}\end{aligned}\ ] ] and it is plotted in fig . 7 as a function of the radial distance @xmath79 from the spicule axis for @xmath149 . as expected , @xmath145 is positive close to the tip of the spicule and negative around the neck where the skeleton has been compressed .
the magnitude of the pressure at the spicule tip is of the order of 20 pn/@xmath1m@xmath12 .
it is useful to compare this pressure to the critical disjoining pressure necessary to break the anchoring and to separate the membrane skeleton from the bilayer .
this critical pressure has not , to the best of our knowledge , been measured ; nevertheless , it can be estimated .
we take 10 pn as a crude estimate of the force required to extract quickly a single anchoring protein from the bilayer . taking into account the density of anchors
, we estimate a detachment pressure of about 2000 pn/@xmath1m@xmath12 ( consistent with measurements reported in waugh and bausserman , 1995 ) , two orders of magnitude larger than what we have calculated above .
this comparison is important , as membrane - cytoskeleton disjoining is another potential mechanism for spheroechinocytosis , an alternative to the one discussed in the paragraph above and has , indeed , been proposed by igli and others ( igli et al . , 1995 , 1996 ) .
we conclude that such direct disjoining seems unlikely , unless it is associated with anchoring defects .
we have seen for the human erythrocyte that the fully developed echinocyte iii , with well developed spicules , first appears when @xmath144 becomes small enough to be comparable to the elastic length scale @xmath0 .
it is a plausible hypothesis that this is a general connection .
if so , there is a direct relation between the spicule dimension at the onset of the echinocyte iii and the ratio ( eq . [ lambda ] ) of the bending rigidity to the elastic moduli .
interestingly , smith et al .
( 1982 ) have studied spiculated shapes for the red cells of a variety of species . despite considerable variation in the volume of the red cells ranging from the goat rbc , which is appreciably smaller than the human one , through the elephant seal , which is appreciably larger
the spicule size is quite stable , suggesting that the elastic length scale is not strongly variable from one species to another .
this means , of course , that the number of spicules on the typical echinocyte is small for the goat ( @xmath150 ) and large for the elephant seal ( @xmath151 ) , as observed . to illustrate this point further , we plot in fig . 8 spicule shapes at different volumes and different ratios @xmath152 but with @xmath0 held fixed .
we find that at smaller volume , the spicules sizes do become smaller but the dependence on volume is weak .
the dependence on the ratio of elastic constants @xmath152 is even weaker . at a value of @xmath10 that is larger by a factor of 2 ( @xmath153 ) ,
the spicules become somewhat broader and their number decreases by around five percent .
of course , the local stretching of the cytoskeleton both at the spicule apex and in the neck region close to the base may be expected to change considerably .
these results are based on the model described in sec .
iii , which assumes spicule axisymmetry , as incorporated in the ( approximate ) boundary condition , eq .
[ boundarycondition ] . in closing this section
, it may be useful to comment on the reliability of this central assumption .
we have two remarks .
the first concerns the boundary condition .
the neighborhood of a given spicule is clearly not axially symmetric , so this condition is clearly approximate . if the calculated shapes depended sensitively on the precise way that this boundary condition is applied , then results based upon it would be suspect .
we motivated the boundary condition ( @xmath154 ) by looking at the point labelled a in fig .
2 , where adjacent spicules are tangent . with equal logic
, we could have argued this boundary condition at the point labelled b. if we made this choice , how different would the calculated spicule shapes turn out to be ? we have run some test cases and find changes in , e.g. , the spicule height of less than @xmath155 , suggesting that the results are strongly insensitive to the details of the boundary condition .
but , this is far from sufficient .
as a second test , we have recently completed a program ( lim et al .
, in preparation ) of direct simulation of the energy functional ( [ totalenergy ] ) on a triangulated surface .
this program allows us to calculate echinocyte shapes without assumptions about spicule axisymmetry .
we find echinocyte shapes in this region of parameter space and can compare the simulated spicule shapes with those calculated here at the same parameter values .
figure 6 shows one such comparison .
although the base of the simulated spicule is a little narrower than that given by the approximate axisymmetric calculation , the overall level of agreement is excellent . at these parameter values ,
the number of spicules in our calculation is 41 , while the number found by simulation is 34 .
similarly , the radius of the central spherical body to which our spicules are fitted is 2.57 @xmath1 m , while that found in the simulations is about 2.5 @xmath1 m ( the central body is only roughly spherical in this case ) .
the approximate calculation presented in this paper appears to be reasonably reliable .
the calculations presented here build on the previous work of igli et al .
( igli , 1997 ; igli et al . , 1998a , b ) and show that a simple mechanical model of a uniform composite membrane can give a good account of observed echinocyte shapes .
the central control parameter in this process is the effective area difference ( or curvature ) of the lipid component , in agreement with the bilayer - couple hypothesis of sheetz and singer ( 1974 ) .
the role of the cytoskeletal elasticity is to suppress the formation of narrow - necked buds which would form in the absence of cytoskeletal shear resistance .
does this good agreement prove that mechanics ( as opposed to biochemistry ) is the sole determinant of echinocyte shape ?
of course , not ; however , it does strongly suggest that simple mechanics will play an important role in any complete picture .
we end this section with a brief discussion of some loose ends and future directions .
there has been much discussion in the literature of what are the `` correct '' elastic constants to use for the membrane skeleton .
we have chosen @xmath11 j / m@xmath12 and @xmath13 .
these values are in rough agreement with recent measurements at low deformation ( hnon et al . , 1999 ;
lenormand et al . ,
2001 ) and somewhat lower than those reported in connection with pipette aspiration of red cells ( evans , 1973a , b ; waugh and evans , 1979 ) , where large deformations play an important role .
discussion of values of the elastic moduli is complicated by the occurrence of nonlinearities in the elastic response . at low deformation ,
@xmath156 with @xmath157 small , there are only two rotational invariants from which to construct the elastic energy , @xmath158 and @xmath159 , whose coefficients define the linear moduli of stretching and shear elasticity . it is easy to verify that these linear moduli are just @xmath10 and @xmath1 , respectively .
however , at large deformation , terms of higher orders in @xmath160 and @xmath161 will generally also appear in the elastic energy , each with its own coefficient .
the expression ( [ elasticenergy ] ) for the elastic energy makes a particular choice for each of these coefficients , one which is simply proportional to the linear moduli . at small extension ratios ,
these higher - order terms are unimportant ; however , for the echinocyte shapes , we find @xmath157 of order unity , so these terms can not be neglected .
unfortunately , present experiments do not constrain the values of the higher order coefficients significantly , although it is probably not an accident that the moduli derived from the pipette aspiration experiments ( evans , 1973a , b ) , which have extension ratios up to nearly 2 ( lee et al . , 1999 ) , are larger than those measured in the linear regime ( henon et al . , 1999 ; lenormand et al . , 2001 ) .
this suggests a hardening of the elasticity at large deformation , consistent , for example , with terms in @xmath162 , @xmath163 , and so forth , coming into play
. we will deal with the issue of nonlinearity at more length in a future publication .
for the present , we simply remark that realistic spicule formation does require a proper balance between stretching and shear energies at large deformation .
our chosen ratio of @xmath164 suffices ; however , a lower ratio plus some additional nonlinear hardening would also achieve the same effect and may be more realistic . any effect that changes spicule shape significantly
may also be expected to modify quantitatively the regions of stability of different shape classes .
previous variational work on echinocyte shapes by igli and coworkers ( igli , 1997 ; igli et al .
, 1998a , b ) has assumed local incompressibility of the cytoskeleton ( @xmath138 ) , thus forcing all the elastic energy into the shear term .
this was done for practical reasons ; we are now in a position to understand the effect of this approximation . figure 8(c )
suggests that increasing @xmath10 at fixed @xmath1 has a modest effect on spicule shape , decreasing the height and the base radius somewhat without changing the mean radius appreciably .
this is consistent with the generally larger number of spicules predicted by igli et al .
( igli , 1997 ; igli et al .
, 1998a , b ) .
because of the limited shapes available in the variational parametrization , these authors can not follow the full spicule - shape evolution exhibited in our fig .
our calculation assumes for the echinocyte regular spicule placement with six - fold coordination on a central spherical body .
of course , this is only an approximation .
first , there is the topological requirement for a net undercoordination of 12 .
more generally , if the energy minimization problem were to be solved exactly , we would expect to find several distinct sheets of locally - minimizing shapes with different spicule numbers @xmath19 and different spicule organization .
for any given set of parameter values , one such shape would have the lowest energy ; however , several others might have nearby energies .
metastability of a shape would be expected , since the energy landscape is on the scale of @xmath2 , which is large on the scale of @xmath165 .
this metastability , in turn , would cause hysteresis in the shape transformations .
some experiments show that , as the red cell is chemically cycled back and forth between discocyte and echinocyte phases , spicules always appear at the same locations on the cell surface ( furchgott , 1940 ; bessis and prenant , 1972 ) , suggesting that spicule placement could be intrinsically related to defects and other inhomogeneities in the cytoskeleton .
this would not happen in a continuum model . without wishing to dispute the experimental evidence
, we can only say that our work shows that spicule formation does not require cytoskeletal inhomogeneity . of course
, if inhomogeneities exist , then they will have an effect in selecting the pattern of spicule placement .
it is not clear a priori whether or not the spicule is a solitonic object with its own intrinsic scale and , if so , whether spicules tend to attract or repel at long distance .
some recent work ( lim et al . , in preparation ) has suggested long - range attraction , which would mean that , under appropriate circumstances , spicule clumping might occur , leaving `` bald '' regions on the echinocyte surface , as is seen in some pictures of echinocyte shapes ( see , for example , igli et al . , 1998a ) .
none of this finer detail is captured in our approximation .
we have focussed on the spontaneous curvature or area difference @xmath131 as the principal parameter controlling the evolution of echinocyte shape .
of course , induced changes in any parameter of the membrane energy expression eqs .
[ totalenergy][elasticenergy ] and the constraints @xmath55 and @xmath32 , singly or in combination , will affect the cell shape .
a ( bio)chemical modification of the red - cell environment presumably changes all these parameters to some extent ; it is just that some changes are larger and more important than others .
if there are situations in which laboratory reagents cause important modification in other parameters e.g . , the elastic constants then addition of these reagents will cause the red cell to be driven along a trajectory in an extended parameter space .
in such a situation , we would need to consider the minimum - energy shape ( or shapes ) as functions of several variables in an extended shape / phase diagram . the work surrounding fig .
8 illustrates the beginning of such a study .
we have summarized in sec .
i the mechanism by which some common reagents affect the sheetz - singer ( 1974 ) parameter @xmath131 .
the ph effect and the associated glass effect are less readily related to @xmath166 but may well be primarily related to another dimension of the generalized phase diagram . for example , gedde et al .
( 1995 ) have shown that lipid asymmetry and the presence of inner leaflet titratable groups , which might be expected to be associated with @xmath131 , do not have appreciable influence on the ph effect ; however , elgsaeter et al .
( 1986 ) and stokke et al .
( 1986 ) have shown that the membrane skeleton expands in vitro in response to high cytoplasmic ph . now , according to the scaling argument ( eq . [ cytoscaling ] ) , we know that such expansion is equivalent to decreasing the shear modulus @xmath1 and , therefore , to increasing the elastic length scale @xmath0 ( [ lambda ] ) .
but , echinocytosis occurs when @xmath144 shrinks below @xmath0 .
thus , at fixed @xmath97 ( roughly equivalent to fixed @xmath167 ) , expanding the membrane skeleton could promote echinocytosis , as observed .
it would be premature to claim that this is the full explanation of the ph effect .
the point is that , as long as changes in biochemical variables produce homogeneous changes in the material parameters that appear in eqs .
[ totalenergy]-[elasticenergy ] , our model continues to apply , although the relevant experimental trajectories may be in a larger parameter space .
description of any effects which lead to inhomogeneity in membrane properties requires modification of our model .
one such effect which is certainly present arises from the inhomogeneous distribution of lipids and/or proteins .
we know that the plasma membrane is composed of a mixture of lipids and proteins .
the bending modulus @xmath2 depends on membrane composition .
as long as the composition remains homogeneous , the model eqs .
[ totalenergy][elasticenergy ] holds ; but , any mechanism that produces compositional inhomogeneity on the scale of rbc shape features would necessitate modification of the model .
thus , for example , observed raft formation ( simons and toomre , 2000 ; pralle et al . , 2000 ) suggests membrane inhomogeneity on the scale of 0.05 - 0.07 @xmath1 m .
when local radii of curvature become comparable to the inhomogeneity scale , then inhomogeneity is expected to influence shape .
this may occur for the echinocyte near the spheroechinocyte limit ; however , there is no evidence that it is a dominant effect elsewhere .
other interesting effects involve the coupling of lipid composition to geometry .
thus , for example , outer - leaflet lipids with large heads may be expected to segregate preferentially to spicule tips , as would inner - leaflet lipids with large tails .
such an effect could be included in a mechanical model by adding a composition field ( or fields ) and including terms coupling composition to curvature .
alternatively , one might imagine mechanisms involving membrane components which preferentially favor ( or avoid ) regions around the cytoskeletal anchoring complexes .
these components would be relatively dilute ( concentrated ) in regions where the cytoskeleton was significantly expanded ( as it is near the spicule tip ) . to represent this effect , one would need to couple composition to local cytoskeletal strain . while these mechanisms may be present to some extent in the red cell , our calculations suggest they are not in any way required for spicule formation .
finally , chemical groups which carry net charge at physiological ph are common both for lipid headgroups and for cytoskeletal components .
as long as these charges are compensated by counterions on scales smaller than the local radii of curvature , they enter the mechanics only via their effect on the mechanical moduli . on the other hand ,
when relevant lengths become comparable to or smaller than the debye length , then charge effects must be treated as nonlocal .
our calculation provides echinocyte sizes and shapes in excellent agreement with experiment .
although shapes are calculated under the approximation of spicule axisymmetry , there is evidence ( e.g. , fig .
6 ) that this approximation is good .
the approximation does not allow us to estimate internally the range of echinocyte iii stability ; however , extrinsic arguments ( sec .
iv ) suggest that observable shapes should be bounded at low @xmath132 in the region where @xmath144 first becomes smaller than @xmath0 and at large @xmath132 in the region where the spicule dimension becomes comparable to the cytoskeletal discreteness . and ,
indeed , outside this region the predicted shapes are no longer seen experimentally . while this consistency lends credibility to the continuum picture ,
the way in which biochemical influences drive the control parameters of the model is not addressed .
the sheetz - singer parameter , @xmath132 , which measures the area difference between the lipid monolayers ( and the related effect of membrane spontaneous curvature ) does seem to be the dominant control parameter for some kinds of induced shape change . on the other hand , other effects such as the ph effect
may probe other dimensions of the shape / phase diagram , e.g. , the role of the cytoskeletal elastic constants and prestress .
we hope that having a mechanical model with predictive power may eventually help to elucidate important biochemical questions . from this perspective
, experiments can now focus on the way in which biochemical probes affect all the mechanical control parameters , including elastic and cytoskeletal variables , in addition to @xmath132 .
acknowledgement : we thank evan evans , ted steck , and narla mohandas for advice and encouragement , and dennis discher for helpful discussions and suggestions .
andrew rutenberg kindly supplied some important references .
this work was supported in part by the natural sciences and engineering research council of canada .
rm acknowledges support for a part of the work from the national science foundation under grant number dmr00 - 96531 .
the euler - lagrange equations obtained by minimizing @xmath59 may be written down as the following set of five coupled ordinary differential equations expressing the variables @xmath86 , @xmath79 , @xmath82 , @xmath168 and @xmath47 as functions of the arclength @xmath80 : @xmath169 - b(s ) \cos \theta + p r^{2 } \sin \theta & = & 0,\end{aligned}\ ] ] @xmath170 + 2 p r \sin \theta.\end{aligned}\ ] ] here @xmath171 is a local variable introduced to impose the constraint , @xmath172 , as first suggested by peterson ( 1985 ) .
the variables @xmath80 , @xmath86 , @xmath79 , @xmath82 and @xmath47 vanish at the north pole .
note also that the surface tension @xmath173 used in these equations is , in general , shifted from the parameter @xmath57 appearing in eq .
[ phi ] .
christiansson , a. , f.a .
kuypers , b. roelofsen , j.a.f .
op den kamp , and l.l.m .
van deenen .
lipid molecular shape effects erythrocyte morphology : a study involving replacement of native phosphatidylcholine with different species followed by treatment of cells with sphingomyelinase c or phospholipase a. _ j. cell biol .
_ 101:1455 - 1462 .
hgerstrand , h. , m. danieluk , m. bobrowska - hgerstrand , a. igli , a. wrbel , b. isomaa , and m. nikinmaa . 2000
. influence of band 3 protein absence and skeletal structures on amphiphile- and ca@xmath174-induced shape alterations in erythrocytes : a study with lamprey ( _ lampetra fluviatilis _ ) , trout ( _ onochorhynchus mykiss _ ) , and human erythrocytes .
_ biochim .
acta _ 1466:125 - 138 .
schwarz , s. , b. deuticke , and c.w .
passive transmembrane redistributions of phospholipids as a determinant of erythrocyte shape change .
studies on electroporated cells . _
_ 16:247 - 255 .
schwarz , s. , c.w .
haest , and b. deuticke .
extensive electroporation abolishes experimentally induced shape transformations of erythrocytes : a consequence of phospholipid symmetrization ?
_ biochim .
1421:361 - 379 .
figure 1 : : \(a ) a budded or vesiculated shape , which would be a low - energy configuration of a lipid membrane without cytoskeleton at sufficiently positive spontaneous curvature .
( b ) the spiculated shape into which ( a ) is transformed by the elastic - energy cost of cytoskeletal deformations in the high - shear neck regions .
figure 2 : : parametrization of a single - spicule shape , showing the boundary @xmath69 where the spicule joins the spherical central body in constructing the full echinocyte shape .
the bases of two adjacent spicules are sketched in to illustrate the points a of spicule tangency and the point b which is symmetrically situated between three neighboring spicules .
figure 3 : : sequence of spicule shapes for increasing values of the reduced effective relaxed area difference @xmath131 , eq .
( [ deltaa0 ] ) , which measures the combined effect of spontaneous curvature and additional area in the outer leaflet in driving the membrane to bend outward .
( a)(d ) correspond to @xmath131 = 0.014 , 0.018 , 0.020 , and 0.022 , respectively .
note the bar shows the elastic length scale @xmath0 .
note how the spicule sharpens as @xmath131 increases and how the neck begins to form when the spicule dimension ( set by @xmath144 ) falls below @xmath0 so that the bending energy begins to dominate .
figure 4 : : calculated equilibrium number of spicules @xmath19 plotted as a function of the reduced effective relaxed area difference @xmath131 .
figure 5 : : calculated effective spontaneous curvature @xmath97 , eq .
( [ effectivecurvature ] ) , plotted as a function of @xmath131 .
figure 6 : : calculated spicules at @xmath135 .
( a ) shape found in the axisymmetric approximation , as discussed in this paper ( @xmath136 , @xmath175 m ) .
( b ) shape found by numerical minimization on a triangulated surface ( @xmath176 , @xmath177 m ) .
figure 7 : : the local pressure @xmath145 , eq .
( [ pressure ] ) , exerted by the cytoskeleton on the bilayer plotted as a function of radial distance , for a spicule corresponding to @xmath178 .
figure 8 : : dependence of spicule shape on parameters at @xmath135 .
( a ) shape with the standard rbc parameters , as discussed in the text .
( b ) effect of starting from ( a ) and increasing @xmath10 by a factor of 2 ( i.e. @xmath179 ) .
the spicules become slightly broader .
( c ) effect of reducing overall cell dimension by a factor of two , so that @xmath180 and @xmath181 .
this change reduces spicule size by about 10% . | we study the shapes of human red blood cells using continuum mechanics . in particular
, we model the crenated , echinocytic shapes and show how they may arise from a competition between the bending energy of the plasma membrane and the stretching / shear elastic energies of the membrane skeleton .
in contrast to earlier work , we calculate spicule shapes exactly by solving the equations of continuum mechanics subject to appropriate boundary conditions . a simple scaling analysis of this competition reveals an elastic length @xmath0 which sets the length scale for the spicules and is , thus , related to the number of spicules experimentally observed on the fully developed echinocyte . |
the extensive mass loss which occurs from stars on the asymptotic giant branch ( agb ) leads to the formation of dusty , extended , circumstellar envelopes ( cse ) and the production of many gas - phase molecular species in the atmospheres and cse ( e.g. olofsson 2008 ) .
the cses of agb stars eventually contribute their heavy elements and dust grains back to the interstellar medium ( ism ) .
molecules from the cse may be returned directly to the ism or by incorporation into or onto the dust grains
. mass loss from agb stars can be as high as 10@xmath6 yr@xmath1 or even higher ( habing 1996 ) and is the dominant process at this stage of evolution .
it determines the stellar lifetime on the agb , the maximum luminosity reached , the post - agb evolution of the star , the gas and dust return to the ism , and the chemical composition of the returned gas ( bloecker 1995 ) . although the mass loss mechanism is complex and dependent on many different physical and chemical processes , the reasonably simple geometry and kinematics of the resulting cses make them excellent astrophysical and astrochemical laboratories .
interferometric imaging of thermal molecular emission in cses have been mostly confined to the nearby carbon - rich object irc+10216 , which has a very high mass - loss rate , although there have been targeted imaging projects of other sources .
if we are interested in a much better understanding of circumstellar chemistry , we must observe more typical sources , compare oxygen- and carbon - rich cses , use high - sensitivity interferometry to provide detailed images of the molecular line distribution , and cover a large number of molecular species transitions . to do all these successfully one line at a time ( or even a few )
would be prohibitive in terms of telescope time .
thus an interferometric imaging line survey , as being carried out by the submillimeter array ( patel et al .
2011 ) is required . in this short _ letter
_ we report on a project that is a pilot for observations of such an imaging line survey using the expanded very large array ( evla ; perley et al .
we proposed to observe one each of a carbon - rich cse ( rw lmi @xmath2 cit6 ) , an oxygen - rich cse ( ik tau @xmath2 nml tau ) , and an s - type agb star ( @xmath7 cyg ) . at the time of observations ,
the widar correlator could process a maximum of 1 ghz bandwidth per polarization with spectral channels of @xmath01 km s@xmath1 at the ka receiver band ( 26.5 40 ghz ) .
we were unable to observe @xmath7 cyg because necessary local sidereal time ( lst ) coverage was unavailable .
ik tau and rw lmi were observed with the evla in the * c * configuration during 2010 december and 2011 january . each target
was observed using four dynamic scheduling blocks ( observing runs ) of 3.5 hours , yielding a total of about 7 hours on source integration time each with a fixed frequency setting simultaneously covering 36.000 to 36.512 and 36.744 to 37.000 ghz . for each subband of 128 mhz bandwidth , 1024 channels with a channel separation of 125 khz ( @xmath0 1 kms@xmath1 ) were produced .
due to limitations during the early commissioning phase , the full one - ghz range was not available at this resolution , and a relatively large amount of time was spent on additional calibration . for ik
tau , 3c 48 was used for absolute flux calibration , 3c 84 for bandpass and delay calibration , and j0409 + 1217 for gain ( amplitude and phase ) calibration . for rw
lmi , 3c 286 was used for absolute flux calibration , 3c 273 for bandpass and delay calibration , and j0958 + 3224 for gain calibration .
standard data editing , flux , bandpass and phase calibration were performed using aips and a modified version of the vla pipeline .
the four data chunks per source were combined using the strongest line by correcting for the approximate doppler shift for the day of observation , although the diurnal variation was ignored for these short observations .
the maximum shifts applied were 5 and 8 channels for ik tau and rw lmi , respectively . before imaging all individual channels using natural weighting and 0.165 pixel size without cleaning ,
the data were averaged to 20 second visibilities and separated by subband , which speeds up the process considerably .
the continuum emission and spectral channels containing interesting spectral line features were selected and re - imaged with cleaning .
the synthesized beams from the imaging were @xmath00.7 x 0.6 arcseconds , depending slightly upon the frequency and the source .
the typical rms noise in a single 125 khz spectral channel is @xmath0500 @xmath8jy beam@xmath1 , and @xmath010 @xmath8jy beam@xmath1 in the continuum .
radio continuum emission at @xmath036.5 ghz was detected from both ik tau ( 970@xmath910 @xmath8jy beam@xmath1 ) and rw lmi ( 890@xmath910 @xmath8jy beam@xmath1 ) .
line emission was detected from the hc@xmath3n ( j @xmath2 4@xmath43 ; rest frequency 36392.332 mhz ) , the sis ( j @xmath2 2@xmath41 ; 36309.627 mhz ) , and the hc@xmath5n ( j @xmath2 32@xmath431 ; 36095.546 mhz ) transitions in the carbon - rich cse in rw lmi .
the only spectral line clearly detected in ik tau was the sis ( j @xmath2 2@xmath41 transition , although the total frequency coverage included the so ( j , n @xmath2 ( 2,3)@xmath4(2,2 ) ; 36202.041 mhz ) transition , and the ocs ( j @xmath2 3@xmath42 ; 36488.813 mhz ) transition . figure 1 shows the emission in nine of the 34 velocity images of the hc@xmath3n line from rw lmi .
these images show striking concentric rings in the hc@xmath3n emission ; very clear evidence of multiple spherical shells of emission . in the large - scale structure of the ring emission ,
there are also noticeable asymmetries in the west and northwest near the central velocities , and to the east in more blue - shifted velocities .
the evla data is more sensitive at higher angular and spectral resolution than the spectrally averaged hc@xmath3n ( j @xmath2 5@xmath44 ) data from the vla at 45.5 ghz presented by dinh - v - trung & lim ( 2009 ) .
the similarities of the two data sets are striking in terms of asymmetries ; however the higher sensitivity evla images show a wealth of detail ( e.g.the several emission shells ) not found in the vla data . figure 2 shows images of the continuum - subtracted , averaged ( in velocity ) emission of the sis line from both rw lmi and ik tau ( all channel images showing emission were averaged together after subtracting the continuum in the _ u - v _ data ) . as compared with the hc@xmath3n emission , the sis emission from rw lmi is more compact , centrally condensed , but slightly elongated in the n - s direction . for ik tau the sis emission is also centrally concentrated toward the star , and is also elongated rather than circular .
although we detected the hc@xmath5n line in rw lmi , the emission is quite weak , and further analysis of this transition will be deferred to another publication .
figure 3 shows the global emission of the hc@xmath3n and sis lines from rw lmi .
these spectra were made by summing the emission in the same rectangular region of each spectral channel image .
the channel with the largest emission extent was used to set the rectangular region for each line .
the spectra for both the sis and hc@xmath3n lines are clearly asymmetric with the red - shifted side being the brightest .
at submillimeter and infrared wavelengths the continuum emission from agb stars is dominated by warm circumstellar dust which has a very steep spectral index ( draine 2006 and references therein ) .
we compared the detections of the radio continuum emission at 36.5 ghz with measurements of other authors at different frequencies .
marshall et al .
( 1992 ) observed both rw lmi and ik tau at frequencies of 264 , 394 , and 685 ghz .
they find the derived spectral index from 100 @xmath8 m ( using _ iras _ data ) through 1.1 mm to be @xmath103.9 ( ik tau ) and @xmath113.5 ( rw lmi ) . at centimeter wavelengths emission from dust
is small compared to that from the radio `` photosphere '' ( reid & menten 1997 ) i.e. blackbody emission from or just outside the stellar photosphere with a temperature approximately 2000 k , and with a spectral index near @xmath112.0 .
radio continuum measurements of rw lmi at @xmath03 mm or longer include 13.3 mjy beam@xmath1 at 112 ghz ( neri et al .
1998 ) , 8 mjy at 90.7 ghz ( lindqvist et al . 2000 ) , and 2.4 mjy beam@xmath1 at 43.3 ghz ( dinh - v - trung & lim 2009 ) . a two - point spectral index derived from 36.5 ghz to 90.7 ghz and from 36.5 ghz to 112 ghz is consistent and equal to @xmath12 @xmath112.3 .
the spectral index of 2.3 is reasonably close to that of thermal blackbody radiation .
spectral indices calculated using the flux measured at 43.3 ghz to both higher and lower frequencies are inconsistent and do not fit either photospheric emission or dust emission .
all other existing continuum data agree with either photospheric or dust emission , and in particular the present evla measurements are likely detecting the photosphere of rw lmi .
a radio continuum measurement of ik tau at 96.7 ghz of 6.1 mjy beam@xmath1 was reported by marvel ( 2005 ) .
the measurement presented here , combined with marvel s , produce a two - point spectral index of @xmath112.0 .
this is just what is expected from a radio photosphere emitting as a blackbody .
we used the aips task iring to azimuthally sum the hc@xmath3n emission in rw lmi in continuous rings of a fixed width centered on the continuum , for each velocity channel . using this method we then measured the angular radii of the flux density peaks ( rings in figure 1 ) in each channel . assuming that we then can follow the radius of each ring in different channels we can plot , for each detected ring , the radius versus the ( lsr ) velocity for each ring . based on this preliminary analysis , we distinguished four such rings of emission ; a more careful examination of the channel images in conjunction with the ring analysis is certainly required .
the plot of radius versus velocity is shown in figure 4 .
the radius in arcseconds is transformed to the radius in au using a distance of 440 pc ( schier et al .
2002 ) .
if the rings of emission actually correspond to spherical shells of molecular emission around the star , the radius of the shell as a function of radial velocity is given by @xmath13 where @xmath14 is the stellar radial velocity , @xmath15 is the shell radius , and @xmath16 is the constant expansion velocity for a given shell ( e.g. bowers et al .
we fit this equation to the data obtained for the four hc@xmath3n shells as shown in figure 4 , using a simple least - squares method , and using @xmath17 , and @xmath16 as ( bounded ) free parameters in the fit .
the parameters for each of the shells are listed in table 1 . for comparison ,
the expansion velocity and stellar radial velocity are determined to be 17 km s@xmath1 and @xmath111.0 km s@xmath1 ( with respect to the lsr ) , respectively , based on @xmath18co observations ( schier & olofsson 2000 ) , and the expansion velocity estimated from the spectra in figure 3 is also 17 18 km s@xmath1 .
given the expansion velocity and the radius , a time can be derived which is the approximate age of the shell .
this age is also listed in table 1 .
the age of shell 3 ( from table 1 ; blue symbols in figure 4 ) is essentially the same as the age of shell 2 ( red symbols in figure 4 ) this is clearly because the fit from the data for shell 3 gives a much larger expansion velocity than that from shell 2 .
we speculate that since these two shells have the same age , they may have a common origin .
if so , this would suggest a rather complex behavior in the mass - loss history for rw lmi .
these fitted parameters are only approximate , since it is not always clear how to associate the data into rings .
( the red symbols in figure 4 are an example of this . )
also , the equation above ( essentially an ellipse ) does not fit the blue - shifted side of the data very well : note that the data on the blue - shifted side of figure 4 are quite linear as compared to the data on the red - shifted side ( which are fit by the velocity equation much better ) .
perhaps a better way of saying this is that , except for shell 1 ( which does not have much emission on the red - shifted side ) the data , if extended to the shell center , would intersect 0 au at approximately the same lsr velocity ( @xmath0@xmath1913 km s@xmath1 ) .
the extensions of the data for the different shells on the blue - shifted side would not intersect 0 au close to the same velocity at all . we note that the velocity of @xmath1913 km s@xmath1 corresponds to the velocity in both the spectra in figure 3 where the steepest part of the spectrum begins .
we also point out that full width at zero power ( fwzp ) of both the sis and hc@xmath3n lines are essentially the same , despite the fact that the spatial distribution and extent is much different .
this suggests that the gas is accelerated quickly in the inner region of the cse to the expansion velocity . the strong asymmetric character ( with the red - shifted side stronger ) of both these transitions as well as others ( e.g. the sis , j @xmath2 5@xmath44 transition ; lindqvist et al .
2000 ) must provide clues to detailed kinematics of the gas . while in general it is clear from figure 1 that the hc@xmath3n emission from rw lmi consists mostly of broken rings ( or shells ) of emission ,
there are some velocity channels which , if the emission is traced carefully , might be considered as a spiral shape ( c.f . the velocity channel at @xmath117.5 km s@xmath1 in figure 1 ) .
dinh - v - trung & lim suggested such a spiral for the 45.5 ghz transition and there is at least one beautiful example of spiral * dust * emission in a cse ( afgl 3068 ; mauron & huggins 2006 ) , but if there is such a spiral in our images it has the opposite handedness than that shown in the discussion of dinh - v - trung & lim .
thus we think that the possibility of molecular spiral emission in rw lmi is unlikely , though we do not completely rule it out .
dinh - v - trung & lim ( 2008 ) have used high angular resolution observations using the vla to trace several incomplete arcs ( or shells ) of hc@xmath3n emission ( at the 45.5 ghz transition ) and hc@xmath20n emission ( transitions at 42.6 and 24.0 ghz ) in the prototypical agb carbon star irc+10216 , similar to what we have observed in rw lmi .
dinh - v - trung & lim ( 2008 ) estimate the kinematic timescale between the shells is in the range of @xmath0120 to @xmath0360 yr .
this is also similar to the timescales shown in table 1 , based on our data for rw lmi .
the molecular arcs in irc+10216 appear to coincide closely with the dust arcs shown by mauron & huggins ( 2000 ) . for rw lmi , schmidt et al .
( 2002 ) have discovered several faint diffuse arcs of reflected light using deep hst nicmos imaging .
these arcs appear at radial distances of 1.0 , 1.3 , 3.8 , and 4.5 arcseconds from the star , and are most prominent to the south - southwest of the star .
only one of these reflected light arcs may be coincident with the hc@xmath3n shell ( at 4.5 arcseconds radius ) based on angular radius alone . in addition , as can be seen in figure 1 , the most prominent parts of the rings seen in hc@xmath3n are typically not in the south - southwest part of the images .
figure 2 shows the sis emission from both agb stars averaged over the line .
we note that there is a large difference in angular size of the sis emission between the two stars
. this may be only a result of the generally weaker emission in ik tau , but the angular size difference is striking , and , assuming a distance to ik tau of 250 pc ( olofsson et al .
1998 ) , the total physical extent of the sis emission is different by a factor 4 ( the emission in rw lmi @xmath02000 au is 4 times larger than in ik tau @xmath0500 au ) .
the peak of the sis emission is coincident with that of the radio continuum emission which defines the stellar position .
the size estimate for rw lmi is consistent with the modelling of the sis molecular abundance by schier et al .
the total sis envelope size based on their table 3 is @xmath01700 au .
the modelling is based on single - dish spectra ; however , for rw lmi the model appears to fit the interferometric data from the sis , j @xmath2 5@xmath44 transition at @xmath090.7 ghz ( lindqvist et al .
2000 ) rather well . for ik
tau , the model gives a total sis envelope size of 2200 au ( a factor 4 larger than our measurement ) .
schier et al .
point out that their single - gaussian abundance distribution gives a poor fit to the mm and submm - wave spectra of sis for ik tau .
the addition of a compact , high abundance component to the model improves the fit , especially for higher j transitions .
we speculate that the oxygen - rich chemistry of ik tau may have some bearing on the detailed radial abundance of sis in oxygen - rich cse .
it is clear that a larger sample of stars ( both oxygen - rich and carbon - rich cses ) observed with interferometric imaging at high angular resolution are needed to provide the best constraints on such models .
we have used the ever - growing capabilities of the expanded very large array to perform a @xmath01 ghz , targeted line survey of two agb stars with @xmath01 km s@xmath1 spectral resolution ( 125 khz at 18 - 50 ghz ) and sub - arcsecond angular resolution .
the stars are : rw lmi ( with a carbon - rich cse ) , and ik tau ( with an oxygen - rich cse ) .
we find the following results in a preliminary analysis of the data : @xmath21 the hc@xmath3n ( j = 4@xmath43 ) emission from rw lmi is bright enough to be imaged in detail and shows strong asymmetries in the global line profile and partial rings ( shells ) of emission .
we have provided a preliminary analysis of the shell emission and find that the multiple shells can be traced from about 400 years to 1200 years .
in addition , we have detected weak hc@xmath5n ( j @xmath2 32@xmath431 ) emission from rw lmi .
@xmath21 the sis ( j = 2@xmath41 ) emission from both stars is compact , compared with the hc@xmath3n emission from rw lmi .
the emission seems elongated for both stars rather than ring - like .
the extent of the detected sis emission in ik tau is only one - quarter ( @xmath0500 au ) that of rw lmi ( @xmath02000 au ) the sis global line profile from rw lmi is also strongly asymmetric and in the same sense ( red - shifted side stronger than blue - shifted side ) as the hc@xmath3n line . by the end of the evla construction period ( in early 2013 )
spectral imaging surveys should be able to use 8 ghz per polarization with 125 khz resolution ( or better ) in the widar correlator simultaneously .
thus the 3 receiver bands which span the 18 - 50 ghz range could be covered with 5 frequency settings .
this will lead to a revolution in spectral imaging surveys of cm molecular transitions in agb stars and related objects .
bloecker , t. 1995 , , 297 , 727 bowers , p. f. , johnston , k. j. , & spencer , j. h. 1983 , , 274 , 733 habing , h. 1996 , , 7 , 978 dinh - v - trung & lim , j. 2008 , , 558 , 204 dinh - v - trung & lim , j. 2009 , , 701 , 292 draine , b. t. 2006 , , 636 , 1114 marshall , c. r. , leahy , d. a. , & kwok , s. 1992 , , 104 , 397 marvel , k. b. 2005 , , 130 , 261 lindqvist , m. , schier , f. l. , lucas , r. , & olofsson , h. 2000 , , 361 , 1036 mauron , n. & huggins , p. j. 2000 , , 359 , 707 mauron , n. & huggins , p. j. 2006 , , 452 , 257 neri , r. , kahane , c. , luca , r. , bjuarrabal , v. , loup , c. 1998 , , 130 , 1 olofsson , h. , lindqvist , m. nyman , l - . , winnberg , a. 1998 , , 329 , 1059 olofsson , h. 2008 , , 313 , 201 patel , n. a. et al .
2011 , , 193 , 17 perley , r. a. , chandler , c. j. , butler , b. j. , & wrobel , j. m. 2011 , , this issue .
reid , m. j. & menten , k. m. 1997 , , 476 , 727 schmidt , g. d. , hines , d. c. , & swift , s. 2002 , , 576 , 429 schier , f. l. & olofsson , h. 2000 , , 359 , 586 schier , f. l. , ryde , n. , & olofsson , h. 2002 , , 391 , 577 schier , f. l. , bast , j. , olofsson , h. , & lindqvist , m. 2007 , , 473 , 871 ccccc shell & @xmath14 ( km s@xmath1 ) & @xmath15 ( au ) & @xmath16 ( km s@xmath1 ) & age ( yr ) + 1 & -2.6 & 880 & 11.3 & 370 + 2 & -3.1 & 2000 & 12.0 & 800 + 3 & -2.0 & 2500 & 15.2 & 780 + 4 & -1.2 & 3650 & 14.8 & 1200 + n ( j @xmath2 4@xmath43 , rest frequency 36392.332 mhz ) emission from rw lmi .
this selection spans the range of emission for the hc@xmath3n transition , and each panel is every fourth channel .
each image is @xmath021 arcseconds on a side and is labeled with the velocity with respect to the lsr .
the cross marks the position of the continuum source ; the size of the cross is about ten times the position error of the continuum measurement .
north is up and east is to the left . ] | we report on a pilot imaging line survey ( 36.0 - 37.0 ghz , with @xmath01 km s@xmath1 spectral channels ) with the expanded very large array for two asymptotic giant branch stars , rw lmi ( @xmath2 cit6 , which has a carbon - rich circumstellar envelope ) and ik tau ( @xmath2 nml tau , with an oxygen - rich circumstellar envelope ) .
radio continuum emission consistent with photospheric emission was detected from both stars . from rw
lmi we imaged the hc@xmath3n ( j @xmath2 4@xmath43 ) emission .
the images show several partial rings of emission ; these multiple shells trace the evolution of the cse from 400 to 1200 years .
sis ( j @xmath2 2@xmath41 ) emission was detected from both rw lmi and ik tau .
for both stars the sis emission is centrally condensed with the peak line emission coincident with the stellar radio continuum emission .
in addition , we have detected weak hc@xmath5n ( j @xmath2 32@xmath431 ) emission from rw lmi . |
a major fraction of the practical problems in the operation of cyclotrons are related to beam extraction : the activation of extraction elements increases the personal dose during maintenance work and sometimes requires to shutoff the beam long before the scheduled work .
the electrostatic elements are frequently the cause of beam interruptions due to high voltage trips , they require regular maintenance like cleaning and conditioning . in order to increase the extraction efficiency , the energy gain per turn must be maximized , which requires to run cavity and resonators at the limit of what can be achieved .
this in turn increases the frequency of cavity trips and amplifier failures . in this work ,
we propose to utilize the mechanism of stripping extraction , which is fairly well - established in many machines worldwide in nearly the complete energy and intensity range that can be achieved by cyclotrons @xcite .
the extraction mechanism that we present here might help to avoid most extraction problems completely . in the case of variable energy extraction as we propose for proton therapy machines , energy degraders and energy selection systems can be omitted , thus reducing the costs and the facility footprint significantly
. since the required beam intensities are typically in the order of @xmath1 at the patient only , it should be possible to keep a cyclotron with variable energy extraction almost free from activation of components .
higher beam currents of up to @xmath2 are mainly required to compensate the losses of energy degradation and collimation @xcite .
variable energy extraction by stripping has been proposed and used at the manitoba cyclotron by stripping of @xmath3-ions @xcite and in the raccam cyclotron @xcite .
unfortunately , the @xmath3-ion is not stable in strong magnetic fields at high energy , so that @xmath3-cyclotron are either limited in energy or restricted magnetic field values .
this requires large radius machines like the triumf cyclotron @xcite .
furthermore , the use of @xmath3 ions in accelerators is more demanding with respect to the machine vacuum and the production of @xmath3 in ion sources . cyclotrons ( and/or ffags ) with reverse bends have been proposed in the past @xcite mainly in order to achieve the focusing conditions that are required for energies of @xmath4 and above .
however there is no publication known to the authors that proposes the use of reverse bends in combination with stripping extraction . in
most ( if not all ) cases where stripping extraction of @xmath0-ions is used , the proposed extraction schemes lead to complicated orbits that circle one or even multiple times within the cyclotron before the beam exits @xcite .
the use of this method for multiple or even for continously variable energy extraction is difficult - if at all possible .
another method to achieve beam extraction at variable energy is the variation of the main field of the cyclotron and to use a sequence of trim coils to achieve isochronism for the desired extraction energy .
this method is the most `` natural '' way and it is known to work .
however the minimal time to switch between energies is given by the ramping of the main field and the magnetic relaxation time of the yoke .
in the optimal case it might be possible to realize energy switching within minutes .
we are aiming for a millisecond range , i.e. energy switching times that are compatable to the time that is required to adjust a beamline with laminated magnets to the new energy .
the goal of this work is to present first basic geometrical and beam dynamical studies in order to investigate the feasibility of variable energy cyclotrons and to explore the energy ranges that could be achieved . concerning the beam dynamics
we restrict ourselves to the minimum , which we consider to be the verification of the stability of motion of a coasting beam and of the extraction mechanism . in order to survey the parameter space for such machines
we restrict ourselves to the so - called hard edge approximation of the magnets .
and we further simplify this approach by assuming homogeneous magnetic fields within sectors and valleys .
isochronism is achieved exclusively by a variation of the azimuthal sector width along the orbit @xcite . in sec .
[ sec_geom ] , we give a description of the geometry and the calculation of the transfer matrices .
the equations given there have been used in mathematica^^ to analyze the orbits and the traces of the transfer matrices in hard edge approximation in order to find stable solutions with the desired extraction orbits .
base on the results , a `` c''-program was used to generate smoothed magnetic field maps , which have then been analyzed with an equilibrium orbit code @xcite and a cartesic tracking code to verify the analytical results of the hard edge approximation numerically . in sections
[ sec_isocyc]-[sec_adscyc ] we present the results of the calculations .
we consider @xmath0-cyclotrons that are composed of @xmath5 identical sections which are each composed of a sequence of homogeneous sector magnets , reverse bends with homogeneous fields and ( optionally ) drifts .
we do not consider beam injection nor other details of central regions in detail .
we are not concerned about the question , if these machines might need pre - accelerators or can be made `` compact '' .
we first consider machines that are composed of exclusively positive and negative bends as shown in fig .
[ fig_h2cyc4 ] , where we call the positive bends `` sector '' and the negative bends `` valley '' .
the absolute values of the sector ( valley ) field is @xmath6 ( @xmath7 ) , the corresponding bending angle is @xmath8 ( @xmath9 ) , then we have for an ion of mass @xmath10 , charge @xmath11 and momentum @xmath12 : & = & _ s-_v + r_s&=&pqb_s = mcqb_s + r_v&=&pqb_v = mcqb_v + l_tot&=&2n(r_s_s+r_v_v)=2nmcq(+ ) . [ eq_cyc1 ]
isochronism requires that the velocity @xmath13 , the orbital angular frequency @xmath14 and the total length of the orbit @xmath15 are related by v&=&l_tott=_ol_tot2 + & = & vc=_ol_tot2c = l_tot2a , [ eq_cyc2 ] where @xmath16 is the cyclotron length unit . in combination with eqn .
( [ eq_cyc1 ] ) this yields :
2a&=&2nmcq(+ ) + a&=&nmcq(+ ) + _ s&=&n(1+)(b_vb_0 + 1 ) , [ eq_cyc3 ] where we used @xmath17 and defined the `` nominal field '' @xmath18 by b_0= mcaq = mq_o . from fig . ( [ fig_h2cyc4 ] )
we pick the following equations r_v&=&r + & = & r(- ) + r_s&=&r , from which we obtain in a few steps = ( 1+)- .
if @xmath19 is the azimuthal angle of the sector center and @xmath20 ( @xmath21 ) are the angles of entrance ( exit ) of the orbit into the sector , then _ i&=&_c- + _
f&=&_c+ + the angles @xmath22 ( @xmath23 ) between the sector edges and the radial direction can be obtained by & = & rd_idr = rd_id / drd + & = & rd_fdr = rd_id / drd + where @xmath24 is the radius of the orbit entering ( exiting ) the sector . the angles between the orbit and the sector edges ( which are required for the transfer matrices ) can be computed by _ i&=&_s+_i- + _ f&=&_s-_f- .
the radii of the arc centers in the sector @xmath25 and the valley @xmath26 are : _
s&=&r_s(- ) + _ v&=&r_v(+ ) + the cartesic coordinates @xmath27 of the orbit inside the sector in dependence of the angle @xmath28 can be written as x_s()&=&(_s+r_s)-r_s + y_s()&=&(_s+r_s)+r_s , where @xmath28 ranges from @xmath29 to @xmath30 .
correspondingly one finds for the valley : x_v()&=&(_v - r_v)-r_v + y_v()&=&(_v - r_v)+r_v , where @xmath28 ranges from @xmath31 to @xmath32 .
if we assume that a stripper foil is placed exactly at the sector exit ( i.e. at radius @xmath24 and azimuthal angle @xmath21 ) , then the center coordinates @xmath33 of the arc described by the extracted orbit are x_c&=&(_s+(r_s+r_v/2 ) ) + & -&(r_s+r_v/2 ) + y_c&=&(_s+(r_s+r_v/2 ) ) + & + & ( r_s+r_v/2 ) + coordinates @xmath34 of the extracted orbit are : x_x()&=&x_c+r_v/2 + y_x()&=&y_c+r_v/2 , where @xmath28 starts at zero . with the above equations
, we analyzed the geometry of the orbit and the extraction for different choices of @xmath6 , @xmath7 , @xmath18 and @xmath35 as a function of @xmath36 .
the horizontal transfer matrices @xmath37 for the sector ( valley ) are given by _
s , v= & r_s , v + -r_s , v&. the horizontal transfer matrix that describes the edge focusing effect is _ i , f= 1&0 + r_&1 . where @xmath38 .
starting with the entrance into the sector magnet the horizontal transfer matrix @xmath39 for a single section is the product _
the radial focusing frequency @xmath40 can be obtained from the parametrization by the twiss - parameters @xmath41 , @xmath42 and @xmath43 : _ sec=*1*+ _
t & _ t + -_t&-_t + , from which one obtains with @xmath44 & = & tr(*m*_sec)/2 + ^2&=&det(*m*_sec-*1*tr(*m*_sec)/2 ) + the matrices for the vertical motion are _
s , v&= & 1&2r_s , v_s , v + 0&1 + + * t*_i , f&= & 1&0 + -r_&1 .
so that correspondingly _ sec=*t*_v*t*_f*t*_s*t*_i .
the motion is stable , if tr(m_sec)/2 & & 1 + tr(t_sec)/2 & & 1 . in case of cyclotrons with reverse bends
, one has a huge flutter @xmath45 due to the negative field regions . if one uses the dimensionless ratios @xmath17 and @xmath46 , then the flutter yields f=(1-)(+)^2 ^ 2 , and can approch easily values of above @xmath47 .
therefore care must be taken to not have too strong focusing , i.e. to avoid the @xmath48-stopband . due to the huge flutter , there is no need for large spiral angles . in contrary
, the spiral angle must be kept sufficiently small to avoid the stopband .
the described simplified cyclotron description allows a first analysis , if extraction at various energies can be combined with stability of axial and horizontal motion . fig .
[ fig_h2isocyc ] shows a topview layout for a isotope production machine with maximal @xmath0-energy of @xmath49 that allows stripping extraction of proton beams with energies between @xmath50 and @xmath51 .
-cyclotron for isotope production with reverse bends .
the @xmath0-beam is stripped at the sector edge ( indicated by symbols ) .
the orbits of stripped proton beams are shown from @xmath52 to @xmath51 in steps of @xmath53 .
the arrows indicate the extracted beams for @xmath50 , @xmath54 , @xmath55 , @xmath56 , @xmath57 and @xmath51 .
the nominal field @xmath18 is @xmath58 , the sector field @xmath6 @xmath59 and the field strength @xmath7 in the reverse bends is @xmath60 . without further provision ,
the directions of the extracted beams differ enough to allow for energy specific targets .
if the beam is partially stripped at lower energy , simultaneous irradiation at several targets should be possible .
if superconducting coils are used to increase the field strength , the size of the cyclotron reduces accordingly .
the sector entrance edge has been chosen straight @xmath61 so that @xmath62 and @xmath63 .
[ fig_h2isocyc],width=321 ] fig .
[ fig_isotunes ] shows the corresponding tune diagram . due to the strong flutter
, the axial tune @xmath64 is very large . as a consequence
the minimum number of sectors is likely @xmath47 , so that the @xmath48-stopband starts at @xmath65 .
higher sector numbers are in principle possible , but more expensive and not required for this energy range . .
the vertical tune @xmath64 starts at low energy at about @xmath66 and decreases smoothly to about @xmath67 at maximal energy . due to the large flutter ,
a stable solution for cyclotron with only 3 sectors has not been found and we assume that stable solutions can not be found for a comparable energy range of the extracted beam .
[ fig_isotunes],width=302 ] partial stripping of the beam could allow simultaneous extraction at multiple energies . for this purpose one would move a stripper foil vertically towards the median plane until it strips off the desired beam current for the corresponding energy . the remaining beam ( with reduced emittance )
could be accelerated to higher energies ( see fig .
[ fig_stripper ] ) . ,
commercially available cyclotrons for proton therapy deliver beams with an energy of @xmath68 @xcite . since the presently available cyclotron technology delivers the beam at fixed energy , the beam energy must be reduced to the value that is required for the treatment .
this is typically done by energy degradation at the cost of significant emittance increase and energy straggling in the degradation process @xcite . in order to deliver a beam of the required quality
most of the degraded beam has to be cut off by collimators and an energy selection system ( ess ) .
the intensity is ( depending on energy ) reduced by up to three orders of magnitude . even though there are strong arguments for the use of cyclotrons in proton therapy , there are also disadvantages of the combination of a fixed - energy - cyclotron , degrader and ess : 1 .
the strong energy dependence of the beam intensity which makes fast and save energy variations ( without intensity variations ) of the beam difficult to achieve .
2 . the activation of the accelerator , the degrader material , the collimators and other components , which could be reduced by orders of magnitude , if one could extract high quality beam at various energies .
3 . the cost for the degrader and the ess which typically consists of two dipoles , eight quadrupoles , moveable slits , beam diagnostics and vacuum components for about @xmath69 beamline .
the need to use large aperture quadrupoles and dipoles in order to achieve a suitable transmission efficiency of beam line and gantry .
the list is certainly incomplete , but it suffices to argue that one has to take the over - all costs of an accelerator concept into account .
a separate sector cyclotron with reverse bends is certainly more expensive than a compact cyclotron .
it will also have a larger footprint and a higher power consumption .
however the footprint of the accelerator itsself is only a small fraction of a complete proton therapy facility .
-cyclotron for proton therapy with reverse bends and drifts .
the equilibrium orbits and stripped proton trajectories for energies from @xmath51 to @xmath70 in steps of @xmath71 are also shown .
they been computed by an equilibrium orbit code @xcite and by runge - kutta tracking , respectively .
the nominal field @xmath18 is @xmath72 , the sector field @xmath6 @xmath59 and the field strength @xmath7 in the reverse bends is @xmath60 . with a bit more fine - shaping of the magnetic field of the reverse bend
, it should be possible to make all extracted beams pass a region small enough to install a fast `` catcher '' magnet , which allows to bend the extracted beams of all energies into the same beamline .
the resulting tunes are shown in the right graph .
the small spiral angle has been introduced to avoid the @xmath73-resonance shown as a solid straight line . using superconducting coils and correspondingly higher field values
, the size could be reduced accordingly .
if partial stripping would be applied , it should be possible to extract up to four beams simultaneously .
[ fig_medcyc],width=302 ] we found that variable energy extraction by a moveable stripper foil before a reverse bend allows _ in principle _ to extract beams with energies between @xmath51 and @xmath70 .
[ fig_medcyc ] shows the layout of an @xmath0-cyclotron with @xmath59 sector magnets and @xmath60 reverse bend magnets , the equilibrium and extraction orbits for energies from @xmath51 to @xmath70 .
the use of superconducting coils would allow to increase the field and the radius would reduce by the same factor . .
[ fig_medcyctunes],width=283 ] the time required for a change of the beam energy is then determined by the ramping time of the beamline magnets and the time for the positioning of the stripper foil .
if a series of foils at different radii would be inserted vertically into the beam , then the actuator would need just a few millimeters of motion for the insertion of the foil as shown in fig .
[ fig_stripper ] .
other mechanisms using radial motion with the advantage of continous energy adjustment are also possible .
even though the design of fast moveable parts in vacuum is not trivial , we believe that mechanisms should be feasible with a response time in the order of @xmath74 or below . since the extracted beam current that is required for radiation therapy is of the order of @xmath1 , cooling of the stripper foil is ( for this application ) not necessary .
more challenging ( in terms of costs and engineering time ) is the design of a central region that allows either to use an internal ion source or a spiral inflector .
an internal ion source causes a higher rest gas pressure compared to an external source .
however the beam current in such a pt machine is very low so that even a high relative beam loss by rest gas stripping could be accepted .
certainly the presented extraction mechanism could also be used in combination with a pre - accelerator , but the stripping process itsself can be used only once .
the preaccelerator would necessarily have a different extraction mechanism .
the design scetched in fig .
[ fig_medcyc ] has four sectors so that with an appropriate design of rf - resonators one might use at maximum four exit ports in four directions .
they might ( but do nt have to ) be used simultaneously in order to deliver beam for four treatment rooms located around the cyclotron bunker .
since a direct beam from the cyclotron has a small emittance and energy spread , the beam transport system does not require magnets with large aperture .
hence beamline and gantry might be smaller and cheaper than those of conventional systems .
if the beam size and energy spread are too small for fast painting of the tumor , one could insert scatterers into the beam path - or one might directly use `` thick '' stripper foils , which increase the beam size by scattering and make the beam shape more gaussian .
we used the flat field design since it allows to calculate the desired properties analytically in very good approximation .
however a cyclotron with a flat field has also practical advantages .
it allows for instance to make precise online field measurements by nmr - probes .
the results could be used to stabilize the magnetic field without beam extraction as it is required for a phase probe @xcite .
this would not only reduce start up time and simplify beam quality management , but it might also reduce activation of an external beam dump .
furthermore the mechanism that places the stripper foil - if fast enough - could be made `` fail - save '' : if a spring retracts the foil off the median plane in case of emergency , extraction immediately stops . without stripper foil but with an appropriate shaping of the edge field with enough phase shift per turn , the cyclotron could operate in a stand - by mode without activation and extraction but with contineous beam in the median plane .
the beam would be accelerated to maximal energy , phase shifted in the fringe field , decelerated back to the cyclotron center and dumped there without activation of components . in this way , the equipment could stay `` warm '' in stand - by mode .
if beam is requested , the only action to be taken is to insert the stripper foil at the desired location for the requested energy .
recently there has been renewed interest in high intensity cyclotrons not only for the potential use in accelerator driven systems ( ads ) for transmutation of nuclear waste or as `` energy amplifier '' @xcite , but also for physical experiments like @xmath75 @xcite .
typically the cyclotron should be able to deliver @xmath76 or more proton beam current at @xmath77 and @xmath78 .
such cyclotrons have never been build , but the psi ring machine which delivers @xmath79 at @xmath80 often serves as a proof - of - principle machine @xcite .
however , there is still a factor of @xmath81 between the beam power of the psi machine ( @xmath82 ) and the desired @xmath83 ( or more ) for an ads driver .
we are not going to discuss this in detail here , but we give an example of an @xmath0-cyclotron with stripping extraction between @xmath84 and @xmath85 .
the major advantage of the proposed extraction method is the increased reliability of extraction without electrostatic elements .
furthermore the concept allows to reduce the turn separation , i.e. the required voltage of the accelerating cavities is not a major design issue .
a flattop system could also be obsolete .
[ fig_adscyc ] shows a machine layout for a homogenous sector ( valley ) field of @xmath86 and @xmath87 . as shown in fig .
[ fig_adscyctunes ] , major resonances could be avoided by an adequat choice of the spiral angle .
-cyclotron for extraction energies in the range between @xmath84 and @xmath85 for ads .
the orbits of the protons after stripping are shown from @xmath88 to @xmath85 .
the nominal field @xmath18 is @xmath89 , the sector field @xmath6 @xmath86 and the field strength @xmath7 in the reverse bends is @xmath87 .
the tune diagram is shown in the right graph and covers the @xmath0-energy range from @xmath90 to @xmath91 .
a more advanced field shaping with reduced flutter at lower energies would keep the vertical @xmath64 below @xmath92 also at lower injection energy and would allow to stay above @xmath40 at higher energies .
the spiral angle has been chosen to be @xmath93 , @xmath94 and @xmath95 . [ fig_adscyc],width=302 ] there are two major differences between the cyclotron design here and the one proposed in ref .
@xcite , the first being the difference in the vertical tune , which is in our design considerably increased by the reverse bends .
the second is the trajectory of the stripped beam .
the design proposed in ref .
@xcite uses the conventional scheme in which the stripped beam is bend inwards and passes the cyclotron median plane at nearly all radii before exiting the field .
there is no principle problem with this scheme , but it has disadvantages .
first the exact position and direction of the extracted beam depends on the cyclotron field all along the extraction orbit which is more or less the complete median plane area .
it is therefore influenced by trim coils , cavities and main field changes .
second , this beam path has to be free from obstacles . fig .
[ fig_adscyc ] shows a machine layout for a homogeneous sector ( valley ) field of @xmath86 and @xmath87 , fig .
[ fig_adscyctunes ] the tune - diagram .
the spiral angle has been optimized to avoid major resonances . .
[ fig_adscyctunes],width=283 ] the extraction path with reverse bends is as short as possible and passes only the area between two sectors .
it is therefore much less sensitive to changes of main field and/or trim coil settings .
the discussed machine layouts allow some further optimization with respect to the direction of the extracted beam by an appropriate shaping of the fields of the sectors magnets and the reverse bends .
this would go beyond the scope of this paper , since our intention was to survey the _ principle _ possiblities of the extraction mechanism .
certainly the flat field approach used above is neither necessary for this extraction scheme to work nor do we consider it to be the optimal choice .
it has been chosen as it allows for a fairly simple analytical description of cyclotron beam optics .
the `` inner region '' of such cyclotrons , i.e. the energy range in which beam extraction is not possible , might be designed very different from what is scetched above .
the negative field in the reverse bends is not required at small radii .
therefore it is possible ( and unavoidable ) to reduce the effect of the reverse bend towards the cyclotron center ( compensating this with reduced sector field or sector width ) .
the beam in a cyclotron like the ones described above should be centered so that the energy and radius are related in a predictable and reproducable way .
this is especially important for pt applications , where the position of the stripper foil selects the beam energy .
since resonant beam extraction is not required , the phase curve may be chosen flat up to the cyclotron fringe field .
this allows to accelerate beams with relatively low cavity voltages .
since neither a low energy spread nor a high turn separation is essential in order to minimize extraction losses ( depending on the acceptance of the beam line transporting the beam to target ) , even the high intensity machine might be operational without flattop cavity .
the space saved this way could be used to improve the vacuum conditions by the installation of cryogenic pumps .
the beam loss by rest gas stripping has to be minimized when such cyclotrons are operated with high currents .
we discussed a long list of advantages of the new extraction mechanism , but the discussed method has it s price : @xmath0 has half the charge to mass ratio of protons and therefore on has to use double size and/or field strength to reach the same final proton energy .
the use of reverse bends has a comparable effect . a discussion , if and when the increase in size or field strength pays off by the mentioned advantages , is beyond the scope of this paper , but depends certainly on the purpose of the machine .
the geometry of cyclotrons with reverse bends has been analyzed and the resulting transfer matrices have been given .
we investigated some of the design options involving the use of reverse bends in combination with stripping extraction of @xmath0 .
we proved the principle feasability of variable energy / multiple beam extraction from cyclotrons with reverse bends and verified the analytical beam stability by a numerical calculation of the tunes .
we presented three potential applications for the described extraction mechanism , an isotope production cyclotron with simultaneous extraction at several energies between @xmath50 and @xmath51 , a medical cyclotron with a variable energy extraction in the range between @xmath96 and @xmath70 and a high intensity ring cyclotron with beam extraction at energies between @xmath84 and @xmath85 .
we thank nada fakhoury for her help in writing the mathematica^^ notebooks used for this work .
software has been written in `` c '' and been compiled with the gnu^^-c++ compiler on scientific linux .
the figures have been generated with the cern library ( paw ) and xfig .
the floor layout of cyclotrons is just a special case of the general problem of the calculation of floor layouts , which itself is a special case of the geometry of curves in the plane . in the general case ,
a planar smooth curve can be described by a `` state vector '' @xmath97 that contains the coordinates and the direction derivatives @xmath98 , where @xmath99 and @xmath100 are the cartesic coordinates of the orbit ( planar curve ) and @xmath101 and @xmath102 are the derivatives with respect to the pathlength @xmath103 . by definition one
has x^2+y^2=1 , so that one may also write @xmath104 with the direction angle @xmath28 of the orbit .
the state vector is a function of the pathlength @xmath103 of the orbit and the general evolution of this vector can be described by a differential equation of the form : & = & d**ds=*f*(1 ) * * + x + y + x + y + & = & 0&0&1&0 + 0&0&0&1 + 0&0&0&- + 0&0&&0 + x + y + x + y + , where @xmath105 is the local bending radius of the curve . in the hard edge approximation
, we assume that @xmath106 is piecewise constant . in this case
, a transfer matrix method can be used and the solution is given by a transfer ( or transport ) matrix @xmath107 : ( s)=*m*(s)**(0 ) , [ eq_solution ] where the matrix @xmath39 is the product of the transfer matrices for the individual segments : ( s)=_k=0^n-1*m*_k in hard edge approximation , there are basically two transfer matrices , the matrix @xmath108 for a drift of length @xmath109 and the matrix @xmath110 for a bending magnet for a bending radius @xmath111 and angle @xmath112 : _ d(l)&= & + * m*_b(,)&= & + the matrix powers of @xmath113 are readily computed : ^2&= & 0&0&0&-1 + 0&0&1&0 + 0&0&-1 ^ 2&0 + 0&0&0&-1 ^ 2 + + * f*^3&=&-1 ^ 2*f * + * f*^4&=&-1 ^ 2*f*^2 + from which one finds within a few steps _ b(,)&= & 1&0&s&-(1-c ) + 0&1&(1-c)&s + 0&0&c&-s + 0&0&s&c + , where @xmath114 , @xmath115 and @xmath116 . a reverse bend ( i.e. a bend into the opposite direction ) , is described by a negative radius _ and _ a negative angle , yielding a positve length @xmath117 . if @xmath118 , then the transfer matrix simplifies to the transfer matrix of a drift : _
d(l)&=&= 1&0&l&0 + 0&1&0&l + 0&0&1&0 + 0&0&0&1 + .
these two matrices are sufficient to compute the floor layout of most accelerator beamlines .
but they are also usefull for the geometrical analysis of separate sector cyclotrons with homogeneous field magnets in hard edge approximation as described above .
in addition to the above transfer matrices , we will use the familiar coordinate rotation matrix @xmath119 : _ rot()= & -&0&0 + & & 0&0 + 0&0&&- + 0&0 & & + . if we consider an accelerator with @xmath5 equal sectors ( or sections ) , then - given an arbitrary starting position @xmath120 - the position and direction change after one sector _
relative to some center point @xmath121 _ can be described by a rotation with an angle of @xmath122 .
hence we may write ( * * ( l)-**_c)=*m*_rot()(**(0)-**_c ) , so that by the use of eqn .
[ eq_solution ] one finds ( * m*_sec**(0)-**_c)&=&*m*_rot()(**(0)-**_c ) + ( * m*_sec-*m*_rot())**(0)&=&(*1 * -*m*_rot())**_c .
the coordinates of the accelerator center can be obtained by : _ c=(*1 * -*m*_rot())^-1(*m*_sec-*m*_rot())**(0 ) .
[ eq_center ] the matrix @xmath123 can be directly computed and is explicitely given by _
x()&=&12*m*_rot(/2-/2 ) + & = & 1&-&0&0 + & 1&0&0 + 0&0&1&- + 0&0&&1 + if one computes the center of motion of a bending magnet for an angle @xmath124 , the result is given by _
c&=&*m*_x()(*m*_b(,)-*m*_rot())**(0 ) + & = & 1&0&0&- + 0&1&&0 + 0&0&0&0 + 0&0&0&0 + * * ( 0 ) + & = & ( x(0)-y(0),y(0)+x(0),0,0)^t , [ eq_centerbend ] which is easy to verify .
the computation of the center coordinates is therefore straightforward - and yields a result even , if the matrix @xmath125 does not describe a `` valid '' sector .
such a non - valid situation is given , if the the `` velocity '' components of @xmath126 do not vanish , which happens , if the sum of the bending angles entering @xmath125 does not equal @xmath124 . in the following we use eqn .
[ eq_center ] to compute the starting conditions for a cyclotron centered at @xmath127 , i.e. the radius and direction of an _ equilibrium orbit_. if we let the orbit start at @xmath128 in arbitrary direction , i.e. we choose for instance @xmath129 , then the orbit with starting position @xmath130 is centered .
hence the starting position ( 0)=(*1 * -*m*_x()(*m*_sec-*m*_rot()))(0,0,0,1)^t .
[ eq_centerstart ] is centered .
the orbit still starts at an `` arbitrary '' angle @xmath131 , i.e. @xmath120 as given by eqn .
[ eq_centerstart ] can be written as ( 0)=(r , r,0,1)^t=(x_0,y_0,0,1)^t .
if one aims for a specific orientation of the orbit with respect to the floor coordinates - for instance on the x - axis - then one may use the rotation matrix with @xmath132 : ( 0)_rot(-_0)**(0)=(r,0,,)^t . the angular width of the magnet can then be calculated by computing the position angle of @xmath133 . and
@xmath134 are the spiral angles of the entrance and exit of the magnet . @xmath135 and @xmath136 are the angles between the sector entrance and exit and the orbit normal vector .
it is obvious from the drawing that @xmath137 .
[ fig_secgeom],width=321 ] if this method is applied to a cyclotron sector composed of a dipole with bending radius @xmath138 and bend angle @xmath139 and a drift of length @xmath109 as shown in fig .
[ fig_secgeom ] , then one obtains : & = & 2rl+ + r&=&. both conditions could also be derived from fig .
[ fig_secgeom ] .
the advantage of the algebraic method is , that it gives an _ algorithm _ at hand that allows to determine the essential geometric conditions directly from the parameters @xmath138 , @xmath5 and @xmath109 , without the need to analyze a `` hand - made '' drawing .
furthermore the algebraic algorithm enables to produce the drawing . in case of the simple situation scetched in fig .
[ fig_secgeom ] , the drawing might do as well .
but in the case of the medical cyclotron as shown in fig .
[ fig_medcyc ] , the symmetrie of the equilibrium orbit for a given energy is broken by the drift between the reverse bend ( valley ) and the next sector . in this case and in case of more complex configurations , the analysis of the layout by a handmade scetch becomes cumbersome due to the increasing number of angles and geometrical relations .
in fact , the geometry of the medical cyclotron presented above has been analyzed by a `` c''-program and a mathematica^^ notebook based on the above algebraic ansatz .
the main reason was the desire to create a map of the magnetic field in cylindrical coordinates for the numerical ( and hence more `` realistic '' ) computation of the tunes . in case of a cyclotron that is composed of @xmath5 sections each containing a sector magnet , a reverse bend and a drift
, it turns out , that the entrance and exit radius for a given energy are not equal . for a given radius of the grid
, we had to determine the energy ( i.e. the @xmath140-value ) of the equilibrium orbit entering the sector , a second @xmath140-value for the orbit existing the sector and a third one at the exit of the reverse bend .
this was done by an iterative numerical interval search .
e. pedroni , d. meer , c. bula , s. safai and s. zenklusen ; eur .
j. plus ( 2011 ) , 126:66 .
craddock and k.r .
symon ; rev . of accel .
vol . 1 ( 2008 ) , 65 - 97 , world scientific . dejan trbojevic ; rev .
vol . 2 ( 2009 ) , 229 - 251 , world scientific .
g. gulbekyan , o.n .
borisov , v.i .
kazacha ; proceedings of hiat2009 ; http://accelconf.web.cern.ch/accelconf/hiat2009/papers/d-02.pdf .
y. jongen _
et al _ ; nucl .
meth . a 624 ( 2010 ) ,
et al _ ; nucl .
meth . a 704 ( 2013 ) ,
l. calabretta _
et al _ ; nucl .
. meth . a 562 ( 2006 ) ,
1009 - 1012 .
borisov , g.g .
gulbekyan and d. solivajs ; proceedings of rupac xix , dubna 2004 ; http://accelconf.web.cern.ch/accelconf/r04/papers/thbp02.pdf .
d. solivajs _ et al _ ; j. of electr . engineering vol .
7 - 8 ( 2004 ) , 201 - 206 .
stefan k. zeisler and vinder jaggi ; nucl .
instr . meth .
a 590 ( 2008 ) , pp .
y. huang , a. kumar and s. oh ; proceedings of pac 1987 ; http://accelconf.web.cern.ch/accelconf/p87/pdf/pac1987_1881.pdf .
k. r. symon , d. w. kerst , l. w. jones , l. j. laslett , and k. m. terwilliger ; phys .
103 ( 1956 ) , pp .
1837 - 1859 j.i.m .
botman , m.k .
craddock and c.j .
kost ; 10th int .
east lansing 1984 , ed .
f. marti , ieee cat .
84ch1996 - 3 , pp .
craddock ; proceedings of pac 2009 ; http://accelconf.web.cern.ch/accelconf/pac2009/papers/fr5rep114.pdf .
g. schatz : orbit dynamics of isochronous cyclotrons ; nim vol .
72 ( 1969 ) , p.29 - 34 .
timmer , h. rcken , t. stephani , c. baumgarten and a. geisler ; nucl .
instrum . meth .
a 568 ( 2006 ) , pp .
532 - 536 .
u . klein _
et al _ ; nucl .
b 241 ( 2005 ) , p. 721
. y. jongen ; proc . of the int .
conf . on cycl .
2010 , lanzhou , china ; http://accelconf.web.cern.ch/accelconf/cyclotrons2010/papers/frm1cio01.pdf .
b. gottschalk ; http://arxiv.org/abs/1204.4470v2 .
b. gottschalk ; med .
37 ( 1 ) , 2010 , p. 352
m. j. van goethem , r van der meer , h.w .
reist and j.m . schippers ;
54 ( 2009 ) 5831 - 5846 .
gordon : computation of closed orbits and basic focusing properties for sector
focused cyclotrons and the design of `` cyclops '' : part .
1984 , vol . 16 ,
39 - 62 . m. seidel et .
al . ; proc of ipac 2010 , isbn 978 - 92 - 9083 - 352 - 9 , p. 1309 - 1313 .
alonso : high current @xmath0 cyclotrons for neutrino physics : the isodar and @xmath75 projects ; http://arxiv.org/abs/1210.3679 .
a. adelmann _
et al _ : cost - effective design options for isodar ; http://arxiv.org/abs/1210.4454 .
j.j yang _ et al _ ; nucl . instr . meth . a 704 ( 2013 ) , pp 84 - 91 .
c. rubbia _
et al _ ; cern - report cern / at/95 - 44 ; ab-atb-eet.web.cern.ch/ab-atb-eet/papers/ea/pdf/95-44.pdf . | we discuss the principle possibility of stripping extraction in combination with reverse bends in isochronous separate sector cyclotrons ( and/or ffags ) . if one uses reverse bends between the sectors ( instead of drifts ) and places stripper foils at the sector exit edges , the stripped beam has a reduced bending radius and it should be able to leave the cyclotron within the range of the reverse bend - even if the beam is stripped at less than full energy .
we are especially interested in @xmath0-cyclotrons , which allow to double the charge to mass ratio by stripping .
however the principle could be applied to other ions or ionized molecules as well . for the production of proton beams by stripping extraction of an @xmath0-beam
, we discuss possible designs for three types of machines : first a low - energy cyclotron for the simultaneous production of several beams at multiple energies - for instance 15 mev , 30 mev and 70 mev - thus allowing to have beam on several isotope production targets . in this case
it is desired to have a strong energy dependence of the direction of the extracted beam thus allowing to run multiple target stations simultaneously .
second we consider a fast variable energy proton machine for cancer therapy that should allow extraction ( of the complete beam ) at all energies in the range of about 70 mev to about 250 mev into the same beam line .
and third , we consider a high intensity high energy machine , where the main design goals are extraction with low losses , low activation of components and high reliability .
especially if such a machine is considered for an accelerator driven system ( ads ) , this extraction mechanism has severe advantages : beam trips by the failure of electrostatic elements could be avoided and the turn separation could be reduced , thus allowing to operate at lower main cavity voltages .
this would in turn reduce the number of rf - trips .
the price that has to be paid for these advantages is an increase in size and/or in field strength compared to proton machines with standard extraction at the final energy .
# 1(#1 ) # 1(#1 ) # 1(#1 ) # 1(#1 ) # 1(#1 ) # 1(#1 ) # 1(#1 ) # 1(#1 ) |
predictive analysis of networked data is a fast - growing research area whose application domains include document networks , online social networks , and biological networks . in this work
we view networked data as weighted graphs , and focus on the task of node classification in the transductive setting , i.e. , when the unlabeled graph is available beforehand .
standard transductive classification methods , such as label propagation @xcite , work by optimizing a cost or energy function defined on the graph , which includes the training information as labels assigned to training nodes .
although these methods perform well in practice , they are often computationally expensive , and have performance guarantees that require statistical assumptions on the selection of the training nodes .
a general approach to sidestep the above computational issues is to sparsify the graph to the largest possible extent , while retaining much of its spectral properties see , e.g. , @xcite .
inspired by @xcite , this paper reduces the problem of node classification from graphs to trees by extracting suitable _ spanning trees _ of the graph , which can be done quickly in many cases .
the advantage of performing this reduction is that node prediction is much easier on trees than on graphs .
this fact has recently led to the design of very scalable algorithms with nearly optimal performance guarantees in the online transductive model , which comes with no statistical assumptions . yet
, the current results in node classification on trees are not satisfactory .
the treeopt strategy of @xcite is optimal to within constant factors , but only on _ unweighted _ trees .
no equivalent optimality results are available for general weighted trees . to the best of our knowledge , the only other comparable result is wta by @xcite , which is optimal ( within log factors ) only on weighted lines .
in fact , wta can still be applied to weighted trees by exploiting an idea contained in @xcite .
this is based on linearizing the tree via a depth - first visit . since linearization loses most of the structural information of the tree , this approach yields suboptimal mistake bounds .
this theoretical drawback is also confirmed by empirical performance : throwing away the tree structure negatively affects the practical behavior of the algorithm on real - world weighted graphs .
the importance of weighted graphs , as opposed to unweighted ones , is suggested by many practical scenarios where the nodes carry more information than just labels , e.g. , vectors of feature values . a natural way of leveraging this side information is to set the weight on the edge linking two nodes to be some function of the similariy between the vectors associated with these nodes . in this work ,
we bridge the gap between the weighted and unweighted cases by proposing a new prediction strategy , called shazoo , achieving a mistake bound that depends on the detailed structure of the weighted tree .
we carry out the analysis using a notion of learning bias different from the one used in @xcite and more appropriate for weighted graphs .
more precisely , we measure the regularity of the unknown node labeling via the weighted cutsize induced by the labeling on the tree ( see section [ s : lower ] for a precise definition ) .
this replaces the unweighted cutsize that was used in the analysis of wta .
when the weighted cutsize is used , a cut edge violates this inductive bias in proportion to its weight .
this modified bias does not prevent a fair comparison between the old algorithms and the new one : shazoo specializes to treeopt in the unweighted case , and to wta when the input tree is a weighted line . by specializing shazoo s analysis to the unweighted case
we recover treeopt s optimal mistake bound .
when the input tree is a weighted line , we recover wta s mistake bound expressed through the weighted cutsize instead of the unweighted one .
the effectiveness of shazoo on any tree is guaranteed by a corresponding lower bound ( see section [ s : lower ] ) .
shazoo can be viewed as a common nontrivial generalization of both treeopt and wta . obtaining this generalization while retaining and extending the optimality properties of the two algorithms
is far from being trivial from a conceptual and technical standpoint . since shazoo works in the online transductive model
, it can easily be applied to the more standard train / test ( or `` batch '' ) transductive setting : one simply runs the algorithm on an arbitrary permutation of the training nodes , and obtains a predictive model for all test nodes .
however , the implementation might take advantage of knowing the set of training nodes beforehand .
for this reason , we present two implementations of shazoo , one for the online and one for the batch setting
. both implementations result in fast algorithms .
in particular , the batch one is linear in @xmath0 .
this is achieved by a fast algorithm for weighted cut minimization on trees , a procedure which lies at the heart of shazoo . finally , we test shazoo against wta , label propagation , and other competitors on real - world weighted graphs . in _
almost all _ cases ( as expected ) , we report improvements over wta due to the better sensitivity to the graph structure . in some cases , we see that shazoo even outperforms standard label propagation methods .
recall that label propagation has a running time per prediction which is proportional to @xmath1 , where @xmath2 is the graph edge set . on the contrary , shazoo can typically be run in _ constant _
amortized time per prediction by using wilson s algorithm for sampling random spanning trees @xcite . by disregarding edge weights in the initial sampling phase , this algorithm is able to draw a random ( unweighted ) spanning tree in time proportional to @xmath0 on most graphs .
our experiments reveal that using the edge weights only in the subsequent prediction phase causes in practice only a minor performance degradation .
let @xmath3 be an undirected and weighted tree with @xmath4 nodes , positive edge weights @xmath5 for @xmath6 , and @xmath7 for @xmath8 .
a binary labeling of @xmath9 is any assignment @xmath10 of binary labels to its nodes .
we use @xmath11 to denote the resulting labeled weighted tree .
the online learning protocol for predicting @xmath11 is defined as follows .
the learner is given @xmath9 while @xmath12 is kept hidden .
the nodes of @xmath9 are presented to the learner one by one , according to an unknown and arbitrary permutation @xmath13 of @xmath14 . at each time
step @xmath15 node @xmath16 is presented and the learner must issue a prediction @xmath17 for the label @xmath18 . then @xmath18 is revealed and the learner knows whether a mistake occurred .
the learner s goal is to minimize the total number of prediction mistakes .
following previous works @xcite , we measure the regularity of a labeling @xmath12 of @xmath9 in terms of @xmath19-edges , where a @xmath19-edge for @xmath11 is any @xmath6 such that @xmath20 .
the overall amount of irregularity in a labeled tree @xmath11 is the * weighted cutsize * @xmath21 , where @xmath22 is the subset of @xmath19-edges in the tree .
we use the weighted cutsize as our learning bias , that is , we want to design algorithms whose predictive performance scales with @xmath23 . unlike the @xmath19-edge count @xmath24 , which is a good measure of regularity for unweighted graphs , the weighted cutsize takes the edge weight @xmath25 into account
typically encodes the strength of the connection @xmath26 .
in fact , when the nodes of a graph host more information than just binary labels , e.g. , a vector of feature velues , then a reasonable choice is to set @xmath25 to be some ( decreasing ) function of the distance between the feature vectors sitting at the two nodes @xmath27 and @xmath28 .
see also remark [ r:2 ] . ] when measuring the irregularity of a @xmath19-edge @xmath26 . in the sequel , when we measure the distance between any pair of nodes @xmath27 and @xmath28 on the input tree @xmath9 we always use the resistance distance metric @xmath29 , that is , @xmath30 , where @xmath31 is the unique path connecting @xmath27 to @xmath28 .
graph cluster : maximal connected subgraph of @xmath32 ( or @xmath9 ) uniformly labeled .
`` closest '' : the default metric distance is the resistance distance metric on the tree .
hinge - node : labeled node or fork ( come in treeopt ) hinge - line : path connecting two hinge nodes such that no internal node is a hinge - node ( come in treeopt ) hinge - tree : each of the component of the forest obtained by removing all edges incident to the hinge nodes .
( quasi come in sel - pred . in sel - pred la foresta era ottenuta eliminando anche gli hinge - nodes ; definendo le cose in questo modo semplifichiamo le cose per questo paper ) @xmath33 : hinge - tree containing @xmath27 .
a connection node of a hinge - tree @xmath34 is any hinge node adjacent to one of node of @xmath34 .
( come in sel - pred ) fork : unlabeled node connected by ( at least ) three edge - disjoint path to ( at least ) three distinct labeled nodes ( come in sel - pred ) .
connection fork : connection node which is fork .
@xmath31 = path connecting node @xmath27 and @xmath28 .
@xmath19-free edge .
@xmath35 : predicted label of @xmath27 .
in this section we show that the weighted cutsize can be used as a lower bound on the number of online mistakes made by any algorithm on any tree . in order to do so ( and unlike previous papers on this specific subject see , e.g. , @xcite ) ,
we need to introduce a more refined notion of adversarial budget " .
given @xmath3 , let @xmath36 be the maximum number of edges of @xmath9 such that the sum of their weights does not exceed @xmath37 , @xmath38 we have the following simple lower bound ( all proofs are omitted from this extended abstract ) .
[ th : lb ] for any weighted tree @xmath3 there exists a randomized label assignment to @xmath14 such that any algorithm can be forced to make at least @xmath39 online mistakes in expectation , while @xmath40 .
specializing ( * ? ? ?
* theorem 1 ) to trees gives the lower bound @xmath41 under the constraint @xmath42 .
the main difference between the two bounds is the measure of label regularity being used : whereas theorem [ th : lb ] uses @xmath23 , which depends on the weights , ( * ? ? ?
* theorem 1 ) uses the weight - independent quantity @xmath43 .
this dependence of the lower bound on the edge weights is consistent with our learning bias , stating that a heavy @xmath19-edge violates the bias more than a light one .
since @xmath44 is nondecreasing , the lower bound implies a number of mistakes of at least @xmath45 .
note that @xmath46 for any labeled tree @xmath11 .
hence , whereas a constraint @xmath47 on @xmath43 implies forcing at least @xmath41 mistakes , a constraint @xmath37 on @xmath23 allows the adversary to force a potentially larger number of mistakes . in the next section
we describe an algorithm whose mistake bound nearly matches the above lower bound on any weighted tree when using @xmath48 as the measure of label regularity .
in this section we introduce the shazoo algorithm , and relate it to previously proposed methods for online prediction on unweighted trees ( treeopt from @xcite ) and weighted line graphs ( wta from @xcite ) .
in fact , shazoo is optimal on any weighted tree , and reduces to treeopt on unweighted trees and to wta on weighted line graphs .
since treeopt and wta are optimal on _ any _ unweighted tree and _ any _ weighted line graph , respectively , shazoo necessarily contains elements of both of these algorithms . in order to understand our algorithm
, we now define some relevant structures of the input tree @xmath9 .
see figure [ fig : hinge - trees_shazoo ] ( left ) for an example .
these structures evolve over time according to the set of observed labels .
first , we call * revealed * a node whose label has already been observed by the online learner ; otherwise , a node is * unrevealed*. a * fork * is any unrevealed node connected to at least three different revealed nodes by edge - disjoint paths . a * hinge node *
is either a revealed node or a fork .
a * hinge tree * is any component of the forest obtained by removing from @xmath9 all _ edges _ incident to hinge nodes ; hence any fork or labeled node forms a @xmath49-node hinge tree . when a hinge tree @xmath34 contains only one hinge node , a * connection node * for @xmath34 is the node contained in @xmath34 . in all other cases ,
we call a connection node for @xmath34 any node outside @xmath34 which is adjacent to a node in @xmath34
. a * connection fork * is a connection node which is also a fork .
finally , a * hinge line * is any path connecting two hinge nodes such that no internal node is a hinge node .
* left : * an input tree .
revealed nodes are dark grey , forks are doubly circled , and hinge lines have thick black edges .
the hinge trees not containing hinge nodes ( i.e. , the ones that are not singletons ) are enclosed by dotted lines . the dotted arrows point to the connection node(s ) of such hinge trees . *
middle : * the predictions of shazoo on the nodes of a hinge tree .
the numbers on the edges denote edge weights . at a given time @xmath50
, shazoo uses the value of @xmath51 on the two hinge nodes ( the doubly circled ones , which are also forks in this case ) , and is required to issue a prediction on node @xmath16 ( the black node in this figure ) .
since @xmath16 is between a positive @xmath51 hinge node and a negative @xmath51 hinge node , shazoo goes with the one which is closer in resistance distance , hence predicting @xmath52 . *
right : * a simple example where the mincut prediction strategy does not work well in the weighted case . in this example
, mincut mispredicts all labels , yet @xmath53 , and the ratio of @xmath23 to the total weight of all edges is about @xmath54 .
the labels to be predicted are presented according to the numbers on the left of each node .
edge weights are also displayed , where @xmath55 is a very small constant . ]
given an unrevealed node @xmath27 and a label value @xmath56 , the * cut function * @xmath57 is the value of the minimum weighted cutsize of @xmath9 over all labelings @xmath58 consistent with the labels seen so far and such that @xmath59 . define @xmath60 if @xmath27 is unrevealed , and @xmath61 , otherwise .
the algorithm s pseudocode is given in algorithm [ alg : shazoo ] . at time @xmath50 , in order to predict the label @xmath18 of node @xmath16 , shazoo calculates @xmath62 for all connection nodes @xmath27 of @xmath63 , where @xmath63 is the hinge tree containing @xmath16 .
then the algorithm predicts @xmath18 using the label of the connection node @xmath27 of @xmath63 which is closest to @xmath16 and such that @xmath64 ( recall from section [ sec : prel ] that all distances / lengths are measured using the resistance metric ) .
ties are broken arbitrarily .
if @xmath65 for all connection nodes @xmath27 in @xmath63 then shazoo predicts a default value ( @xmath66 in the pseudocode ) .
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * let @xmath67 be the set of the connection nodes @xmath27 of @xmath63 for which @xmath64 let @xmath28 be the node of @xmath67 closest to @xmath16 set @xmath68 set @xmath52 [ alg : shazoo ] * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * if @xmath16 is a fork ( which is also a hinge node ) , then @xmath69 . in this case
, @xmath16 is a connection node of @xmath63 , and obviously the one closest to itself .
hence , in this case shazoo predicts @xmath70 simply by @xmath71 .
see figure [ fig : hinge - trees_shazoo ] ( middle ) for an example .
on unweighted trees , computing @xmath62 for a connection node @xmath27 reduces to the fork label estimation procedure in ( * ? ? ?
* lemma 13 ) . on the other hand ,
predicting with the label of the connection node closest to @xmath16 in resistance distance is reminiscent of the nearest - neighbor prediction of wta on weighted line graphs @xcite .
in fact , as in wta , this enables to take advantage of labelings whose @xmath19-edges are light weighted .
an important limitation of wta is that this algorithm linearizes the input tree .
on the one hand , this greatly simplifies the analysis of nearest - neighbor prediction ; on the other hand , this prevents exploiting the structure of @xmath9 , thereby causing logaritmic slacks in the upper bound of wta .
the treeopt algorithm , instead , performs better when the unweighted input tree is very different from a line graph ( more precisely , when the input tree can not be decomposed into long edge - disjoint paths , e.g. , a star graph ) .
indeed , treeopt s upper bound does not suffer from logaritmic slacks , and is tight up to constant factors on any unweighted tree .
similar to treeopt , shazoo does not linearize the input tree and extends to the weighted case treeopt s superior performance , also confirmed by the experimental comparison reported in section [ s : exp ] . in figure
[ fig : hinge - trees_shazoo ] ( right ) we show an example that highlights the importance of using the @xmath51 function to compute the fork labels .
since @xmath51 predicts a fork @xmath16 with the label that minimizes the weighted cutsize of @xmath9 consistent with the revealed labels , one may wonder whether computing @xmath51 through mincut based on the number of @xmath19-edges ( rather than their weighted sum ) could be an effective prediction strategy .
figure [ fig : hinge - trees_shazoo ] ( right ) illustrates an example of a simple tree where such a @xmath51 mispredicts the labels of all nodes , when both @xmath23 and @xmath43 are small . [ r:1 ] we would like to stress that shazoo can also be used to predict the nodes of an arbitrary _ graph _ by first drawing a random spanning tree @xmath9 of the graph , and then predicting optimally on @xmath9 see , e.g. , @xcite .
the resulting mistake bound is simply the expected value of shazoo s mistake bound over the random draw of @xmath9 . by using a fast spanning tree sampler @xcite , the involved computational overhead amounts
to constant amortized time per node prediction on `` most '' graphs .
[ r:2 ] in certain real - world input graphs , the presence of an edge linking two nodes may also carry information about the extent to which the two nodes are _ dissimilar _ , rather than similar .
this information can be encoded by the sign of the weight , and the resulting network is called a _ signed
graph_. the regularity measure is naturally extended to signed graphs by counting the weight of _ frustrated edges _ ( e.g.,@xcite ) , where @xmath26 is frustrated if @xmath72 .
many of the existing algorithms for node classification @xcite can in principle be run on signed graphs .
however , the computational cost may not always be preserved .
for example ,
mincut @xcite is in general np - hard when the graph is signed @xcite .
since our algorithm sparsifies the graph using trees , it can be run efficiently even in the signed case .
we just need to re - define the @xmath51 function as @xmath73 , where @xmath74 is the minimum total weight of frustrated edges consistent with the labels seen so far . the argument contained in section [ s : impl ] for the positive edge weights ( see , e.g. , eq .
( [ eq : cut ] ) therein ) allows us to show that also this version of @xmath51 can be computed efficiently .
the prediction rule has to be re - defined as well : we count the parity of the number @xmath75 of negative - weighted edges along the path connecting @xmath16 to the closest node @xmath76 , i.e. , @xmath77 . in
@xcite the authors note that treeopt approximates a version space ( halving ) algorithm on the set of tree labelings .
interestingly , shazoo is also an approximation to a more general halving algorithm for weighted trees .
this generalized halving gives a weight to each labeling consistent with the labels seen so far and with the sign of @xmath78 for each fork @xmath79 .
these weighted labelings , which depend on the weights of the @xmath19-edges generated by each labeling , are used for computing the predictions .
one can show ( details omitted due to space limitations ) that this generalized halving algorithm has a mistake bound within a constant factor of shazoo s .
we now show that shazoo is nearly optimal on every weighted tree @xmath9 .
we obtain an upper bound in terms of @xmath23 and the structure of @xmath9 , nearly matching the lower bound of theorem [ th : lb ] .
we now give some auxiliary notation that is strictly needed for stating the mistake bound . given a labeled tree @xmath11 , a * cluster * is any maximal subtree whose nodes have the same label .
an * in - cluster line graph * is any line graph that is entirely contained in a single cluster .
finally , given a line graph @xmath80 , we set @xmath81 , i.e. , the ( resistance ) distance between its terminal nodes .
[ th : ub ] for any labeled and weighted tree @xmath82 , there exists a set @xmath83 of @xmath84 edge - disjoint in - cluster line graphs such that the number of mistakes made by shazoo is at most of the order of @xmath85 the above mistake bound depends on the tree structure through @xmath83 .
the sum contains @xmath84 terms , each one being at most logarithmic in the scale - free products @xmath86 .
the bound is governed by the same key quantity @xmath87 occurring in the lower bound of theorem [ th : lb ] .
however , theorem [ th : ub ] also shows that shazoo can take advantage of trees that can not be covered by long line graphs .
for example , if the input tree @xmath9 is a weighted line graph , then it is likely to contain long in - cluster lines .
hence , the factor multiplying @xmath87 may be of the order of @xmath88 .
if , instead , @xmath9 has constant diameter ( e.g. , a star graph ) , then the in - cluster lines can only contain a constant number of nodes , and the number of mistakes can never exceed @xmath84 . this is a log factor improvement over wta which , by its very nature , can not exploit the structure of the tree it operates on . ) and lower ( theorem [ th : lb ] ) bounds exists due to the extra factors depending on @xmath89 .
one way to get around this is to follow the analysis of wta in @xcite .
specifically , we can adapt here the more general analysis from that paper ( see lemma 2 therein ) that allows us to drop , for any integer @xmath47 , the resistance contribution of @xmath47 arbitrary non-@xmath19 edges of the line graphs in @xmath83 ( thereby reducing @xmath90 for any @xmath80 containing any of these edges ) at the cost of increasing the mistake bound by @xmath47 .
the details will be given in the full version of this paper . ] as for the implementation , we start by describing a method for calculating @xmath91 for any unlabeled node @xmath92 and label value @xmath93 .
let @xmath94 be the maximal subtree of @xmath9 rooted at @xmath92 , such that no internal node is revealed .
for any node @xmath27 of @xmath94 , let @xmath95 be the subtree of @xmath94 rooted at @xmath27 .
let @xmath96 be the minimum weighted cutsize of @xmath95 consistent with the revealed nodes and such that @xmath59 .
since @xmath97 , our goal is to compute @xmath98 .
it is easy to see by induction that the quantity @xmath99 can be recursively defined as follows , where @xmath100 is the set of all children of @xmath27 in @xmath94 , and @xmath101 if @xmath102 is revealed , and @xmath103 , otherwise : @xmath104 now , @xmath98 can be computed through a simple depth - first visit of @xmath94 . in all backtracking steps of this visit
the algorithm uses ( [ eq : cut ] ) to compute @xmath99 for each node @xmath27 , the values @xmath105 for all children @xmath28 of @xmath27 being calculated during the previous backtracking steps .
the total running time is therefore linear in the number of nodes of @xmath106 .
next , we describe the basic implementation of shazoo for the on - line setting .
a batch learning implementation will be given at the end of this section .
the online implementation is made up of three steps .
* 1 . find the hinge nodes of subtree * @xmath107 . recall that a hinge - node is either a fork or a revealed node .
observe that a fork is incident to at least three nodes lying on different hinge lines .
hence , in this step we perform a depth - first visit of @xmath107 , marking each node lying on a hinge line . in order to accomplish this task , it suffices to single out all forks marking each labeled node and , recursively , each parent of a marked node of @xmath107 . at the end of this process
we are able to single out the forks by counting the number of edges @xmath26 of each marked node @xmath27 such that @xmath28 has been marked , too .
the remaining hinge nodes are the leaves of @xmath107 whose labels have currently been revealed .
* 2 . compute * @xmath108 * for all connection forks of @xmath63*. from the previous step we can easily find the connection node(s ) of @xmath63 . then
, we simply exploit the above - described technique for computing the cut function , obtaining @xmath108 for all connection forks @xmath27 of @xmath63 . *
3 . propagate the labels of the nodes of @xmath109 ( only if @xmath16 is not a fork)*. we perform a visit of @xmath63 starting from every node @xmath110 . during these visits
, we mark each node @xmath28 of @xmath63 with the label of @xmath111 computed in the previous step , together with the length of @xmath112 , which is what we need for predicting any label of @xmath63 at the current time step . the overall running time is dominated by the first step and the calculation of @xmath62 .
hence the worst case running time is proportional to @xmath113 .
this quantity can be quadratic in @xmath0 , though this is rarely encountered in practice if the node presentation order is not adversarial .
for example , it is easy to show that in a line graph , if the node presentation order is random , then the total time is of the order of @xmath114 . for a star graph
the total time complexity is always linear in @xmath0 , even on adversarial orders . * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * let @xmath107 be the maximal subtree of @xmath9 containing node @xmath16 such that none of its internal node labels have been revealed before time @xmath50 . hence ,
the leaves of @xmath107 are the only nodes of @xmath107 whose label may be revealed before time @xmath50 .
we now consider tree @xmath107 as rooted at @xmath16 and we denote by @xmath115 the subtree of @xmath107 formed by @xmath27 and all its descendants .
notice that , by definition of @xmath115 , at time @xmath50 the algorithm does not know any label of its internal nodes .
let @xmath116 be the minimal cutsize of @xmath115 consistent with the labels of its leaves that have been revealed before time @xmath50 , conditioned on the the fact that label @xmath117 is revealed and it is equal to @xmath93 .
let now @xmath118 the set of all children of @xmath27 in @xmath107 .
it is easy to see that @xmath119 depends on @xmath120 and @xmath121 for each @xmath122 .
in fact , one can formally give a recursive definition of @xmath119 .
for each tree @xmath9 , @xmath123 , label @xmath124 and node @xmath27 of @xmath107 , @xmath116 satisfies @xmath125 when @xmath27 is a leaf of @xmath107 .
@xmath126 when @xmath27 is an internal node of @xmath107 , where @xmath101 if @xmath102 is already revealed , and @xmath127 otherwise . by induction on the height of @xmath27 .
term @xmath128 takes into account the contribution of edge @xmath26 to the cutsize .
let now @xmath129 be the set of labels that one can assign to node @xmath27 for minimizing the cutsize contribution of ( i ) the edge set of @xmath130 , together with ( ii ) the edge connecting @xmath27 with its parent @xmath131 in @xmath107 , if we set @xmath132 , i.e. @xmath133 let @xmath134 be the version space of @xmath115 consistent with all labels revealed at time @xmath50 , setting @xmath59 . as stated by the following lemma
, the cardinality of @xmath134 can be expressed in terms of @xmath135 for all @xmath122 .
for each tree @xmath9 , @xmath123 , label @xmath124 and unlabeled node @xmath27 of @xmath115 , @xmath136 satisfies @xmath137 when @xmath27 is a leaf of @xmath107 .
@xmath138 when @xmath27 is an internal node of @xmath107 .
by induction on the height of @xmath27 the basic implementation of this algorithm consists of a simple depth - first visit of tree @xmath115 . in all backtracking steps of this visit
the algorithm calculates , for each node @xmath27 , @xmath119 and @xmath136 using equation [ cutsize ] and [ card_s ] respectively .
in fact , equation [ cutsize ] defines @xmath119 in terms of @xmath139 for all children @xmath28 of @xmath27 , which have therefore been already calculated during the previous backtracking steps .
similarly , equation [ card_s ] defines @xmath136 in terms of @xmath119 and @xmath140 for all children @xmath28 of @xmath27 ( see figure [ f : halving_online ] ) .
label @xmath70 is predicted with @xmath141 if @xmath142 , or @xmath143 otherwise .
[ f : halving_online ] * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * + in many real - world scenarios , one is interested in the more standard problem of predicting the labels of a given subset of _ test _ nodes based on the available labels of another subset of _ training _ nodes .
building on the above on - line implementation , we now derive an implementation of shazoo for this train / test ( or `` batch learning '' ) setting .
we first show that computing @xmath144 and @xmath145 for all unlabeled nodes @xmath27 in @xmath9 takes @xmath146 time .
this allows us to compute @xmath147 for all forks @xmath92 in @xmath146 time , and then use the first and the third steps of the on - line implementation .
overall , we show that predicting _ all _ labels in the test set takes @xmath146 time
. consider tree @xmath148 as rooted at @xmath27 .
given any unlabeled node @xmath27 , we perform a visit of @xmath148 starting at @xmath27 . during the backtracking steps of this visit we use ( [ eq : cut ] ) to calculate @xmath149 for each node @xmath28 in @xmath148 and label @xmath124 .
observe now that for any pair @xmath150 of adjacent unlabeled nodes and any label @xmath124 , once we have obtained @xmath151 , @xmath152 and @xmath153 , we can compute @xmath154 in constant time , as @xmath155 .
in fact , all children of @xmath28 in @xmath148 are descendants of @xmath27 , while the children of @xmath27 in @xmath148 ( but @xmath28 ) are descendants of @xmath28 in @xmath156 .
shazoo computes @xmath151 , we can compute in constant time @xmath154 for all child nodes @xmath28 of @xmath27 in @xmath148 , and use this value for computing @xmath157 . generalizing this argument , it is easy to see that in the next phase we can compute @xmath158 in constant time for all nodes @xmath159 of @xmath148 such that for all ancestors @xmath160 of @xmath159 and all @xmath124 , the values of @xmath161 have previously been computed .
the time for computing @xmath162 for all nodes @xmath163 of @xmath148 and any label @xmath93 is therefore linear in the time of performing a breadth - first ( or depth - first ) visit of @xmath148 , i.e. , linear in the number of nodes of @xmath148 .
since each labeled node with degree @xmath29 is part of at most @xmath29 trees @xmath148 for some @xmath27 , we have that the total number of nodes of all distinct ( edge - disjoint ) trees @xmath148 across @xmath164 is linear in @xmath0 .
finally , we need to propagate the connection node labels of each hinge tree as in the third step of the online implementation . since also this last step takes linear time , we conclude that the total time for predicting all labels is linear in @xmath0 .
let now @xmath28 any child of @xmath27 in @xmath148 and let @xmath165 be the subtree obtained by eliminating @xmath166 and edge @xmath167 from @xmath165 , where @xmath159 is a child of @xmath28 .
let @xmath168 be the minimal cutsize of @xmath169 consistent with the labels of its leaves that have been revealed at time @xmath50 , setting @xmath170 , i.e. @xmath171 notice that @xmath168 can be easily calculated in constant time once we obtain @xmath172 , @xmath173 and @xmath174 .
since all children of @xmath28 in @xmath148 are descendants of @xmath27 , we can calculate @xmath175 for each label @xmath124 reusing the value of @xmath176 for all nodes @xmath163 of @xmath148 previously computed , in the following way : @xmath177 where in the second equality we used the fact that ( i ) @xmath178 and ( ii ) the children of @xmath27 in @xmath148 are descendants of @xmath28 in @xmath156 .
term @xmath179 takes into account the contribution of edge @xmath26 to the cutsize .
hence we can calculate the value of @xmath180 in constant time for all children @xmath28 of @xmath27 . applying the same formulas it is easy to verify that , in a subsequent phase
, we can compute @xmath158 in constant time for all nodes @xmath159 of @xmath148 such that @xmath181 .
more precisily we can calculate @xmath162 for each node @xmath163 in constant time once we obtain @xmath182 for all nodes @xmath183 such that @xmath184 .
the time for computing @xmath162 for all nodes @xmath163 of @xmath148 is therefore linear in the time of performing a breadth - first visit of @xmath148 , i.e. linear in the number of nodes of @xmath148 .
since each labeled node with degree @xmath29 is part of at most @xmath29 tree @xmath148 for some @xmath27 , we deduce that the sum of the number of nodes of all trees @xmath148 for @xmath164 is linear in @xmath0 .
hence , we can conclude that the total time for predicting all labels is linear in @xmath0 .
we tested our algorithm on a number of real - world weighted graphs from different domains ( character recognition , text categorization , bioinformatics , web spam detection ) against the following baselines : * online majority vote * ( omv ) .
this is an intuitive and fast algorithm for sequentially predicting the node labels is via a weighted majority vote over the labels of the adjacent nodes seen so far .
namely , omv predicts @xmath18 through the sign of @xmath185 , where @xmath163 ranges over @xmath186 such that @xmath187 . both the total time and space required by omv are @xmath188 .
* label propagation * ( labprop ) .
@xmath189 @xcite is a batch transductive learning method computed by solving a system of linear equations which requires total time of the order of @xmath190 .
this relatively high computational cost should be taken into account when comparing labprop to faster online algorithms .
recall that @xmath191 can be viewed as a fast `` online approximation '' to labprop . * weighted tree algorithm * ( wta ) .
as explained in the introductory section , wta can be viewed as a special case of shazoo .
when the input graph is not a line , wta turns it into a line by first extracting a spanning tree of the graph , and then linearizing it . the implementation described in @xcite runs in constant amortized time per prediction whenever the spanning tree sampler runs in time @xmath192 .
the graph perceptron algorithm @xcite is another readily available baseline .
this algorithm has been excluded from our comparison because it does not seem to be very competitive in terms of performance ( see , e.g. , @xcite ) , and is also computationally expensive . in our experiments , we combined shazoo and wta with spanning trees generated in different ways ( note that omv and labprop do not need to extract spanning trees from the input graph ) .
* random spanning tree * ( rst ) .
following ch . 4 of @xcite , we draw a weighted spanning tree with probability proportional to the product of its edge weights
. we also tested our algorithms combined with random spanning trees generated uniformly at random ignoring the edge weights ( i.e. , the weights were only used to compute predictions on the randomly generated tree ) we call these spanning trees nwrst ( no - weight @xmath193 ) . on most graphs
, this procedure can be run in time linear in the number of nodes @xcite .
hence , the combinations shazoo+nwrst and wta+nwrst run in @xmath194 time on most graphs . * minimum spanning tree * ( mst ) .
this is the spanning tree minimizing the sum of the resistors on its edges .
this tree best approximates the original graph in terms of the trace norm distance of the corresponding laplacian matrices . following @xcite , we also ran shazoo and wta using committees of spanning trees , and then aggregating predictions via a majority vote .
the resulting algorithms are denoted by @xmath159*shazoo and @xmath159*wta , where @xmath159 is the number of spanning trees in the aggregation .
we used either @xmath195 or @xmath196 , depending on the dataset size .
for our experiments , we used five datasets : rcv1 , usps , krogan , combined , and webspam .
webspam is a big dataset ( 110,900 nodes and 1,836,136 edges ) of inter - host links created for the web spam challenge 2008 @xcite .
krogan ( 2,169 nodes and 6,102 edges ) and combined ( 2,871 nodes and 6,407 edges ) are high - throughput protein - protein interaction networks of budding yeast taken from @xcite see @xcite for a more complete description .
finally , usps and rcv1 are graphs obtained from the usps handwritten characters dataset ( all ten categories ) and the first 10,000 documents in chronological order of reuters corpus vol . 1 ( the four most frequent categories ) , respectively . in both cases
, we used euclidean @xmath197-nearest neighbor to create edges , each weight @xmath198 being equal to @xmath199 .
we set @xmath200 , where @xmath201 is the average squared distance between @xmath27 and its @xmath197 nearest neighbours .
following previous experimental settings @xcite , we associate binary classification tasks with the five datasets / graphs via a standard one - vs - all reduction . each error rate is obtained by averaging over ten randomly chosen training sets ( and ten different trees in the case of rst and nwrst ) .
webspam is natively a binary classification problem , and we used the same train / test split provided with the dataset : 3,897 training nodes and 1,993 test nodes ( the remaining nodes being unlabeled ) . in the below table
, we show the macro - averaged classification error rates ( percentages ) achieved by the various algorithms on the first four datasets mentioned in the main text . for each dataset we trained ten times over a random subset of 5% , 10% and 25% of the total number of nodes and tested on the remaining ones .
in boldface are the lowest error rates on each column , excluding labprop which is used as a `` yardstick '' comparison .
standard deviations averaged over the binary problems are small : most of the times less than 0.5% . [ cols="<,^,^,^,^,^,^,^,^,^,^,^,^ " , ] our empirical results can be briefly summarized as follows : * 1 .
* without using committees , shazoo outperforms wta on all datasets , irrespective to the type of spanning tree being used . with committees ,
shazoo works better than wta almost always , although the gap between the two reduces .
the predictive performance of shazoo+mst is comparable to , and sometimes better than , that of labprop , though the latter algorithm is slower .
* @xmath159*shazoo , with @xmath202 ( or @xmath203 on webspam ) seems to be especially effective , outperforming labprop , with a small ( e.g. , 5% ) training set size .
nwrst does not offer the same theoretical guarantees as rst , but it is extremely fast to generate ( linear in @xmath0 on most graphs e.g. , @xcite ) , and in our experiments is only slightly inferior to rst . 10 n. alon , c. avin , m. kouck , g. kozma , z. lotker , and m.r .
many random walks are faster than one . in _ proc .
20th symp . on parallel algo . and architectures _ , pages 119128 .
springer , 2008 .
m. belkin , i. matveeva , and p. niyogi .
regularization and semi - supervised learning on large graphs . in _ proceedings of the 17th annual conference on learning theory _ , pages 624638 .
springer , 2004 .
y. bengio , o. delalleau , and n. le roux . label propagation and quadratic criterion . in _ semi - supervised learning _ , pages 193216 .
mit press , 2006 .
a. blum and s. chawla .
learning from labeled and unlabeled data using graph mincuts . in _ proceedings of the 18th international conference on machine learning_. morgan kaufmann , 2001 . n. cesa - bianchi , c. gentile , and f.vitale .
fast and optimal prediction of a labeled tree . in _ proceedings of the 22nd annual conference on learning theory _
, 2009 .
n. cesa - bianchi , c. gentile , f. vitale , and g. zappella .
random spanning trees and the prediction of weighted graphs . in _ proceedings of the 27th international conference on machine learning _
, 2010 .
n. cesa - bianchi , c. gentile , f. vitale , and g. zappella .
active learning on trees and graphs . proc . of the 23rd conference on learning theory ( colt 2010 ) . c. altafini g. iacono .
monotonicity , frustration , and ordered response : an analysis of the energy landscape of perturbed large - scale biological networks . , 4(83 ) , 2010 .
m. herbster and g. lever .
predicting the labelling of a graph via minimum @xmath131-seminorm interpolation . in _ proceedings of the 22nd annual conference on learning theory_. omnipress , 2009 .
m. herbster , g. lever , and m. pontil .
online prediction on large diameter graphs . in _ advances in neural information processing systems
22_. mit press , 2009 .
m. herbster , m. pontil , and s. rojas - galeano . fast prediction on a tree . in _ advances in neural information processing systems 22_. mit press , 2009 .
f. r. kschischang , b. j. frey , and h. a. loeliger .
factor graphs and the sum - product algorithm .
, 47(2):498519 , 2001 .
r. lyons and y. peres .
probability on trees and networks .
manuscript , 2008 .
s. t. mccormick , m. r. rao , and g. rinaldi .
easy and difficult objective functions for max cut . , 94(2 - 3):459466 , 2003 .
g. pandey , m. steinbach , r. gupta , t. garg , and v. kumar .
association analysis - based transformations for protein interaction networks : a function prediction case study . in _ proceedings of the 13th acm sigkdd international conference on knowledge discovery and data mining _ ,
pages 540549 .
acm press , 2007 . yahoo ! research and laboratory of web algorithmics university of milan
. web spam collection .
/ webspam / datasets/. d. a. spielman and n. srivastava .
graph sparsification by effective resistances . in _ proc .
of the 40th annual acm symposium on theory of computing ( stoc 2008)_. acm press , 2008 .
wilson . generating random spanning trees more quickly than the cover time . in
_ proceedings of the 28th acm symposium on the theory of computing _ , pages 296303 .
acm press , 1996 .
x. zhu , z. ghahramani , and j. lafferty .
semi - supervised learning using gaussian fields and harmonic functions . in _ proceedings of the 20th international conference on machine learning _ , 2003 .
pick any @xmath204 such that @xmath205 .
let @xmath206 be the forest obtained by removing from @xmath9 all edges in @xmath207 .
draw an independent random label for each of the @xmath208 components of @xmath206 and assign it to all nodes of that component .
then any online algorithm makes in expectation at least half mistake per component , which implies that the overall number of online mistakes is @xmath209 in expectation . on the other hand
, @xmath40 clearly holds by construction .
we first give additional definitions used in the analysis , then we present the main ideas , and finally we provide full details . recall that , given a labeled tree @xmath11 , a * cluster * is any maximal subtree whose nodes have the same label . let @xmath210 be the set of all clusters of @xmath9 . for any cluster @xmath211 ,
let @xmath212 be the subset of all nodes of @xmath213 on which shazoo makes a mistake .
let @xmath214 be the subtree of @xmath9 obtained by adding to @xmath213 all nodes that are adjacent to a node of @xmath213 .
note that all edges connecting a node of @xmath215 to a node of @xmath213 are @xmath19-edges .
let @xmath216 be the set of @xmath19-edges in @xmath214 and let @xmath217 .
let @xmath218 be the total weight of the edges in @xmath216 .
finally , recall the notation @xmath81 , where @xmath80 is any line graph .
recall that an * in - cluster line graph * is any line graph that is entirely contained in a single cluster .
the main idea used in the proof below is to bound @xmath219 for each @xmath211 in the following way .
we partition @xmath212 into @xmath220 groups , where @xmath221 .
then we find a set @xmath222 of edge - disjoint in - cluster line graphs , and create a bijection between lines in @xmath222 and groups in @xmath212 .
we prove that the cardinality of each group is at most @xmath223 , where @xmath224 is the associated line .
this shows that the subset @xmath225 of nodes in @xmath9 which are mispredicted by shazoo satisfies @xmath226 where @xmath227 .
then we show that @xmath228 by the very definition of @xmath44 , and using the bijection stated above , this implies @xmath229 thereby resulting in the mistake bound contained in theorem 2 .
according to shazoo prediction rule , when @xmath16 is not a fork and @xmath230 , the algorithm predicts @xmath18 using the label of any @xmath76 closest to @xmath16 . in this case , we call @xmath28 an * r - node * ( reference node ) for @xmath16 and the pair @xmath231 , where @xmath232 is the edge on the path between @xmath28 and @xmath16 , an * rn - direction * ( reference node direction )
. we use the shorthand notation @xmath233 to denote an r - node for @xmath27 . in the special case when all connection nodes @xmath27 of the hinge tree containing @xmath16 have @xmath65 ( i.e. , @xmath234 ) , and @xmath16 is not a fork , we call any closest connection node @xmath235 to @xmath16 an r - node for @xmath16 and we say that @xmath236 is a rn - direction for @xmath16 .
clearly , we may have more than one node of @xmath212 associated with the same rn - direction .
given any rn - direction @xmath231 , we call * r - line * ( reference line ) the line graph whose terminal nodes are @xmath28 and the first ( in chronological order ) node @xmath237 for which @xmath231 is a rn - direction , where @xmath232 lies on the path between @xmath235 and @xmath28 .. ] we denote such an r - line by @xmath238 . in the special case
where @xmath239 and @xmath240 we say that the r - line is associated with the @xmath19-edge of @xmath216 included in the line - graph .
in this case we denote such an r - line by @xmath241 , where @xmath242 .
figure [ fig : rn - direction ] gives a pictorial example of the above concepts .
we illustrate an example of r - node , rn - direction and r - line .
the numbers near the edge lines denote edge weights . in order to predict @xmath243 , shazoo uses the r - node @xmath244 and the rn - direction @xmath245 . after observing @xmath243 ,
the hinge line connecting @xmath244 with @xmath246 ( the thick black line ) is created , which is also an r - line , since at the beginning of step @xmath247 the algorithm used @xmath245 . in order to predict @xmath248 , we still use the r - node @xmath244 and the rn - direction @xmath245 . after the revelation of @xmath248 ,
node @xmath79 becomes a fork . ]
* @xmath249 is the set of all forks in @xmath212 .
* @xmath250 is the subset of @xmath212 containing the nodes @xmath27 whose reference node @xmath233 belongs to @xmath213 ( if @xmath27 is a fork , then @xmath251 ) . note that this set may have a nonempty intersection with the previous one
* @xmath252 is the subset of @xmath212 containing the nodes @xmath27 such that @xmath233 does not belong to @xmath213 .
* @xmath253 is the subset of all forks @xmath254 such that @xmath255 at some step @xmath50 . since we assume the cluster label is @xmath256 ( see below ) , and
since a fork @xmath257 is mistaken only if @xmath258 , we have @xmath259 .
* @xmath260 is the subset of all nodes in @xmath212 that , when revealed , create a fork that belongs to @xmath253 .
since at each time step at most one new fork can be created , a new fork @xmath28 is created when the number of edge - disjoint paths connecting @xmath28 to the labeled nodes increases .
this event occurs only when a new hinge line @xmath261 is created .
when this happens , the only node for which the number of edge - disjoint paths connecting it to labeled nodes gets increased is the terminal node @xmath28 of the newly created hinge line .
] we have @xmath262 .
the proof of the theorem relies on the following sequence of lemmas that show how to bound the number of mistakes made on a given cluster @xmath263 .
a major source of technical difficulties , that makes this analysis different and more complex than those of treeopt and wta , is that on a weighted tree the value of @xmath62 on forks @xmath27 can potentially change after each prediction .
for the sake of contradiction , assume @xmath267 .
let @xmath268 be the maximal subtree of @xmath9 rooted at @xmath79 such that no internal node of @xmath268 is revealed . now , consider the cut given by the edges of @xmath269 belonging to the hinge lines of @xmath268 .
this cut separates @xmath79 from any revealed node labeled with @xmath66 .
the size of this cut can not be larger than @xmath218 . by definition of @xmath270 , this implies @xmath271 .
however , also @xmath272 can not be larger than @xmath218 . because @xmath273 must hold independent of the set of nodes in @xmath274 that are revealed before time @xmath50 , this entails a contradiction .
let now @xmath275 be the restriction of @xmath44 on the subtree @xmath214 , and let @xmath276 be the set of all distinct rn - directions which the nodes of @xmath250 can be associated with .
the next lemmas are aimed at bounding @xmath277 and @xmath278 .
we first need to introduce the superset @xmath279 of @xmath276 .
then , we show that for any @xmath213 both @xmath280 and @xmath277 are linear in @xmath281 . in order to do so ,
we need to take into account the fact that the sign of @xmath51 for the forks in the cluster can change many times during the prediction process .
this can be done via lemma [ lm : delta ] , which shows that when all labels in @xmath215 are revealed then , for all fork @xmath282 , the value @xmath78 does not increase .
thus , we get the largest set @xmath276 when we assume that the nodes in @xmath215 are revealed before the nodes of @xmath213 . given any cluster @xmath213 , let @xmath283 be the order in which the nodes of @xmath214 are revealed .
let also @xmath284 be the permutation in which all nodes in @xmath213 are revealed in the same order as @xmath283 , and all nodes in @xmath215 are revealed at the beginning , in any order .
now , given any node revelation order @xmath283 , @xmath279 can be defined by describing the three types of steps involved in its incremental construction supposing @xmath284 was the actual node revelation order . 1 .
after the first @xmath285 steps , @xmath279 contains all node - edge pairs @xmath286 such that @xmath27 is a fork and @xmath26 is an edge laying on a hinge line of @xmath214 .
recall that no node in @xmath213 is revealed yet .
2 . for each step
@xmath287 when a new fork @xmath79 is created such that @xmath255 just after the revelation of @xmath18 , we add to @xmath279 the three node - edge pairs @xmath288 , where the @xmath289 are the edges contained in the three hinge lines terminating at @xmath79 .
3 . let @xmath163 be any step where : ( i ) a new hinge line @xmath290 is created , ( ii ) node @xmath291 is a fork , and ( iii ) @xmath292 at time @xmath293 . on each such step
we add @xmath294 to @xmath279 , for @xmath28 in @xmath290 .
it is easy to verify that , given any ordering @xmath283 for the node revelation in @xmath214 , we have @xmath295 .
in fact , given an rn - direction @xmath296 , if @xmath26 lies along one of the hinge lines that are present at time @xmath297 according to @xmath298 , then @xmath286 must be included in @xmath279 during one of the steps of type 2 above , otherwise @xmath286 will be included in @xmath279 during one of the steps of type 2 or type 3 .
assume nodes are revealed according to @xmath284 .
let @xmath301 be the subtree of @xmath214 made up of all nodes in @xmath214 that are included in any path connecting two nodes of @xmath215 . by their very definition ,
the forks at time @xmath302 are the nodes of @xmath303 having degree larger than two in subtree @xmath301 .
consider @xmath301 as rooted at an arbitrary node of @xmath215 .
the number of the leaves of @xmath301 is equal to @xmath304 .
this is in turn @xmath305 because @xmath306 now , in any tree , the sum of the degrees of nodes having degree larger than two can not is at most linear in the number of leaves .
hence , at time @xmath299 both the number of forks in @xmath213 and the cardinality of @xmath279 are @xmath307 .
[ incr_cutsize ] let @xmath50 be a step when a new hinge line @xmath310 is created such that @xmath311 .
if just after step @xmath50 we have @xmath312 , then @xmath313 , where @xmath314 is the lightest edge on @xmath310 .
since @xmath312 and @xmath310 is completely included in @xmath213 , we must have @xmath312 just before the revelation of @xmath18 .
this implies that the difference @xmath315 can not be smaller than the minimum cutsize that would be created on @xmath310 by assigning label @xmath66 to node @xmath316 .
[ dc_second ] assume nodes are revealed according to @xmath284 .
then the cardinality of @xmath253 and the total number of elements added to @xmath279 during the steps of type 2 above are both linear in @xmath281 .
let @xmath317 be the set of forks in @xmath274 such that @xmath318 at some time @xmath319 .
recall that , by definition , for each fork @xmath320 there exists a step @xmath321 such that @xmath255 .
hence , lemma [ lm : delta ] implies that , at the same step @xmath321 , for each fork @xmath320 we have @xmath318 .
since @xmath253 is included in @xmath317 , we can bound @xmath277 by @xmath322 , i.e. , by the number of forks @xmath264 such that @xmath323 , under the assumption that @xmath284 is the actual revelation order for the nodes in @xmath214 .
now , @xmath322 is bounded by the number of forks created in the first @xmath285 steps , which is equal to @xmath300 plus the number of forks @xmath79 created at some later step and such that @xmath255 right after their creation . since nodes in @xmath214 are revealed according to @xmath284 , the condition @xmath324 just after the creation of a fork @xmath79 implies that we will never have @xmath255 in later stages .
hence this fork @xmath79 belongs neither to @xmath317 nor to @xmath253 .
in order to conclude the proof , it suffices to bound from above the number of elements added to @xmath279 in the steps of type 2 above . from lemma [ incr_cutsize ]
, we can see that for each fork @xmath79 created at time @xmath50 such that @xmath255 just after the revelation of node @xmath16 , we must have @xmath325 , where @xmath314 is the lightest edge in @xmath261 .
hence , we can injectively associate each element of @xmath253 with an edge of @xmath326 , in such a way that the sum of the weights of these edges is bounded by @xmath218 . by definition of @xmath44
, we can therefore conclude that the total number of elements added to @xmath279 in the steps of type 2 is @xmath300 . with the following lemma we bound the number of nodes of @xmath327 associated with every rn - direction and
show that one can perform a transformation of the r - lines so as to make them edge - disjoint .
this transformation is crucial for finding the set @xmath83 appearing in the theorem statement .
observe that , by definition of r - line , we can not have two r - lines such that each of them includes only one terminal node of the other .
thus , let now @xmath328 be the forest where each node is associated with an r - line and where the parent - child relationship expresses that ( i ) the parent r - line contains a terminal node of the child r - line , together with ( ii ) the parent r - line and the child r - line are not edge - disjoint .
@xmath328 is , in fact , a forest of r - lines .
we now use @xmath329 for bounding the number of mistakes associated with a given rn - direction @xmath330 or with a given @xmath19-edge @xmath232 .
given any connected component @xmath331 of @xmath328 , let finally @xmath332 be the total number of nodes of @xmath327 associated with the rn - directions @xmath333 of all r - lines @xmath334 of @xmath331 .
* the number of nodes in @xmath327 associated with a given rn - direction @xmath335 is of the order of @xmath336 . *
the number of nodes in @xmath337 associated with a given @xmath19-edge @xmath338 is of the order of @xmath339 .
* let @xmath340 be the r - line associated with the root of any connected component @xmath331 of @xmath328 .
@xmath332 must be at most of the same order of @xmath341 where @xmath342 is a set of @xmath343 edge - disjoint line graphs completely contained in @xmath340 .
we will prove only ( i ) and ( iii ) , ( ii ) being similar to ( i ) .
let @xmath16 be a node in @xmath327 associated with a given rn - direction @xmath335 .
there are two possibilities : ( a ) @xmath16 is in @xmath238 or ( b ) the revelation of @xmath18 creates a fork @xmath79 in @xmath238 such that @xmath344 for all steps @xmath345 .
let now @xmath346 be the next node ( in chronological order ) of @xmath327 associated with @xmath335 .
the length of @xmath347 can not be smaller than the length of @xmath348 ( under condition ( a ) ) or smaller than the length of @xmath349 ( under condition ( b ) ) .
this clearly entails a dichotomic behaviour in the sequence of mistaken nodes in @xmath327 associated with @xmath335 .
let now @xmath131 be the node in @xmath238 which is farthest from @xmath28 such that the length of @xmath350 is not larger than @xmath23 . once a node in @xmath350 is revealed or becomes a fork @xmath79 satisfying @xmath344 for all steps @xmath345 , we have @xmath351 for all subsequent steps
( otherwise , this would contradict the fact that the total cutsize of @xmath9 is @xmath23 ) .
combined with the above sequential dichotomic behavior , this shows that the number of nodes of @xmath327 associated with a given rn - direction @xmath335 can be at most of the order of @xmath352 part ( iii ) of the statement can be now proved in the following way .
suppose now that an r - line @xmath353 , having @xmath28 and @xmath235 as terminal nodes , includes the terminal node @xmath354 of another r - line @xmath355 , having @xmath354 and @xmath356 as terminal nodes .
assume also that the two r - lines are not edge - disjoint .
if @xmath355 is partially included in @xmath353 , i.e. , if @xmath356 does not belong to @xmath353 , then @xmath355 can be broken into two sub - lines : the first one has @xmath354 and @xmath159 as terminal nodes , being @xmath159 the node in @xmath353 which is farthest from @xmath354 ; the second one has @xmath159 and @xmath356 as terminal nodes .
it is easy to see that @xmath353 must be created before @xmath355 and @xmath235 is the only node of the second sub - line that can be associated with the rn - direction @xmath357 .
this observation reduces the problem to considering that in @xmath331 each r - line that is not a root is completely included in its parent .
consider now the simplest case in which @xmath331 is formed by only two r - lines : the parent r - line @xmath359 , which completely contains the child r - line @xmath360 .
let @xmath163 be the step in which the first node @xmath160 of @xmath359 becomes a hinge node .
after step @xmath163 , @xmath359 can be vieved as broken in two edge - disjoint sublines having @xmath361 and @xmath362 as terminal node sets , where @xmath235 is one of the terminal of @xmath359 .
@xmath363 generalizing this argument for every component @xmath331 of @xmath328 , and using the above observation about the partially included r - lines , we can state that , for any component @xmath331 of @xmath328 , @xmath332 is of the order of @xmath364 where @xmath365 .
this entails that we can define @xmath366 as the union of @xmath367 and @xmath368 , which concludes the proof .
assume nodes are revealed according to @xmath284 , and let @xmath163 be any type-3 step when a new element is added to @xmath279 .
there are two cases : ( a ) @xmath292 at time @xmath163 or ( b ) @xmath369 at time @xmath163 .
case ( a ) .
lemma [ incr_cutsize ] combined with the fact that all hinge - lines created are edge - disjoint , ensures that we can injectively associate each of these added elements with an edge of @xmath326 in such a way that the total weight of these edges is bounded by @xmath218 .
this in turn implies that the total number of elements added to @xmath326 is @xmath307 .
case ( b ) . since we assumed
that nodes are revealed according to @xmath284 , we have that @xmath370 is positive for all steps @xmath371 .
hence we have that case ( b ) can occur only once for each of such forks @xmath291 .
since this kind of fork belongs to @xmath253 , we can use lemma [ dc_second ] and conclude that ( b ) can occur at most @xmath372 times . theorem 2
let @xmath375 be the union of @xmath328 over @xmath376 .
using lemma [ lm : number_rnd ] we deduce @xmath377 , where the term @xmath378 takes into account that at most one r - line of @xmath328 may be associated with each @xmath19-edge of @xmath214 .
let now @xmath382 be the set of components of @xmath375 .
given any tree @xmath383 , let @xmath384 be the r - line root of @xmath331 .
recall that , by part ( iii ) of lemma [ lm : m_l ] for any tree @xmath383 we can find a set @xmath385 of @xmath343 edge - disjoint line graphs all included in @xmath384 such that @xmath332 is of the order of @xmath386 .
let now @xmath387 be equal to @xmath388 .
thus we have @xmath389 observe that @xmath387 is not an edge disjoint set of line graphs included in @xmath9 only because each @xmath19-edge may belong to two different lines of @xmath387 . by definition of @xmath390 , for any line graphs @xmath80 and @xmath391 , where @xmath391 is obtained from @xmath80 by removing one of the two terminal nodes and the edge incident to it , we have @xmath392 .
if , for each @xmath19-edge shared by two line graphs of @xmath387 , we shorten the two line graphs so as no one of them includes the @xmath19-edge , we obtain a new set of edge - disjoint line graphs @xmath83 such that @xmath393 .
hence , we finally obtain @xmath394 , where in the last equality we used the fact that @xmath395 for all line graphs @xmath391 . | predicting the nodes of a given graph is a fascinating theoretical problem with applications in several domains .
since graph sparsification via spanning trees retains enough information while making the task much easier , trees are an important special case of this problem . although it is known how to predict the nodes of an unweighted tree in a nearly optimal way , in the weighted case a fully satisfactory algorithm is not available yet .
we fill this hole and introduce an efficient node predictor , shazoo , which is nearly optimal on any weighted tree . moreover ,
we show that shazoo can be viewed as a common nontrivial generalization of both previous approaches for unweighted trees and weighted lines .
experiments on real - world datasets confirm that shazoo performs well in that it fully exploits the structure of the input tree , and gets very close to ( and sometimes better than ) less scalable energy minimization methods . |
toric residues are fundamental invariants of sparse polynomial systems .
they were first studied by cox @xcite who defined the residue of @xmath0 sections of an ample line bundle on a toric variety @xmath3 .
the definition was extended by cattani , cox , and dickenstein to sections of @xmath0 arbitrary line bundles @xcite .
there are numerous applications to sparse resultants and resultant or subresultant complexes @xcite , mixed hodge structures @xcite , and mirror symmetry @xcite .
the related notion of global residue in the torus , a sum of grothendiek local residues , was studied by gelfond , khovanskii , and soprounov @xcite .
cattani , cox and dickenstein @xcite showed that the global residue could always be computed as an instance of the toric residue .
applications of the toric and global residue include gkz hypergeometric systems @xcite and computations on sparse polynomial systems such as counting the number of real roots and computing elementary symmetric functions on the roots @xcite .
given @xmath0 arbitrary sparse laurent polynomials @xmath4 in @xmath1 affine variables , let @xmath5 be their corresponding newton polytopes .
the minkowski sum @xmath6 determines a toric variety @xmath3 , and each @xmath7 corresponds to a semi - ample divisor class . in the homogeneous coordinate ring @xmath8 of @xmath3
each @xmath9 can be homogenized to a polynomial @xmath10 of degree @xmath11 corresponding to the divisor class of @xmath7 . the toric residue @xmath12 is a linear function on homogeneous polynomials of a certain critical degree corresponding to the interior of @xmath13 which vanishes on the ideal of the @xmath10 . in many cases of interest , for example when all of the @xmath7 are full dimensional , the ideal of the @xmath10 has codimension @xmath14 in the critical degree . hence knowing a single element of non - zero residue will allow a full computation of the residue map .
more generally , we show in section [ s : essential ] that there is an element of non - zero residue whenever the polytopes form an essential family .
the goal of this paper is a general framework for the construction of specific elements whose residue we can compute .
the construction depends only on the combinatorics and affine geometry of the polytopes @xmath7 .
[ t : main1 ] let @xmath3 be a complete toric variety of dimension @xmath1 .
fix @xmath0 semi - ample degrees @xmath15 on @xmath3 and let @xmath16 be their polytopes .
let @xmath17 be a collection of partitions of the lattice points of the @xmath7 such that 1 .
for any lattice point @xmath18 , at least one vertex of the minimal face of @xmath7 containing @xmath19 lies in @xmath20 , 2 . for any permutation @xmath21 of @xmath22 :
@xmath23 given a collection of laurent polynomials @xmath24 supported on @xmath16 @xmath25 define polynomials @xmath26 then @xmath27 is a laurent polynomial supported on @xmath28 .
the toric residue @xmath29 of the corresponding homogeneous polynomial @xmath30 of critical degree for the homogenized @xmath31 is an integer that depends only on the combinatorics of the @xmath7 and the partitions of their lattice points . using this theorem we are able to find an element of residue @xmath32 , i.e. find an appropriate collection of partitions , in two important cases .
the first is when @xmath7 share a complete flag of faces .
this will generalize earlier results of dandrea and khetan when all of the @xmath11 were ample degrees .
the second application is a complete analysis when @xmath2 .
we show that , except for one degenerate family of supports , we can always find a collection of partitions yielding an element of residue @xmath33 .
the proof of the theorem makes use of some very elegant combinatorics . starting with a partition of the lattice points
we will show that there are induced colorings of the faces of the polytope @xmath34 .
moreover , the matrix will yield a canonical coloring of the facets of the barycentric refinement of @xmath13 .
such a facet coloring will allow us to reduce the computation to that of the residue of a monomial with respect to a monomial ideal . by an earlier theorem of soprounov @xcite ,
the residue is the _ combinatorial degree _ of the coloring which can be computed by counting the number of flags of certain colors .
the paper is organized as follows .
section [ s : prelim ] provides the definitions of the toric residue and some basic properties .
section [ s : essential ] proves the existence of elements of non - zero residue if and only if the polytopes are essential .
section [ s : cdeg ] introduces facet colorings of polytopes and their connection to the toric residue of monomials . the residue for general polynomials
is reduced to the monomial case via the global transformation law .
section [ s : partition ] and section [ s : canoncolor ] discuss the relationships between partitions , colorings , and residue matrices used to complete the proof of theorem [ t : main1 ] .
section [ s : locallyunmixed ] uses the previous results to give an explicit element of residue 1 when the polytopes @xmath7 share a complete flag of faces .
section [ s : dim2 ] is a complete analysis when @xmath3 is of dimension 2 .
finally , section [ s : conclusion ] discusses progress in dimensions three and higher .
we begin by setting up the notation and reviewing some basic definitions and facts about toric varieties and toric residues . for details and proofs
we refer the reader to @xcite .
consider an @xmath1-dimensional complete toric variety @xmath3 determined by a rational complete fan @xmath35 .
let @xmath36 denote the set of 1-dimensional cones ( rays ) of @xmath37 .
each ray @xmath38 determines a @xmath39-invariant irreducible divisor @xmath40 on @xmath3 .
as introduced by cox in @xcite the variety @xmath3 has the homogeneous coordinate ring @xmath41 $ ] graded by the chow group @xmath42 so that a monomial @xmath43 has degree @xmath44\in a_{n-1}(x).\ ] ] denote by @xmath45 the graded piece of @xmath8 consisting of all polynomials of degree @xmath46 .
let @xmath47 be a representative of @xmath46 .
it defines a continuous piecewise linear function @xmath48 on the support @xmath49 such that @xmath50 for all @xmath38 , where @xmath51 denotes the primitive generator of @xmath52 ( see ( * ? ? ?
* section 3.3 ) ) .
it also determines a convex polytope @xmath53 to every lattice point @xmath19 of @xmath54 we can assign a monomial @xmath55 in @xmath8 of degree @xmath56 : @xmath57 one can check that this map is a bijection . furthermore , given a laurent polynomial @xmath58 supported in @xmath54 its _ @xmath54-homogenization _ is the homogeneous polynomial @xmath59 notice that if @xmath60 is supported in the interior of @xmath54 then the @xmath54-homogenization is divisible by the product of all the variables @xmath61 , @xmath38 .
it is easy to see that if @xmath62 and @xmath63 are linearly equivalent then @xmath64 is a linear function , and @xmath54 and @xmath65 are the same up to a translation .
therefore , @xmath54-homogenization is independent of the choice of the representative @xmath62 of the divisor class @xmath56 . in
what follows the _ polytope of @xmath56 _ will mean the polytope of any representative of @xmath56 and will be denoted by @xmath66 .
recall the construction of the euler form @xmath67 from @xcite .
let @xmath68 be a basis for @xmath69 and for every subset @xmath70 of size @xmath1 denote @xmath71 then the _ euler form _ on @xmath3 is the sum over all size @xmath1 subsets @xmath70 : @xmath72 now we recall the definition of the toric residue @xcite .
consider @xmath0 homogeneous polynomials @xmath73 , for @xmath74 .
their critical degree is defined to be @xmath75 then for every polynomial @xmath30 of degree @xmath76 consider a meromorphic @xmath1-form on @xmath3 : @xmath77 where @xmath67 is the euler form .
we use @xmath78 to denote the list @xmath79 .
suppose that the @xmath10 do not vanish simultaneously on @xmath3 .
then @xmath3 has an open cover @xmath80 by the @xmath0 sets @xmath81 and @xmath82 defines a ech cohomology class @xmath83\in h^n(x,\widehat\omega_x^n)$ ] relative to the cover @xmath80 . here
@xmath84 denotes the sheaf of zariski @xmath1-forms on @xmath3 .
one can check that the class @xmath83 $ ] is alternating in the order of the @xmath10 and is zero if @xmath30 belongs to the ideal of @xmath31 .
therefore , @xmath83 $ ] depends on the equivalence class of @xmath30 modulo the ideal @xmath85 . the _ toric residue map _
@xmath86 is given by @xmath87)$ ] , where @xmath88 is the trace map on @xmath3 .
when there is no danger of confusion we will write @xmath29 instead of @xmath89 .
let @xmath3 be a complete @xmath1-dimensional toric variety defined by a complete fan @xmath37 in @xmath90 . recall that a @xmath39-cartier divisor @xmath62 on @xmath3 is called _ semi - ample _ if the corresponding line bundle @xmath91 is generated by global sections .
equivalently , @xmath62 is semi - ample if and only if the corresponding piecewise linear function @xmath48 is convex ( * ? ? ?
* section 3.4 ) . consider the ( generalized ) normal fan @xmath92 of the polytope @xmath54 of @xmath62 , i.e. a complete fan whose cones are @xmath93 for every face @xmath94 of @xmath54 .
it follows that if @xmath62 is semi - ample then @xmath37 refines @xmath92 .
indeed , by the convexity of @xmath48 for any maximal cone @xmath95 the restriction of @xmath48 to @xmath96 defines a a vertex @xmath19 of @xmath54 .
then @xmath97 , @xmath98 .
we will say that a degree @xmath99\in a_{n-1}(x)$ ] is _ semi - ample _ if @xmath62 is semi - ample . consider a collection of @xmath0 semi - ample degrees @xmath15 on @xmath3 .
let @xmath16 be their polytopes ( defined up to translations ) and @xmath100 the normal fans of the polytopes . by above @xmath37 refines each @xmath101 and , thus , refines the minimal common refinement of the @xmath101 , which is the normal fan @xmath102 of the minkowski sum @xmath103 by ( * ?
* chapter 5 , theorem 4.8 ) .
now let @xmath104 be a birational morphism defined by a refinement @xmath105 .
if @xmath62 is a @xmath39-cartier divisor on @xmath3 then the pull - back @xmath106 has the same piecewise linear function @xmath48 and the same polytope @xmath54 .
it follows that @xmath107 is semi - ample on @xmath108 if @xmath56 is semi - ample on @xmath3 .
also if @xmath78 is a homogeneous polynomial in @xmath45 and @xmath60 the corresponding laurent polynomial supported in @xmath66 then the pull - back @xmath109 is the @xmath66-homogenization of @xmath60 in the homogeneous coordinate ring @xmath110 of @xmath108 , and hence @xmath111 .
next we will see how the toric residue @xmath112 behaves under the birational morphism @xmath104 .
[ p : birational ] let @xmath3 be a complete @xmath1-dimensional toric variety defined by a complete fan @xmath37 .
let @xmath104 be a birational morphism induced by a refinement @xmath105 .
suppose @xmath15 are semi - ample degrees with polytopes @xmath16 and consider @xmath0 polynomials @xmath73 not vanishing simultaneously on @xmath3 .
then the polynomials @xmath113 do not vanish simultaneously on @xmath108 .
furthermore , let @xmath114 be any laurent polynomial supported in the interior of @xmath115 , and @xmath116 ( resp .
@xmath117 ) be the @xmath13-homogenization of @xmath114 in @xmath8 ( resp .
@xmath118 ) . then the homogeneous polynomials @xmath119 and @xmath120 are of critical degree for the @xmath10 and the @xmath121 , respectively , and satisfy @xmath122 here @xmath123 denotes the product of the homogeneous variables @xmath124 .
first the sets @xmath125 form a covering of @xmath108 since it is the pull - back of the covering @xmath80 of @xmath3 . in particular ,
the @xmath121 do not vanish simultaneously on @xmath108 .
now let @xmath67 and @xmath126 be the euler forms on @xmath3 and @xmath108 , respectively .
we have @xmath127 since both are rational extensions of the @xmath39-invariant regular @xmath1-form @xmath128 on the torus , where the @xmath129 are affine coordinates .
therefore @xmath130 since @xmath131 both @xmath82 and @xmath132 have the same toric residue .
the proposition follows .
[ d : essential ] a collection of polytopes @xmath5 is said to be _ essential _ if for every @xmath133 the dimension of the polytope @xmath134 is at least @xmath135 . given a toric variety @xmath3 of dimension @xmath1 ,
a collection of semi - ample degrees @xmath136 is called _ essential _ if the corresponding polytopes @xmath5 are essential .
the goal of this section is to prove that the toric residue is not identically zero if and only if the degrees @xmath11 are essential .
[ t : nonzero ] consider degrees @xmath136 on a complete toric variety @xmath3 .
the toric residue with respect to polynomials @xmath137 , viewed as a rational function in the coefficients of the @xmath10 , is identically zero if and only if the @xmath11 are not essential . for essential @xmath11
there is a polynomial @xmath30 of critical degree and homogeneous of degree @xmath14 in the coefficients of each @xmath10 such that @xmath138 .
the first implication is that for the non - essential degrees the toric residue is identically 0 . by proposition [ p : birational ] , we can refine @xmath3 to a simplicial variety without changing the toric residue .
so assume @xmath3 is simplicial .
in this case the toric residue @xmath29 is the sum of the grothendieck local residues of any @xmath139 with respect to the common zeros of the remaining @xmath10 ( * ? ? ? * theorem 0.4 ) .
suppose there exists a proper subset @xmath140 such that @xmath141 has dimension less than @xmath135 .
let @xmath142 be the toric variety corresponding to @xmath143 , and @xmath144 the morphism defined by the natural map of fans @xmath145 .
the polynomial @xmath10 for @xmath146 is the pullback of a polynomial of semi - ample degree on @xmath142 with polytope @xmath7 .
clearly , generic polynomials supported on the @xmath7 , @xmath146 , do not have a common zero on @xmath142 since @xmath147 .
thus the corresponding @xmath148 do not have a common zero on @xmath3 for generic coefficients .
extend @xmath140 to a subset of size @xmath1 , without loss of generality we take it to be @xmath149 . for generic coefficients
@xmath150 do not have a common root .
in particular there are no local residues in the sum .
so for generic coefficients the toric residue is @xmath151 .
since , @xmath12 is a rational function of the coefficients of the @xmath10 it must be identically zero . for the converse ,
the main tool is the following dual koszul complex of sheaves with respect to @xmath152 which appears in numerous places including @xcite , @xcite , and @xcite . given a subset @xmath153 let @xmath154 .
we have an exact sequence of sheaves @xmath155 where as before @xmath76 is the critical degree for @xmath78 and @xmath156 .
one can take the ech cohomology double complex and then pass to a spectral sequence .
the @xmath157 terms of this spectral sequence are : @xmath158 because the @xmath11 are essential , a result of @xcite gives us : @xmath159 as a consequence there is a unique top differential @xmath160 which must be an isomorphism since the spectral sequence is exact . moreover , @xmath161 and @xmath162 is a quotient of @xmath163 .
so we have an induced map @xmath164 which is the composition of the projection onto @xmath162 and the inverse of the isomorphism @xmath165 .
we also have an isomorphism @xmath166 sending a local section @xmath167 to @xmath168 where @xmath67 is the euler form .
finally there is the trace isomorphism @xmath169 .
composing all of these maps we obtain a map @xmath170 .
the maps are illustrated via the diagram below : s _
= e_1^n+1,0 & & & & & & + & & & & & & + e_n+1^n+1 , 0 & ^d_n+1 _ & h^n(x , ( -_0 ) ) & & h^n(x , ^n ) & ^_x & c. we will prove that this composition is precisely the toric residue map . in that case if we started with a differential form @xmath171 such that @xmath172 , it would correspond to an element of @xmath173 which is mapped to an element @xmath174 .
let @xmath30 be any element of @xmath163 lifting @xmath175 . from the above constructions it would follow that @xmath176 and the toric residue is not identically zero as desired .
moreover , by a theorem of weyman ( * ? ? ?
* chapter 3 , theorem 4.11 ) , the differential @xmath165 , and therefore the element @xmath30 above , can be lifted up to a ( non - unique ) map @xmath177 which is polynomial of degree @xmath14 in the coefficients of each @xmath10 . to prove that the residue map coincides with the one constructed above we compute the cohomology terms using the ech resolutions given by the open cover @xmath81 .
more generally , given @xmath178 define @xmath179 . in this way we have the @xmath180 terms of our spectral sequence @xmath181 where @xmath182 denotes the complement of @xmath140 . in terms of the cover , given a polynomial @xmath183 we have @xmath184 the residue map is defined to be the trace of the cohomology class of this latter element ( after multiplying by the euler form ) .
so it is enough to show that @xmath185 ) = [ h ] \in e_{n+1}^{n+1,0}$ ] . to compute this differential we start with @xmath186 and map it via @xmath187 to @xmath188
this can be lifted via the ech differential @xmath189 to an element of @xmath190 which is further mapped to @xmath191 and lifted to @xmath192 and so on . at the end
we obtain an element of @xmath193 which is mapped via @xmath187 to @xmath194 .
let @xmath195 be the basis of @xmath196 and @xmath197 .
we have the following lemma .
[ l : lift ] in the above mapping and lifting process , a valid choice for the element in @xmath198 is @xmath199 .
the base case @xmath200 is our starting element . for the inductive step
we need to show that @xmath201 .
however , by the definitions of the koszul and ech morphisms it is easy to see that both of the above elements are : @xmath202 where @xmath203 . therefore we get the element @xmath204 .
the final koszul differential is multiplication by @xmath10 in each factor so we are left with @xmath205 which corresponds to the global section @xmath206 .
so @xmath30 is a valid lifting of the image of the class of @xmath207 under @xmath165 completing the proof .
the theorem from the previous section guaranteed the existence of an element of toric residue one but was completely nonconstructive .
this section and the next one provide the framework for an explicit combinatorial construction of such elements . here
we recall the definition of the facet coloring of a polytope and the relation between the combinatorial degree of a facet coloring and the toric residue for monomial ideals .
we will also obtain the generalized global transformation law that allows us to reduce the computation of the toric residue for semi - ample degrees to the monomial case .
consider an @xmath1-dimensional polytope @xmath13 in @xmath90 .
we let @xmath208 denote the boundary of @xmath13 and @xmath209 the partially ordered set ( poset ) by inclusion of all proper faces of @xmath13 .
we also let @xmath210}$ ] denote the set of all subsets of @xmath211=\{0,\dots , n\}$ ]
. it will be convenient for us to equip @xmath210}$ ] with the inverse partial order @xmath212 , i.e. @xmath213 if and only if @xmath214 for @xmath215}$ ] .
a map of posets @xmath216},<)$ ] is called a _ coloring of @xmath13 into @xmath0 colors _ ( or simply _ coloring _ ) . the image @xmath217 is called the _ set of colors _ of a face @xmath218 .
we will also say that _
@xmath94 is colored by @xmath217_. a coloring is called _ simplicial _ if every face @xmath218 is colored by a non - empty proper subset of @xmath211 $ ] . the poset @xmath219},<)$ ] can be identified with the poset of faces of the standard @xmath1-simplex : @xmath220 indeed , each non - empty proper subset @xmath221 $ ] defines the codimension @xmath222 face of @xmath223 : @xmath224 therefore , any simplicial coloring is , in fact , a map of posets @xmath225 .
fix orientations of @xmath13 and @xmath223 . given a simplicial coloring @xmath226 consider a continuous piecewise linear map @xmath227 such that @xmath228 for any @xmath218 .
one can show that such a map @xmath229 always exists and the topological degree @xmath230 does not depend on the choice of @xmath229 ( see @xcite ) .
we call it the _ combinatorial degree _
@xmath231 of the simplicial coloring @xmath226 .
the combinatorial degree is alternating in the ordering of the elements of @xmath211 $ ] as every such ordering defines an orientation of the corresponding simplex @xmath223 .
we have the following property of the combinatorial degree .
let @xmath226 , @xmath232 be two simplicial colorings of @xmath13 .
we say that @xmath232 _ refines _
@xmath226 if @xmath233 for any @xmath218 .
[ p : refinement]@xcite let @xmath226 , @xmath232 be two simplicial colorings of @xmath13 . if @xmath232 refines @xmath226 then @xmath234 .
the combinatorial degree can be computed explicitly as a signed number of certain complete flags of faces of @xmath13 . to state the precise formula we will need the following definition .
consider a complete flag @xmath78 of faces of @xmath13 @xmath235 for every @xmath236 choose a vector @xmath237 that begins at @xmath238 and points strictly inside @xmath239 .
define the _ sign _ of the flag to be @xmath240 if @xmath68 gives a positive oriented frame for @xmath13 , and @xmath241 otherwise .
it is easy to see that the sign is independent of the choice of the @xmath242 .
[ t : flags ] let @xmath226 be a simplicial coloring of an @xmath1-dimensional polytope @xmath243 . fix any permutation @xmath21 on the elements of @xmath211 $ ] .
then the combinatorial degree of @xmath226 equals the sign of @xmath21 times the number of complete flags @xmath244 counted with signs , such that for every @xmath245 the face @xmath246 is colored by @xmath247 .
this is a particular case of @xcite , theorem 2.2
. in particular , this theorem says that the combinatorial degree is zero unless for every @xmath248 there is a facet in @xmath13 colored by @xmath249 , for every @xmath250 there is a codimension two face in @xmath13 colored by @xmath251 , and so on .
one way to define a coloring @xmath252}$ ] is to give the colors to every facet of @xmath13 and then extend it by taking intersections , i.e. if @xmath253 for some facets @xmath254 then @xmath255 ( remember we have the reversed order in the target ) .
the coloring obtained in this way is called a _
facet coloring_. in the present paper we will only be interested in facet colorings . next , let @xmath13 be a polytope in @xmath90
. then one can consider the poset of all flags of faces of @xmath13 ( chains in @xmath209 ) .
the partial order is defined as follows : if @xmath256 and @xmath257 are two flags of faces of @xmath13 then @xmath258 if and only if @xmath259 is a subset of @xmath260 .
this poset can be realized as the poset of faces of a simple polytope @xmath261 whose normal fan is the barycentric subdivision of the normal fan of @xmath13 .
indeed , one can easily see that there is a 1 - 1 order preserving correspondence between codimension @xmath222 faces of @xmath261 and length @xmath222 flags of faces of @xmath13 .
in particular , facets of @xmath261 correspond to flags of faces of length one , i.e. to faces of @xmath13 .
later on we will be concerned with facet colorings not of the polytope @xmath13 itself , but the polytope @xmath261 associated with it . by the above , to define a facet coloring of @xmath261 we need to assign a non - empty proper subset of @xmath211 $ ] to every facet of @xmath261 , hence , to every face of @xmath13 .
therefore , any map @xmath252}$ ] defines a facet coloring @xmath262}$ ] .
( we should warn the reader , however , the map @xmath226 may not be a map of posets , in general . ) clearly , for every flag @xmath263 the union @xmath264 is the set of colors of the face of @xmath261 corresponding to this flag .
we thus say that a _ flag @xmath263 is colored by @xmath264_. furthermore , @xmath265 is simplicial if and only if for any flag @xmath263 the union @xmath264 is proper .
let @xmath3 be a projective toric variety of dimension @xmath1 defined by a lattice polytope @xmath13 , and let @xmath37 denote the normal fan of @xmath13 .
consider a collection of @xmath0 ( monic ) monomials @xmath266 in the homogeneous coordinate ring @xmath41 $ ] of @xmath3 .
assume that the product of the variables @xmath267 divides the product of the monomials @xmath268 .
then the quotient @xmath269 has critical degree with respect to @xmath266 . on the other hand ,
since the variables @xmath61 correspond to the facet normals of @xmath13 , any collection of monomials @xmath270 with @xmath271 defines a facet coloring of @xmath13 : @xmath272},\quad c_z(q_\rho)=\{i\in [ n+1]\ , : \ , x_\rho|z_i\},\ ] ] where @xmath273 is the facet of @xmath13 whose inner normal generates @xmath52 .
conversely , any facet coloring @xmath226 of @xmath13 defines a collection of squarefree monomials in @xmath8 whose product is divisible by the product of the variables : @xmath274 .
if @xmath266 do not vanish simultaneously on @xmath3 then the corresponding coloring @xmath275 is simplicial . indeed , if @xmath275 is not simplicial then there is a vertex @xmath19 of @xmath13 which is colored by @xmath276 , i.e. @xmath277 for some facets @xmath278 , such that @xmath278 contains @xmath279 as one of its colors .
but this implies that the corresponding point @xmath280 on @xmath3 lies on the irreducible divisors @xmath281 , where each @xmath282 is a component of the zero locus of @xmath283 on @xmath3 , a contradiction .
the next theorem asserts that the combinatorial degree of @xmath275 equals the toric residue of the quotient @xmath269 .
[ t : monomres]@xcite let @xmath3 be an @xmath1-dimensional projective toric variety defined by a lattice polytope @xmath13 .
let @xmath266 be monomials in the homogeneous coordinate ring @xmath8 such that 1 .
@xmath284 , 2 .
@xmath266 do not vanish simultaneously on @xmath3 .
then @xmath285 where @xmath275 is the simplicial coloring of @xmath13 defined by @xmath266 . to reduce the computation of the toric residue for arbitrary polynomials to the case of monomials
we will need the following generalized version of the global transformation law @xcite .
[ t : gtl ] let @xmath286 and @xmath287 for @xmath288 .
suppose @xmath289 where @xmath290 and @xmath291 are homogeneous of degree @xmath292 and @xmath293 respectively for some fixed degrees @xmath294 .
assume that neither @xmath295 nor @xmath296 vanish simultaneously on @xmath3 .
let @xmath297 , @xmath298 , @xmath299 , and @xmath300 .
then for any @xmath301 , the polynomials @xmath302 and @xmath303 are of critical degree for @xmath78 and @xmath116 respectively , and @xmath304 for any @xmath301 the degree of @xmath302 is @xmath305 , which is the critical degree for @xmath31 . consider the @xmath0 homogeneous polynomials @xmath306 .
according to the global transformation law ( theorem 0.1 , @xcite ) @xmath307 on the other hand , @xmath308 and @xmath303 has critical degree for @xmath309 .
therefore , @xmath310 again by the global transformation law .
the theorem follows .
our reduction is then based on the following assertion .
[ c : reduction ] let @xmath3 be an @xmath1-dimensional projective toric variety .
let @xmath286 be homogeneous polynomials not vanishing simultaneously on @xmath3 .
suppose @xmath311 and @xmath266 are squarefree monomials such that 1 .
@xmath312 , 2 .
@xmath313 for some @xmath314 , @xmath74 , 3 .
@xmath266 do not vanish simultaneously on @xmath3 .
then we have @xmath315 where @xmath275 is the simplicial facet coloring defined by @xmath266 .
the first statement in ( [ e : reduction ] ) follows from theorem [ t : gtl ] and the second statement follows from theorem [ t : monomres ] .
let @xmath13 be a lattice polytope in @xmath90 . consider any partition of the set of vertices of @xmath13 into @xmath0 disjoint ( possibly empty ) subsets : @xmath316 extend this partition to a partition of the set of lattice points of @xmath13 by adding to @xmath317 lattice points in the relative interior of faces containing a vertex from @xmath317 :
@xmath318 any such extension ( [ e : indpart ] ) will be called an _ induced partition _ of @xmath319 defined by the _
vertex partition _
( [ e : vertpart ] ) . now consider @xmath0 lattice polytopes @xmath16 in @xmath90 .
for each polytope @xmath7 fix an ( ordered ) vertex partition @xmath320 we say that these partitions are _ compatible _ if for any permutation @xmath21 of @xmath276 @xmath321 where @xmath322 denotes the relative interior of @xmath13 .
let @xmath16 be lattice polytopes in @xmath90 .
then subsets @xmath323 , @xmath324 , form a _ partition matrix _ for @xmath16 if @xmath325 is a collection of induced partitions defined by a compatible collection of vertex partitions @xmath326 it is not hard to see that the compatibility condition on the @xmath327 ( [ e : compatibility ] ) implies the same condition on any induced partitions : @xmath328 [ ex : partition ] consider three polygons @xmath329 , @xmath330 and @xmath331 in figure [ f : partition ] .
we partition their lattice points in accordance with the labels : the set @xmath20 consists points of @xmath7 labeled with @xmath332 ( @xmath333 ) . clearly these are induced partitions . to show that they are compatible it is enough to check that for any linear functional @xmath334 any three vertices @xmath335 , @xmath336 and @xmath337 that minimize @xmath338 on @xmath329 , @xmath330 and @xmath331 , respectively , will not have all different labels .
let @xmath13 be a polytope in @xmath90 . recall that every vector @xmath338 in the dual space @xmath339 defines a face @xmath340 of @xmath13 on which @xmath338 restricted to @xmath13 attains its minimal value .
let @xmath341 be a partition matrix for @xmath16 .
define a map from @xmath339 to the set of @xmath342-matrices of dimension @xmath343 : @xmath344 where the value of @xmath345 at @xmath346 is the matrix @xmath347 whose @xmath348-th entry is @xmath349 the matrix @xmath347 is called the _ coloring matrix _ of @xmath338 . informally speaking , the coloring matrix @xmath347 `` encodes '' the partitions of the lattice points of the @xmath7 restricted to the corresponding faces @xmath350 .
the compatibility condition implies that for any non - zero @xmath338 the coloring matrix @xmath347 has permanent zero .
indeed , if the permanent is non - zero then there exists a permutation @xmath21 of @xmath276 such that @xmath351 for all @xmath74 . by the definition of @xmath347
this implies that for each @xmath279 there is a point @xmath352 in @xmath353 that lies on the face @xmath354 .
but then the sum @xmath355 gives a point on the face @xmath340 of the minkowski sum @xmath356 , which contradicts the compatibility condition ( [ e : compatibility2 ] ) .
the following statement is known as the frobenius - knig theorem ( it is also equivalent to hall s marriage theorem @xcite ) .
[ t : f - k ] let @xmath357 be a @xmath342-matrix of dimension @xmath358 with zero permanent
. then @xmath357 has a submatrix of zeroes of dimension @xmath359 for some positive @xmath360 such that @xmath361 . by the above theorem for every non - zero @xmath338
the @xmath343 matrix @xmath347 has a zero submatrix ( not unique , in general ) of dimension @xmath359 with @xmath362 .
the rows ( resp .
columns ) of the submatrix are indexed by a subset of @xmath276 which we denote by @xmath363 ( resp .
@xmath364 ) .
we thus have @xmath365 for all non - zero @xmath338 . now consider a polytope @xmath13 whose normal fan @xmath37 is a common refinements of the normal fans of @xmath16 .
clearly , @xmath347 is the same for all @xmath338 in the intersection of the cones of the @xmath354 .
therefore , @xmath345 is constant on the cones of @xmath37 .
since cones of @xmath37 correspond to faces of @xmath13 we arrive at the following definition .
let @xmath341 be a partition matrix for @xmath16 .
let @xmath13 be a polytope whose normal fan is a common refinements of the normal fans of @xmath16 . given a face @xmath94 of @xmath13 define its _ coloring matrix @xmath366 _ to be the coloring matrix of any @xmath367 , where @xmath368 is the cone of @xmath94 .
we will need the following simple observation .
let @xmath369 be faces of @xmath13 .
then @xmath370 consider a projective @xmath1-dimensional toric variety @xmath3 defined by a projective fan @xmath37 .
let @xmath15 be @xmath0 semi - ample degrees on @xmath3 and let @xmath16 be their polytopes .
[ d : resmatrix ] consider a collection of @xmath0 homogeneous polynomials @xmath371 of degrees @xmath15 : @xmath372 given a partition matrix @xmath341 for @xmath16 define the _ residue matrix _ @xmath373 of @xmath78 to be the matrix whose entries are the homogeneous polynomials @xmath374 the determinant @xmath375 is a homogeneous polynomial of degree @xmath376 . since the @xmath11 are semi - ample , @xmath56 is also semi - ample and its polytope is the minkowski sum @xmath377 . as follows from the definition of homogeneous coordinates ( see ( [ e : homocoord ] ) ) a monomial @xmath55 of degree @xmath56 is divisible by all the variables if and only if the corresponding lattice point @xmath19 lies in the interior of the polytope of @xmath56
. therefore , by the compatibility condition ( [ e : compatibility2 ] ) every monomial in @xmath375 is divisible by all the variables , and hence the quotient @xmath378 is a homogeneous polynomial of critical degree @xmath379 .
[ p : resmatrix ] let @xmath15 be semi - ample degrees on @xmath3 with polytopes @xmath16 . fix a partition matrix @xmath341 for @xmath16 . for
every coloring matrix @xmath380 , @xmath38 , make any choice of an @xmath359 zero submatrix with @xmath362 and let its rows and columns be indexed by subsets @xmath381 and @xmath382 of @xmath276 , respectively .
define squarefree monomials @xmath383 then for any homogeneous polynomials @xmath31 of degrees @xmath15 1 .
@xmath312 , 2 .
@xmath313 for some @xmath314 , @xmath74 .
moreover , @xmath291 can be chosen so that @xmath384 where @xmath373 is the residue matrix defined by the partition matrix @xmath341 .
\(1 ) for every @xmath38 the variable @xmath61 appears in the product @xmath268 with multiplicity @xmath385 and in @xmath386 with multiplicity @xmath387 since @xmath388 .
\(2 ) for every @xmath74 we have @xmath389 we need to show that every monomial @xmath390 is divisible by at least one of @xmath266 .
since every monomial @xmath55 is divisible by a vertex monomial we can assume that @xmath19 is a vertex of @xmath7 .
recall that in homogeneous coordinates @xmath391 where @xmath392 is a representative of @xmath11 ( see ( [ e : homocoord ] ) ) .
therefore @xmath61 divides @xmath55 if and only if @xmath393 , where @xmath394 is the cone of @xmath101 corresponding to @xmath19 .
the vertex @xmath19 is contained in @xmath20 for some @xmath288 .
we show that @xmath395 divides @xmath390 .
indeed , take any @xmath61 with @xmath396 containing @xmath332 . if @xmath397 then @xmath398 . from the definition of @xmath399 it follows that @xmath400 does not contain the vertex @xmath19 , i.e. @xmath393 and so @xmath401 by above .
if @xmath402 then @xmath403 by the definition of @xmath404 . the above argument shows that @xmath405 for some homogeneous polynomial @xmath291 . taking the determinant we obtain @xmath406 .
now the last statement follows from part ( 1 ) .
the above proposition shows that given a partition matrix @xmath341 , any choice of zero submatrices in @xmath399 , for @xmath38 , defines a collection of squarefree monomials @xmath311 and @xmath266 that satisfy the conditions ( 1 ) and ( 2 ) of corollary [ c : reduction ] . if the facet coloring @xmath275 defined by the monomials @xmath266 is simplicial the condition ( 3 ) of corollary [ c : reduction ] is satisfied and that would imply the result of theorem [ t : main1 ] , namely that the residue of @xmath373 equals the combinatorial degree of @xmath275 .
however , there are examples of @xmath16 when the condition ( 3 ) fails no matter how one chooses a partition matrix and zero submatrices . to avoid this obstruction we are going to change the variety @xmath3 by taking the barycentric refinement of its fan @xmath407 .
this gives a birational morphism @xmath408 which allows us to transfer our construction to the variety @xmath409 ( see proposition [ p : birational ] ) .
the advantage of this is that for any partition matrix @xmath341 there is a canonical choice of a zero submatrix in every coloring matrix @xmath399 , for @xmath410 , which guarantees that the corresponding monomials @xmath411 do not vanish simultaneously on @xmath409 .
let @xmath16 be @xmath0 lattice polytopes in @xmath90 and @xmath100 their normal fans .
let @xmath13 be any polytope whose normal fan @xmath37 is a common refinement of the @xmath101 .
given a partition matrix we will define a canonical facet coloring of a polytope @xmath261 whose normal fan @xmath412 is the barycentric refinement of @xmath37 .
we will then prove that this coloring is simplicial .
this will allow us to define monomials @xmath411 on the toric variety corresponding to @xmath261 that satisfy all the conditions of corollary [ c : reduction ] and thus obtain our main result ( theorem [ t : main1 ] ) .
let @xmath341 be a partition matrix for polytopes @xmath16 and let the polytopes @xmath13 and @xmath261 be as above .
as mentioned in section [ s : facetcolor ] , to define a facet coloring of @xmath261 it suffices to assign a subset @xmath413 to every face @xmath94 of @xmath13 .
we will start by describing all possible candidates for @xmath217 , so called admissible colorings of @xmath94 .
let @xmath94 be a face of @xmath13 and @xmath366 its coloring matrix .
a subset @xmath414 is called an _ admissible coloring of @xmath94 _ if @xmath366 contains an @xmath359 zero submatrix with @xmath362 whose columns are indexed by @xmath415 .
it turns out that the set of admissible colorings of a face possesses very nice properties .
first , for any flag of faces of @xmath13 we have the reversed inclusion of the corresponding sets of admissible colorings : @xmath416 where @xmath417 is the set of admissible colorings of @xmath418 . indeed , ( [ e : reverse ] ) implies that every zero submatrix in the coloring matrix of @xmath418 is also a zero submarix of the coloring matrix of @xmath419 .
second , let @xmath366 be the coloring matrix of a face @xmath420 .
( in what follows we will only use that @xmath366 is an @xmath343 matrix with @xmath342-entries and zero permanent . )
denote by @xmath421 the set of all zero submatrices @xmath422 in @xmath366 of dimension @xmath359 such that @xmath362 .
this set is non - empty by theorem [ t : f - k ] .
for @xmath423 we let @xmath424 ( resp .
@xmath425 ) denote the subset in @xmath276 of indices of rows ( resp .
columns ) of @xmath422 .
we have the following lemma .
[ l : lattice ] let @xmath426
. then there is @xmath423 such that either @xmath427 or @xmath428 .
let @xmath429 be the submatrix whose rows are indexed by @xmath430 and whose columns are indexed by @xmath431 .
clearly @xmath429 is a zero submatrix .
similarly , let @xmath432 be the zero submatrix with rows indexed by @xmath433 and columns indexed by @xmath434 .
denote @xmath435 , @xmath436 , and @xmath437 . by the inclusion / exclusion formula @xmath438 .
similarly , @xmath439 , where @xmath440 , @xmath441 , and @xmath442 . summing up these two equations we obtain @xmath443 therefore , either @xmath444 or @xmath445 . in other words , either @xmath429 or @xmath432 contains a zero submatrix @xmath422 with @xmath362 , as required .
the above lemma means that if @xmath446 and @xmath447 are two admissible colorings of @xmath94 then either @xmath448 or @xmath449 is also an admissible coloring .
as follows from the proof , a slightly stronger statement is true : if @xmath449 is not an admissible coloring then any single color can be removed from @xmath448 and the remaining set will still be an admissible coloring of @xmath94 .
[ l : canonical ] let @xmath421 be as above and consider the partially ordered by inclusion set @xmath450 let @xmath451 be the set of maximal elements , and @xmath452 the set of minimal elements of @xmath453 . then the subsets @xmath454 belong to @xmath453 and satisfy @xmath455 . to prove @xmath456 we show that @xmath457 for any @xmath458 , @xmath459 .
we proceed by induction .
the case @xmath460 is trivial .
assume @xmath461 and let @xmath462 .
if @xmath463 we are done , otherwise @xmath464 by lemma [ l : lattice ] .
since @xmath465 is maximal we have @xmath466 , i.e. @xmath467 .
similar arguments show that @xmath468 . to show @xmath455 it is enough to notice that for any @xmath469 and any @xmath470 we have @xmath471 . indeed , either @xmath472 and so @xmath473 by minimality of @xmath415 , or @xmath474 and so @xmath475 by maximality of @xmath476
the above lemma supplies us with two canonical coloring of a face @xmath94 : let @xmath366 be the coloring matrix of a face @xmath420 and @xmath477 the set of all admissible colorings of @xmath94 .
maximal ( minimal ) elements of @xmath477 are called _ maximal ( minimal ) colorings _ of @xmath94 .
the union @xmath478 of minimal colorings is called the _ minimal canonical coloring _ of @xmath94 .
the intersection @xmath217 of maximal colorings is called the _ maximal canonical coloring _ of @xmath94 .
consider the three polygons @xmath329 , @xmath330 and @xmath331 from example [ ex : partition ] .
let @xmath94 be the horizontal edge of the minkowski sum @xmath479 .
then it has the coloring matrix @xmath480.\ ] ] we get @xmath481 and @xmath482 .
next let @xmath483 be the edge of @xmath13 with @xmath484 slope .
( it is the sum of the highest vertex of @xmath329 and the two edges of @xmath330 and @xmath331 of slope @xmath484 . )
its coloring matrix is @xmath485.\ ] ] this time we have @xmath486 and @xmath487 , @xmath488 . to obtain less trivial example we need to consider the case @xmath489 .
here is an example of a coloring matrix whose set of admissible colorings has more than one maximal ( and minimal ) element .
@xmath490.\ ] ] indeed , the set of admissible colorings is @xmath491 .
therefore , @xmath492 and @xmath493 .
according to the discussion in section [ s : facetcolor ] the maps @xmath494 and @xmath495 defined above give rise to two facet coloring @xmath496 and @xmath265 of @xmath261 which we call the minimal and maximal canonical facet colorings of @xmath261 , respectively .
it is easy to see that under @xmath265 no facet of @xmath261 gets all the colors .
indeed , for any face @xmath420 its coloring matrix @xmath366 can not contain zero rows , thus every maximal coloring @xmath497 is a proper subset of @xmath276 .
the next theorem shows that an even stronger statement is true : no face of @xmath261 gets all the colors .
[ t : simplicial ] the maximal and minimal canonical facet colorings of @xmath261 are simplicial and have the same combinatorial degree .
recall from section [ s : facetcolor ] that to prove @xmath265 simplicial we need to show that for any maximal flag of faces of @xmath13 @xmath498 the union @xmath499 is a proper subset of @xmath276 .
we will prove by induction that for any @xmath500 @xmath501 for some maximal coloring @xmath502 of @xmath503 .
for @xmath504 this implies the statement of the theorem , since @xmath505 is a proper subset of @xmath276 .
the base @xmath506 is clear since @xmath507 is the intersection of maximal elements of @xmath508 . for the inductive step assume that ( [ e : ih ] ) is true for some maximal @xmath509 . by ( [ e : reverse2 ] )
@xmath510 , thus there exists a maximal element @xmath511 such that @xmath512
. also @xmath513 , by definition .
this together with ( [ e : ih ] ) gives @xmath514 as required .
by lemma [ l : canonical ] @xmath515 for any face @xmath420
. therefore , @xmath496 is also simplicial .
finally , @xmath516 follows from proposition [ p : refinement ] .
when the maximal canonical coloring of a face consists of a single element we can say more about admissible colorings of this face : [ l : singlecolor ] suppose a face @xmath420 is maximally canonically colored by a single color @xmath517 .
then this is the only admissible coloring of @xmath94 .
moreover , any face containing @xmath94 is also singly canonically colored by @xmath249 while every subface of @xmath94 is canonically colored by a set containing @xmath222 .
suppose @xmath518 , @xmath519 , are maximal colorings of @xmath94 such that @xmath520 , but @xmath521 . by the proof of lemma [ l : canonical
] @xmath522 is an admissible coloring of @xmath94 . by the remark after lemma [ l : lattice ] @xmath94 can either be colored by @xmath523 or else by @xmath524 with any single color removed .
the first is a coloring strictly larger than @xmath525 which is impossible since @xmath525 is maximal .
the second is empty since @xmath524 is already a single color .
both are contradictions .
thus , the unique maximal coloring of @xmath94 is @xmath249 which is therefore the only admissible coloring . if @xmath526 then @xmath527 by ( [ e : reverse2 ] ) and , hence , @xmath528 .
if @xmath529 then @xmath249 is an admissible coloring of @xmath530 . but @xmath249 can be appended to any coloring of @xmath530 .
thus @xmath222 is contained in every maximal coloring of @xmath530 , i.e. in the maximal canonical coloring of @xmath530 .
we now turn back to residues and prove the result of theorem [ t : main1 ] . as before @xmath3 is a complete @xmath1-dimensional toric variety defined by a fan @xmath37 and @xmath15 are semi - ample degrees with polytopes @xmath16 .
we can assume that @xmath3 is projective and take @xmath13 to be the polytope of an ample divisor on @xmath3 ( if @xmath3 is not projective it can be dominated birationally by a projective toric variety .
this will not affect the toric residue computation by proposition [ p : birational ] . )
we also let @xmath261 denote a polytope whose normal fan is the barycentric subdivision of @xmath37 .
let @xmath341 be a partition matrix for @xmath16 . according to section [ s : canonical ] @xmath341 produces a map @xmath252}$ ] which assigns to every proper face @xmath94 of @xmath13 its maximal canonical coloring @xmath217 .
the induced canonical facet coloring @xmath265 of @xmath261 is simplicial by theorem [ t : simplicial ] .
the next theorem says that for any @xmath31 of degrees @xmath15 the determinant of the residue matrix @xmath373 ( see definition [ d : resmatrix ] ) gives an element whose residue is the combinatorial degree of @xmath265 .
let @xmath3 be a complete toric variety of dimension @xmath1 .
let @xmath15 be semi - ample degrees and @xmath16 their polytopes .
consider a partition matrix @xmath341 for @xmath16 .
for any collection of homogeneous polynomials @xmath31 of degrees @xmath15 consider the corresponding residue matrix @xmath373 .
then the residue of @xmath378 is equal to the combinatorial degree of the canonical facet coloring of @xmath261 : @xmath531 first notice that we can work on the variety @xmath409 defined by the polytope @xmath261 .
indeed , let @xmath532 be the birational morphism defined by the barycentric refinement @xmath407 and @xmath533 the induced homomorphism of homogeneous coordinate rings .
then each polynomial @xmath534 is of semi - ample degree @xmath535 and by proposition [ p : birational ] @xmath536 where @xmath537 and @xmath538 , for @xmath539 . since the degrees
@xmath540 have the same polytopes
@xmath7 we did not change the partition matrix and the pullback @xmath541 is the residue matrix @xmath542 for the @xmath543 .
therefore we can apply proposition [ p : resmatrix ] for the canonical facet coloring of @xmath261 to obtain squarefree monomials @xmath544 and @xmath411 in @xmath545 which satisfy 1 .
@xmath546 , 2 .
@xmath547 for some @xmath548 , @xmath74 , 3 .
@xmath411 do not vanish simultaneously on @xmath409 , 4 .
@xmath549 .
( part ( 3 ) follows since the @xmath550 define the canonical facet coloring @xmath265 of @xmath261 which is simplicial according to theorem [ t : simplicial ] . ) by corollary [ c : reduction ] @xmath551 which completes the proof .
in this section we consider the special case when the @xmath0 polytopes share a complete flag of faces .
an essential family of degrees with such collection of polytopes is called locally unmixed .
we show that for any family of locally unmixed degrees one can write an explicit partition matrix yielding an element of residue @xmath33 ( theorem [ t : loc_unmixed ] ) .
polytopes @xmath552 are said to _ share a complete flag _ if for each @xmath7 there is a complete flag of faces : @xmath553such that the sums of the corresponding entries @xmath554 form a complete flag of faces of @xmath555 : @xmath556 an immediate consequence of the above definition is that if @xmath140 is any non - empty subset of @xmath557 we can similarly define @xmath558 such that the @xmath559 also form a complete flag of faces of @xmath560 : @xmath561 let @xmath3 be a complete toric variety of dimension @xmath1 .
an essential family of semi - ample degrees @xmath136 is said to be _ locally unmixed _ if the corresponding polytopes @xmath5 share a complete flag .
note that the @xmath7 themselves may be only @xmath562 dimensional , although at least two of them must be @xmath1-dimensional since the family is essential .
[ t : loc_unmixed ] let @xmath136 be locally unmixed degrees on @xmath3 .
define partitions : @xmath563 this is a compatible collection of partitions and the corresponding residue matrix gives an element of residue @xmath32 for any homogeneous polynomials @xmath564 not vanishing simultaneously on @xmath3 .
before we begin the proof let us illustrate the partition using the following 3-dimensional example .
consider four 3-dimensional polytopes @xmath329 , @xmath330 , @xmath331 and @xmath565 as in figure [ f : polytopes ] .
they share a complete flag of faces .
indeed , each of them has a face with inner normal @xmath566 and this face has an edge along the vector @xmath567 .
then for each @xmath568 set @xmath569 consists of the vertex of the flag ( point marked as `` 0 '' ) , set @xmath570 consists of the other lattice points on the edge of the flag ( lattice points marked as `` 1 '' ) , set @xmath571 consists of the lattice points on the face of the flag , but not on the edge ( points marked as `` 2 '' ) , and the rest of the lattice points constitute @xmath572 ( points marked as `` 3 '' ) .
@xmath573 we start with a simple lemma . [
l : sum ] let @xmath13 be a polytope of dimension @xmath1 and @xmath574 a polytope such that @xmath575 is also a polytope of dimension @xmath1 . for any facet @xmath576 of @xmath13
there is a unique face @xmath483 of @xmath574 such that @xmath577 is a proper face ( in fact a facet ) of @xmath575 .
hence , if @xmath578 is a point in the relative interior of @xmath576 and @xmath579 is not on the corresponding face @xmath483 of @xmath574 then @xmath580 is in the interior of @xmath581 .
let @xmath90 be the affine span of @xmath13 .
since @xmath575 is also @xmath1-dimensional , we must have @xmath582 and @xmath583 .
for any facet @xmath576 of @xmath13 there is a unique linear functional ( up to scaling ) @xmath584 minimized on @xmath576 in @xmath13 .
let @xmath483 be the unique maximal face of @xmath574 on which @xmath585 is minimized .
the minkowski sum @xmath577 is the facet of @xmath575 on which @xmath585 is minimized and conversely any face of @xmath575 with @xmath576 a summand must minimize @xmath585 and so must be @xmath577 . for the second statement ,
note that @xmath576 is the only face of @xmath13 containing @xmath19 . by the first part
, every face @xmath530 of @xmath574 such that @xmath586 is contained in a proper face of @xmath575 must have @xmath587 . as a consequence
if @xmath588 is not on @xmath483 , hence not on any such @xmath530 , @xmath580 is not contained in any proper face of @xmath575 .
the lattice point partitions @xmath20 are induced from the vertex partitions obtained from the same rule restricted to the vertices of the @xmath7 . to show that @xmath341 is a partition matrix
we must show that : @xmath589 for any permutation @xmath21 of @xmath22 .
we will show by induction that @xmath590 for @xmath591 .
the case @xmath592 is our desired result .
let @xmath593 .
hence , the right hand side is @xmath594 , a polytope of dimension @xmath332 .
the case @xmath595 is trivial . for the induction , we assume @xmath596 next
, @xmath597 is a facet of @xmath598 ( the case @xmath592 requires that @xmath599 is actually @xmath1-dimensional ) , so we apply lemma [ l : sum ] .
any point in @xmath600 lies in the interior of @xmath597 , and any point in @xmath601 does not lie on the associated face @xmath602 of @xmath603 .
therefore , by lemma [ l : sum ] , any point in @xmath604 is in the ( relative ) interior of @xmath605 as desired . to show that the combinatorial degree of the maximal canonical coloring of @xmath606 is @xmath33 we apply theorem [ t : flags ] . recall that a face of codimension @xmath222 of @xmath606 is a flag of @xmath222 faces @xmath607 of @xmath13 .
we show that there is only one complete flag of faces of @xmath606 colored @xmath608 , namely @xmath609 . to do this we prove a few simple lemmas : [ l : lem1 ] the maximal canonical coloring of the face @xmath610 for @xmath611 is @xmath612 .
the polytope @xmath239 is the minkowski sum of @xmath613 , and each @xmath614 contains precisely all of the lattice points in @xmath615 for @xmath616 .
thus , the corresponding coloring matrix for @xmath239 has all @xmath14 s in columns @xmath617 and all @xmath151 s in columns @xmath618 .
it follows immediately that the only maximal coloring is @xmath612 , as desired .
[ l : lem2 ] the maximal canonical coloring of any proper subface of @xmath239 other than @xmath619 contains some color @xmath222 with @xmath620 .
let @xmath94 be a proper subface of @xmath239 .
we decompose @xmath94 as the minkowski sum @xmath621 where each @xmath418 is a subface of @xmath622 . if @xmath592 and if @xmath94 were a counterexample to the lemma it would have to be colored just @xmath623 .
the last column of its coloring matrix is @xmath151 .
consequently @xmath418 contains no points of @xmath624 and so is entirely contained in @xmath625 .
so we can reduce to the case @xmath611 and assume that each @xmath418 is a _ proper _ subface of @xmath622 .
now assume @xmath626 .
we show that we can take @xmath222 to be the smallest number such that for all @xmath279 , @xmath627 .
if @xmath628 for some @xmath279 then as @xmath418 is a proper face of @xmath622 we must have @xmath629 .
this is a facet of @xmath622 , so repeated applications of lemma [ l : sum ] show that every other summand @xmath630 and so @xmath631 , a contradiction .
therefore , @xmath632 . by hypothesis , for some @xmath279 , @xmath633 but @xmath634 .
in particular @xmath635 .
if @xmath636 for some @xmath637 , then another application of lemma [ l : sum ] shows that @xmath638 contains a point in the relative interior of @xmath639 and so must contain the entire face .
but this would imply @xmath418 contains @xmath640 , a contradiction .
therefore , in the coloring matrix of @xmath94 coming from @xmath341 , the entire column @xmath641 is 0 and so @xmath222 is part of the canonical maximal coloring of @xmath94 as desired .
our desired result now follows by induction . by lemma [ l : lem1 ] , @xmath642 is colored just @xmath623 and by lemma [ l : lem2 ] it is the only such face of @xmath13 ( facet of @xmath606 ) .
inductively , the face of @xmath606 given by the flag of faces @xmath643 in @xmath13 is colored @xmath644 . for the next step we must add a subface of @xmath239 to the flag with @xmath332 the only new color .
but by lemma [ l : lem1 ] and lemma [ l : lem2 ] , the only such subface is @xmath619 .
in this section we prove that matrices whose determinant have residue @xmath33 can be found for almost all essential , two - dimensional families of degrees .
recall the definition of essential in definition [ d : essential ] . in the special case @xmath2
, essential means that no @xmath7 is zero dimensional , and while some or all of the @xmath7 may be one dimensional line segments , no two such are parallel line segments .
we will show we can always find a residue matrix that gives an element of residue @xmath33 in all but one exceptional case .
degrees @xmath645 are _ exceptional _ if for two of them , @xmath11 and @xmath646 , the corresponding polygons @xmath7 and @xmath647 are 1-dimensional , and the third @xmath648 is an ample divisor on the toric variety defined by @xmath649 .
let @xmath645 be an essential , non exceptional family of degrees on a toric surface @xmath3 .
there exists a partition matrix for the @xmath11 which yields an element of residue @xmath33 for every set of @xmath564 without a common root .
note that the codimension 1 theorem for the critical degrees has been proved by cox and dickenstein @xcite when all @xmath11 are full dimensional .
such a case , of course , will never be exceptional .
it is , however possible for the critical degree to be of codimension 1 , in which case the residue map is an isomorphism , and still be exceptional .
see example [ ex : exbad ] below .
let @xmath650 be the corresponding polygons and @xmath651 their minkowski sum .
every edge @xmath652 of @xmath13 is the sum of edges from one or more of the @xmath7 and vertices from the others .
label an edge by a subset of @xmath653 consisting of those polygons for which the summand of @xmath652 is an edge .
now consider consecutive edges of @xmath13 .
proceed until we have a sequence containing all three labels @xmath654 .
take the smallest subsequence with this property .
we then have the following cases : 1 .
the sequence has length @xmath14 , so there is a single edge labeled @xmath655 .
this will be the locally unmixed case .
2 . the sequence has length @xmath656 . up to relabeling and change of direction
the sequence will be either : 1 .
@xmath657 or 2 .
+ such sequences will be called _ partially unmixed_. 3 .
the sequence has length @xmath659 or more .
all such sequences can be represented as + @xmath660{0.4}{1}}\,,\ , \boxed{1}\,,\ , \textcolor[gray]{0.4}{\boxed{1}}\ , , \dots,\ , \textcolor[gray]{0.4}{\boxed{1}}\,,\ , \boxed{\textcolor[gray]{0.4}{1}\ , 2}$ ] .
+ the numbers in gray may or may not occur .
that is to say , the first term must contain label @xmath151 but may or may not contain @xmath14 .
similarly for the final term must contain label @xmath656 and possibly also label @xmath14 .
there is at least one term labeled just @xmath14 in the middle , but there may be others .
altogether , sequences of this type will be called _
generically mixed_. case 1 : : the three degrees share an edge , hence share a complete flag consisting of this edge and either of the two vertices .
so the polygons are locally unmixed in the sense of the previous section .
therefore , we know we can always find a partition yielding a residue @xmath14 matrix .
we illustrate the partition via the diagram in figure [ f : lup ] .
+ + each of the three figures represents one of the three polygons .
the edge @xmath661 on the bottom is shared by all three polygons .
the white vertex in each @xmath7 is in partition set @xmath569 hence marked as `` @xmath151 '' in the diagram , the rest of the lattice points on @xmath661 are in set @xmath570 , and finally any lattice points off of @xmath661 are in set @xmath571 .
hence the partition matrix @xmath20 is exactly the one constructed in the previous section . since the polygons are essential , at most one of them is @xmath14-dimensional , thus two of them , say @xmath329 and @xmath331 as shown , have at least one point off the edge @xmath661 marked `` 2 '' .
case 2a : : the polygons are partitioned according to figure [ f : case2a ] .
+ + there are two distinguished edges @xmath661 and @xmath662 .
polygons @xmath329 and @xmath330 share a complete flag along edge @xmath661 and are partitioned accordingly into three sets @xmath569 , @xmath570 , and @xmath571 for @xmath663 as shown .
but , this time the third polygon @xmath331 has only one point on @xmath661 represented by the dotted line .
this point is put into set @xmath664 and all other points of @xmath331 are put into set @xmath665 . notice that in the third polygon @xmath666 .
also note that @xmath329 and @xmath330 each have only one point ( marked `` 1 '' ) on the dotted lines representing edges parallel to @xmath662 . + to see that this is a partition matrix we must show that the sum of lattice points in @xmath569 , @xmath667 , and @xmath668 with @xmath669 is in the interior of @xmath13 . from the diagram this corresponds to taking three points marked `` 0 '' , `` 1 '' and `` 2 '' respectively from three different polygons and showing they can not all
lie on parallel edges with the same inner normal .
if `` 0 '' and `` 1 '' come from the first two polygons then their sum is in the interior of the edge @xmath661 of the two - dimensional ( by essentiality ) sum of @xmath329 and @xmath330 . by definition
however a point marked `` 2 '' from @xmath331 is not on this edge .
+ if instead `` 0 '' comes from the @xmath331 and `` 1 '' comes from @xmath329 or @xmath330 the minkowski sum of these two points is either the vertex lying only on edges @xmath661 and @xmath662 or in the interior of @xmath661 . however , the points marked `` 2 '' from the third polygon @xmath329 or @xmath330 are not on either of these two edges .
+ for the combinatorial degree we apply lemma [ l : singlecolor ] .
this shows that any face of @xmath13 ( facet of @xmath606 ) maximally canonically colored by @xmath670 must only be colored by @xmath670 .
such a face must have its coloring matrix with @xmath14 s only in the first two columns .
it is easy to see from the picture that this can only happen for the bottom edge @xmath661 .
it is also easy to see that the two vertices of this edge are colored @xmath671 and @xmath672 respectively . in particular
there is a unique complete flag colored @xmath673 , as desired .
case 2b : : the difference between case 2a and case 2b is that @xmath330 now also contains an edge parallel to @xmath662 .
one attempt to account for this would be to use the same partition as in case 2a above except all the lattice points in @xmath330 along the edge @xmath662 are placed in @xmath674 .
+ if this were a partition matrix , edge @xmath661 would remain the only edge colored just @xmath675 and its vertices would still be colored @xmath671 and @xmath672 .
+ to check if this is a partition matrix , most of the arguments from the previous case go through .
if we take points marked `` 0 '' from @xmath331 and `` 1 '' from @xmath329 or @xmath330 , the sum is either the vertex between edge @xmath661 and @xmath662 or in the interior of one of these edges .
taking `` 0 '' from @xmath330 and `` 1 '' from @xmath329 again yields a point in the interior of edge @xmath661 .
however , if we take `` 0 '' from @xmath329 and `` 1 '' from @xmath330 there is a problem if there is some edge other than @xmath661 passing through both these points . but
this can only happen if the edge @xmath676 of @xmath329 directly before @xmath661 passes through the endpoint of @xmath662 in @xmath330 , marked `` 1 '' as in figure [ f : case2b1 ] .
+ + in this case the partition fails , so we try to partition in a different way , reversing the roles of @xmath661 and @xmath662 as in figure [ f : case2b2 ] .
+ + just as before this partition works as long as the next edge @xmath677 of @xmath331 after @xmath662 does not pass through the left endpoint of @xmath661 in @xmath330 . if both of the above attempts fail , we have that the edge @xmath676 of @xmath329 before @xmath661 passes through the end point of @xmath662 in @xmath330 , and the edge @xmath677 of @xmath331 after @xmath662 passes through the left endpoint of @xmath661 in @xmath330 .
+ in this final case , we show that we can find a partition matrix unless we are in the exceptional case .
first , assume that one of @xmath329 or @xmath331 is actually two - dimensional .
assume without loss of generality it is @xmath329 .
take the partition in figure [ f : case2b3 ] . + + once again , a sum of points marked `` 0 '' from @xmath329 and `` 1 '' from @xmath330 must lie on the interior of @xmath661 which contains no points marked `` 2 '' from @xmath331 . if we take `` 0 '' from @xmath330 then all the edges on which it lies are between @xmath661 and @xmath676 .
but since @xmath329 was assumed two - dimensional , the only such edge that could pass through a point marked `` 1 '' is @xmath661 itself .
+ if we take `` 1 '' from @xmath331 we must lie on an edge of @xmath331 on or before @xmath662 but on or after @xmath677 . the edge @xmath662 passes only through points marked `` 1 '' or `` 2 '' in @xmath329 and @xmath330 .
the next edge before @xmath662 is @xmath661 which passes through only `` 0 '' and `` 1 '' in @xmath329 and @xmath330 .
furthermore all edges before @xmath661 and on or after @xmath677 pass only through `` 0 '' in @xmath330 and , since @xmath676 is before @xmath677 , only through `` 0 '' in @xmath329 .
in every case we do not get three points from different partition sets from three different polygons all lying on the same edge
. + a similar argument with an analogous partition applies if @xmath331 is two - dimensional .
finally , if both @xmath329 and @xmath331 are one - dimensional , the only way the partition above can fail is if the edge @xmath676 , parallel to @xmath661 , which is known to go through a point marked `` 2 '' in @xmath330 , also goes through a point marked `` 0 '' . but this can only happen if this is the only other point of @xmath330 and moreover the edge connecting this point to the endpoint of @xmath661 is parallel to @xmath677 which is also parallel to @xmath662 .
in other words , we must have @xmath330 have the same normal fan as the minkowski sum of the two non - parallel segments that are @xmath329 and @xmath331 .
that is to say we are in the exceptional case .
case 3 : : the mixed case is somewhat easier and is partitioned as in figure [ f : mixed ] .
+ + there are now three edges @xmath661 , @xmath662 , and @xmath676 .
there may actually be several edges of @xmath330 between @xmath662 and @xmath676 in which case @xmath661 is the left most such edge .
the edge @xmath662 intersects @xmath330 in either a point or a whole edge , marked as a dashed line in the diagram .
all of the points on this edge are placed in set @xmath678 .
the edge @xmath662 together with @xmath661 and all other edges before @xmath676 intersects @xmath331 in a single point , represented by the dotted lines , placed in partition set @xmath664 hence also marked `` 0 '' in the diagram .
the edge @xmath676 together with @xmath661 and all other edges before @xmath676 intersects @xmath329 in a single point placed in set @xmath679 finally @xmath676 intersects @xmath330 in either a single point or a whole edge marked by a dashed line and all of the points are in set @xmath674 .
all the points in @xmath330 on the edges between @xmath662 and @xmath676 are also in set @xmath674 .
all other lattice points in each @xmath7 is in partition set @xmath571 thus marked as `` 2 '' .
+ to show that this is a partition matrix we must take a point marked `` 1 '' from @xmath329 or @xmath151 from @xmath331 ( otherwise we would have to take `` 2 '' from both ) . first if we took both `` 0 '' from @xmath331 and `` 1 '' from @xmath329 , then their minkowski sum is on both @xmath662 and @xmath676 .
however any point marked `` 2 '' from @xmath330 is on neither edge nor any edge in between .
if we took `` 0 '' from @xmath331 and `` 1 '' from @xmath330 , the minkowski sum is a point lying only on edges on or between @xmath661 and @xmath676 .
however , any point marked `` 2 '' from @xmath329 is not on any of these edges .
the case of `` 1 '' from @xmath330 is similar .
+ for combinatorial degree we note that the only edge colored just @xmath675 is the edge @xmath661 . every other edge either intersects a point marked `` 2 '' or , if it is one of the other edges missing `` 2 '' , can be colored @xmath680 .
the two vertices of this edge are colored @xmath672 , and @xmath681 respectively , completing the combinatorial degree computation .
to finish we show that an exceptional set of degrees _ never _ has a compatible vertex partition resulting in an element of residue one .
[ p : exceptional ] let @xmath645 be an exceptional family of degrees on a two - dimensional toric variety @xmath3 . any compatible vertex partition yields a residue matrix with determinant zero .
we can assume that the corresponding polygons @xmath329 and @xmath330 are one - dimensional and @xmath331 has the same normal fan as @xmath682 .
therefore @xmath329 and @xmath330 each have two vertices which we denote by @xmath335 , @xmath336 and @xmath683 , @xmath684 , and @xmath331 has four vertices @xmath685 , @xmath686 , @xmath687 , @xmath688 . for each pair
@xmath352 , @xmath689 there is a unique @xmath690 such that @xmath691 .
now suppose we have a compatible vertex partition with associated partition matrix @xmath341 .
if @xmath335 and @xmath336 are in the same partition set , by the above we know there do not exist @xmath689 , @xmath690 such that both @xmath692 and @xmath693 are in @xmath694 . hence , there are no non - zero terms in the expansion of the determinant of the induced residue matrix .
if @xmath335 , @xmath336 , @xmath683 , @xmath684 all lie in two columns of the partition matrix , then again there is no possible compatible choice of @xmath690 in the complementary entry and the induced residue matrix will have determinant zero .
therefore , up to relabeling we are left with only one choice of partition matrix of the form : @xmath695 for compatibility each of @xmath696 can only contain a unique vertex @xmath690 .
but this implies there is some vertex @xmath690 which can not lie in any of the @xmath696 , a contradiction .
hence there are no non - trivial compatible partition matrices .
in particular there is no partition yielding an element of residue one .
this last result shows that while we know for essential degrees there must always exist a polynomial of residue one by theorem [ t : nonzero ] , it can not always be obtained as the determinant of a matrix . on the flip side the method does work for all but one quite degenerate situation .
we illustrate this by constructing residue matrices for some examples .
consider the polynomials : @xmath697 the newton polygons are shown in figure [ f : dim2example1 ] with the lattice points labeled by their corresponding coefficients .
this falls under case 3 of the previous theorem so applying the partition as in figure [ f : partition ] yields the following residue matrix : @xmath698 the determinant is @xmath699 this is a polynomial supported on the interior of the minkowski sum @xmath700 .
the homogenization up to critical degree has toric residue equal to 1 .
consider the polynomials : @xmath701 the corresponding newton polygons are shown in figure [ f : dim2example2 ] .
these polygons can be classified under case 2a of the above theorem . as such we get the following residue matrix : @xmath702
the determinant is @xmath703 which is supported in the interior of @xmath700 ( consisting of two points ) .
once again the homogenization up to critical degree yields the desired element of residue 1 .
[ ex : exbad ] let us now consider an exceptional system : @xmath704 the newton polygons consists of two line segments and their minkowski sum ( a square ) .
since we are in the exceptional case the theorem does not apply .
however , there is a unique interior point of the minkowski sum , so the critical degree is trivial .
thus , there is a unique element of residue @xmath14 , namely the resultant itself which in this case is : @xmath705 by proposition [ p : exceptional ] this polynomial is not expressible as the determinant of a residue matrix .
given a collection of @xmath0 semi - ample divisors on a toric variety @xmath3 which do not have a common zero , there exists a toric residue map which is not identically zero if and only if the degrees of the divisors are essential .
the goal of this work was to explicitly construct an element of residue one .
we have shown how compatible partitions of the newton polytopes lead to matrices whose determinant is an element of critical degree with toric residue equal to a certain integer constant , namely the combinatorial degree of a canonical induced coloring . in the case
the polytopes share a complete flag of faces and in almost every case in dimension 2 we have shown how to choose this partition to yield an element of residue exactly one .
the most obvious open question is to find compatible partitions yielding elements of residue one in higher dimensions when the polytopes do not necessarily share a complete flag .
we have computed a large number of examples in dimension three where the four polytopes are simplices . in every case
we have found working partitions .
of course there will be exceptional families , as in dimension two , where no such partitions exist .
however , it is hoped that these will be relatively rare and perhaps nonexistent in the most important case when the polytopes are all full dimensional .
this project was inspired by discussions with eduardo cattani and alicia dickenstein .
we would like to especially thank david cox for his careful proofreading of the text and all of his advice and suggestions . | the toric residue is a map depending on @xmath0 divisors on a complete toric variety of dimension @xmath1 .
it appears in a variety of contexts such as sparse polynomial systems , mirror symmetry , and gkz hypergeometric functions . in this paper
we investigate the problem of finding an explicit element whose toric residue is equal to one .
such an element is shown to exist if and only if the associated polytopes are essential .
we reduce the problem to finding a collection of partitions of the lattice points in the polytopes satisfying a certain combinatorial property .
we use this description to solve the problem when @xmath2 and for any @xmath1 when the polytopes of the divisors share a complete flag of faces .
the latter generalizes earlier results when the divisors were all ample . |
an opinion is commonly accepted that quantum gravity should manifest itself only on the planck scale of energies , i.e. it is a high - energy phenomenon .
the value of the planck energy @xmath0 gev has been got from dimensional reasonings . in this contribution
, i would like to describe a very unexpected possibility to consider gravity as a very - low - energy stochastic process .
i enumerate those discoveries and observations which may support this my opinion . \1 . in 1998 , anderson s team reported about the discovery of anomalous acceleration of nasa s probes pioneer 10/11 @xcite ; this effect is not embedded in a frame of the general relativity , and its magnitude is somehow equal to @xmath1 , where @xmath2 is the hubble constant , @xmath3 is the light velocity . \2 . in the same 1998
, two teams of astrophysicists , which were collecting supernovae 1a data with the aim to specificate parameters of cosmological expansion , reported about dimming remote supernovae @xcite ; the one would be explained on a basis of the doppler effect if at present epoch the universe expands with acceleration .
this explanation needs an introduction of some `` dark energy '' which is unknown from any laboratory experiment .
\3 . in january 2002 ,
nesvizhevsky s team reported about discovery of quantum states of ultra - cold neutrons in the earth s gravitational field @xcite . observed energies of levels ( it means that and their differences too ) in full agreement with quantum - mechanical calculations turned out to be equal to @xmath4 ev .
if transitions between these levels are accompanied with irradiation of gravitons then energies of irradiated gravitons should have the same order - but it is of 40 orders lesser than the planck energy .
an alternative model of redshifts @xcite which is based on a conjecture about an existence of the graviton background gives us odds to see on the effect of supernova dimming as an additional manifestation of low - energy quantum gravity .
the main results of author s research in this approach are described here briefly ( it is a short version of my summarizing paper @xcite ) . starting from a statistical model of the graviton background with a low temperature
, it is shown - under the very important condition that gravitons are super - strong interacting particles - that if a redshift would be a quantum gravitational effect then one can get from its magnitude an estimate of a new dimensional constant characterizing a single act of interaction in this model .
if the isotropic graviton background exists , then it is possible photon scattering on gravitons , if one of the gravitons is virtual . due to forehead collisions with gravitons ,
an energy of any photon should decrease when it passes through the sea of gravitons . from another side , none - forehead collisions of photons with gravitons of the background will lead to an additional relaxation of a photon flux , caused by transmission of a momentum transversal component to some photons .
it will lead to an additional dimming of any remote objects , and may be connected with supernova dimming .
we deal here with the uniform non - expanding universe with the euclidean space , and there are not any cosmological kinematic effects in this model .
we shall take into account that a gravitational `` charge '' of a photon must be proportional to @xmath5 ( it gives the factor @xmath6 in a cross - section ) and a normalization of a photon wave function gives the factor @xmath7 in the cross - section .
also we assume here that a photon average energy loss @xmath8 in one act of interaction is relatively small to a photon energy @xmath9 then average energy losses of a photon with an energy @xmath10 on a way @xmath11 will be equal to @xcite : @xmath12 where @xmath13 is a constant . if a _ whole _
redshift magnitude is caused by this effect , we must identify @xmath14 where @xmath3 is the light velocity , to have the hubble law for small distances . the expression ( 1 )
is true if the condition @xmath15 takes place .
photons with a very small energy may lose or acquire an energy changing their direction of propagation after scattering . early or late such photons should turn out in the thermodynamic equilibrium with the graviton background , flowing into their own background . perhaps , the last one is the cosmic microwave background .
photon flux s average energy losses on a way @xmath16 due to non - forehead collisions with gravitons should be proportional to @xmath17 where @xmath18 is a new constant of the order @xmath19 these losses are connected with a rejection of a part of photons from a source - observer direction .
we get for the factor @xmath18 ( see @xcite ) : @xmath20 both redshifts and the additional relaxation of any photonic flux due to non - forehead collisions of gravitons with photons lead in our model to the following luminosity distance @xmath21 @xmath22 where @xmath23 . to compare a form of this predicted dependence @xmath24 by unknown , but constant @xmath2 , with the latest observational supernova data by riess et al .
@xcite , we can use the fact that @xmath25 is the luminosity distance in units of @xmath26 . in figure 1 ,
the graph of @xmath25 is shown ; observational data ( 82 points ) are taken from table 5 of @xcite .
observations of @xcite are transformed as @xmath27 with the constant @xmath28 .
the predictions fit observations very well for roughly @xmath29 .
it excludes a need of any dark energy to explain supernova dimming .
discrepancies between predicted and observed values of @xmath30 are obvious for higher @xmath31 : we see that observations show brighter sne that the theory allows , and a difference increases with @xmath31 .
= 12.98 cm it would be explained in the model as a result of specific deformation of sn spectra due to a discrete character of photon energy losses .
today , a theory of this effect does not exist .
in this model , the hubble constant may be computed .
let us consider that a full redshift magnitude is caused by an interaction with single gravitons , and @xmath32 is a cross - section of interaction by forehead collisions of a photon with an energy @xmath5 with a graviton , having an energy @xmath33 let us introduce a new dimensional constant @xmath34 , so that for forehead collisions : @xmath35 then @xmath36 where @xmath37 is an average graviton energy .
assuming @xmath38 and @xmath39 we get the following rough estimate for @xmath40 @xmath41 that gives us the phenomenological estimate of cross - section by the same and equal @xmath5 and @xmath42 : @xmath43 it follows from a universality of gravitational interaction , that not only photons , but all other objects , moving relative to the background , should lose their energy , too , due to such a quantum interaction with gravitons . if @xmath14 it turns out that massive bodies must feel a constant deceleration of the same order of magnitude as a small additional acceleration of nasa cosmic probes ( the pioneer anomaly ) .
we get for the body acceleration @xmath44 by a non - zero velocity : @xmath45 for small velocities : @xmath46 if the hubble constant @xmath2 is equal to @xmath47 ( it is the theoretical estimate of @xmath2 in this approach ) , a modulus of the acceleration will be equal to @xmath48 that has the same order of magnitude as a value of the observed additional acceleration @xmath49 for nasa probes @xcite .
it was shown by the author @xcite that screening the background of super - strong interacting gravitons creates for any pair of bodies both attraction and repulsion forces due to pressure of gravitons .
for single gravitons , these forces are approximately balanced , but each of them is much bigger than a force of newtonian attraction .
if single gravitons are pairing , an attraction force due to pressure of such graviton pairs is twice exceeding a corresponding repulsion force if graviton pairs are destructed by collisions with a body .
in such the model , the newton constant is connected with the hubble constant that gives a possibility to obtain a theoretical estimate of the last .
we deal here with a flat non - expanding universe fulfilled with super - strong interacting gravitons ; it changes the meaning of the hubble constant which describes magnitudes of three small effects of quantum gravity but not any expansion or an age of the universe .
if masses of two bodies are @xmath50 and @xmath51 ( and energies @xmath52 and @xmath53 ) , @xmath54 is a cross - section of interaction of body @xmath55 with a graviton with an energy @xmath56 where @xmath57 is a graviton frequency , @xmath58 is the same cross - section for body @xmath59 then the following attractive force will act between bodies @xmath55 and @xmath60 @xmath61 if @xmath62 is described with the planck formula , and @xmath63 is an average number of gravitons in a flat wave with a frequency @xmath57 ( on one mode of two distinguishing with a projection of particle spin ) , @xmath64 is a probability of that in a realization of flat wave a number of gravitons is equal to @xmath65 we shall have for @xmath66 the following expression ( for more details , see @xcite ) : @xmath67 a quantity @xmath66 is another average energy of running gravitons with a frequency @xmath57 taking into account a probability of that in a realization of flat wave a number of gravitons may be equal to zero , and that not all of gravitons ride at a body .
then an attractive force @xmath68 will be equal to : @xmath69 @xmath70 where @xmath71 when @xmath72 the constant @xmath73 is equal to : @xmath74 by @xmath75 @xmath76 that is three order greater than the newton constant , @xmath77 but if single gravitons are elastically scattered with body @xmath78 then our reasoning may be reversed : the same portion of scattered gravitons will create a repulsive force @xmath79 acting on body @xmath80 and equal to @xmath81 so , for bodies which elastically scatter gravitons , screening a flux of single gravitons does not ensure newtonian attraction .
but for black holes which absorb any particles and do not re - emit them , we will have @xmath82
it means that such the object would attract other bodies with a force which is proportional to @xmath73 but not to @xmath83 i.e. einstein s equivalence principle would be violated for them .
this conclusion stays in force for the case of graviton pairing , too . to ensure an attractive force which is not equal to a repulsive one
, particle correlations should differ for _ in _ and _ out _ flux .
for example , single gravitons of running flux may associate in pairs @xcite .
if such pairs are destructed by collision with a body , then quantities @xmath66 will be distinguished for running and scattered particles .
graviton pairing may be caused with graviton s own gravitational attraction or gravitonic spin - spin interaction .
left an analysis of the nature of graviton pairing for the future ; let us see that gives such the pairing . to find an average number of pairs @xmath84 in a wave with a frequency @xmath57 for the state of thermodynamic equilibrium
, one may replace @xmath85 by deducing the planck formula .
then an average number of pairs will be equal to : @xmath86 and an energy of one pair will be equal to @xmath87 it is important that graviton pairing does not change a number of stationary waves , so as pairs nucleate from existing gravitons .
the question arises : how many different modes , i.e. spin projections , may graviton pairs have ?
we assume here that the background of initial gravitons consists of two modes .
for massless transverse bosons , it takes place as by spin @xmath55 as by spin @xmath59 if graviton pairs have maximum spin @xmath88 then single gravitons should have spin @xmath19 but from such particles one may constitute four combinations : @xmath89 ( with total spin @xmath80 ) , and @xmath90 ( with total spin @xmath91 all these four combinations will be equiprobable if spin projections @xmath92 and @xmath93 are equiprobable in a flat wave ( without taking into account a probable spin - spin interaction ) .
but it follows from the energy conservation law that composite gravitons should be distributed only in two modes .
so as @xmath94 then by @xmath95 we have @xmath96 i.e. all of gravitons are pairing by low frequencies .
an average energy on every mode of pairing gravitons is equal to @xmath97 the one on every mode of single gravitons - to @xmath98 these energies are equal by @xmath99 because of that , the numbers of modes are equal , too , if the background is in the thermodynamic equilibrium with surrounding bodies .
the above reasoning does not allow to choose a spin value @xmath80 or @xmath100 for composite gravitons .
a choice of namely spin @xmath80 would ensure the following proposition : all of gravitons in one realization of flat wave have the same spin projections .
from another side , a spin - spin interaction would cause it .
the spectrum of composite gravitons is also the planckian one , but with a smaller temperature @xmath101 it is important that the graviton pairing effect does not change computed values of the hubble constant and of anomalous deceleration of massive bodies : twice decreasing of a sub - system particle number due to the pairing effect is compensated with twice increasing the cross - section of interaction of a photon or any body with such the composite gravitons .
non - pairing gravitons with spin @xmath55 give also its contribution in values of redshifts , an additional relaxation of light intensity due to non - forehead collisions with gravitons , and anomalous deceleration of massive bodies moving relative to the background .
if running graviton pairs ensure for two bodies an attractive force @xmath103 then a repulsive force due to re - emission of gravitons of a pair alone will be equal to @xmath104 it follows from that the cross - section for _ single additional scattered _ gravitons of destructed pairs will be twice smaller than for pairs themselves ( the leading factor @xmath105 for pairs should be replaced with @xmath106 for single gravitons ) . for pairs , we introduce here the cross - section @xmath107 where @xmath108 is an average pair energy with taking into account a probability of that in a realization of flat wave a number of graviton pairs may be equal to zero , and that not all of graviton pairs ride at a body ( @xmath108 is an analog of @xmath66 ) .
replacing @xmath109 and @xmath110 where @xmath111 we get for graviton pairs : @xmath112 this expression does not take into account only that beside pairs there may be single gravitons in a realization of flat wave . to reject cases when , instead of a pair , a single graviton runs against a body ( a contribution of such gravitons in attraction and repulsion is the same ) , we add the factor @xmath113 into @xmath114 @xmath115 then a force of attraction of two bodies due to pressure of graviton pairs , @xmath116 , - in the full analogy with ( 19 ) - will be equal to and @xmath2 were overestimated of @xmath117 times ] : @xmath118 @xmath119 where @xmath120 the difference @xmath121 between attractive and repulsive forces will be equal to : @xmath122 where the constant @xmath123 is equal to : @xmath124 both @xmath73 and @xmath123 are proportional to @xmath125 ( and @xmath126 so as @xmath127 ) .
if one assumes that @xmath128 then it follows from ( 17 ) that by @xmath129 the constant @xmath34 should have the value : @xmath130 we can use ( 5 ) and ( 17 ) to establish a connection between the two fundamental constants , @xmath102 and @xmath2 , under the condition that @xmath131 we have for @xmath40 @xmath132 then @xmath133 so as the value of @xmath102 is known much better than the value of @xmath134 let us express @xmath2 via @xmath135 @xmath136 or in the units which are more familiar for many of us : @xmath137 this value of @xmath2 is in the good accordance with the majority of present astrophysical estimations @xcite ( for example , the estimate @xmath138 km / s / mpc has been got from sn1a cosmological distance determinations in @xcite ) , but it is lesser than some of them @xcite and than it follows from the observed value of anomalous acceleration of pioneer 10 @xcite .
the described quantum mechanism of classical gravity gives newton s law with the constant @xmath123 value ( 17 ) and the connection ( 19 ) for the constants @xmath123 and @xmath139 we have obtained the rational value of @xmath2 ( 20 ) by @xmath140 if the condition of big distances is fulfilled : @xmath141 because it is known from experience that for big bodies of the solar system , newton s law is a very good approximation , one would expect that this condition is fulfilled , for example , for the pair sun - earth .
but assuming @xmath142 and @xmath143 we obtain assuming for rough estimation @xmath144 @xmath145 it means that in the case of interaction of gravitons or graviton pairs with the sun in the aggregate , the considered quantum mechanism of classical gravity could not lead to newton s law as a good approximation .
this `` contradiction '' with experience is eliminated if one assumes that gravitons interact with `` small particles '' of matter - for example , with atoms .
if the sun contains of @xmath146 atoms , then @xmath147 where @xmath148 is an average energy of one atom .
for rough estimation we assume here that @xmath149 where @xmath150 is a proton rest energy ; then it is @xmath151 i.e. @xmath152 this necessity of `` atomic structure '' of matter for working the described quantum mechanism is natural relative to usual bodies . but would one expect that black holes have a similar structure ? if any radiation can not be emitted with a black hole , a black hole should interact with gravitons as an aggregated object , i.e. this condition for a black hole of sun mass has not been fulfilled even at distances @xmath153 for bodies without an atomic structure , the allowances , which are proportional to @xmath154 and are caused by decreasing a gravitonic flux due to the screening effect , will have a factor @xmath155 or @xmath156 these allowances break the equivalence principle for such the bodies . for bodies with an atomic structure , a force of interaction is added up from small forces of interaction of their `` atoms '' : @xmath157 where @xmath158 and @xmath159 are numbers of atoms for bodies @xmath55 and @xmath80 .
the allowances to full forces due to the screening effect will be proportional to the quantity : @xmath160 which can be expressed via the full masses of bodies as @xmath161 or @xmath162 by big numbers @xmath158 and @xmath159 the allowances will be small .
the allowance to the force @xmath163 acting on body @xmath88 will be equal to : @xmath164 @xmath165 ( for body @xmath55 we shall have the similar expression if replace @xmath166 ) , where @xmath167 let us find the ratio : @xmath168 using this formula , we can find by @xmath169 @xmath170 an analogical allowance to the force @xmath68 has by the same conditions the order @xmath171 or @xmath172 one can replace @xmath150 with a rest energy of very big atom - the geometrical approach will left a very good language to describe the solar system .
we see that for bodies with an atomic structure the considered mechanism leads to very small deviations from einstein s equivalence principle , if the condition of big distances is fulfilled for microparticles , which prompt interact with gravitons . for small distances we shall have :
@xmath173 it takes place by @xmath174 for @xmath175 this quantity is many orders larger than the planck length .
the equivalence principle should be broken at such distances .
if the described model of redshifts is true , what is a picture of the universe ?
it is interesting that in a frame of this model , every observer has two own spheres of observability in the universe ( two different cosmological horizons exist for any observer ) @xcite .
one of them is defined by maximum existing temperatures of remote sources - by big enough distances , all of them will be masked with the cmb radiation .
another , and much smaller , sphere depends on their maximum luminosity - the luminosity distance increases with a redshift much quickly than the geometrical one .
the ratio of the luminosity distance to the geometrical one is the quickly increasing function of @xmath176 @xmath177 which does not depend on the hubble constant .
an outer part of the universe will drown in a darkness .
we can assume that the graviton background and the cosmic microwave one are in a state of thermodynamical equilibrium , and have the same temperatures .
cmb itself may arise as a result of cooling any light radiation up to reaching this equilibrium .
then it needs @xmath178 to get through the very edge of our cosmic `` ecumene '' .
some other possible cosmological consequences of an existence of the graviton background were described in @xcite .
the graviton background may turn up `` a perpetual engine '' of the universe , pumping energy from any radiation to massive objects .
an equilibrium state of the background will be ensured by such a temperature @xmath179 for which an energy profit of the background due to an influx of energy from radiation will be equal to a loss of its energy due to a catch of virtual massive gravitons with `` black holes '' or other massive objects .
in such the picture , the chances are that `` black holes '' would turn out `` germs '' of galaxies . after accumulation of a big enough energy by a `` black hole '' ( to be more exact , by a super - compact massive object ) by means of a catch of virtual massive gravitons
, the one would be absolved from an energy excess in via ejection of matter , from which stars of galaxy should form .
it awaits to understand else in such the approach how usual matter particles form from virtual massive gravitons .
there is a very interesting but non - researched possibility : due to relative decreasing of an intensity of graviton pair flux in an internal area of galaxies ( pairs are destructed under collisions with matter particles ) , the effective newton constant may turn out to be running on galactic scales .
it might lead to something like to the modified newtonian dynamics ( mond ) by mordehai milgrom ( about mond , for example , see @xcite ) . but to evaluate this effect , one should take into account a relaxation process for pairs , about which we know nothing today .
it is obvious only that gravity should be stronger on a galactic periphery .
the renormalization group approach to gravity leads to modifications of the theory of general relativity on galactic scales @xcite , and a growth of newton s constant at large distances takes place , too .
kepler s third law receives quantum corrections that may explain the flat rotation curves of the galaxies .
i would like to show here ( see @xcite ) a full realizability at present time of verifying my basic conjecture about the quantum gravitational nature of redshifts in a ground - based laser experiment .
of course , many details of this precision experiment will be in full authority of experimentalists .
it was not clear in 1995
how big is a temperature of the graviton background , and my proposal @xcite to verify the conjecture about the described local quantum character of redshifts turned out to be very rigid : a laser with instability of @xmath180 hasnt appeared after 10 years .
but if @xmath181 , the satellite of main laser line of frequency @xmath182 after passing the delay line will be red - shifted at @xmath183 ev / h and its position will be fixed ( see fig .
2 ) . it will be , and its intensity should linear rise with a path of photons in a delay line , @xmath184 . a center - of - mass of both lines
is expected to be approximately near @xmath185.,width=490 ] caused by the fact that on a very small way in the delay line only a small part of photons may collide with gravitons of the background .
the rest of them will have unchanged energies .
the center - of - mass of laser radiation spectrum should be shifted proportionally to a photon path . then due to the quantum nature of shifting process
, the ratio of satellite s intensity to main line s intensity should have the order : @xmath186 where @xmath184 is a path of laser photons in a vacuum tube of delay line .
it gives us a possibility to plan a laser - based experiment to verify the basic conjecture of this approach with much softer demands to the equipment .
an instability of a laser of a power @xmath187 must be only @xmath188 if a photon energy is of @xmath189 .
it will be necessary to compare intensities of the red - shifted satellite at the very beginning of the path @xmath184 and after it . given a very low signal - to - noise ratio , one could use a single photon counter to measure the intensities .
when @xmath190 is a quantum output of a cathode of the used photomultiplier ( a number of photoelectrons is @xmath190 times smaller than a number of photons falling to the cathode ) , @xmath191 is a frequency of its noise pulses , and @xmath192 is a desired ratio of a signal to noise s standard deviation , then an evaluated time duration @xmath193 of data acquisition would have the order : @xmath194 assuming @xmath195 we would have the estimate : @xmath196 years , that is unacceptable . but given @xmath197 , we get : @xmath198 days , that is acceptable for the experiment of such the potential importance .
of course , one will rather choose a bigger value of @xmath184 by a small laser power forcing a laser beam to whipsaw many times between mirrors in a delay line - it is a challenge for experimentalists .
from thermodynamic reasons , it is assumed here that the graviton background has the same temperature as the microwave background .
also it follows from the condition of detail equilibrium , that both backgrounds should have the planckian spectra .
composite gravitons will have spin @xmath80 , if single gravitons have the same spin as photons .
the question arise , of course : how are gravitons and photons connected ? has the conjecture by adler et al .
@xcite ( that a graviton with spin @xmath80 is composed with two photons ) chances to be true ?
intuitive demur calls forth a huge self - action , photons should be endued with which if one unifies the main conjecture of this approach with the one by adler et al . - but one may get a unified theory on this way . to verify this combined conjecture in experiment , one would search for transitions in interstellar gas molecules caused by the microwave background , with an angular momentum change corresponding to absorption of spin @xmath80 particles ( photon pairs ) .
a frequency of such the transitions should correspond to an equivalent temperature of the sub - system of these composite particles @xmath199 if @xmath200 is a temperature of the microwave background . from another side , one might check this conjecture in a laser experiment , too ( see @xcite ) . taking two lasers with photon energies @xmath201 and @xmath202
, one may force laser beams to collide on a way @xmath203 ( see fig .
if photons are self - interacting particles , we might wait that photons with energies @xmath204 , if @xmath205 , would arise after collisions of initial photons . .
an expected beam of photons with energies @xmath204 falls to a photoreceiver.,width=490 ] if we _
assume ( only here ) _ that single gravitons are identical to photons , it will be necessary to take into account the following circumstances to calculate an analog of the hubble constant for this experiment : an average graviton energy should be replaced with @xmath202 , the factor @xmath206 in ( 5 ) should be replaced with @xmath207 , where @xmath208 is a divergence of laser beam 2 , and one must use a quantity @xmath209 instead of @xmath210 in ( 5 ) , where @xmath187 is a laser 2 power and @xmath211 is a cross - section of its beam . together
all it means that we should replace the hubble constant with its analog for a laser beam collision , @xmath212 : @xmath213 taken @xmath214 , @xmath215 , @xmath216 , and @xmath217 , that is characterizing a he - ne laser , we get the estimate : @xmath218
. then photons with energies @xmath204 would fall to a photoreceiver with a frequency which should linearly rise with @xmath203 ( proportionally to @xmath219 ) , and it would be of @xmath220 if both lasers have equal powers @xmath221 , and @xmath222 .
it is a big enough frequency to give us a possibility to detect easy a flux of these expected photons in ir band .
if this tentative non - linear vacuum effect exists , it would lead us far beyond standard quantum electrodynamics to take into account new non - linearities ( which are not connected with the electron - positron pair creation ) and an essential impact of such a non - locally born object as the graviton background .
it follows from the above consideration that the geometrical description of gravity should be a good idealization for any pair of bodies at a big distance by the condition of an `` atomic structure '' of matter .
this condition can not be accepted only for black holes which must interact with gravitons as aggregated objects .
in addition , the equivalence principle is roughly broken for black holes , if the described quantum mechanism of classical gravity is realized in the nature .
because attracting bodies are not initial sources of gravitons , a future theory must be non - local in this sense to describe gravitons running from infinity .
the described quantum mechanism of classical gravity is obviously asymmetric relative to the time inversion . by the time inversion
, single gravitons would run against bodies to form pairs after collisions with bodies .
it would lead to replacing a body attraction with a repulsion .
but such the change will do impossible the graviton pairing .
cosmological models with the inversion of the time arrow were considered by sakharov @xcite .
penrose has noted that a hidden physical law may determine the time arrow direction @xcite ; it will be very interesting if namely realization in the nature of newton s law determines this direction .
a future theory dealing with gravitons as usual particles should have a number of features which are not characterizing any existing model to image the considered here features of the possible quantum mechanism of gravity .
if this mechanism is realized in the nature , both the general relativity and quantum mechanics should be modified .
any divergencies , perhaps , would be not possible in such the model because of natural smooth cut - offs of the graviton spectrum from both sides .
gravity at short distances , which are much bigger than the planck length , needs to be described only in some unified manner .
references anderson , j.d .
_ 1998 , _ 81 , _ 2858 ; _ phys .
d65 , _ 082004 ; [ gr - qc/0104064 v4 ] .
riess , a.g .
aj _ 1998 , _ 116 , _ 1009 .
perlmutter , s. et al .
_ apj _ 1999 , _ 517 , _ 565 .
nesvizhevsky , v.v .
nature _ 2002 , _ 415 , _ 297 .
ivanov , m.a .
_ general relativity and gravitation _ 2001 , _ 33 , _ 479 ; erratum : 2003 , _ 35 , _ 939 ; [ astro - ph/0005084 v2 ] .
ivanov , m.a .
gravitons as super - strong interacting particles , and low - energy quantum gravity [ hep - th/0506189 ] , [ http://ivanovma.narod.ru/nova04.pdf ] .
ivanov , m.a .
[ gr - qc/0009043 ] ; proc . of the int .
`` frontiers of fundamental physics 4 '' ( 9 - 13 dec 2000 , hyderabad , india ) , sidharth , b.g . , altaisky , m.v .
kluwer academic / plenum publishers : august 2001 ; proc . of the 4th edoardo amaldi conference on gravitational waves ( perth , western australia , 8 - 13 july 2001 ) _ class .
quantum grav .
_ 2002 , _ 19 , _ 1351 .
ivanov , m.a .
screening the graviton background , graviton pairing , and newtonian gravity [ gr - qc/0207006 ] .
ivanov , m.a .
another origin of cosmological redshifts [ astro - ph-0405083 ] .
riess , a.g .
type ia supernova discoveries at @xmath223 from the hubble space telescope ... [ astro - ph/0402512 ] ( to appear in _ apj , _ 2004 ) .
ivanov , m.a .
contribution to the quantum electronics and laser science conference ( qels95 ) , may 21 - 26 , 1995 , baltimore , usa ; paper number : qthg1 .
ivanov , m.a .
contribution to the sixth international symposium `` frontiers of fundamental and computational physics '' ( ffp6 ) , 26 - 29 september 2004 , udine , italy ; [ astro - ph/0409631 ] .
ivanov , m.a .
contribution to the tenth marcel grossmann meeting ( mg10 ) , 20 - 26 july 2003 , rio de janeiro , brazil ; [ gr - qc/0307093 ] .
ivanov , m.a .
contribution to the conference `` thinking , observing and mining the universe '' ( thinking2003 ) , 22 - 27 sep 2003 , sorrento , italy ; [ astro - ph/0309566 ] .
freedman , w. l. et al .
_ apj _ 2001 , _ 553 , _ 47 .
filippenko , a.v .
[ astro - ph/0410609 ] ; to be published in _
white dwarfs : probes of galactic structure and cosmology , _ sion , e. m. , shipman , h. l. , vennes , s. , eds .
; kluwer : dordrecht . willick , j.a . , puneet batra .
_ apj _ 2001 , _ 548 , _ 564 .
ivanov , m.a . in _
searches for a mechanism of gravity , _ ivanov , m.a .
, savrov , l.a .
nikolaev , yu.a .
publisher : nizhny novgorod , 2004 , pp 266 - 273 ( in russian ) .
ivanov , m.a . a quantum gravitational model of redshifts [ astro - ph/0409111 ] .
ivanov , m.a .
model of graviton - dusty universe [ gr - qc/0107047 ] .
milgrom , m. in the proceedings of the ii int .
workshop on the identification of dark matter , buxton , england , 1998 ; world scientific : singapore , 1999 ; [ astro - ph/9810302 ] .
shapiro , i.l . ,
sola , j. , stefancic , h. running g and @xmath224 at low energies from physics at @xmath225 : possible cosmological and astrophysical implications [ hep - ph/0410095 ] .
reuter , m. , weyer , h. running newton constant , improved gravitational actions , and galaxy rotation curves [ hep - th/0410117 ] .
ivanov , m.a .
contribution to the 14th workshop on general relativity and gravitation ( jgrg14 ) , nov 29 - dec 3 2004 , kyoto , japan ; [ gr - qc/0410076 ] .
adler , s.l .
d14 , _ 359 .
adler , s.l .
d14 , _ 379 .
sakharov , a.d .
_ jetp _ 1980 , _ 79 , _ 689 .
r. in the einstein survey _ general relativity , _ hawking , s.w .
, israel , w. , eds . ; cambridge university press : 1979 . | if gravitons are super - strong interacting particles and the low - temperature graviton background exists , the basic cosmological conjecture about the dopplerian nature of redshifts may be false . in this case , a full magnitude of cosmological redshift would be caused by interactions of photons with gravitons .
non - forehead collisions with gravitons will lead to a very specific additional relaxation of any photonic flux .
it gives a possibility of another interpretation of supernovae 1a data - without any kinematics .
these facts may implicate a necessity to change the standard cosmological paradigm . a quantum mechanism of classical gravity based on an existence of this sea of gravitons
is described for the newtonian limit .
this mechanism needs graviton pairing and `` an atomic structure '' of matter for working it , and leads to the time asymmetry .
if the considered quantum mechanism of classical gravity is realized in the nature , then an existence of black holes contradicts to einstein s equivalence principle .
it is shown that in this approach the two fundamental constants - hubble s and newton s ones - should be connected between themselves .
the theoretical value of the hubble constant is computed . in this approach
, every massive body would be decelerated due to collisions with gravitons that may be connected with the pioneer 10 anomaly .
some unsolved problems are discussed , so as possibilities to verify some conjectures in laser - based experiments . |
let @xmath8 be a partition of the integer @xmath5 , i.e. , @xmath9 and @xmath10 .
the _ length _ @xmath11 of a partition @xmath4 is the number of nonzero parts of @xmath4 .
the ( durfee or frobenius ) _ rank _ of @xmath4 , denoted rank(@xmath4 ) , is the length of the main diagonal of the diagram of @xmath4 , or equivalently , the largest integer @xmath12 for which @xmath13 .
the rank of @xmath4 is the least integer @xmath14 such that @xmath4 is a disjoint union of @xmath14 border strips ( defined below ) .
nazarov and tarasov ( * ? ? ?
1 ) , in connection with tensor products of yangian modules , defined a generalization of rank to skew partitions ( or skew diagrams ) @xmath15 .
the paper ( * ? ? ?
* proposition 2.2 ) gives several simple equivalent definitions of rank(@xmath15 ) .
one of the definitions is that rank(@xmath15 ) is the least integer @xmath14 such that @xmath16 is a disjoint union of @xmath14 border strips .
it develops a general theory of minimal border strip tableaux of skew shapes , introducing the concepts of the snake sequence and the interval set of a skew shape @xmath15 .
these tools are used to count the number of minimal border strip decompositions and minimal border strip tableaux of @xmath15 .
in particular , the paper @xcite gives an explicit combinatorial formula for the coefficients of the @xmath17 , where @xmath18 , which appear in the expansion of @xmath19 .
the paper @xcite considered a degree operator deg@xmath20 and defined the _ bottom schur functions _ to be the sum of the terms of lowest degree which appear in the expansion of @xmath19 as a linear combination of the @xmath17 .
we study the bottom schur functions in detail when @xmath21 .
in particular , in section [ sec : spacebotschur ] we give a basis for the vector space they span . in section [ sec : identity ] we show that when we substitute @xmath22 for @xmath23 in the expansion of a bottom schur function in terms of power sums , then the resulting symmetric function has the same coefficients when expanded in terms of power sums or augmented monomial symmetric functions .
in general we follow ( * ? ? ? * ch . 7 ) for notation and terminology involving symmetric functions .
let @xmath4 be a partition of @xmath5 with frobenius rank @xmath24 . recall that @xmath24 is the length of the main diagonal of the diagram of @xmath4 , or equivalently , the largest integer @xmath12 for which @xmath13 .
let @xmath25 , the number of parts of @xmath4 equal to @xmath12 .
define @xmath26 .
a _ border strip _ ( or _ rim hook _ or _ ribbon _ ) is a connected skew shape with no @xmath27 square .
an example is @xmath28 whose diagram is illustrated in figure [ fig : borderstrip ] .
define the _ height _
ht(@xmath29 ) of a border strip @xmath29 to be one less than its number of rows .
, width=192 ] let @xmath30 be a weak composition of @xmath5 , i.e. , @xmath31 and @xmath32 .
define a _ border strip tableau _ of shape @xmath4 and type @xmath33 to be an assignment of positive integers to the squares of @xmath4 such that : 1 .
every row and column is weakly increasing , 2 .
the integer @xmath12 appears @xmath34 times , and 3 .
the set of squares occupied by @xmath12 forms a border strip .
equivalently , one may think of a border - strip tableau as a sequence @xmath35 of partitions such that each skew shape @xmath36 is a border - strip of size @xmath34 .
for instance , figure [ fig : borderstriptableau ] shows a border strip tableau of @xmath37 of type @xmath38 . of type @xmath38,width=144 ]
it is easy to see ( in this nonskew case ) that the smallest number of strips in a border - strip tableau is rank(@xmath4 ) .
define the _ height _
ht(@xmath39 ) of a border - strip tableau @xmath39 to be @xmath40 where @xmath41 are the ( nonempty ) border strips appearing in @xmath39 .
in the example we have ht@xmath42 .
now we can define @xmath43 summed over all border - strip tableaux of shape @xmath4 and type @xmath44 .
since there are at least rank@xmath45 strips in every tableau , we have that @xmath46 if @xmath47 rank@xmath45 .
the numbers @xmath48 for @xmath49 are the values of the irreducible characters @xmath50 of the symmetric group @xmath51 .
finally we can express the schur function @xmath1 in terms of power sums @xmath17 , viz .
, @xmath52 define deg(@xmath23 ) @xmath53 , so deg(@xmath54 . the _ bottom schur function _
@xmath0 is defined to be the lowest degree part of @xmath1 , so @xmath55 also write @xmath56 .
for instance , @xmath57 hence @xmath58 we identify a partition @xmath4 with its _ diagram _ @xmath59 let @xmath60 be an edge of the lower envelope of @xmath4 , i.e. , no square of @xmath4 has @xmath60 as its upper or left - hand edge .
we will define a certain subset @xmath61 of squares of @xmath4 , called a _
snake_. if @xmath60 is horizontal and @xmath62 is the square of @xmath4 having @xmath60 as its lower edge , define @xmath63 if @xmath60 is vertical and @xmath62 is the square of @xmath4 having @xmath60 as its right - hand edge , define @xmath64 in figure [ fig : snakes ] the nonempty snakes of the shape @xmath65 are shown with dashed paths through their squares , with a single bullet in the two snakes with just one square .
the _ length _ @xmath66 of a snake @xmath67 is one fewer than its number of squares ; a snake of length @xmath68 ( so with @xmath12 squares ) is call an _ i - snake_. call a snake of * even * length a _
left snake _ if @xmath60 is horizontal and a _
right snake _
if @xmath60 is vertical .
it is clear that the snakes are linearly ordered from lower left to upper right . in this linear ordering , replace a left snake with the symbol @xmath69 , a right snake with @xmath70 , and a snake of odd length with @xmath71 . the resulting sequence ( which does not determine @xmath4 )
is called the _ snake sequence _ of @xmath4 , denoted ss@xmath45 .
for instance , from figure [ fig : snakes ] we see that @xmath72 [ lemma2.1 ] the @xmath69 s in the snake sequence correspond exactly to horizontal edges of the lower envelope of @xmath4 which are below the line @xmath73 .
the @xmath70 s correspond exactly to vertical edges of the lower envelope of @xmath4 which are above the line @xmath73 .
all other edges of the lower envelope of @xmath4 are labelled by @xmath71 s .
clearly we could have defined the snake sequence this way ; however , the definitions above also hold for skew shapes .
lemma [ lemma2.1 ] only holds when @xmath4 is a straight ( i.e. , nonskew ) shape .
let @xmath60 be an edge of the lower envelope of @xmath4 below the line @xmath74 .
let @xmath62 be the square of @xmath4 having @xmath60 as its lower edge .
the last square in the snake is some square in the first column of @xmath4 .
so if @xmath60 is horizontal then the last square is @xmath75 , the snake has an odd number of squares and so has even length , and is labelled by @xmath69 .
if @xmath60 is vertical then the last square is @xmath76 , the snake has an even number of squares , so has odd length , and is labelled by @xmath70 .
the case when @xmath60 is above @xmath73 is proved similarly .
[ cor : lsfirst ] in the snake sequence of @xmath4 , the @xmath69 s occur strictly to the left of the @xmath70 s .
the number of horizontal edges of the lower envelope of @xmath4 which are below the line @xmath77 equals the length of the main diagonal of the diagram of @xmath4 , which is the rank of @xmath4 .
similarly the number of vertical edges of the lower envelope of @xmath4 which are above the line @xmath73 also equals the rank of @xmath4 .
henceforth we fix @xmath78 .
let ss@xmath79 , and define an _
interval set _ of @xmath4 to be a collection @xmath80 of @xmath24 ordered pairs , @xmath81 satisfying the following conditions : 1 .
the @xmath82 s and @xmath83 s are all distinct integers , 2 .
@xmath84 , 3 .
@xmath85 and @xmath86 .
figure [ fig : iset ] illustrates the interval set @xmath87 of the shape 533322 .
given an interval set @xmath88 , define the _ crossing number _
@xmath89 to be the number of _ crossings _ of @xmath80 , i.e. the number of pairs @xmath62 for which @xmath90 .
let @xmath39 be a border strip tableau of shape @xmath4 . recall that @xmath91 where @xmath29 ranges over all border strips in @xmath39 and @xmath92 is one less than the number of rows of @xmath29 .
define @xmath93 to be the height @xmath94 of a `` greedy border strip tableau '' @xmath39 of shape @xmath4 obtained by starting with @xmath4 and successively removing the largest possible border strip .
( although @xmath39 may not be unique , the set of border strips appearing in @xmath39 is unique , so @xmath94 is well - defined . )
the connection between bottom schur functions and interval sets was given by stanley ( * ? ? ?
* theorem 5.2 ) : @xmath95 where @xmath80 ranges over all interval sets of @xmath44 . for example
the shape @xmath96 has snake sequence @xmath97 .
there are two interval sets , @xmath98 with crossing number @xmath99 , and @xmath100 with crossing number @xmath101 .
so as we saw before @xmath102
[ l : shapeord ] the lexicographic order on shapes @xmath44 whose length @xmath103 equals their rank @xmath24 is equal to the reverse lexicographical order ( with respect to the ordering l@xmath104r@xmath104o ) on their snake sequences . since @xmath105 , the snake sequence begins with @xmath24 l s . if the length of the @xmath12th row of @xmath44 is @xmath106 , then there are @xmath107 o s to the left of the @xmath108st r. denote the _ complete homogeneous symmetric functions _ by @xmath109 .
recall that the jacobi - trudi identity expresses the @xmath1 s in terms of the @xmath110 s : @xmath111 where we define @xmath112 for @xmath113 .
for example @xmath114.\ ] ] since @xmath115 , the term of lowest degree ( in @xmath116 ) in the expansion of a given @xmath117 in terms of the @xmath118 is just @xmath119 . for a product
@xmath120 the term of lowest degree in the expansion in terms of the @xmath118 is just @xmath121 .
so we have that @xmath122 terms of lowest order in @xmath123 ( since the @xmath124 are algebraically independent , and since @xmath125 , this determinant will not vanish ) .
for example @xmath126.\ ] ] since @xmath127 , the terms of lowest order are those which contain the most number of 1 s .
row @xmath12 of the matrix will have a @xmath99 in position @xmath62 if @xmath128 , i.e. if @xmath129 ( this shows that the number of rows of @xmath130 which do not contain a @xmath99 is another definition of @xmath131 ( * ? ? ? * prop .
2.2 ) ) .
let @xmath132 be the matrix obtained from the original jacobi - trudi matrix by removing every row and column which contains a 1 and replacing the @xmath133 with @xmath134 .
we show below that this matrix is not singular and so we have @xmath135 for example @xmath136.\ ] ] any minor of the jacobi - trudi matrix for a shape @xmath4 is the jacobi - trudi matrix for some skew shape @xmath137 . for
let @xmath138 be some minor of size @xmath139 of some jacobi - trudi matrix @xmath140 .
if the entry in position @xmath62 is @xmath141 put @xmath142 .
now we can set @xmath143 and @xmath144 again note that since the @xmath124 are algebraically independent and @xmath145 , we have @xmath146 . in our running example
, we have @xmath147 and @xmath148 .
hence @xmath149 . also @xmath150 and @xmath151 .
thus @xmath152 .
therefore we have that @xmath153 equals the determinant of the jacobi - trudi matrix of @xmath154 with the @xmath155 s replaced by @xmath156 s .
[ l : squarek ] if the skew shape @xmath137 has the jacobi - trudi matrix @xmath138 obtained by removing all rows and columns with a 1 from a jacobi - trudi matrix @xmath140 of a shape @xmath4 with rank @xmath24 , then @xmath157 contains a square of size @xmath24 .
the rank of @xmath158 is @xmath159 , and the diagram of @xmath154 does indeed contain a square of size 4 : we give a proof due to christine bessenrodt , greatly improving our original proof . define @xmath160 and @xmath161 .
we give a diagrammatic definition of @xmath162 and @xmath163 which also illustrates that the skew diagram @xmath164 contains a square of size @xmath24 .
consider @xmath4 as a @xmath165 square with two partitions @xmath33 and @xmath166 glued to it , i.e. @xmath167 . flip @xmath33 over its anti - diagonal and then glue the bottom right corner of the result to the bottom left corner of the square .
the final diagram is the skew diagram of @xmath164 .
we show that @xmath168 and @xmath169 .
the @xmath24 rows of @xmath138 are contained in the first @xmath24 rows of @xmath140 , so @xmath170 for some constant @xmath171 .
the last column of @xmath140 does not have a 1 in it , so it will not be removed , and its first @xmath24 entries will be the last column of @xmath138 . hence @xmath172 . since @xmath173 , we have @xmath174 .
the first @xmath24 entries of the last column of @xmath140 are retained .
then we remove the next @xmath175 columns to its left , do not remove the next column , remove the next @xmath176 columns to the left , and so on .
formally we have @xmath177 .
combining this with @xmath178 gives us @xmath179 .
before we use the above results to give a basis for the space spanned by the bottom schur functions , we must first recall some classical tableaux theory . if @xmath180 is a skew shape , then a _ standard young tableau _ ( syt ) of shape
@xmath180 is a labelling of the squares of @xmath180 with the numbers @xmath181 , each number appearing once , so that every row and column is increasing .
a _ semistandard young tableau _ ( ssyt ) of shape @xmath180 is a labelling of the squares of @xmath180 with positive integers that is weakly increasing in every row and strictly increasing in every column .
we say that @xmath39 has _ type _
@xmath182 if @xmath39 has @xmath34 parts equal to @xmath12 .
now we define an operation ( of schtzenberger ) on standard young tableaux called a _
jeu de taquin slide_. given a skew shape @xmath15 , consider the squares @xmath183 that can be added to @xmath15 , so that @xmath183 shares at least one edge with @xmath184 , and @xmath185 is a valid skew shape .
suppose that @xmath183 shares a lower or right edge with @xmath15 ( the other situation is completely analogous ) .
there is at least one square @xmath186 in @xmath15 that is adjacent to @xmath183 ; if there are two such squares , then let @xmath186 be the one with a smaller entry .
move the entry occupying @xmath186 into @xmath183 . then repeat this procedure , starting at @xmath186 .
the resulting tableau will be a standard young tableau .
analogously if @xmath183 shares an upper or left edge , the operation is the same except we let @xmath186 be the square with the bigger entry from two possibilities .
for example we illustrate both situations in figure [ fig : jeudetaquin ] ; the tableau on the right results from playing jeu de taquin beginning at the square marked by a bullet on the tableau on the left ( and vice versa ) .
two tableaux @xmath39 and @xmath187 are called _ jeu de taquin equivalent _ if one can be obtained from another by a sequence of jeu de taquin slides .
given an syt @xmath39 of shape @xmath180 , there is exactly one syt @xmath188 of straight shape , denoted jdt@xmath189 , that is jeu de taquin equivalent to @xmath39 ( * ? ? ?
a1.2.4 ) .
the _ reading word _ of a ( semi)standard young tableau is the sequence of entries of @xmath39 obtained by concatenating the rows of @xmath39 bottom to top .
for example , the tableau on the left in figure [ fig : jeudetaquin ] has the reading word @xmath190 .
reverse reading word _ of a tableau is simply the reading word read backwards .
lattice permutation _ is a sequence @xmath191 such that in any initial factor @xmath192 , the number of @xmath12 s is at least as great as the number of @xmath193 s ( for all @xmath12 ) .
for example @xmath194 is a lattice permutation
. the _ littlewood - richardson coefficients _
@xmath195 are the coefficients in the expansion of a skew schur function in the basis of schur functions : @xmath196 the littlewood - richardson rule is a combinatorial description of the coefficients @xmath195 . we will use two different versions of the rule . fix an syt @xmath188 of shape @xmath44 .
the littlewood - richardson coefficient @xmath195 is equal to the number of syt of shape @xmath15 that are jeu de taquin equivalent to @xmath188 . for example
, let @xmath197 and @xmath198 .
consider the tableau @xmath188 of shape @xmath44 shown above .
there are exactly two syts @xmath39 of shape @xmath199 such that @xmath200 , namely , the littlewood - richardson coefficient @xmath195 is equal to the number of semistandard young tableaux of shape @xmath201 and type @xmath44 whose reverse reading word is a lattice permutation .
for example , with @xmath197 and @xmath198 as above , there are exactly two ssyts @xmath39 of shape @xmath16 and type @xmath44 whose reverse reading word is a lattice permutation : now we have enough machinery to state and prove this section s main theorem .
fix @xmath5 and @xmath24 .
the set @xmath202 and @xmath203 is a basis for the space @xmath204 .
for example if @xmath205 and @xmath206 , we have that @xmath207 is a basis for @xmath208 .
first we prove that the @xmath209 are linearly independent .
we show that given any such @xmath44 , there is some term in the expansion of @xmath209 which does not occur in the expansion of any @xmath210 for @xmath211 lexicographically less than @xmath44 . from ( * ?
* theorem 5.2 ) we have that @xmath95 where @xmath80 ranges over all interval sets of @xmath44 .
let @xmath212 be the term corresponding to the noncrossing interval set @xmath80 of the snake sequence of @xmath44 .
we claim that @xmath213 does not occur in the expansion of any @xmath210 for @xmath211 lexicographically less than @xmath44 .
assume by way of contradiction that it does occur for some such @xmath211 with corresponding interval set @xmath214 .
assume inductively that the first @xmath68 l s are matched with the last @xmath68 r s without crossings in @xmath214 .
let @xmath215 ( and @xmath216 respectively ) be the position of the @xmath107th r in the snake sequence of @xmath44 ( @xmath211 respectively ) . by lemma [ l : shapeord ] @xmath217 .
but the length of the interval matching the @xmath12th r from the right in @xmath80 is @xmath218 .
so for there to be an interval of this length in @xmath214 we must match the @xmath12th r from the right with the @xmath12th l ; this interval has no crossing . proceeding by induction we see that @xmath219 is also noncrossing , and so must equal @xmath80 .
this shows that the snake sequences corresponding to @xmath44 and @xmath211 are equal .
identical snake sequences and equal lengths guarantee that @xmath220 , a contradiction .
now we prove that the @xmath209 span the space of all @xmath221 .
we have shown that @xmath222 .
expand @xmath223 in terms of ( straight ) schur functions using the littlewood richardson rule @xmath224 we need to show that @xmath225 unless @xmath44 is of rank @xmath24 and length @xmath24 .
fix an syt p of shape @xmath44 .
the littlewood - richardson coefficient @xmath226 is equal to the number of syt of shape @xmath157 that are jeu de taquin equivalent to p. playing jeu de taquin on a straight - shape tableau of shape @xmath44 can only increase the length of the shape . hence if @xmath227 .
the littlewood - richardson coefficient @xmath226 is also equal to the number of semistandard young tableaux of shape @xmath137 and type @xmath44 whose reverse reading word is a lattice permutation .
but we know that @xmath137 contains a square of size @xmath24 ( by lemma [ l : squarek ] )
. therefore the bottom @xmath24 boxes of this square must have labels at least @xmath24 .
since @xmath228 , the labels are exactly @xmath24 .
so @xmath229 , i.e. rank(@xmath230 .
since @xmath231 , we must have rank(@xmath232 .
taking terms of lowest degree on both sides of @xmath233 we have that @xmath234 where the sum is over @xmath44 of length @xmath24 and rank @xmath24 as required .
let @xmath235 be the number of partitions of @xmath5 with length at most @xmath24 , and define @xmath236 .
a partition @xmath237 of length @xmath24 and rank @xmath24 decomposes into a @xmath165 square of boxes and a partition of @xmath238 of length at most @xmath24 .
[ cor : botdim ] the dimension of the space of bottom schur functions + @xmath239 is @xmath240 for example , the first 27 terms in this sequence ( beginning with @xmath241 ) are @xmath242 there is a nice bijection between the above partitions and the set of partitions @xmath243 . for , given a @xmath24 and a partition @xmath244 with fewer than @xmath24 parts , we can set @xmath245 .
this gives a partition of @xmath5 with @xmath24 rows with @xmath246 as required .
this is clearly a bijection .
this classical sequence also gives the number of partitions of @xmath5 into parts congruent to @xmath99 or @xmath247 ; equivalently these numbers are the coefficients in the expansion of the rogers - ramanujan identity @xmath248
we have shown that a basis for the space spanned by the bottom schur functions consists of the @xmath0 where @xmath249 .
it is natural to define for fixed @xmath250 the @xmath107-_bottom schur function _
@xmath251 to be the sum of those terms of degree at most rank@xmath252 in the expansion ( [ eq : char ] ) ( with @xmath253 as usual ) . when @xmath254 we have verified ( using stembridge s sf package for maple @xcite ) that for @xmath255 the dimension of the space spanned by @xmath256 equals the number of @xmath257 satisfying @xmath258 .
this suggests the following conjecture . a basis for the space spanned by the 2-bottom schur functions consists of all 2-bottom schur functions @xmath259 , where @xmath4 is a partition of n satisfying @xmath260 .
however in the @xmath261 case , the dimensions of the spaces spanned by the 3-bottom schur functions are @xmath262 .
we have computed that the numbers of @xmath257 satisfying @xmath263 are given by @xmath264 .
unfortunately these sequences do not agree .
we prove a surprising identity satisfied by a variant of the bottom schur functions related to their expansion in terms of power sum and monomial symmetric functions .
fix a shape @xmath4 of rank @xmath24 .
given an interval set @xmath265 of @xmath4 and a labelling of the intervals @xmath266 such that @xmath267 , define @xmath268 recall that @xmath269 is the number of crossings of the interval set @xmath80 .
figure [ fig : intervalset ] shows a labelled interval set of the shape 533322 with the snake sequence @xmath270 . for this interval set @xmath271 and for this labelling @xmath272 .
for the shape @xmath273 with snake sequence @xmath274 , figure [ fig : involution ] depicts some of the labelled interval sets . in the top
left we have @xmath275 . in the top right we have @xmath276 .
in fact in every row the term @xmath277 in the left column is exactly the negative of the corresponding term @xmath277 in the right column .
we give a sign reversing involution on these labelled interval sets .
examine a specific labelled interval set @xmath80 .
since we are dealing with straight ( non - skew ) shapes , we know by corollary [ cor : lsfirst ] that the snake sequence has all the l s before any of the r s , or @xmath279 .
so any two intervals @xmath12 and @xmath107 ( @xmath280 say ) either intersect ( @xmath281 ) or are nested ( @xmath282 ) .
let @xmath283 be the smallest label which is repeated .
the intervals in @xmath80 are ordered by where they start , so identifying the first two intervals @xmath12 and @xmath107 ( @xmath280 ) labelled by @xmath283 is well - defined
. simply change the interval @xmath284 to @xmath285 and the interval @xmath286 to @xmath287 , while preserving the label @xmath283 on both . where the intervals start remains unchanged , so these intervals remain the first two intervals labelled by @xmath283 .
hence this operation is an involution .
note that if the two intervals initially nested , they now intersect , and if they initially intersected , they now nest .
the parity of the number of crossings of these two with any other interval is preserved under this operation . so the parity of the total number of crossings has changed , and this involution is sign reversing .
the rows of figure [ fig : involution ] are some examples of this involution ( if @xmath288 ) .
[ e : powersums ] consider as before the shape @xmath296 with snake sequence @xmath274 .
in particular , consider the interval set @xmath297 of type @xmath298 labelled by @xmath299 .
this interval set is illustrated in figure [ fig : labiset ] .
if @xmath300 then @xmath301 and the sum over all such labellings will give @xmath302 .
if @xmath303 the sum over all such labellings will give @xmath304 .
similarly if @xmath305 we will get @xmath306 , and if @xmath307 we will get @xmath308 . finally if the three labels are all different , the sum will give @xmath309 . so the sum over all such labellings is @xmath310 .
we need to show that for every interval set @xmath80 of type @xmath291 , @xmath311 , where the sum is over all labellings of @xmath80 .
first note that the intervals can be ordered largest first and left to right among intervals of the same length .
so the @xmath12th interval is well defined , and has length @xmath312 . by definition @xmath313 .
but if we expand this product into monomials @xmath314 , each monomial corresponds uniquely to the labelling of @xmath80 where the @xmath107th interval is labelled by @xmath315 , and so occurs exactly once in @xmath316 as required .
fix a specific interval set @xmath80 of type @xmath291 .
we need to show @xmath322 where the sum is over all labellings with no label repeated . as before we can order the intervals and say that the @xmath12th interval is labelled by @xmath34 .
note that if @xmath323 , the two labellings @xmath324 and @xmath325 both produce the same term @xmath326 .
so we have @xmath327 where we impose the condition that if @xmath323 , then @xmath328 .
recall that @xmath329 by definition .
so we have @xmath330 as required .
from lemma [ l : powersums ] we have @xmath335 , where the sum is over all labelled interval sets of type @xmath291 .
but by lemma [ l : indistinct ] @xmath336 if we sum over all labelled interval sets with a repeated label , while by lemma [ l : monomials ] we have @xmath337 if we sum over all labelled interval sets with no label repeated .
so @xmath332 .
let @xmath338 denote the space of all symmetric functions @xmath339 with rational coefficients such that if @xmath340 , then @xmath341 theorem [ th : identity ] shows that @xmath342 . however , there are elements of @xmath338 that are not linear combinations of @xmath0 s , such as @xmath343 .
let @xmath70 denote the transition matrix from monomial symmetric functions to the power sums , i.e. , @xmath344 let @xmath345 denote the diagonal matrix
whose diagonal coincides with that of @xmath70 , so @xmath346 it is easy to see that @xmath347 , the kernel ( or null space ) of @xmath348 . let @xmath349 denote the elements of @xmath338 that are homogeneous of degree @xmath5 , and let @xmath350 .
we have computed that @xmath351 compare this with the dimension @xmath352 of the space spanned by the bottom schur functions of degree @xmath5 , given in general by corollary [ cor : botdim ] and for @xmath353 just below this corollary .
in particular , the least @xmath5 for which @xmath354 is @xmath355 .
we do nt have a conjecture for the value of @xmath356 . | we give a basis for the space spanned by the sum @xmath0 of the lowest degree terms in the expansion of the schur symmetric functions @xmath1 in terms of the power sum symmetric functions @xmath2 , where deg@xmath3 .
these lowest degree terms correspond to minimal border strip tableaux of @xmath4 .
the dimension of the space spanned by @xmath0 , where @xmath4 is a partition of @xmath5 , is equal to the number of partitions of @xmath5 into parts differing by at least 2 . applying the rogers - ramanujan identity
, the generating function also counts the number of partitions of @xmath5 into parts @xmath6 and @xmath7 .
we also show that a symmetric function closely related to @xmath0 has the same coefficients when expanded in terms of power sums or augmented monomial symmetric functions . * bottom schur functions * + _ cnri , _ _ dublin institute of technology , ireland _
+ ` [email protected] ` + 0.5 cm * richard p. stanley * + _ department of mathematics , massachusetts institute of technology _
+ _ cambridge , ma 02139 , usa _
+ ` [email protected] ` + submitted : nov 19 , 2003 ; accepted 27 aug , 2004 + mr subject classifications : 05e05 , 05e10 + |
let @xmath0 be a graph .
a _ drawing _ of @xmath0 means a representation of the graph in the euclidean plane or the @xmath1-sphere , where vertices are points and edges are simple arcs joining their end - vertices .
since each edge is simple , no edge admits self crossings .
furthermore , we assume that the interiors of edges do not contain vertices , and that two edges do not intersect if they have a common vertex , and that two edges without common end - vertex intersect at most once , and if so , then they intersect transversally .
these requirements are essential in this paper .
a drawing is called a _ regular drawing _ ( resp . _ semi - regular drawing _ ) if it has only double ( resp .
triple ) crossing points . from the requirements , we know that a graph has at least @xmath2 vertices if it admits a semi - regular drawing with at least one triple crossing point .
the crossing number @xmath3 of @xmath0 is defined to be the minimal number of crossing points over all regular drawings of @xmath0 . in particular , @xmath4 if @xmath0 is planar . in this paper
, we introduce a new variation of the crossing number .
triple crossing number _
@xmath5 is zero if @xmath0 is planar , and @xmath6 if @xmath0 does not admit a semi - regular drawing .
otherwise , @xmath5 is defined to be the minimal number of triple crossing points over all semi - regular drawings of @xmath0 . in particular , @xmath7 if and only if @xmath0 is planar .
the triple crossing number can be regarded as a specialization of the degenerate crossing number introduced by pach and tth @xcite .
in addition , for example , the petersen graph is known to have the crossing number two ( and thus non - planar ) , and hence has the triple crossing number one from figure [ fig : petersen ] . in general
, we have the inequality @xmath8 for these two notions , since we obtain a regular drawing from a semi - regular drawing by perturbing each triple crossing point into three double crossing points . in this paper ,
we determine the triple crossing numbers for all complete multipartite graphs .
a complete multipartite graph is a graph whose vertex set can be partitioned into at least two , mutually disjoint non - empty sets , called the partite sets , so that two vertices @xmath9 and @xmath10 are adjacent if and only if @xmath9 and @xmath10 belong to different sets of the partition .
if the partite sets are of sizes @xmath11 @xmath12 , then the graph is denoted by @xmath13 .
we always assume that @xmath14 if @xmath15 .
in particular , if all @xmath16 , then the graph @xmath17 is the complete graph @xmath18 with @xmath19 vertices . here
is how the paper is organized .
after we describe basic lemmas , used in the paper repeatedly , in section [ sec : pre ] , we show that the triple crossing number of a complete @xmath19-partite graph is @xmath6 if @xmath20 in section [ sec:5partite ] . in the successive sections , we work on the cases when @xmath21 . here we should mention that the hardest part is the case where @xmath22 , in particular , long , but elementary , geometric arguments are needed to show that @xmath23 , @xmath24 , @xmath25 and @xmath26 with @xmath27 do not admit a semi - regular drawing .
this is treated in sections [ sec : k54 ] , [ sec : k44 ] and [ sec : kn3 ] . after concluding the case where @xmath22 in section [ sec : bipartite ] , the cases where @xmath28 and @xmath29 are established in sections [ sec:4partite ] and [ sec : tripartite ] , respectively .
section [ sec : comment ] contains some remarks on our requirements for drawings and a generalization of triple crossing number .
we would like to thank the referee for careful reading and suggestions making the paper more readable .
basic terms of graph theory can be found in textbooks such as @xcite . [
lem : nonplanar ] the complete bipartite graph @xmath30 and the complete graph @xmath31 with five vertices are non - planar . also , a graph is non - planar if it contains @xmath30 or @xmath31 as a subgraph .
[ lem : plane ] let @xmath0 be a plane graph with @xmath32 vertices and @xmath33 edges
. then @xmath0 has a vertex of degree less than @xmath2 and the inequality @xmath34 holds .
furthermore , @xmath35 if and only if each region of @xmath0 is @xmath36-sided . for the proofs of these lemmas , see @xcite .
[ lem : alg - condition ] let @xmath0 be a graph with @xmath32 vertices and @xmath33 edges . if @xmath0 admits a semi - regular drawing , then @xmath34 .
thus , if @xmath37 , then @xmath38 .
let @xmath39 be a semi - regular drawing of @xmath0 , and let @xmath40 be the number of triple crossing points in @xmath39 .
if a new vertex is added to each triple crossing point , then we obtain a ( simple ) plane graph @xmath41 .
since @xmath41 has @xmath42 vertices and @xmath43 edges , we have @xmath44 by lemma [ lem : plane ] , from which we have the conclusion .
[ lem : faces ] let @xmath0 be a connected plane graph with @xmath32 vertices , @xmath33 edges and @xmath45 faces .
let @xmath46 . 1 .
if @xmath47 , then one face is @xmath48-sided , and the others are @xmath36-sided .
2 . if @xmath49 , then either 1
. one face is @xmath50-sided , and the others are @xmath36-sided ; or 2 .
two faces are @xmath48-sided , and the others are @xmath36-sided .
3 . if @xmath51 , then either 1
. one face is @xmath2-sided , and the others are @xmath36-sided ; or 2 .
one face is @xmath50-sided , another face is @xmath48-sided , and the others are @xmath36-sided ; or 3 .
three faces are @xmath48-sided , and the others are @xmath36-sided .
we remark that @xmath52 , and @xmath53 if and only if all faces are @xmath36-sided by lemma [ lem : plane ] . by euler s formula ( @xcite ) ,
let @xmath55 denote the number of @xmath56-sided faces of @xmath0 .
then @xmath57 thus @xmath58 .
since @xmath59 and @xmath60 , we have @xmath61 in particular , the difference between the second and third terms , which is @xmath62 , is at most @xmath63 .
we remark that @xmath64 .
when @xmath65 , @xmath66 is the only multiple of four within the interval @xmath67 $ ] .
since @xmath68 is a multiple of four , we see @xmath69 , giving @xmath70 . furthermore , if @xmath71 , then ( [ eq : eq ] ) reduces to @xmath72 .
combining this with @xmath73 gives @xmath74 .
\(1 ) since @xmath75 , we have @xmath71
. then @xmath76 , and thus @xmath77 .
\(2 ) since @xmath78 , we have @xmath79 and @xmath80 .
if @xmath81 , then @xmath82 , giving @xmath77 .
this is the conclusion ( a ) . if @xmath83 , then @xmath84 , and thus @xmath85
this is the conclusion ( b ) .
\(3 ) since @xmath86 , we have @xmath87 .
if @xmath88 , then @xmath89 , and thus @xmath77 .
this is the conclusion ( a ) .
suppose @xmath79 . since @xmath90 , we see @xmath80 .
if @xmath81 , then @xmath91 . from ( [ eq : eq ] ) , @xmath92 . combining this with @xmath93 gives @xmath94 . since @xmath95
, we have @xmath76 .
hence @xmath96 and @xmath85 .
this is the conclusion ( b ) .
finally suppose @xmath83 .
then @xmath97 , and thus @xmath98 .
this is the conclusion ( c ) .
[ thm:5multi ] if @xmath20 , then no complete @xmath19-partite graph @xmath0 admits a semi - regular drawing .
thus , @xmath38 . assume for a contradiction that @xmath0 admits a semi - regular drawing @xmath39 .
let @xmath100 .
if a new vertex is added to each triple crossing point , then we have a plane graph @xmath41 . however , the original vertices have degree at least @xmath101 , and the new vertices have degree @xmath2 .
this contradicts lemma [ lem : plane ] .
let @xmath102 .
then @xmath0 has @xmath103 vertices and @xmath104 edges .
then @xmath105 this contradicts lemma [ lem : alg - condition ] .
finally , let @xmath106 .
as above , @xmath107 this contradicts lemma [ lem : alg - condition ] again .
[ cor : complete ] let @xmath108 be the complete graph with @xmath109 vertices . then @xmath110 if @xmath111 , then @xmath108 is planar , and thus @xmath112 by definition .
the rest follows from theorem [ thm:5multi ] .
throughout this section , we will assume that @xmath113 admits a semi - regular drawing .
we will show that this is impossible .
let @xmath114 and @xmath115 be the partite sets of @xmath0 . for convenience , we refer to vertices of @xmath116 ( resp .
@xmath117 ) as _ black _ ( resp . _
white _ ) vertices .
we denote the edge @xmath118 by @xmath119 for @xmath120 .
these are called _
@xmath121-lines_. similarly , define @xmath122 , and call them @xmath123- , @xmath124- , @xmath39-lines , respectively .
in particular , each black vertex is of degree four , and is incident with four distinct classes of lines .
we fix a semi - regular drawing of @xmath0 hereafter , which is denoted by the same symbol @xmath0 .
notice that any line in @xmath0 connects a white vertex and a black vertex , and there may be triple crossing points on it . from our requirements for drawings ,
each triple crossing point of @xmath0 arises from three distinct classes of lines .
this fact will be referred to as _ property _ ( @xmath125 ) throughout the paper .
property ( @xmath125 ) is very useful and powerful .
for example , if an @xmath121-line and a @xmath123-line intersect at a triple crossing point , then we can conclude that the remaining line through the triple crossing point is either a @xmath124- or @xmath39-line .
let @xmath40 be the number of triple crossing points .
add a new vertex to each triple crossing point .
then we have a plane graph @xmath41 with @xmath126 vertices and @xmath127 edges .
since @xmath128 , the faces of the plane graph @xmath41 are all @xmath36-sided , except a single @xmath48-sided face by lemma [ lem : faces ] . for the semi - regular drawing @xmath0
, a face means that of @xmath41 , although it is an abuse of words .
a @xmath36-sided face is also called a triangle .
take a look around vertex @xmath121 .
there are five faces of @xmath0 , since @xmath121 is not a cut vertex .
we may assume that all five faces around vertex @xmath121 are triangles without loss of generality , since the @xmath48-sided face is incident with at most two white vertices .
there are two types of triangle around vertex @xmath121 as shown in figure [ fig : triangle ] .
type i triangle _ is incident with two triple crossing points , and a _ type ii triangle _ is incident with a black vertex and a triple crossing point . and an adjoint pair of type ii triangles ] notice that type ii triangles appear in pairs .
more precisely , this means that every type ii triangle at @xmath121 shares an @xmath121-line fully with another type ii triangle .
such a pair of type ii triangle is referred to as an _
adjoint pair _ of type ii triangles .
see figure [ fig : triangle ] .
hence the number of type ii triangles at vertex @xmath121 is either @xmath129 , @xmath1 or @xmath48 .
we will eliminate these three possibilities .
[ lem : typeii-4 ] the number of type ii triangles at vertex @xmath121 is not four .
suppose that there are four type ii triangles at @xmath121 .
then we can assume that the local configuration at @xmath121 is as shown in figure [ fig:4triangle](1 ) by renaming @xmath130 , if necessary .
( recall that each black vertex is incident with four distinct classes of lines . ) then , by property ( @xmath125 ) , the horizontal line is a @xmath123- or @xmath39-line , and the right lower line is a @xmath124- or @xmath39-line .
see figure [ fig:4triangle](2 ) , where the symbol @xmath131 , for example , indicates the class of the horizontal line .
it does not mean that vertex @xmath123 or @xmath39 locates the left side of the horizontal line .
thus there are four cases as shown in figure [ fig : case1-sub ] . here , the class of the right upper line is determined by property ( @xmath125 ) and the fact that each black vertex is incident with four distinct classes of lines . for figure [ fig : case1-sub](4 ) , the right upper line may be a @xmath124-line .
but then , it can be reduced to ( 3 ) by renaming @xmath123 and @xmath39 and symmetry . ]
[ cl : case1 ] figure [ fig : case1-sub](1 ) is impossible . consider the @xmath123-line @xmath132 .
let @xmath133 and @xmath134 be the faces incident with @xmath132 and vertex @xmath135 .
see figure [ fig : case1 - 1 - 2 ] .
-line @xmath132 and two faces @xmath133 , @xmath134 ] if @xmath133 is @xmath48-sided , then @xmath134 is @xmath36-sided , because there is only one @xmath48-sided face .
since two @xmath123-lines can not intersect at a triple crossing point by property ( @xmath125 ) , the horizontal @xmath123-line and @xmath132 meet at vertex @xmath123 as shown in the second of figure [ fig : case1 - 1 - 2 ] in order to make @xmath134 @xmath36-sided .
however , then @xmath133 can not be @xmath48-sided .
thus we can conclude that @xmath133 is @xmath36-sided .
then vertex @xmath123 is located around @xmath133 as shown in the first of figure [ fig : case1 - 1 ] by the same reason as above .
is @xmath36-sided ] also , then @xmath134 can not be @xmath48-sided , and thus @xmath36-sided .
hence the @xmath123-line , intersecting two @xmath121-lines @xmath137 and @xmath138 , turns out to be @xmath139 .
that is , it goes to black vertex @xmath140 as shown in the third of figure [ fig : case1 - 1 ] .
( at this point , @xmath139 may contain triple crossing points on it after crossing @xmath138 . )
consider the face @xmath141 , which is adjacent to @xmath133 along the @xmath123-line @xmath142 .
if @xmath141 is @xmath36-sided , then the @xmath123-line @xmath143 , which must be @xmath144 or @xmath145 , intersects the @xmath121-line @xmath146
. then @xmath143 can reach neither @xmath147 nor @xmath148 , because @xmath143 can not cross @xmath139 or meet @xmath146 twice .
therefore , we found that @xmath141 is the only @xmath48-sided face .
thus @xmath149 is @xmath36-sided .
we see that the line going through the upper triple crossing point of @xmath149 is either an @xmath121- or @xmath123-line by property ( @xmath125 ) .
since @xmath150 is also @xmath36-sided , the @xmath123-line @xmath151 goes to @xmath147 , or crosses the @xmath121-line @xmath137 . in any case
, @xmath152 can not be @xmath36-sided as described in figure [ fig : case1 - 1 - 3 ] .
this is a contradiction .
goes to @xmath147 or crosses @xmath137 ] figure [ fig : case1-sub](2 ) is impossible .
first , since the horizontal @xmath39-line intersects @xmath137 , @xmath138 , @xmath154 , it is either @xmath155 or @xmath156 from our requirements for drawings . by symmetry
, we can assume that @xmath133 is @xmath36-sided as in figure [ fig : case1 - 2 ] .
then vertex @xmath39 is located around @xmath133
. then the horizontal @xmath39-line turns out to be @xmath155 .
however , neither @xmath134 nor @xmath141 is @xmath36-sided , which contradicts that there is only one @xmath48-sided face . is @xmath36-sided ] figure [ fig : case1-sub](3 ) is impossible . in figure
[ fig : case1 - 3 ] , either @xmath141 or @xmath149 is @xmath48-sided .
( for , if @xmath141 is not @xmath48-sided , then it is @xmath36-sided .
then vertex @xmath39 is located there , which implies that @xmath149 is not @xmath36-sided . )
thus both @xmath133 and @xmath134 are @xmath36-sided .
then the @xmath123-line @xmath139 is determined . since @xmath150 is @xmath36-sided , the @xmath123-line @xmath143 , which is @xmath144 or @xmath145 , crosses the @xmath121-line @xmath146 .
then @xmath143 can reach neither @xmath147 nor @xmath148 , a contradiction . and
@xmath134 are @xmath36-sided ] figure [ fig : case1-sub](4 ) is impossible . in figure
[ fig : case1 - 4 ] , if @xmath133 is not @xmath36-sided , then @xmath134 is @xmath36-sided , and then vertex @xmath39 is located there .
but then , @xmath133 can not be @xmath48-sided .
hence @xmath133 is @xmath36-sided .
similarly , so is @xmath134 . then the @xmath39-line @xmath155 is determined . and
@xmath134 are @xmath36-sided ] if @xmath141 is @xmath36-sided , then the @xmath39-line @xmath63 , which is @xmath157 or @xmath158 , crosses the @xmath121-line @xmath146 .
then @xmath63 can not reach any black vertex as before .
hence @xmath141 is @xmath48-sided . as in the proof of claim [ cl : case1 ] ,
examining @xmath159 leads to a contradiction .
this completes the proof of lemma [ lem : typeii-4 ] .
[ lem : typeii-2 ] the number of type ii triangles at vertex @xmath121 is not two .
suppose that there are two type ii triangles at @xmath121 .
then we can assume that the local configuration at @xmath121 is as shown in figure [ fig : case2start](1 ) , up to renaming . by property ( @xmath125 )
, the left upper line is a @xmath123- or @xmath39-line .
similarly , the right upper line is a @xmath124- or @xmath39-line .
see figure [ fig : case2start](2 ) . ]
then there are three cases , up to symmetry and relabeling of vertices , as shown in figure [ fig : case2 ] , where the class of the horizontal line is determined by property ( @xmath125 ) .
figure [ fig : case2](1 ) is impossible .
first , assume that @xmath133 is @xmath48-sided in figure [ fig : case2 - 1 - 1 ] .
then the others are all @xmath36-sided .
is @xmath48-sided ] thus @xmath134 and @xmath141 , and then @xmath149 , @xmath150 are determined as in figure [ fig : case2 - 1 - 1 ] .
( if an @xmath121-line goes through the left triple crossing point of @xmath134 , then the face sharing a @xmath39-line with @xmath134 can not be @xmath36-sided . similarly for @xmath141 . )
consider the @xmath123-line @xmath143 .
it goes to @xmath135 or crosses the @xmath121-line @xmath160 .
suppose that the former happens
. then @xmath161 are determined as in figure [ fig : case2 - 1 - 1a ] . moreover ,
the @xmath39-line @xmath162 is also determined .
goes to @xmath135 ] then the @xmath124-line @xmath164 can not go to @xmath165 , since it crosses @xmath162 .
hence it crosses the @xmath121-line @xmath166 .
this forces the @xmath39-line @xmath63 to cross the same @xmath166 .
then it can not reach any black vertex , a contradiction .
therefore , @xmath143 crosses the @xmath121-line @xmath160 . by the same reason , @xmath164 crosses @xmath166 . repeating the same argument
, we obtain the configuration as shown in figure [ fig : case2 - 1 - 1b ] .
is @xmath48-sided ] if the @xmath123-line @xmath151 crosses the @xmath121-line @xmath138 , then both @xmath151 and @xmath167 go to @xmath147 , a contradiction .
thus @xmath151 goes to @xmath148 .
similarly , the @xmath124-line @xmath168 is determined .
see the second of figure [ fig : case2 - 1 - 1b ]
. then @xmath133 can be incident with neither vertex @xmath123 nor @xmath124 .
for example , if @xmath133 is incident with @xmath123 , then the left face of @xmath133 can not be @xmath36-sided .
hence @xmath133 is incident with two more triple crossing points .
then the upper horizontal line of @xmath133 is an @xmath121- or @xmath39-line by property ( @xmath125 ) . from our requirements
, it can not be an @xmath121-line .
thus we have the third of figure [ fig : case2 - 1 - 1b ] , but then vertex @xmath39 can not be located .
next , assume that @xmath133 is @xmath36-sided .
we see that a @xmath39-line goes through the upper triple crossing point of @xmath133 from our requirements and property ( @xmath125 ) . by symmetry
, we can assume that @xmath134 is @xmath36-sided .
see figure [ fig : case2 - 1 - 2 ] . is @xmath36-sided ] if @xmath141 is also @xmath36-sided , then the @xmath123-line @xmath169 and the @xmath124-line @xmath170 meet twice , a contradiction . hence @xmath141 turns out to be @xmath48-sided . also , no @xmath121-line is adjacent to @xmath141 , because each of the four @xmath121-lines @xmath171 meets @xmath169 or @xmath170 .
thus vertex @xmath123 is located as in figure [ fig : case2 - 1 - 2 ] .
again , examining @xmath159 leads to a contradiction as in the proof of claim [ cl : case1 ] .
figure [ fig : case2](2 ) is impossible . in figure
[ fig : case2 - 2 ] , suppose that @xmath133 is @xmath48-sided
. then @xmath134 is @xmath36-sided , and so vertex @xmath39 appears there .
then @xmath141 is not @xmath36-sided , a contradiction .
hence @xmath133 is @xmath36-sided .
is @xmath36-sided ] if @xmath134 is @xmath36-sided , then @xmath141 is @xmath48-sided as above .
otherwise , @xmath134 is @xmath48-sided . in any case
, @xmath149 is @xmath36-sided , and vertex @xmath124 appears .
also , @xmath150 and @xmath152 are @xmath36-sided . but
this is impossible as in the proof of claim [ cl : case1 ] again .
figure [ fig : case2](3 ) is impossible . in figure
[ fig : case2 - 3 ] , at least two of @xmath172 are @xmath36-sided .
if @xmath133 and @xmath134 are @xmath36-sided , then vertex @xmath39 can not be located correctly .
similarly for the case where @xmath133 and @xmath141 are @xmath36-sided .
hence @xmath134 and @xmath141 are @xmath36-sided .
then the @xmath39-line @xmath63 goes to @xmath135 or @xmath147 , and another @xmath39-line @xmath173 goes to @xmath148 or @xmath165 . but
this is impossible . and
@xmath141 are @xmath36-sided ] this completes the proof of lemma [ lem : typeii-2 ] .
[ lem : typeii-0 ] the number of type ii triangles at vertex @xmath121 is not zero .
assume that there are no type ii triangles at @xmath121 .
up to symmetry and relabeling of vertices , the local configuration at @xmath121 can be assumed as in the first of figure [ fig : case3 ] . ] by symmetry , we can assume that the right hand side does not contain a @xmath48-sided face .
more precisely , @xmath174 are all @xmath36-sided .
thus vertex @xmath123 is located .
then examining @xmath175 , as in the proof of claim [ cl : case1 ] , leads to a contradiction .
( in this case , @xmath141 can be incident with @xmath165 .
then @xmath149 can not be @xmath36-sided likewise . )
[ thm : k45 ] @xmath23 does not admit a semi - regular drawing .
this follows from lemmas [ lem : typeii-4 ] , [ lem : typeii-2 ] and [ lem : typeii-0 ] .
throughout this section , we will assume that @xmath176 admits a semi - regular drawing .
we will show that this is impossible .
let @xmath177 and @xmath178 be the partite sets of @xmath0 . as in section
[ sec : k54 ] , we refer to vertices of @xmath116 ( resp .
@xmath117 ) as black ( resp .
white ) vertices , and use the same notions as @xmath121-lines , type i or ii triangles , so on . also , property ( @xmath125 ) holds from our requirements .
we fix a semi - regular drawing of @xmath0 , which is denoted by @xmath0 again .
let @xmath40 be the number of triple crossing points .
add a new vertex to each triple crossing point .
then we have a plane graph @xmath41 with @xmath179 vertices and @xmath180 edges . since @xmath181 , either 1 .
one face of @xmath41 is @xmath50-sided , and the others are @xmath36-sided ; or 2 .
two faces of @xmath41 are @xmath48-sided , and the others are @xmath36-sided by lemma [ lem : faces ] . as in section
[ sec : k54 ] , a face of @xmath0 means that of @xmath41 .
we treat the case where one face of @xmath0 is @xmath50-sided , and the others are @xmath36-sided . at most two white vertices appear in the @xmath50-sided face .
hence we can assume that four faces at vertex @xmath121 are all @xmath36-sided .
thus the number of type ii triangles at vertex @xmath121 is either @xmath129 , @xmath1 or @xmath48 .
we will eliminate these three possibilities .
[ lem : i4 ] the number of type ii triangles at vertex @xmath121 is not four .
assume that there are four type ii triangles at @xmath121 .
we may assume that the local configuration at @xmath121 is as shown in figure [ fig : casei-0](1 ) , up to renaming . by property ( @xmath125 ) ,
the right upper line is a @xmath124- or @xmath39-line , and the left upper line is a @xmath123- or @xmath39-line ( see figure [ fig : casei-0](2 ) ) . ] up to symmetry and relabeling , there are two possibilities as shown in figure [ fig : casei-1 ] .
figure [ fig : casei-1](1 ) is impossible . by symmetry
, we can assume that both @xmath133 and @xmath134 are @xmath36-sided .
since @xmath133 is @xmath36-sided , vertex @xmath39 is located .
then @xmath134 can not be @xmath36-sided , a contradiction .
[ cl : cc ] figure [ fig : casei-1](2 ) is impossible . by symmetry
, we can assume that @xmath133 , @xmath134 are @xmath36-sided again .
then the @xmath123-line @xmath144 meets the @xmath124-line @xmath168 , a contradiction .
this complete the proof of lemma [ lem : i4 ] .
[ lem : i2 ] the number of type ii triangles at vertex @xmath121 is not two .
assume that there are two type ii triangles at @xmath121 .
as before , we may assume that the local configuration at @xmath121 is as shown in figure [ fig : casei-2start](1 ) . by property ( @xmath125 ) ,
the right upper line is a @xmath124- or @xmath39-line , and the left upper line is a @xmath123- or @xmath39-line ( see figure [ fig : casei-2start ] ) . up to symmetry and relabeling
, there are two possibilities as shown in figure [ fig : casei-2 ] . ]
figure [ fig : casei-2](1 ) is impossible .
assume that @xmath134 is not @xmath36-sided .
then @xmath133 is @xmath36-sided , and so vertex @xmath123 is located there . thus the @xmath123-line @xmath132 is determined .
see the first of figure [ fig : casei-2 - 1 ] .
another @xmath123-line @xmath143 crosses the @xmath121-line @xmath160 , but then it can not reach any black vertex . crosses @xmath160 ] hence @xmath134 is @xmath36-sided , so vertex @xmath39 is located
. then @xmath141 can not be @xmath36-sided .
thus @xmath133 is @xmath36-sided , so vertex @xmath123 is located , and the @xmath123-line @xmath132 is determined again as in the second of figure [ fig : casei-2 - 1 ] . examining @xmath143 leads to a contradiction as above .
[ cl : i-2 ] figure [ fig : casei-2](2 ) is impossible . by symmetry
, we may assume that both @xmath133 and @xmath134 are @xmath36-sided .
then vertex @xmath123 is located .
see the first of figure [ fig : casei-2 - 2 ] . and
@xmath134 are @xmath36-sided ] if @xmath141 is @xmath36-sided , then the @xmath123-line @xmath143 and the @xmath124-line @xmath164 meet twice , a contradiction .
hence @xmath141 is not @xmath36-sided , and @xmath143 goes to @xmath135 as in figure [ fig : casei-2 - 2 ] .
then a line @xmath151 turns out to be a @xmath123-line . but this @xmath123-line can not reach any black vertex , otherwise it crosses @xmath160 twice .
this completes the proof of lemma [ lem : i2 ] .
[ lem : i0 ] the number of type ii triangles at vertex @xmath121 is not zero .
assume that there is no type ii triangle at @xmath121 .
up to symmetry and relabeling , there are two possibilities as shown in figure [ fig : casei-3 ] . in any case
, we can assume that @xmath133 is @xmath36-sided by symmetry .
then vertex @xmath123 is located there .
the @xmath123-line @xmath143 goes to @xmath140 or @xmath135 , and the @xmath123-line @xmath151 goes to @xmath147 or @xmath148 .
this is impossible .
[ prop : casei ] case ( 1 ) is impossible .
this immediately follows from lemmas [ lem : i4 ] , [ lem : i2 ] and [ lem : i0 ] .
we treat the case where two faces of @xmath0 are @xmath48-sided , and the others are @xmath36-sided .
there are two subcases .
* all white vertices are incident with a @xmath48-sided face .
* there is a white vertex which is not incident with a @xmath48-sided face . around each
@xmath48-sided face , just two white vertices appear .
hence there is just one @xmath48-sided face around each white vertex .
also , it implies that if a face is incident with two adjacent triple crossing points or an adjacent pair of a triple crossing point and a black vertex , then it must be @xmath36-sided . in this subcase , then the number of type ii triangle at vertex @xmath121 is @xmath129 , @xmath183 , @xmath1 or @xmath36 .
[ lem : ii3 ] the number of type ii triangles at @xmath121 is not three .
assume that there are thee type ii triangle at @xmath121 .
up to symmetry and renaming , the situation is as shown in figure [ fig : caseii3](1 ) , where @xmath184 is @xmath48-sided . ] as remarked above , @xmath133 is @xmath36-sided .
if @xmath184 is incident with vertex @xmath39 , then @xmath133 is incident with vertex @xmath39 .
this is impossible .
thus @xmath184 is incident with vertex @xmath124 .
see figure [ fig : caseii3](2 ) .
since @xmath134 is also @xmath36-sided , the left upper line of @xmath134 is a @xmath123-line .
we have the configuration as in figure [ fig : caseii3a](1 ) .
( continued ) ] then @xmath141 is @xmath36-sided , but this forces @xmath149 to be neither @xmath36-sided nor @xmath48-sided , a contradiction .
see figure [ fig : caseii3a](2 ) .
[ lem : ii2 ] the number of type ii triangles at @xmath121 is not two .
assume that there are two type ii triangles at @xmath121 .
up to symmetry and renaming , there are two possibilities as shown in figure [ fig : caseii-1 - 1 ] , where @xmath184 is @xmath48-sided . notice that @xmath133 and @xmath134 are @xmath36-sided . for figure
[ fig : caseii-1 - 1](1 ) , either vertex @xmath124 or @xmath39 appears around @xmath184 by property ( @xmath125 ) . since @xmath133 is @xmath36-sided , the former is impossible .
the latter is also impossible , because @xmath134 is @xmath36-sided . for figure [ fig : caseii-1 - 1](2 )
, vertex @xmath124 appears around @xmath184 by property ( @xmath125 ) . then @xmath133 gives a contradiction , again .
[ lem : ii1 ] the number of type ii triangles at @xmath121 is not one .
suppose that there is one type ii triangle at @xmath121 .
up to symmetry and renaming , the configuration is as shown in figure [ fig : caseii1 ] , where @xmath184 is @xmath48-sided . ]
notice that @xmath133 is @xmath36-sided .
if @xmath184 is incident with vertex @xmath124 , then @xmath133 is incident with @xmath124 , an impossible .
hence @xmath184 is incident with vertex @xmath39 .
thus we have the configuration as in figure [ fig : caseii1 ] .
the fact that @xmath134 is @xmath36-sided forces @xmath141 to be @xmath48-sided .
then @xmath141 is incident with vertices @xmath124 and @xmath39 , which contradicts the fact that @xmath39 is incident with only one @xmath48-sided face .
[ lem : ii0 ] the number of type ii triangles at @xmath121 is not zero .
assume that there is no type ii triangle at @xmath121 .
up to symmetry and renaming , there are two possibilities as shown in figure [ fig : caseii-1 - 2 ] , where @xmath184 is @xmath48-sided . for figure
[ fig : caseii-1 - 2](1 ) , @xmath133 is @xmath36-sided , so vertex @xmath123 is located there .
then @xmath143 goes to @xmath140 or @xmath135 , and @xmath151 goes to @xmath147 or @xmath148 .
this is impossible . for figure
[ fig : caseii-1 - 2](2 ) , vertex @xmath124 appears around @xmath184 by property ( @xmath125 ) . on the other hand
, @xmath133 is @xmath36-sided .
this is impossible . in this subcase
, we may assume that vertex @xmath121 is not incident with a @xmath48-sided face without loss of generality .
thus the number of type ii triangles at @xmath121 is @xmath129 , @xmath1 or @xmath48 .
[ lem : ii-2 - 4 ] the number of type ii triangles at @xmath121 is not four .
suppose that there are four type ii triangles at @xmath121 .
then there are two possibilities as in figure [ fig : casei-1 ] .
figure [ fig : casei-1](1 ) is impossible . among the four faces @xmath174 ,
at least two are @xmath36-sided .
furthermore , if @xmath133 ( resp .
@xmath141 ) is @xmath36-sided , then @xmath134 ( resp .
@xmath149 ) is @xmath48-sided , and vice versa . up to symmetry , there are three possibilities : * @xmath133 and @xmath141 are @xmath36-sided . *
@xmath133 and @xmath149 are @xmath36-sided .
* @xmath134 and @xmath141 are @xmath36-sided .
\(a ) in this case , @xmath134 and @xmath149 are @xmath48-sided .
thus the others are all @xmath36-sided .
see figure [ fig : caseii-2 - 1-a](1 ) . by examining @xmath150 ,
the left upper line of @xmath149 is not an @xmath121-line .
since @xmath150 and @xmath152 are @xmath36-sided , two @xmath39-lines @xmath63 and @xmath173 go to @xmath135 , or cross , a contradiction .
\(b ) in this case , @xmath134 and @xmath141 are @xmath48-sided . by examining @xmath63 and @xmath173 shown in figure [ fig : caseii-2 - 1-a](2 ) , the same argument as ( a ) leads to a contradiction .
\(c ) in this case , @xmath133 and @xmath149 are @xmath48-sided . as above
, we see that neither @xmath133 nor @xmath149 is incident with an @xmath121-line .
see the first of figure [ fig : caseii-2 - 1-c ] .
thus we have two @xmath123-lines @xmath143 and @xmath151 as shown there .
since @xmath150 and @xmath152 are @xmath36-sided , vertex @xmath123 is located as in the second of figure [ fig : caseii-2 - 1-c ] .
but then @xmath185 can not be @xmath36-sided , a contradiction .
figure [ fig : casei-1](2 ) is impossible .
if both @xmath133 and @xmath134 are @xmath36-sided , then we have a contradiction as in the proof of claim [ cl : cc ] .
hence either of @xmath133 or @xmath134 is @xmath48-sided . similarly ,
either @xmath141 or @xmath149 is @xmath48-sided .
then there are two possibilities , up to symmetry .
* @xmath133 and @xmath141 are @xmath48-sided . *
@xmath133 and @xmath149 are @xmath48-sided .
\(d ) then @xmath134 and @xmath149 are @xmath36-sided as in the first of figure [ fig : caseii-2 - 1 - 1 ] . by property ( @xmath125 ) , the right upper line @xmath186 of @xmath141 is an @xmath121- or @xmath123-line .
if @xmath186 is an @xmath121-line , then it is either @xmath160 or @xmath138 .
but if @xmath186 is @xmath138 , then @xmath186 meets @xmath187 twice , impossible .
if @xmath186 is @xmath160 , then @xmath186 meets @xmath139 twice , impossible .
thus @xmath186 is a @xmath123-line . by the same reason ,
the right lower line of @xmath133 is not an @xmath121-line , and so a @xmath124-line .
thus vertices @xmath123 and @xmath124 are located as in the second of figure [ fig : caseii-2 - 1 - 1 ] . after locating @xmath150 and @xmath152 as in the third of figure [ fig : caseii-2 - 1 - 1 ] ,
consider the @xmath123-line @xmath143 and the @xmath124-line @xmath164 .
if @xmath143 crosses the @xmath121-line @xmath160 , then so does @xmath187 through the same triple crossing point on @xmath160
. then @xmath185 can not be @xmath36-sided .
hence @xmath143 , and then @xmath164 , go to @xmath135 .
then @xmath185 and @xmath188 can not be @xmath36-sided simultaneously .
\(e ) then @xmath134 and @xmath141 are @xmath36-sided .
see figure [ fig : caseii-2 - 1 - 2 ] . by the same argument as ( d ) , the right lower line of @xmath133 turns out to be a @xmath124-line . similarly , the upper line of @xmath149 is a @xmath124-line .
since both @xmath133 and @xmath149 are @xmath48-sided , vertex @xmath124 is located as in the second of figure [ fig : caseii-2 - 1 - 2 ] .
then @xmath150 is not @xmath36-sided , a contradiction .
this completes the proof of lemma [ lem : ii-2 - 4 ] .
[ lem : ii-2 - 2 ] the number of type ii triangles at @xmath121 is not two .
suppose that there are two type ii triangles at @xmath121 .
then the local configuration at @xmath121 is figure [ fig : casei-2](1 ) or ( 2 ) .
[ cl : ii-2 - 2 ] figure [ fig : casei-2](1 ) is impossible .
we claim that @xmath133 is @xmath48-sided .
assume that @xmath133 is @xmath36-sided .
then vertex @xmath123 is located , and the @xmath123-line @xmath132 is determined as in figure [ fig : caseii-2 - 2 - 1 ] . is @xmath36-sided ] suppose further that @xmath134 is @xmath48-sided .
if vertex @xmath39 is incident with @xmath134 , then @xmath134 can not be @xmath48-sided .
hence @xmath134 is incident with two more triple crossing points .
then the fourth line of @xmath134 is an @xmath121- , @xmath123- or @xmath124-line by property ( @xmath125 ) .
however , the existence of @xmath132 implies that it is neither an @xmath121- nor @xmath123-line .
thus the right lower line of @xmath134 is a @xmath124-line .
then the second of figure [ fig : caseii-2 - 2 - 1 ] is the only possible configuration for @xmath134 .
but this is impossible , because two lines meet at most once .
is @xmath36-sided ( continued ) ] thus we see that @xmath134 is @xmath36-sided , so vertex @xmath39 is located as in figure [ fig : caseii-2 - 2 - 1 - 1 ] .
then @xmath141 is @xmath48-sided , and the @xmath39-lines @xmath157 and @xmath158 , and thus @xmath156 , are determined .
but , @xmath149 can be neither @xmath36-sided nor @xmath48-sided , because the existence of @xmath132 disturbs an @xmath121-line and a @xmath123-line as above .
we have thus shown that @xmath133 is @xmath48-sided .
next , we claim that @xmath134 is @xmath48-sided .
assume not .
then vertex @xmath39 is located , and thus @xmath158 is determined .
also , @xmath141 is @xmath48-sided .
if @xmath141 is incident with @xmath135 , then the @xmath39-line @xmath156 is determined , and thus @xmath157 can not be drawn .
hence @xmath141 is incident with another triple crossing point as in the first of figure [ fig : caseii-2 - 2 - 1a ] , where an @xmath121- or @xmath124-line goes through .
if it is a @xmath124-line , then @xmath149 can not be @xmath36-sided .
hence it is an @xmath121-line , in particular , @xmath160 .
see the second of figure [ fig : caseii-2 - 2 - 1a ]
. then @xmath156 can not be drawn .
thus we have specified two @xmath48-sided faces @xmath133 and @xmath134 .
is @xmath36-sided ] now , @xmath141 is @xmath36-sided . by property ( @xmath125 ) ,
the right lower line of @xmath134 is an @xmath121- or @xmath124-line .
see the first of figure [ fig : caseii-2 - 2 - 1a - final0 ] .
if it is an @xmath121-line , then it is @xmath160 .
but this is impossible , because the right upper line of @xmath134 already meets @xmath160 .
thus the right lower line of @xmath134 is a @xmath124-line .
furthermore , if @xmath134 is incident with @xmath148 , then the situation is drawn as in the second of figure [ fig : caseii-2 - 2 - 1a - final0 ] . then @xmath132 is determined , and thus vertex @xmath123 is located .
however , @xmath139 can not be drawn . and
@xmath134 are @xmath48-sided ] thus @xmath134 is incident with a triple crossing point at its right .
see the first of figure [ fig : caseii-2 - 2 - 1a - final ] . and
@xmath134 are @xmath48-sided ( continued ) ] by property ( @xmath125 ) , an @xmath121- or @xmath123-line goes through the triple crossing point .
but it can not be an @xmath121-line by examining the ( @xmath36-sided ) face right above @xmath134 . after locating vertex @xmath123 ,
the @xmath123-line @xmath132 is determined
. then @xmath139 can not be drawn .
figure [ fig : casei-2](2 ) is impossible .
we claim that @xmath133 and @xmath149 are @xmath48-sided .
assume that @xmath133 is @xmath36-sided .
then vertex @xmath123 is located , and then the @xmath123-line @xmath132 is determined .
see the first of figure [ fig : caseii-2 - 3 ] .
is @xmath36-sided ] suppose further that @xmath134 is @xmath48-sided .
then @xmath134 is incident with either vertex @xmath124 , or vertex @xmath39 , or two more triple crossing points .
if @xmath134 is incident with vertex @xmath124 , then @xmath189 is determined , and then @xmath187 can not be drawn .
if @xmath134 is incident with vertex @xmath39 , then an @xmath121- or @xmath123-line appears at the triple crossing point where a @xmath124-line meets a @xmath39-line .
but this is impossible by the existence of @xmath132 as in the proof of claim [ cl : ii-2 - 2 ] .
if @xmath134 is incident with two more triple crossing points , then a similar argument to the proof of claim [ cl : ii-2 - 2 ] gives a contradiction again ( see figure [ fig : caseii-2 - 3 ] ) .
thus @xmath134 is @xmath36-sided .
then the situation is as in the proof of claim [ cl : i-2 ] , leading to a contradiction .
thus @xmath133 is @xmath48-sided . by the same argument
, @xmath149 is @xmath48-sided
. then @xmath134 and @xmath141 are @xmath36-sided . but
this is impossible , because there are a @xmath123-line and @xmath124-line meeting twice .
this completes the proof of lemma [ lem : ii-2 - 2 ] .
[ lem : ii-2 - 0 ] the number of type ii triangles at @xmath121 is not zero .
suppose that there is no type ii triangle at @xmath121 .
then the local configuration at @xmath121 is figure [ fig : casei-3](1 ) or ( 2 ) . in figure
[ fig : casei-3](1 ) , if @xmath133 or @xmath134 is @xmath36-sided , then we have a contradiction as in the proof of lemma [ lem : i0 ] .
thus both are @xmath48-sided . by the same reason , @xmath141 and @xmath149
are @xmath48-sided , a contradiction . in figure
[ fig : casei-3](2 ) , if @xmath133 or @xmath134 is @xmath36-sided , then we have a contradiction as above .
hence @xmath133 and @xmath134 are @xmath48-sided .
thus @xmath141 and @xmath149 are @xmath36-sided .
then the @xmath124-line @xmath164 meets the @xmath39-line @xmath63 twice , a contradiction .
[ prop : caseii ] case ( 2 ) is impossible .
this follows from lemmas [ lem : ii3 ] , [ lem : ii2 ] , [ lem : ii1 ] , [ lem : ii0 ] , [ lem : ii-2 - 4 ] , [ lem : ii-2 - 2 ] and [ lem : ii-2 - 0 ] . [ thm : k44 ]
@xmath24 does not admit a semi - regular drawing .
this follows from propositions [ prop : casei ] and [ prop : caseii ] .
let @xmath190 with @xmath191 . in this section , we show that if @xmath192 then @xmath0 does not admit a semi - regular drawing .
hereafter , we assume that @xmath191 and @xmath192 .
let @xmath116 and @xmath193 be the partite sets of @xmath0 . as before , we refer to a vertex of @xmath116 ( resp .
@xmath117 ) as a black ( resp .
white ) vertex .
any black vertex is incident with an @xmath121-line , @xmath123-line and @xmath124-line .
suppose that @xmath0 admits a semi - regular drawing .
fix such a drawing , denoted by @xmath0 again .
property ( @xmath125 ) holds .
that is , at each triple crossing point , an @xmath121-line , a @xmath123-line and a @xmath124-line meet .
let @xmath40 be the number of triple crossing points .
add a new vertex to each triple crossing point .
then we have a plane graph @xmath41 with @xmath194 vertices and @xmath195 edges .
since @xmath196 , the faces of @xmath41 are @xmath36-sided , except at most three faces , by lemma [ lem : faces ] .
we refer to a non - triangular face as an _
exceptional face_. more precisely , lemma [ lem : faces ] claims that either 1 .
@xmath41 has only one exceptional face , which is @xmath2-sided ; or 2 .
@xmath41 has just two exceptional faces , which are @xmath50-sided and @xmath48-sided , respectively ; or 3 .
@xmath41 has just three exceptional faces , which are @xmath48-sided . as before ,
a face of @xmath0 means that of @xmath41 .
let @xmath197 be the number ( counted with multiplicities ) of white vertices which are incident with exceptional faces .
then @xmath198 , because two white vertices are not adjacent in @xmath41 .
since a white vertex is not a cut - vertex of @xmath41 , a white vertex can not appear around one exceptional face twice .
[ lem : same - pair - exc ] two exceptional faces are not incident with the same pair of white vertices .
suppose that two exceptional faces @xmath184 and @xmath199 are incident with white vertices @xmath121 and @xmath123 , say .
then both @xmath184 and @xmath199 are @xmath48-sided , or one is @xmath48-sided and the other @xmath50-sided .
see figure [ fig : k3p - n6 ] . recall that any black vertex is incident with a @xmath124-line .
thus , in any case , we can not place @xmath124-lines .
recall that there are two types of triangles at a white vertex as shown in figure [ fig : triangle ] .
let @xmath200 or @xmath124 . at vertex @xmath121 ,
if a type i triangle is bounded by two @xmath121-lines and an @xmath201-line , then it is said to be of _ type _ i-@xmath201 .
see figure [ fig : typei ] .
furthermore , a type i-@xmath201 triangle is said to be _ good _ if the face sharing the @xmath201-line with the type i-@xmath201 triangle is @xmath36-sided .
otherwise , it is _ bad_. in particular , a bad type i triangle is adjacent to an exceptional face , which is referred to as its _ associated exceptional face_. triangle and a type i-@xmath124 triangle ] [ lem : typeix - ax ] let @xmath202 . if there is a good type i-@xmath201 triangle at vertex @xmath121 , then there is neither an exceptional face incident with both @xmath121 and @xmath203 , nor another good type i-@xmath201 triangle at @xmath121 . in particular
, the number of good type i triangles is at most two .
let @xmath133 be a good type i-@xmath124 triangle at @xmath121 .
then the face @xmath134 sharing the @xmath124-line with @xmath133 is @xmath36-sided , so vertex @xmath123 is located there .
suppose that an exceptional face @xmath184 is incident with vertex @xmath121 and @xmath123 .
then we have a similar situation to the proof of lemma [ lem : same - pair - exc ] .
hence we can not place @xmath124-lines .
the existence of another good type i-@xmath124 triangle is excluded by a similar argument .
see figure [ fig : k3p - case1a ] . here
, we can not place the @xmath124-lines going to the left upper black vertex and the left lower black vertex simultaneously .
triangles at @xmath121 ] [ lem : badi ] suppose that there is a bad type i triangle @xmath184 at vertex @xmath121 .
let @xmath205 be the associated exceptional face of @xmath184 .
if @xmath205 is @xmath40-sided ( @xmath206 ) , then @xmath205 is incident with at most @xmath207 white vertices .
we may assume that @xmath184 is a bad type i-@xmath123 triangle without loss of generality .
as shown in the first of figure [ fig : typei ] , the boundary of @xmath205 contains a sequence of @xmath208 the existence of this sequence forces @xmath205 to admit at most @xmath207 white vertices .
[ lem : badshare ] two bad type i triangles at vertex @xmath121 can not have the same associated exceptional face .
let @xmath133 and @xmath134 be bad type i triangles at @xmath121 whose associated exceptional faces coincide .
let @xmath205 be the common associated exceptional face .
when @xmath205 is @xmath48- or @xmath50-sided , the situation is as shown in figure [ fig : n4a ] , where labels @xmath123 and @xmath124 may be exchanged . - or @xmath50-sided exceptional face ] then we can not place @xmath124-lines as before .
if @xmath205 is @xmath2-sided , then the situation is as shown in figure [ fig : n4b ] .
similarly , we can not draw @xmath124-lines
. -sided exceptional face ] [ lem : badimore ] suppose that @xmath0 satisfies one of the following conditions . 1 .
@xmath0 has a @xmath50-sided exceptional face incident with exactly one white vertex and a @xmath48-sided exceptional face incident with at least one white vertex .
@xmath0 has three @xmath48-sided exceptional faces , only one of which is incident with no white vertex .
then there is at most one bad type i triangle at vertex @xmath121 .
let @xmath209 be a bad type i triangle at @xmath121 , and let @xmath210 be the associated exceptional face for @xmath211 .
by lemma [ lem : badi ] , we have @xmath212 . but
this contradicts lemma [ lem : badshare ] .
[ lem : bad3 ] there is at most three bad type i triangles at vertex @xmath121 .
suppose that there are four bad type i triangles at vertex @xmath121 .
since @xmath0 has at most three exceptional faces and each bad type i triangle is adjacent to an exceptional face , there exist two bad type i triangles whose associated exceptional faces coincide , contradicting lemma [ lem : badshare ] .
recall that an adjoint pair of type ii triangles at vertex @xmath121 is a pair of type ii triangles sharing an @xmath121-line fully
. see figure [ fig : k3p - case1b](1 ) . ]
[ lem : typeiipair ] suppose that @xmath0 satisfies one of the following conditions . 1 .
@xmath0 has a single @xmath2-sided exceptional face , which is incident with all white vertices .
2 . @xmath0 has a @xmath50-sided exceptional face incident with two white vertices and a @xmath48-sided exceptional face incident with a white vertex .
@xmath0 has three @xmath48-sided exceptional faces , each of which is incident with a white vertex .
then there is no adjoint pair of type ii triangles at vertex @xmath121 .
hence , if there is a type ii triangle at @xmath121 , then it shares an @xmath121-line fully with an exceptional face ( figure [ fig : k3p - case1b](2 ) ) .
suppose that there is an adjoint pair of type ii triangle at @xmath121 .
then @xmath184 , indicated in figure [ fig : k3p - case1b](1 ) , is not @xmath36-sided .
however , @xmath184 can not be an exceptional face from the assumption , a contradiction .
[ lem : adjointpair ] there is at most one adjoint pairs of type ii triangle at vertex @xmath121 . as in the proof of lemma [ lem : typeiipair
] , each adjoint pair of type ii triangles yields an exceptional face .
if two such pairs share the same exceptional face , then the exceptional face must be a @xmath2-sided face without a white vertex .
the situation is as shown in figure [ fig : k3p - pair ]
. then we can not draw @xmath124-lines to the left two black vertices from vertex @xmath124 .
hence the number of adjoint pairs of type ii triangles is no greater than the number of exceptional faces .
when @xmath0 has a @xmath2-sided exceptional face , we have the conclusion .
assume that @xmath0 has at least two , then two or three , exceptional faces .
suppose that there are two adjoint pairs of type ii triangle .
then these pairs correspond to distinct exceptional faces as above . in any case
, there exists an adjoint pair of type ii triangles which yields a @xmath48-sided exceptional face @xmath184 .
then the situation is as shown in figure [ fig : n2 ] .
-sided face ] this implies that both @xmath205 and @xmath213 are exceptional .
another adjoint pair of type ii triangles yields one more exceptional face .
thus @xmath0 would have @xmath48 exceptional faces , a contradiction .
[ lem : odd ] if @xmath109 is odd , then each white vertex is incident with an exceptional face , hence @xmath214 .
assume that only triangles appear at a white vertex , @xmath121 , say .
each triangle at @xmath121 is incident with either a @xmath123-line or a @xmath124-line .
moreover , such triangles appear alternatively around @xmath121
. hence @xmath109 must be even .
[ lem : even ] suppose that vertex @xmath121 is incident with only one exceptional face @xmath184 .
if @xmath184 is @xmath48-sided and incident with two white vertices , then @xmath109 is even .
we may assume that @xmath184 is incident with @xmath121 and @xmath123 .
then each triangle adjacent to @xmath184 at @xmath121 is incident with a @xmath124-line . by the same reason as the proof of lemma [ lem :
odd ] , @xmath109 is even .
[ lem : n6 ] @xmath215 .
let @xmath216 .
this happens only when @xmath0 has three @xmath48-sided exceptional faces , each of which is incident with two white vertices . in particular , a face incident with two adjacent triple crossing points is @xmath36-sided . by lemma [ lem : same - pair - exc ] , there is one exceptional face for each pair of @xmath121 , @xmath123 , @xmath124 .
we examine the local configuration at vertex @xmath121 . by lemmas [ lem : typeix - ax ] and [ lem : badi ]
, there is no type i triangle . by lemma [ lem : typeiipair ]
, each type ii triangle is adjacent to an exceptional face .
if two exceptional faces at @xmath121 share an @xmath121-line , then there are at most two type ii triangles .
this implies that @xmath111 , a contradiction .
( we remark that this situation can happen when @xmath217 . ) otherwise , there are at most four type ii triangles .
in fact , both sides of a type ii triangle can not be exceptional faces , because there is only one exceptional ( @xmath48-sided ) face for each pair of @xmath121 , @xmath123 , @xmath124 .
( see figure [ fig : k3p - case1b](2 ) . )
hence there are exactly four type ii triangles and two exceptional faces around @xmath121 , giving @xmath218 , a contradiction .
[ lem : n5 ] @xmath219 .
assume @xmath220 .
this happens only when @xmath0 has three @xmath48-sided exceptional faces @xmath221 , two of which are incident with two white vertices , the other to one white vertex . by lemma [ lem : same - pair - exc ] , we may assume that @xmath133 is incident with @xmath121 and @xmath123 , @xmath134 is incident with @xmath123 and @xmath124 , and @xmath141 is incident with @xmath123 or @xmath124 .
we examine the local configuration at @xmath121 . by lemma [ lem : badi ]
, there is no bad type i triangle .
assume that @xmath141 is incident with @xmath123 .
by lemma [ lem : typeix - ax ] , a good type i-@xmath124 triangle is impossible , and at most one good type i-@xmath123 triangle is possible . by lemma [ lem : typeiipair ]
, there are at most two type ii triangles , which are adjacent to @xmath133 .
hence we have @xmath111 , a contradiction . the case where @xmath141 is incident with @xmath124 is similar .
[ lem : n4 ] @xmath222 .
. then @xmath0 has at least two exceptional faces .
first , suppose that @xmath0 has a @xmath50-sided face @xmath184 and a @xmath48-sided face @xmath199 .
then both @xmath184 and @xmath199 are incident with two white vertices .
we may assume that @xmath184 are incident with @xmath121 , @xmath123 , and @xmath199 to @xmath123 , @xmath124 .
this case is handled by the same argument as in the second paragraph of the proof of lemma [ lem : n5 ] .
next , suppose that @xmath0 has three @xmath48-sided faces @xmath133 , @xmath134 , @xmath141 .
by lemma [ lem : same - pair - exc ] , there are five possibilities for three exceptional faces as shown in table [ tb : n4 ] , up to renaming . by lemmas [ lem : odd ] and [
lem : even ] , @xmath224 in any case .
.five possibilities [ cols="^,^,^,^",options="header " , ] the graph @xmath0 has @xmath225 vertices and @xmath104 edges .
then @xmath226 if @xmath227 , then @xmath228
. thus we obtain @xmath38 by lemma [ lem : alg - condition ] .
let @xmath229 .
then @xmath230 . if @xmath231 , then @xmath228 . if @xmath232 , then @xmath228 , except when @xmath233
for these cases , @xmath38 by lemma [ lem : alg - condition ] again . since @xmath234 is planar , @xmath235
. let @xmath236 .
we have @xmath237 .
if @xmath238 , or if @xmath239 and @xmath240 , then @xmath228 . for these cases , @xmath38 . since @xmath241 is not planar ( it contains @xmath30 as a subgraph ) and it admits a semi - regular drawing with one triple crossing as shown in figure [ fig : k331 ] , we have @xmath242 . ]
the remaining cases are when @xmath243 . if @xmath244 or @xmath245 , @xmath0 is planar .
therefore consider the case where @xmath232 and @xmath246 .
let @xmath116 and @xmath117 be the partite sets of @xmath0 with @xmath247 and @xmath248 elements , respectively , and let @xmath249 be the remaining vertex .
assume that @xmath0 admits a semi - regular drawing .
notice that no edge connecting @xmath249 and a vertex of @xmath117 contains a triple crossing point , since @xmath250 .
thus we obtain a semi - regular drawing of a complete bipartite graph @xmath251 by removing two edges between @xmath249 and @xmath117 .
then we have that @xmath252 , @xmath48 or @xmath2 from theorem [ thm : bipartite ] . in either case
, @xmath0 is not planar , since @xmath0 contains @xmath30 as a subgraph .
thus we have that @xmath253 by figure [ fig:4partite ] ( after removing the edge @xmath254 ) .
since @xmath255 by theorem [ thm : bipartite ] , we obtain that @xmath256 by figure [ fig:6111 ] ( after removing the edge @xmath254 ) .
in this paper , we require that two edges intersect at most once , and two edges with a common end - vertex do not intersect .
this is one natural standpoint in the study of the crossing number ( @xcite ) , but this might be so strong that most complete multipartite graphs do not admit semi - regular drawings .
if we relax it , then @xmath24 , for example , admits a semi - regular drawing . in general , for @xmath257 , we can define the @xmath109-fold crossing number for a graph @xmath0 to be the minimal number of @xmath109-fold crossing points over all drawings with only @xmath109-fold crossings .
clearly , theorem [ thm:5multi ] holds for the @xmath109-fold crossing number .
furthermore , if @xmath0 is a non - planar complete @xmath19-partite graph with @xmath258 , then we can show that @xmath0 does not admit a drawing with only @xmath109-fold crossings by similar arguments to those of sections [ sec : pre ] , [ sec:4partite ] and [ sec : tripartite ]
. it might be possible to determine the values of this invariant for complete bipartite graphs . | we introduce the triple crossing number , a variation of the crossing number , of a graph , which is the minimal number of crossing points in all drawings of the graph with only triple crossings .
it is defined to be zero for planar graphs , and to be infinite for non - planar graphs which do not admit a drawing with only triple crossings . in this paper , we determine the triple crossing numbers for all complete multipartite graphs which include all complete graphs . |
the hadron hose is a current - carrying wire used to instrument the decay volume of a conventional neutrino beam with a toroidal focusing field for a sign - selected meson beam .
conventional neutrino beams are tertiary beams , with the primary proton beam producing @xmath0 and @xmath1 meson secondaries in a target .
beamline elements downstream of the target , called horns @xcite , produce a toroidal magnetic field to sign- and momentum- select the mesons , focusing them toward an evacuated or helium - filled volume , where the mesons decay into the tertiary neutrino beam .
steel and earth shielding absorb remnant protons , mesons , and muons at the end of the decay volume ( see figure [ numinohose ] ) . in conventional neutrino beams
to date @xcite the mesons freely propagate through the decay volume .
the toroidal field of the hadron hose provides continuous focusing of the secondary meson beam throughout the decay volume .
mesons spiral around the hose wire and are drawn away from the decay volume walls where they might interact before decaying .
a pion orbit in the numi beamline with the hadron hose is shown in figure [ numihose ] .
the spiraling motion randomizes the decay angles of the mesons , which reduces the need to model detector and beamline acceptances in monte carlo calculations of the expected neutrino fluxes at the detectors .
toroidal magnetic fields have been used for plasma beams @xcite and the possibility of using toroidal fields to focus charged particle beams was noted by van der meer @xcite . in this article , we study the utility of such a focusing system for neutrino beams and the technical feasibility of placing such a wire in an evacuated region with high particle fluences . while we have developed this device for the neutrinos at the main injector ( numi ) project , such a system might be of utility at future neutrino facilities such as the jhf @xcite or conventional neutrino `` super beams '' @xcite .
this article procedes as follows : we discuss the focusing properties of the hose in sections [ numi ] - [ hose ] .
sections [ requirements]-[elecdesign ] present the mechanical and electrical design which was proposed for numi .
sections [ breakdown ] through [ radiological ] discuss some of the hardware studies to demonstrate that the hose wire can survive in the radiation field of a neutrino beamline without significant failure of wire segments .
section [ conclude ] concludes .
for numi , 120 gev protons will be extracted from the main injector @xcite and focused downward by 58 mrad , to strike a 0.94 m long graphite target .
the bunch length is @xmath2sec , and the cycle time @xmath3 sec .
the beamline is designed for @xmath4 protons / pulse and @xmath5 protons / year .
numi will have two focusing horns @xcite pulsed at 200 ka after the target .
the secondary hadrons enter a 675 m long , 1 m radius evacuated decay volume .
minos @xcite is a 2-detector neutrino experiment .
a 980 ton near detector measures the neutrino energy spectrum and rate produced at fermilab . a 5400 ton
far detector is located in the soudan mine in minnesota , 735 km from fermilab .
the numi beam can be adapted to produce a low ( @xmath6 gev ) , medium ( @xmath7 gev ) , or high ( @xmath8 gev ) energy neutrino event spectrum .
figure [ lespectra ] shows the expected spectra in the two detectors for the low - energy and high - energy beam options assuming no new physics resulting in neutrino disappearance . in figure
[ lespectra ] , the vertical axis is the number of expected charged current neutrino interactions expected per kiloton of detector mass per @xmath9 protons on target ( @xmath101 numi year ) .
the neutrinos in the peaks of the spectra of figure [ lespectra ] come from pions focused by the horns , whereas the neutrinos in the high energy tail come from poorly focused pions ( pions that travel through the necks of the horns ) .
the high energy tail is nonetheless important to the experiment because it provides us a control sample to demonstrate a region without oscillation effects .
oscillation with @xmath11 ev@xmath12 results in a depletion of neutrinos centered around 1.8 - 3.0 gev , while the spectrum above 10 gev is unchanged . ]
it is important to make the best prediction of the far - field energy spectrum to search for any kind of spectral distortions .
it is desirable to have the neutrino energy spectra at the two detector sites as similar as possible so that the near detector site strongly constrains the calculation of the spectrum at the far site .
any relative difference observed at the far detector is evidence for new physics such as neutrino oscillations . in practice ,
the two detector spectra are not identical : the vastly different solid angles subtended by the two detectors results in an energy - dependent acceptance difference .
this difference is evident in the neutrino spectra of figure [ lespectra ] . )
minos detectors , computed without ( above ) and with ( below ) the hadron hose .
the left plots are for the numi low - energy beam and the right are for the high energy beam . , title="fig:",width=264 ] )
minos detectors , computed without ( above ) and with ( below ) the hadron hose .
the left plots are for the numi low - energy beam and the right are for the high energy beam .
, title="fig:",width=264 ] the acceptance difference between the two detectors is accounted for in the `` far - to - near ratio '' , which is the factor by which the near detector spectrum must be multiplied to predict the far detector spectrum . it is defined as : @xmath13 where @xmath14 is the observed number of events in the @xmath15 energy bin in the near detector , and @xmath16 is the predicted number of events in the @xmath15 bin in the far in the absence of new physics .
uncertainties in the calculation of @xmath17 lead to systematic uncertainties in the predicted far detector spectrum which limit the ultimate reach of physics searches .
the calculation of @xmath17 is complicated by the extended length and finite aperture of the neutrino beam geometry .
if the neutrinos were produced from a point source , then @xmath17 could be estimated by @xmath18 , where @xmath19 ( @xmath20 ) is the distance from the target to the near ( far ) neutrino detector . considering that neutrino beamlines are an extended source
, one could weight this extrapolation factor by the pion lifetime along the length of the decay tunnel : @xmath21 where the integral is over the length of the decay tunnel ( 725 m in the case of the numi beamline ) and the substitution @xmath22 has been made .
.,title="fig:",width=264 ] .,title="fig:",width=264 ] however , even this estimate of the near - far extrapolation is a simplification : not all decaying pions produce neutrinos within the finite acceptances of the two detectors and not all pions are able to decay before interacting along the decay pipe walls .
furthermore , the correct relation between the pion and neutrino energy is @xmath23 where @xmath24 is the pion relativistic boost and @xmath25 is the decay angle of the pion . in fact , because low - momentum pions tend to enter the decay volume at wider angles and decay far upstream in the decay volume , while high - momentum pions tend to propagate much further , the acceptance of the near detector for high momentum neutrinos is larger than for low - momentum neutrinos .
the acceptance correction @xmath26@xmath27 given in equation [ rfneqn ] is incorrect in an energy - dependent manner because of the correlation in equation [ enrgycorrel ] .
a more detailed extrapolation of the measured near spectrum to the far detector requires a monte carlo calculation which tracks secondary hadrons produced in the target through the beamline optics .
a comparison of the naive calculation of equation [ rfneqn ] to a geant @xcite simulation of the numi beamline is shown in figure [ modelspread ] .
the hadron hose consists of a wire at the center of the decay tunnel and carries a 1000 a peak current pulse .
the current provides a toroidal magnetic field which continually focuses positive particles back toward the center of the decay pipe . a typical meson executes 2 - 3 orbits around the wire over the length of the decay pipe .
the hadron hose contributes three essential features .
first , @xmath0 s and @xmath1 s that otherwise diverge out to the decay pipe walls and interact before they decay are given a restoring force back to the decay pipe center .
thus , @xmath0s/@xmath1 s travel farther and have a greater chance to decay , so the neutrino event rate in the far detector is increased ( see table [ fluxes ] ) .
second , the pion decay distribution along the beamline direction more nearly follows a simple exponential , which reduces the difference in acceptance of the near detector to low - momentum _ vs. _ high - momentum pions .
thus , the extrapolation factor @xmath26@xmath27 more closely follows that given by the pion lifetime in equation [ rfneqn ] , as can be seen in figure [ modelspread ] .
third , the pion orbits effectively randomize the decay angle between the pion direction and the neutrino that hits the minos near and far detectors .
this randomization washes out the kinematic correlation between neutrino energy and decay angle in equation [ enrgycorrel ] , which otherwise always produces a softer spectrum in the near detector than the far detector .
the difference is especially visible at the high energy edge of the high energy beam in figure [ lespectra ] .
the randomization is a larger effect for the high energy beam because the lorentz boost at high energy otherwise increases the sensitivity to the decay angle in equation [ rfneqn ] .
the effect is also demonstrated in figure [ nearfarangles ] , which plots the ratio of neutrino energies in the near and far detector that would arise from the decay of a given pion in the low - energy or high - energy numi beam configurations .
.charged current event yields in the far minos detector in the low and high energy beam configurations , with and without hadron hose .
the peak refers to the regions @xmath286 and 30 gev , for low and high energy beams , while overall refers to @xmath29 gev .
the units are events per kiloton of far detector mass per @xmath30 protons on target . [
cols="<,^,^",options="header " , ]
we have investigated the potential impact of a new focusing system , the hadron hose , for conventional neutrino beams .
this system was developed for the numi beam at fermilab , and may be of benefit to future conventional super beams because of its increase in neutrino event yield and ability to control systematic uncertainties due to particle production in the target or imperfections in the rest of the neutrino beam elements .
we thank our numi and minos colleagues , particularly karol lang , doug michael , and stan wojcicki , for discussions and encouragement .
we thank ron dimelfi , john hall , nikolai mokhov , howie pfeffer , steve sansone , and sergei striganov for helpful consultations .
we acknowledge the valuable contributions of the university of texas physics department mechanical support shops , the fermilab particle physics , beams , survey / alignment , and mechanical support divisions , and in particular virgil bocean , cary kendziora , stephen pordes , bob webber , and james lackey .
this work was supported by the u.s .
department of energy , de - ac02 - 76ch3000 , de - fg03 - 93er40757 and de - fg02 - 91er40654 , the national science foundation , nsf / phy-0089116 , and the fondren foundation .
s. van der meer , cern yellow report cern-61 - 07 ( 1961 ) .
et al . _ ,
lett . * 9 * , 36 ( 1962 ) .
g.bernardini _ et al .
_ , _ proc .
intl conf . on elementary particles _ ,
sienna , vol . 1 ( 1963 ) 523 ; m. giesch _
_ , _ proc .
intl conf . on elementary particles _ , sienna , vol . 1 ( 1963 ) 536 ; c. angelini _ et al . _
( bebc collaboration ) , phys .
b179 * , 307 ( 1986 ) ; f. bergsma _
( charm collaboration ) , phys .
* b142 * , 103 ( 1984 ) ; f. dydak
_ et al . _
( cdhs collaboration ) , phys .
* b134 * , 281 ( 1984 ) . g. acquistapace _
_ , the west area neutrino facility , cern - ecp/95 - 14 ( 1995 ) ; p. astier _
( nomad collaboration ) , nucl .
b611 * , 3 ( 2001 ) ; a.g .
( chorus collaboration ) , _ proc .
xxxvi recontres de moriond _
l. ahrens _ et al . _
( bnl e734 ) , phys . rev . * d31 * , 2732 ( 1985 ) ; b. blumenfeld _
( bnl e776 ) , phys .
* 19 * , 2237 ( 1989 ) .
( k2k collaboration ) , phys .
b511 * , 178 ( 2001 ) .
taylor , _ et al . _
( fnal e338 collaboration ) , phys . rev .
* d28 * , 2705 ( 1983 ) .
brucker , _ et al . _
( fnal e564 collaboration ) , phys .
rev . * d34 * , 2183 ( 1986 ) ; i.e. stockdale , _ et al .
_ ( ccfr collaboration ) , phys . rev
. lett . * 52 * , 1384 ( 1985 ) ; n. ushida , _ et al . _
( fnal e531 collaboration ) , phys .
lett . * 57 * , 2897 ( 1986 ) .
stepa , _ motion of a relativistic charged particle in the magnetic field produced by a constant cylindrical current of a rarefied plasma _ ,
soviet phyics techn .
physics * 4 * , 1237 ( 1960 ) . ,
f. naterf . * 14a * , 47 ( 1959 ) . s. van der meer , _ the beam guide _ ,
cern-62 - 16 , april 17 , 1962 .
e. regenstreif , _ contributions to the theory of the beam guide _ , cern 64 - 41 , september 24 , 1964 .
_ , `` the jhf - kamioka neutrino project '' , letter of intent , kek report 2001 - 4 , icrr report 477 - 2001 - 7 , triumf report tri - pp-01 - 05 , june , 2001 . v. barger et al .
, hep - ph/0103052 ( fermilab - fn-703 ) , addendum to report fn-692 to the fermilab directorate , march 5 , 2001 . v. d. barger , d. marfatia and k. whisnant , `` neutrino superbeam scenarios at the peak , '' in _ proc . of the aps / dpf / dpb summer study on the future of particle physics ( snowmass 2001 ) _ ed .
r. davidson and c. quigg , arxiv : hep - ph/0108090 .
j. j. gomez - cadenas _ et al .
_ [ cern working group on super beams collaboration ] , physics potential of very intense conventional neutrino beams , arxiv : hep - ph/0105297 .
j. hylen _ et al .
_ `` conceptual design for the technical components of the neutrino beam for the main injector ( numi ) , '' fermilab - tm-2018 , sept . , 1997 .
the minos collaboration , `` the minos detectors technical design report , '' fermilab numi - l-337 , oct .
1998 , s. wojcicki , spokesman . g. ambrosini _ et al . _
( spy collaboration ) , phys .
* b420 * ( 1998 ) 225 ; phys . lett .
* b425 * ( 1998 ) 208 ; eur .
j. * c10 * ( 1999 ) 605 .
m. bonesini , a. marchionni , f. pietropaolo , and t. tabarelli de fatis , `` on particle production for high energy neutrino beams , '' eur .
j. * c20 * , 13 - 27 ( 2001 ) .
mokhov , `` the mars monte carlo '' , fermilab fn-628 ( 1995 ) ; o.e .
et al . _ ,
proc . of the third and fourth workshops on simulating accelerator radiation environments ( sare3 and sare4 ) , fermilab - conf-98/043(1998 ) and fermilab - conf-98/379(1998 ) .
malensek , `` empirical formula for thick target particle production , '' fermilab fn-341 , oct . 1981 .
, cern program library , w5013 ( 1994 ) .
a. ferrari and p.r .
sala , fluka monte carlo, atlas internal note atl - phys-97 - 113 , proc .
of the workshop on nuclear reaction data and nuclear reactors physics , design and safety , ictp , trieste , italy 1996 .
publ . by world
scientific , a.gandini , g.reffo , eds .
l. nagel , `` spice2 , a computer program to simulate semiconductor circuits , '' memo - no .
erl - m520 , university of california , berkeley , may , 1975 ; d. o. pederson , `` a historical review of circuit simulation , '' ieee transactions on circuits and systems , vol .
cas-31 , no .
1 , january 1984 ; t. quarles , a. r. newton , d. o. pederson , a. sangiovanni - vincentelli , `` spice 3b1 user s guide , '' university of california , berkeley , april , 1987 .
e. kuffeland and m. abdullah , `` high voltage engineering '' , 1970 , pergamon press .
icru report 31 `` average energy required to produce an ion pair '' ( international commission on radiation units and measurements , 1 may , 1979 ) . | we have developed a new focusing system for conventional neutrino beams .
the `` hadron hose '' is a wire located in the meson decay volume , downstream of the target and focusing horns .
the wire is pulsed with high current to provide a toroidal magnetic field which continuously focuses mesons .
the hose increases the neutrino event rate and reduces differences between near - field and far - field neutrino spectra for oscillation experiments .
we have studied this device as part of the development of the neutrinos at the main injector ( numi ) project , but it might also be of use for other conventional neutrino beams .
-.25 cm accelerator , neutrino , beamline , focusing 41.75.l , 52.75.d , 29.27 , 14.60.p , 52.55.e , |
the theory of quantum entanglement has provided a useful and novel perspective for the analysis of correlations and quantum phase transitions in interacting many body systems @xcite . at the same time
, it is essential for determining the capability of such systems for performing different quantum information tasks @xcite .
more recently , a general theory of quantum resources , similar to that of entanglement but based on the degree of coherence of a quantum system with respect to a given reference basis , was proposed @xcite .
thus , entanglement and coherence provide a means to capture the degree of quantumness of a given quantum system .
in particular , spin systems constitute paradigmatic examples of strongly interacting many body systems which enable to study in detail the previous issues , providing at the same time a convenient scalable scenario for the implementation of quantum information protocols .
interest on spin systems has been recently enhanced by the significant advances in control techniques of quantum systems , which have permitted the simulation of interacting spin models with different type of couplings by means of trapped ions , josephson junctions or cold atoms in optical lattices @xcite .
accordingly , interacting spin systems have been the object of several relevant studies .
entanglement and discord - type correlations @xcite in spin chains with heisenberg couplings under uniform fields were intensively investigated , specially for spin @xmath7 systems @xcite .
the effects of non - uniform fields have also received attention , mostly for spin @xmath7 systems @xcite , although some results for higher spins in non - uniform fields are also available @xcite .
the aim of this work is to analyze in detail the effects of a non - uniform magnetic field on the entanglement and coherence of a spin-@xmath1 pair interacting through an @xmath0 coupling , both at zero and finite temperature .
we examine the interplay between the non - uniform magnetic field and temperature and their role to control quantum correlations .
we also study the critical behaviour and the development of different phases as the spin increases , when the field , temperature , and coupling anisotropy are varied .
analytical rigorous results are also provided . in particular
, the @xmath5 phase diagram will be characterized by ground states of definite magnetization @xmath8 , all reachable through non - uniform fields for any value of the couplings , with entanglement decreasing with increasing @xmath9 .
special critical points will be discussed . on the other hand , the limit temperature for entanglement
will be shown to depend . for any value of @xmath1 , only on the difference between the field applied at each spin , leading to a thermal phase diagram characterized by a separability stripe in field space .
finally , we will analyze the relative entropy of coherence @xcite in the standard basis , which coincides here exactly with the entanglement entropy at @xmath5 but departs from entanglement as @xmath6 increases .
the model and results are presented in sections ii
iv , starting with the basic spin @xmath7 case and considering then the @xmath10 and the general spin-@xmath1 cases .
conclusions are finally given in section v.
we consider a spin @xmath1 pair interacting through an @xmath0-type coupling , immersed in a transverse magnetic field @xmath11 not necessarily uniform .
the hamiltonian can be written as @xmath12 where @xmath13 ( @xmath14 ) denote the ( dimensionless ) spin operators at site @xmath15 and @xmath16 , @xmath17 the exchange couplings , with @xmath18 the anisotropy ratio .
this hamiltonian commutes with the total spin along the @xmath19 axis , @xmath20 , having then eigenstates with definite magnetization @xmath8 along @xmath19 . without loss of generality
, we will set in what follows @xmath21 , as its sign can be changed by a local rotation of angle @xmath22 around the @xmath19 axis of one of the spins , which will not affect the energy spectrum nor the entanglement of its eigenstates .
the ferromagnetic ( fm ) case @xmath23 , @xmath24 is then equivalent to @xmath21 , @xmath24 .
we also remark that a hamiltonian with an additional dzyaloshinskii - moriya coupling along @xmath19 @xcite , @xmath25 , can be transformed back exactly into a hamiltonian ( [ h1 ] ) with @xmath26 , by means of a rotation of angle @xmath27 around the @xmath19 axis at the second spin @xcite
. hence , its spectrum and entanglement properties will also coincide _ exactly _ with those of eq .
( [ h1 ] ) for @xmath28 .
we first review the @xmath29 case , providing a complete study with analytical results and including coherence in the standard basis , which allows to understand more easily the general spin @xmath1 case , considered in the next subsections . entanglement and
discord - type correlations under non homogeneous fields in a spin @xmath7 pair were studied in @xcite for an @xmath30-type coupling , in @xcite for an isotropic coupling , in @xcite for an @xmath0 coupling and in @xcite for an @xmath31 coupling .
using qubit notation , the eigenstates of the hamiltonian ( [ h1 ] ) for @xmath29 are the separable aligned states @xmath32 and @xmath33 , with magnetization @xmath34 and energies @xmath35 and the entangled @xmath36 states @xmath37 , with energies @xmath38 and @xmath39 .
the concurrence @xcite of these states is given by @xmath40 and is a decreasing function of @xmath41 .
their entanglement entropy , @xmath42 with @xmath43 the reduced state of one of the spins , can then be obtained as @xmath44 and is an increasing function of @xmath45 . in the uniform case @xmath46 ,
@xmath47 become the bell states @xmath48 and @xmath49 . through a non - uniform field it is then possible to tune the entanglement of the @xmath36 eigenstates , decreasing it by applying a field difference . on the other hand
, such difference also decreases the energy @xmath50 of @xmath51 and increases that of @xmath52 , without affecting that of the aligned eigenstates if the average field is kept constant , enabling to have the entangled state @xmath53 as a non - degenerate ground state ( gs ) for any value of @xmath16 or @xmath17 . a similar effect
can be obtained by increasing @xmath17 , which increases the gap between the entangled and the aligned states , in this case without affecting their concurrence .
( [ e1])([e0 ] ) then lead to the phase diagrams of fig .
[ f1 ] . for clarity
we have considered the whole field plane , although the diagrams are obviously symmetric under reflection from the @xmath54 line ( and spectrum and entanglement also from the @xmath55 line ) .
the gs will be either the entangled state @xmath53 ( red sector ) or one of the aligned states ( @xmath56 if @xmath57 or @xmath58 if @xmath59 , white sectors ) , with @xmath53 a non - degenerate gs ( @xmath60 ) if and only if @xmath61 this equation is equivalent to the following conditions : @xmath62 which show that the borders of the entangled sector are displaced hyperbola branches .
top : antiferromagnetic ( afm ) case @xmath63 ( left ) and ferromagnetic ( fm)-type case @xmath64 ( right ) .
bottom : fm cases @xmath65 ( left ) and @xmath66 ( right ) . ] in the afm case @xmath67 , the diagram has the form of the top left panel . here
the gs is entangled at zero field and if one of the fields is sufficiently weak ( @xmath68 ) the gs remains entangled for any value of the other field .
however , in the fm case @xmath24 two distinct diagrams can arise ( top and bottom right panels ) , separated by the limit diagram of the bottom left panel ( @xmath69 ) .
if @xmath70 ( top right ) , the system is still entangled at zero field but now if one of the fields is sufficiently weak ( @xmath71 , dashed vertical lines ) entanglement is confined to a finite interval of the other field .
control of just one field then allows to switch entanglement on and off for any value of the other field . on the other hand ,
if @xmath72 , the gs is aligned for _ any _ uniform non - zero field , and @xmath53 becomes gs only above a _
value of the field difference , @xmath73 ( eq.([c0 ] ) ) , within the limits determined by eqs .
( [ c0])([c2 ] ) .
these limits imply that the sign of the field at each site must be _ different _ , as seen in the bottom right panel .
hence , gs entanglement is in this case switched on ( rather than destroyed ) by field application , provided it has opposite signs at each spin .
in addition , a gs transition between the aligned states @xmath58 and @xmath56 takes place at the line @xmath55 for @xmath74 , with the gs degenerate in this interval along this line . the gs concurrence for the same cases of fig .
[ f1 ] is depicted on fig.[f2 ] . as seen in the top panels ,
the maximum @xmath75 is reached for @xmath76 provided @xmath77 and @xmath78 . for @xmath72 ,
the maximum value is @xmath79 , attained at the edges @xmath80 of the entangled sector . .
top : @xmath63 ( left ) and @xmath81 ( right ) .
bottom : @xmath65 ( left ) and @xmath82 ( right ) . ]
let us now consider a finite temperature @xmath6 . as @xmath6 increases from 0 , an entangled gs will become mixed with other excited states , leading to a decrease of the entanglement which will vanish beyond a limit temperature .
however , if the gs is separable , the thermal state can become entangled for @xmath3 ( below some limit temperature ) due to the presence of entangled excited states , implying that the entanglement phase diagram for @xmath3 may _ differ _ from that at @xmath5 even for low @xmath6 . in the present case the thermal state @xmath83 , with @xmath84 the partition function and @xmath85 , has in the standard basis the form @xmath86 where @xmath87 , @xmath88 and @xmath89 .
its concurrence @xcite is then given by @xmath90={\textstyle 2z^{-1}{\rm max}[\frac{j}{\delta}e^{\beta j_z/4}\sinh\frac{\beta \delta}{2}-e^{-\beta j_z/4},0]}\,.\end{aligned}\ ] ] thus , for @xmath3 @xmath91 is entangled iff @xmath92 eq .
( [ jzc ] ) implies a limit temperature for entanglement that will depend on @xmath17 , @xmath16 and @xmath2 only .
it also implies a _
threshold value _ of @xmath17 for entanglement at _ any _ @xmath3 , @xmath93 hence , it is always possible to entangle the thermal state by increasing @xmath17 , since it will effectively cool down the system to the state @xmath53 , as previously stated .
the same effect occurs if the field difference @xmath2 is increased .
the l.h.s . of eq .
( [ jzc ] ) is an _ increasing _ function of @xmath94 and hence of @xmath2 for any @xmath3 and @xmath17 , so that at any @xmath3 there will also exist a _ threshold value _
@xmath95 of the field difference @xmath2 above which the thermal state will become entangled : @xmath96 eq . ( [ hcz ] ) gives rise to a _ separability stripe _ @xmath97 , as depicted in fig .
here @xmath98 , with @xmath99 and @xmath100 the inverse of the increasing function @xmath101 ( @xmath102 ) .
pair at @xmath103 . left : afm case @xmath63 .
right : fm case @xmath64 .
red sectors indicate entanglement .
the whole plane remains entangled for @xmath104 if @xmath105 and @xmath106 if @xmath107 . above these temperatures ,
a separability stripe @xmath108 arises . for @xmath109
the separability stripe arises for any @xmath3 . ]
( [ hcz ] ) implies that the thermal entanglement phase diagram in the field plane _ differs _ from the @xmath110 phase diagram even for small temperatures @xmath3 , as it is determined just by the field difference @xmath2 .
entanglement will be turned on in @xmath5 separable sectors outside the stripe as soon as @xmath6 becomes finite .
in particular , for @xmath77 , eq.([jzc ] ) shows that in contrast with the @xmath5 case , the _ whole _
@xmath111 plane becomes entangled for @xmath112 , with @xmath113 determined by @xmath114 such that @xmath115 if @xmath116 .
the separability stripe arises then for @xmath117 . for @xmath118 , @xmath119 while if @xmath120 , @xmath121 . however , for @xmath72 a separability stripe will be present for all @xmath3 , with @xmath122 for @xmath123 .
the thermal phase diagram in the field plane is then characterized , for any value of @xmath17 , by a separability stripe whose width increases with increasing @xmath6 , and vanishes for @xmath77 if @xmath116 . and @xmath124 at finite temperature .
top panels : afm case @xmath63 ( left ) and fm case @xmath64 at temperature @xmath103 .
bottom panels : same diagrams for the fm cases @xmath65 ( left ) and @xmath66 ( right ) at @xmath125 . ] at @xmath126 ( a ) , @xmath127 ( b ) and @xmath128 ( c ) .
the reentry of entanglement for @xmath3 in case ( b ) is clearly seen , with the limit temperature for entanglement independent of the field if @xmath54 ( cases a , b ) and increasing with increasing values of @xmath2 ( case c ) , as follows from eq .
( [ jzc ] ) .
in contrast , the coherence remains non - zero @xmath129 @xmath6 , decreasing uniformly as @xmath130 for high @xmath6 ( eq . [ cha ] ) and approaching the entanglement entropy ( [ s ] ) for @xmath131 .
right : entanglement of formation @xmath132 ( solid lines ) and relative entropy of coherence ( dashed lines ) as a function of the field difference @xmath133 at fixed temperature @xmath103 for @xmath134 ( a , e ) , @xmath135 ( b , f ) , @xmath136 ( c , g ) and @xmath137 ( d , h ) , at @xmath54 . as @xmath138 increases , all curves coalesce and become independent of @xmath17 , approaching the entanglement entropy of the gs @xmath53 determined by eq .
( [ con1 ] ) . ]
the thermal concurrence is shown in fig .
it is verified that it is strictly zero just within the separability stripe ( [ hcz ] ) , becoming small but non - zero in the @xmath5 separable regions outside it ( dark blue in fig .
nonetheless , such reentry of entanglement for @xmath3 can become quite noticeable in some cases , as seen in the top panel of fig .
it is also verified that through non - uniform fields it becomes possible to preserve entanglement up to temperatures higher than those in the uniform case ( which lies at the center of the separability stripe ) , as also seen in fig .
[ f5 ] . and @xmath124 at finite temperatures , for the same cases of fig .
top panels : afm case @xmath63 ( left ) and fm case @xmath64 for @xmath103 .
bottom panels : fm cases @xmath65 ( left ) and @xmath66 ( right ) for @xmath139 . ] we now analyze the coherence of the thermal state ( [ r12 ] ) with respect to the standard product basis @xmath140 where @xmath141 is the von neumann entropy and @xmath142 its diagonal part in the previous basis . it is a measure of the strength of the off - diagonal elements in this basis , and would obviously vanish if @xmath143 .
it will be here driven just by the coefficient @xmath144 in ( [ r12 ] ) .
a series expansion of ( [ ch ] ) for @xmath145 in ( [ r12 ] ) leads in fact to @xmath146 .
the exact expression is @xmath147 , where @xmath148 . in the zero temperature limit ,
@xmath149 vanishes while @xmath150 and hence @xmath151 become the _ entanglement entropy _
@xmath132 of the gs , eq . ( [ s ] ) , since the standard basis is here the schmidt basis for @xmath53 .
however , for @xmath3 , @xmath151 becomes everywhere non - zero due to the non - vanishing weight of the entangled states @xmath152 , as seen in figs .
[ f5 ] and [ f6 ] .
in fact , for @xmath153 $ ] , a series expansion leads to the asymptotic expression @xmath154}\,,\label{cha}\end{aligned}\ ] ] showing that it ultimately decreases uniformly as @xmath155 in this limit .
it then exhibits a reentry for @xmath3 in all @xmath5 separable sectors , as seen in figs .
[ f5 ] and [ f6 ] . as previously mentioned , by applying sufficiently strong opposite fields at each site it is possible to effectively `` cool down '' the thermal state @xmath91 at any @xmath3 , bringing it as close as desired to the entangled state @xmath156 .
this behaviour is shown in the bottom panel of fig .
it is seen that the entanglement of formation @xmath132 obtained from the thermal concurrence @xmath45 by the same expression ( [ s ] ) @xcite , and the relative entropy of coherence , initially different and dependent on @xmath17 , merge for increasing values of @xmath2 , approaching a common @xmath17 independent limit which is the entanglement entropy @xmath132 of the pure state state @xmath53 . the vanishing difference between @xmath132 and @xmath157 for high
@xmath2 is a clear signature that @xmath91 has become essentially pure .
we now consider the @xmath10 case .
the behaviour is essentially similar to that for @xmath29 , the main difference being the appearing of an
_ intermediate @xmath34 magnetization step _ in the @xmath5 diagrams , between the entangled @xmath36 gs and the aligned separable @xmath159 states .
this effect leads to an entanglement step since the @xmath34 gs is less entangled than the @xmath36 gs . using now the notation @xmath160 for the states of the standard basis , with @xmath161 the eigenvalues of @xmath162 , the gs of the spin @xmath158 @xmath0 pair can be one of the @xmath163 aligned states @xmath164 , one of the @xmath165 states , which will be of the form @xmath166 with @xmath167 and @xmath168 , and one of the @xmath36 states , of the form @xmath169 all coefficients are independent of @xmath170 , but those of @xmath171 depend now on @xmath17 . for @xmath120 they can be still written down concisely : @xmath172 , @xmath173 .
their energies are @xmath174 the border of the @xmath5 entangled region , determined by that between the @xmath163 and @xmath165 gs , @xmath175 , is then given again by eqs.([c0])([c2 ] ) with @xmath176 , @xmath177 .
the gs phase diagrams have then the same forms as those of fig . [ f1 ] except for the previous rescaling and the magnetization step .
the @xmath120 case is shown in fig .
[ f7 ] .
pair with @xmath120 as a function of the applied fields at each spin .
colored sectors on the left indicate entanglement .
bottom : the negativity ( left ) and relative entropy of coherence ( right ) of the thermal state of the spin @xmath158 pair at temperature @xmath178 as a function of the applied fields . ] as entanglement measure , we will now use the negativity @xcite , a well - known entanglement monotone which is computable in any mixed state .
it is minus the sum of the negative eigenvalues of the partial transpose @xcite @xmath179 of @xmath91 : @xmath180 a non - zero negativity implies entanglement , whereas for mixed states , the converse is not necessarily true ( except for two qubit states @xcite or special states ) , vanishing for bound entangled states . nonetheless it is normally used as an indicator of useful entanglement . for pure states it reduces to a special entanglement entropy @xcite , being a function of the one - spin reduced state @xmath181 ( or @xmath182 , isospectral with @xmath181 for a pure state ) : @xmath183 $ ] .
it is then non - zero iff the state is entangled .
its maximum value for a spin @xmath1 pair is @xmath184 .
we then obtain @xmath185 , @xmath186\\ & = & { \textstyle\frac{1+\sqrt{2 } \sum\limits_{\nu=\pm}\sqrt{1+\eta(\eta+\nu\sqrt{2+\eta^2}}}{2(2+\eta^2 ) } } \;\;\;\;\;(j_z=0)\,.\end{aligned}\ ] ] they are decreasing functions of @xmath168 , reaching for @xmath187 the value @xmath188 for @xmath165 ( maximum value for schmidt rank @xmath189 ) and @xmath190 for @xmath36 .
the negativity of the @xmath36 gs depends now on @xmath17 , reaching the maximum @xmath191 for @xmath192 .
it is verified in fig .
[ f7 ] that the @xmath5 phase diagram and entanglement for @xmath120 is similar to that for @xmath29 except for the @xmath34 magnetization and negativity steps .
remarkably , the finite temperature negativity diagram _ is again characterized by a separability stripe _
@xmath108 for @xmath193 , with the boundary of the non - zero negativity sector _ independent of @xmath170 _ ( as demonstrated in the next section ) . at @xmath120
the stripe emerges for @xmath194 ( root of the critical equation @xmath195 ) , with the whole field plane entangled for @xmath196 .
it should be also mentioned that the @xmath5 negativity step gives rise to a negativity `` valley '' for low finite @xmath6 due to the convexity of @xmath197 , as will be seen in the next section .
the relative entropy of coherence in the standard basis behaves in the same way as before .
it approaches the gs entanglement entropy for @xmath131 , while for @xmath198 , it decreases uniformly at leading order , becoming , for @xmath120 , @xmath199}\,.\label{cap1}\ ] ]
let us finally consider the main features of the general spin @xmath1 @xmath0 pair in a non - uniform field .
the gs phase diagram remains similar to the previous cases , but now with _ @xmath200 magnetization steps _ , from @xmath36 up to @xmath201 .
these steps originate @xmath200 steps in the @xmath5 entanglement and negativity , since they decrease with increasing @xmath9 .
this behaviour can be seen in the top panels of fig .
[ f8 ] for an @xmath202 pair .
+ pair with @xmath120 as a function of the applied fields at each spin .
colored sectors on the left indicate entanglement , with the number denoting the magnetization of the state .
bottom : the negativity of the spin 2 pair at temperature @xmath139 ( left ) and @xmath203 ( right ) . ] the border of the entangled region in the field plane is determined by that between the aligned gs with @xmath204 and the entangled gs with @xmath205 . remarkably , _
it is the same as that for @xmath29 , eq .
( [ c0 ] ) _ , with the rescaling @xmath206 , @xmath207 : @xmath208 the border are then the hyperbolas
( [ c1])([c2 ] ) with the previous scaling and give rise to the same possibilities depicted in fig . [ f1 ] , with the additional inner magnetization steps .
+ _ proof : _ considering first @xmath209 , the energies of the @xmath210 aligned state @xmath211 and the lowest @xmath212 state , which is @xmath213 , with @xmath214 and @xmath168 , are @xmath215 the condition @xmath216 leads then to eq .
( [ crits ] ) . if @xmath217 , the result is similar with @xmath170 replaced by @xmath218 and @xmath219 by @xmath220 .
pair with @xmath221 as a function of the applied fields at each spin .
colored sectors indicate an entangled gs , with the color and number identifying the distinct magnetizations .
bottom : the negativity ( left ) and relative entropy of coherence ( right ) of the spin 2 pair for @xmath221 at temperature @xmath222 . ] in fig .
[ f9 ] we plot an example for @xmath202 of the interesting fm case @xmath72 , where the gs is fully aligned for any uniform field @xmath76 , as in the @xmath29 case , with a transition @xmath223 to @xmath210 at @xmath224 .
however , _ it can again be entangled with a non - uniform field _ , by applying _
opposite _ fields at each spin .
( [ crits ] ) implies that in this case gs entanglement will arise for @xmath225 within the limits ( hyperbolas ) determined by eq .
( [ crits ] ) , which entail that no entangled gs will arise for fields of equal sign if @xmath72 , as in the @xmath29 case . moreover , the edges of the @xmath5 entangled sector , @xmath226 are actually _ critical points _ in which @xmath227 distinct gs s , corresponding to all magnetizations @xmath228 , _ coalesce and become degenerate _ , as verified in the top left panel of fig . [ f9 ] . at these points ,
their common energy is @xmath229 independent of @xmath16 and @xmath8 .
their entanglement decreases , however , with @xmath8 , as seen in the top right panel through the negativity . along the line @xmath55 a gs transition from the lowest non - degenerate @xmath36 state to the aligned states @xmath230 ( degenerate along this line ) occurs at @xmath231 , although precisely at these points the gs becomes @xmath227-fold degenerate .
actually the transition region with intermediate gs magnetizations @xmath232 is rather narrow in the @xmath233 field plane , as seen in the top left panel , collapsing at the critical points .
a final comment is that the maximum gs entanglement of a spin @xmath1 @xmath0 pair is reached at the @xmath36 gs and depends on @xmath17 for @xmath234 .
for @xmath72 it is reached at the previous critical points ( fig .
[ f9 ] ) , while for @xmath77 it is reached along the line @xmath54 ( fig .
[ f8 ] ) . in the uniform afm case @xmath69
, the @xmath36 eigenstate will be _ maximally entangled _ for @xmath54 , leading to maximum negativity @xmath184 , while for @xmath235 , the @xmath36 negativity will be smaller and proportional to @xmath236 for large @xmath1 , due to a gaussian profile of width @xmath237 of the expansion coefficients in the standard basis @xcite .
we also mention that some internal magnetization steps may disappear for large @xmath238 and small @xmath2 . as @xmath6 increases ,
the @xmath5 negativity steps become initially negativity _ valleys _ , as clearly seen in the bottom left panel of fig .
[ f8 ] , since convexity of @xmath197 implies that its value for the mixture of two entangled states will be smaller than the average negativity of the states .
these valleys are rapidly smoothed out as @xmath6 increases further . on the other hand ,
it is also seen in figs .
[ f8][f9 ] that entanglement diffuses outside the @xmath5 entangled region as @xmath6 increases , covering initially the whole field plane for @xmath77 and the whole plane outside the stripe @xmath239 for @xmath72 , although the negativity will be small in the @xmath5 aligned sectors .
a striking feature for finite temperatures is the _ persistence of a separability stripe _ @xmath108 in the field plane when considering the thermal entanglement , as seen in the bottom right panel of fig .
[ f8 ] for @xmath120 , where the stripe emerges for @xmath240 , and in the bottom left panel of fig .
[ f9 ] for @xmath241 , where the stripe is present @xmath129 @xmath3 .
this result will now be shown to hold for arbitrary spin , following the arguments of @xcite for @xmath0 systems in uniform fields .
_ lemma 1 .
the limit condition for entanglement and non - zero negativity of a spin @xmath1 pair with an @xmath0 coupling in a non - uniform transverse field at temperature @xmath3 , depends only on the field difference @xmath242 .
this result applies also to any coupling independent of the field that commutes with the total spin along @xmath19 ( @xmath243=0 $ ] ) .
_ proof : _ we first rewrite the hamiltonian of a spin @xmath1 pair in a non - uniform field as @xmath244 where @xmath245 denotes the ( field independent ) interaction between the spins , assumed to satisfy @xmath246=0 $ ] ( @xmath20 ) . the first term in ( [ h2 ] ) is the uniform field component and commutes with the rest of the hamiltonian .
consequently , the thermal state for average field @xmath247 , @xmath248 , can be written as @xmath249 where @xmath250 depends just on @xmath242 and commutes with @xmath251 .
( [ r0 ] ) implies that @xmath252 will be separable , i.e. , a convex combination of product states @xcite , _ if and only if @xmath250 is separable _ : if @xmath253 , with @xmath254 and @xmath255 local mixed states , then @xmath256 with @xmath257 also local mixed states , so that it is separable as well . similarly , @xmath252 separable implies @xmath250 separable ( @xmath258 ) .
hence , the limit condition for exact separability depends only on @xmath242 .
let us now consider the negativity .
the non - zero matrix elements of @xmath252 are @xmath259 .
its partial transpose will then have matrix elements @xmath260 , such that it can also be written as @xmath261 although @xmath262 will no longer commute with @xmath251 , @xmath263 will be positive definite ( i.e. , with positive eigenvalues ) if and only if @xmath262 is positive definite , since @xmath264 is positive definite and @xmath265 , @xmath266 are positive .
this result demonstrates the lemma for the negativity .
more explicitly , the onset for non - zero negativity occurs when the lowest eigenvalue of @xmath267 becomes negative , implying a vanishing eigenvalue at the onset , i.e. , @xmath268=0 $ ] .
but eq . ( [ t2 ] ) implies @xmath269= ( \frac{z_0}{z_h})^{(2s+1)^2}{\rm det } [ \rho_{12}^{t_2}(0)]$ ] ( as @xmath270 ) , so that the critical conditions at @xmath271 and @xmath224 are equivalent .
in addition , for an @xmath0 coupling as well as for any coupling invariant under permutation of the spins , the limit condition will obviously depend only on the _ absolute value _
@xmath2 of the field difference , as those for @xmath4 and @xmath272 should be identical . therefore , even though the negativity for the @xmath0 pair does depend on the average field @xmath247 ( through the relative weights of the distinct eigenstates ) , as was seen in previous figures , the limit temperature for non - zero negativity at fixed exchange couplings , and the threshold values of @xmath17 or @xmath16 for non - zero negativity at fixed temperature , will depend just on @xmath2 . in the @xmath233 field plane
, the set of zero negativity states will then be stripes , i.e. , typically a stripe @xmath239 .
of course , @xmath197 can be exponentially small outside the stripe , but not strictly zero .
the previous features of the relative entropy of coherence remain also valid .
the standard basis of states @xmath160 continues to be the schmidt basis for definite magnetization eigenstates , i.e. latexmath:[$|\psi_m\rangle=\sum_{m}c_m coherence will approach the gs entanglement entropy ( for a non - degenerate gs ) adopting qualitatively the same form as the @xmath5 negativity .
nevertheless , as @xmath6 increases the @xmath5 steps will become rapidly smoothed out in the coherence , without exhibiting minima or valleys .
it will also rapidly occupy the @xmath5 separable sectors , becoming in particular prominent along the line @xmath55 for @xmath72 , as seen in the bottom right panel of fig .
[ f9 ] . on the other hand , for sufficiently high temperatures if will approach a uniform decay pattern for all @xmath1 .
a series expansion for @xmath274 leads to @xmath275}{2 d\ln 2}=\frac{1}{9\ln 2 } \left(\frac{s(s+1)j}{kt}\right)^2\,,\label{cas}\ ] ] where the first result holds in a system of finite dimension @xmath276 and the last one is the leading asymptotic expression for a spin @xmath1 @xmath0 pair .
it reproduces the leading term of previous asymptotic results ( [ cha ] ) and ( [ cap1 ] ) .
we have discussed in detail the entanglement and coherence of the @xmath0 spin @xmath1 pair in a transverse non - uniform field at both zero and finite temperatures .
the general spin @xmath1 case exhibits interesting features which can already be seen in the basic @xmath29 case . in the latter ,
while the @xmath5 diagram in the field plane is characterized by an @xmath36 entangled region bounded by hyperbola branches , reachable through non - uniform fields even in the fm case @xmath72 , the thermal state is characterized by a _ separability stripe _ @xmath108 in the @xmath277 field plane for _ any _ @xmath3 , with the system becoming pure and entangled for large values of @xmath2 .
analytic expressions were provided .
remarkably , these features were shown to remain strictly valid for any value of the spin @xmath1 .
the boundaries of the @xmath5 entangled sector are given by the same expressions with a simple rescaling , while the conditions for non - zero thermal entanglement and negativity were rigorously shown to depend just on the field difference @xmath2 for any @xmath1 , entailing that for @xmath3 strict separability will still be restricted to a stripe @xmath278 .
the main difference with the spin @xmath7 case is the emergence of @xmath200 magnetization and entanglement steps at @xmath5 , which lead to deep valleys in the negativity at low temperatures but which disappear as @xmath6 increases .
another interesting aspect emerging for non - uniform fields for increasing spin is the appearing of a critical point along the line @xmath55 , which determines the onset of gs entanglement for @xmath72 and where all @xmath227 gs s with magnetizations @xmath279 coalesce .
the relative entropy of coherence in the standard basis approaches the entanglement entropy for @xmath131 , although for @xmath280 it stays non - zero for all fields .
the exact asymptotic expression for high @xmath6 was derived , which shows that it ultimately decays uniformly as @xmath281 for sufficiently high temperatures . in summary ,
the present results show that the @xmath0 pair in a non - uniform field is an attractive simple system with potential for quantum information applications .
its entangled eigenstates , having definite magnetization , admit a variable degree of entanglement which can be controlled by tuning the fields at each spin .
moreover such tuning enables to select the magnetization of the gs at @xmath5 for any anisotropy , while at @xmath3 it allows one to effectively cool down the system to an entangled state . at @xmath5 entanglement
itself can be detected and approximately measured through the magnetization , since it decreases with increasing @xmath8 and vanishes just for maximum @xmath8 .
the possibility of simulating @xmath0 systems with tunable couplings and fields by different means enhances the interest in this type of models .
this work was supported in part by the departamento de ingeniera qumica utn - fra ( er ) , comisin de investigaciones cientficas ( cic ) ( rr ) , and conicet(nc ) of argentina .
authors also acknowledge support from conicet grant pip 11220150100732 .
999 amico l , fazio r , osterloh a , vedral v 2008 517 eisert j , cramer m , plenio m b 2010 277 osborne t j , nielsen m a 2002 a * 66 * 032110 vidal g latorre j i , rico e , kitaev a 2003 l * 90 * 227902 nielsen m a and chuang
i l 2000 _ quantum computation and quantum information _ ( cambridge univ .
press , cambridge , uk ) haroche s and raimond j m 2006 _ exploring the quantum : atoms , cavities and photons _ ( oxford , univ .
press , oxford , uk ) divincenzo d p _ et .
_ 2000 , nature ( london ) * 408 * , 339 vogel w and sperling j 2014 a * 89 * , 052302 baumgratz t , cramer m and plenio m 2014 l * 113 * , 140401 girolami d 2014 _ phys .
lett . _ * 113 * , 170401 ; smyth c and scholes g d 2014 _ phys .
rev . a _ * 90 * , 032312 matera j m , egloff d , killoran n and plenio m b 2016 _ j. quantum science and technology _ * 1 * , 01lt01 .
d. porras , j.i .
cirac 2004 l * 92 * 207901 georgescu i m , ashhab s , nori f 2014 153 blatt r and roos c f 2012 nat
. phys .
* 8 * 277 senko c et al 2015 x * 5 * 021026 barends r et al 2016 nature * 534 * 222 ; 2013 a * 111 * 080502 lewenstein m , sanpera a , ahufinger v 2012 _ ultracold atoms in optical lattices _ ( oxford univ
. press , uk ) arnesen m c , bose s and vedral v 2001 l * 87 * , 017901 gunlycke d et al 2001 a * 64 * , 042302 wang x 2001 a * 64 * , 012313 ; 2002 a * 66 * , 034302 ; 2002 a * 66 * , 044305 wang x , fu h and solomon a 2001 , 11307 kamta g l and starace a f 2002 l * 88 * , 107901 glaser u , bttner h and fehske h 2003 a * 68 * , 032318 canosa n and rossignoli r 2004 a * 69 * 052306 rossignoli r and canosa n 2005 a * 72 * 012335 ; canosa n and rossignoli r 2006 a * 73 * 022347 sadiek g , kais s 2013 245501 ollivier h and zurek w h 2001 l * 88 * 017901 ( 2001 ) ; henderson l and vedral v 2001 * 34 * , 6899 modi k , brodutch a , cable h , paterek t , vedral v , 2012 1655 rossignoli r , canosa
n , ciliberti l 2010 a * 82 * 052342 ; ciliberti l , rossignoli r , canosa n 2010 a * 82 * 042316 sun y , chen y and chen h 2003 a * 68 * 044301 asoudeh m and karimipour v 2005 a * 71 * 022308 zhang g f , li s s 2005 a * 72 * 034302 .
hu z n , yi k s , park k s , 7284 hassan a s m , lari b , joag p s 2010 * 43 * 485302 guo j l , mi y j , zhang j , song h 2011 065504 abliz a et al 2006 a * 74 * 052105 zhang g f _ et al _ 2011 , 02694 .
albayrak e 2010 , _
chinese phys .
b _ * 19 * , 090319 guo k t et al 2010 505301 hu li _
_ 2014 _ ch . phys .
lett . _ * 31 * , 040301 moriya t 1960 91 ; dzialoshinskii t e 1957 sov . phys . * 5 * 1259 oshikawa m , affleck i 1997 2883 ; derzhko o , richter j zaburannyi o 2000 8661 ( 2000 ) .
wootters w k 1998 * 80 * 2245 ; hill s and wootters w k 1997 5022
vidal g and werner r f 2002 a * 65 * , 032314 zyczkowski k , horodecki p , sanpera a and peres a 1996 * 77 * , 1413 horodecki m , horodecki p and horodecki r 1996 werner r f 1989 a * 40 * , 04277 boette a , rossignoli r , canosa n , matera j m 2016 b * 94 * 214403 | we investigate entanglement and coherence in an @xmath0 spin-@xmath1 pair immersed in a non - uniform transverse magnetic field .
the ground state and thermal entanglement phase diagrams are analyzed in detail in both the ferromagnetic and antiferromagnetic cases .
it is shown that a non - uniform field enables to control the energy levels and the entanglement of the corresponding eigenstates , making it possible to entangle the system for any value of the exchange couplings , both at zero and finite temperatures .
moreover , the limit temperature for entanglement is shown to depend only on the difference @xmath2 between the fields applied at each spin , leading for @xmath3 to a separability stripe in the @xmath4 field plane such that the system becomes entangled above a threshold value of @xmath2 .
these results are demonstrated to be rigorously valid for any spin @xmath1 . on the other hand ,
the relative entropy of coherence in the standard basis , which coincides with the ground state entanglement entropy at @xmath5 for any @xmath1 , becomes non - zero for any value of the fields at @xmath3 , decreasing uniformly for sufficiently high @xmath6 .
a special critical point arising at @xmath5 for non - uniform fields in the ferromagnetic case is also discussed . +
* keywords : * quantum entanglement , quantum coherence , spin systems , non - uniform fields |
beam steering is one of the most important technologies in the modern optics due to its numerous applications in such fields as optical imaging , laser machining , and free space communication .
various physical mechanisms have been explored for deflection of beams by inducing refractive index gradient @xcite .
they include mechanical motion @xcite , thermal gradient @xcite , the acousto - optical interaction @xcite , and the electro - optic effect @xcite . fast beam steering in photonic crystals @xcite and phased arrays @xcite were also proposed .
the direction of the light propagation can also be changed by another beam of light through interaction with matter . in particular ,
a medium exhibiting electromagnetically induced transparency ( eit ) @xcite can provide large probe - beam deflection since the refractive index changes significantly near the transparency center @xcite . among the related studies we mention the refractive index measurements by probe - beam @xcite , investigation of slow light deflection by magneto - optically controlled atomic media @xcite , and exploring a scheme of all - optical beam steering @xcite .
yet another example is the electromagnetically induced waveguiding which uses the control field as a waveguide to confine the probe field @xcite .
however , the schemes proposed in the most of the previous studies are restricted to the linear regime
. they usually result in a spread of the probe pulse because the refractive index gradient depends on both the probe frequency and the spatial coordinates .
fortunately , such spread can be suppressed by the enhancement of the wave localization through the formation of solitons .
large intrinsic nonlinearity in an eit - based media allows for existence of probe beam solitons with extremely weak light intensities @xcite .
more specifically , ultraslow solitons @xcite , spatial solitons @xcite , and gap solitons @xcite at low light intensity can exist in such kind of media .
in this work , we propose a scheme to achieve efficient all - optical steering of light in a resonant three - level atomic system under eit regime .
the scheme is based on the phenomenon of nonlinear long - living bloch oscillations ( bos ) of gap solitons , recently reported in ref .
the system at hand is a gas of @xmath0-atoms loaded in a planar waveguide created by two photonic crystals .
the control field used in the system is a standing - wave laser beam that originates a linear force , as well as _ linear _ and _ nonlinear _ optical lattices ( ols ) affecting a weak probe beam .
ols in such a system are flexible : their parameters can be adjusted either by changing the geometry and/or intensity of the control field , or by varying one and/or two photon detunings .
briefly , the mechanism of the steering consists of two ingredients : the spatially periodic linear force induces dynamics of the probe beam in the transverse direction , while the combined effect of the lattices controls the direction of the beam propagation introducing desirable and controllable deviations . more specifically , the linear ol provides the band structure necessary for existence of the bloch states , while the nonlinear lattice is used to control the stability properties of the bloch states necessary for existence of gap solitons at both edges of each band @xcite . due to the enhancement of the kerr nonlinearity induced by the eit mechanism ,
stable spatial gap solitons , representing the beam channels , can be formed even subject to extremely weak probe light intensity and subsequently strongly deflected without undergoing appreciable deformations and attenuations .
the paper is organized as follows . in the next section ,
the model is introduced . in sec .
iii , the nonlinear equation governing the evolution of the probe field amplitude is derived .
we show that stable spatial gap solitons can exist under properly chosen parameters of the control field .
we also show how one can implement an efficient all - optical steering of the gap solitons .
all results in this work are obtained under a set of experimentally feasible parameters .
the outcomes are summarized in the conclusion .
we consider a cold gas of lifetime - broadened @xmath0-type atoms , loaded into an anti - relaxation - coated planar hollow - core photonic crystal waveguide , schematically shown in fig .
( it is relevant to note that both room - temperature and ultracold atoms have been successfully loaded into hollow - core photonic crystal fibers @xcite , and have further been used to study eit @xcite and all - optical switching @xcite ) . as in any @xmath0-system
, the atoms can populate three states : the ground state @xmath1 , the excited state @xmath2 , and the low - energy state @xmath3 [ see fig .
[ fig1 ] ( b ) ] .
the transitions between the states @xmath1 and @xmath3 are forbidden .
-direction ) is orthogonal to the ol axis ( @xmath4-direction ) .
@xmath5 is the half angle between the two input control beams and @xmath6 is the output angle of the probe beam .
( b ) left panel : @xmath0-type atomic system interacting with two laser beams . @xmath7 and @xmath8 are respectively the rabi frequencies of the probe and control fields . @xmath9 and @xmath10 are respectively the two - photon and the one - photon detunings .
right panel : the waveguide cross section .
the dotted region denotes the core filled in with the atomic gas and limited by the bragg mirrors .
@xmath11 , @xmath12 , and @xmath13 are the thicknesses of the core and different materials of the bragg mirrors , respectively . ] for the particular choice of the atomic gas and geometry of the system , we take into account the two facts .
first , in order to contribute to the nonlinear polarization the matrix elements of the transition between the lower states and the excited state must be nonzero , and hence the population of the excited state must also be nonzero ( see e.g. @xcite ) .
thus , one has to explore a gas with weak atomic losses due to spontaneous emission from the excited state .
a particular system having such properties is the laser - cooled alkaline earth metal atoms , such as strontium atoms @xcite where the atomic states @xmath14 , @xmath15 , and @xmath16 can be chosen as @xmath17 , and @xmath18 , and @xmath19 , respectively .
one of the advantages of such a system is that the atoms possess a long life time even in their excited states : the life time of the state @xmath2 is about @xmath20 @xmath21s corresponding to the decay rate @xmath22 khz .
therefore , in what follows we use this system for the sake of the numerical studies ( see figs .
[ fig2 ] , [ fig3 ] , and [ fig4 ] ) .
the second fact to be taken into account is the smallest possible size of the system .
our aim is to execute effective steering of light using ols . in order to make the periodicity appreciable in a finite structure
, one has to require the system to have the width of at least ten lattice periods or more .
hence , the desirable ol should be created by the beam with the highest possible frequency . in our case , there are only two electromagnetic waves in the system , which provide the transitions @xmath23 and @xmath24 , respectively .
thus we are interested in the geometry where the strong control field connects the ground @xmath1 and the excited @xmath2 states , while the probe weak wave links the states @xmath1 and @xmath3 . in the case of the strontium atoms @xcite
the wavelength of the probe laser beam is @xmath25 @xmath21 m while the length of the coupling field is @xmath26 nm . requiring the modulation of the nonlinear polarization to have at least 20 lattice periods , we conclude that the width of the waveguide should be of order of 14 @xmath21 m .
now we can describe the geometry of the system at hand as shown in fig .
the waveguide created by two parallel bragg mirrors is placed in the ( @xmath27)-plane . a strong control field consisting of a superposition of counter - propagating waves with the frequency @xmath28 and the wavevectors @xmath29 ( @xmath30 ) , couples the excited state @xmath2 and the ground state @xmath1 .
the excited state @xmath2 is also coupled to the low energy state @xmath3 by a weak probe field of the frequency @xmath31 and the wavevector @xmath32 ( i. e. it propagates along @xmath33-axis ) .
we consider a planar waveguide with the transverse ( i.e. along @xmath34-axis ) width small enough so that only one transverse mode of the probe beam needs to be considered ( see appendix [ ap ] ) .
we concentrate on the @xmath35th mode designated as @xmath36 .
therefore the probe field can be presented in the form @xmath37 where , @xmath38 is the polarization vector and @xmath39 is the slowly varying envelope of the probe field .
the frequency ( amplitude ) of the control field @xmath40 ( @xmath41 ) is much higher than that of the guided probe beam @xmath42 ( @xmath43 ) .
this leads to a the effective refractive index of the bragg mirrors for the control field in the claddings to be of order one ( see appendix [ ap ] ) , allowing us neglecting its transverse distribution ( we however emphasize that this design is introduced only for the sake of simplicity ) .
the control field can be presented in the form @xmath44 here , @xmath45 stands for the polarization vector , the function @xmath46 describes the slowly varying distribution of the control field in the @xmath4 direction , and @xmath47 is the complex amplitude of the control field . in experiments , the desired space dependence of @xmath46 can be achieved with the help of beam masks . the total electric - field can be written down as @xmath48 and considered classically .
large difference between the frequencies of the probe and control fields ( @xmath49 ) allows us to split the equations for @xmath50 and @xmath51 , while large difference in the field amplitudes ( @xmath52 ) allows us to consider @xmath51 to be constant in the equation for the probe beam @xcite .
now the equation governing @xmath50 can be written down as follows : [ max ] ( ^2- ) e_p= p. here @xmath53 is the probe beam polarization defined by the properties of the atomic gas in the waveguide core and by the properties of the bragg mirrors in the waveguide claddings , i.e. p= p_clad , ( y<-b , y > b ) + p_core.(-b < y < b ) following the standard procedure @xcite @xmath54 can be determined by calculating the effective refractive index of the bragg mirrors @xmath55 from eq .
( [ app_p ] ) : @xmath56 . in order to compute @xmath57
one introduces the bosonic field operators @xmath58 , @xmath59 of the states @xmath60 ( @xmath61 ) , as well as the electric dipole matrix elements @xmath62 associated with the transitions between the states @xmath63 and @xmath60 . then the polarization can be obtained as @xmath64 the steering of the light we are interested in is performed through the proper spatial and/or temporal modulations of @xmath57 . in order to link it directly to the probe field @xmath65
we have to address the heisenberg equations for the operators @xmath58 describing the atomic medium .
this can be done neglecting the atomic kinetic energy , what leads to the system as follows ( its derivation repeats the steps outlined say in @xcite ) [ heisen ] & & i_1=-_1 _ 1 - 2_c^f(x)(k x)_2 , + & & i_2=(_2-i_2 ) _ 2 - 2_c f(x)(k x)_1 + & & -_p_3 , + & & i_3= -_p^_2 .
[ heisen_a ] here @xmath66 and @xmath67 , are the rabi frequencies of the respective fields @xmath68 , and @xmath69 present one- and two - photon detunings , respectively ( see fig . [ fig1 ] ) . in eqs .
( [ heisen ] ) , we keep only the dissipation of the state @xmath2 due to spontaneous emission , by adding the decay rate @xmath70 , phenomenologically .
usually , the decay rates of the lower states @xmath1 and @xmath3 are much smaller than that of the exited state and can be safely neglected . for a particular choice of the atomic gas and assuming that the decay rate of the excited state is much less than the two - photon detuning ( @xmath71 ) , one may consider the loss of the atoms from the system as a small perturbation ( see also the discussion in @xcite ) .
since the two - photon detuning is nonzero ( i.e. @xmath72 ) , the pure dark state ( corresponding to @xmath73 ) can not exist .
however , we explore the situation where the atomic system is close enough to the dark state . to this end
we look for a solution of eqs .
( [ heisen ] ) in the form @xmath74 , where the exponent @xmath75 is slowly varying in space and can be obtained as a root of the respective characteristic equation .
we are interested in a root having the smallest imaginary part @xcite , i.e. [ lambda_r ] = _ r+i_i , ( @xmath76 ) with & & _ r= , + & & _ i=_r^2.then the expectation value @xmath77 can be readily computed from ( [ heisen ] ) and ( [ lambda_r ] ) : @xmath78 here we introduced the dimensionless functions @xmath79 and = = _ r+i_i , where _ r&= & , + _ i&=&_r^2 , as well as the parameters @xmath80 ( @xmath81 ) and @xmath82 . since the probe field ( decay rate of the excited state )
is assumed to be weak in comparison with the control field , we have that @xmath83 ( @xmath84 , see eq .
( [ g_1 ] ) and the related discussions ) .
an important conclusion follows from eq .
( [ heisen_a ] ) , the quasi - dark state , characterized by the small population of the excited state @xmath2 , requires @xmath85 and hence @xmath86 . in
what follows we will be interested in a particular choice @xmath87 where @xmath88 is the mask parameter considered to be small .
more specifically , we will consider the beam propagation ( i.e. distributions of @xmath89 in space ) characterized by finite deviations from the axis @xmath4 . as we have already announced in the introduction
, the mentioned deviation will occur due to bos of the beam . thus if we assume that the amplitude of bos measured in units @xmath90 [ see eq .
( [ max3 ] ) below ] is @xmath91 , the smallness of @xmath88 is determined by the requirement that @xmath92 .
this allows us to achieve further simplification for the polarization by defining [ v ] v(k x)_1_2 + 4 ^2(k x ) .
performing the expansion of the right hand side of eq .
( [ g_0 ] ) in the taylor series with respect to @xmath89 we get : [ g_1 ] & & _ 3^ _ 2 + i. here we have neglected the @xmath93-order terms and assumed that in the domain of bos the condition @xmath94 is satisfied . notice that the requirement @xmath95 holds for in the whole space provided [ v_cond ] _
1_2>0 _ 1_2<-4
. this condition will be imposed in what follows .
we are particularly interested in the beam dynamics resulting from the interplay of diffraction and nonlinearity when the probe field passes through the atomic cloud .
therefore , we concentrate on stationary solutions , requiring @xmath89 to be independent on time . substituting eqs .
( [ ep ] ) and ( [ g_1 ] ) into eq .
( [ max ] ) , taking into account ( [ app_s ] ) , and leaving only the leading order terms , we arrive at the dimensionless equation for the slowly varying function @xmath89 : [ max3 ] & & i+ + u()+f ( ) -g()||^2 + & & + i()=0 . here
@xmath96 and @xmath97 are the dimensionless independent variables , @xmath98 is given by ( [ v ] ) , and we defined u()&= & , f()= , + g()&= & , ( ) = , with u_0=_1 , and @xmath99 .
the constant @xmath100 depending on the particular choice of the bragg mirrors is computed in the appendix [ ap ] . the obtained dimensionless eq .
( [ max3 ] ) implies that all the terms are of the order one or less .
more specifically , it is necessary to require that @xmath101 .
this last constrain can be satisfied by different means .
notice that the parameter @xmath102 can be experimentally controlled by the geometry of the laser beams ( see fig .
[ fig1 ] ) , we thus fix @xmath103 and adjust other parameters ( in particular @xmath104 and @xmath8 ) to satisfy the constrain . for a typical set of data explored below ,
i.e. the atomic concentration @xmath105 @xmath106 and @xmath107 s@xmath108 ( @xmath109 ) , we obtain @xmath110 satisfying the desired order of magnitude .
meanwhile , the dissipation is very small ( @xmath111 ) in the considered system since @xmath112 .
we study the situation when the effects of the linear force , nonlinearity , and dissipation are small enough .
therefore we first notice that for @xmath113 , @xmath114 , and @xmath115 one arrives at the familiar framework of the band theory , where by means of substitution @xmath116 eq .
( [ max3 ] ) is reduced to the eigenvalue problem [ linear ] u_,q()+k_(q)u_,q()=0 . here
@xmath117 and @xmath118 is a bloch function with indexes @xmath119 and @xmath120 standing respectively for the band index and for the wave vector in the first brillouin zone ( bz ) , i.e. for @xmath121 $ ] in the case at hand .
the physical meaning of the eigenvalue @xmath122 is the propagation constant . in what follows we concentrate on the lowest energy band ( @xmath123 ) and therefore , for the sake of brevity ,
omit the band index .
the effect of the periodic linear force originated by the control beam can be described by the equations [ motion ] = = , = -f ( ) , where @xmath124 and @xmath125 denote the coordinates of the center of the bloch wave packet in the real and reciprocal spaces , respectively .
@xmath126 is the refraction angle with @xmath127 ( see figs .
[ fig1 ] ( a ) and [ fig3 ] ( a ) ) . a peculiarity of our case is that @xmath128 is an oscillating function , which has a nonzero mean value @xmath129 and hence results in the oscillation behavior of the bloch wave packet in the real and reciprocal spaces ( i.e. in bos ) .
the dimensionless amplitude ( along the @xmath4 axis ) and period ( along the @xmath33 axis ) of such oscillations can be calculated as @xmath130 , with @xmath131 denoting the width of the lowest allowed band of the periodic potential @xmath132 , and @xmath133 , respectively . in order to describe the effect of the nonlinearity we employ the standard multiple - scale expansion assuming smallness of the averaged force , i.e. @xmath134 . this last requirement is equivalent to the condition @xmath135 , which is consistent with the constrains on the bos amplitude discussed in sec .
ii c provided @xmath136 .
next we introduce the scaled variables @xmath137 ( @xmath138 , 2 , @xmath139 ) , where @xmath21 is a small parameter estimated by @xmath140 .
the probe field can be written in the form of the expansion @xmath141 where in the arguments of the functions @xmath142 we have indicated only the most rapid variables .
@xmath143 . substituting this ansatz into eq .
( [ max3 ] ) and collecting the terms at each order of @xmath144 , we obtain a set of equations which can be solved order by order ( see e.g. @xcite ) . omitting the details we turn directly to the equation for the slowly varying amplitude @xmath145 , which is defined through the relation @xmath146 , where @xmath147\zeta_1 $ ] and the dependence @xmath148 and @xmath149
are obtained from eqs .
( [ motion ] ) . in the third order of the asymptotic expansion , @xmath150
, we obtain [ nls ] i+d(q)-g(q)|a|^2a+ir(q)a=0 , where the coefficients of the diffraction @xmath151 and the effective nonlinearity @xmath152 are given respectively by & & d(q)=12 , g(q)=_0^ g ( ) |u_q()|^4 d , + & & r(q)=_0^ r ( ) |u_q()|^2 d. we notice that both @xmath151 and @xmath152 can be either positive or negative depending on the signs of detunings and on the position of the center of the wave packet ( i.e. on @xmath153 ) .
therefore , in a general case bos oscillations will be accompanied by significant change of the width of the beam , due to the diffraction .
this undesired effect , however can be dramatically reduced , if the conditions for self - focusing of the beam will be satisfied ( and hence will compensate the diffraction ) at any coordinate @xmath153 . these requirements ( which are also referred to as conditions for the gap soliton existence or , alternatively , as conditions of the modulational instability of the bloch waves ) , are well known and read @xcite @xmath154 if condition ( [ main_cond ] ) is satisfied for all ( or almost all ) wave - vectors @xmath125 , the self - focusing of the beam will occur along the whole ( or almost whole ) trajectory of the beam . )
is valid for 90% of wavevectors in the first bz ; in the black domains the requirement ( [ v_cond ] ) is not satisfied and thus the approximation eq .
( [ max3 ] ) fails ; the gray domains show the parameter region where our approximation is valid , but the conditions ( [ main_cond ] ) are not satisfied for more than 10% of the wavevectors in the bz .
the system parameters are given in sec .
ii a. ( b ) the coefficients @xmath155 ( dashed line ) , @xmath156 ( dash - dotted line ) , and @xmath157 ( solid line ) _ vs _ @xmath125 with the parameters @xmath110 , @xmath107 s@xmath108 , and @xmath158 s@xmath108 [ the point @xmath159 in panel ( a ) ] .
( c ) the same as in panel ( b ) but with @xmath160 s@xmath108 [ the point @xmath161 in panel ( a ) ] .
the dynamics of the bloch wave packets for the points @xmath159 and @xmath161 is shown below in the panels ( a ) and ( b ) of fig . [ fig3 ] . ]
thus , as the third step , for achieving the best performance of the device one has to design the system ( geometry of the beams or photon detunings ) in a way to ensure the condition ( [ main_cond ] ) for all @xmath125 .
the procedure of performing this task was described in the previous works @xcite and here we report only the final result .
more specifically , in fig . [ fig2 ] ( a ) , we show the domains on the plain ( @xmath162 ) where the optimal conditions are satisfied ( empty regions ) , not satisfied ( gray regions ) , and the parameter requirement ( [ v_cond ] ) is violated ( black regions ) . in fig . [ fig2 ] ( b ) and ( c ) , we plot the curves @xmath163 , @xmath164 , and @xmath157 versus @xmath125 with the almost optimal design [ point @xmath159 in panel ( a ) ] and non optimal design [ point @xmath161 in panel ( a ) ] of the ols .
we mention that always there exists a small range of the wavectors where spatial gap solitons do not exist ( i.e. where @xmath165 ) .
this range , however , can be made enough small [ less than @xmath166 of the whole bz , like this happens in the point @xmath159 , see fig .
[ fig2 ] ( b ) ] , and therefore has no appreciable destructive effect on the beam dynamics [ see fig .
[ fig3 ] ( a ) ] . on the contrary
, the point @xmath161 in panel ( a ) corresponds to the case in which the bos of the beam are rapidly destroyed leading to strong defocussing of the outcome beam [ see fig .
[ fig3 ] ( b ) ] .
now we turn to the final step of numerical descriptions of the bos of a spatial soliton and of the ight steering . in fig .
[ fig3 ] ( a ) , we show the dynamics of the bloch wave packet launched normally in the atomic media with the almost optimal design of the ols [ point @xmath159 in fig .
[ fig2 ] ( a ) ] .
the result was obtained by employing numerical integration of eq .
( [ max3 ] ) .
the existence of the long - living bos of the gap soliton , with the amplitude and period well matching the theoretical estimates , can be observed . due to the very small dissipation
, there is only a tiny decrease of the soliton intensity due to the absorption of the medium .
more specifically , one can see the highly concentrated beam in the output with the direction of propagation determined by the nonzero angle @xmath126 . for the sake of comparison , in fig .
[ fig3 ] ( b ) , we show the dynamics of the same bloch wave packet in the non optimally designed ol [ the parameters used in this panel are the same with those used in fig .
[ fig2 ] ( c ) ] . in this case
, the beam undergoes strong spreading out due to the alternating diffractive and self - defocussing regeems .
the intensity profiles of the output beam vs the input beam are shown in the panels ( c ) and ( d ) .
one observes that there is no apparent deformation between the output beam and input beam in panel ( c ) while an obvious destruction appears in panel ( d ) . in fig .
[ fig2 ] ( a ) ] obtained by direct numerical integration of eq .
( [ max3 ] ) .
the amplitude and the period of the bos are @xmath167 ( @xmath168 @xmath21 m ) and @xmath169 ( @xmath170 @xmath21 m ) .
( b ) bos of the wave packet in the non optimally designed ols [ point @xmath161 in fig .
[ fig2 ] ( a ) ] . in the both cases
the initial condition is a stationary gap soliton located at @xmath171 .
the intensity profiles of the output beam ( blue , dashed line ) vs the input beam ( red , solid line ) corresponding to the cases ( a ) and ( b ) are respectively shown in the panels ( c ) and ( d ) obtained at @xmath172 . ]
the maximum input intensity of the probe beam in fig .
[ fig3 ] ( a ) can be estimated by the formula @xmath173 . using the parameters given in fig .
[ fig2 ] ( b ) , we obtain @xmath174 @xmath21w @xmath175 .
thus , the generation of the spatial gap soliton in the system at hand requires only very low input light intensity .
we remark that the intensity of a single 500-nm photon per nanosecond on an area of 1 @xmath21m@xmath176 is @xmath177 w @xmath175 .
this estimation shows that our system makes it possible to manage the weak beams characterized by single - photon wavepackets .
this is drastically different from the spatial optical soliton generation in a conventional waveguide where an input laser pulse with very high peak power @xmath178 kw is needed in order to bring out a sufficient nonlinear effect @xcite .
the existence of long - living bos of spatial optical gap solitons can be used to implement efficient all - optical steering of probe beams .
different from all linear systems where the spread and attenuation of the probe pulse are unavoidable because the refractive index gradient depends on both light frequency and spatial coordinates , the scheme proposed here can greatly improve the device performance due to the self - focusing of the beam . in experiments , the optical masks and atomic cells are usually chosen in advance .
however , one can exploit the possibility of changing the geometry or parameters of the control field to steer the output probe beam .
as an example , by changing the angle between the input control fields @xmath5 ( see fig .
[ fig1 ] ) , one can efficiently control the refraction angle of the output probe beam . in fig .
[ fig4 ] , we show the tangent of the refraction angle @xmath179 versus @xmath180 with a particular length of atomic cells ( i.e. @xmath172 ) . a wide range of angle , @xmath181 $ ] , can be achieved with the given parameters .
_ vs _ @xmath180 at @xmath172 for the almost optimally designed ols [ corresponding to point @xmath159 in fig . [ fig2 ] ( a ) and panel ( a ) of fig . [ fig3 ] ]
in conclusion , we studied ( both analytically and numerically ) the dynamics of weak probe optical beams in a gas of three - level @xmath0-atoms subjected to a control laser beam . the standing wave originates periodic modulations of the linear and nonlinear parts of the dielectric permittivity of the medium , which is `` seen '' by the probe beam .
if the amplitude of the control field is smoothly modulated , the probe beam undergoes the bloch oscillations .
these oscillations correspond to periodic change of the direction of the propagation of the beam .
we have shown that it is possible to choose the parameters of the atomic system , say rabi frequency of the control field and photon detunings , to reduce the diffraction of the probe beam , and thus the output beam is approximately the same with the input one . in this case , the probe field distributions is nothing but a spatial gap soliton undergoing long - living bloch oscillations . due to the significant enhancement of nonlinearity , such solitons can be formed with extremely weak light intensity , below single - photon wavepacket level .
thus , through changing the geometry or parameters of the system one can control the direction of the output beam to carry out an efficient all - optical steering of light .
the work of c.h . was supported by the fundao para a cincia e a tecnologia ( fct ) under grant no .
sfrh / bpd/36385/2007 and estmulo investigao 2009 de fundao calouste gulbenkian .
the research of vvk was partially supported by the grant piif - ga-2009 - 236099 ( nomatos ) .
the transverse profiles of the probe field can be determined from the eigenvalue problem @xcite [ app_s ] & & ( + -k_l^2)s_l(y)=0 , ( |y|<b ) [ app_s1 ] + & & ( + -k_l^2)s_l(y)=0 , ( |y| > b ) [ app_s2 ] where @xmath182 stands for either @xmath42 or @xmath40 , @xmath183 stands for the mode index ( @xmath184 ) , @xmath185 is a normalized , i.e. @xmath186 , profile of the respective mode , and @xmath11 is the width of the waveguide core . the eigenvalue @xmath187 and effective refractive index of the bragg mirrors @xmath55 can be obtained from the dispersion relations of the planar waveguide [ app_w ] ^2= , and the expression valid the finite photonic structures @xcite [ app_p ] n()=(_t - t ( ) ) , where @xmath188 and @xmath189 are the total phase and the transmittance of the complex transmission coefficient for the structure @xmath190 with @xmath191 ( the integer @xmath192 is uniquely defined assuming @xmath193 is a monotonically increasing function ) and @xmath194 .
@xmath195 is the total thickness of the @xmath196-period crystal @xmath197 with @xmath12 and @xmath13 being the thicknesses of two different materials [ see fig .
[ fig1 ] ( b ) ] . as an example we consider the probe field constrained of quarter - wave bragg mirrors , @xmath198 with @xmath199 and @xmath200 being the refractive indexes of the two different materials , for claddings , @xmath201 , and @xmath202 mm
, we obtain that @xmath203 cm@xmath108 and @xmath205 . at the same time , for the control field we have @xmath206 cm@xmath108 and the effective refractive index of the cladding is @xmath207 .
the assumption that only @xmath208 takes into account requires that the energy for exciting @xmath209 [ corresponding to the eigenvalue of eq .
( [ app_s1 ] ) ] is higher than that of the self - action ( nonlinearity ) in eq .
( [ max3 ] ) . with the parameters given in fig .
[ fig2 ] , we obtain @xmath210 ( @xmath211 cm@xmath108 for the probe field ) , hence we can safely neglect the higher transverse modes ( i.e. @xmath212 for @xmath213 ) . here , the integral @xmath214 .
however , the much higher frequency of the control field allows us to neglect the difference among transverse profiles of the control field in the waveguide core .
e. g. spencer , p. v. lenzo , and a. a. ballman , proc .
ieee * 55 * , 2074 ( 1967 ) ; t. c. lee and j. d. zook , ieee j. quantum electron . * 4 * , 442 ( 1968 ) ; j. f. lotspeich , ieee spectrum * 5 * , 45 ( 1968 ) ; j. d. zook , appl
. opt . * 13 * , 875 ( 1974 ) ; d. scrymgeour , n. malkova , s. kim , and v. gopalan , app .
, * 82 * , 3176 ( 2003 ) . s. e. harris , j. e. field , and a. imamoglu , phys . rev . lett .
* 64 * , 1107 ( 1990 ) ; m. o. scully and s. y. zhu , opt .
commun . * 87 * , 134 ( 1992 ) ; m. xiao , y. q. li , s. z. jin , and j. gea - banacloche , phys . rev
74 * , 666 ( 1995 ) .
r. holzner , p. eschle , s. dangel , r. richard , h. schmid , u. rusch , b. rhricht , r. j. ballagh , a. w. mccord , and w. j. sandle , phys .
rev . lett . * 78 * , 3451 ( 1997 ) ; d. l. zhou , l. zhou , r. q. wang , s. yi , and c. p. sun , phys .
a * 76 * , 055801 ( 2007 ) .
r. r. moseley , s. shepherd , d. j. fulton , b. d. sinclair , and m. h. dunn , phys .
* 74 * , 670 ( 1995 ) ; a. g. truscott , m. e. j. friese , n. r. heckenberg , and h. rubinsztein - dunlop , phys .
lett . * 82 * , 1438 ( 1999 ) ; r. kapoor and g. s. agarwal , phys .
rev . a * 61 * , 053818 ( 2000 ) ; j. a. andersen , m. e. j. friese , a. g. truscott , z. ficek , p. d. drummond , n. r. heckenberg , and h. rubinsztein - dunlop , phys . rev .
a * 63 * , 023820 ( 2001 ) .
s. ghosh , a. r. bhagwat , c. k. renshaw , s. goh , a. l. gaeta , and b. j. kirby , phys .
97 * , 023603 ( 2006 ) ; p. s. light , f. benabid , f. couny , m. maric , and a. n. luiten , opt
32 * , 1323 ( 2007 ) ; t. takekoshi and r. j. knize , phys . rev
* 98 * , 210404 ( 2007 ) ; c. a. christensen , s. will , m. saba , g. jo , y. shin , w. ketterle , and d. pritchard , phys .
a * 78 * , 033429 ( 2008 ) . | a standing - wave control field applied to a three - level atomic medium in a planar hollow - core photonic crystal waveguide creates periodic variations of linear and nonlinear refractive indexes of the medium .
this property can be used for efficient steering of light . in this work
we study , both analytically and numerically , the dynamics of probe optical beams in such structures . by properly designing the spatial dependence of the nonlinearity it is possible to induce long - living bloch oscillations of spatial gap solitons , thus providing desirable change in direction of the beam propagation without inducing appreciable diffraction . due to the significant enhancement of the nonlinearity ,
such self - focusing of the probe beam can be reached at extremely weak light intensities . |
when the `` new '' star in andromeda appeared in 1885 with its striking association with the great andromeda nebula , it was still a long way to go until the recognition that these phenomena represent , like nothing else , our deep relation with the universe .
shortly after 1920 , astronomers learned that the andromeda nebula was located far outside our milky way and that the `` nova '' s andromedae must have been much more luminous than any other nova .
thus fritz zwicky created the term _ supernova _ ( sn ) for these events , of which by 1930 only a handful had been detected in external galaxies . today
it is common wisdom that supernovae are largely responsible for the enrichment of the universe with heavy elements and thus also enable the existence of life .
supernova detections are no longer a matter of incident alone : sophisticated supernova searches with automatic telescopes are now routinely conducted and , thanks to these efforts , we have detected 1387 sn events until february 1999 .
morphological classification is one of the first scientific steps on the way to the understanding of a new phenomenon .
it is amazing that the first attempts to classify sne still provide the basis , although our classification scheme has become more subtle and we now know that the morphological phenomenon `` supernova '' embraces physically distinct processes ( for the early years see the excellent review of trimble @xcite , a recent review on sne ia as distance indicators is from branch @xcite ) . as did minkowski in 1936 , we still basically distinguish two types of sne , types i and ii . this classification is purely spectroscopic : type i events do not show hydrogen lines in their spectra , while type ii events do so .
an important observation can be made with respect to the parent population of sne : type i appears in all sorts of galaxies , while type ii seems to be restricted to galaxies with a young stellar population , i.e. spiral galaxies and irregulars . for many years
, the canonical interpretation was that the progenitors of type i stem from an old stellar population .
a popular model is a matter - accreting white dwarf in a binary system which becomes unstable when the chandrasekhar mass limit is reached .
then the star explodes by thermonuclear detonation or deflagration @xcite ( which is just another expression for super- or subsonic movement of the burning front but is nevertheless very important in modelling the event ) , but see kundt @xcite for difficulties of this model . on the other hand ,
type ii was thought to be the explosion of a massive star , whose core collapses at the end of its evolution , forms a neutron star ( already zwicky had this vision ) , and the released gravitational energy ( about @xmath2erg ) drives the expansion of the outer shells ( in a complicated and not yet fully understood manner ) . regarding the energetics of type ii events , it soon became clear that radiation and the kinetic energy of the shell must be only a tiny fraction of the total energy ( @xmath3erg ) , the rest being emitted as neutrinos .
the detection of neutrinos from sn1987a was a triumph to this theory ( e.g. @xcite ) .
the coarse classification depending on the spectral hydrogen features is further refined .
the subclassification of type ii by the shape of the light curve appears useful , distinguishing sne ii with ( ii__p _ _ for _ _ p__lateau ) or without plateaus ( ii__l _ _ for _ _ l__inear ) , which basically depends on the thickness of the hydrogen shell and the speed at which the @xmath4h - ionisation front moves inward during expansion . but type
i sne are also subclassified into ia , ib and ic . by 1983 ,
a few type i sne had been observed which developed strong radio emission a few weeks after maximum light .
this is now commonly interpreted as the core collapse and subsequent explosion of the outer layers of a massive evolved star ( for instance a wolf - rayet star ) which lost its hydrogen shell prior to the explosion .
this new type of sn without hydrogen lines and massive progenitor was named ib or ic , depending on whether helium lines were strong or not .
the remaining type ia is spectroscopically distinct from ib by the occurrence of @xmath4si lines in an otherwise very similar spectrum . for more information on sn classification
refer to wheeler & harkness @xcite and filippenko @xcite . in a modern compilation ( asiago catalogue @xcite )
, we now have 403 classified ia events at the time of writing ( may 1999 ) . their distribution among the host galaxies
is shown in fig .
[ fig : hosts ] .
the abscissa is de vaucouleurs classification , where the extreme numbers are @xmath5 ( elliptical galaxies ) and @xmath6 ( late - type irregulars ) .
it is apparent that ia in elliptical galaxies are rarer than in spiral galaxies .
if we normalise to luminosity , then the difference becomes even more obvious .
table [ tab : snstat ] shows the sn ia rate in different galaxy types according to et al .
if we recall that a typical @xmath7-ratio in elliptical galaxies is about 8 and in spiral galaxies about 0.05 , then it becomes clear that a unit mass in spiral galaxies produces many more sne than in elliptical galaxies .
this would be understandable if spiral galaxies had more opportunities to let ia s explode than elliptical galaxies have .
indeed , the former hypothesis , that all ia s are explosions of chandrasekhar mass white dwarfs in old population environments , now stands on very weak grounds .
recent investigations have revealed that there exists a weak statistical correlation with spiral arms in spiral galaxies @xcite , which speaks in favour of the contribution of a parent population of massive stars for at least some ia s . on the other hand , in elliptical galaxies , at best very few massive stars
exist , so one is driven towards the conclusion that ia s originate in young _ and _ old populations . a further , quite irritating observation is that ia s seem to avoid the central regions of ellipticals and spirals @xcite , so that the old bulge populations apparently are not effective producers of ia events .
some researchers even propose that every supernovae has a massive progenitor , e.g. kundt @xcite .
however , our theme is not the physics of ia s but how they can be exploited to obtain the hubble constant . in the next section
, we shall see that in spite of a probably non - uniform origin , ia s exhibit stunningly homogeneous properties which make them excellent standard candles .
.sn rates for various galaxy types in , according to cappellaro et al .
rates in snu scale as @xmath8 [ cols="^,^",options="header " , ] ajhar e.a . , blakeslee j.p . , tonry j.l .
( 1994 ) @xmath9 photometry of globular clusters in virgo and leo ellipticals .
aj 108,2087 arnett w.d . , bahcall j.a . , kirshner r.p
, woosley s.e .
( 1989 ) supernova 1987a .
araa 29,629 ashman k.m . , conti a. , zepf s.e .
( 1995 ) globular cluster systems as distance indicators metallicity effects on the luminosity function . aj 110,1164 barbon r. , buondi v. , capellaro e. , turatto m. ( 1999 ) the asiago supernova catalogue-10 years after .
astro - ph/9908046 , http://athena.pd.astro.it/~supern/ bartunov o.s . ,
tsvetkov d.y .
, filimonova i.v .
( 1994 ) distribution of supernovae relative to spiral arms and h ii regions .
pasp 106,1276 bevington p.r . , robinson d.k .
( 1992 ) data reduction and error analysis for the physical sciences .
mcgraw - hill , 2nd edition branch d. ( 1998 ) type ia supernovae and the hubble constant .
araa 36,17 cappellaro e. , turatto m. , benetti s. , tsvetkov d.y . ,
bartunov o.s .
, makarova i.n .
( 1993 ) the rate of supernovae ii .
the selection effect and the frequencies per unit blue luminosity .
273,383 colgate s.a . ,
mckee c. ( 1969 ) early supernova luminosity .
apj 157 , 623 della valle m. , kissler - patig m. , danziger j. , storm j. ( 1998 ) globular cluster calibration of the peak brightness of the type ia supernova 1992a and the value of @xmath10 .
mnras 299,267 drenkhahn g. , richtler t. ( 1999 ) sn1994d in ngc4526 : a normally bright type ia supernova .
a&a , in press filippenko a.v .
( 1997 ) optical spectra of supernovae .
araa 35,309 gieren w.p .
, richtler t. , hilker m. ( 1994 ) the distance to the large magellanic cloud cluster ngc1866 from its cepheid members .
apj 433,l73 gieren w.p .
, fouqu p. , gomez m. ( 1998 ) cepheid period - radius and period - luminosity relations and the distance to the large magellanic cloud .
apj 496,17 gmez m. , richtler t. , infante l. , drenkhahn g. , ( 1999 ) the peculiar globular cluster system of ngc 1316 ( fornax a ) .
agm 15,p62 grillmair c.j .
, forbes d.a .
, brodie j.p . , elson r.a.w . ( 1999 ) hst imaging of the globular clusters in the fornax cluster : color and luminosity distribution .
aj 117,167 hamuy m. , phillips m.m . ,
maza j. , wischnjewsky m. , uomoto a. , landolt a.u . , khatwani r. ( 1991 ) the optical light curves of sn1980d and sn1981n in ngc1316 ( fornax a ) .
aj 102,208 hamuy m. , phillips m.m . , suntzeff n.b .
, schommer r.a . , maza j. , aviles r. , ( 1996 ) the hubble diagram of the caln / tololo type ia supernovae and the value of @xmath10 .
aj 112,2398 harris w.e . ( 1996 ) a catalog of parameters for globular clusters in the milky way . aj 112,1487 , http://physun.physics.mcmaster.ca/globular.html heithausen a. , bensch f. , stutzki j. , falgarone e. , panis j.f .
the iram key project : small - scale structure of star forming regions . combined mass spectra and scaling laws .
a&a 331,l65 .
heithausen a. , stutzki j. , bensch f. , falgarone e. , panis j.f .
results from the iram key project : `` the small - scale structure of pre - star - forming regions '' laws . rev . of mod .
12,101 hflich p. , khokhlov a. , wheeler j.c .
, phillips m.m .
, suntzeff n.b .
, hamuy m. ( 1996 ) maximum brightness and postmaximum decline of light curves of type ia supernovae : a comparison of theory and observations .
apj 472,l81 jacoby g.h .
, ciardullo r. , ford , h.c .
( 1990 ) planetary nebulae as standard candles .
v - the distance to the virgo cluster . apj 356,332 jensen j.b .
, tonry j.l . , luppino g.a . (
1988 ) measuring distances using infrared surface brightness fluctuations .
apj 505,111 kissler - patig m. , richtler t. , storm j. , della valle m. ( 1997 ) halo and bulge / disk globular clusters in the s0 galaxy ngc1380 .
a&a 327,503 kohle s. , kissler - patig m. , hilker m. , richtler t. , infante l. , quintana h. ( 1996 ) the distance of the fornax cluster based on globular cluster luminosity functions .
309,l39 kuchner m.j .
, kirshner r.p . , pinto p.a . , leibundgut b. ( 1993 ) evidence for ni-56 yields co-56 yields fe-56 decay in type ia supernovae .
apj 426 , l89 kundt w. ( 1990 ) supernova explosions and their ejected shells . in : kundt w. ( ed . )
neutron stars and their birth events .
nato asi c300 , kluwer , p. 40 kundt w. ( 1998 ) astrophysics of neutron stars - facts and fiction about their formation and functioning . fund . of cosm .
20,1 madore b.f .
, freedman w.l . , silbermann n. et al .
( 1998 ) a cepheid distance to the fornax cluster and the local expansion rate of the universe .
nature 395,47 mccray r. ( 1993 ) sn1987a revisited .
araa 31,175 mclaughlin d.e .
( 1994 ) an analytical study of the globular cluster luminosity function .
pasp 106,47 mcmillan r. , ciardullo r. , jacoby g.h .
( 1993 ) planetary nebulae as standard candles .
the distance to the fornax cluster .
apj 416,62 phillips m.m .
( 1993 ) the absolute magnitudes of type ia supernovae .
apj 413,l105 pierce m.h .
, jacoby g.h .
( 1995 ) `` new '' @xmath11 and @xmath12 photometry of the `` old '' type ia supernova 1937c .
aj 110,2885 richmond m.w . , treffers r.r . , filippenko a.v
, van dyk s.d . , paik y. , peng c. , marschall l.a .
, laaksonen b.d . , macintosh b. , mclean i.s .
( 1995 ) @xmath13 photometry of the type ia sn1994d in ngc4526 .
aj 109,2121 richtler t. , grebel e.k . , domgoergen h. , hilker m. , kissler m. ( 1992 ) the globular cluster system of ngc1404 .
a&a 264,25 richtler t. , drenkhahn g. , gmez m. , seggewiss w. ( 1999 ) the hubble constant from the fornax cluster distance . in : science in the vlt era and beyond .
eso vlt opening symposium .
springer - verlag , berlin , in press rieke g.h .
, lebofsky m.j .
( 1985 ) the interstellar extinction law from 1 to 13 microns .
apj 288,618 robichon n. , arenou f. , mermilliod j.c .
, turon c. ( 1999 ) open clusters with hipparcos .
i. mean astrometric parameters .
345,471 saha a. , labhardt l. , schwengeler h. , macchetto f.d . , panagia n. , sandage a. , tammann g.a .
( 1994 ) discovery of cepheids in ic4182 : absolute peak brightness of sn1937c and the value of @xmath10 .
apj 425,14 saha a. , sandage a. , tammann g.a . , labhardt l. , schwengeler h. , panagia n. , macchetto f.d .
( 1995 ) discovery of cepheids in ngc5253 : absolute peak brightness of sn1895b and sn1972e and the value of @xmath10 .
apj 438,8 saha a. , sandage a. , labhardt l. , tammann g.a . , schwengeler h. , macchetto f.d . , panagia n. ( 1996 ) cepheid calibration of the peak brightness of type ia supernovae : sn1960f in ngc4496a .
apjs 107,693 saha a. , sandage a. , labhardt l. , tammann g.a . , schwengeler h. , macchetto f.d . , panagia n. ( 1996 ) cepheid calibration of the peak brightness of type ia supernovae : sn1981b in ngc4536 .
apj 466,55 sandage a. , tammann g.a .
( 1995 ) steps toward the hubble constant .
x. the distance of the virgo cluster core using globular clusters .
apj 446,1 sandage a. , saha a. , tammann g.a .
, labhardt l. , panagia n. , macchetto f.d .
( 1996 ) cepheid calibration of the peak brightness of type ia supernovae : calibration of sn1990n in ngc4639 averaged with six earlier type ia supernova calibrations to give @xmath10 directly .
apj 460,15 schaefer b.e .
( 1995 ) the peak brightness of sn1981f in ngc4536 and the hubble constant .
apj 449,l9 schaefer b.e .
( 1996 ) the peak brightness of sn1960f in ngc4496 and the hubble constant .
apj 460,l19 schaefer b.e .
( 1998 ) the peak brightness of sn1974 g in ngc4414 and the hubble constant .
apj 509,80 schweizer f. ( 1980 ) an optical study of the giant radio galaxy ngc1316 ( fornax a ) .
apj 237,303 secker j. ( 1992 ) a statistical investigation into the shape of the globular cluster luminosity distribution
. aj 104,1472 secker j. , harris w.e .
( 1993 ) a maximum likelihood analysis of globular cluster luminosity distributions in the virgo ellipticals .
aj 105,1358 seggewiss w. ( 1998 ) the distance scale of the universe before and after hipparcos . in : brosche
p. , dick w.r .
, schwarz o. , wielen r. ( eds . )
the message of the angles - astrometry from 1798 to 1998 . proc . of the international spring meeting of the astronomische gesellschaft gotha 1998 , verlag harri deutsch , thun und frankfurt am main 1998 , p. 150
suntzeff n.b . ,
schommer r.a . , olszewski e.w .
, walker a.r .
( 1992 ) spectroscopy of giants in lmc clusters iii .
velocities and abundances for ngc1841 and reticulum and the properties of the metal - poor clusters .
aj 104,1743 suntzeff n.b . , phillips m.m . , covarrubias r. et al .
( 1999 ) optical lightcurve of the ia supernova sn1998bu in m96 and the supernova calibration of the hubble constant
. aj 117,1175 trimble v.l .
( 1982 ) supernovae : a survey of current research , part i : the events .
reviews of modern physics .
54 turner a. , ferrarese l. , saha a. ( 1998 ) the hubble space telescope project on the extragalactic distance scale : the cepheids in ngc4414 . apj 505,207 vandenberg d.a . , stetson p.b . , bolte m. ( 1996 ) the age of the galactic globular cluster system .
araa 34,461 van den bergh s. ( 1996 ) the luminosities of type ia supernovae derived from cepheid calibration .
apj 472,431 walker a.r .
( 1992 ) the absolute magnitudes of lmc rrlyrae variables and the ages of galactic globular clusters .
apj 390,l81 wang l. , hflich p. , wheeler j.c .
( 1997 ) supernovae and their host galaxies .
apj 483,l29 wheeler j.c . , harkness r.p .
( 1990 ) classification of supernovae . in : petschek a.g .
( ed . ) `` supernovae '' .
a&a library , springer 1990 whitmore b.c .
( 1996 ) globular clusters as distance indicators . in : livio m. , donahue m. , panagia n. ( eds . ) the extragalactic distance scale .
stsci symp .
10 , cambridge university press woosley s.e .
( 1990 ) type i supernovae : carbon deflagration and detonation . in : petschek a.g .
a&a library , springer , ney york
we thank wilhelm seggewiss for many useful discussions and the permission to use his literature compilation .
we also thank gordon ogilvie for manuscript reading and bruno leibundgut for helpful and improving comments . | type ia supernovae ( sne ) are the best standard candles available today in spite of an appreciable intrinsic variation of their luminosities at maximum phase , and of probably non - uniform progenitors . for an unbiased use of type ia sne as distance indicators it is important to know accurately how the decline rate and colour at maximum phase correlate with the peak brightness . in order to calibrate the hubble diagram of type ia sne ,
i.e. to derive the hubble constant , one needs to determine the absolute brightness of nearby type ia sne .
globular cluster systems of early type ia host galaxies provide suitable distance indicators .
we discuss how ia sne can be calibrated and explain the method of globular cluster luminosity functions ( gclfs ) . at present , the distance to the fornax galaxy cluster is most important for deriving the hubble constant .
our present data indicate a hubble constant of @xmath0kms@xmath1mpc@xmath1 . as an appendix
, we summarise what is known about absolute magnitudes of ia s in late - type galaxies . |
nuclear mass plays an important role not only in studying the knowledge of nuclear structure @xcite , but also in understanding the origin of elements in the universe @xcite . with the construction and
upgrade of radioactive ion beam facilities , the measurements of nuclear masses have made great progress in recent years . during the last decade
, hundreds of nuclear masses were measured for the first time or with higher precisions @xcite .
the astrophysical calculations involve thousands of nuclei far from @xmath2-stability line . however , most of these nuclei are still beyond the experimental reach
. one could use the local mass relations such as the garvey - kelson ( gk ) relations @xcite and the residual proton - neutron interactions @xcite to predict unknown masses .
however , the intrinsic error grows rapidly when the local mass relations are used to predict the nuclear masses in an iterative way @xcite .
therefore , the theoretical predictions for nuclear masses are inevitable to astrophysical calculations .
the early theoretical studies of nuclear masses are mainly macroscopic models , such as the famous weizscker mass formula @xcite .
it is known that this kind of mass model neglects the microscopic effects , and hence shows systematic deviations for the nuclei near the shell closure or those with large deformations . in order to better describe the nuclear ground - state properties ,
the macroscopic - microscopic and microscopic theoretical models are developed for mass predictions . by including the microscopic correction energy to the macroscopic mass formula , the macroscopic - microscopic mass model can well take into account the important microscopic corrections . during the past decades ,
a number of macroscopic - microscopic mass models have been developed , such as the finite - range droplet model ( frdm ) @xcite , the extended thomas - fermi plus strutinsky integral ( etfsi ) @xcite , and the koura - tachibana - uno - yamada ( ktuy ) @xcite .
these macroscopic - microscopic mass models have similar accuracy for mass prediction and their the root - mean - square ( rms ) deviation with respect to data in the atomic mass evaluation of 2012 ( ame12 ) @xcite is about @xmath3 mev .
guiding by the skyrme energy density functional , a semiempirical nuclear mass formula , the weizscker - skyrme ( ws ) model , was proposed based on the macroscopic - microscopic method @xcite . for the latest version of ws model ( ws3 ) @xcite , the rms deviation with respect to @xmath4 known nuclear masses in ame12 is significantly reduced to @xmath5 mev .
on the other hand , great progress has been achieved for microscopic mass models with the rapid development of the computer technology in the new century .
based on the hartree - fock - bogoliubov ( hfb ) theory with skyrme or gogny force , a series of microscopic mass models have been proposed with the accuracy comparable with the traditional macroscopic - microscopic mass models @xcite .
apart from the non - relativistic microscopic model , the relativistic mean - field ( rmf ) model has also received wide attention due to many successes achieved in describing lots of nuclear phenomena @xcite as well as successful applications in astrophysics @xcite .
a systematic study of the ground - state properties for all nuclei from the proton drip line to the neutron drip line with @xmath6 and @xmath7 was performed for such model several years ago , and the rms deviation with respect to known masses is about @xmath8 mev @xcite .
however , it should be noted that the effective interaction of this rmf mass model was only optimized with the properties of a few selected nuclei . by carefully adjusting the effective interaction of rmf model with the properties of more selected nuclei ,
the deviation can be remarkably reduced . for the @xmath9
even - even nuclei with @xmath10 , the rms deviation with respect to known masses in atomic mass evaluation of 2003 ( ame03 ) is reduced to @xmath11 mev for the effective interaction pc - pk1 @xcite . moreover , the pc - pk1 predictions well reproduce the new and accurate mass measurements from sn to pa @xcite with the rms deviation of @xmath12 mev @xcite , and also successfully describe the coulomb displacement energies between mirror nuclei @xcite .
in addition , inspired by the shell model , the duflo - zuker ( dz ) mass model @xcite has made considerable success in describing nuclear masses with accuracy of about @xmath13 mev .
although these theoretical models can well reproduce the experimental data , there are still large deviations among the mass predictions of different models , even in the region close to known masses .
a number of investigations on the accuracy and predictive power of these nuclear mass models have been made so far in the literatures , e.g. refs .
@xcite . to further improve the accuracy of nuclear mass model , the image reconstruction techniques based on the fourier transform
is applied to the nuclear mass models and significantly reduces the rms deviation to the known masses with the clean algorithm @xcite .
later on , the radial basis function ( rbf ) approach was developed to improve the mass predictions of several theoretical models @xcite . comparing with the clean reconstruction ,
the rbf approach more effectively reduces the rms deviations with respect to the masses first appearing in ame03 @xcite . to improve the mass prediction of a nucleus
, thousands of nuclei with known masses are involved in the rbf approach @xcite . however , do all the nuclei involved play effective roles in the improvement of mass prediction for this nucleus ? what are the key nuclei that have to be included in the rbf approach ?
in other words , how far away from the measured region of nuclear mass could we predict with satisfactory accuracy in the rbf approach ?
these questions were not addressed in previous investigations @xcite .
therefore , it is interesting to investigate the mass correlations between a certain nucleus and those nuclei involved in the rbf approach , and hence to evaluate the predictive power of the rbf approach . in this work , we will carefully evaluate the predictive power of the rbf approach based on @xmath0 widely used nuclear mass models , ranging from macroscopic - microscopic to microscopic types .
special attention will be paid to the mass correlations among various nuclei .
the paper is organized as follows . in sec .
ii , a brief introduction to the rbf approach including numerical details is given . in sec .
iii , the mass correlations are first carefully investigated , and then the predictive power of the rbf approach based on different mass models will be evaluated .
finally , the summary is presented in sec .
the rbf approach has been widely applied in surface reconstruction and its solution is written as @xmath14 where @xmath15 denotes the point from measurement , @xmath16 is the weight of center @xmath15 , @xmath17 is the radial basis function , @xmath18 is the euclidean norm , and @xmath19 is the number of the data to be fitted . given @xmath19
samples @xmath20 , one wishes to reconstruct the smooth function @xmath21 with @xmath22 , i.e. , @xmath23 where @xmath24 . then the rbf weights are determined to be @xmath25 once the weights are obtained with the @xmath19 samples @xmath20 , the reconstructed function @xmath21 can be calculated with eq .
( [ eq : sx ] ) for any point @xmath26 . as in ref .
@xcite , the euclidean norm is defined to be the distance between nuclei @xmath27 and @xmath28 in nuclear chart : @xmath29 the basis function @xmath30 is adopted in this work , since the mass deviation can be reconstructed relatively better with @xmath30 than other basis functions @xcite
. then the mass difference @xmath31 between the experimental data @xmath32 and those predicted with nuclear mass models @xmath33 could be reconstructed with eq .
( [ eq : dx ] ) .
once the weights are obtained , the reconstructed function @xmath34 can be calculated with eq .
( [ eq : sx ] ) for any nucleus @xmath35 .
then the revised mass for nucleus @xmath35 is given by @xmath36 for training the rbf with eq .
( [ eq : dx ] ) , only those nuclei between the minimum distance @xmath37 and maximum distance @xmath38 are involved , i.e. @xmath39 . if the reconstructed function @xmath34 for a nucleus is obtained by training the rbf including itself , i.e. @xmath40 , it is clear that @xmath34 is just the @xmath41 and hence @xmath42 .
therefore , to test the predictive power of rbf approach , the function @xmath34 for a known nucleus should be reconstructed with @xmath43 . to evaluate the predictive power of rbf approach , the rms deviation , i.e. , @xmath44 is employed , where @xmath45 and @xmath46 are the theoretical and experimental nuclear masses , respectively , and @xmath47 is the number of nuclei contained in a given set . in this investigation , we only consider nuclei with @xmath48 and @xmath49 and the experimental data are taken from ame12 @xcite , unless otherwise specified . for the theoretical mass models , we take rmf @xcite , hfb-21 @xcite , dz10 @xcite , dz31 @xcite , etfsi-2 @xcite , ktuy @xcite , frdm @xcite , and ws3 @xcite mass models as examples , with the rms deviation spanning from @xmath50 mev to @xmath51 mev with respect to experimental data in ame12 . for convenience ,
the @xmath52 and @xmath53 of nuclear mass models are denoted as @xmath54 and @xmath55 , hereafter .
.the rms deviations in mev between known masses in ame12 and predictions of various nuclear mass models without ( the second column ) and with ( the third column ) the rbf approach .
the fourth column gives the reduction of the rms deviations after combining the rbf approach . [ cols="^,^,^,^,^,^,^",options="header " , ] .
the squares and circles represent the experimental uncertainties in ame03 and ame12 , respectively.,title="fig:",width=283 ] + by using the systematic trends in the mass surface and its derivative , the mass evaluation method in ame provides the best short - range mass extrapolation @xcite .
therefore , it is interesting to compare the accuracy between the rbf approach and the method used in ame . taking the nuclei whose
masses are evaluated values in ame03 ( marked by # " in the mass table of ame03 ) as an example , the rms deviations with respect to the new experimental data in ame12 are @xmath56 mev . for comparison with the method in ame at the same foot , the experimental data employed in training the rbf should be taken from ame03 as well , and the corresponding results based on different mass models are given in the third column of table [ tb2 ] .
in addition , the rms deviations @xmath57 for these nuclei are given in the second column of table [ tb2 ] .
clearly , the rbf approach also remarkably improves the predictive power of various mass models .
it should be pointed out that the rms deviation @xmath58 based on the ws3 mass model is even smaller than that from the method in ame . in fig .
[ fig6 ] , the experimental uncertainties in ame12 and ame03 are presented as a function of the isospin asymmetry @xmath59 . it is found that the experimental uncertainties are significantly improved in ame12 comparing with those in ame03 , especially for those nuclei around the border region of experimental data with @xmath60 and @xmath61 . therefore , we further update the data in training the rbf with those in ame12 and the corresponding results are shown in the fourth column of table [ tb2 ] .
it is clear that @xmath58 can be significantly reduced for most mass models .
based on the ws3 mass model , the predictive accuracy has been reduced to @xmath62 mev .
therefore , the new high - precision experimental data are also very important to improve the nuclear mass models with rbf approach . based on the measured masses in the ame12 for the rmf [ panel ( a ) ] and ws3 [ panel ( b ) ] mass models .
the dotted lines denote the magic numbers.,title="fig:",width=321 ] + between those based on the measured masses in the ame12 and those based on the measured masses in ame03 .
the panels ( a ) and ( b ) correspond to the rmf and ws mass models , respectively .
the new " nuclei in ame12 are indicated by the black contours.,title="fig:",width=321 ] + to investigate the radial basis function corrections in detail , the reconstructed functions @xmath34 based on the measured masses in the ame12 are shown in fig .
[ fig7 ] by taking the rmf and ws3 mass models as examples .
it is clear that the reconstructed functions are sensitive to the the nuclear mass models . for the rmf mass model ,
the @xmath34 are about @xmath63 mev for the nuclei near @xmath64 , @xmath65 , @xmath66 , and @xmath67 .
this just corresponds to the nuclei whose masses are underestimated in the rmf model @xcite .
moreover , the overestimation of nuclear masses in the regions near @xmath68 and @xmath69 in rmf model are also well improved by the rbf approach with the @xmath70 mev in these two regions . for the ws3 mass model
, it better describes the nuclear masses than the rmf mass model , while there still exists small but systematically correlated errors @xcite . with the rbf approach
, these systematic correlations can be well extracted as well . to further investigate the influence of the new " masses in ame12 on improving the nuclear mass models with the rbf approach , the differences of the reconstructed functions @xmath71 between those based on the measured masses in the ame12 and those based on the measured masses in ame03 are shown in fig .
it is found that the differences of reconstructed functions @xmath71 are generally within @xmath72 kev for most nuclei , while it is relatively larger for those nuclei around the border region of experimental data , especially for those new " nuclei in ame12 . this can be well understood since
significant improvements in the mass measurements are made for nuclei near the border region in recent years @xcite .
therefore , the rbf approach is sensitive to the experimental masses and it is necessary to adopt the high - precision experimental data to improve the nuclear mass models .
in this work , the mass correlations in the radial basis function ( rbf ) approach are carefully investigated based on @xmath0 widely used nuclear mass models , ranging from macroscopic - microscopic to microscopic types .
the mass correlations usually exist between a nucleus and its surrounding nuclei with distance @xmath73 .
however , the correlation distance is dependent on the nuclear mass models , which can go up to the distance of @xmath74 for the mass models with larger rms deviations , such as the rmf model . to extract these mass correlations
, it is shown that the nuclei at distance @xmath75 are necessary to include in the training of rbf approach . in this way
, the rbf approach can make significant improvements in the mass predictions for different mass models .
the ame95 - 03 - 12 test further shows that the rbf approach provides a very effective tool to improve mass predictions significantly in regions not far from known nuclear masses .
based on the latest weizscker - skyrme mass model , the rbf approach can achieve an accuracy comparable with the extrapolation method used in atomic mass evaluation , which can be further improved by the incorporation of new measurements . as claimed in the introduction , the effective interaction pc - pk1 remarkably improves the mass prediction of the rmf model .
therefore , it is interesting to investigate the predictive power of pc - pk1 mass model with the help of rbf approach when the calculated masses with pc - pk1 for all nuclei in ame12 are available in the future .
in addition , considering the success in improving the nuclear mass predictions , the rbf approach has a great potential to improve theoretical calculations of other physical quantities , such as nuclear @xmath2-decay half - lives , fission barriers , and excitation spectra .
this work was partly supported by the national natural science foundation of china ( grants no .
11205004 , no .
11235002 , no .
11175001 , no . 11105010 , no .
11128510 , and no .
11035007 ) , the 211 project of anhui university under grant no
. 02303319 - 33190135 , the program for new century excellent talents in university of china under grant no .
ncet-09 - 0031 , the key research foundation of education ministry of anhui province of china under grant no .
kj2012a021 , and the natural science foundation of anhui province under grant no .
11040606m07 .
99 d. lunney , j. m. pearson , c. thibault , rev .
phys . * 75 * , 1021 ( 2003 ) .
e. margaret burbidge , g. r. burbidge , william a. fowler , and f. hoyle , rev .
* 29 * , 547 ( 1957 ) .
a. arcones and g. f. bertsch , phys .
lett . * 108 * , 151101 ( 2012 ) .
g. audi , f. g. kondev , m. wang , b. pfeiffer , x. sun , j. blachot , and m. maccormick , chin .
c * 36 * , 1157 ( 2012 ) . g. t. garvey and i. kelson , phys .
* 16 * , 197 ( 1966 ) .
g. t. garvey , w. j. gerace , r. l. jaffe , i. talmi , and i. kelson , rev .
phys . * 41 * , s1 ( 1969 ) .
j. y. zhang , r. f. casten , and d. s. brenner , phys .
b227 * , 1 ( 1989 ) .
g. j. fu , h. jiang , y. m. zhao , s. pittel , and a. arima , phys .
c * 82 * , 034304 ( 2010 ) .
h. jiang , g. j. fu , y. m. zhao , and a. arima , phys .
c * 82 * , 054317 ( 2010 ) .
et al_. , phys .
c * 85 * , 054303 ( 2012 ) g. j. fu , y. lei , h. jiang , y. m. zhao , b. sun , and a. arima , phys .
c * 84 * , 034311 ( 2011 ) .
i. o. morales , j. c. lopez vieyra , j. g. hirsch , and a. frank , nucl .
a828 * , 113 ( 2009 ) . c. f. von weizscker , z. phys . *
96 * , 431 ( 1935 ) .
p. mller , j. r. nix , w. d. myers , and w. j. swiateck , atom .
data nucl .
data tables * 59 * , 185 ( 1995 ) .
s. goriely , aip conf
529 * , 287 ( 2000 ) .
h. koura , t. tachibana , m. uno , and m. yamada , prog .
theor . phys . *
113 * , 305 ( 2005 ) .
n. wang , m. liu , and x. z. wu , phys .
c * 81 * , 044322 ( 2010 ) .
n. wang , z. y. liang , m. liu , and x. z. wu , phys .
c * 82 * , 044304 ( 2010 ) .
m. liu , n.wang , y. deng , and x. wu , phys .
c * 84 * , 014333 ( 2011 ) .
s. goriely , n. chamel , and j. m. pearson , phys .
lett . * 102 * , 152503 ( 2009 ) .
s. goriely , s. hilaire , m. girod , and s. pru , phys .
lett . * 102 * , 242501 ( 2009 ) .
s. goriely , n. chamel , and j. m. pearson , phys .
c * 82 * , 035804 ( 2010 ) .
j. meng , h. toki , s. g. zhou , s. q. zhang , w. h. long , and l. s. geng , prog .
* 57 * , 470 ( 2006 ) .
d. vretenar , a. v. afanasjev , g. a. lalazissis , and p. ring , phys . rep . * 409 * , 101 ( 2005 ) .
p. w. zhao , z. p. li , j. m. yao , and j. meng , phys .
c * 82 * , 054319 ( 2010 ) .
x. m. hua , t. h. heng , z. m. niu , b. h. sun , and j. y. guo , sci .
china phys .
astron . * 55 * , 2414 ( 2012 )
. s. g. zhou , j. meng , p. ring , and e .-
zhao , phys .
c * 82 * , 011301(r ) ( 2010 ) .
w. h. long , p. ring , n. van giai , and j. meng , phys .
c * 81 * , 024308 ( 2010 ) .
z. m. niu , q. liu , y. f. niu , w. h. long , and j. y. guo , phys .
c * 87 * , 037301 ( 2013 ) .
h. mei , j. xiang , j. m. yao , z. p. li , and j. meng , phys .
c * 85 * , 034321 ( 2012 ) .
h. z. liang , n. van giai , and j. meng , phys .
* 101 * , 122502 ( 2008 ) .
y. f. niu , n. paar , d. vretenar , and j. meng , phys .
* b681 * , 315 ( 2009 ) .
d. vretenar , y. f. niu , n. paar , and j. meng , phys .
c * 85 * , 044317 ( 2012 ) .
h. z. liang , t. nakatsukasa , z. m. niu , and j. meng , phys .
c * 87 * , 054310 ( 2013 ) .
z. m. niu , y. f. niu , q. liu , h. z. liang , and j. y. guo , phys .
rev c * 87 * , 051303(r ) ( 2013 ) .
b. sun , f. montes , l. s. geng , h. geissel , yu .
a. litvinov , and j. meng , phys .
c * 78 * , 025806 ( 2008 ) .
b. sun and j. meng , chin .
* 25 * , 2429 ( 2008 ) .
z. m. niu , b. sun , and j. meng , phys .
c * 80 * , 065806 ( 2009 ) .
y. f. niu , n. paar , d. vretenar , and j. meng , phys . rev .
c * 83 * , 045807 ( 2011 ) .
z. m. niu and c. y. gao , int .
e * 19 * , 2247 ( 2010 ) .
j. meng , z. m. niu , h. z. liang , and b. sun , sci .
chin . ser .
g * 54 * , 119 ( 2011 ) .
w. h. zhang , z. m. niu , f. wang , x. b. gong , and b. sun , acta phys . sin .
* 61 * , 112601 ( 2012 ) .
x. d. xu , b. sun , z. m. niu , z. li , y .- z .
qian , and j. meng , phys .
c * 87 * , 015805 ( 2013 ) .
z. m. niu , y. f. niu , h. z. liang , w. h. long , t. niki , d. vretenar , and j. meng , phys . lett . *
b723 * , 172 ( 2013 ) .
l. s. geng , h. toki , and j. meng , prog .
. phys . * 113 * , 785 ( 2005 ) .
q. s. zhang , z. m. niu , z. p. li , j. m. yao , and j. meng , arxiv : 1305.1736 . l. chen
_ et al_. , nucl . phys . *
a882 * , 71 ( 2012 ) .
p. w. zhao , l. s. song , b. sun , h. geissel , and j. meng , phys .
c * 86 * , 064324 ( 2012 ) .
b. sun , p. zhao , and j. meng , sci .
china phys .
astron . * 54 * , 210 ( 2011 ) .
j. duflo and a. p. zuker , phys .
c * 52 * , r23 ( 1995 ) .
a. p. zuker , rev .
s * 54 * , 129 ( 2008 ) .
j. mendoza - temis , i. morales , j. barea , a. frank , j. g. hirsch , j. c. lpez vieyra , p. van isacker , v. velzquez , nucl
* a812 * , 28 ( 2008 ) .
a. litvinov , a. sobiczewski , a. parkhomenko , and e. a. cherepanov , int . j. mod
e * 21 * , 1250038 ( 2012 ) .
a. sobiczewski and yu .
a. litvinov , phys .
scr . * t154 * , 014001 ( 2013 ) .
irving o. morales , p. van isacker , v. velzquez , j. barea , j. mendoza - temis , j. c. lpez vieyra , j. g. hirsch , and a. frank , phys .
c * 81 * , 024304 ( 2010 ) .
n. wang and m. liu , phys .
c * 84 * , 051303(r ) ( 2011 ) .
g. audi and a. h. wapstra , nucl .
* a595 * , 409 ( 1995 ) . g. audi , a. h. wapstra , and c. thibault , nucl .
a729 * , 337 ( 2003 ) .
n. wang and m. liu , j. phys .
* 420 * , 012057 ( 2013 ) . | the radial basis function ( rbf ) approach is applied in predicting nuclear masses for @xmath0 widely used nuclear mass models , ranging from macroscopic - microscopic to microscopic types .
a significantly improved accuracy in computing nuclear masses is obtained , and the corresponding rms deviations with respect to the known masses is reduced by up to @xmath1 . moreover
, strong correlations are found between a target nucleus and the reference nuclei within about three unit in distance , which play critical roles in improving nuclear mass predictions .
based on the latest weizscker - skyrme mass model , the rbf approach can achieve an accuracy comparable with the extrapolation method used in atomic mass evaluation . in addition , the necessity of new high - precision experimental data to improve the mass predictions with the rbf approach is emphasized as well . gbksong |
in this contribution we present a novel study of chaos and anomalous diffusion in the dynamics of the single - particle motion in the hamiltonian mean field ( hmf ) model .
this model describes n classical particles moving on the unit circle and interacting through an infinite range potential .
it has been recently studied both analitycally and numerically @xcite . at low energy the particles
are clustered and , as the energy is increased , the system undergoes a second - order phase transition , reaching a high temperature gaseous " phase where the particles are homogeneously distributed on the circle . the motion is chaotic and the lyapunov exponent reaches a maximum at the critical point .
particle motion is superdiffusive for a range of energies close to the critical point , but only in a transient quasi - stationary regime preceeding equilibration .
particles can evaporate from the cluster showing lvy walks .
close to the critical point , if the continuum ( @xmath0 ) limit is performed before the @xmath1 limit , canonical equilibrium is never reached and quasi - stationary superdiffusive states live forever @xcite .
though anomalous diffusion and lvy walks have been widely studied in the recent literature in models with a few degrees of freedom @xcite , there are very few studies for many - degrees - of - freedom systems @xcite .
the hmf model has the advantage of having an exact solution in the canonical ensemble , and therefore microscopic dynamics can be put in connection to macroscopic features , thus offering a very interesting perspective @xcite . in the following sections we remind the reader the details of the model and then we discuss superdiffusion in connection to microscopic chaotic dynamics .
hmf describes a system of @xmath2 classical particles moving on a circle , characterized by the angles @xmath3 and the conjugate momenta @xmath4 .
each particle interact with all the others according to the following hamiltonian : @xmath5 where @xmath6\ ] ] are the kinetic and potential energy . if we define a spin vector associated to each particle @xmath7 $ ] and
a total magnetization @xmath8 the hamiltonian then describes @xmath2 classical fully coupled spins , similarly to the xy model . with this interpretation ,
the system is a ferromagnet at low energy and shows a second - order phase transition at the critical energy @xmath9 ( corresponding to the critical temperature @xmath10 in the canonical ensemble ) @xcite ) .
we show in fig.1 the total magnetization vs. energy @xmath11 .
numerical simulations ( dots ) are compared with the canonical ensemble result ( full line ) .
pictorial views of the low energy ferromagnetic ( clustered ) phase and high energy paramagnetic ( gaseous ) phases are shown in the insets .
the equations of motion for the @xmath2 particles are @xmath12 where @xmath13 are respectively the modulus and the phase of the total magnetization vector @xmath14 .
these equations are formally equivalent to those of a perturbed pendulum . to study relaxation to canonical equilibrium , we solve these equations on the computer using fourth order symplectic algorithms ( the details can be found in refs @xcite ) .
we start the system in a given initial state and we compute @xmath15 at each time step , and from them the total magnetization @xmath16 and temperature @xmath17 ( through the relation @xmath18 ) .
we consider systems with an increasing size @xmath2 and different energies @xmath11 .
the single - particle motion is given by eq .
( [ eqmoto ] ) , if we consider the mean - field pair @xmath13 as an _ external _ time dependent forcing . in the next section
we focus our attention on the dynamics of the single - particle motion and we show that it is chaotic and superdiffusive .
in this section we discuss the numerical results . in fig .
2 we show single - particle phase spaces for three different energies : below ( 2a ) , close to ( 2b ) and above ( 2c ) the critical energy @xmath19 . while for @xmath20 the particle never escapes the cluster and moves on the chaotic web near the center of the nonlinear resonance , close to the critical point ( at @xmath21 ) it escapes frequently from the cluster and is then trapped again , exhibiting the well known phenomenon of _ separatrix crossing_. above the critical energy one has mainly a ballistic regime , separatrix crossings are rarely observed and particles interact very weakly .
we show in fig .
3 the largest lyapunov exponent @xmath22 for the single - particle motion as a function of the total energy per particle .
the motion is chaotic and has a peak at the critical point , but @xmath22 is much smaller than the largest lyapunov exponent calculated for the full dynamics in refs .
the reason is that in the calculation of the lyapunov of the full system one gets the additional contribution to the instability of the off - diagonal terms of the jacobian .
we show in the inset a log - log plot with a fit at energies less than @xmath23 , which is a power law , i.e. @xmath24 ; again quite far from the @xmath25 law found for the full dynamics .
a very weak @xmath2 dependence is present is this region . in the high energy region @xmath22
vanishes with @xmath2 , with the universal " law @xmath26 , which is found in many different contexts @xcite .
this result is displayed in fig.3(b ) for @xmath27 . .
in the inset we show the behavior at small energy in a log - log plot , see text .
( b ) we report numerical simulations ( open circles ) that show that @xmath22 scales like @xmath28 ( straight line ) .
, title="fig:",scaledwidth=50.0% ] . in the inset
we show the behavior at small energy in a log - log plot , see text .
( b ) we report numerical simulations ( open circles ) that show that @xmath22 scales like @xmath28 ( straight line ) . , title="fig:",scaledwidth=50.0% ] diffusion and transport of a particle in a medium or in a fluid flow are characterized by the mean square displacement @xmath29 in the long - time limit . in general
, one has @xmath30 with @xmath31 for normal diffusion .
all the processes with @xmath32 are termed anomalous diffusion , namely subdiffusion for @xmath33 and superdiffusion for @xmath34 . in order to study anomalous diffusion in hmf
we follow the dynamics of each particle starting the system in a `` water bag '' , i.e. a far - off - equilibrium initial condition obtained by putting all the particles at @xmath35 and giving them a uniform distribution of momenta with a finite width centered around zero .
we compute the variance of the one - particle angle @xmath36 according to the expression @xmath37 where @xmath38 indicates the average over the @xmath2 particles , and we fit the value of the exponent @xmath39 in eq .
( [ anoma ] ) . in fig .
4(a ) we show the motion of four particles in the same simulation as a function of time for @xmath40 and @xmath41 .
the figure shows that the erratic motion is not brownian and lvy walks are present .
each particle experiences in time localized motion and free walks . in fig.4(b ) we show the mean square displacement as a function of time . for energies close to the critical one
the exponent @xmath39 differs from 1 , as it should be for normal diffusion , and takes values close to 1.4 .
one has therefore superdiffusion , which however turns again into normal after a crossover time .
this behavior does not seem to depend strongly on energy , although the superdiffusion exponent @xmath39 is slightly energy dependent . in ref . @xcite we have shown that this crossover time coincides with the equilibration one and that by using the model of ref .
@xcite it is possible to interpret this dynamics in terms of lvy walks . .
( b ) variance of the single - particle displacement as a function of time for four different energies .
, title="fig:",scaledwidth=50.0% ] .
( b ) variance of the single - particle displacement as a function of time for four different energies .
, title="fig:",scaledwidth=50.0% ]
we have discussed chaotic and superdiffusive motion in a n - body hamiltonian system which exhibits a second - order phase transition .
the single - particle motion is shown to be chaotic for energies smaller than the critical one .
superdiffusion is present in a transient out - of - equilibrium regime for energies close to the critical one , where the system is strongly chaotic and quasi - stationary states exist .
the crossover time from enhanced diffusion to normal coincides with the equilibration time which diverges with @xmath2 .
this is true for fully coupled systems , but should appear also if the interaction decays with the distance among the particles with a power law @xcite .
our model lies in the class of self - consistent many - degrees - of - freedom dynamics , to which also belongs the single - wave model described in ref .
this is the closest contact of our results to the lvy - walk behavior found in fluid experiments @xcite .
we dedicate this work to the memory of giovanni paladin who was an enthusiastic pioneer in the study of chaos and statistical mechanics , and the first one to explain to one of the authors ( s.r . )
the concept of anomalous diffusion .
we thank c. anteneodo , m. antoni , m .
baranger , e. barkai , d. del - castillo - negrete , p. grigolini , a. torcini and c.tsallis for useful discussions.s.r .
thanks ens - lyon for financial support . | we discuss recent results obtained for the hamiltonian mean field model .
the model describes a system of n fully - coupled particles in one dimension and shows a second - order phase transition from a clustered phase to a homogeneous one when the energy is increased . strong chaos is found in correspondence to the critical point on top of a weak chaotic regime which characterizes the motion at low energies . for a small region around the critical point ,
we find anomalous ( enhanced ) diffusion and lvy walks in a transient temporal regime before the system relaxes to equilibrium . and hamiltonian dynamics , deterministic chaos , lyapunov exponents , relaxation to equilibrium , anomalous diffusion , lvy walks + _ pacs numbers : _ 05.45.pq , 05.20.-y , 05.60.cd , 05.70.fh |
the search for extrasolar planets has revealed that there are many terrestrial planets out of our solar system ( e.g. , lger et al .
2009 ) . by detecting these extrasolar terrestrial planets ,
many researches have tried to determine the planets thermal and orbital histories ( henning et al .
2009 ; barnes et al . 2010 ; lger et al . 2011 ; wagner et al . 2012 ) .
the thermal and orbital states of extrasolar terrestrial planets are important for the planets evolution .
in addition , the thermal and orbital states affect the habitability , making analysis of these factors of key importance .
recently , in the search for habitable extrasolar planets , m stars were suggested for observation ( scalo et al .
2007 ) because the number of m stars is much larger than that of g stars .
conventionally , m stars are not considered to be likely hosts of habitable planets because their planets orbits are too close , which results in the planets being tidally rocked ( synchronously rotated ) and their atmosphere freezing on the dark side
. however , joshi et al . (
1997 ) hypothesized that carbon dioxide can make a planet habitable .
thus , the importance of m star extrasolar planet analysis continues to increase .
focusing on the terrestrial planets in our solar system , the main heat sources are the insolation from the sun and the radioactive decay in the planets themselves . however
, if extrasolar planets orbit around small stars such as an m star , the tidal heating has a significant effect on the habitability .
compared with the sun , m stars have low temperatures , and thus the magnitude of the insolation becomes low .
therefore , the habitable zone is closer to the star .
the rate of change of a planet s orbit due to tidal dissipation increases with decreasing orbital distance .
for this reason , a planet orbiting an m star may not be able to stay within the habitable zone for a long time .
in addition , assuming that the planets orbit according to kepler s law , the tidal frequency is high , producing tidal heating that may exceed the radiogenic and insolation heating .
in contrast to the terrestrial planets in our solar system , such tidal heating is an important factor for the thermal and orbital evolutions of planets orbiting m stars .
the effects of tidal heating on terrestrial planets orbiting m stars have been analyzed ( jackson et al . 2008 ; barnes et al . 2008 ) .
although these researches revealed that tidal heating influences a planet habitability , they assumed that the magnitude of the dissipation of the planet is constant .
the tidal heating mechanism is complicated and interacts with the internal thermal state and orbit of the planet ( e.g. , hussmann & spohn 2004 ) .
when a planet is tidally heated , structure and thermal states of the planet changes , which affects the tidal heating by the feed - back .
in addition , the orbit and heating rates also interact with each other .
tidal heat dissipation is dependent on the eccentricity of the orbit ; the eccentricity decreases with heating rate because orbital energy is lost as tidal heat .
thus , tidal heating should be coupled with orbital and thermal evolution . in this work
, we evaluate the coupled calculation of tidal heating considering the thermal and orbital states .
however , coupled calculations are very complex . in addition
, the internal structure of extrasolar terrestrial planets is not constrained well .
thus , as a fist step , we consider a martian - sized terrestrial planet .
an advantage of this choice is that the internal pressure of planets of this size is small compared with earth - sized planets , due to their small radii , which allows a simple model of the internal structure to be used
. a martian - sized planet has been discovered ( muirhead et al . 2012 ) ; however , this planet is outside the habitable zone . despite this
, we can expect similarly sized planets to be discovered within habitable zones in future observations .
considering the observation , earth - sized planets are easier to detect compared to martian - sized planet .
however , we try the coupling theoretical calculation with martian structure . in our solar system , io is a tidally locked rocky satellite with high tidal heating ( segatz et al .
thus , we reference the thermal model of io as well as mars .
if accurate internal structures of super - earths are derived in the future , we can apply the results of this work to such models with more complicated structures .
in this work , for simplicity , we assume that the internal structure of the planet is composed of three layers : a liquid core , a mantle and a stagnant lid ( fig .
[ structure ] a ) .
stagnant lid convection occurs in the mantle . because stagnant lid convection is the most typical convection system , scaling laws which relate the magnitude of the convection to the structure have been analyzed by many researches ( e.g. , reese et al .
although mobile lid convection ( e.g. , oneill & lenardic 2007 ) or heat pipe state ( moore 2001 ; moore & webb 2013 ) are suggested as alternative heat transfer mechanisms , we assume the stagnant lid convection as the heat transfer model in this work .
rheology and temperature distributions depend only on the radial distance ( i.e. , are independent of latitude and longitude ) . the mantle is differentiated into three parts : two thermal boundaries and a well mixed convective section ( fig .
[ structure ] b ) . at the two boundaries ,
heat is transferred by conduction . in this work ,
we define the lid and the upper boundary of the mantle as the lithosphere .
the lithosphere is assumed to be elastic , and tidal heat is generated only in the mantle by viscoelastic dissipation ( hussmann & spohn 2004 ) .
this assumption is valid because the surface temperature is less than 300 k in the habitable zone , which is much smaller than the mantle temperature ( more than 1600 k ) . in the convective part , for simplicity
, we assume that convective temperature is @xmath0 and independent of depth . @xmath0
changes with time both by the heating and cooling .
the rate of change of the mantle and core temperatures @xmath0 and @xmath1 are calculated based on the energy balance ( breuer & spohn 2006 ) .
@xmath2 @xmath3 here , the subscript @xmath4 and @xmath5 refer to properties of the mantle and core , respectively . @xmath6 and @xmath7 are the densities , @xmath8 and @xmath9 are the heat capacities , @xmath10 and @xmath11 are the volumes , @xmath12 and @xmath13 are the radial distances from the center to the convective mantle and the core , respectively .
@xmath14 is the ratio between the average mantle and upper boundary temperatures ( grott & breuer 2008 b ) . * in the case of adiabatic thermal model , temperature changes with the depth .
thus , effect of @xmath15 is required .
@xmath15 ranges between 1.0 and 1.05 ( breuer & spohn 2006 ; grott & breuer 2008 b ) . because this range is around 1.0 , eq .
( [ tm ] ) is applicable to the isothermal mantle model which we consider in this work .
we conducted the calculations with different values of @xmath15 , and confirmed that the value of @xmath14 little affects the results .
* @xmath16 is the ratio between the average core and lower boundary temperatures , which is 1.1 ( breuer & spohn 2006 ; grott & breuer 2008 a ) .
the values of each parameter are shown in table [ value ] .
@xmath17 is the stefan number , which accounts for the consumption or release of latent heat by the mantle ( breuer & spohn 2006 ) .
@xmath17 is represented as the ratio of latent heat to the heat capacity of the mantle , and is given by @xmath18 where @xmath19 is the latent heat of the mantle ( breuer & spohn 2006 ) .
@xmath20 is the * volume * melt fraction , which is discussed below .
@xmath21 is the volume of the melt zone .
* effect of melt in the rock has been studied by many researches ( e.g. , rosenberg & handy 2005 ) .
when mantle begins to melt , melt is distributed to the frame and junctions of grain boundaries . by increasing the melt fraction ,
melts among grains begin to be interconnected each other .
however , as long as the melt fraction is under the critical value called the rheological critical melt percentage ( rcmp ) , grain framework is maintained ( rosenberg & handy 2005 ) .
when the melt fraction exceeds the rcmp , melt covers the grain completely and effect of liquid dominates .
considering these effects , if the melt fraction is under the rcmp , we assume that the convective area of the mantle is well mixed and that melt is not separated from the solid mantle ; thus , @xmath22 .
* this assumption may not be valid if the melt fraction exceeds the rcmp because the generated melt segregates from the solid mantle .
however , we found that melt fraction of our coupled calculation is less than the 25 % at every condition , and moore ( 2003 ) estimates that the segregation of melt occurs at the 40 % to 60 % of melt fraction .
thus the assumption of mantle with partial melt state is reasonable .
@xmath23 is the heating due to radioactive decay . for simplicity , we use the averaged radiogenic heating in the martian mantle estimated by grott & breuer ( 2008 b ) . the volumetric radiogenic heating @xmath24 is given by @xmath25 where @xmath26 w m@xmath27 and @xmath28 s@xmath29 ( grott & breuer 2008 b ) .
thus , @xmath30 . we do not consider the radiogenic heating from other parts of the planet . in the core of the earth
, potassium can be the large heat sources in the core , and mars may contain potassium , but the stability of potassium in iron is extremely small at low pressure such as the case for the martian core ( gessmann & wood 2002 ) .
thus , we ignored the heat generation in the core .
* @xmath31 is the heat flux from the mantle , which is given by ( nimmo & stevenson 2000 ) . in the case of stagnant lid convection
, @xmath31 can be related to the nusselt number @xmath32 as * @xmath33 * where @xmath34 is the thermal conductivity of the mantle .
@xmath35 is the temperature difference between top and bottom of the mantle .
@xmath36 is the thickness of the mantle . assuming the newtonian viscosity , @xmath32 can be scaled by the rayleigh number @xmath37 and frank - kamenetskii parameter @xmath38 ( reese et al .
1999 a ) as * @xmath39 * @xmath37 is given by *
@xmath40 * where @xmath41 , @xmath42 and @xmath43 are acceleration due to gravity , thermal expansivity and thermal diffusivity of the mantle , respectively .
@xmath44 is the viscosity of the mantle , which is discussed in the next section . in this work
, we assume the arrhenius type viscosity , which mainly depends on the temperature .
@xmath38 is given by *
@xmath45 * where @xmath46 and @xmath47 are the activation energy and the gas constant , respectively .
@xmath48 is the internal temperature and defined as the mean temperature at the upper thermal boundary ( reese et al .
1999 b ) . using eqs .
( [ flux])-([theta ] ) @xmath31 is given ( nimmo & stevenson 2000 ) as * @xmath49 where @xmath50 using the typical value of @xmath48=1800 k and @xmath51 kj / mol , @xmath52 becomes 0.011 ( nimmo & stevenson 2000 ) . in order to calculate the accurate value of @xmath52 , numerical calculation of convection
is required ( reese et al .
1999 b ) . however , in this work , we fixed @xmath52 at 0.011 for simplicity .
the heat flux from the core to the mantle @xmath53 is given by @xmath54 where @xmath55 is the thickness of the lower boundary .
thickness of the upper and lower thermal boundaries ( @xmath56 and @xmath55 ) are derived by nimmo & stevenson ( 2000 ) . through the upper thermal boundary , temperature drops by 8/@xmath52 and @xmath31 of heat flux
is transported from the mantle to the surface ( nimmo & stevenson 2000 ) .
they mention that 8/@xmath52 of temperature drop is found by solomatov ( 1995 ) .
thus , @xmath56 is given by @xmath57 in the case of the thickness of the lower boundary @xmath55 , they assume that local rayleigh number of the lower thermal boundary is the same to that of the upper thermal boundary . under this assumption , local rayleigh numbers are related as @xmath58)\kappa\eta_m}=ra_c , \label{du_db}\ ] ] where @xmath59 and @xmath60 are the napier s constant and critical rayleigh number ( nimmo & stevenson 2000 ) . from eq .
( [ du_db ] ) , the ratio between @xmath55 and @xmath56 is given ( nimmo & stevenson 2000 ) as @xmath61)^{-1/3}\exp\left(-\frac{\gamma(t_c - t_m)}{6}\right ) .
\label{ratio}\ ] ] @xmath55 can be calculated from eqs .
( [ d_u ] ) and ( [ ratio ] ) .
the thickness of the lithosphere @xmath62 and the surface heat flux @xmath63 are related by @xmath64 where @xmath65 and @xmath66 are the thermal conductivity of the lithosphere and the surface temperature , respectively . in this work
, we assume that tidal and radiogenic heatings are generated in the mantle .
* heat generation and loss in the lithosphere are ignored .
* under this assumption , @xmath63 equals @xmath31 .
thus , from eqs ( [ qm ] ) and ( [ qs ] ) , the lithosphere thickness @xmath62 can be calculated , which is required for the calculation of tidal heating .
if the surface of the planet is assumed to be in equilibrium with the insolation heat flux uniformly distributed over the surface , the surface temperature @xmath66 is given by ( lger et al .
2009 ; barnes et al . 2010 ) @xmath67 if the heat transport is limited to the light side , @xmath68 @xmath69 is the albedo of the planet , which is assumed to be constant at 0.3 ( paige et al .
@xmath70 is the effectiveness of the heat transport .
@xmath71 is the orbital distance , which is discussed in sec .
@xmath72 is the radius of the star , which can be related to the mass of the star through @xmath73 where @xmath74 and @xmath75 are the mass and radius of the sun , respectively ( gorda & svechnikov 1999 ) .
@xmath76 is the mass of the star .
@xmath77 in eq .
( [ ts ] ) is the effective temperature of the star , which is given by @xmath78 where @xmath79 and @xmath80 are the stefan - boltzman constant and luminosity of the star , respectively ( barnes et al .
@xmath80 is related to the mass of the star by @xmath81 where @xmath82 ( scalo et al .
@xmath83 is the luminosity of the sun .
using eqs .
( [ ts])-([l_st ] ) , @xmath66 can be determined .
we use the constant values as shown in table [ value ] except for @xmath76 and @xmath71 . @xmath71
changes with time according to the evolution of the orbit . in the case of @xmath76 , we conduct the calculations by varying @xmath84 from 0.1 to 0.5 , which is a reasonable range for an m star . melt fraction @xmath20 depends on the solidus and liquidus temperatures .
the melt fraction is assumed to be 1.0 at the liquidus temperature and increases linearly from the solidus to liquidus temperature ( moore 2003 ) .
the solidus @xmath85 and liquidus @xmath86 temperatures of the mantle are given as functions of pressure @xmath87 ( takahashi 1990 ) : @xmath88 @xmath89 pressure depends on the depth @xmath90 and can be calculated as @xmath91 where @xmath92 is the radius of the planet .
@xmath93 is the density at @xmath90 .
@xmath94 is the acceleration due to gravity at @xmath90 , which can be expressed as @xmath95 where @xmath96 and @xmath97 are radial distance ( @xmath96=@xmath92@xmath90 ) and the gravitational constant , respectively .
@xmath98 is the mass enclosed by a sphere of radius @xmath96 , which is @xmath99 we assume that the solidus and liquidus temperatures of the mantle are averaged between the top and bottom of the mantle ( hussmann & spohn 2004 ) : @xmath100 @xmath101 when @xmath0 exceeds @xmath102 , the mantle begins to melt .
thus , @xmath20 can be expressed as @xmath103 for the calculation of the tidal heating , the mantle is assumed to behave viscoelastically .
the tidal heating rate @xmath104 can be obtained as follows ( segatz et al .
1988 ) .
@xmath105 @xmath106 is the mean motion of the planet . assuming that the planet orbits according to kepler s law , @xmath106 can be related to the orbital distance @xmath71 by @xmath107 .
@xmath59 is the eccentricity of the orbit , which is important for the coupled calculation because it interacts with the tidal heating .
the evolution of @xmath59 is discussed in sec .
@xmath108 is the second degree complex love number , which largely depends on the internal structure and rheology of the planet . in the case of spherical - shell structured models
, the complex love number can be calculated using the methods based on the gravity potential of the planet ( segatz et al .
1988 ; tobie et al . 2005 ; roberts & nimmo 2008 ) . detailed theory and the precise method are reviewed by tobie et al .
( 2005 ) and roberts & nimmo ( 2008 ) . here
, we mention the outline . assuming the spherical - shell structure , displacement , stress and potential
can be separated into radial function part and spherical harmonics ( alterman et al . 1959 ; takeuchi & saito 1972 ) .
assuming the radial functions of displacement ( @xmath109 and @xmath110 ) , stress ( @xmath111 and @xmath112 ) , and gravity potential ( @xmath113 and @xmath114 ) of a spherical shell model , they are related each other via propagation matrix @xmath115 as @xmath116 where @xmath96 is the radial distance from the center of the planet . for bodies with liquid cores , sabadini & vermeersen ( 2004 )
derived the components of @xmath115 , which we use in this work .
each radial function is solved by the boundary conditions at the core and the surface .
love number of the spherical shell structure can be calculated ( tobie et al . 2005 ) by @xmath117 note that love number in eq .
( [ love ] ) is not complex number because the radial function theory is developed for the elastic body .
however , viscoelastic rheology model can be related to the elastic model by fourier transformation , which is called correspondence principle ( bilot , 1954 ) . by the fourier transformation ,
radial functions of viscoelastic model becomes complex value in the domain of the frequency of applied force .
complex radial functions also can be related linearly via propagation matrix although each component becomes complex number .
by solving the radial functions by the viscoelastic model , complex love number can be calculated as @xmath118 where @xmath119 is the complex radial function of potential . as for the rheology
, we consider the andrade model . an andrade body is proven to fit well to the experimental results for the response of mantle rock ( jackson et al . 2002 ) .
the creep function for an andrade body @xmath120 is given by @xmath121 where @xmath122 , @xmath123 and @xmath44 are time , the shear modulus and viscosity of the mantle , respectively ( jackson et al . 2002 ; mccarthy & castillo - rogez 2013 ) . @xmath124 and @xmath125 are the experimental parameters , which represent the effects of the anelasticity of the material such as heterogeneity of the grain .
if @xmath125 is zero , the creep function becomes the conventional viscoelastic rheology ( maxwell model ) .
@xmath125 can be expressed as ( castillo - rogez et al .
2011 ) @xmath126 this relationship is valid if the dissipation is dominated mainly by diffusion creep ( castillo - rogez et al .
@xmath124 is the unconstrained empirical parameter , which ranges from 0.2 to 0.5 . in this work ,
we use @xmath127 , which is based on the laboratory expedient of olivine ( jackson et al .
creep function at the frequency domain can be given ( mccarthy & castillo - rogez 2013 ) by the fourier transformation as @xmath128 , \label{creep_complex}\ ] ] where @xmath129 is the frequency , which is the same to the mean motion in the case of tidal heating .
@xmath130 and @xmath131 are the gamma function and @xmath132 , respectively .
complex shear modulus @xmath133 of the andrade model can be calculated as 1/@xmath134 .
propagation matrix @xmath115 contains the complex shear modulus , which affects the tidal dissipation of the planet . in the calculation of the tidal heating ,
the shear modulus and viscosity of the mantle are the most important factors . at temperatures under @xmath102
, the shear modulus is assumed to be constant at 50 gpa , which is a typical value for mantle rock . as for the viscosity
, we adopt an arrhenius function , which is the typical description for the mantle viscosity ( e.g. , breuer & spohn 2006 ) : @xmath135\right ) .
\label{eta}\ ] ] @xmath136 k is the reference temperature ( breuer & spohn 2006 ) and @xmath137 is the reference viscosity at @xmath138 .
when the temperature becomes higher than @xmath102 , the mantle begins to melt . if the melt is contained within the mantle , the shear modulus and viscosity drop drastically .
as mentioned above , we assume that the partially molten mantle is well mixed .
reduction of the shear modulus and viscosity is controlled by the melt fraction @xmath20 .
moore ( 2003 ) gives the viscosity over the solidus temperature .
when the melt fraction @xmath20 is under rcmp ( 40% to 60% ) , the viscosity of the mantle is multiplied by @xmath139 where @xmath140 is the dimensionless melt fraction coefficient .
thus , @xmath44 becomes @xmath135\right)\exp(-b\phi ) .
\label{eta_melt}\ ] ] in this work , we assume that rcmp is 40 @xmath141 . if the melt fraction exceeds the disaggregation point , the viscosity follows the rascoe - einstein relationship , which can be written ( moore 2003 ) as @xmath142 the shear modulus over the solidus temperature was determined by fischer and spohn ( 1990 ) to be @xmath143 pa , where @xmath144 k. they also found that @xmath145 due to the continuous condition at the solidus temperature ( 1600 k in their work ) . in our calculation , the solidus temperature was calculated using eq .
( [ t_sol_ave ] ) .
thus , the shear modulus over the solidus temperature is @xmath146 } ~~\mathrm{pa } , \label{mu_melt}\ ] ] where @xmath147 .
in addition , if @xmath20 exceeds the disaggregation point ( 0.4 ) , the shear modulus is assumed to drop to @xmath148 pa ( moore 2003 ) .
figure [ heat ] shows the tidal heating rate in the case of two layers ( liquid core and mantle ) as a function of viscosity and shear modulus of the mantle .
heating rates are calculated by eqs .
( [ tide])-([creep_complex ] ) with @xmath149 and @xmath150 . in the case of a 50 gpa shear modulus ,
the heating rates are maximized at a viscosity of @xmath151 pa s. if we assume that the initial viscosity is @xmath152 pa s , two types of heating state are possible .
one evolution type is runaway cooling .
if the heat radiated from the surface is larger than the tidal heating , the mantle cools and the viscosity increases . when the viscosity increases , the heating rates decrease as shown in fig . [ heat ] .
thus , mantle cools more .
the other heating state is runaway heating . if the tidal heating is large , the mantle temperature increases and the viscosity decreases according to eq .
( [ eta ] ) , which results in increased tidal dissipation .
however , runaway heating stops when the solidus temperature is reached . because of the melt , the shear modulus and viscosity decrease , which results in a reduction of the heating rate ( fig .
[ heat ] ) . the coupled calculations in this work
consider the three - layer model ( the elastic lithosphere is included ) .
the lithosphere works to decrease the heating rate because the displacement is reduced .
an elastic layer can be included by considering the large viscosity of a viscoelastic layer . in our calculations
, we set the viscosity of the lithosphere to @xmath153 pa s. it is important to set a characteristic time for the viscoelastic body ( @xmath154 ) that is sufficiently larger than the tidal period to produce the desired elastic behavior .
the shear modulus is set to 50 gpa .
thus , a viscosity of @xmath153 pa s is sufficient to approximate the elastic body .
the interaction between tidal heating and orbital evolution is very complicated . in this work ,
following the work by barnes et al .
( 2008 ) , we use the classical orbital equations , which are based on the works of goldreich & soter ( 1966 ) .
the rates of change of orbital distance @xmath71 and eccentricity @xmath59 are as follows .
@xmath155 @xmath156 here @xmath157 and @xmath158 are the q values of the planet and the star .
note that @xmath158 is divided by two thirds and includes the value of the love number of the star ( barnes et al .
the first terms represent the effect of the dissipation of the planet .
the second terms show the dissipation of the star .
@xmath159 is the love number of the planet , which is not complex but a real number . in the case of the viscoelastic rheology model
, @xmath159 can be approximated by @xmath160 .
in addition @xmath157 can be represented by re(@xmath108)/im(@xmath108 ) .
the imaginary part of love number is much smaller than the real part .
thus , we can write @xmath161 .
the complex love number depends on the internal structure and thermal state of the planet , which affects the heating rate as shown in eq .
( [ tide ] ) . using the approximation shown above
, we can relate the orbital , thermal , and heating evolution of the planet .
we have to consider that eqs .
( [ a ] ) and ( [ e ] ) are not valid when the orbital period of the planet is shorter than the rotational period of the star . in this case , the signs of the second terms change ( barnes et al . 2008 ) .
however , due to the large mass of the star compared to the martian - sized planet , the values of the second terms are significantly smaller than those of the first terms .
thus , eqs .
( [ a ] ) and ( [ e ] ) are valid in this work .
if the planet s mass is relatively large we need to consider the rotational and orbital periods , which may be important for the case of a super - earth .
@xmath158 represents the magnitude of the star s dissipation , which also affects the orbital evolution of the planet .
we tried the calculation with values between @xmath162 and @xmath163 .
however , even at @xmath164 , the first terms are larger than the second terms by more than six orders of magnitude .
thus , @xmath158 does not affect the orbital evolution of the planet , and the effect of the dissipation of the planet is dominant unless @xmath158 is unreasonably small .
thus , we use the constant value of @xmath165 in every calculation shown in this paper .
the purpose of this work is a habitability analysis through coupled tidal calculations .
thus , as an initial condition for the orbital distance , we set the center of the planet between the inner and outer edges of the habitable zone ( jackson et al . 2008 ) .
the inner and outer edges of the habitable zone ( @xmath166 and @xmath167 , respectively ) were calculated by barnes et al .
( 2008 ) : @xmath168 @xmath169 where @xmath170 au k@xmath29 , @xmath171 au k@xmath172 , @xmath173 au k@xmath29 and @xmath174 au k@xmath172 .
@xmath175 k. @xmath77 can be calculated using eq .
( [ t_st ] ) .
however , if @xmath77 is less than 3700 k , it should be 3700 k when considering @xmath176 ( barnes et al .
2008 ) . @xmath177 and @xmath178 are the inner and outer edges of the solar habitable zone , respectively .
they depend on the condition of cloud .
@xmath177 is @xmath1790.89 au , @xmath1790.72 au and @xmath1790.49 au when the cloud cover is 0% , 50% and 100% , respectively .
@xmath178 is @xmath1791.67 au , @xmath1791.95 au and @xmath1792.4 au when the cloud cover is 0% , 50% and 100% , respectively . in this work ,
we take into account the habitable zone estimated by barnes et al .
( 2008 ) as shown above .
however , habitable zone shown in eqs .
( [ l_in ] ) and ( [ l_out ] ) is based on the conventional estimations by kasting et al .
( 1993 ) and selsis et al .
recently , kopparapu et al . (
2013 ) revised the habitable zone based on new datas .
in addition , habitable zone changes with time due to the evolution of the central star .
we discuss the effect of the estimation and evolution of the habitable zone in the next section .
although many equations are required for the coupled calculation , eqs .
( [ tm ] ) , ( [ tc ] ) , ( [ tide ] ) , ( [ a ] ) and ( [ e ] ) are the main equations needed .
once the initial values of @xmath0 , @xmath1 , @xmath59 and @xmath71 are set , we can calculate properties of the internal structure such as the thickness of each layer and the thermal properties such as the surface temperature of the planet . using the determined conditions ,
the complex love number can be calculated , which is needed to calculate the tidal heating rate ( eq .
( [ tide ] ) ) and orbital evolution ( eqs . ( [ a ] ) and ( [ e ] ) ) . using the calculated love number and tidal heating rate ,
the rates of change of @xmath59 , @xmath71 , @xmath0 and @xmath1 can be determined , following which the internal structure and thermal properties are updated . continuing this process , evolutional calculations based on the coupled calculations become possible . in our calculation , we integrate each value forward with a time step of @xmath162 years for 10 gyr . we tried other time step and confirmed that the original time step is valid for calculations . in some cases ,
mantle temperature exceeds the core temperature because the mantle is heated by tidal heating . in this case
, we calculate the heat flux from the mantle to the core following nimmo & stevenson ( 2000 ) .
we separate the core into spherical shell and solved the conduction equation with 10@xmath180 years of time step .
however , if the stagnant lid thickness becomes very large , the convection in the mantle stops . compared with the convective layer
, the temperature of the conductive layer is very low .
thus , the magnitude of the tidal heating is much smaller than that of a convective planet , and the tidal heating has little effect on the evolution .
our simple thermal and structural models are not valid for a conductive planet . in this work ,
the tidal heating in a conductive structure is out of our scope .
thus , when the convective mantle thickness becomes zero , we stop the calculations .
integrations were performed using the runge - kutta method . as for the initial values of eccentricity @xmath181
, we considered a range of values less than 0.5 .
the initial values of @xmath0 and @xmath1 are not constrained .
thus we tried a wide range of temperatures between 1600 k and 2000 k. in addition , because the tidal heating in the mantle strongly depends on the mantle viscosity , we consider the reference viscosity @xmath137 as a free parameter between @xmath152 pa s and @xmath182 pa s.
the thermal and orbital evolutions of the three - layer model are shown in figs .
[ fig3 ] and [ fig4 ] .
the initial values of @xmath0 and @xmath1 were 1800 k. the initial value of the semi - major axis was calculated considering habitability with 0% cloud cover . surface temperature is determined by eq .
( [ ts ] ) . in the case of @xmath149 ( fig .
[ fig3 ] ) , two types of evolution can be seen , depending on the eccentricity . due to the initial conditions ,
a heating rate of more than @xmath183 w was generated with every eccentricity . because of this large heat , the mantle temperature increases , inducing runaway heating .
when the temperature exceeds the solidus temperature , the mantle begins to melt and the heating rates decrease as shown in fig .
because of the melt , runaway heating stops , and an equilibrium state is reached in which the radiated - heat rate is equal to the tidal heating . if @xmath181 is large enough , large heat is generated as shown in eq .
( [ tide ] ) .
thus , the melt fraction becomes larger , which results in a large value of q. due to the large q value , the eccentricity and semi - major axis do not change significantly even though @xmath76 is small ( the habitable zone is close to the star ) .
although semi major axis decreases by the tidal dissipation , it is far from the inner edge of the habitable zone in every case , which means that the planet stays in the habitable zone for more than 10 gyr .
one caveat is that habitable zone itself changes with time because the luminosity of the central star changes .
thus , our estimation for the stability at the habitable zone of the planet may be too rough . however , kopparapu et al .
( 2013 ) calculate the `` continuous '' habitable zone in which planets are habitable for 5 gyr . in their calculation ,
continuous habitable zone of 0.1@xmath74 mass star is between 0.2 au and 0.5 au .
they also show the continuous habitable zone based on the model by selsis et al .
our calculation results of the semi major axis are in the range of the continuous habitable zone of both models .
thus , we can say that martian - sized planets can stay in the habitable zone for at least 5 gyr .
if the value of q is constant of the order of 10 , the eccentricity is rapidly reduced and the tidal heating does not largely affect the thermal evolution of the planet .
the orbital distance also changes significantly .
however , considering the melt , the q value increases and the large eccentricity and semi - major axis are maintained , which is a difference between constant q value calculations and the coupled calculations . due to the large eccentricity
, relatively large heat generation occurs and the high temperature is maintained . this tidal heating is not too large because the melt reduces the viscosity and shear modulus .
thus , an equilibrium state occurs .
however , at @xmath184 , the heating rate is relatively small , which reduces the tidal heating , making an equilibrium state impossible to maintain .
when the heating rate is reduced , runaway cooling occurs and the melt fraction also decreases .
radiogenic heating is not sufficient to heat up the mantle .
the decreasing melt fraction causes the q value to drop drastically .
however , once the melt fraction becomes zero , the q value increases again because the mantle temperature drops .
the results in fig .
[ fig3 ] are for the case in which the initial value of @xmath1 and @xmath0 were both 1800 k. we also tried with @xmath185 i.e. , with an initial temperature difference of 1200 k. in that case the core temperature rapidly reaches the mantle temperature .
thus , the effect of a large @xmath1 is small in terms of the evolution of the planet .
barnes et al .
( 2009 ) analyzed the habitability focusing on the surface heat flux .
conventionally , the habitable zone is defined as the area where h@xmath186o is kept in its liquid phase from models which assume that the ground is water saturated ( e.g. , kasting et al .
. however , as well as liquid water , the surface state also plays an important role in determining habitability .
barnes et al .
( 2009 ) gave a condition for habitability based on the heat fluxes and surface states of mars and the jovian satellite io .
in their estimation , a heat flux of between 0.04 w m@xmath172 and 2 w m@xmath172 is required for habitability .
if the magnitude of the heat flux is similar to io , active volcanoes make the surface inhabitable . in our calculations ,
the hot equilibrium state of the planet radiates more than 0.5 w m@xmath172 when @xmath149 and @xmath181 is 0.3 , which is the suitable value for the habitability . at @xmath181=0.5 , although it does not exceed 2 w m@xmath172 , the surface heat flux reaches the maximum limit .
thus , the planet is predicted to have active volcanos like io , which may make the planet inhabitable .
if @xmath181 is less than 0.1 , tidal heating is not enough to maintain the melt in the mantle , which results in cooling of the planet . from the coupled calculations
, we can say that a martian - sized planet that orbits in the habitable zone of an @xmath149 star is suitable for the habitability if the initial eccentricity is between 0.2 and 0.5
. however , when the eccentricity is around 0.1 , the heat flux exceeds 0.5 w m@xmath172 at @xmath181=0.1 .
thus , when we observe the martian sized planet at the habitable zone of 0.1m@xmath187 star , it is important whether the planet has the eccentricity more than 0.1 for the habitability . if observed eccentricity is over 0.1 , the planet is predicted to have partial melt and the relatively large heat flux .
figure [ fig4 ] shows the case for @xmath188 .
compared to the star with @xmath149 , the melt fraction is small . at @xmath184 ,
melt is not generated . because the planet is assumed to orbit according to kepler s law ,
the mean motion in the habitable zone is small . as shown in eq .
( [ tide ] ) , the heating rate increases with @xmath189 .
thus , tides can not produce enough heat to generate large amount of melt in the case of @xmath188 star .
surface heat flux at the hot state is around 0.3 w m@xmath172 and 0.4 w m@xmath172 when @xmath181 is 0.3 and 0.5 , respectively , which is suitable range for the habitability .
the eccentricity does not change significantly for @xmath188 even though the q value is relatively small because the orbital distance is large , which is different from the case of a @xmath190 mass star . in the case of @xmath149 ,
the orbital distance is small .
however , the small rate of change of eccentricity is caused by the large @xmath157 due to the melt . through the coupled calculations
, it has been determined that the thermal state of small extrasolar planets in habitable zones depends mainly on their initial eccentricity and the mass of their star .
the initial value of the core temperature does not affect the heating state because tidal heating soon heats up the mantle anyway .
figure [ phase ] shows the thermal state as a function of the mass ratio of the star ( @xmath84 ) and the initial eccentricity @xmath181 for different initial temperatures ( @xmath191 ) and reference viscosities ( @xmath137 ) .
as can be seen in fig .
( [ phase ] ) , the reference viscosity and initial temperature do not affect the heating states of the planet .
if the mass of the star is around 0.1@xmath74 , the planet reaches a hot state when @xmath192 , independently of @xmath191 and @xmath137 . as mentioned above
, the planet can stay in the habitable zone for a long time in the hot state .
if the mass of the star is around 0.2m@xmath187 , a hot state occurs if @xmath193 .
if @xmath84 is more than 0.3 , a hot state is not formed even though the eccentricity is quite large .
the habitable zone of a star in this mass range is too far from the star for tidal heating large enough for runaway heating and melting of the mantle to occur .
since the cold planet has low temperature and rigid , stars of less than @xmath194 in mass are good targets for finding habitable planets .
if the tidal heating is sufficiently large , the mantle is heated and the surface heat flux exceeds the martian heat flux . in this work , we assume stagnant lid convection as a heat transport mechanism . as an alternative mechanism
, plate tectonics may occur in planets in a hot state .
although it is currently only possible to observe the plate tectonics of earth , it has been suggested that plate tectonics occurs in large , rocky extrasolar planets ( e.g. , valencia et al .
however , plate tectonics is a very complicated system and the main induction mechanism is not well understood .
the size of the planet may affect the stress ( valencia et al .
2007 ; oneill & lenardic 2007 ) ; however , korenaga ( 2010 ) suggests that the effect of surface water is more important than the size of the planet . in our model
, we do not include surface oceans and the size of the planets considered is relatively small .
thus , stagnant - lid convection is more reasonable .
however , in the case of mars , plate tectonics has been suggested to be the origin of the early internal dynamo ( nimmo & stevenson 2000 ) .
thus , plate tectonics can not be ruled out for martian - sized extrasolar planets .
delamination may be more probable in small - sized extrasolar planets than plate tectonics . in a hot state ,
the melt in the mantle reduces the mantle viscosity .
thus , the bottom of the stagnant lid is delaminated and absorbed into the mantle .
delamination displaces the surface of the planet , which is the suggested cause of the corona of venus ( smrekar & stofan 1997 ) .
thus , we can suggest that the planet is heated by tides , and the morphology of the planet is mainly a product of localized delamination processes caused by partial melting . as an alternative heat transfer model , heat pipe mode is suggested in io and the earth ( moore 2001 ; moore & webb 2013 ) , which can be the important heat transfer mechanism for the tidally heated extrasolar planets .
thus , the effects of delamination and heat pipe should be considered in future work .
it was revealed that mars had an intrinsic magnetic field in the past ( acua et al .
the most probable origin of the intrinsic magnetic field is the dynamo in the core .
the mechanism of the martian dynamo is not well understood .
however , it is required that more heat than that transported by adiabatic heat flux is transported from the core to the mantle in order to induce the core convection ( nimmo & stevenson 2000 ) .
if tidal heat is not generated in the mantle , heat is transported sufficiently by the mantle and a dynamo may be induced .
however , in our calculations , tidal heating heats up the mantle . in some cases
, heat is even transported from the mantle to the core .
thus , in the hot state of a planet , it is unlikely the dynamo is activated .
an intrinsic magnetic field is not predicted in a tidally heated planet ; however , if the orbital plane of the planet is oblique to the magnetic axis of the star , an induced magnetic field may be possible because of the melt in the mantle .
in the case of io , an induced magnetic field is observed ( khurana et al .
the induced magnetic field of io is consistent with a mantle of at least 20 % partial melt ( khurana et al . 2011 ) .
this melt fraction is comparable to our calculation results . for simplicity , we assume that the internal structure of the planet is spherical shells and depends only on the radial distance . as for the habitable zone
, the spherical - shell structure should provide a valid approximation because the temperature in light side is not so high .
however , the heterogeneity of the planet is an interesting topic for future study , especially for planets that orbit significantly close to their stars .
tidal heating in a planet with a heterogeneous structure was calculated for the case of enceladus , though it is extremely complicated ( tobie et al .
2008 ; bhounkov et al .
if coupled calculations including a heterogeneous structure are possible in future work , an important comparison to the spherical - shell model of this work can be made .
one caveat is that the coupled calculations in this work do not consider the resonance of the planet .
if the planet is in a resonant state with other planets , the orbital evolution becomes significantly more complicated .
resonance among planets increases the planets eccentricity , as exhibited by the jovian satellites io , europa and ganymede ( murray & dermott 1999 ) .
some satellites around saturn also interact with each other ( murray & dermott 1999 ) .
hence , if there are multiple planets around a star , we must consider the resonance between them .
through thermal and orbital coupled calculations , it was found that two types of evolution are probable for martian - sized planets in the habitable zones of m stars .
one type is the runaway cooling state in which the radiated heat flux is larger than the tidal heating , decreasing the mantle temperature through positive feedback . as a result of the cooling , convection in the mantle
is reduced and the planet becomes rigid .
the other type of evolution is the hot state . in a hot state
, tidal heating exceeds the radiated heat flux and runaway heating occurs .
however , melt in the mantle reduces the tidal heating by decreasing the viscosity and shear modulus , and an equilibrium state is reached .
the eccentricity and orbital distance of hot state planets do not change significantly due to the large q value caused by the melt .
thus , the planet in hot state can stay the habitable zone for more than a few billion years . in the equilibrium state ,
the heat flux is in the range suitable for habitability ( barnes et al .
we performed the calculations and by changing many parameters , found that those that have the most significant effect on the evolution are the mass of the star and the initial value of the eccentricity .
if the mass of the star is less than 0.2@xmath74 and the initial eccentricity is more than 0.2 , the planet should be in a hot state ( fig .
[ phase ] ) .
a hot state planet contains melt in the mantle ; thus , a magnetic field may be induced .
an intrinsic magnetic field , however , is unlikely to occur because the mantle temperature is large and sufficient heat is not transported from the core . among extrasolar planets , super - earths in particular
are considered for habitability . in the case of a super - earth ,
our simple structural model is not valid and the effects of pressure are likely large ( wagner et al . 2012 )
. we must also consider the phase of the core because the core may be solidified by the large pressure .
however , our simple coupled calculations provide the first step towards these more complicated and accurate analyses of extrasolar planets .
and @xmath59 are @xmath190 and 0.2 , respectively . assuming the initial state at the circle point
, two types of heating evolution ( runaway heating or cooling ) can be expected .
if the temperature exceeds the solidus temperature , the melt reduces the mantle shear modulus and viscosity , which results in decreased tidal heating and ends runaway heating.,width=491 ] is 0.1@xmath74 . @xmath195
pa s. initial values of the mantle temperature and core are 1800 k. initial value of the eccentricity is 0.1 ( solid line ) , 0.3 ( dashed line ) and 0.5 ( dotted line).,width=491 ] is 0.2@xmath74 . @xmath195
pa s. initial values of the mantle temperature and core are 1800 k. initial value of the eccentricity is 0.1 ( solid line ) , 0.3 ( dashed line ) and 0.5 ( dotted line).,width=491 ] ) and initial eccentricity of the planet ( @xmath181 ) .
triangle points show the hot state in which runaway heating and the melt induced equilibrium state occur .
circle points show the cold state , in which runaway cooling occurs .
top panels ( a ) and ( b ) show the results for @xmath191=1800 k. bottom panels ( c ) and ( d ) show the results for @xmath196 k. @xmath197 is the same as @xmath191 . in the left panels ( a ) and ( c ) , @xmath195 pa s , and in the right panels ( b ) and ( d ) @xmath198 pa s. , width=491 ] | m type stars are good targets in the search for habitable extrasolar planets . because of their low effective temperatures ,
the habitable zone of m stars is very close to the star itself . for planets close to their stars
, tidal heating plays an important role in thermal and orbital evolutions , especially when the planet orbit has a relatively large eccentricity .
although tidal heating interacts with the thermal state and orbit of the planet , such coupled calculations for extrasolar planets around m star have not been conducted .
we perform coupled calculations using simple structural and orbital models , and analyze the thermal state and habitability of a terrestrial planet . considering this planet to be martian sized
, the tide heats up and partially melts the mantle , maintaining an equilibrium state if the mass of the star is less than 0.2 times the mass of the sun and the initial eccentricity of the orbit is more than 0.2 .
the reduction of heat dissipation due to the melted mantle allows the planet to stay in the habitable zone for more than 10 gyr even though the orbital distance is small .
the surface heat flux at the equilibrium state is between that of mars and io .
the thermal state of the planet mainly depends on the initial value of the eccentricity and the mass of the star . |
quite a number of physical , chemical , and biological systems manifest some type of modulation in their spatial ordering @xcite .
such structures are stripes and bubbles in two - dimensional ( 2d ) systems , or lamellae , hexagonally packed cylinders , and cubic arrays of spheres in three - dimensional ( 3d ) cases as well as more complex structures such as gyroids .
examples of such systems include ferromagnetic layers @xcite , magnetic garnet films @xcite , ferrofluids @xcite , dipolar langmuir films @xcite , rippled phases in lipid bilayers @xcite , and block copolymers @xcite .
modulated phases may also occur in systems described by two ( or more ) coupled order parameters , each favoring a different equilibrium state @xcite .
the observed spatial patterns exhibit striking similarity even for systems that are very different in their nature .
it is generally understood that the modulated structures are formed spontaneously due to the competition between short- and long - range interactions . in the case of 2d ferromagnetic layers , for example
, the short - range interaction arises from magnetic domain wall energy , while the long - range interaction is due to magnetic dipole - dipole interaction which induces a demagnetizing field @xcite . adding both contributions and minimizing the total free energy with respect to the wavenumber @xmath0 , one obtains the most stable mode @xmath1 .
this description is valid in the weak segregation limit ( close to a critical point ) , where the equilibrium domain size is given by @xmath2 . in general , this quantity depends on temperature and/or other external fields . in this paper , we shall consider two modulated monolayers that are jointly coupled .
our motivation is related to recent experiments by collins and keller @xcite who investigated montal - mueller planar bilayer membranes @xcite composed of lipids and cholesterol . in this technique ,
a bilayer is constructed by preparing separately two independent monolayers and then combining them into one joint bilayer across a hole at the air / water interface .
the experiments addressed specifically the question of liquid domains in the two leaflets , and the mutual influence of the monolayers in terms of their domain phases . in the experiment , asymmetric bilayers are prepared in such a way that one leaflet s composition would phase - separate in a symmetric bilayer and the other s would not
. in some cases , one leaflet may induce phase separation in the other leaflet , whereas in other cases , the second leaflet suppresses domain formation in the original leaflet .
these results imply that the two leaflet coupling is important ingredient in determining the bilayer phase state .
motivated by these experiments , the coupled bilayer system was investigated theoretically . the coupling mechanism arises through interactions between lipid tails across the bilayer midplane , and the phase behavior of such a bilayer membrane was computed using either regular solution theory @xcite or landau theory @xcite .
the theoretical results are in accord with several of the experimental observations .
it should be noted that all previous models dealt with the coupling between two _ macro phase _ separated leaflets , while it is also of interest to investigate the coupling between two _ micro phase _ separated ( modulated ) leaflets .
furthermore , one might also consider the interplay between a macro and a micro phase separation .
in the present work , we suggest a model describing the coupling between two modulated systems , and , in particular , we analyze the influence of this coupling on the phase behavior of two coupled 2d monolayers .
when the two monolayers have the same preferred periodicity of modulation , we obtain the mean - field phase diagrams which exhibit various combinations of micro phase separated structures . in some cases
, the periodic structure in one of the monolayers will induce a modulation in the other monolayer .
interesting situations take place when the two monolayers have different preferred wavelengths of modulation . here
the frustrations between the two competing modulated structures need to be optimized . these structures and their dynamical behavior
are examined using numerical simulations .
although there has been so far no experiment which directly corresponds to the proposed model , our predictions may be verified , for example , by constructing montal - mueller bilayers @xcite out of two lipid monolayers that exhibit a striped phase near the miscibility critical point @xcite . in the next section
, we present a phenomenological model describing the coupling between two modulated lipid monolayers . in sec .
[ identical ] , we discuss the case when the two monolayers have the same preferred wavelength of modulation .
monolayers having different preferred wavelengths will be considered in sec .
[ different ] , and some related situations are further discussed in sec . [ discussion ] .
although we limit our present analysis to 2d systems , the suggested model can be generalized to 3d systems as well .
in order to illustrate the coupling effect between two modulated systems , we imagine a pair of lipid monolayers forming a coupled bilayer . each of the monolayer can undergo separately a micro phase separation . as shown in fig .
[ fig1 ] , we assume that each monolayer is a mixture of two lipid species , say lipid a and lipid b. their area fractions are defined by @xmath3 and @xmath4 , where @xmath5 is the 2d positional vector . by assuming that the monolayer is incompressible , @xmath6 , the monolayer composition can be characterized by a single order parameter defined by the relative a / b composition @xmath7 .
let us denote this local order parameter of the upper and lower monolayers by @xmath8 and @xmath9 , respectively .
the coarse - grained free - energy functional for the coupled modulated bilayer is written as : @xmath10 & = f_{\rm u}[\phi]+f_{\rm \ell}[\psi]- \lambda \int { \rm d}\mathbf{r}\ , \phi\psi\nonumber\\ & = \int { \rm d}\mathbf{r } \biggl[2 ( \nabla^2 \phi)^2 - 2 ( \nabla \phi)^2 + \frac{\tau}{2}\phi^2 + \frac{1}{4}\phi^4 - \mu_{\phi}\phi \nonumber \\ & + 2 d(\nabla^2 \psi)^2 - 2 c(\nabla \psi)^2 + \frac{\tau}{2}\psi^2 + \frac{1}{4}\psi^4 - \mu_{\psi}\psi -\lambda \phi \psi \biggr ] .
\label{freeenergy}\end{aligned}\ ] ] this is a modified ginzburg - landau free energy expanded in powers of the order parameters @xmath11 and @xmath12 and their derivatives . the @xmath13 $ ] free energy has five terms depending only on @xmath11 and its derivatives .
it describes the upper monolayer and its possible modulations , while the coefficients of the laplacian squared , the gradient squared and the @xmath14 terms are taken to be numbers , for simplicity .
similarly , @xmath15 $ ] describing the lower monolayer contains the next five terms that are only functions of @xmath12 and its derivatives .
the last term represents the coupling between the two leaflets as will be explained later .
the coefficients of the two gradient squared terms are both negative ( @xmath16 ) favoring spatial modulations , whereas the coefficients of the laplacian squared terms are positive ( @xmath17 ) to have a stable modulation at finite wavenumbers .
the @xmath18 , @xmath14 , @xmath19 and @xmath20 terms in @xmath21 are the usual landau expansion terms with @xmath22 being the reduced temperature ( @xmath23 is the critical temperature ) . for simplicity ,
the two leaflets are taken to have the same critical temperature @xmath23 ( and hence the same @xmath24 ) .
finally , the linear term coefficients , @xmath25 and @xmath26 , are the chemical potentials which regulate the average values of @xmath11 and @xmath12 , respectively . in the absence of the coupling term ( @xmath27 )
, each of the two leaflets can have its own modulated phase .
free energy functionals such as @xmath28 have been used successfully in the past to describe a variety of modulated systems : magnetic garnet films @xcite , langmuir films @xcite , diblock copolymers @xcite , and amphiphilic systems @xcite .
furthermore , interfacial properties between different coexisting phases have been investigated using a similar model @xcite . in the above expression for the free energy @xmath21
, the @xmath11-leaflet has a dominant wavenumber @xmath29 , and so has the @xmath12-leaflet with @xmath30 .
the modulation wavenumbers and amplitudes of the two monolayers coincide when @xmath31 and the average compositions are the same .
next we address the physical origin of the coupling term @xmath32 .
we first note that this quadratic term is invariant under the exchange of @xmath33 . when @xmath34 , this term can be obtained from a @xmath35 term @xcite , which represents a local energy penalty when the upper and lower monolayers have different compositions . in the case of mixed lipid bilayers ,
such a coupling may result from the conformational confinement of the lipid chains , and hence would have entropic origin @xcite . by estimating the degree of the lipid chain interdigitation
, the magnitude of the coupling parameter @xmath36 was recently estimated by may @xcite . in general , the coupling constant @xmath36 can also be negative depending on the specific coupling mechanism @xcite .
however , it will be explained later that the phase diagram for @xmath37 can easily be obtained from the @xmath34 one .
hence , it is sufficient to consider only the @xmath34 case without loss of generality .
although the microscopic origin of the coupling may differ between systems , we will regard @xmath36 as a phenomenological parameter and investigate its role on the structure , phase behavior and dynamics of coupled modulated bilayers .
the phase behavior for uncoupled case , @xmath27 , can be obtained from the analysis of @xmath13 $ ] @xcite and is only briefly reviewed here ( see also fig .
[ fig2 ] ) . for a 2d system
, the mean - field phase diagram can be constructed by comparing the free energies of striped ( s ) and hexagonal ( h ) phases . in terms of the @xmath11 order parameter
, the stripe phase is described by @xmath38 where @xmath39 is the spatially averaged composition ( imposed by the chemical potential @xmath25 ) , and @xmath40 is the amplitude of the @xmath41-mode in the @xmath42-direction .
similarly , the composition of the hexagonal phase is given by a superposition of three 2d modes of equal magnitude , @xmath43 @xmath44 where @xmath45 and @xmath46 . in the above
, only the most unstable wavenumber @xmath41 is used within the single - mode approximation .
this can be justified for the weak segregation region close to the critical point @xcite .
averaging over one spatial period , we obtain the free energy densities of the striped , hexagonal , and disordered phases , respectively @xmath47 @xmath48 @xmath49 in fig .
[ fig2](a ) , we reproduce the original phase diagram of ref . @xcite . the striped , hexagonal , and disordered phases are separated by first - order phase - transition lines .
regions of two - phase coexistence do exist but are omitted from the figure for clarity sake @xcite .
thus , the transition lines indicate the locus of points at which the free energies of two different phases cross each other , and are not the proper phase boundaries ( binodals ) . the critical point ( filled circle )
is located at @xmath50 .
having introduced the free energy and explained the phase behavior of the uncoupled case , we shall now explore the equilibrium and non - equilibrium properties of two coupled modulated monolayers , @xmath52 .
first we consider the case when @xmath31 so that the preferred wavenumbers are the same for both monolayers , @xmath53 .
the mean - field phase diagram is calculated within the single - mode approximation .
various combinations of 2d modulated structures appearing in the two monolayers are possible .
the first example is the striped - striped ( ss ) phase , in which both monolayers exhibit the striped phase .
this can be expressed as @xmath38 @xmath54 where @xmath39 and @xmath55 are the average compositions , @xmath40 and @xmath56 are the respective amplitudes .
these composition profiles are substituted into the free energy of eq .
( [ freeenergy ] ) .
averaging over one spatial period , we obtain the free energy density of the ss phase : @xmath57 where @xmath58 is defined in eq .
( [ energys ] ) .
we then minimize @xmath59 with respect to both @xmath40 and @xmath56 for given @xmath60 , @xmath61 , @xmath24 and @xmath36 .
when either @xmath40 or @xmath56 vanishes , the corresponding monolayer is in its disordered phase and the mixed bilayer state will be called the striped - disordered ( sd ) or the disordered - striped ( ds ) phase .
note that we use the convention that the first index is of the @xmath11leaflet and the second of the @xmath12one . when both @xmath40 and @xmath56 are zero , the free energy density of the disordered - disordered ( dd ) phase is given by @xmath62 where @xmath63 is defined in eq .
( [ energyd ] ) .
this free energy @xmath64 was analyzed in ref .
@xcite in order to investigate the macro phase separation of a bilayer membrane with coupled monolayers .
it was shown that the bilayer can exist in four different phases , and can also exhibit a three - phase coexistence .
similar to the stripe case , the order parameters of the hexagonal - hexagonal ( hh ) phase can be represented as @xmath65 @xmath66 where the basis of the three @xmath67 was defined in eq .
( [ three_q ] ) . by repeating the same procedure as for the ss phase ,
the free energy density of the hh phase is obtained as @xmath68 where @xmath69 is defined in eq .
( [ energyh ] ) .
when either @xmath40 or @xmath56 vanishes , one of the monolayers is in the disordered phase and the bilayer will be called the hexagonal - disordered ( hd ) phase or the disordered - hexagonal ( dh ) phase . when the normal hexagonal phase in one leaflet is coupled to the inverted hexagonal phase in the other leaflet
, it is energetically favorable to have a particular phase shift of @xmath70 between the two hexagonal structures .
the order parameters which represent such a different type of hexagonal - hexagonal ( hh@xmath71 ) phase can be written as @xmath72 @xmath73 the free energy density of the hh@xmath71 phase is then obtained as @xmath74 another combination which should be considered in the present model is the asymmetric case where one monolayer exhibits the striped phase and the other the hexagonal phase . this striped - hexagonal ( sh ) phase
is expressed as @xmath75 @xmath76 the free energy density of this sh phase is calculated to be @xmath77 the phase in which @xmath78 and @xmath79 in eqs .
( [ shphi ] ) and ( [ shpsi ] ) are interchanged with @xmath80 and @xmath81 is called the hexagonal - striped ( hs ) phase , and its free energy is obtained from the sh phase by noting the @xmath33 symmetry .
in addition to these phases , we have also taken into account the square - square ( qq ) phase expressed by @xmath82 , \label{qqphi}\ ] ] @xmath83 .
\label{qqpsi}\ ] ] then its free energy density is given by @xmath84 where @xmath85 however , we will show below that this qq phase can not be more stable than the other phases . minimizing eqs .
( [ energyss ] ) , ( [ energyhh ] ) , ( [ energyhh ] ) , ( [ energysh ] ) and ( [ energyqq ] ) with respect to both @xmath40 and @xmath56 , we obtain the phase diagram for the coupled bilayer . as a reference
, we first show in fig .
[ fig2](b ) the phase diagram in the decoupled case ( @xmath27 ) for @xmath86 .
this can easily be obtained from fig .
[ fig2](a ) by combining its two cross - sections ( one for @xmath60 and one for @xmath61 ) at @xmath86 .
figure [ fig3 ] gives the phase diagram for a coupled bilayer when ( a ) @xmath87 and ( b ) @xmath88 , while the temperature is fixed to @xmath86 as before . on the @xmath89-plane
, we have identified the phase which has the lowest energy , whereas possible phase coexistence regions between different phases have been ignored .
all the boundary lines indicate first - order transitions . since the free energy eq .
( [ freeenergy ] ) is invariant under the exchange of @xmath33 , the phase diagrams are symmetric about the diagonal line @xmath90 as the upper and lower leaflets have been chosen arbitrarily .
these phase diagrams are also symmetric under the rotation of 180 degrees around the origin because eq .
( [ freeenergy ] ) is invariant ( except the linear terms ) under the simultaneous transformations of @xmath91 and @xmath92 .
this is reasonable as the labels of `` a '' or `` b '' for the two lipids have been assigned arbitrarily . as a consequence ,
the phase diagrams are also symmetric about the diagonal line @xmath93 .
the symmetries with respect to both @xmath94 and @xmath95 in fig .
[ fig2](b ) for @xmath27 are now broken because of the coupling between the two leaflets .
when the coupling parameter is small ( @xmath87 ) , the global topology of the phase diagram resembles that of the uncoupled case presented in fig .
[ fig2](b ) .
close to the origin , @xmath96 , there is a region of ss phase surrounded by eight other phases : two sh , two hs , two hh , and two hh@xmath71 phases .
the hh phase appearing in the region of @xmath97 and @xmath98 is the combination of the two inverted hexagonal structures on each monolayer .
one sees that the hh@xmath71 phase appears in the regions of @xmath99 , where the hexagonal and the inverted hexagonal structures are coupled to each other .
a remarkable feature of this phase diagram is the existence of the ss and hh phases in the regions where either @xmath100 or @xmath101 are large .
these outer ss and hh phases extend up to the maximum or the minimum values of the compositions .
these regions of the ss and hh phases with @xmath102 roughly correspond to those of the sd ( ds ) and hd ( dh ) phases , respectively , in fig .
[ fig2](b ) with @xmath103 .
hence the modulated structure in one of the monolayers induces the same modulated phase in the other monolayer due to the coupling term .
notice that the sd ( ds ) phase and hd ( dh ) phase do not exist in fig .
[ fig3](a ) .
we further remark that the extent of the four dd phase regions is almost unaffected by the coupling .
even when the temperature is lowered by decreasing @xmath24 , only the phases located close to the origin ( @xmath96 ) would expand , and the global topology does not change substantially . for a larger value of the coupling parameter ( @xmath88 ) , the five regions of the ss phase merge together forming one single continuous ss region .
the four hh regions and still distinct and separate the ss region from four dd phase regions .
note that in fig .
[ fig3](b ) , all phases have a symmetric combination of phase modulation such as ss or hh .
the asymmetric combination such as the sh phase does not appear , because the large coupling parameter strongly prefers symmetric phases of equal modulations in the two monolayers , although the @xmath40 and @xmath56 amplitudes of the two modulated monolayers are not the same in the stripe ss phase ( or the hexagonal hh phase ) .
as the value of @xmath36 is increased from @xmath104 to @xmath105 , first the sh phase disappears , followed by the disappearance of the hh@xmath71 phase . when the value of @xmath36 is further increased , the regions of the ss and hh phases expand on the expense of the dd phase regions .
this means that the coupling between the monolayers causes more structural order in the bilayer .
finally we remark that the qq phase was never found to be more stable than any of the other phases considered above .
although we have so far assumed that @xmath36 is positive , the phase diagrams for @xmath37 can be easily obtained from those for @xmath34 by rotating them by 90 degrees around the origin .
this is because the free energy eq .
( [ freeenergy ] ) is invariant under the simultaneous transformations of either @xmath91 and @xmath106 , or @xmath107 and @xmath106 . in order to check the validity of the obtained phase diagram and to investigate the dynamics of coupled modulated bilayers
, we consider now the time evolution of the coupled equations for @xmath11 and @xmath12 : @xmath108 here we have assumed that both @xmath11 and @xmath12 are conserved order parameters in each of the monolayer ( model b in the hohenberg - halperin classification @xcite ) . for simplicity , the kinetic coefficients @xmath109 and @xmath110 are taken to be unity , and both the hydrodynamic effect and thermal fluctuations are neglected .
we solve the above equations numerically in 2d using the periodic boundary condition .
each simulation starts from a disordered state with a small random noise around the average compositions @xmath60 and @xmath61 . in fig .
[ fig4 ] , we show typical equilibrium patterns of @xmath11 , @xmath12 , @xmath111 and @xmath112 for two choices of parameters .
the @xmath111 pattern is presented here because this quantity can be directly observed in the experiment on montal - mueller bilayers using fluorescence microscopy @xcite . the quantity
@xmath112 measures the concentration contrast between the @xmath11 and @xmath12 leaflets .
time is measured in discrete time steps , and @xmath113 corresponds to a well equilibrated system . in all the simulations below , the temperature is fixed to be @xmath86 corresponding to the weak segregation regime .
notice that all the patterns in fig .
[ fig4 ] are presented with the same gray scale .
figure [ fig4](a ) illustrates the coupling between a hexagonal phase with @xmath114 and an inverted hexagonal phase with @xmath115 in the weak coupling regime ( @xmath87 ) .
being consistent with the phase diagram of fig .
[ fig3](a ) , this parameter choice yields the hh@xmath71 phase as seen from the pattern of @xmath111 where the two hexagonal structures are superimposed .
we note that the difference in the order parameter @xmath112 also exhibits a hexagonal structure .
figure [ fig4](b ) shows the equilibrium patterns for @xmath116 and @xmath95 in the strong coupling regime ( @xmath88 ) .
if there were no coupling , the @xmath11-monolayer would not exhibit any modulation ( as it is in its own disordered phase ) , whereas the @xmath12-monolayer is in the striped phase .
we clearly see that , due to the coupling effect , the stripe structure is induced in the pattern of @xmath11 .
this corresponds to the ss phase shown in fig .
[ fig3](b ) .
the periodicities of the two striped structures are the same , although their amplitudes differ .
notice that the modulation phase of @xmath111 is shifted by @xmath117 relatively to that of @xmath112 . since the patterns in fig .
[ fig4 ] would correspond to the equilibrium configurations , they can be compared with the phase diagrams in fig .
we conclude that these simulation results indeed reproduce the predicted equilibrium modulated structures .
we consider next the more general case in which the preferred wavelengths of modulation in the two uncoupled leaflets are different , @xmath118 .
the free energy densities can not be obtained analytically as was done in sec
. [ density ] , because there is not a single periodicity on which one can average @xmath8 and @xmath9 .
due to such a difficulty in the analytical treatment , we present below the results of numerical simulations , relying on eq .
( [ evolution ] ) for the time evolution of the two coupled order parameters . in fig .
[ fig5 ] , we show the patterns for @xmath86 and @xmath96 , when both monolayers exhibit the striped phase without the coupling . as a reference
, we show in fig .
[ fig5](a ) the case when @xmath31 and @xmath87 corresponding to the ss phase in fig .
[ fig3](a ) .
the patterns of @xmath11 and @xmath12 match each other as the composition difference @xmath112 vanishes throughout the system . in fig .
[ fig5](b ) , the parameters are chosen to be @xmath119 , @xmath120 and @xmath87 .
the preferred wavenumbers of the two monolayers are different : @xmath121 for uncoupled leaflets .
the above set of parameters , especially @xmath122 and @xmath123 , are chosen in such a way that the amplitudes of the two stripes are nearly equal . as long as the coupling parameter is small ,
the two stripes of different periodicities are formed rather independently .
the superposition of the two striped structures produces an interference pattern resulting in a new modulation as seen from the pattern of @xmath111 .
the striped modulations of @xmath11 and @xmath12 are almost parallel or perpendicular to each other .
when @xmath36 is made larger , up to @xmath88 as in fig .
[ fig5](c ) , the @xmath12 field exhibits a complex pattern in which two different length scales coexist ( reflecting @xmath124 and @xmath125 ) , whereas the pattern of @xmath11 is characterized by a single mode ( reflecting @xmath124 ) .
the patters of @xmath11 and @xmath12 almost match each other when @xmath126 as seen in fig .
[ fig5](d ) . in this case
, the modulation with a longer wavelength ( @xmath124 ) dominates both monolayers .
figure [ fig5](b ) , ( c ) , ( d ) provide a typical sequence of morphological changes , i.e. , interference pattern @xmath127 two - mode pattern @xmath127 single - mode pattern , as the coupling constant @xmath36 is increased . to further analyze the temporal correlations of the two order parameters , @xmath11 and @xmath12 , we have plotted in fig .
[ fig6 ] the time evolution of the quantity @xmath128 ^ 2 , \label{delta}\ ] ] where @xmath129 is the linear system size in the simulations .
the solid , dashed , dotted , and dashed - dotted lines correspond to the time evolutions of @xmath130 in fig .
[ fig5](a ) , ( b ) , ( c ) , and ( d ) respectively . the solid line ( a ) first increases and then approaches to zero since the patterns of the two stripes coincide in the late stage .
the dashed line ( b ) increases in two separate stages . at first because the growth rate of the modulation having smaller wavelength ( corresponding to @xmath12 ) is faster than that having larger periodicity ( corresponding to @xmath11 ) as is also revealed from the linear stability analysis of eq .
( [ evolution ] ) which will be published elsewhere . at late temporal stages ,
the value of @xmath131 remains large for small coupling parameter ( @xmath87 ) .
when the coupling becomes even stronger ( @xmath88 ) as for the dotted line ( c ) , the value of @xmath131 is suppressed compared to the dashed line ( b ) , because @xmath11 and @xmath12 tend to have more overlap for larger @xmath36 .
the same applies for the dashed - dotted line ( d ) with @xmath126 as compared to the dotted line ( c ) .
we propose a minimal model describing the coupling phenomena between two modulated bilayers . considering 2d case , we obtain the mean - field phase diagram when the two coupled and spatially modulated monolayers have the same preferred periodicity .
various combinations of modulated phases can exist such as the ss , hh , hh@xmath71 and sh ( hs ) phases as described in sec .
[ identical ] .
we have seen that modulations in one of the monolayers induces a similar modulations in the other .
the region of the induced modulated phase expands as the coupling parameter becomes larger . when the two monolayers have different inherent wavelengths in the decoupled case , we have conducted numerical simulations to investigate the morphologies and dynamics of the coupled system .
we obtain several complex patterns arising from the frustration induced by the two incommensurate structures .
as the coupling constant @xmath36 is made larger , the two different modes start to interfere with each other and eventually coincide . the time evolution of the striped structures can take place in two steps reflecting the different growth rate of the two modulations .
it is instructive to rewrite the free energy eq .
( [ freeenergy ] ) in terms of the sum and the difference of @xmath11 and @xmath12 with @xmath132 . when @xmath31 , we obtain @xmath133 & = \int { \rm d}\mathbf{r } \biggl[(\nabla^2 \eta_+)^2 - ( \nabla \eta_+)^2 + \frac{1}{4}(\tau-\lambda)\eta_+^2 + \frac{1}{32}\eta_+^4 -\mu_+ \eta_+ \nonumber \\ & + ( \nabla^2 \eta_-)^2 - ( \nabla \eta_-)^2 + \frac{1}{4}(\tau+\lambda)\eta_-^2 + \frac{1}{32}\eta_-^4 -\mu_- \eta_- + \frac{3}{16}\eta_+^2 \eta_-^2 \biggr ] , \label{freeenergy2}\end{aligned}\ ] ] where @xmath134 . hence the coupling term between @xmath135 and @xmath136 in the free energy
takes the form of an @xmath137 term with a numerical positive coefficient .
the original coupling parameter @xmath36 enters in the coefficients of @xmath138 and @xmath139 terms ( but not in the @xmath137 coupling term ) .
it shifts the respective transition temperatures of @xmath135 and @xmath136 in opposite directions .
when the gradient terms are absent and @xmath140 , a similar model was considered by mackintosh and safran who studied transitions between lamellar and vesicle phases in two - component fluid bilayers @xcite . the proposed free energy eq .
( [ freeenergy ] ) has some analogies to the previous model for the rippled phase in lipid bilayers @xcite .
it was argued that the coupling between the membrane curvature and the asymmetry in the area per molecule between the two monolayers would induce a structural modulation of a bilayer . by considering a similar mechanism , kumar _
et al . _
@xcite investigated various modulated phases in two - component bilayer membranes .
they claimed that the phase behavior of two - component bilayers resembles that of three - component monolayer .
this is because the three different local combinations of upper / lower composition in bilayers ( a / b , b / a , and a / a for excess of a ) , would correspond to three different types of molecules for the monolayer .
one of the new aspects in our model is that the preferred wavelengths of the two monolayers can , in general , be different from one another leading to a frustrated bilayer state .
we also point out that there are some similarities between coupled modulated structures and the problem of atoms adsorbed on a periodic solid substrate .
the latter topic has been extensively studied within the frenkel - kontorova ( fk ) model which provides a simple description of the commensurate - incommensurate transition @xcite .
our model and the fk model are analogous in the sense that there are two natural length scales whose ratio changes as a function of other model parameters . in the fk model , however , these length scales are quenched , whereas in our model they are annealed .
another related experimental system can be seen for surface - induced ordering in thin film of diblock copolymers @xcite .
when the surface is periodically patterned , a tilt of the lamellae is induced in order to match the surface periodicity .
the situation becomes more complex if a copolymer melt is confined between two surfaces .
an interesting case arises when the spacing between the two surfaces is incommensurate with the lamellar periodicity @xcite . for systems out of equilibrium
, spatial resonances and superposition patterns combining stripes and/or hexagons were investigated in a reaction - diffusion model with interacting turing modes of different wavelengths @xcite .
these models were successful in reproducing hexagonal superlattice patterns which are known as `` black - eyes '' .
although the mechanism of pattern formation is different than in our model , we observe similar superposition patterns as reported in ref .
@xcite , such as hexagons on stripes or hexagons on hexagons ( not shown in this paper ) .
a more detailed study of the present model and several interesting extensions will be published elsewhere .
one possible extension is to consider vector order parameters describing , for example , the molecular tilt for coupled bilayers @xcite .
when the two order parameters are vectors , the nature of the transitions between different phases can be different , and even the square phase may exist in thermodynamic equilibrium @xcite .
we thank t. kato , s. l. keller , m. schick , and k. yamada for useful discussions .
yh acknowledges a research fellowship for young scientists no .
215097 from the japan society for the promotion of science ( jsps ) .
sk acknowledges support by kakenhi ( grant - in - aid for scientific research ) on priority areas `` soft matter physics '' and grant no .
21540420 from the ministry of education , culture , sports , science and technology of japan
. da acknowledges support from the israel science foundation ( isf ) under grant no .
231/08 and the us - israel binational foundation ( bsf ) under grant no .
2006/055 .
d. andelman and r. e. rosensweig , _ polymers , liquids and colloids in electric fields : interfacial instabilities , orientation and phase transitions _ , edited by y. tsori and u. steiner ( world scientific , singapore , 2009 ) . | we propose a model addressing the coupling mechanism between two spatially modulated monolayers .
we obtain the mean - field phase diagrams of coupled bilayers when the two monolayers have the same preferred modulation wavelength .
various combinations of the monolayer modulated phases are obtained and their relative stability is calculated . due to the coupling , a spatial modulation in one of the monolayers induces a similar periodic structure in the second one .
we have also performed numerical simulations for the case when the two monolayers have different modulation wavelengths .
complex patterns may arise from the frustration between the two incommensurate but annealed structures . |
heavy - ion collision experiments running at rhic have provided various information about the new state of matter , the quark - gluon plasma ( qgp ) . to remove theoretical uncertainties in understanding the nature of qgp , first principle calculations by lattice qcd are indispensable .
so far , most lattice studies at finite temperature have been performed using staggered - type quark actions with the fourth - root trick , whose theoretical basis is not fully established yet .
thus , a crosscheck with other actions such as the wilson - type quark actions are important to control and estimate systematic errors due to lattice discretization .
we investigate thermodynamic properties of qgp using the improved wilson quark action . adopting the fixed scale approach developed in @xcite , we perform finite - temperature simulations @xcite , while the corresponding zero - temperature configurations are taken from the results of the cp - pacs and jlqcd collaborations @xcite . at @xmath1 ,
interaction between a static quark and an antiquark can be studied by the heavy - quark potential evaluated from the wilson loop operator .
the resulting potential takes the coulomb form at short distances due to perturbative gluon exchange , while it takes the linear form at long distances due to confinement : v(r ) = - + r + v_0 .
[ eq : v ] for @xmath2 , inter - quark interaction may be studied by the heavy - quark free energy @xmath3 evaluated from a polyakov - line correlation in the color - singlet channel with coulomb gauge fixing @xcite : ( r , t ) = - t ^ ( ) ( ) , r=|- | , where @xmath4 with @xmath5 being the link variable in the temporal direction . at zero temperature
, we expect @xmath6 , while at high temperature , we can extract the debye screening mass @xmath7 from ( r , t ) = - e^-m_d(t ) r + 2f_q .
[ eq : md ] in the conventional fixed @xmath8 approach where @xmath9 is varied by changing the lattice spacing @xmath10 , @xmath11 and @xmath12 receive different renormalization at each @xmath9 , and thus are usually adjusted by hand such that @xmath11 and @xmath12 coincide with each other at a short distance , assuming that the short distance properties are insensitive to the temperature . on the other hand , in the fixed scale approach
, the temperature @xmath13 is varied by changing the temporal lattice size @xmath8 at fixed @xmath10 @xcite . because the spatial volume and the renormalization factors are common to all temperatures , we can directly compare the free - energies at different temperatures without any adjustment .
we show below that @xmath12 s for different @xmath9 approach to @xmath11 at short distances , which proves the expected insensitivity of the short distance physics to the temperature .
we employ the renormalization group improved gluon action and @xmath14 flavors of nonperturbatively @xmath15-improved wilson quark actions .
zero - temperature configurations are given by the cp - pacs and jlqcd collaborations @xcite at @xmath16 and @xmath17 .
finite temperature simulations with the same parameters are performed on @xmath18 lattices with @xmath19 , 10 , 8 , 6 and 4 , which correspond to @xmath20700 mev @xcite .
the absolute scale is estimated from the sommer parameter , @xmath21 fm .
was calculated by the cp - pacs and jlqcd collaborations @xcite.,width=291 ] [ fig:1 ] figure [ fig:1 ] shows the results of the heavy - quark potential @xmath11 at @xmath1 and the heavy - quark free energies @xmath12 at various temperatures as functions of @xmath22 . at @xmath1
, @xmath11 shows coulomb - like and linear - like behaviors at short and long distances , respectively .
a fit with eq .
( [ eq : v ] ) , as shown by the dashed line in fig . [ fig:1 ] , gives @xmath23 and @xmath24 gev . for @xmath2 , we note that the heavy - quark free energies @xmath12 at all temperatures converge to @xmath11 at short distances .
this means that the short distance physics is insensitive to temperature .
as stressed above , unlike the case of the conventional fixed-@xmath8 approach in which this insensitivity is assumed and used to adjust the constant terms of @xmath12 , our fixed scale approach enabled us to directly confirm this theoretical expectation . at large @xmath22
, @xmath12 departs from @xmath11 and eventually becomes flat due to debye screening . in fig .
[ fig:1 ] , the asymptotic values of @xmath12 at long distance are also compared with @xmath25 denoted by the arrows : here the thermal average of a single polyakov - line is defined as @xmath26 .
we find that @xmath12 converges to @xmath25 quite accurately at long distances . in order to extract the screening mass @xmath7
, we fit @xmath12 by the screened coulomb form , eq .
( [ eq : md ] ) .
the fit range is chosen to be 0.38 fm @xmath27 0.57 fm .
the left panel of fig .
[ fig:2 ] shows @xmath28 which does not have strong dependence on @xmath9 . to make a quantitative comparison to the result of the thermal perturbation theory
, we define the 2-loop running coupling by @xmath29 $ ] with the qcd scale parameter @xmath30 mev @xcite , where we assume a degenerated @xmath31 case .
then , the debye mass in the leading - order ( lo ) thermal perturbation theory is given by @xmath32 neglecting the quark mass effects .
a formula in the next - to - leading - order ( nlo ) is also available from the resummed hard thermal loop calculation @xcite : = g_2l(t ) , [ eq : m_d_nlo ] where @xmath33 is the magnetic screening mass . since the factor @xmath34 can not be determined in perturbation theory due to the infrared problem , we adopt @xmath35 calculated in a quenched lattice simulation @xcite as a typical value . in fig
. [ fig:2](left ) , the dashed lines represent the lo results for @xmath31 and the bold lines represent the nlo results for @xmath31 , for a range of the renormalization point @xmath36@xmath37 .
we find that the lo debye mass does not reproduce the lattice data in magnitude , while the nlo debye mass is approximately 50 % larger than the lo debye mass and shows a better agreement with the lattice data .
obtained from the heavy - quark free energy with those of the leading order ( lo ) and next - to - leading order ( nlo ) thermal perturbation theory for @xmath31 .
right : comparison among the debye masses for @xmath38 , @xmath39 @xcite and @xmath40 @xcite extracted from the heavy - quark free energy on the lattice .
arrows in the right panel indicate critical temperatures in @xmath39 and 0 cases.,title="fig:",width=234 ] obtained from the heavy - quark free energy with those of the leading order ( lo ) and next - to - leading order ( nlo ) thermal perturbation theory for @xmath31 .
right : comparison among the debye masses for @xmath38 , @xmath39 @xcite and @xmath40 @xcite extracted from the heavy - quark free energy on the lattice .
arrows in the right panel indicate critical temperatures in @xmath39 and 0 cases.,title="fig:",width=234 ] [ fig:2 ] finally we study the flavor dependence of @xmath7 .
[ fig:2](right ) shows @xmath7 in @xmath38 qcd ( this work ) , that in @xmath39 qcd with an improved wilson - quark action at @xmath41 @xcite , and that in the quenched qcd ( @xmath40 ) @xcite .
arrows on the horizontal axis indicate critical temperatures for @xmath39 ( @xmath42 mev at @xmath41 ) and @xmath40 ( @xmath43 mev ) .
we find that @xmath7 for @xmath44 is comparable to that for @xmath39 , whereas it is larger than that for @xmath40 .
a similar result was obtained with a staggered - type quark action @xcite .
we studied the free energy between a static quark and an antiquark at finite temperature in lattice qcd with @xmath14 flavors of improved wilson quarks on @xmath04 lattices in the high temperature phase .
we adopted the fixed scale approach which enables us to compare the free - energies at different temperatures directly without any adjustment of the overall constant . at short distances , the heavy - quark free energies , evaluated from the polyakov - line correlations in the color - singlet channel ,
show universal coulomb - like behavior common to that of the heavy - quark potential at zero temperature .
this is in accordance with the expected insensitivity of short distance physics to the temperature . at long distances ,
the heavy - quark free energies approach to twice the single - quark free energies calculated from the thermal average of a polyakov - line . also , we extracted the debye screening mass @xmath45 from the heavy - quark free energy and found that the next - to - leading order thermal perturbation theory is required to explain the magnitude of @xmath45 on the lattice .
comparison to the previous results at @xmath39 and @xmath40 , shows that the dynamical light quarks have sizable effects on the value of @xmath45 .
we thank the members of cp - pacs and jlqcd collaborations for providing us with the data at zero temperature .
this work is partially supported by grants - in - aid of the japanese ministry of education , culture , sports , science and technology , ( nos .
17340066 , 18540253 , 19549001 , 20105001 , 20105003 , 20340047 , 21340049 ) .
se is supported by u.s .
department of energy ( de - ac02 - 98ch10886 ) .
this work is supported also by the large - scale numerical simulation projects of ccs / accc , univ . of tsukuba , and by the large scale simulation program of high energy accelerator research organization ( kek ) nos.08 - 10 and 09 - 18 . | the free energy between a static quark and an antiquark is studied by using the color - singlet polyakov - line correlation at finite temperature in lattice qcd with 2 + 1 flavors of improved wilson quarks . from the simulations on @xmath0 , 10 , 8 , 6 , 4 lattices in the high temperature phase , based on the fixed scale approach , we find that , the heavy - quark free energies at short distance converge to the heavy - quark potential evaluated from the wilson loop at zero temperature , in accordance with the expected insensitivity of short distance physics to the temperature . at long distance ,
the heavy - quark free energies approach to twice the single - quark free energies , implying that the interaction between heavy quarks is screened . the debye screening mass obtained from the long range behavior of the free energy
is compared with the results of thermal perturbation theory . |
porous poly(lactic acid ) ( pla ) has been developed for tissue engineering scaffolds for decades [ 1 - 3 ] .
including poly(l - lactic acid ) ( plla ) , poly(d - lactic acid ) ( pdla ) and pla - based copolymers like poly(lactic - co - glycolic ) acid ( plga ) , these bio - based resins have been proved to be a successful candidate of scaffold materials with excellent biocompatibility and biodegradablity .
the tissue engineering requires sufficient interconnecting inner space in the scaffold for biofactor delivery , tissue growth , and the scaffold should be degradable after tissue s growth meanwhile providing proper mechanical strength to support the tissue engineering system . therefore the balance of growth space , degradation behavior and mechanical properties is the main concern of constructing a scaffold . with the respect to the particular requirements of certain tissue engineering ,
nowadays designed preparation and modification techniques of porous pla scaffold materials become an intensive interests - drawing subject [ 4 - 13 ] , which requires more understanding of the basic principles of physical and chemical properties , particular in the form of scaffold .
the crystallization of pla plays an important role in its mechanical properties and degradability .
generally the crystallized polymers have higher strength and mechanical modulus [ 14 ] . in the case of pla
, the crystallinity also significantly affects the degradability , with the general behavior that the degradation time is longer with higher crystallinity , as the crystal segments are more stable than amorphous area and prevent water permeation into it . for example , it was reported that plla takes more than 5 years for total degradation , whereas only about 1 year for the amorphous pla or pdlla [ 15 ] .
however unlike the crystallinity control for the inorganic components in the tissue engineering scaffold [ 16 ] , very rare reports concerned the crystallinity of the polymer scaffold materials .
one probable reason is that the preparation of porous scaffolds is a delicately process , where the control of crystallinity is usually difficult or unavailable . except the solvent casting / particle leaching method , in other widely used preparation methods such as electrospun fiber , phase separation , membrane lamination and gas foaming , the polymers are not able to experience a thermal treating step , i.e. the most common way to control the crystallinity [ 1 ] . in some works
the crystallinity is controlled by the raw materials itself , i.e. selecting raw materials of different molecular weight associated with different crystallization behaviors , or a particular processing procedure for chain cleavage to control the crystallinity [ 17 ] . and in some methods , even polymer with high crystallinity
can not be served as the raw materials to prepare the scaffold , for example the gas foaming technique reported by david j. mooney et al . [ 18 ] .
nevertheless , the study of the crystallization of the scaffold should hold considerable practical merits in tissue engineering , for example to fit the scaffold degradation time with the expected tissue growth time by the control of crystallinity ( if possible ) .
also , the scaffold structure may vary with crystallinity and influent the biological behavior of living tissue leaning on it .
park et al . reported a research on the sustained release of human growth hormone from semi - crystalline poly(l - lactic acid ) and amorphous poly(d , l - lactic - co - glycolic acid ) microspheres , which reveals that the morphological effect is important on protein release [ 19 ] .
crystallinity can be tailored in solvent casting / particulate leaching technique . however in this method the salt as porogens are solved in the solution and flow with the polymers during the casting , therefore without the immobilization of porogens the crystallinity behavior in the space - limited gap can not be tracked [ 1 ] . in this work , we modified the solvent casting / particulate leaching technique by diffusing the pla solution into a steady salt stack instead of solving the porogens .
the control of crystallinity was made possible by inserting a thermal treating step before leaching , while maintaining the stable structure of salt stack .
we have investigated the morphological effect of limited space on the crystallization of pla , and the porous structure with different crystallinity under thermal treatments .
the pla of label 4032d is purchased from natureworks , with l / d ratios from @xmath1:@xmath2 to @xmath3:@xmath2 .
the porogens is nacl of analytical grade .
the @xmath2:@xmath2 mixture of dichloromethane and chloroform is served as solvent .
the pla pellets are solved into dichloromethane and chloroform ( @xmath2:@xmath2 ) with the concentration of @xmath4 .
the nacl powder is thoroughly grounded and sieved with @xmath5 then @xmath6 sieve to screen the particles of sizes in between , and paved onto a petri dish where it forms a @xmath7 thick disc .
the pla solution is very slowly poured into the dish at the edge .
the pouring is as slow as that the solution diffuses inside the salt stack instead of flowing over the surface , also the slow diffusing guarantees the salt particles are not considerably moved by the liquid flowing to keep the inner structure of the salt stack stable . after pouring , the salt - pla solution composite is then placed in vacuum for @xmath8 for drying out the solvent .
the product is a solid dry pla - glued salt composite ready for thermal treatment .
the composite is placed in water for @xmath9 to leach out the salts after the thermal treatment for recrystallization .
the leached samples are freeze - dried for @xmath10 and stored in vacuum ready for characterizations .
the recrystallization of composite is processed by the heating and cooling process .
four samples were made to have different crystallinity .
one sample was kept as the original composite without thermal treatment for reference ( sample r ) .
the other three composites are heated in oven at @xmath11 for @xmath12 , then one composite ( sample a ) was immediately quenched in liquid nitrogen ; sample b was linearly cooled down at the rate @xmath13 @xmath14 . ; sample c was linearly cooled at @xmath13 @xmath14 till @xmath15 , kept at @xmath15 for @xmath16 , then cooled down with the same rate to room temperature . for comparison we also made two bulk pla samples with the same crystallization process of sample b and c and they were labeled as sample b and c. the crystallinity of samples were characterized by x - ray diffraction ( xrd ) ( d / max-2200/pc , rigaku corporation ) .
the porous structure of samples were revealed by scanning electron microscope ( sem ) ( nova nanosem 450 , fei ) .
the xrd results firstly dispel the doubt about the possibility of crystallization in confined geometry .
fig.1(1 ) shows the xrd of three thermal treated samples a , b and c ( note that the sample r is not included because its xrd curve almost overlap with the sample a , i.e. amorphous ) , and the bulk samples of b and c are shown in fig.1(2 ) . with the comparison of the bulk behavior , clear crystal peaks at @xmath17 and other minor peaks indicate that the heat treatment makes the crystallization possible but we can see significant impact of confined geometry , the crystallization of pla chains in porogens slits is harder and the crystallinity is lower than the bulk sample under the same treatment condition .
the crystallinity is calculated to be @xmath18 and @xmath19 in b and c , comparing to @xmath20 and @xmath21 in b and c. it is clear that with the same thermal treatment the bulk samples are much easier to crystallize with higher crystallinity .
this agrees with our expectation that the crystallization is more difficult in confined geometry , even the size of fibers and walls confining the pores is still in the magnitude of micrometers ( see the sem paragraph ) which far overweighted the diameter of chain segments , the chain movement and rearrangement is obstructed by the limited space .
another evidence that confined space impairs the crystallibility is the comparable crystallinity of b and c. the small difference is trivial due to many factors such as sample preparation or the baseline selection .
it implies that @xmath22 linear cooling process may reach the upper limit crystallinity for the porogens confined environment , while for bulk sample , the effects of staying at @xmath15 for longer time is significant on crystallinity .
although the samples are leached for @xmath9 there are unavoidable porogens residues left in the sample .
the peaks on @xmath23 and @xmath24 are nacl crystals [ 20 ] , whereas these two peaks do not exist in the bulk samples b and c. the sem results in fig.2 show the macro - structures of unheated original casting sample r and three samples with heat treatment a , b and c. a general observation of macro - structure confirms that the thermal treatment does not impressively affect the porous structure forming .
the pores and caves structure in each sample can be clearly observed with the pore size of around @xmath25 , which accords to the @xmath26 sieving process .
nevertheless the morphological effect of heat treatment is obvious . in fig.2(2 ) , ( 3 ) and ( 4 ) the reheated samples present the features of thinner pore walls , branches and fragments , while in the reference sample the pore wall is thicker with rod - like branches .
regardless of the crystallinity , heating the samples to @xmath11 ( the melting temperature of pla ) enables the polymer chains to remobilize and diffuse into the slits between salt particles where the solved chains had not diffused into and occupied .
the thinner pore wall indicates that the chain remobilization also moves the porogens and makes narrower space among them .
although we employ the steady salt stack to confine the recrystallization within the limited space , the porogens are only relatively `` stable '' comparing to typical solvent casting technique .
the quenching sample a has more fragmental structures than b and c , it is clear to understand the phenomenon that in sample a the polymer melt diffused into thinner slits is quenched to solidify its diffusing state of fragmental features . for sample b and c ,
the recrystallization process offers sufficient time for the diffused chains to mobilize and rearrange themselves to be more ordered , crystal structure .
this rearrangement provides a less fragmental structure on the macro - scale .
no obvious difference of macro - structures between sample b and c is observed , i.e. @xmath27 hours linear cooling is sufficient for recrystallization to achieve this structural effect , longer recrystallization time plays very little more effects on the crystallinity and the macro - structure , and this observation also agrees to the crystallinity results of b and c as indicated above .
the pla porous matrix for potential use in tissue engineering have been prepared by modified salt casting and particulate leaching technique .
the pla solution is diffused into relative stable salt stack , instead of solving salts with polymers in typical method . because the raw salt casting solidifies the salt - pla composite , we are able to insert a step of thermal treatment to recrystallize the polymer matrix before leaching process . in this way we are able to : 1 ) investigate the crystallization behavior of pla confined in limited space ;
2 ) develop an available crystallinity control option in porous pla scaffold preparation .
the xrd results indicate the crystallization of porous foams , in a manner of lower crystallibility than the bulk materials .
the marco - structure of porous samples are observed by sem , by obtaining the pores of around @xmath25 , it is revealed that the polymer foam may crystallize without significant structure damage .
the features of thinner pore walls , branches and fragments confirmed the effect of heating treatment .
both xrd and sem results of sample b and c indicate that 7 hours linear cooling is sufficient to achieve certain crystallinity and marco - structure .
thomson rc , shung ak , yaszemski mj , mikos ag .
polymer scaffold processing . in :
lanza rp , langer r , vacanti jp , editors .
_ principles of tissue engineering _ , 2nd ed . san diego : academic press , 2000 .
[ chapter 21 ] .
chihiro mochizuki , yuji sasaki , hiroki hara , mitsunobu sato , tohru hayakawa , fei yang , xixue hu , hong shen , shenguo wang . journal of biomedical materials research part b : applied biomaterials volume 90b , issue 1 , pages 290301 , july 2009 . | the porous pla foams potential for tissue engineering usage are prepared by a modified solvent casting / particulate leaching method with different crystallinity . since in typical method
the porogens are solved in the solution and flow with the polymers during the casting and the crystallinity behavior of pla chains in the limited space can not be tracked , in this work the processing is modified by diffusing the pla solution into a steady salt stack . with a thermal treatment before leaching while maintaining the stable structure of the porogens stack ,
the crystallinity of porous foams is made possible to control .
the characterizations indicate the crystallization of porous foams is in a manner of lower crystallibility than the bulk materials .
pores and caves of around 250@xmath0 size are obtained in samples with different crystallinity .
the macro - structures are not much impaired by the crystallization nevertheless the morphological effect of the heating process is still obvious .
porous materials , biomaterials , poly(lactic acid ) , tissue engineering scaffold , crystallization |
organic electronics now rapidly enters the consumer markets .
field - effect mobilities in polycrystalline organic semiconductors are comparable to the mobilities in hydrogenated amorphous silicon and the organic semiconductors can be deposited by thermal evaporation or from solution at low costs on large areas .
pentacene and its soluble derivatives are among the most promising organic semiconductors for organic thin - film transistor applications.@xcite organic semiconductors are distinct in that they generally consist of neutral molecules which interact by rather weak forces ( predominantly van der waals forces ) . the energy associated with a van der waals bond is 10@xmath3 - 10@xmath4ev , i.e. orders of magnitude smaller than the energy of a covalent bond ( several ev).@xcite therefore , organic semiconductors are soft , have a low mechanic strength and a high compressibility , low melting and sublimation temperatures and a large thermal expansion coefficient .
the weak interaction forces between the molecules not only enable low - cost processing but also lead to peculiarities such as the existence of at least four different pentacene polymorphs or an even negative thermal expansion coefficient.@xcite as in the case of amorphous inorganic materials , the charge transport in organic semiconductors can be described with a mobility edge or transport level concept . in the case of crystalline small molecule organic semiconductors such as pentacene , the mobility edge may be identified with the valence or conduction band edge and the charge carriers are transported in extended states .
however , the charge carriers are transported in states above a mobility edge and are multiply trapped by and thermally released from trap states below the mobility edge .
consequently , the conduction of charge in organic semiconductors and thus the transistor characteristics critically depend on trap states .
energetically deep traps influence the subthreshold swing of a device which needs to be sufficiently steep for a low power operation .
the density of shallow traps , on the contrary , can dominate the effective field - effect mobility .
trap states in organic semiconductors have long been a subject of investigation.@xcite this topic is currently attracting much attention due to the crucial importance of trap states for the emerging applications of organic field - effect transistors.@xcite the shallow traps in particular are poorly understood .
the density of trap states as a function of energy can be derived from the transistor characteristics . on the one hand ,
a density of states function can be postulated a priori and the corresponding transistor characteristic can be simulated by means of a suitable computer program .
the density of states function is then iteratively refined until , after a number of predictor - corrector loops , good agreement between the measured characteristic and the simulated curve is achieved.@xcite on the other hand , the density of states function can be calculated from the linear regime transfer characteristics in a straightforward fashion .
a number of such extraction schemes has been suggested.@xcite this approach has the advantage of giving an unambiguous result but , depending on the complexity of the extraction scheme , spurious errors may result from simplifying assumptions .
organic thin - film transistors ( tfts ) are most often characterized after the samples have been exposed to ambient air and in situ electrical characterization is very rare.@xcite the organic semiconductors , however , are generally presumed to be sensitive to water vapor and oxygen . moreover ,
the transistors are often characterized by two - terminal measurements which do not allow to distinguish between contact effects and effects of the semiconducting layer .
the gated four - terminal method yields the field - effect conductivity and the effective field - effect mobility free from contact effects .
it also allows for an extraction of the device contact resistance.@xcite for this work pentacene served as a prototypical small molecule semiconductor and we have investigated the performance of pentacene thin - film transistors by in situ gated four - terminal measurements .
the corresponding trap states functions were derived with a straightforward extraction scheme which had been successfully used to understand trap states in hydrogenated amorphous silicon .
pentacene from aldrich ( purum ) was sublimation purified twice and was introduced into the evaporation chamber soon after the purification .
as substrates we used heavily doped si wafers with a 260 nm thick sio@xmath5 layer .
the substrates were cleaned with hot acetone and hot isopropanol ( mos grade ) in an ultrasonic bath . immediately after the cleaning
, the substrates were mounted on a sample holder and were introduced into the in situ device fabrication and characterization system via the load lock ( fig .
[ insitu ] ) . the evaporation chamber and
the prober station were both separated from the load lock with a gate valve and were constantly kept under vacuum ( base pressure @xmath6 mbar ) .
the vacuum in the evaporation chamber and in the prober station was maintained with a cryopump and a turbopump , respectively .
the substrates were introduced into the evaporation chamber with transfer rod 1 and were placed on a shadow mask for the pentacene evaporation . a high precision mask positioning mechanism allowed for a proper adjustment of the mask with respect to the substrates . in situ device fabrication and characterization system :
pentacene thin - film transistors were fabricated and characterized electrically by gated four - terminal measurements at a base pressure of order 10@xmath0 mbar without breaking the high vacuum . ]
the substrates were kept in high vacuum for approximately 24h prior to the device fabrication .
after that time , also the pressure in the turbo - pumped load lock was @xmath6 mbar .
a 50 nm thick film of pentacene was evaporated through the shadow mask onto the si / sio@xmath5 substrates while the substrates were kept at room temperature .
after the pentacene evaporation , the sample holder was positioned on a shadow mask for the gold evaporation with transfer rod 1 and the pentacene tfts were completed by evaporating 40 nm thick gold electrodes .
the resulting transistor test structures are shown schematically in fig .
[ device ] .
the transistors consisted of a well - defined stripe of pentacene and had voltage sensing electrodes with little overlap to the pentacene film .
it has been shown that the use of a `` masked '' pentacene film and a proper alignment of the electrodes is important for the four - terminal measurement.@xcite the channel length and width of the devices was @xmath7@xmath8 m and @xmath9@xmath8 m , respectively . the voltage sensing electrodes were situated at @xmath10 and @xmath11 , such that the distance between these electrodes was @xmath12@xmath8 m . transistor test structure for the gated four - terminal measurements .
the alignment was achieved by means of a high precision in situ mask positioning mechanism .
the channel length and width were @xmath7@xmath8 m and @xmath9@xmath8 m and the distance between the voltage sensing electrodes was @xmath12@xmath8 m . with the voltage sensing electrodes , the potentials @xmath13 and @xmath14 were measured .
] after the completion of the device fabrication , the samples were transported to the prober station by means of transfer rod 1 and transfer rod 2 ( fig . [ insitu ] ) . for the electrical measurements we used a hp 4155a semiconductor parameter analyzer connected to five microprobers at the prober station
transfer characteristics in the linear regime were measured in steps of 0.2v ( drain voltage @xmath15v ) . in all cases ,
the forward and the reverse sweeps were measured .
the integration time was 20ms and the delay time was 0s .
in addition to the drain current @xmath16 , the voltage drops @xmath13 and @xmath14 between the grounded source and the two voltage sensing electrodes were measured at each gate voltage @xmath17 ( gated four - terminal measurement ) .
all electrical measurements were carried out in the dark .
a device was measured initially @xmath18h after the completion of the pentacene evaporation .
subsequently , the same device was measured regularly ( normally twice a day ) for approximately one week while being kept in the prober station at @xmath6 mbar in the dark . in some experiments
, the effect of oxygen or nitrogen exposure on the device characteristics was investigated .
this was done by introducing a continuous flow of gas ( pureness @xmath19vol.-@xmath20 ) into the prober station through a leak valve , thus adjusting the partial pressure of the gas within the prober station .
a few days prior to the device fabrication for the gas exposure experiments , the prober station was filled with the respective gas through the leak valve in order to flush the gas supply line and the valve .
the prober station was then re - evacuated and the supply lines were held at an overpressure of @xmath21atm until the leak valve was opened in the experiment . in other experiments ,
the influence of annealing was explored by means of an electrical heating element on the sample holder tray in the prober station .
in this section , we elucidate the extraction of key parameters from the device characteristics .
linear regime transfer characteristics are particularly interesting for the study of semiconductor physics , because the gradual channel approximation is valid which leads to a drastic simplification of the device physics . the two - dimensional problem is approximated by two one - dimensional equations , i.e. possion s equation for the charge distribution perpendicular to the insulator - semiconductor interface and a simple ohmic current - voltage relationship . in single crystalline inorganic mosfets above a threshold voltage ,
essentially all the trap states are filled and the charge which is induced by the gate appears in the valence / conduction band.@xcite both the constant mobility in the above threshold regime and the threshold voltage are important device parameters .
this approach may also be valid in the case of organic single crystal transistors , i.e. organic field - effect transistors with a very low trap density .
however , contact effects often severely affect the characteristics of these devices because of the low channel resistance.@xcite as in the case of amorphous silicon field - effect transistors , the equations developed for single crystalline mosfets are not suitable to describe organic tfts due to the increased trap density.@xcite depending on the density of trap states , the majority of charge carriers induced by the gate may be trapped even at relatively high gate voltages.@xcite provided that contact effects are negligible , the drain current of an organic tft in the linear regime is given by @xmath22 @xmath23 is the field - effect conductivity which is the effective field - effect mobility @xmath24 multiplied with the total gate induced charge per unit area @xmath25 , i.e @xmath26 @xmath27 is the flatband voltage and @xmath28 the capacitance per unit area .
the effective field - effect mobility is one of the most important device parameters and , for a p - type semiconductor such as pentacene , @xmath24 can be written as @xmath29 where @xmath30 and @xmath31 are respectively the density of free and trapped holes per unit area .
@xmath32 is the extended state mobility . the effective field - effect mobility @xmath24 is expected to increase with gate voltage even at relatively high gate voltages , because the ratio @xmath33 increases as the valence band is bent towards the fermi energy.@xcite the flatband voltage is the gate voltage which needs to be applied in order to enforce flat bands at the insulator - semiconductor interface .
a non - zero flatband voltage can result from a difference of the semiconductor and the gate fermi levels .
more importantly , the flatband voltage is influenced by charge that is permanently trapped at the interface or within the gate dielectric and only the latter contribution can explain the large shifts of the transfer characteristics that are sometimes observed in experiments.@xcite the flatband voltage is approximately equal to the onset voltage of the device .
the onset voltage is defined as the gate voltage , where the drain current , as a function of gate voltage , rises sharply if plotted on a logarithmic scale . in the present work
we assume that the flatband voltage is equal to the onset voltage . from a technological point of view
it is useful to define a threshold voltage which marks the transition between the regime below threshold to the regime above threshold . in the above - threshold regime ,
the deep traps are filled and the field - effect mobility is less strongly dependent on gate voltage . the above - threshold regime in an organic tft can be understood as being in between the below - threshold regime and the above - threshold regime of a single crystalline mosfet.@xcite it can be shown that the field - effect mobility in the above - threshold regime follows a power law @xmath34 and this allows for a refined extraction of the field - effect mobility and the threshold voltage @xmath35.@xcite in this study we do , however , follow the ieee standard for the characterization of organic transistors and materials .
the field - effect conductivity can be calculated from @xmath36 with eq .
[ sigma1 ] and eq .
[ sigma0 ] @xmath24 can be approximated as @xmath37 since this approach is frequently used it has the advantage that the values of @xmath24 can easily be compared .
moreover , the definition and extraction of a threshold voltage is not necessary .
however , small errors in the absolute value of the field - effect mobility and its gate voltage dependence are to be expected since the derivation rests on the assumption of a weak dependence of the field - effect mobility on gate voltage.@xcite in an organic field - effect transistor , a significant fraction of the drain voltage @xmath38 may drop at the contacts which can introduce significant errors when extracting the field - effect conductivity and the field - effect mobility . from gated - four terminal measurements ,
the conductivity can be derived without error with @xmath39 .
@xmath40 is the distance between the voltage sensing electrodes and @xmath41 the voltage drop between these electrodes ( fig .
[ device ] ) . the contact - corrected field - effect conductivity is then given by @xmath42 the effective field - effect mobility @xmath24 is not influenced by contact effects when calculated from @xmath43 in the following we use the expressions `` two - terminal conductivity '' and `` two - terminal mobility '' as short hand for eq .
[ sigma0 ] and eq .
we furthermore use the abbreviations `` four - terminal conductivity '' and `` four - terminal mobility '' for the quantities defined in eq .
[ sigma2 ] and eq .
[ mu ] .
the device contact resistance @xmath44 was extracted from the four - terminal measurement and was compared to the device channel resistance @xmath45 .
we now assume a linear voltage drop all along the channel , i.e. from the source to the drain . with this assumption ,
the voltage drop across the transistor channel is @xmath46 and the voltage drop at the contacts is @xmath47 .
the contact resistance is thus given by @xmath48 and the channel resistance by @xmath49 the gate voltage dependence of the field - effect mobility reflects the spectral density of trap states close to the valence band .
a number of direct extraction schemes has been suggested to obtain the underlying density of states function.@xcite in these approaches , the relevant energy scale is derived from the activation energy @xmath50 of the current ( i.e. the field - effect conductivity ) which is obtained from temperature dependent measurements . if , however , the electrical characteristics of a transistor change on a timescale comparable to the time of a temperature dependent measurement ( hours ) , this approach is not suitable .
grnewald et al . have suggested an extraction scheme of high accuracy for amorphous silicon thin - film transistors which allows to convert a single linear regime transfer characteristic into the underlying density of states function.@xcite it is based on surprisingly few simplifying assumptions including : a. the gradual channel approximation , b. the semiconductor is homogenous normal to the dielectric - semiconductor interface , and c. insulator surface states only introduce an initial band bending without applied field , i.e. contribute to a non - zero flatband voltage @xmath27 .
extraction schemes often rest on the abrupt approximation : all the charge is assumed to reside in a region close to the dielectric - semiconductor interface of depth @xmath51 .
grnewald s method , however , is not based on this simplification but takes proper account of the gate - induced band bending . in the following ,
the key equations of the extraction scheme are given .
within the boltzmann approximation , the field - effect conductivity is given by @xmath52 now @xmath17 is the gate voltage above the flatband voltage , @xmath53 is the thickness of the pentacene film , @xmath54 is the band shift as a function of the distance @xmath55 from the insulator - semiconductor interface and @xmath56 is the conductivity at flatband .
@xmath57 can be approximated as @xmath58 @xmath59 is an effective density of extended states and @xmath60 the energy of the valence band edge far from the insulator - semiconductor interface .
the situation is depicted in fig .
[ bandbending ] .
the complicated dependence of the band shift on the space coordinate @xmath55 in eq .
[ sigma3 ] can be eliminated and an equation can be derived , which implicitly contains the interface potential @xmath61 as a function of gate voltage : @xmath62\end{aligned}\ ] ] @xmath63 is the thickness of the gate insulator and @xmath64 and @xmath65 are the dielectric constants of the insulator and the semiconductor , respectively ( see ref . for a derivation of this equation ) . for each gate voltage , eq . [ implicit ] can be evaluated numerically and a value for the interface potential @xmath66 is obtained .
eventually , we have the complete function @xmath67 .
sketch of the energetics near the sio@xmath5/pentacene interface : the application of a gate voltage @xmath17 above the flatband voltage leads to a bending of the valence band ( vb ) and of the conduction band ( cb ) . at the interface ( @xmath68 ) , the band shift is @xmath69 . under these conditions ,
the energy @xmath70 of specific trap states ( dashed line ) is raised at the interface such that it coincides with the fermi energy @xmath71 of the sample . ] in most general terms , the total hole density @xmath72 ( volume density ) is @xmath73de,\ ] ] where n(e ) is the density of states function .
the total hole density @xmath74 can be calculated as @xmath75 i.e. by a numerical differentiation of the previously obtained set of data @xmath67.@xcite differentiating eq .
[ total1 ] with respect to the band shift yields @xmath76 several deconvolution methods exist to solve this type of equation for @xmath77 , e.g. with cubic spline functions.@xcite however , for a slowly varying density of states function ( absence of monoenergetic states ) ,
the difference between @xmath77 and @xmath78 is expected to be relatively small on a logarithmic scale.@xcite consequently , we obtain the final result by a numerical derivation of @xmath79 with respect to the interface potential @xmath66 according to @xmath80 within this zero temperature approximation , the band shift at the interface @xmath81 is equal to the energy of the respective traps relative to the fermi energy @xmath71 of the sample , i.e. @xmath82 ( fig .
[ bandbending ] ) .
we have used grnewald s method to interpret the current voltage characteristics from pentacene tfts .
we used the `` four - terminal conductivity '' as a starting point which allows to extract a density of states function free from contact artifacts . a simple matlab@xmath83 coding allowed for the calculation of the density of states
function from eq .
[ implicit ] , [ total2 ] and [ total3 ] .
fig . [ currents ] shows the transfer characteristic from a pentacene tft measured 4h and 140h after the completion of the pentacene evaporation .
the device was kept at @xmath84 mbar all along .
after 140h , the device shows an increased on - current . in addition , the current hysteresis is reduced : at a current level of @xmath85a , the difference between the forward and the reverse sweep is 3.8v after 4h and 1.2v after 140h .
the subthreshold swing is essentially unaffected by the high vacuum storage .
there is a small shift of the onset voltage to more positive voltages from -6.4v after 4h to -4.9v after 140h .
transfer characteristic from a pentacene tft measured 4h and140h after the completion of the pentacene evaporation .
the storage under high vacuum conditions leads to an increased on - current and to a reduced current hysteresis . ]
an increase in on - current can be due to changes of the pentacene film and/or to a reduction of the device contact resistance .
the gated four - terminal method can disentangle the field - effect conductivity and the device contact resistance . in fig .
[ sigma ] we show the `` four - terminal conductivity '' after 4h and after 140h , as derived from the forward sweeps ( eq . [ sigma2 ] ) .
the `` two - terminal conductivity '' ( eq . [ sigma0 ] ) is shown for comparison .
the `` four - terminal conductivity '' is increased after 140h , which reveals changes of the pentacene film .
the difference between the `` four - terminal conductivity '' and the `` two - terminal conductivity '' is reduced after 140h , indicative of an additional contact resistance reduction .
the `` four - terminal conductivity '' increases with time .
the graph shows the `` four - terminal conductivity '' ( circles ) after 4h and after 140h .
the dashed lines indicate the `` two - terminal conductivity '' for comparison . ]
[ mobility ] shows the `` four - terminal mobility '' derived with eq .
[ mu ] and the `` two - terminal mobility '' calculated from eq .
as expected , the mobility increases monotonically with gate voltage . when comparing both measurements ,
a significant improvement in mobility can be ascertained . at @xmath86v ,
the mobility is @xmath87@xmath1/vs after 4h and @xmath88@xmath1/vs after 140h , i.e. @xmath8 has increased by a factor of 2 .
`` four - terminal mobility '' ( circles ) as a function of gate voltage .
there is a significant improvement in mobility with time . at @xmath86v ,
the mobility increases from @xmath87@xmath1/vs after 4h to @xmath88@xmath1/vs after 140h , i.e. by a factor of two .
the dashed lines show the `` two - terminal mobilities '' for comparison . ] in fig .
[ resistance ] we show the width - normalized contact resistance @xmath89 according to eq .
[ rcontact ] and , for comparison , the width - normalized channel resistance @xmath90 ( eq . [ rchannel ] ) .
there is a drastic reduction in contact resistance . at @xmath91v
the contact resistance decreases by a factor of @xmath92 from @xmath93@xmath94 cm to @xmath95@xmath94 cm .
the channel resistance decreases by a factor of @xmath96 .
the channel resistance is always higher than the contact resistance : at @xmath91v and after 4h the channel resistance is @xmath97 times larger than the contact resistance and after 140h it is @xmath98 times the contact resistance .
thus , the device is always dominated by the channel resistance . width - normalized contact resistance ( triangles ) after 4h and after 140h .
the contact resistance is drastically reduced : at @xmath91v it decreases by a factor of @xmath92 .
the graph also contains the respective width - normalized channel resistances ( circles ) for comparison .
the contact resistance is always lower than the channel resistance . ] in sec .
[ evstart ] and [ evend ] we show , that the performance improvement is not due to doping by a residual gas but due to a healing of defects at room temperature . to investigate the energetic position of these defects , we have applied the scheme described in sec .
[ advanced ] .
the first step was to calculate the interface potential @xmath66 as a function of gate voltage with eq .
[ implicit ] . we have assumed a dielectric constant of @xmath99 for sio@xmath5 and @xmath100 for pentacene .
[ together](a ) shows the result of the extraction for the measurement after 4h and after 140h . in a second step ,
the density of states was calculated by two numerical differentiations of @xmath101 according to eq .
[ total2 ] and eq .
[ total3 ] .
some degree of data smoothing was applied in order to obtain a reasonably smooth density of states function .
[ together](b ) shows the final result for the measurement after 4h and after 140h .
a step width of 0.2v in the gate voltage sweeps leads to a good resolution of the deep states .
the fermi energy of the sample coincides with the zero point of the energy scale in fig .
[ together](b ) . at high gate voltages ,
the valence band edge at the interface is close to the fermi level .
consequently , the fermi level is approximately @xmath102ev from the valence band edge at flatband .
for comparison , the bandgap of pentacene is 2.24 - 2.5ev.@xcite the interface potential in fig .
[ together](a ) reflects the spectral density of trap states . at low gate voltages bands
bend easily , which indicates a low trap density . at high gate voltages , however , band bending is more difficult and this is indicative of a high trap density very close to the valence band edge .
the energetics in fig .
[ together](a ) are in good agreement with activation energies @xmath50 from temperature dependent measurements of pentacene - based transistors.@xcite in these studies the activation energy is derived from @xmath103 by means of arrhenius plots . assuming the abrupt approximation , eq .
[ sigma3 ] can be approximated as @xmath104 and we have @xmath105 .
the activation energy from the arrhenius plots is approximately equal to the energetic difference between the fermi level and the valence band at the interface .
the activation energy is found to be @xmath106ev near the onset voltage of the devices and @xmath107ev when the devices are turned on completely .
@xmath50 shows a similar functional dependence on gate voltage as @xmath101 in fig .
[ together](a).@xcite from fig .
[ together](b ) it is clear that it is the shallow traps with energies approximately @xmath108ev from the valence band edge which are reduced by the high vacuum storage .
it is the density of these states that influences the value of the field - effect mobility @xmath24 .
a relatively small reduction leads to a significant improvement in field - effect mobility .
the traps which are deeper in energy are essentially unaffected , resulting in an almost identical subthreshold swing of the transfer characteristics .
both after 4h and after 140h , the density of states function can reasonably well be approximated by a single exponential function @xmath109 it is however slightly steeper than exponential .
fitting the curves in fig .
[ together](b ) for @xmath110ev to eq .
[ expo ] yields the parameters @xmath111mev for the measurement after 4h and @xmath112mev for the measurement after 140h .
this is in good agreement with results for pentacene tfts obtained with a device simulation program.@xcite in ref .
, characteristic slopes of @xmath113mev and 37mev are specified and in ref .
a slope of @xmath111mev was determined .
the effects described above , i.e. a significant increase in the `` four - terminal mobility '' , a drastic reduction in the contact resistance and a reduction in the current hysteresis , have been observed in all experiments .
[ nointerference ] shows the evolution of the `` four - terminal mobility '' and the contact resistance with time for four different runs .
the values are for @xmath114 and are normalized by the value obtained after @xmath18h .. for the mobility , an average in the range @xmath115v to @xmath116v was taken since in some cases the derivative of the drain current was more noisy . ] a time span of at least 4h was allowed in between subsequent measurements .
the absolute values for the mobility and the contact resistance from the initial measurement are summarized in table [ table1 ] .
.[table1 ] `` four - terminal mobility '' @xmath117 , at @xmath86v , width - normalized contact resistance @xmath118 at @xmath91v and onset voltage @xmath119 from the initial measurements of different experiments .
the mobility increased by as much as a factor of 2 in the course of an experiment such that mobilities up to 0.45@xmath1/vs were achieved . [ cols="<,^,^,>",options="header " , ] upper panel : `` four - terminal mobility '' at @xmath114 normalized by the mobility @xmath117 of the initial measurement for four different runs .
the lower panel shows , on a logarithmic scale , the respective values for the contact resistance @xmath120 at @xmath121 relative to the contact resistance @xmath122 from the initial measurement .
the field - effect mobility increases with time and the contact resistance reduces with time in all experiments .
there is some variation in the rate of the effects . ]
table [ table1 ] also contains the respective onset voltages .
initially , the onset voltage is between @xmath123v and @xmath124v and shifts in all cases by less than 2.5v during the course of an experiment .
the reduction of the current hysteresis takes place in the early stages of the experiments . in all four runs ,
the current hysteresis is significantly reduced after the first @xmath125h to @xmath126v at a current level of @xmath85a .
subsequently , there is only a small further reduction of the current hysteresis .
we now proceed by providing experimental evidence that the performance improvement is not due to doping of the pentacene thin films by residual oxygen or nitrogen and show that the performance improvement is a thermally promoted process . even at a pressure of order @xmath127 mbar
the time for the formation of a monolayer of residual gas molecules is less than ten minutes.@xcite in the case of semiconducting polymers experimental evidence indicates that doping leads to an increased field - effect mobility.@xcite this may be understood if one assumes the charge transport to take place by variable - range hopping and the doping to increase the density of states / hopping sites.@xcite in crystalline small molecule organic semiconductors , however , charge transport in extended states has to be considered .
doping raises the fermi level of the sample .
this results in an increase of the ( flatband ) conductivity , since the density of free holes is increased ( eq . [ sigmanull ] ) . when the device is turned on , however , the current flow is due to the gate induced holes and the holes from a chemical doping are negligible in eq . [ mu0 ] .
thus it would not immediately be obvious how chemical doping would increase the field - effect mobility in pentacene .
both oxygen and nitrogen have been reported to have the capability of doping pentacene .
gas exposure is found to lead to a shift of the transistor transfer characteristic to more positive voltages which corresponds to a shift of the fermi level.@xcite in other studies , an increase in conductivity of gas exposed pentacene thin - films or single crystals was observed in two - terminal measurements without gate - field induced charge . the increased conductivity is ascribed to an increased charge carrier density caused by doping.@xcite there seems to be no evidence that doping leads to an increased effective mobility in pentacene .
recent measurements on rubrene single crystals also show doping not to increase the field - effect mobility.@xcite the issue of reversibility and the importance of light for the doping is not completely clear .
doping is observed when pentacene thin films are exposed to relatively high partial pressures ( 0.01atm-1atm ) of oxygen or nitrogen in the presence of light.@xcite the doping effect by oxygen is reported to be negligible in the absence of light.@xcite ultraviolet photoelectron spectroscopy ( ups ) before and after exposure to @xmath128 mbar oxygen could not detect a lasting effect on the position of the energy levels.@xcite according to ref . , exposing pentacene tfts to an oxygen partial pressure of @xmath129 mbar leads to a doping of the films .
while the mobility increase in crystalline / polycrystalline samples can not be easily linked to chemical doping , a reduction of the contact resistance might be reconciled with doping . from inorganic semiconductor physics
it is well known that doping close to a contact reduces the contact resistance .
moreover , it has been shown with ups that oxygen , at a low partial pressure , can lead to a lowering of the injection barrier for holes at an au / pentacene interface.@xcite the results of our studies concerning the effect of gas exposure on the field - effect mobility and on the contact resistance are discussed in the following . if e.g. oxygen was responsible for the increase in mobility and/or reduction in contact resistance , an increase in the oxygen partial pressure should accelerate the rate of the respective process .
[ oxygengrey ] shows the influence of oxygen on the `` four - terminal mobility '' and on the contact resistance ( see table .
[ table1 ] for the initial parameters ) .
after 41h , the oxygen partial pressure was raised from a total turbo - pumped background pressure of order @xmath127 mbar to @xmath130 mbar .
after 69h , the partial pressure was increased to @xmath131 mbar and eventually , after 98h , the chamber was re - evacuated .
the increase in oxygen partial pressure does not accelerate the gradual increase in mobility or the reduction in contact resistance . the dashed red line in fig .
[ oxygengrey ] is an estimate for the time - dependence of the contact resistance if the sample had not been exposed to oxygen .
consequently , the increase in oxygen partial pressure even leads to a delay of the reduction in contact resistance .
it is noteworthy that neither the small current hysteresis nor the subthreshold swing of the devices are effected by the oxygen exposure .
since oxygen at a partial pressure of @xmath130 mbar for 28h and at a partial pressure of @xmath131 mbar for 29h appears not to accelerate the development of the device parameters with time one would conclude that oxygen at the base pressure is not responsible for the performance improvement .
effect of oxygen exposure : the graph shows the `` four - terminal mobility '' at @xmath114 ( upper panel ) and the contact resistance at @xmath121 ( lower panel ) as a function of time .
the dashed red line is an estimate of the time - dependence of the contact resistance without oxygen exposure .
the drastic increase in oxygen partial pressure neither leads to a more rapid increase in mobility nor to a more rapid decrease in contact resistance . on the contrary
, the oxygen exposure slows down the decrease in contact resistance . in the inset
we show the device onset voltage as a function of time .
the increase in oxygen partial pressure does not lead to a sudden shift of the onset voltage which would be indicative of doping . ]
the inset in fig .
[ oxygengrey ] shows the onset voltage for each of the measurements .
the overall shift of the onset voltage is small and smooth ( @xmath132v over @xmath133h ) .
in particular , there is no sudden shift of the transfer characteristic after an increase in oxygen partial pressure .
therefore , there is no evidence that oxygen at partial pressures @xmath134 mbar leads to a doping of the films .
[ nitrogengrey ] shows the analogous experiment with nitrogen .
after 50h the nitrogen partial pressure was increased to @xmath130 mbar , followed by an increase to @xmath131 mbar after 69.5h .
after 98h , the system was pumped down to the base pressure of order @xmath127 mbar .
the increase in nitrogen partial pressure does not accelerate the increase in mobility or the reduction in contact resistance .
similarly to the oxygen experiment we have estimated the time - dependence of the contact resistance if the sample had not been exposed to nitrogen ( dashed red line in fig .
[ nitrogengrey ] ) . clearly , the nitrogen exposure leads to a slowing down of the decrease in contact resistance .
the nitrogen exposure leaves the small current hysteresis and the subthreshold swing unaffected .
residual nitrogen is not responsible for the performance improvement .
effect of nitrogen exposure : a nitrogen partial pressure of @xmath131 mbar leads to a slowing of the increase in mobility and decrease in contact resistance .
the dashed red line is an estimate for the time - dependence of the contact resistance if the sample had not been exposed to nitrogen .
the inset shows the device onset voltage .
when the nitrogen partial pressure is increased to @xmath130 mbar , there is a sudden shift of the onset voltage which is indicative of doping . ] the inset in fig .
[ nitrogengrey ] shows the device onset voltage as a function of time . when the nitrogen partial pressure is increased to @xmath130 mbar
there is a sudden shift of the onset voltage to more positive voltages by 1.2v .
we take this as evidence for doping by nitrogen at a partial pressure of @xmath130 mbar .
when the pressure is increased to @xmath131 mbar , there is no additional marked shift of the onset voltage . annealing pentacene thin films at moderate temperatures ( e.g. 50@xmath135c ) results in an improved crystallinity of the films as seen by xrd.@xcite
when concentrating on the channel region adjacent to the gate dielectric in a field - effect transistor , the annealing is found to leave the mobility unchanged ( ref . ) or to lead to an increased mobility ( ref . ) .
since the interaction between the pentacene molecules is of the weak van der waals type , it is not too surprising that annealing at moderate temperatures leads to an improved crystallinity of the films .
moreover , it seems plausible that structural defects can be `` annealed '' even at room temperature , which will account for the effects we observe . in this scenario
we would expect an acceleration of the performance improvement at higher temperatures .
[ annealinggrey ] shows the influence of such an annealing at 320k ( see table [ table1 ] for the initial parameters ) .
after 44.5h , the sample temperature was slowly raised from rt to 320k at a rate of @xmath136@xmath135/min .
transfer characteristics were measured after the temperature of 320k had been reached .
after 104.5h , the heating was switched off , followed by a slow cooling of the sample to room temperature .
two effects can be discerned in the annealing process .
first , the overall level of the mobility and the contact resistance are affected by the increase in temperature since both quantities depend on temperature .
second , and more significant in the present context , is the rate of change : after raising the temperature , the increase in mobility and the decrease in contact resistance is accelerated significantly .
this is indicated by the dashed red lines in fig .
[ annealinggrey ] .
effect of annealing at 320k : the increased temperature accelerates the increase in mobility and the decrease in contact resistance as indicated by the dashed red lines . ] in an additional experiment ( not shown ) , the sample temperature was raised from room temperature to 310k after 45.5h and was increased to 320k after 74h . even at 310k , which is not much above room temperature ( @xmath137 ) , the increase in mobility and the decrease in contact resistance is accelerated noticeably .
all these results taken together clearly indicate an improvement of the electronic parameters driven by a thermally promoted process , and not by chemical doping .
we suggest this process to be a healing of structural defects .
we now discuss the microscopic origin of the relevant traps and suggest how a reduced trap density can lead to a reduced contact resistance .
pentacene thin - films on si / sio@xmath5 substrates are known to have a layered structure and the layers are parallel to the substrate . within these layers ,
the molecules are arranged in a herringbone pattern and are oriented almost perpendicular with respect to the substrate .
it has been shown by high impedance stm that some of the pentacene molecules in the layers are displaced along their long molecular axis , while the two - dimensional packing is not disturbed.@xcite with electronic structural calculations it could be shown that the displaced molecules result in traps @xmath138ev from the valence band edge.@xcite in sec .
[ secdos ] we have shown that only the shallow traps @xmath2ev from the valence band edge are reduced during the high vacuum storage at room temperature .
we suggest that a major cause of the shallow traps in pentacene thin films are pentacene molecules within the grains that are slightly misplaced , i.e. various types of structural point defects .
some of these defects are in a metastable state before relaxation .
they require only a small amount of energy in order to align which can be provided by the thermal energy at room temperature .
this is a direct manifestation of the weak intermolecular interaction which is characteristic of small molecule organic semiconductors .
the `` annealing '' of shallow traps at room temperature can easily explain the increase in the effective field - effect mobility . in a simplistic view ,
the contact resistance is given by the energetic difference between the work function of the metal and the ionization energy of the pentacene .
in reality , however , a clear correlation between the metal work function and the contact resistance is often not observed.@xcite interface states can significantly affect the energetics at the metal - pentacene interface.@xcite in a top contact device , also the film resistance should contribute to the contact resistance .
the film resistance in pentacene devices with gold top contacts has even been suggested to dominate the contact resistance.@xcite the situation is illustrated in fig .
[ transistor ] : the hole injection at the source / pentacene interface is good , but the holes must cross the pentacene film in order to reach the channel at the insulator - semiconductor interface . the intrinsic mobility perpendicular to the molecular layers is lower than parallel to the layers . combined with high trap densities , this can result in a large resistance between the gold electrodes and the transistor channel .
if the contact resistance is dominated by the film resistance , a reduction in the density of hole traps is therefore expected to lead to a reduced contact resistance .
sketch of the different contributions to the total contact resistance .
the holes must overcome a barrier associated to the gold / pentacene interface .
the resistance of the 50 nm thick pentacene film adds to the total contact resistance both at the source and at the drain . ]
pentacene thin - film transistors were made by thermal evaporation , employing a high precision in situ mask alignment mechanism .
the devices were characterized electrically by gated four - terminal measurements without breaking the high vacuum ( base pressure of order 10@xmath0 mbar ) . under the high vacuum conditions ,
the device performance is found to improve with time . within approximately one week
, the contact - corrected field - effect mobility improves by a factor of up to two and the device contact resistance typically decreases by more than an order of magnitude . in addition , the current hysteresis reduces significantly .
we have shown that an increased partial pressure of oxygen or nitrogen does not accelerate the performance improvement .
on the contrary , the gas exposure delays the decrease in contact resistance .
nitrogen was found to dope pentacene thin films at partial pressures as low as 10@xmath139 mbar .
annealing at a slightly elevated temperature ( e.g. 320k ) , on the contrary , leads to an acceleration of the performance improvement . some defects within the pentacene `` anneal '' even at room temperature .
this is a peculiarity of the physics of organic semiconductors , which is governed by the weak van der waals type interaction between the molecules .
we have derived the spectral density of trap states from the `` four - terminal conductivity '' .
the calculations show shallow traps @xmath2ev from the valence band edge to be significantly affected by the defect healing .
we suggest these traps to originate from structural point defects , i.e. slightly misaligned molecules within the grains of the polycrystalline film .
the effective field - effect mobility critically depends on the number of these shallow traps and a relatively small reduction results in a significant improvement of the mobility .
the contact resistance is likely to be dominated by the film resistance and also depends on the active traps within the film .
the method to calculate the spectral density of traps is a powerful tool to further elucidate the origin of trap states in organic semiconductors , provided that contact effects are properly taken into account .
it is particularly suitable to study metastable defects in organic semiconductors , because the density of states function can be derived from a single transfer characteristic in an unambiguous fashion and with a minimal set of simplifying assumptions . | we report on a healing of defects at room temperature in the organic semiconductor pentacene .
this peculiar effect is a direct consequence of the weak intermolecular interaction which is characteristic of organic semiconductors .
pentacene thin - film transistors were fabricated and characterized by in situ gated four - terminal measurements . under high vacuum conditions ( base pressure of order 10@xmath0 mbar ) ,
the device performance is found to improve with time .
the effective field - effect mobility increases by as much as a factor of two and mobilities up to 0.45@xmath1/vs were achieved .
in addition , the contact resistance decreases by more than an order of magnitude and there is a significant reduction in current hysteresis . oxygen / nitrogen exposure and
annealing experiments show the improvement of the electronic parameters to be driven by a thermally promoted process and not by chemical doping . in order to extract the spectral density of trap states from the transistor characteristics ,
we have implemented a powerful scheme which allows for a calculation of the trap densities with high accuracy in a straightforward fashion .
we show the performance improvement to be due to a reduction in the density of shallow traps @xmath2ev from the valence band edge , while the energetically deeper traps are essentially unaffected .
this work contributes to an understanding of the shallow traps in organic semiconductors and identifies structural point defects within the grains of the polycrystalline thin films as a major cause . |
ogorodnikov ( 1958 ) and vorontsov - velyaminov ( 1967 ) were among the first to draw attention to a subset of edge - on spiral galaxies with disk axial ratios as extreme as 30:1 and having little or no discernable bulge component .
goad & roberts ( 1981 ) dubbed these galaxies `` superthins '' and recognized that spiral galaxies selected on the basis of their thin stellar disks and extreme axial ratios also tend to possess a variety of other intriguing properties .
for example , goad & roberts ( 1981 ) discovered that superthins have low emission - line intensity ratios ( [ ] /h@xmath2 and [ ] / [ ] ) , which in a modern perspective suggests that superthins are moderate ionization , low metallicity galaxies more like irregulars than spirals in terms of their region properties ( see also bergvall & rnnback 1995 ) .
similar trends were confirmed for larger samples by karachentsev & xu ( 1991 ) and karachentsev ( 1991 ) . in spite of having `` classic '' double - horned global profiles ( e.g. , matthews & van driel 1999 ) , often superthins exhibit slowly - rising , solid - body rotation curves throughout most or all of their stellar disks ( goad & roberts 1981 ; karachentsev & xu 1991 ; karachentsev 1991 ; cox _ et al .
_ 1996 ; makarov 1997,1998 ; abe _ et al .
such rotation curves are generally characteristic of late - type dwarf and irregular galaxies rather than normal spirals ( cf . casertano & van gorkom 1991 ) .
not only are superthins fascinating objects in their own right , but the very nature of these galaxies makes them extremely valuable objects for exploring a variety of fundamental astrophysical problems , ranging from disk formation and evolution scenarios , to the interpretation of high - redshift galaxy populations , to constraining the stability and dark matter contents of disks ( e.g. , karachentsev 1989 ; zasov 1991 ; dalcanton & schectman 1996 ) . in spite of this ,
to date only a handful of detailed studies of individual superthins have been undertaken ( e.g. , gallagher & hudson 1976 ; bergvall & rnnback 1995 ; cox 1996 ; matthews 1998 ; abe _ et al .
_ 1999 ) .
measurements of the vertical structure and disk color gradients are of particular interest in the case of superthin spirals , as these can provide important clues to their evolutionary history .
vertical structure measurements provide vital constraints on disk stability , dynamical heating mechanisms , and dark matter contents ( matthews 1998 ; matthews 1999 , hereafter paper ii ) , and when combined with measures of global properties and disk color gradients , information on the star - formation histories and dynamical evolutionary histories of the galaxies can be gleaned .
ugc 7321 is a superthin galaxy ideally suited to such investigations . since it is nearby ( d@xmath710 mpc ; matthews _ et al . _
1999b ) , its structure can be reasonably resolved from the ground .
in addition , ugc 7321 appears to suffer little from internal extinction ( see below ) , and it is nearly exactly edge - on ( @xmath588@xmath6 ) , hence projection effects will have a minimal influence on derived vertical structural parameters ( e.g. , de grijs , peletier , & van der kruit 1997 ) .
finally , the analysis of ugc 7321 is simplified compared with other edge - on spirals due to the absence of a bulge component .
here we present measurements of the global properties of ugc 7321 based on new multiwavelength observations including @xmath0- and @xmath1-band imaging and photometry , narrow - band h@xmath2 imaging , near - infrared @xmath3-band imaging , and 21-cm pencil beam spectroscopy .
we also examine aperture synthesis data obtained by rots , roberts , & goad ( private communication ) .
we use our datasets to measure the radial and vertical color gradients in the disk of ugc 7321 ( sect . 3 ) and
briefly discuss what the color gradients reveal about the evolutionary history of this galaxy .
in a second paper ( paper ii ) , we present measurements of the vertical scale height of the ugc 7321 disk .
there we combine vertical structural measurements with results from the present work in order to further constrain the stability and dynamical history of ugc 7321 .
we obtained near - infrared ( nir ) @xmath3-band imaging observations of ugc 7321 in may 1997 using irim on the 2.1-m telescope at the kitt peak national observatory .
irim employs a 256@xmath19256 hgcdte nicmos3 detector with @xmath20 pixels , yielding a field - of - view of @xmath21 per side .
the detector has a gain of 10.5 @xmath22/adu and a readnoise of @xmath737 @xmath22 per pixel rms .
the observations were taken in the presence of moderate cirrus , hence our @xmath3-band data can not be reliably photometrically calibrated .
we obtained sets of 9 co - added , 20-@xmath23 exposures of ugc 7321 at each of 35 telescope pointings . to insure accurate flatfielding and sky subtraction , we shifted the galaxy to a different position on the detector before each series of exposures .
in addition , interspersed throughout the target observations , we obtained several series of `` blank '' sky exposures at regions adjacent to the galaxy .
linearization corrections to the data were made automatically during data acquisition .
data reduction was accomplished using standard iraf tasks . to remove the sky background , a median was made of several groups of temporally consecutive object and blank sky exposures . from these , a median - averaged combination of 25 dark frames was subtracted to remove the structure present in the detector dark current .
finally , the dark - corrected sky frame was subtracted from all dark - corrected object frames .
flatfielding was accomplished by dividing by a night sky flat composed of the median of all dark - corrected sky and object exposures .
lastly , the flatfielded object exposures were aligned and co - added .
our resulting @xmath3-band image of ugc 7321 is shown in fig . 1 .
note the striking thinness of the stellar disk and the complete absence of a bulge component even in the nir . at the disk center ,
the vertical exponential scale height is only @xmath7140 pc ( paper ii ) ! fig .
2 shows the @xmath3-band data in the form of a contour map . because the disk is so thin , the contours
have been stretched in the @xmath24-direction for clarity .
note that the contours are quite regular , even near the galaxy center , suggesting that there is not significant dust absorption in the @xmath3-band . in fig .
3 we plot a major axis profile extracted from the @xmath3-band data . from this profile
, we see that the disk brightness of ugc 7321 falls off somewhat more steeply on the east side ( the righthand side of fig .
3 ) than on the west side .
it is unlikely this can be attributed to internal extinction effects ( which should be small at nir wavelengths ) , or to flatfield errors ( which are quite small in our @xmath3-band images ) , hence this suggests a slight lopsidedness in the stellar disk of ugc 7321 .
arrows overplotted on fig .
3 denote `` zones '' between which the slope of the brightness profile change fairly abruptly .
these zones can also be seem from visual inspection of both our nir and the optical images and are an indication that the disk surface brightness of ugc 7321 deviates somewhat from a smooth , exponential distribution ( see also sect .
this makes it difficult to characterize the @xmath3-band major axis profile in terms of a single exponential scale length . as shown by matthews & gallagher ( 1997 ) ,
this phenomenon is common among extreme late - type spiral galaxies , although the effect is generally lost if the light profile is azimuthally averaged .
matthews & gallagher ( 1997 ) suggested these surface brightness zones may be a signature of limited dynamical evolution in some of the smallest and faintest spirals ( see also sect .
2.3.3 ) . we obtained photometrically calibrated @xmath0- and @xmath1-band ccd imaging observations of ugc 7321 in june 1997 using the 3.5-m wiyn telescope at kitt peak .
in addition , we obtained narrow - band h@xmath2 and uncalibrated @xmath0-and @xmath1-band observations of the galaxy with the wiyn telescope in april 1997 .
the direct imaging camera on the wiyn telescope employs a thinned stis 2048@xmath192048 ccd with @xmath25 pixels , yielding a field - of - view of @xmath77@xmath26 per side .
the ccd gain was 2.8 @xmath22 per adu , and readnoise was @xmath78 @xmath22 per pixel rms .
exposure times were 900@xmath23 in @xmath0 , 600@xmath23 in @xmath1 , and 1500@xmath23 in h@xmath2 .
seeing during both the april and june observations was @xmath7@xmath27 fwhm .
the wiyn telescope takes advantage of the best seeing on kitt peak through a combination of thermal controls , active optics , and dome ventilation systems .
this , combined with the small pixel scale of the wiyn ccd imaging camera , makes wiyn an ideal facility for obtaining the data necessary for detailed studies of disk structure and color gradients of galaxies .
our ccd images were reduced using standard iraf tasks .
individual frames were overscan corrected and bias subtracted in the usual manner .
flatfielding was accomplished by dividing by a mean of 5 dome flats taken in the appropriate filter .
the wiyn ccd contains a bad column which was corrected in our images by interpolation from the two adjacent columns . both the april and the june observations were obtained during dark time .
photometric calibration for the june run was accomplished by monitoring two standard star fields from landolt ( 1992 ) at two different air masses and then applying standard color term corrections to convert from wiyn natural magnitude system to the landolt system .
the total formal dispersion in our wiyn magnitude zero points is @xmath280.03 magnitudes ; the validity of our transformations was checked via comparisons with other recent wiyn photometric solutions .
errors in the colors are approximately @xmath280.02 magnitudes .
sky brightnesses during our observations were @xmath2922.1 mag arcsec@xmath8 and @xmath3019.8 mag arcsec@xmath8 . our @xmath1-band image of ugc 7321
is shown in fig . 4 & 5 .
note the very diffuse , low surface brightness ( lsb ) appearance of the disk in spite of its edge - on geometry .
several background galaxies are clearly visible through the disk ( e.g. , fig .
6 ) , implying internal extinction is low . our continuum - subtracted h@xmath2 image of ugc 7321
is shown in fig .
7 . due to cirrus
, we were unable to obtain a flux calibration for this frame .
however , qualitatively , this image shows that h@xmath2 emission is present out to radii of at least @xmath31@xmath32 ( where @xmath33 is the distance from the galaxy center measured along the disk major axis in the plane of the sky ) .
we see in fig .
7 that the bulk of the h@xmath2 emission in ugc 7321 is confined to a region near the disk midplane roughly 4@xmath34 thick , although diffuse emission is clearly visible at @xmath24-heights of up to @xmath35 at galactocentric radii @xmath36@xmath37 .
we note however that there is a relative dearth of emission within a region offset 8@xmath34 to the east of the disk center ( fig .
this `` gap '' is @xmath38 wide and contains faint plumes of emission extending to @xmath397@xmath34 out of the galaxy plane .
we also see marginal evidence of several other thin filaments which are more extended in @xmath24-height .
deeper narrow - band images are clearly needed to establish whether the central h@xmath2 feature could be a signature of blowout from the disk , and whether ugc
7321 may indeed contain a diffuse , ionized medium extended to large @xmath24-heights , analogous to that seen in more luminous edge - on spirals ( cf .
dettmar 1995 and references therein ) .
we performed surface photometry on ugc 7321 using routines from the iraf stsdas analysis package . with the `` ellipse '' program ,
we fit a series of 9 concentric ellipses to the ugc 7321 images . position angle and ellipticity of the ellipses
were determined from the outermost galaxy isophotes and were kept fixed throughout the fitting .
the galaxy center was chosen to be the position of peak brightness ; this location was identical in the @xmath0 and @xmath1 ( and @xmath3 ) frames .
foreground stars , cosmic rays , and background galaxies within the aperture were removed via background interpolation using the iraf task `` imedit '' .
sky values were determined by measuring the sky counts in each of several rectangular apertures at various locations on the image ( see matthews & gallagher 1997 ) . because ugc 7321 is so close to edge - on , elliptical apertures are not ideal fits to the galaxy isophotes .
nonetheless , these apertures allow us to measure aperture magnitudes and colors for ugc 7321 , and permit an estimate of the mean exponential scale length of its disk .
although the ellipticity of the fitted isophotes is an unreliable method of determining the inclination of near edge - on galaxies , our derived value ( @xmath40 ) is consistent with the inclination we estimate from the slight asymmetry of the dust features along the galaxy major axis .
our derived photometric parameters for ugc 7321 are given in table 1 .
errors for the aperture magnitude measurements were computed following matthews & gallagher ( 1997 ) and take into account sky , flatfield and poisson errors , as well as the scatter in the photometric solution .
large - scale flatfield errors are the dominant source of uncertainty ; the maximum amplitude of the flatfield variations was @xmath71% of sky in both @xmath0 and @xmath1 .
our aperture magnitudes are in good agreement with previous photoelectric values reported by tully ( 1988 ; @xmath41=13.86 ) and de vaucouleurs & longo ( 1988 ; @xmath42=12.99 ) .
9 shows our azimuthally averaged @xmath0-band radial surface brightness profile of ugc 7321 . although we noted in sect .
2.1 that the disk of ugc 7321 is slightly lopsided and shows deviations from a perfect exponential , an azimuthally - averaged brightness profile permits determination of a `` mean '' scale length which is useful for offering a global characterization of the size scale of the disk . because ugc 7321 is viewed close to edge - on
, one must take into account projection effects in analyzing the projected light profile and in deriving an exponential scale length .
for a disk with an exponential radial brightness distribution viewed exactly edge - on ( @xmath43 ) , the projected radial brightness profile along the galaxy midplane ( @xmath24=0 ) is expressed as @xmath44 where @xmath45 is the modified bessel function of first order ( e.g. , van der kruit & searle 1981 ) . at small radii ( @xmath46 ) , this expression can be approximated as @xmath47.\ ] ] note this implies a slight flattening of the light profile will be observed at small radii compared with a simple , unprojected @xmath48 function . at large radii ( @xmath49 ) , one can write @xmath50 ( see van der kruit & searle 1981 ) .
because of the @xmath51 term , this produces a slightly less steep light profile than a pure exponential function with the same scale length . as a result ,
a scale length derived from fitting a simple , unprojected exponential profile to an edge - on disk will be overestimated . at @xmath52-@xmath53
this effect is @xmath710% ( e.g. , van der kruit & searle 1981 ) . to derive the scale length for ugc 7321 ,
we have fitted a model projected exponential light profile to our data .
our model is overplotted on fig .
9 . for the purpose of deriving a scale length ,
the slight deviation of ugc 7321 from the edge - on ( @xmath43 ) case is not significant . from our fit
we derive a scale length of @xmath54=44@xmath55 ( compared with @xmath54=51@xmath56 we derive from simply fitting the function @xmath57 to the disk at intermediate @xmath33 ) . a comparison between the data and our model ( fig .
9 ) shows that in spite of lacking a bulge component , ugc 7321 exhibits a light excess at small radii compared with the prediction of an exponential disk .
the @xmath0-band _ measured _ central surface brightness of ugc 7321 ( before correction to a face - on value ) is 21.6 mag arcsec@xmath8 , while that predicted from extrapolation of the exponential fit to small @xmath33 is @xmath70.35 magnitudes fainter .
interestingly , this excess is even more pronounced in the @xmath1-band and appears to correspond to a distinct region in the color map of ugc 7321 ( see sect . 3 ) .
this strengthens the suggestion that ugc 7321 appears to have a multi - component disk that is more complex than a simple , pure exponential ( see also sect .
in addition , we note that at @xmath58 our observed light profile falls off faster than the projected exponential model , suggesting the stellar disk of ugc 7321 may be truncated ( cf .
barteldrees & dettmar 1994 ) .
however , this latter trend should be confirmed with deeper observations on a wide - field ccd where sky subtraction and flatfielding can be accomplished with greater accuracy . for an optically thin disk which is exponential in both the @xmath33 and @xmath24 directions , and which is observed inclined at 90@xmath6 , the face - on central surface brightness will be @xmath59 times the observed edge - on value . here
@xmath60 is the exponential scale height ( cf .
, van der kruit & searle 1981 ) .
transformations from observed surface brightness values to face - on values at other radii may be derived from equations 1 - 3 above . using a thin - disk approximation ( see the appendix ) we have derived the additional corrections to these values required for the case where a is disk observed at an inclination slightly less than 90@xmath6 .
after also taking into account internal extinction corrections ( see sect .
3.2.2 ) , we then find that if projected to face - on , the disk of ugc 7321 would have an _ observed _ @xmath0-band central surface brightness of @xmath723.4 mag arcsec@xmath8 , an extrapolated central surface brightness of @xmath723.8 mag arcsec@xmath8 , and a mean total disk surface brightness @xmath6127.6 mag arcsec@xmath8 .
thus ugc 7321 is a very lsb galaxy , and much of its disk would likely be nearly invisible if viewed closer to face - on . as shown below , the internal extinction in ugc 7321 appears to be quite low , thus its lsb appearance must result from a rather low current star formation rate .
further evidence that the seemingly `` anemic '' nature of ugc 7321 is due to minimal current star formation comes from comparing its blue luminosity to its far - infrared luminosity . using the _ iras _ 60@xmath62 m and 100@xmath62 m fluxes for ugc 7321 from the ned database and sage ( 1993 ) , respectively , and following the prescription of rice ( 1988 )
, we derive @xmath63=7.8@xmath64 , a value nearly two orders of magnitude fainter than the mean for scd / sd spirals in the ugc catalogue ( roberts & haynes 1994 ) . using our blue luminosity ( corrected for internal extinction ) , we find that @xmath650.08 for ugc 7321 .
a comparison with rice ( 1988 ) reveals that such low @xmath66 ratios are commonly found only in two classes of galaxies : ( 1 ) old , red early - type spirals , de s and s0 s with few young stars ; ( 2 ) gas - rich , extreme late - type spirals with diffuse , extremely lsb disks .
examples of the latter group of objects include the nearby galaxies ngc 4395 ( sd
iv ) , ngc 45 ( sdm iv ) , and ic 2574 ( sm iv - v ) .
all three of these galaxies have @xmath670.10 , @xmath16@xmath681 , blue optical colors , and highly transparent disks ( e.g. , sandage 1961 ; matthews 1999a ) .
using our @xmath0- and @xmath1-band wiyn data , we have produced a @xmath9 color map of ugc 7321 ( fig .
a variety of structure is evident in this map , including significant @xmath9 color gradients in both the vertical and radial directions .
we discuss these features in detail below .
the global @xmath9 color of ugc 7321 is 0.99 , as measured within the outermost observed isophote ( 25.2 mag arcsec@xmath8 in @xmath0 before correction for inclination ) , and after correction for galactic foreground extinction .
this @xmath9 color is a typical value for a normal , late - type spiral ( cf . lauberts & valentijn 1988 ; de blok , van der hulst , & bothun 1995 ) .
this suggests that in spite of its edge - on geometry , the disk of ugc 7321 does not suffer severely from internal reddening due to dust .
we return to this issue below . near the center of the ugc 7321 disk ,
our color map reveals a small , very red ( @xmath691.5 ) nuclear feature , only a few arcseconds across ( fig .
this feature is offset @xmath75@xmath34 to the east of the disk center as determined from the brightness peak of the @xmath1- and @xmath0-band images .
one possibility is that we are seeing the signature of an embedded nucleus .
however , although compact star cluster nuclei are common in late - type low - luminosity spirals with diffuse , lsb disks ( matthews & gallagher 1997 ; matthews 1999a ) , we do not see any direct evidence of such a feature in ugc 7321 in either our wiyn images or in images obtained with the _ hubble space telescope _ ( matthews 1999b ) .
nonetheless , the location of the compact red feature in the @xmath9 color map is near to that of the peculiar h@xmath2 emission features shown in fig . 8 .
the optical longslit spectrum of ugc 7321 obtained by goad & roberts ( 1981 ) also shows a possible kinematic disturbance in the h@xmath2 emission at this location .
these are hints that some interesting physical processes are at work near the center of the ugc 7321 disk .
further investigation of this region via high resolution spectroscopy , deep h@xmath2 imaging , and nir colors may be fruitful .
surrounding the compact red feature , our color maps reveal a more extended red region ( @xmath701.2 ) , visible to @xmath71 on either side of the disk center and showing a rather abrupt boundary ( see also below ) .
intriguingly , the extent of this region corresponds very closely to the region over which we observe a light excess over a pure exponential disk fit ( sect .
this raises the possibility that this red central region might possibly represent an ancient central starburst , the core of the original protogalaxy , or perhaps even a kinematically distinct disk subsystem analogous to the bulge of normal spirals .
cutting into the red central region discussed above are thin blue bands of stars visible along the midplane of the galaxy .
these bands grow both thicker and bluer with increasing distance from the galaxy center . at @xmath7220@xmath34 ,
this layer has @xmath691.05 , reaching @xmath690.85 at @xmath73=@xmath37 , @xmath690.80 at @xmath72@xmath74 , @xmath690.45@xmath75 at @xmath33=@xmath76 , and @[email protected] at @xmath77@xmath76 .
thus , from the nuclear region ( where @xmath78 ) to the outer disk edge we see a total @xmath9 color change of up to 1.05 magnitudes along the major axis of ugc 7321 .
this is illustrated in fig .
11@xmath79 , which shows @xmath9 color profiles of ugc 7321 extracted along the major axis , as well as @xmath80 north ( through the red nuclear feature ) , @xmath81 south , and @xmath82 south relative to the major axis , respectively .
the extracted profiles were averaged over 12-pixel - wide strips and then smoothed , hence the red color of the compact red central disk feature is slightly subdued in the profiles shown in fig .
because we are interested in accurately measuring colors in the faintest regions of the outer disk , we extracted our color profiles from our our uncalibrated @xmath0 and @xmath1 images , which have slightly superior flatfields to our photometrically - calibrated data ; maximum amplitude variations of the flatfields are @xmath71% of the sky in @xmath0 and @xmath70.5% of the sky in @xmath1 .
we then used comparisons between high signal - to - noise inner disk regions to calibrate the absolute colors from the photometrically - calibrated images .
the maximum uncertainties expected from the combination of sky subtraction and flatfield errors are overplotted as dotted lines on fig .
although signal - to - noise in the outer disk is low , the profiles extracted at all four positions parallel to the major axis show regions with @xmath830.60 , suggesting these blue outer disk colors are real . finally , we draw attention to the presence of a faint , thicker , but highly flattened disk of unresolved stars visible in our color map surrounding the ugc 7321 disk at @xmath36@xmath84 .
this component has @xmath691.1 and shows little change in color with galactocentric distance ( fig .
12 ) . based upon the observations of galactic globular clusters ( see secker 1995 ) ,
old , metal - poor stellar populations are expected to have @xmath130.8 , hence the highest @xmath24-height stars in ugc 7321 appear to represent an `` old disk '' population ( see also below ) .
bluing as a function of increasing galactocentric distance is a well - known feature of spiral galaxy disks ( e.g. , de jong 1996 ) .
de jong ( 1996 ) has demonstrated through monte carlo simulations that in face - on spirals , this observed radial bluing generally can not be accounted for by dust .
he shows that for realistic models , dust creates color gradients of less than 0.3 magnitudes in @xmath9 , and argues instead that the observed radial color changes in spirals are indicative of star formation progressing radially outwards in disks over time , leading to stellar age and metallicity gradients .
this picture is consistent with a number of semi - analytic galaxy formation models where galaxy disks are built `` from the inside out '' ( e.g. , white & frenk 1991 ; mo 1998 ) .
nonetheless , the radial color gradient we observe in ugc 7321 ( @xmath85 ) is significantly larger than typical gradients in the de jong sample ( where typically @xmath860.6 magnitudes , even when bulge light is included ) .
moreover , in most spirals observed edge - on , radial color gradients are generally found to be small ( e.g. , sasaki 1987 ; aoki 1991 ; wainscoat , freeman , & hyland 1989 ; bergvall & rnnback 1995 ) or negligible ( e.g. , hamabe 1980 ; jensen & thuan 1982 ; de grijs 1998 ) . unfortunately , in edge - on spiral galaxies , color gradients become more difficult to interpret physically , particularly since internal extinction can be quite significant near the galactic plane .
the net effect can be both an alteration in the galaxy luminosity profile , and a reddening of the observed optical colors which can vary as a function of @xmath33 and @xmath24 . therefore in order to determine what fraction of the color gradients we observe in ugc 7321 are due to true age and/or metallicity gradients , we must include an assessment of the role of dust in this galaxy . in practice , accurately quantifying the effects of dust on the observed colors and luminosity distribution in a galaxy is a complicated problem ( e.g. , de jong 1996 ; xilouris 1999 )
. moreover , standard assumptions about the nature and distribution of dust in normal giant spirals are unlikely to be applicable to low - metallicity lsb galaxies ( cf .
han 1992 ) . as a result
, the effects of internal extinction in a given galaxy are ideally derived using a combination of sophisticated radiative transfer models and empirical measurements ( e.g. , xilouris 1999 ) . such modelling beyond the scope of the present work , so here we derive some fundamental constraints on the effects of dust in ugc 7321 using a very simple model .
although some authors have argued that the internal extinction in lsb galaxies may be almost negligible ( e.g. , mcgaugh 1994 ; tully 1998 ) , these claims have only rarely been tested in edge - on systems ( e.g. , goad & roberts 1981 ; kodaira & yamashita 1996 ; bergvall & rnnback 1995 ; karachentsev 1999 ) .
we have already mentioned several lines of evidence that the internal extinction in ugc 7321 is low ( regularity of the @xmath3-band isophotes ; visibility of background galaxies through the disk ; low @xmath66 ratio ; correspondance between the galaxy center in @xmath3- and @xmath0-band ) .
nonetheless , visible inspection of our wiyn images shows the galaxy is clearly not completely devoid of dust , and a number of individual clumps ( possibly molecule - rich dark clouds ) can be seen in our optical images ( e.g. , fig . 5 ; see also matthews 1999b ) .
in spite of this , the fact that we resolve many of these individual dust clumps instead of seeing a uniform dust lane implies that the disk of ugc 7321 is not optically thick .
another test of the optical thickness of the ugc 7321 disk comes from a comparison between its optical rotation curve ( derived from longslit spectroscopic measurements of h@xmath2 emission from regions ) , with a rotation curve derived from measurements . from optical longslit measurements ,
goad & roberts ( 1981 ) found ugc 7321 to have a slowly rising , solid - body rotation curve throughout much of its stellar disk .
if a galaxy disk is optically thick , the rotation curve may appear solid body regardless of its intrinsic shape , as an artifact of one s inability to observe regions at small galactocentric radii ( goad & roberts 1981 ; byun 1993 ) .
however , a rotation curve of ugc 7321 derived from aperture synthesis measurements ( sect .
4.2 ) confirms that the slow rise of the rotation curve throughout the stellar disk is indeed intrinsic ( see fig .
19 , discussed below ) .
once again we conclude the disk of ugc 7321 is not optically thick .
dust affects the color and luminosity profiles of galaxies through both the scattering and absorption of optical photons .
however , the maximum color change ( i.e. , maximum reddening ) occurs in the pure extinction case .
in addition , byun , freeman , & kylafis ( 1994 ) have shown that scattering effects become decreasingly important with increasing inclination ( see also bianchi , ferrara , & giovanardi 1996 ) .
since here we are primarily interested in estimating the effects of dust on the color gradient , and since we are in the optically thin regime , we can simplify the problem by considering a model of extinction due to a foreground dust screen ( see disney , davies , & phillips 1989 ) .
for a given amount of dust , the screen model produces more reddening than a more realistic model of stars mixed with dust , hence it establishes an upper limit to @xmath87 as a function of radius . to evaluate @xmath87 , we begin by attempting to attribute as much of our observed color gradient as possible to dust .
we further assume that , to first order , dust extinction in the @xmath3-band is negligible ( e.g. , fig . 2 ) .
in the case of zero intrinsic radial color gradient , a dust - free version of ugc 7321 would exhibit no radial color gradient in @xmath88 , hence we can use the changes we do observe in the @xmath88 color to estimate the amount of extinction at a given radius . we do not attempt to reproduce the clumpy nature of the dust in our model ( which again , would decrease any reddening effects ) , but assume the dust distribution is roughly exponential in the radial and vertical directions ( e.g. , wainscoat , freeman , & hyland 1989 ; xilouris 1999 ) . in this case
the extinction along an interval from @xmath33 to @xmath89 can be expressed as @xmath90 where @xmath91 is the absorption in magnitudes per unit length at a given wavelength , @xmath92 corresponds to the edge of the stellar disk , @xmath93 is the scale length of the dust , and @xmath94 is the scale height of the dust ( see also wainscoat , hyland , & freeman 1989 ) .
13 shows a plot of the `` pseudo '' @xmath88 color profile of ugc 7321 along the major axis interval @xmath33=0120@xmath34 . here
the @xmath1-band data were smoothed to match the resolution of the @xmath3-band observations . because our @xmath3-band data are not photometrically - calibrated , we have chosen an arbitrary normalization such that our `` pseudo '' @xmath88 color is 0 near @xmath33=80@xmath34 . in fig .
13 , we see that near @xmath33=0 , the @xmath1-band light is depressed by roughly 40% relative to the @xmath3-band , corresponding to an @xmath1-band extinction @xmath950.55 magnitudes per kiloparsec .
assuming the extinction law of bouchet ( 1985 ) , we derive @xmath96=0.89 magnitudes per kiloparsec . from fig .
13 we estimate a scale length of the dust distribution of @xmath9740@xmath34 , which appears consistent with the observed dust distribution in our images .
although in giant spirals , it is often assumed the dust has a similar scale length to the old stellar disk ( e.g. , kylafis & bahcall 1987 ) , or even extends beyond it ( e.g. , xilouris 1999 ) , this does not appear to hold in ugc 7321 .
we see an abrupt end to the presence of resolved dust clumps beyond radii @xmath98 in both our wiyn images , and in _ hubble space telescope _
images obtained by matthews ( 1999b ) . moreover , background galaxies are readily visible through the ugc 7321 disk beyond these radii ( matthews 1999b ) , implying extinction has dropped appreciably . for the vertical scale height of the dust ,
we adopt @xmath94=@xmath99 , which is half the scale height of the old stellar disk ( sect .
3.3 ; see also paper ii )
. using equation 4 , we can now derive an extinction - corrected @xmath9 color profile for ugc 7321 along its major axis ( fig 14@xmath100 ) and offset @xmath80 to the north ( fig 14@xmath101 ) .
we see that even the extinction - corrected radial color gradient remains pronounced ( @xmath102 mag ) .
this demonstrates that _ a significant fraction of the observed radial color gradient in ugc 7321 appears to be due to stellar population and/or metallicity gradients .
_ the presence of the observed radial disk color gradients in ugc 7321 has several implications .
regions with @xmath103 are too blue to be accounted for by old , metal - poor stellar populations ( e.g. , secker 1995 ) , but instead must result from a mixture of ages . without resolving these populations
, we can not uniquely constrain the mix of stars present . however , the bluest colors we reliably measure in ugc 7321 ( @xmath104 ) are comparable , for example , to young regions in the moderate metallicity giant magellanic galaxy ngc 4449 ( bothun 1986 ) and to the predicted colors of the younger outer regions of a smoothly evolving galaxy with a high halo spin parameter and an age of @xmath710 gyr , as in the models of jimenez ( 1998 ) . in either case , it is likely that youthful stars are a significant contributor to the blue colors of the outer disk of ugc 7321 ( see also bell 1999 ) .
this assertion is strengthened by the presence of h@xmath2 emission in the outer disk ( fig .
7 ) and the existence of resolved supergiant stars at these radii ( matthews 1999b ) .
we note that edge - on , superthin lsb galaxies like ugc 7321 are especially valuable for studies of the outer regions of disks , as they may contain some of the most pristine , unevolved galaxian disk environments observable in the nearby universe .
however , because these disk regions are so optically thin , they would become very difficult to detect in galaxies viewed at lower inclinations .
for example , if viewed near face - on , the portions of the ugc 7321 disk where @xmath830.6 would have an observed @xmath0-band surface brightnesses of only @xmath10527.6 mag arcsec@xmath8 .
such faintness levels are only rarely reliably achieved in galaxy imaging surveys due to combinations of short integration times , scattered light , limited detector fields - of - view , and flatfield uncertainties ( cf .
morrison , boronson , & harding 1994 ; lequeux 1996 ) . in spite of the potentially very young outer disk regions of ugc 7321 , its moderately red central disk ( where @xmath691.2 ) and its somewhat thicker `` old disk '' ( where @xmath691.1 ) both indicate that ugc 7321 is likely to contain stellar populations with ages in excess of 10 gyr ( see jimenez 1998 ) . as demonstrated above , accounting for a reasonable amount of reddening due to dust does not affect this conclusion .
the red stellar components in ugc 7321 imply that this is not a young galaxy presently undergoing its first epoch of star formation .
our observed colors and color gradients in ugc 7321 are similar to those found in other lsb spirals by de blok , van der hulst , & bothun ( 1995 ; see also impey & bothun 1997 ) .
a consistent model for ugc 7321 is then that of a disk galaxy which has made stars slowly and is now seen to be under - evolved relative to typical giant spirals ( see also oneil 1997 ; jimenez 1998 ) .
this would account for the low gas metallicity of ugc 7321 ( goad & roberts 1981 ) , the large mass fraction ( sect . 4 ) , and the low density of stars ( i.e. , the low optical surface brightness disk ) .
thus while lsb disks like ugc 7321 may be _ unevolved _ galaxies , we concur with the suggestion of jimenez _ et al . _
( 1998 ) that _ lsb galaxies are not necessarily young . _ in order to preserve its relatively strong radial color gradients over many gyr , it appears that viscous evolution in the disk of ugc 7321 has been minimal . in giant spirals , viscous evolution ( which redistributes angular momentum in the disk )
is often argued to be responsible for the relatively smooth exponential luminosity profiles of most disk galaxies ( e.g. , lin & pringle 1987 ) .
this process is also expected to partially smooth color gradients due to stellar age and metallicity changes with radius ( e.g. , firmani , hernndez , & gallagher 1996 ) .
however , such a mechanism may not be able to work efficiently in galaxies like ugc 7321 , due to both the low disk surface density ( which considerably extends dynamical timescales ) and the slowly rising rotation curve [ which produces limited shear in the inner disk regions ( lin & pringle 1987 ; see also matthews & gallagher 1997 ) ] .
the gradients in the @xmath3-band major - axis profile of the ugc 7321 disk ( sect .
2.1 ) and the deviation of the azimuthally - averaged profile from a pure exponential disk at small @xmath33 ( sect .
2.2.3 ) may be additional signs that viscous evolution has been limited in this galaxy compared with normal spirals .
finally , we note that at intermediate galactic radii , the eastern side of ugc 7321 has a slightly bluer mean color than the western side .
this result is insensitive to reasonable sky subtraction and flatfield errors .
one explanation could be that we are seeing the projection of a spiral arm(s ) , although it remains uncertain how much spiral structure a thin disk like ugc 7321 might actually sustain , since spiral structure is believed to be an efficient disk heating mechanism ( e.g. , lacey 1991 and references therein ) .
moreover , at least some self - gravity is needed for spiral arm formation , but if the disk of ugc 7321 were completely self - gravitating , it would be highly unstable to `` firehose '' instabilities and hence to vertical thickening ( zasov , makarov , & mikailova 1991 ; matthews 1998 ) .
an alternative is that there are asymmetries in the stellar distribution of ugc 7321 due to patchy star - formation ( see also gerritsen & de blok 1999 ) .
the presence of vertical color gradients in galaxies offers an important clue toward the dynamical evolution of galaxian disks , since such gradients are predicted to occur as a consequence of dynamical heating processes ( e.g. , just , fuchs , & wielen 1996 ) .
unfortunately , vertical color gradients are very difficult to measure in most edge - on spirals due to dust , contamination from the bulge component , and the effects of atmospheric seeing . to date , only a few analyses of edge - on spirals have uncovered non - negligible vertical color gradients ( e.g. , wainscoat , freeman , & hyland 1989 ; bergvall & rnnback 1995 ) , while many workers have reported such gradients to be very small or negligible ( e.g. , jensen & thuan 1982 ; de grijs , peletier , & van der kruit 1997 ) .
15@xmath100-@xmath106 illustrates the vertical color profiles of the ugc 7321 disk at various galactocentric distances . at @xmath107
, we see very little @xmath24 color gradient : @xmath701.2 at a range of @xmath24 values ( fig .
15@xmath100 ) , although a slight asymmetry is visible .
at @xmath108 ( fig 15@xmath101 ) , we see the addition of a peak with @xmath701.45 .
note its displacement from the disk midplane ; this peak is in the vicinity of the red nuclear feature discussed in sect .
3.1 . at @xmath33=10@xmath34 ( fig 15@xmath109 ) , the vertical color gradient is again fairly flat ( @xmath701.2 ) , with a very slight bluing visible near @xmath24=@xmath110 . at @xmath111@xmath112 ( fig 15@xmath113 ) ,
the high @xmath24 regions of the disk have nearly the same color as those at @xmath33=0 ( this is due to the `` old disk '' of red stars discussed above ) , but at small @xmath24 values , the disk has become much bluer than at smaller galactocentric radii . finally , with further increasing radius , while the color at high @xmath24 continues to stay nearly constant , the color at small @xmath24 grows increasingly bluer ( fig .
15@xmath114-@xmath106 ) , consistent with the radial bluing observed along the major axis of the galaxy in fig .
vertical color gradients of the type we observe over much of the disk of ugc 7321 ( i.e. redder with increasing @xmath24 height ) are predicted to occur due to dynamical disk evolution in which older stellar populations acquire higher velocity dispersions over time due to heating processes ( e.g. , fuchs & wielen 1987 ; just , fuchs , & wielen 1996 ) .
ugc 7321 represents one of the few examples of this type of vertical color structure being directly observed in a galaxy disk .
thus even the most dynamically cold examples of nearby galaxy disks appear to have undergone some dynamical heating . nonetheless , the exact mechanism by which older stars are dynamically heated and how they are redistributed as a function time remain uncertain ( see review by lacey 1991 ) and different mechanisms may operate in different galaxies , depending upon factors like environment , degree of self - gravity , and the size and number of molecular clouds .
the relative simplicity of ugc 7321 , and its status as a relatively unevolved galaxy make this an ideal system to help place important constraints on these issues .
we explore this further in paper ii . using the extinction parameters derived in sect . 3.2.2 , we can also estimate the effects of dust reddening on our observed vertical color gradients .
we adopt for a scale height of the dust half the scale height of the old stellar disk ( xilouris 1999 ) . using the stellar scale height of the old disk derived in paper ii , this
yields @xmath94=@xmath99 .
16 shows an example of one of our vertical color profiles corrected for dust reddening .
we see that at the galactocentric radii where the vertical color gradients become the strongest , correction for dust reddening acts to slightly increase @xmath11 along the vertical direction .
in june 1997 we used the nanay decimetric radio telescope to obtain a 7-point pencil - beam map of ugc 7321 in the 21-cm line of neutral hydrogen ( ) . at the declination of ugc 7321
, the nanay radio telescope has a fwhm beam size of approximately 4@xmath26e - w@xmath1923@xmath26n - s .
other details regarding the nanaytelescope may be found in , e.g. , matthews , van driel , & gallagher ( 1998 ) .
ugc 7321 was observed at 7 different telescope pointings , including a position corresponding to the optical center of the galaxy , and positions offset 2@xmath26 , 4@xmath26 , and 6@xmath26 east and west of center , respectively .
the observations were obtained in total power ( position - switching ) mode using consecutive pairs of two - minute on- and two - minute off - source integrations .
total integration times were 2 - 3 hours at each pointing . for these observations ,
the autocorrelator was divided into two pairs of cross - polarized receiver banks , each with 512 channels and a 6.4 mhz bandpass .
this yielded a channel spacing of 2.6 , for an effective velocity resolution of @xmath73.3 .
fortuitously , ugc 7321 is oriented along nearly an e - w line , so the nanay telescope provides sufficient spatial resolution along this direction to obtain crude information about the distribution and kinematics of the galaxy .
the individual spectra obtained at each telescope pointing are shown in fig .
17 and the resulting global spectrum in fig . 18 .
note the global profile appears quite symmetric .
table 1 gives our derived global parameters .
errors were computed following matthews , van driel , & gallagher ( 1998 ) .
our values are in good agreement with the recent single - dish observations of ugc 7321 obtained by haynes ( 1998 ) using the green bank 140-ft telescope .
our nanay observations show that the gas in ugc 7321 clearly extends beyond the optical galaxy , confirming the results of hewitt , haynes , & giovanelli ( 1983 ) .
significant flux is detected at the pointings both 4@xmath26e and 4@xmath26w of the galaxy , giving a lower limit for the for -to - optical diameter ratio of @xmath1151.25 .
in addition , we detect a small amount of flux ( @xmath71.1 jy ) in our observation 6@xmath26w of the galaxy center . from the nanay data
alone , it is difficult to assess whether this flux is due to a sidelobe contamination or real extended emission , since the strength of the nanay telescope sidelobes varies significantly with the hour angle of the source ( guibert 1973 ) . after our nanay observations were obtained , we had the opportunity to examine aperture synthesis data of ugc 7321 obtained by rots , roberts , & goad ( private communication ) using the very large array ( vla ) .
these observations were obtained in c array in 1981 using 18 antennas , hence they do not achieve the same sensitivity limits as modern vla data , but they still offer a useful complement to our nanaydataset .
thirty - one independent velocity channels were used for the observations , with a velocity resolution of @xmath710.3 per channel .
the fwhm beamwidth was @xmath712@xmath34 . using the vla data
, we derived a position - velocity diagram for ugc 7321 . by assuming that at each point along the major axis the maximum velocity traces the rotation , we have derived the rotation curve shown in fig . 19 .
this figure confirms the slowly - rising , solid body nature of the rotation of ugc 7321 throughout its stellar disk , as was first seen in the optical rotation curve of goad & roberts ( 1981 ) .
although beam - smearing may be expected to slightly decrease the amplitude of our derived rotational velocities , this effect is not expected to exceed a few kilometers per second , or to be significant beyond 22.5 beam diameters from the disk center ( e.g. , swaters 1999 ) , hence it can not explain the shallow shape of the inferred rotation curve .
finally , we note that fig .
19 suggests that the rotation curve of ugc 7321 does not begin to flatten until near the edge of the stellar disk .
this is consistent with ugc 7321 being a galaxy whose dynamics are dominated by a dark halo even at small galactocentric radii ( e.g. , matthews 1998 ) . however , deeper , more sensitive aperture synthesis measurements would be valuable for further constraining the detailed shape of the outer rotation curve .
the integrated map of ugc 7321 derived by rots , roberts , & goad ( private communication ) confirms a slight extension of the on the west side of the galaxy compared with the east side , consistent with our nanaydata .
it also permitted us to measure an diameter of @xmath116@xmath117 for ugc 7321 at the limiting column density of the observations ( @xmath118 atoms cm@xmath8 ) .
thus @xmath1151.2 , in agreement with our nanay lower limit .
it is somewhat difficult to accurately compare the extent of ugc 7321 with other spirals since the commonly - quoted @xmath120 ratio ( where @xmath121 is measured at a face - on - corrected isophote ) in not particularly meaningful for a galaxy like ugc 7321 , whose face - on central surface brightness is only @xmath122 mag arcsec@xmath8 .
moreover , we can not unambiguously translate our limiting observed column density into an surface density due to projection effects , thus it is difficult to accurately measure @xmath123 at some canonical surface density ( e.g. , 1 pc@xmath8 ) . in terms of optical scale length ,
the in ugc 7321 extends to at least 5.5@xmath18 .
although this value may increase with more sensitive vla observations , such an extent is still quite normal for both high and low surface brightness late - type spirals having similar rotational velocities to ugc 7321 ( see fig . 8 of de blok , mcgaugh , & van der hulst 1996 ) .
if we assume both the stellar and the disks of ugc 7321 are optically thin , and hence both similarly enhanced in surface brightness due to their edge - on projection , then we see that the maximum extent of the disk relative to the maximum optical extent is typical for sd spirals ( e.g. , hewitt , haynes , & giovanelli 1983 ) .
we can more directly compare with two other sd superthins previously mapped in : @xmath1241.2 was also found by cox ( 1996 ) for the superthin spiral ugc 7170 and by abe ( 1999 ) for the superthin ic 5249 , both at similar limiting column densities and optical surface brightnesses . ignoring internal extinction effects , the @xmath16 ratios of ugc 7321 , ugc 7170 , and ic5249 are 2.6 , 3.4 , and 2.3 ( in solar units ) , respectively , and the superthin spiral eso 146 - 014 studied by bergvall & rnnback ( 1995 ) has @[email protected] .
if we assume all have internal extinctions similar to ugc 7321 ( @xmath125 mags ; sect .
3.2.2 ) , the implied @xmath16ratios are still significantly larger than typical values for normal sd spirals derived by roberts & haynes ( 1994 ; @[email protected] ) , although not as extreme as some of the unevolved dwarf galaxies studied by van zee ( 1995 ) or some of the extreme late - type lsb spirals in the sample of matthews & gallagher ( 1997 ) which have @xmath126 as high as 10 ( see also salzer 1991 ; matthews , van driel , & gallagher 1998 ) .
van der hulst ( 1993 ) suggested that low gas surface densities are responsible for the low star - formation efficiencies in lsb galaxies ( see also de blok , mcgaugh , & van der hulst 1996 ; gerritsen & de blok 1999 ) .
although we can not unambiguously recover the radial surface density distribution in ugc 7321 due to its edge - on geometry ( cf . olling 1996 ) , we can compute a mean surface density within the stellar disk . adopting the definition @xmath127@xmath128 , where @xmath129 is the linear optical diameter of the disk ( e.g. , roberts & haynes 1994 ) we find @xmath130=5.3 pc@xmath8 .
this value is considerably lower than the median of this quantity found by roberts & haynes ( 1994 ) for scd / sd spirals in the ugc catalogue : @xmath130=9.80 pc@xmath8 , but is consistent with the @xmath130 ratios of several of the lsb galaxies in the sample of de blok , mcgaugh , & van der hulst ( 1996 ) .
we have presented @xmath0- and @xmath1-band imaging and photometry , near - infrared @xmath3-band imaging , narrow - band h@xmath2 imaging , and 21-cm line measurements of the nearby , edge - on sd spiral galaxy ugc 7321 .
ugc 7321 is a `` superthin '' galaxy , with an extremely dynamically cold stellar disk and no discernible bulge component , even in the near - infrared . in spite of its edge - on orientation , ugc 7321
is visibly quite diffuse and it is clear that its intrinsic optical surface brightness is quite low .
the dust content of ugc 7321 also appears to be small , hence we argue that the `` anemic '' appearance of the galaxy results from the low - level of current star formation rather than severe internal extinction .
ugc 7321 exhibits significant @xmath9 color gradients in the radial direction : measured from the disk center to its edges edge , @xmath1311.05 magnitudes along the galaxy major axis .
dust alone can not explain the large gradient .
using a simple extinction model we find @xmath1320.80 after correction for internal reddening .
this is somewhat larger than the color gradients typically observed in normal giant spirals ( cf .
de jong 1996 ) , and suggests significant stellar population gradients in the disk of ugc 7321 .
the outermost disk regions of ugc 7321 have @xmath830.6 , suggesting they are composed of stellar populations that include a significant fraction of young stars .
however , ugc 7321 also contains a population of old stars with @xmath1331.1 , indicating it is not a young or recently - formed galaxy . the rather strong radial segregation of these populations suggests that the galaxy has evolved quite slowly and perhaps that viscous evolution has not operated efficiently in this system .
ugc 7321 also exhibits appreciable vertical color gradients : @xmath134 as large as 0.45 magnitudes was measured parallel to the minor axis .
this is a reflection of a concentration of young blue stars along the galaxy midplane , and a population of the older , red stars at larger scale heights .
this type of age segregation is predicted to occur due to dynamical heating processes in spiral disks , but ugc 7321 represents one of the few examples of it being directly observed in an external galaxy .
this implies that even dynamically cold disks like ugc 7321 have undergone some dynamical heating . the stellar light distribution of the ugc 7321 disk can not be characterized by a single exponential function .
a light excess over the prediction of an exponential model is seen at small radii , and in both optical and nir wavelengths , the disk appears to contain distinct surface brightness `` zones '' between which the slope of the brightness profile changes . this may be an additional signature of minimal viscous evolution .
it also brings into question the prediction of some semi - analytic galaxy models ( e.g. , dalcanton , spergel , & summers 1997 ) that exponential stellar disks are a natural product of the disk formation process . in spite of its regular , organized disk , many of the global properties of ugc 7321 , including its luminosity , content , rotational velocity , and @xmath16 ratio ( table 1 ) , as well as its slowly - rising , solid - body rotation curve ( fig .
19 ) are more typical of a dwarf irregular galaxy than a normal sd spiral .
the origin of such vastly different disk morphologies in an otherwise similar physical parameter space remains unclear , but it may place an important constraint on galaxy formation and evolution models .
together the properties of ugc 7321 suggest that it is an under - evolved galaxy in both a dynamical and in a star - formation sense .
nonetheless , this galaxy clearly demonstrates that even seemingly simple `` pure disk '' galaxies like the superthins are highly complex systems .
when deriving central surface brightness measurements of galaxies , it is desirable to convert measured values to face - on values in order that galaxies observed at different inclinations can be meaningfully compared . however , for highly - inclined galaxies , the approximation that @xmath135 ( e.g. , freeman 1970 ) breaks down , and alternative corrections must be used . here
we derive a simple formula for this purpose , using a `` thin disk '' approximation . in our model
we assume the galaxy disk is optically thin and has an exponential brightness profile along the @xmath33 direction of the form @xmath57 , where @xmath136 is the luminosity volume density and @xmath18 is the disk scale length .
we also assume the disk has a finite but non - negligible thickness of 2@xmath60 , and a constant brightness along the @xmath24 direction ( see fig . 20 ) .
for such a disk , the central surface brightness if viewed edge - on is expressed as @xmath137 if viewed face - on , the central surface brightness would be simply @xmath138 .
now suppose our model disk is viewed at some arbitrary angle @xmath139 through a line of sight along a vector * x * ( see fig .
we define @xmath140 as the angle between the line of sight and a normal to the disk .
now the integrated brightness of the disk along a path from @xmath141 to @xmath142 becomes @xmath143 taking the integral , and making the substitution @xmath144 we find that @xmath145.\ ] ] to apply this approximation to a real galaxy , one can substitute the measured radial and vertical scale heights for @xmath18 and @xmath60 respectively , provided that the conditions @xmath146 and @xmath147 both hold .
dettmar , r. -j .
1995 , in physics of the interstellar medium and the intergalactic medium , edited by a. ferrara , c. f. mckee , c. heiles , & p. r. shapiro , asp conference series , vol .
80 , ( astronomical society of the pacific , san francisco ) , p. 398
de vaucouleurs , a. & longo , g. 1988 , catalogue of visual and infrared photometry of galaxies from 0.5@xmath62 m to 10@xmath62 m ( 1961 - 1985 ) , university of texas monographs in astronomy no .
5 ( austin , university of texas at austin ) | we present @xmath0- and @xmath1-band optical imaging and photometry , h@xmath2 narrow - band imaging , near - infrared @xmath3-band imaging , and 21-cm spectroscopy of the nearby ( @xmath4=407 km / s ) , sd spiral galaxy ugc 7321 .
ugc 7321 exhibits a remarkably thin stellar disk with no discernible bulge component .
the galaxy has a very diffuse , low surface brightness disk , which appears to suffer relatively little internal extinction in spite of its nearly edge - on geometry ( @xmath588@xmath6 ) .
if seen face - on , ugc 7321 would have an observed central @xmath0-band surface brightness of only @xmath723.4 mag arcsec@xmath8 .
the ugc 7321 disk shows significant @xmath9 color gradients in both the radial and vertical directions : @xmath100.80 magnitudes along the galaxy major axis , and @xmath11 as large as 0.45 magnitudes is observed parallel to the galaxy minor axis .
these color gradients can not be explained solely by dust and are indicative of changes in the mix of stellar ages and/or metallicity as a function of both radius and height above the galaxy plane .
the outer regions of the ugc 7321 disk are too blue to be explained by low metallicity alone ( @xmath12 ) , and must be relatively young .
however , the galaxy also contains stellar populations with @xmath131.1 , indicating it is not a young or recently - formed galaxy .
the disk of ugc 7321 is not a simple exponential , but exhibits a light excess at small radii , as well as distinct surface brightness zones . despite its organized disk structure , many of the global properties of ugc 7321 ( @xmath14 ; = 1.1@xmath15 ; @xmath16=1.1 / ; @xmath17=233 ; @xmath18=2.1 kpc )
are reminiscent of a dwarf galaxy .
together the properties of ugc 7321 imply that it is an under - evolved galaxy in both a dynamical and in a star - formation sense
. 0.5 cm |
fractal conception @xcite has become widespread idea of contemporary science ( see refs . @xcite
@xcite , for review ) .
characteristic feature of fractal sets is known to be the self - similarity : if one takes a part of the whole set , it looks like the original set after a deformation @xmath2 .
formal basis of the self - similarity is the power - law function @xmath3 with a hurst exponent @xmath4 ( for time series , value @xmath5 is reduced to the fluctuation amplitude and @xmath6 is the interval size within which this amplitude is determined ) .
while the simple case of monofractal is characterized by a single exponent @xmath4 , a multifractal system is described by a continuous spectrum of exponents ( singularity spectrum ) @xmath7 @xcite . along this way
, a self - similarity degree @xmath8 represents the exponent of a homogeneous function being characteristic function of self - similar systems @xcite ( so , within nonextensive thermostatistics , this exponent expresses the escort probability @xmath9 throughout the original one @xmath10 @xcite ) . in physical applications , a central role
is played by the partition function @xmath11 ( @xmath12 is a characteristic size of boxes covering multifractal ) to be determined by the mass exponent @xmath13 connected with the generalized hurst exponent @xmath7 by the relation @xmath14 .
since fractals are scale invariant sets , it is natural to apply the quantum calculus to description of multifractals because quantum analysis is based on the jackson derivative which yields variation of a function with respect to the deformation of its argument @xcite .
this letter is devoted to such a description with using deformed series of expansions of the partition function , the mass exponent and the averages of random variables over deformed powers of the difference @xmath0 . as shows the consideration in section [ sec.2 ] , the coefficients of the partition function expansion
are reduced to binomial - type combinations of the tsallis entropies related to manifold deformations @xmath15 with growing powers @xmath16 respectively , section [ sec.3 ] shows that escort probability is generated by the mass exponent whose expansion into deformed power series represents the generalization of well - known relation @xmath1 with fractal dimension spectra @xmath17 corresponding to manifold deformations .
according to section [ sec.4 ] , above expansion permits also to expresses physical averages over the escort probability @xmath18 in terms of related averages based on the ordinary one @xmath10 .
section [ sec.5 ] contains discussion of results obtained , and in appendix we adduce necessary information from the quantum calculus .
following the standard scheme @xcite , we consider a multifractal set covered by boxes @xmath19 with number @xmath20 .
the peculiarity of self - similar sets is that the probability to occupy a box @xmath21 is determined by its size @xmath22 according to the power - law relation @xmath23 characterized by a hlder exponent @xmath24 .
all properties of a multifractal is known to be determined by the partition function @xmath25 which according to the normalization condition takes the value @xmath26 at @xmath27 . because @xmath28 for all boxes @xmath21 , the function ( [ z ] ) decreases monotonically from maximum magnitude @xmath29 related to @xmath30 . in the simplest case of the flat distribution @xmath31 fixed by the statistical weight @xmath32
, one has the exponential decay @xmath33 .
our approach is based on the expansion of the partition function ( [ z ] ) into the deformed series ( [ 7 ] ) in powers of the difference @xmath0 @xcite @xmath34_\lambda!}(q-1)_\lambda^{(n)}. \label{z1}\ ] ] here , the deformed powers @xmath35 are determined by the binomial ( [ 8 ] ) , while the kernels @xmath36 are defined by the @xmath37-fold action of the jackson derivative : @xmath38 by using the method of induction with accounting the definitions ( [ 1 ] ) and ( [ z ] ) , one obtains the explicit expression @xmath39 where @xmath40 are binomial coefficients . in absence of deformation @xmath41 , all coefficients ( [ kernel ] ) equal zero , apart from the single term @xmath42 .
inserting these coefficients into the series ( [ z1 ] ) gives @xmath43 as desired .
it is easily to see that the set of kernels ( [ kernel ] ) with @xmath44 is expressed in terms of the tsallis entropy @xcite @xmath45 with growth of the deformation parameter @xmath6 , this entropy decreases monotonically with inflection in the point @xmath46 where the expression ( [ s ] ) takes the boltzmann - gibbs form @xmath47 . in the case
@xmath48 , the coefficient ( [ kernel ] ) does not depend on the deformation parameter @xmath6 to give the value @xmath49 . at @xmath50 ,
one obtains @xcite @xmath51 respectively , in the second order @xmath52 , eq.([kernel ] ) yields @xmath53 with accounting the definition ( [ s ] ) one has @xmath54 in arbitrary order @xmath37 , combination of equations ( [ kernel ] ) and ( [ s ] ) arrives at the sum @xmath55 which expresses in explicit form a generalized entropy of a multifractal by means of contributions of the tsallis entropies related to different powers of the set deformation . in the limit @xmath56 , when @xmath57 , accounting of all terms in the sum over @xmath58 arrives at cancellation of the denominator @xmath59 so that @xmath60 ; similarly , at @xmath61 , one obtains the boltzmann - gibbs limit @xmath62 as desired . finally , in the case @xmath63 , where @xmath64 , the term related to @xmath65 is reduced to zero so that main contribution gives the term with @xmath66 to arrive at the sign - changing asymptotics @xmath67 .
for the flat distribution , when @xmath68 , the dependence ( [ s ] ) is characterized by the asymptotics @xmath69 $ ] in the limit @xmath70 and @xmath71 at @xmath72 . as a result , with the @xmath6 growth the coefficients ( [ kernel ] ) increase from the value @xmath73 to the boltzmann magnitude @xmath74 and then decay to the asymptotics @xmath67
following pseudo - thermodynamic picture of multifractal sets @xcite , let us define effective values of the free energy @xmath75 , the internal energy @xmath76 and the entropy @xmath77 : @xmath78 here , @xmath79 is the size of characteristic box in phase space , @xmath10 and @xmath18 are original and escort probabilities to be connected with the definition @xmath80 inserting this equation into the second expression ( [ fa ] ) , one obtains the well - known legendre transformation @xcite @xmath81 where @xmath8 plays the role of the inverse temperature and the internal energy is specified with the state equation @xmath82 it is easily to convince the escort probability ( [ prob ] ) is generated by the characteristic function being the mass exponent @xmath83 really , one has @xmath84 by analogy with eq.([z1 ] ) , one can expand the function ( [ tau ] ) into the deformed series @xmath85_\lambda!}(q-1)_\lambda^{(n ) } \label{tau1}\ ] ] being the generalization of the known relation @xmath1 connecting the mass exponent @xmath75 with the multifractal dimension spectrum @xmath17 @xcite . similarly to the equations ( [ z2 ] ) and ( [ kernel ] ) , the kernel @xmath86 is determined by the @xmath37-fold action of the jackson derivative : @xmath87 noteworthy , the term with @xmath48 is absent in the series ( [ tau1 ] ) . at @xmath50 , one reproduces the ordinary relation @xmath88 , while the kernels @xmath86 with @xmath89 include terms being proportional to @xmath90 to correspond to manifold deformations @xmath15 , @xmath91 .
generally , the definition ( [ dd ] ) yields a hierarchy of the multifractal dimension spectra related to different multiplication factors @xmath37 of the set deformation . in trivial case of the flat distribution )
is not applicable in this case because fixation of the statistical weight @xmath92 makes impossible variation of the probability @xmath31 .
] , when the partition function is @xmath68 , the definition ( [ tau ] ) yields the mass exponent @xmath93 with the fractal dimension @xmath94 which tends to @xmath95 when the size of covering boxes @xmath12 goes to the inverse statistical weight @xmath96 .
thus , we can conclude the flat distribution relates to a monofractal with dimension @xmath97 tending to a smooth one - dimensional set at deformation @xmath98 .
let us consider an observable @xmath99 distributed over a multifracal set . according to ref .
@xcite , its mean value @xmath100 is determined by the escort probability ( [ prob ] ) . with accounting eqs .
( [ p ] ) ( [ dd ] ) , the average ( [ o ] ) can be expressed in terms of the sum @xmath101_\lambda!}\frac{(q-1)_\lambda^{(n)}}{(\lambda-1)^{n } } \sum\limits_{m=0}^n(-1)^{n - m}{n\choose m}\lambda^m\left<\phi\right>_{\lambda^m}\ ] ] where the specific average @xmath102 is introduced . at @xmath65 ,
the escort distribution @xmath103 is reduced to the original one @xmath10 so that related terms in the sum ( [ oo ] ) are proportional to the ordinary average @xmath104 . on the other hand , in absence of deformation @xmath41 , the only term with @xmath48 determines the mean value ( [ o ] ) : @xmath105 . in the limit @xmath56 , the escort probability reduced to flat distribution @xmath106 for arbitrary form of the ordinary distribution so that average ( [ o ] ) is determined by the simple expression @xmath107 .
finally , in the case @xmath63 , main contribution is given by the terms with @xmath65 and @xmath66 so that the average ( [ oo ] ) takes the form @xmath108
a principle peculiarity of above consideration is that the expansions ( [ z1 ] ) , ( [ tau1 ] ) and ( [ oo ] ) depend of both deformation parameter @xmath6 and self - similarity degree @xmath8 . as shows the example of using the flat distribution in the end of section [ sec.3 ] , the former determines the fractal dimension @xmath97 in the simple case of the monofractal generated by deformation @xmath72 of initial cell of size @xmath79 .
with passage to general case of multifractal , we obtain the fractal dimension spectrum defined by the series ( [ dd ] ) . along this line , the self - similarity degree @xmath8 plays the role of a free parameter whose variation describes the multifractal spectrum , while value of the deformation parameter @xmath6 fixes the scale of the multifractal resolution to be determined by external conditions in the case of natural objects .
as pointed out in introduction , our study is based on the jackson derivative ( [ 1 ] ) which gives variation of a function with respect to scale choice of its argument .
this derivative expresses the coefficients of the deformed taylor expansion ( [ 7 ] ) in powers of the deformed difference @xmath0 defined by the binomial ( [ 8 ] ) . in this way
, the expansion ( [ z1 ] ) of the partition function ( [ z ] ) is determined by the coefficients ( [ k ] ) being binomial - type combinations of the tsallis entropies ( [ s ] ) related to manifold deformations @xmath15 with growing powers @xmath16 according to definitions ( [ fa ] ) and ( [ prob ] ) the mass exponent ( [ tau ] ) is expressed by the legendre transformation ( [ l ] ) to generate the escort probability ( [ p ] ) . on the other hand , the expansion of the mass exponent into deformed power series ( [ tau1 ] ) represents generalization of known relation @xmath1 with fractal dimension spectra ( [ dd ] ) corresponding to manifold deformations . moreover , making use of the deformed expansion ( [ 7 ] ) arrives at the expression ( [ oo ] ) of the physical averages ( [ o ] ) corresponding to the escort probability @xmath18 in terms of the averages ( [ a ] ) related to multiple deformations . under deformation @xmath6 , above
multifractal characteristics vary in the following manner .
the generalized entropies ( [ k ] ) tend to the statistical weight @xmath92 in the limit @xmath56 , take the boltzmann - gibbs magnitude @xmath109 at @xmath46 and arrive at the sign - changing asymptotics @xmath67 when @xmath63 .
what about the relation ( [ oo ] ) between the physical average @xmath110 and the ordinary one @xmath111 , one has the simple expression @xmath107 in the limit @xmath56 , the trivial relation @xmath105 in absence of deformation @xmath41 , and the difference @xmath112 being proportional to the factor @xmath0 at @xmath63 . from physical point of view
, the most tractable is the expansion of the mass exponent ( [ tau1 ] ) whose coefficients represent fractal dimension spectra . as shows the equality ( [ dd ] ) , the fractal dimensions @xmath113 related to different powers @xmath37 of the series ( [ tau1 ] ) can change their signs with variation of the deformation @xmath6 .
because , within the pseudo - thermodynamic formalism , the mass exponent @xmath114 and the self - similarity degree @xmath8 relate to the free energy and the inverse temperature respectively , such a change means that thermodynamic potential can obtain a singularity at some deformations @xmath6 . within framework of the physical presentation @xcite
, this singularity means the phase transition in the course of the deformation of the multifractal set .
quantum analysis is known to be based on the jackson derivative @xcite @xmath115 where @xmath6 is a deformation parameter .
the deformed taylor expansion reads : @xmath116_\lambda!}\left.\mathcal{d}_x^\lambda f(x)\right|_{x = a}+\frac{(x - a)_\lambda^{(2)}}{[2]_\lambda!}\left.\big(\mathcal{d}_x^\lambda\big)^2 f(x)\right|_{x = a}+\dots\\=\sum\limits_{n=0}^\infty\frac{(x - a)_\lambda^{(n)}}{[n]_\lambda ! } \left.\big(\mathcal{d}_x^\lambda\big)^n
f(x)\right|_{x = a}. \end{split}\ ] ] here , the deformed binomial @xmath117_\lambda \lambda^{\frac{m(m-1)}{2}}x^m y^{n - m } \end{split}\ ] ] is determined by the coefficients @xmath118_\lambda=\frac{[n]_\lambda!}{[m]_\lambda![n - m]_\lambda ! } \label{9}\ ] ] where generalized factorials @xmath119_\lambda!=[1]_\lambda[2]_\lambda\dots[n]_\lambda$ ] are given by the basic deformed numbers @xmath120_\lambda=\frac{\lambda^n-1}{\lambda-1}. \label{10}\ ] ] b.b .
mandelbrot , the fractal geometry of nature , freeman , san francisco , 1982 .
j. feder , fractals , plenum publishers , new york , 1988 .
barnsley , r. hawley , fractals everywhere , academic press professional , boston , 1993 .
gouyet , physics and fractal structures , springer - verlag , new york , 1996 .
d. sornette , critical phenomena in natural sciences , springer , new york , 2001 .
halsey , m.h .
jensen , l.p .
kadanoff , i. procaccia , b.i .
shraiman , _ fractal measures and their singularities : the characterization of strange sets _ , phys .
a 33 ( 1986 ) 1141 - 1151 .
a. erzan , j .-
eckmann , _ @xmath8-analysis of fractal sets _ , phys . rev . lett . 78
( 1997 ) 3245 - 3248 . c. tsallis , introduction to nonextensive statistical mechanics approaching a complex world , springer , new york , 2009 . c. beck , f. schlgl , thermodynamics of chaotic system , cambridge university press , cambridge , 1993 .
v. kac , p. cheung , quantum calculus , springer , new york , 2002 . c. kassel , quantum groups , springer , new york , 1995 .
s. abe , _ a note on the @xmath8-deformation - theoretic aspect of the generalized entropies in nonextensive physics _ , phys .
lett . a 24 ( 1997 ) 326 - 330 . | within framework of the quantum calculus , we represent the partition function and the mass exponent of a multifractal , as well as the average of random variables distributed over self - similar set , on the basis of the deformed expansion in powers of the difference @xmath0 . for the partition function , such expansion is shown to be determined by binomial - type combinations of the tsallis entropies related to manifold deformations , while the mass exponent expansion generalizes known relation @xmath1 .
we find the physical average related to the escort probability in terms of the deformed expansion as well .
it is demonstrated the mass exponent can acquire a singularity that relates to a phase transition of the multifractal set in the course of its deformation .
multifractal set ; deformation ; power series |
currently there are about 70 multiply - imaged systems due to galactic mass scale gravitational lenses .
the statistics of gravitational lenses depend on three key ingredients , namely , the cosmology , the number density of potential lenses as a function of redshift and the dynamical properties of galaxies ( e.g. , velocity dispersions and the surface mass densities ) .
gravitational lenses hence encode information of the cosmology , the galaxy mass profiles and the evolution history of galaxies . at present , the lens sample is too small to constrain all the ingredients simultaneously .
most previous studies concentrated on constraining the cosmological constant assuming non - evolving populations of lenses ( e.g. , fukugita et al . 1992 ; kochanek 1996 ; helbig et al .
1999 ) . under this assumption ,
the most recent lens statistics study of chae et al .
( 2002 ) finds that the lens statistics are best - fitted by a present - day matter density of @xmath12 and a cosmological constant of @xmath13 .
this result is consistent with results from a variety of other studies , including the cosmic microwave background radiation ( e.g. , de bernardis et al .
2000 ; spergel et al . 2003 ) , type ia supernovae at cosmological distances ( e.g. , riess et al .
1998 ; perlmutter et al . 1999 ) , and the large - scale structures in the universe ( e.g. , peacock et al . 2001 ) . in light of the convergence of the cosmological model
, it becomes important to use gravitational lenses for a different purpose : to study the evolution of galaxies ( their number density and dynamical properties ) in the @xmath14 cosmology ( hereafter @xmath15cdm ) .
lensing limits on galaxy evolution have been explored by mao ( 1991 ) , mao & kochanek ( 1994 ) , rix et al .
( 1994 ) , and jain et al .
( 2000 ) , all of which used optically - selected lenses .
lensing is sensitive to the evolution of galaxy properties as the lensing probability is @xmath16 , while separations are @xmath17 ; here @xmath18 is the number density and @xmath19 is the velocity dispersion of typical lenses .
for example , ( for a fixed cosmological model ) a decreasing number density of galaxies with redshift(@xmath20 ) lowers the lensing rate and the mean redshift of lenses while a decreasing velocity dispersion with @xmath20 lowers the lensing rate , the mean lens redshift and the mean angular size of image separations .
therefore , through a careful analysis of the lens redshifts , image separations and lensing probability we can constrain the evolution of the number density and dynamical properties of galaxies .
most lensing galaxies are massive early - type galaxies as they dominate the lensing cross - sections due to their larger central mass concentrations .
gravitational lenses therefore provide a unique mass - selected sample to study the evolution of early - type galaxies , independent of and complementary to the traditional redshift surveys of galaxies ( e.g. , fried et al .
2001 ; i m et al .
this is a much debated research area .
there exist two different views on the formation and evolution of early - type galaxies , namely a monolithic collapse model ( eggen , lynden - bell , & sandage 1962 ) and a merger hypothesis ( toomre & toomre 1972 ) . in the monolithic collapse model ,
early - type galaxies are thought to have formed rapidly at high redshift and then evolve passively to the present - day .
the merger model is a natural consequence of the hierarchical structure formation theory .
semi - analytic implementations of this model predict a continuous formation of ellipticals and hence a certain fraction of massive early - type galaxies must have formed since @xmath0 ( e.g. kauffmann 1996 ; baugh et al .
1996 ; kauffmann et at .
1999 ) ; the fraction depends on the assumed cosmology and other assumptions and is typically one third or more ( for more , see 4 ) .
an observational way of probing galaxy evolution is using redshift survey of galaxies .
however , no consensus has been reached either with this method .
for example , kauffmann , charlot , & white ( 1996 ) found rapid evolutions in the number density of ellipticals while schade et al .
( 1999 ) advocated the opposite conclusion using similar samples ( see 4 for more details ) .
the purpose of this work is to use data from the cosmic lens all - sky survey ( class ) to provide independent constraints on galaxy evolution . as we were completing this work ,
a complementary study has been carried out by ofek , rix , & maoz ( 2003 ) ; their results are compared with our results in 2 .
we use data from the cosmic lens all - sky survey ( class ) for our study .
the survey is described extensively in myers et al .
( 2003 ) and browne et al .
we refer the readers to those papers for details , and here we only give a brief summary .
the class well - defined statistical sample contains 8958 radio sources including 13 multiply - imaged sources .
class sources contain further 9 multiply - imaged sources which , however , do not fall into our statistical sample ( myers et al .
2003 ; browne et al .
the advantage of the class survey is that it uses very well - defined observational selection criteria and does not suffer from the effect of dust extinction in lenses .
it is also the largest completed survey for gravitational lenses , so it is the best sample for our purposes .
we assume the galaxy population is described by a schechter luminosity function ( lf ) , n(l ) d(l l _ ) = n_(ll_)^ ( -l / l _ ) d(l l _ ) .
[ eq : schechter ] lensing galaxies are modelled as singular isothermal ellipsoids , which are described by two parameters , the velocity dispersion and the axial ratio ( or equivalently , the ellipticity ) .
the galaxy luminosity is related to the velocity dispersion via [ eq : sigma ] l l _ = ( _ ) ^. we divide galaxies into two populations , namely the early - type ( ellipticals and s0 s ) population and the late - type population and assume that each population is described by its own lf .
the parameter values we take are identical to those in chae et al .
( 2002 ) ( see also chae 2003 for the details of the analysis and further results ) ; we refer the readers to those two papers .
in particular , we adopt the type - specific lfs based on the galaxy classifications from the second southern sky redshift survey ( ssrs2 : marzke et al .
1998 ) ; see chae ( 2003 ) for the details .
our adaptation of maximum likelihood analyses ( kochanek 1993 ) is identical to that of chae et al .
( 2002 ) and chae ( 2003 ) . under the @xmath1 cosmology we test two simple models of the evolution of early - type galaxies .
the evolution of late - type galaxies is not considered , as we can not obtain any useful limits due to the small number of late - type lens galaxies in the current sample ( i.e. 1 or 2 ) .
in the first model we adopt , we assume that the shape of the luminosity function ( @xmath21 ) is a constant given by the ssrs2 ( chae 2003 ) and the non - evolving ( characteristic ) velocity dispersion @xmath22 ( defined in eq .
[ eq : sigma ] ) is a constant to be determined from the data , but the ( characteristic ) number density of galaxies ( defined in eq .
[ eq : schechter ] ) evolves as a function of redshift ; we choose a power - law evolutionary shape of @xmath23 , [ eq : nz ] n_(z ) = n_,0 ( 1+z)^_n , where @xmath24 is the present - day value .
the no - evolution model corresponds to @xmath25 .
1 shows confidence limits on the parameter @xmath26 and @xmath22 .
the @xmath27 in fig . 1 refers to @xmath28 where @xmath29 is the likelihood function ( chae et al .
2002 ; chae 2003 ) . as one can see , the lens statistics are consistent with a no - evolution model in a @xmath1 cosmology at @xmath5 level .
we find @xmath30 ; namely , the number density of galaxies at redshift @xmath7 can not be smaller by a factor of two or larger by @xmath31 than the present day number density . from fig . 1 the non - evolving characteristic velocity dispersion is @xmath32 km s@xmath33(@xmath5 ) .
this value is in good agreement with the values from recent analyses of lensing statistics assuming no evolution of galaxies ( chae et al .
2002 ; chae 2003 ; davis , huterer , & krauss 2003 ) . in the second model
, we allow the velocity dispersion @xmath22 ( as well as the number density ) to vary as a function of redshift : [ eq : nv ] _ ( z ) = _ , 0 ( 1+z)^_v , where @xmath34 is the present - day characteristic velocity dispersion .
the no - evolution model corresponds to @xmath35 .
2 shows the limits in the parameter space of @xmath26 , @xmath36 and @xmath34 .
2(a ) , ( b ) and ( c ) are the three projected parameter planes .
2(d ) shows the limits in the plane of @xmath36 and @xmath34 based only on the image separations and the available lens redshifts ( 6 of them ) of the nine single - galaxy induced multiply - imaged systems ( see section 3.1 of chae 2003 ) , namely without using the lensing rate [ see below for a discussion of the fig .
the contours shown on each plane represent 68% , 90% , 95% and 99% confidence levels for one parameter .
the contours in fig .
2(a ) are elongated parallel to a line @xmath37 .
this is because along the @xmath37 line the optical depth ( @xmath16 ) is a constant .
in other words , there is a degeneracy in determining the evolutions of the velocity dispersion and the number density from the optical depth alone .
notice , however , that the degeneracy is in part broken by the observed image separations as a function of redshift .
the @xmath5 limits on the two evolutionary indices are : @xmath38 and @xmath8 .
the limit on @xmath36 is particularly interesting : lensing statistics demand that the velocity dispersion for an @xmath39 galaxy at @xmath7 must be between @xmath9 and @xmath10 of the present - day value .
this implies that dynamically , the population of lensing galaxies can not be much different from the present - day population . from fig .
2(b ) or ( c ) the characteristic velocity dispersion of the present - day early - type population is @xmath40 km s@xmath33 ( @xmath5 ) .
this value has a relatively large uncertainty and is consistent with the value for the non - evolving case shown in fig . 1 .
however , it is of interest to note that the best - fit value of @xmath34 is somewhat larger than the non - evolving value .
this is then consistent with the best - fit value of @xmath36 being negative .
recently , ofek et al . (
2003 ) have used the redshifts of the lensing galaxies in moderate - size source - redshift(@xmath41)limited samples to constrain the galaxy mass evolution .
our work is different as we use the well - defined uniform class sample and we include all the lensing information ( lensing rate , image separations and lens redshifts ) .
notice that our sample ( 13 lenses in total ) includes 6 systems with both lens and source redshifts measured while the ofek et al .
( 2003 ) samples have up to 17 .
they also conclude that there is little evidence for rapid evolution of early - type galaxies .
they parameterize the evolutions in different forms from ours . but
equivalently , they find that at 95% confidence level , @xmath22 at @xmath7 should be at least 63% of the present value ; this is similar to our @xmath5 limit . despite the small number of the measured redshifts in our sample our limits on the evolution of @xmath22
are relatively strong because of the additional constraint of the lensing rate .
this can be seen from the comparison of fig .
2(c ) and ( d ) . without the lensing rate
, we have @xmath42 [ fig .
2(d ) ] , which is significantly broader .
the ofek et al .
( 2003 ) limit on the number density evolution is given by @xmath43 , translating into a @xmath5 lower limit of the number density of lenses at @xmath7 of 30% of the present value ; our limit ( 57% ) is significantly stronger because of the strong effects of the lensing rate .
our results are consistent with an early - formation / passive - evolution picture ( e.g. @xmath44 ) of early - type galaxies , as also inferred from studies of the fundamental plane of lensing galaxies ( see kochanek et al .
2000 ; rusin et al .
we have used the statistical properties of the class strong lens sample ( i.e. , the rate of multiple - imaging and the image separations as a function of redshift ) to constrain the evolution of galaxies .
the lens sample is unique as it is mass - selected and hence the constraints obtained from it are independent of those from redshift surveys .
the method we use is based on chae et al .
( 2002 ) and chae ( 2003 ) and has some uncertainties , such as the adopted luminosity function of early - type galaxies and the redshift distribution of the source population in the class survey ( see chae 2003 for details ) . however , these uncertainties are smaller than those arising from the moderate - size class sample of lenses .
we find that the ( comoving ) number density of lensing galaxies at redshift @xmath7 must be within 50% to 164% of the present - day number density ; their characteristic velocity dispersion also must be within @xmath9 and @xmath10 of the present value .
the lensing statistics are therefore consistent with a slow evolution of galaxies in both their number density and their dynamical properties .
these results are inconsistent with very fast evolution of early - type galaxies , where the number of early - types at @xmath45 is only 20%-40% or less of the present - day values ( e.g. , lin et al .
1999 ; fried et al . 2001 ; wolf et al .
our results are , however , consistent with several other studies , in particular the study based on the hubble space telescope observations of the groth strip ( i m et al .
2002 ) where there find a number density evolution of @xmath46 ( see their table 6 ) in the same underlying cosmology . in the standard cold dark matter model with @xmath47 , the number density of bright ellipticals at @xmath7
is predicted to be a factor of 2 - 3 smaller than the local value ( kauffmann 1996 ; baugh , cole & frenk 1996 ) , inconsistent with our results .
however , the evolution of galaxies is expected to be slower in the @xmath1 cosmology .
indeed , our results are consistent with the modest evolution out to redshift @xmath48 predicted by kauffmann et al .
( 1999 ; see their fig .
9 ) for the same cosmology .
quantitatively , for ellipticals with a stellar mass @xmath49 , the space density at @xmath7 is about 30% lower than the present day value in the fiducial model of cole et al .
this result is robust to changes in the amount of star formation occurring in bursts at high redshift ( c. baugh , private communication ; baugh et al .
our results are consistent with these predictions
. the effects of galaxy evolution on the lens statistics will be more dramatic for very high - redshift quasars ( @xmath11 , fan et al .
2003 ) . for a source at @xmath50 , half of the multiple - imaging cross - section ( with separations between 0.3 to 5 arcseconds )
is contributed by galaxies with redshifts greater than 1.2 if there was no evolution of galaxies .
hence the lensing probability will be reduced significantly if the comoving number density of early - type galaxies is much smaller at redshifts @xmath51 compared with the local number density .
a large sample of very high - redshift quasars is therefore an independent and effective way of probing galaxy evolution .
the lensing constraints on the galaxy evolution are already quite competitive compared with those from other methods .
however , they still suffer from the small number of lenses ( @xmath52 ) in our class statistical sample . with planned upgrades in major radio instruments ( such as evla and e - merlin ) , it is possible to obtain a radio lens sample that is an oder of magnitude larger .
when such a sample becomes available , the lensing constraint will become more stringent and more physical evolution ( rather than our toy ) models can be tested .
we thank i. browne and c. baugh for discussions and the anonymous referee for constructive comments .
khc acknowledges support from the astrophysical research center for the structure and evolution of the cosmos ( arcsec ) which was established under the kosef src program .
khc also acknowledges the hospitality provided by the jodrell bank observatory during the summer of 2003 .
baugh , c. m. , cole , s. , frenk , c. s. 1996 , , 283 , 1361 browne , i. w. a. , et al .
2003 , mnras , 341 , 13 chae , k .- h .
2003 , mnras , in press ( astro - ph/0211244 ) chae , k .- h . ,
2002 , phys .
lett . , 89 , 151301 cole , s. , lacey , c. g. , baugh , c. m. , & frenk , c. s. 2000 , , 319 , 168 davis , a. n. , huterer , d. , & krauss , l. m. 2003 , , 344 , 1029 de bernardis , p. et al .
2000 , nature , 404 , 955 eggen , o. j. , lynden - bell , d. , & sandage , a. 1962 , , 136 , 748 fan , x. h. , et al .
2003 , , 125 , 1649 fried , j. w. , et al . 2001 , a&a , 367 , 788 fukugita , m. , futamase , t. , kasai m. , & turner , e. l. 1992 , apj , 393 , 3 helbig , p. , marlow , d. , quast , r. , wilkinson , p. n. , browne , i. w. a. , & koopmans , l. v. e. 1999 , a&as , 136 , 297 i m , m. , et al .
2002 , apj , 571 , 136 jain , d. , panchapakesan , n. , mahajan , s. , & bhatia , v. b. 2000 , mod .
phys . lett .
a , 15 , 41 kauffmann , g. 1996 , , 281 , 487 kauffmann , g. , charlot s. , & white s. d. m. 1996 , mnras , 283 , l117 kauffmann , g. , colberg , j. m. , diaferio , a. , & white , s. d. m. 1999 , mnras , 307 , 529 kochanek , c. s. 1993 , apj , 419 , 12 kochanek , c. s. 1996 , apj , 466 , 638 kochanek , c. s. , et al .
2000 , apj , 543 , 131 lin , h. , et al . 1999 ,
apj , 518 , 533 mao , s. 1991 , apj , 380 , 9 mao , s. , & kochanek , c. s. 1994 , mnras , 268 , 569 marzke , r. o. , da costa , l. n. , pellegrini , p. s. , willmer , c. n. a. , & geller , m. j. 1998 , apj , 503 , 617 myers , s. t. et al .
2003 , mnras , 341 , 1 ofek , e. o. , rix , h .- w . , maoz , d. 2003 , , 343 , 639 peacock , j. a. , et al .
2002 , , 410 , 169 perlmutter , s. , turner , m. s. , white , m. 1999 , phys .
lett . , 83 , 670 riess , a. g. , et al .
1998 , aj , 116 , 1009 rix , h .- w . ,
maoz , d. , turner , e. l. , & fukugita , m. 1994 , apj , 435 , 49 rusin , d. , et al .
2003 , apj , 587 , 143 schade , d. , et al .
1999 , , 525 , 31 spergel , d. n. s. , et al .
2003 , apjs , 148 , 175 toomre , a. , & toomre , j. 1972 , , 178 , 623 wolf , c. , meisenheimer , k. , rix , h .- w . ,
borch , a. , dye , s. , kleinheirich , m. 2003 , , 401 , 73 | we use gravitational lenses from the cosmic lens all - sky survey ( class ) to constrain the evolution of galaxies since redshift @xmath0 in the current @xmath1 cosmology .
this constraint is unique as it is based on a mass - selected lens sample of galaxies .
our method of statistical analysis is the same as in chae ( 2003 ) .
we parametrise the early - type number density evolution in the form of @xmath2 and the velocity dispersion as @xmath3 .
we find that @xmath4 ( @xmath5 ) if we assume @xmath6 , implying that the number density of early - type galaxies is within 50% to 164% of the present - day value at redshift @xmath7 .
allowing the velocity dispersion to evolve , we find that @xmath8 ( @xmath5 ) , indicating that the velocity dispersion must be within @xmath9 and @xmath10 of the present - day value at @xmath7 .
these results are consistent with the early formation and passive evolution of early - type galaxies .
more stringent limits from lensing can be obtained from future large lens surveys and by using very high - redshift quasars ( @xmath11 ) such as those found from the sloan digital sky survey . |
the queueing theory has been considerably studied since erlang started designing telephone exchanging system in 1909 @xcite . in @xcite
, he developed the theory of the call - loss system , which significantly contributed to the progress of telephones and electric communication systems .
after his study , kendall presented his first paper @xcite .
since then , the study on the queueing theory including kendall s notation @xcite , burke s theorem @xcite , jackson networks @xcite , and little s theorem @xcite has been accelerated .
nowadays , it is applied to study on social systems such as the internet , resource management systems , vehicular traffic systems and pedestrian traffic systems .
let us review here the discrete - time m / m/1 queueing process , which will be extended in the next section .
m / m/1 means that the queue has a markovian entry of customers and a markovian exit at 1 server .
each particle ( customer ) enters the system with a probability @xmath0 and receives service with a probability @xmath1 , see figure [ fig : mm1q ] .
the system is characterized only by the number @xmath2 of particles .
the probability @xmath3 of finding @xmath4 particles at time @xmath5 is governed by the following master equation : @xmath6 for @xmath7 . when @xmath8 , a unique stationary - state solution to the master equation can be easily obtained as the following geometric distribution : @xmath9 the average number of particles can be easily calculated as @xmath10 on the other hand , the asymmetric simple exclusion process ( asep ) in one dimension is one of typical interacting particle systems and admits exact analyses of non - equilibrium properties @xcite . it has been vigorously studied in the recent two decades . the stationary state of the totally asep with open boundaries ( open tasep ) was found in a matrix product form in @xcite . since then , matrix product stationary states of various generalizations of the asep containing discrete - time versions of the open tasep have been found @xcite .
the tasep is one of the basic models of the vehicular traffic . although both the m / m/1 queueing process and the open tasep are markov processes with particle entry and exit , they are different in the following two points .
first , the number of particles in the m / m/1 queueing process does not have an upper limit , whereas that in the open tasep can not be greater than the fixed number of sites . in other words , the state space of the m / m/1 queueing process is infinite and that of the open tasep is finite .
it should be noticed that the open tasep is a call - loss system , which has been seldom discussed .
second , the tasep has a spatial structure , and particles are affected by their excluded volume ( hard - core repulsion ) . by contrast , the m / m/1 queueing process has no spatial structure , and only the number of particles characterizes the system .
an extension of the m / m/1 queueing process with the excluded - volume effect ( in other words , the tasep with a new boundary condition ) was introduced in @xcite . in this model , particles enter the system at the left site next to the leftmost occupied site and exit at the rightmost site of the chain , see fig .
[ fig : eq ] .
when we ( pedestrians ) make a queue , we usually proceed if there is a space in front of us ( excluded - volume effect ) , which is a motivation to treat this model .
let us call the model _ exclusive queueing process_. in @xcite , a stationary state of the exclusive queueing process was given by a slightly arranged matrix product form for the open tasep , where the probability of finding each configuration is proportional to @xmath11 the row vector , the matrices and the column vector are independent of the entry rate @xmath0 , whereas @xmath12 is independent of the exit rate @xmath1 .
the exponent @xmath13 denotes the position of the leftmost particle . in this paper
, we consider a discrete - time version of the exclusive queueing process with the entry , hopping and exit rates ( @xmath14 and @xmath1 , respectively ) replaced by probabilities within one time step .
although there are some choices of an update scheme , we take the fully - parallel - update rule . the special case where the bulk hopping is deterministic ( i.e. , the hopping probability is 1 ) has been analyzed in @xcite . to find a stationary - state solution is one of basic problems
the idea to find it in this paper is reducing the balance equation for the exclusive queueing process to that for the open tasep .
then , a stationary - state solution will be obtained in a similar form to that for the continuous - time model as .
we calculate the critical line which separates the parameter space into the regions with and without the stationary state by evaluating the convergence of the normalization constant .
the critical line will be written in terms of the stationary current of the open tasep .
we calculate the average length of the system ( the position of the leftmost particle ) and the average number of particles in the system .
we also calculate the average number of particles in the queue .
this paper is organized as follows . in sect .
[ model ] , we define the exclusive queueing process with discrete time . in sect .
[ stat - state ] , we obtain a stationary state of the model . in sect .
[ pdav ] , we identify the region where the stationary state exists in the parameter space , and calculate the average length of the system and the average number of particles . in sect . [
sec : p=1 ] , we treat the model in the case where the bulk hopping probability is 1 . section [ conc ]
is devoted to the conclusion of this paper .
detailed calculations are done in appendices .
we introduce an extension of the discrete - time m / m/1 queueing process with the excluded - volume effect on the semi - infinite chain ( exclusive queueing process ) , see fig .
[ fig : eq ] , where each site is labeled by a natural number from right to left .
each particle enters the chain at the left site next to the leftmost occupied site with a probability @xmath0 , hops to its right nearest neighbor site with a probability @xmath15 , if it is empty , and exits at the right end of the chain with a probability @xmath1 .
if there is no particle on the chain , a particle enters at site 1 .
these transitions occur simultaneously within one time step . in other words , we take the fully - parallel - update scheme .
( we call it simply parallel update hereafter . )
the model is formulated as a discrete - time markov process on the state space @xmath16 where 0 and 1 correspond to unoccupied and occupied sites , respectively .
in particular , @xmath17 denotes the state that there is no particle on the chain .
note that we do not write infinite 0s located left to the leftmost 1 .
we define @xmath18 by the length of a sequence @xmath19 .
in particular , for each element @xmath19 of @xmath20 , @xmath18 stands for the length of the system ( the position of the leftmost occupied site ) .
let us write the probability of finding a configuration @xmath19 at time @xmath5 as @xmath21 .
the master equation governing the model is written as @xmath22 where @xmath23 is the transition - probability matrix of the process and @xmath24 is the transition probability from @xmath25 to @xmath19 within one time step .
in particular , for @xmath26 and @xmath27 , @xmath28 it is difficult to write down the action of @xmath23 explicitly for the general configuration @xmath19 because of the parallel update .
for example , for @xmath29 , @xmath30 see fig .
[ fig:101101 ] , and for @xmath31 , @xmath32 see fig .
[ fig:100100 ] . with a non - zero probability within one time step.,height=215 ] with a non - zero probability within one time step.,height=170 ] note that this model is not equivalent to the parallel - update tasep with the ordinary open boundary condition ( open tasep ) @xcite , see fig .
[ fig : opentasep ] . in the open tasep , particles can enter the system at the fixed leftmost site of the finite chain only if it is empty .
this means that the open tasep is a call - loss system .
on the other hand , in our model , particles can always enter the system .
a stationary state ( stationary distribution ) is a solution to the balance equation @xmath33 with the constraint @xmath34 let us begin with making an assumption that the stationary - state solution has the following form : @xmath35 where @xmath36 is a function on @xmath37 and satisfies @xmath38 from @xmath39 , we get @xmath40 from @xmath41 , we get @xmath42 thus , @xmath43 we reduce the balance equation for the general configuration to that for the open tasep .
one can obtain , for all @xmath44 with length @xmath13 , @xmath45 for all @xmath46 with length @xmath13 , @xmath47 and for all @xmath48 with length @xmath13 , @xmath49 see appendix [ appa ] for detailed calculation .
the sets @xmath50 are defined as @xmath51 for example , @xmath52 equations - give the balance equation for the open tasep on the @xmath13-site chain with the entry probability equal to 1 if we regard @xmath53 as the probability of finding a configuration @xmath19 .
thus , if we set @xmath36 to be proportional to the stationary - state solution to the open tasep , the form and gives a stationary - state solution to our model . substituting the matrix product solution to the open tasep which was found in @xcite into
, we obtain @xmath54 where @xmath55 and @xmath56 are matrices , @xmath57 is a row vector and @xmath58 is a column vector
. they should satisfy @xmath59 which is the algebraic relation found in @xcite with @xmath60 .
this algebraic relation is compatible with the assumption - .
a representation of the relation is given in appendix [ appb ] .
in view of , we have to chose the normalization as @xmath61 note that , from the algebraic relation , we can derive @xmath62 , which is compatible with .
the normalization constant @xmath63 is expressed as @xmath64 when the right - hand side converges .
this form is similar to the generating function of the normalization constant of the open tasep .
the function @xmath65 is useful to calculate the average length of the system and the average number of particles .
the case where @xmath66 and @xmath67 corresponds to the normalization constant : @xmath68 we can easily solve the balance equation of the usual m / m/1 queue recursively .
( the equation is just a three - term recurrence formula . ) in our model , however , we have not found a recursive way to solve the balance equation . in this sense , our assumption and
is truly an ansatz ( or working hypothesis ) . in the continuous - time case
, a similar form gives its stationary state except that @xmath69 is replaced by @xmath0 , and an overall constant @xmath70 appears instead that the first @xmath71 in the matrix product disappears , see ( 25 ) in @xcite .
however , the continuous - time limit ( first replace @xmath72 and @xmath73 , and then take the limit @xmath74 ) of the solution is exactly the same as the solution to the continuous - time model . at this stage , we do not know if the power series for @xmath63 ( the right - hand side of ) converges and the solution is meaningful . note that the form without the normalization constant always gives a stationary measure of our model , which is a solution to the balance equation without the constraint .
if a process on a countable state space is irreducible and recurrent , and has two stationary measures @xmath75 and @xmath76 , then @xmath75 is proportional to @xmath76 ( proposition ii.1.3 of @xcite ) .
furthermore , if an irreducible process has a stationary state , the process is ( positive ) recurrent and thus its stationary state is unique ( theorem ii.1.1 of @xcite ) .
our process is easily seen to be irreducible for generic @xmath77 and @xmath15 .
thus the stationary state is unique if the power series for @xmath63 converges and there is no ( normalizable ) stationary state if it diverges ( corollary ii.1.2 of @xcite ) .
in the next section , we obtain the condition on the parameters such that the power series for @xmath63 converges and a more explicit closed form for @xmath63 .
we can derive the following closed form for @xmath78 , see appendix [ appb ] for details : @xmath79 where @xmath80 the power series for @xmath81 ( the right - hand side of with @xmath67 ) converges to @xmath82 when @xmath83 replacing @xmath84 by @xmath85 , we obtain the region where the power series for @xmath63 ( the right - hand side of ) converges as @xmath86 and its closed form as @xmath87 where @xmath88 . as we see fig .
[ fig : pdgen ] , the region is embedded in the region @xmath8 where the usual m / m/1 queueing process converges . the critical line @xmath89 consists of two parts ; a curve and a straight line .
mathematically , the curve is given by a solution to the denominator of the form @xmath90 , and the straight line by a solution to @xmath91 .
physically , the two parts correspond to the stationary current of the open tasep in the high - density phase ( @xmath92 ) and the maximal - current phase ( @xmath93 ) , respectively , see table [ tab : disccont ] .
this property is due to the form .
in fact , the stationary current of the open tasep on @xmath13-site chain with the entry probability 1 is given by @xmath94 where @xmath95 is the coefficient of @xmath96 in @xmath81 .
note that the critical line for the continuous - time version of the model can be also written in terms of the stationary current of the continuous - time open tasep @xcite , see table [ tab : disccont ] again .
ll discrete time ( parallel update ) & continuous time + entry probability @xmath0 & entry rate @xmath0 + exit probability @xmath1 & exit rate @xmath1 + hopping probability @xmath15 & hopping rate @xmath15 + @xmath97 & @xmath97 + @xmath98 & @xmath99 + @xmath0 & @xmath0 + @xmath100 & @xmath101 + let us calculate some average values in the stationary state .
we can calculate the average length of the system ( the position of the leftmost particle ) and the average number of particles in the system by differentiating @xmath102 with respect to @xmath84 and @xmath103 , respectively , as @xmath104 the number @xmath105 of particles in the queue ( i.e. , @xmath106 for @xmath107 and @xmath108 for @xmath17 ) is also one of the typical values in the queueing theory .
we can calculate the average of @xmath105 as @xmath109_{\xi=\frac{\alpha}{p(1-\alpha)},\zeta=1 } \\ = & \langle n\rangle-\frac{\alpha}{\beta}.\end{aligned}\ ] ]
derivation of the second equality is given in appendix [ appc ] .
( solid line ) divides the parameter space into the convergent phase and the divergent phase .
we drew the dashed line representing the critical line of the usual m / m/1 queueing process as well .
note that there is a region @xmath110 where the m / m/1 queueing process converges but our new model diverges .
, height=226 ]
in this section we treat the model with @xmath111 , where the bulk hopping rule is nothing but one of the typical deterministic cellular automata , i.e. , rule 184 .
this case was analyzed in @xcite .
the matrices and vectors have a two - dimensional representation @xmath112 see @xcite . in the case where @xmath111 , each particle necessarily hops if its right neighbor site is empty .
thus , any configurations containing at least one sequence 00 vanish in the stationary state .. but the process on the subset of the state space @xmath20 consisting of @xmath17 and all the configurations without the sequence 00 is irreducible . ] in fact , we can see @xmath113 one can reduce the stationary - state probability to the following simpler expression : for @xmath114 ( @xmath115 ) , @xmath116 thanks to the two dimensional representation , calculating @xmath78 is an easy exercise : @xmath117 the square root vanishes from @xmath63 , and thus the critical line loses the straight line part , see fig . [
fig : pd184 ] : @xmath118 the average length and the average numbers of particles in the system and the queue are simplified as @xmath119 it is also easy to calculate the probability distributions of @xmath13 and @xmath4 by expanding @xmath81 and @xmath120 around @xmath121 and @xmath122 , respectively : @xmath123 \\ & = \frac{1}{z } \left(\frac{\alpha}{1-\alpha}\right)^l \frac{1}{l!}\left.\frac { \partial^l { { \mathcal z}}(\xi,1 ) } { \partial \xi^l } \right|_{\xi=0 } = \frac{\beta-\alpha\beta-\alpha}{\beta(1-\alpha ) } \left(\frac{\alpha}{\beta(1-\alpha)}\right)^l , \end{split } \\
\begin{split } & { \rm prob}[\text{the number of particles in the system}=n ] \\ & = \frac{1}{z } \frac{1}{n ! } \left . \frac{\partial^n { { \mathcal z}}(\frac{\alpha}{1-\alpha},\zeta)}{\partial \zeta^n } \right|_{\zeta=0 } \\ & = \begin{cases } \displaystyle \frac{\beta-\alpha\beta-\alpha}{\beta(1-\alpha ) } & n=0 , \\ \displaystyle \frac{\beta-\alpha\beta-\alpha } { \beta(1-\beta+\alpha\beta)(1-\alpha ) } \left(\frac{\alpha(1-\beta+\alpha\beta ) } { \beta(1-\alpha)^2}\right)^n & n\ge 1 .
\end{cases } \end{split}\end{aligned}\ ] ] @xmath124 .
the critical line @xmath89 ( solid line ) divides the parameter space into the convergent phase and the divergent phase .
we drew the dashed line representing the critical line of the usual m / m/1 queueing process as well .
note that there is a region @xmath125 where the m / m/1 queueing process converges but our new model diverges.,title="fig:",height=245 ]
in this paper , we introduced an extension of the discrete - time m / m/1 queueing process with excluded - volume effect ( exclusive queueing process ) .
we took the parallel - update scheme .
a stationary - state solution was obtained in a slightly arranged matrix product form for the parallel - update open tasep .
we found that the critical line which separates the parameter space into regions with and without the stationary state is given by the stationary current of the open tasep in the maximal - current and high - density phases .
particularly , we should note that the entry rate can not be greater than the maximal current of the tasep if the model converges .
we calculated the average length of the system ( the position of the leftmost particle ) and the average number of particles in the system .
these results recover the stationary state , the critical line and the average values of the continuous - time version of the exclusive queueing process @xcite , in the limit where the time interval @xmath126 . in this sense ,
our discrete - time model is a generalization of the continuous - time model .
we calculated the number of particles in the queue as well .
when @xmath111 , i.e. the bulk hopping is deterministic , the matrices and the vectors constructing the stationary - state solution have a two - dimensional representation .
the probability distributions of the length of the system and the average number of particles are both geometric in this case .
we leave calculations of the density profile and correlation functions in the stationary state as future works . in the tasep with the ordinary open boundary condition
, a domain wall theory explains its phase diagram successfully @xcite .
investigating how our model converges to the stationary state or diverges with a similar argument will be also an interesting study .
this work is supported by global coe program `` education and research hub for mathematics - for - industry '' and grants - in - aid ( for young scientists ( b ) 22740106 and for jsps fellows
20@xmath12710918 ) from jsps and ministry of education , culture , sports , science and technology .
in this section , we derive eqs . - . we calculate the action of @xmath23 on @xmath128 under the assumption and .
let us use a short hand notation @xmath129 .
for a configuration @xmath130 with @xmath131 , the action of @xmath23 is calculated as @xmath132 where we used @xmath133 and @xmath134 from @xmath135 , we get . for a configuration @xmath136 with @xmath137 ,
the action of @xmath23 is calculated as @xmath138 where we used @xmath139 and @xmath140 from @xmath141 , we get . for a configuration @xmath142 with @xmath143 ,
the action of @xmath23 is calculated as @xmath144 where we used @xmath145 and @xmath146 from @xmath147 , we get .
according to @xcite , @xmath148 and @xmath58 can be reduced as @xmath149 where matrices @xmath150 and @xmath151 , and vectors @xmath152 and @xmath153 satisfy @xmath154 we can easily show that @xmath148 and @xmath58 satisfy the relations from the relation - . in view of the normalization , we impose @xmath155 .
the following gives a representation for these matrices and vectors , which will not be used in calculating @xmath78 : @xmath156 where we arranged the representation ( 6.6)-(6.11 ) in @xcite . to calculate @xmath78 , we imitate the calculation for @xmath111 as @xmath157 where @xmath158 .
using the relation , we can eliminate @xmath159 and @xmath151 in @xmath160 as @xmath161 furthermore , using the relation , we can eliminate @xmath162 and @xmath153 in @xmath78 as @xmath163 in the last equality , we used the second relation of and @xmath164 .
now we borrow an idea from section 4.3 of @xcite , where the generating function of the tasep with a single defect particle on a ring was obtained . set @xmath165 with @xmath166 and @xmath167 ,
and we get @xmath168 noting the relations and .
set @xmath169 , and we get @xmath170 where @xmath171 set @xmath172 , and we get @xmath173 finally , we achieve @xmath174 which can be simplified as and expanded as @xmath175 .
\end{split}\end{aligned}\ ] ]
in this section , we show @xmath177 the left - hand side is calculated as @xmath178 recall the formula @xmath179 using this and the relations and , we obtain @xmath180 finally , noting @xmath181 , @xmath155 and the relation , we achieve .
rogsch , c. , schadschneider , a. , seyfried , a. , klingsch , w. : how to select the
right one " - update schemes for pedestrian movement in simulation and reality .
proceedings of the traffic and granular flow 09 ( to be published ) | in a recent study @xcite , an extension of the m / m/1 queueing process with the excluded - volume effect as in the totally asymmetric simple exclusion process ( tasep ) was introduced . in this paper
, we consider its discrete - time version .
the update scheme we take is the parallel one .
a stationary - state solution is obtained in a slightly arranged matrix product form of the discrete - time open tasep with the parallel update .
we find the phase diagram for the existence of the stationary state .
the critical line which separates the parameter space into regions with and without the stationary state can be written in terms of the stationary current of the open tasep .
we calculate the average length of the system and the average number of particles . |
it is known that the global potential energy surface ( pes ) of a system containing @xmath0 interacting water molecules can be formally expressed in terms of the many - body expansion of the interaction energy as a sum over @xmath1-body terms with @xmath2,@xcite @xmath3 here , @xmath4 collectively denotes the coordinates of all atoms in the @xmath5-th water molecule , @xmath6 is the one - body potential describing the energy required to deform an individual water molecule from its equilibrium geometry , and all higher @xmath1-body interactions , @xmath7 , are defined recursively through
@xmath8 the rapid convergence of equation ( [ eq : mb_exp ] ) , demonstrated in several studies of water clusters,@xcite suggests that the pes associated with a system containing @xmath0 water molecules can in principle be represented as a sum of low - order interactions that are amenable to accurate calculations using correlated electronic structure methods [ e.g. , coupled cluster with single , double , and perturbative triple excitations , ccsd(t ) , which currently represents the gold standard in quantum chemistry ] .
the theoretical modeling of many - body effects in water began in the late 1960s when self - consistent field ( scf ) calculations were carried out on small water clusters.@xcite these studies concluded that nonadditive effects in water are generally nonnegligible , with 3b contributions being as large as 10 15% of the total pair interaction for ring structures .
4b effects were estimated to be , on average , @xmath91% of the total pair interaction .
the first attempt to derive potential energy functions for water from _ ab initio _ data was made by clementi and co - workers who fitted dimer energies calculated at the hartree - fock ( hf ) level of theory to an analytical representation of the 2b interactions.@xcite later , configuration interaction ( ci ) calculations were used to derive the pairwise additive mcy potential,@xcite to which 3b and 4b terms were subsequently added through a classical description of polarization effects.@xcite the first water potential including explicit 2b and 3b terms , derived respectively from 4@xmath10 order mller - plesset ( mp4 ) and hf calculations , along with a classical description of higher - body polarization interactions , was reported in ref . .
following these pioneering efforts , the asp potentials with rigid monomers were derived using intermolecular perturbation theory.@xcite in the 2000s , xantheas and co - workers introduced the ttm ( thole - type model ) water potentials , which for the first time made use of a highly accurate 1b term derived from _ ab initio _ calculations.@xcite the latest versions ( ttm3-f@xcite and ttm4-f@xcite ) were shown to reproduce the properties of water clusters , liquid water , and ice reasonably well , although some inaccuracies were identified in the calculations of vibrational spectra.@xcitearound the same time , 2b and 3b potentials with rigid water monomers were derived by szalewicz and co - workers from symmetry - adapted perturbation theory ( sapt).@xcite these studies eventually led to the development of the rigid - monomer cc - pol potential , a 25-site model with explicit 2b and 3b terms fitted to ccsd(t)-corrected mp2 dimer energies and sapt trimer energies , respectively , with higher - body terms being represented through classical polarization.@xcite cc - pol was shown to accurately reproduce the vibration - rotation tunneling spectrum of the water dimer and to predict the structure of liquid water in reasonable agreement with the experimental data . within the cc - pol scheme , a refined 2b term with explicit dependence on the monomer flexibility
has also been developed.@xcite more recently , the full - dimensional whbb@xcite and mb - pol@xcite many - body potentials with flexible monomers were developed .
whbb includes explicit 2b and 3b terms fitted to ccsd(t ) and mp2 data , respectively , with long - range many - body effects being represented by the same thole - type model used in the ttm3-f potential .
mb - pol was derived from fits to ccsd(t ) energies calculated for both water dimers and trimers in the complete basis set ( cbs ) limit and includes many - body effects within a modified version of the polarization model employed by the ttm4-f potential .
although whbb has successfully been applied to static calculations of several water properties , to date mb - pol is the only many - body potential that has been consistently employed in quantum molecular dynamics ( md ) simulations that accurately predicted the properties of water from the gas to the condensed phase .
while the calculations with many - body potentials reported in the literature demonstrate that equation ( [ eq : mb_exp ] ) can be effectively used to develop accurate molecular - level representations of the water interactions , the role of short- and long - range contributions at different orders has not been fully investigated . here
, we seek to address this issue by analyzing water potentials that treat the leading terms of the many - body expansion through implicit ( e.g. , ttm3-f and ttm4-f ) and explicit ( e.g. , whbb and mb - pol ) representations , with a specific focus on how these terms are defined and incorporated as a function of the intermolecular separations .
the article is organized as follows : the four potentials are briefly described in section 2 , an analysis of the energetics as well as of the vibrational frequencies and intensities for water systems ranging from the gas - phase dimer to the liquid phase is reported in section 3 , and the conclusions are given in section 4 .
the four potentials studied here were chosen because they treat the intramolecular distortions on equal footing , i.e. , through the 1b pes developed by partridge and schwenke.@xcite in this way , the ability of the models to predict water properties is a direct reflection of the treatment of the intermolecular interactions . in addition
, whbb and mb - pol effectively share the same induction scheme employed by ttm3-f and ttm4-f , respectively , which thus allows for a systematic investigation of the effects of explicit low - order terms of the many - body expansion .
the dominant terms of the many - body water interactions are the 2b and 3b terms , which can be conveniently split into short- and long - range contributions , @xmath11 . at the two - body level ,
long - range interactions are dominated by electrostatics and dispersion , while exchange - repulsion and charge - transfer become increasingly important at short - range .
3b interactions in water , on the other hand , arise primarily from induced interactions at long - range and exchange - repulsion when the monomer electron densities overlap.@xcite both ttm3-f and ttm4-f include permanent and induced electrostatics as well as dispersion and repulsion at the 2b level , while all higher order terms are represented through many - body induction.@xcite although whbb and mb - pol seek to exploit the range separation of the low - order water interactions , the two potentials are intrinsically different both in philosophy and by construction . in whbb , the short - range 2b term is described by a 7@xmath12-degree permutationally invariant polynomial that smoothly transitions for an oxygen - oxygen separation between 6.5 and 7.5 into the long - range component described by the 2b ttm3-f potential .
the 3b term only includes a short - range component , which is represented by either a 5@xmath12- ( whbb5 ) or a 6@xmath12-degree ( whbb6 ) permutationally invariant polynomial that dies off as the largest oxygen - oxygen separation between two molecules of the trimer approaches 8.0 . all higher - order interactions ( @xmath13 )
are described by the polarization model employed in the ttm3-f potential . by construction ,
whbb thus employs a strict separation at the 2b and 3b levels between short- and long - range interactions that are described in a completely independent way and are essentially disentangled from all higher - order contributions .
it should be noted that a simplified version of whbb including only 1b , 2b , and 3b interactions with shorter cutoffs was also used,@xcite which , however , can not be implemented in a straightforward way in standard software for molecular simulations in periodic boundary conditions . for this reason , only calculations with the full whbb5 implementation of ref .
are reported in the following analysis . on the other hand
, mb - pol can be viewed as a classical polarizable model supplemented by short - range 2b and 3b terms that effectively represent quantum - mechanical interactions arising from the overlap of the monomer electron densities .
specifically , at all separations , @xmath14 includes ( damped ) dispersion forces derived from _ ab initio _ computed asymptotic expansions of the dispersion energy along with electrostatic contributions due to the interactions between the molecular permanent and induced moments . at short - range ,
@xmath14 is supplemented by a 4@xmath12-degree permutationally invariant polynomial that smoothly switches to zero as the oxygen - oxygen separation in the dimer approaches 6.5 .
similarly , the mb - pol 3b term , @xmath15 , includes a 3b polarization term at all separations , which is supplemented by a short - range 4@xmath12-degree permutationally invariant polynomial that effectively corrects for the deficiencies of a purely classical representation of the 3b interactions in regions where the electron densities of the three monomers overlap .
this short - range 3b potential is smoothly switched off once the oxygen - oxygen separation between any water molecule and the other two water molecules of a trimer reaches a value of 4.5 . in mb - pol ,
all induced interactions are described through many - body polarization using a slightly modified version of ttm4-f.@xcite mb - pol thus contains many - body effects at all monomer separations as well as at all orders , in an explicit way up to the third order and in a mean - field fashion at all higher orders .
it should be noted that alternative functions to the permutationally invariant polynomials ( e.g. , gap@xcite ) have been suggested and successfully employed in modeling short - range many - body interactions.@xcite
the effects of the different representations of the many - body interactions employed by ttm3-f , ttm4-f , whbb , and mb - pol are investigated through the analysis of the energetics of water systems ranging from the gas - phase dimer to the liquid phase .
figure 1a shows a comparison of the dimer binding energies calculated for the global minimum configuration of each pes .
although all potentials predict binding energies within @xmath90.2 kcal / mol of the ccsd(t)/cbs reference value ( independently of whether the dimer geometry is optimized within each pes or the ccsd(t)/cbs dimer geometry is used in the calculations),@xcite the analysis of the many - body water interactions reported in ref .
shows that the ttm3-f and ttm4-f interaction energies deviate significantly from the ccsd(t ) reference data when dimer configurations different from the corresponding equilibrium geometries are considered .
specifically , root mean square deviations ( rmsds ) of 0.73 kcal / mol and 0.44 kcal / mol were obtained in ref . for ttm3-f and ttm4-f ,
respectively , from comparisons with @xmath91,400 ccsd(t)/aug - cc - pvtz 2b energies corrected for the basis set superposition error .
it was instead shown that both whbb and mb - pol provide a highly accurate representation of the whole multidimensional dimer pes as demonstrated by rmsds of 0.15 kcal / mol and 0.05 kcal / mol , respectively , obtained from comparisons with more than 40,000 ccsd(t)/cbs 2b energies.@xcite due to the different treatment of many - body effects , the relative accuracies of the ttm3-f , ttm4-f , whbb , and mb - pol potentials are expected to differ as the number of molecules in the system of interest increases .
this is shown in figure 1b , where the relative energies of low - lying hexamer isomers calculated with the four potentials are compared with the corresponding _ ab initio _
values.@xcite the hexamer isomers hold a special place in the experimental and theoretical studies of water because they are the smallest clusters in which the lowest energy configurations correspond to fully three - dimensional structures . for this reason ,
the hexamer serves as a prototypical model for the hydrogen - bond networks observed in condensed phases .
while ttm3-f , whbb , and mb - pol ( qualitatively ) reproduce the isomer energy ordering , large deviations are found between the ttm4-f predictions and the reference _ ab initio _ values .
similar results are also obtained for other small water clusters , including the tetramer and pentamer isomers , shown in figures s1 and s2 of the supplemental material.@xcite .
since it was shown that the thole - type polarization scheme employed by the ttm4-f potential correctly captures higher - order water interactions,@xcite the low accuracy shown by ttm4-f in describing the relative energies of the hexamer isomers is possibly due to intrinsic deficiencies in the 2b term . in this context , the performance of the ttm3-f potential , which provides the best agreement with the reference _ ab initio _ values , is somewhat surprising given the large deviations already seen at both the 2b and 3b levels,@xcite and may suggest some fortuitous cancellation of errors .
based on this , it is interesting to note that the isomer energies predicted by whbb , which employs the same electrostatic model as ttm3-f , deviate from the reference data by as much as 2.0 kcal / mol .
the origin of these deviations may result from inaccuracies in the short - range 2b and 3b whbb polynomials , incompatibility between the effective many - body representation encoded in the ttm3-f electrostatic model with the short - range whbb polynomials , or both . on the other hand ,
mb - pol , which employs a slightly modified version of the ttm4-f electrostatic model , predicts isomer energies in good agreement with the reference data , suggesting that the mb - pol short - range 2b and 3b terms accurately correct for the deficiencies that are intrinsic to the purely classical representation of these interactions encoded in the ttm4-f electrostatic model . to provide more quantitative insights into the different performance of the four potentials in describing the relative energies of the hexamer isomers , a many - body decomposition of the corresponding interaction energies ( table [ tab : mbdecomp ] )
was carried out at the ccsd(t)-f12b / vtz - f12 level of theory@xcite with the vtz - f12 basis set,@xcite which was shown to effectively provide the same accuracy as ccsd(t)/cbs calculations at a reduced computational cost.@xcite as has been well established,@xcite the interaction energies are dominated by the 2b and 3b terms , with 4b and higher terms making largest ( but still relatively small ) contributions for cyclic structures.@xcite it is interesting to note that , while the hexamer binding energies predicted by ttm3-f are in excellent agreement with the reference calculations ( figure 1b ) , the results of table [ tab : mbdecomp ] demonstrate that this agreement is clearly achieved by a fortuitous cancellation of errors .
for example , in the prism isomer , the ttm3-f 2b interaction is too attractive by @xmath93.2 kcal / mol while the 3b interaction is too repulsive by @xmath93.8 kcal / mol , resulting in a total interaction energy that differs from the reference value by merely 0.79 kcal / mol .
in contrast , ttm4-f , which accurately represents the electrostatic interactions ( as shown by 3b contributions that lie within 1 kcal / mol of the reference values for all low - lying hexamer isomers ) , is unable to correctly describe the 2b interactions , which have significant contributions from exchange / repulsion and charge transfer .
thus , without the benefit of fortuitous cancellation of errors ( that can , to some extent , be encoded into models through empirical parametrization ) , potentials with accurate electrostatics but without a comparably accurate treatment of short - range quantum mechanical effects may be expected to have relatively poor performance . having been explicitly derived from many - body expansions of the interaction energies obtained from correlated electronic structure data , both whbb and mb - pol closely reproduce the ccsd(t)-f12 reference many - body terms . however , nonnegligible discrepancies in the 3b and 4b terms result in whbb being appreciably less accurate than mb - pol at reproducing the hexamer interaction energies .
the differences between the four potentials become more evident when their accuracy is assessed against benchmark quantum monte carlo ( qmc ) interaction energies calculated for liquid water configurations extracted from path - integral molecular dynamics simulations of the vdw - df and vdw - df2 functionals.@xcite qmc has been shown to be a reliable benchmark in the study of small water clusters , producing relative energies with an accuracy comparable to that of ccsd(t).@xcite as a measure of accuracy , the mean absolute deviation ( mad ) between the energies @xmath16 obtained with each of the four potentials and the reference qmc energies @xmath17 was calculated as @xmath18 here , @xmath19 is the number of water configurations and @xmath20 is the average energy difference for all configurations , effectively aligning the zero of energy with the reference qmc data .
figure 1c shows that the mean absolute deviation associated with whbb is @xmath94 times larger than the corresponding values obtained with both ttm3-f and ttm4-f , and @xmath915 times larger than the mb - pol value .
interestingly , mb - pol also achieves better accuracy than all dft models analyzed in ref . .
the ttm3-f and ttm4-f results indicate that effective representations of the many - body interactions through classical polarization models can provide a reasonable description of the energetics associated with the liquid phase ( i.e. , comparable with that provided by most of the dft models currently used in water simulations ) , albeit through empirical parameterization . it should be noted that , as shown by the results reported in table [ tab : mbdecomp ] and the many - body analysis of different dft models carried out in ref .
, cancellation of errors between different terms of the many - body expansion of the interaction energies may also affect the energetics of bulk systems described by both dft and classical polarizable models .
on the other hand , the increased accuracy of mb - pol relative to ttm4-f indicates that short - range many - body interactions beyond a purely classical electrostatic representation are necessary for a correct , molecular - level description of the liquid phase .
the significantly different accuracies associated with the whbb / ttm3-f and mb - pol / ttm4-f potentials also suggest that the correct integration of explicit short - range and effective long - range many - body interactions is critical for ensuring the transferability of the potential across different phases . the ability of ttm3-f , ttm4-f , whbb , and mb - pol to represent the global multidimensional pess of water systems with an increasing number of molecules is directly reflected in the accuracy with which the four potentials predict the associated vibrational frequencies . harmonic frequencies for the water hexamer have been reported at the ccsd(t)/aug - cc - pvdz level.@xcite more recently , _
ab initio _ reference harmonic frequencies for small water clusters have been obtained through two - body : many - body ccsd(t):mp2 calculations using the cc - pvqz and aug - cc - pvqz basis sets for h and o atoms , respectively , which enables a more complete treatment of the basis set through the hybrid two - body : many - body approach.@xcite as shown in figure 2 , the whbb and mb - pol pess result in harmonic frequencies for the water dimer that deviate by less than 10 @xmath21 from the _ ab initio _ frequencies reported in ref . .
these results are consistent with the analysis of the dimer vibration - rotation tunneling spectra of whbb and mb - pol , both of which exhibit nearly perfect agreement with the corresponding experimental data.@xcite in contrast , relatively large deviations are seen for vibrational frequencies of both the ttm3-f and ttm4-f potentials .
the different performance of the four potentials in predicting the relative energies of the hexamer isomers clearly leads to different levels of agreement with the _ ab initio _ harmonic frequencies .
independently of the isomer , the mb - pol values consistently lie within 50 @xmath21 of the reference values while the whbb harmonic frequencies can deviate , in some cases , by more than 100 @xmath21 .
substantially larger deviations , up to @xmath9200 @xmath21 , are instead obtained with both ttm3-f and ttm4-f , reinforcing the notion that purely classical representations of the many - body water interactions are likely not sufficient to fully reproduce the complexity of the underlying multidimensional pess .
interestingly , all potentials predict somewhat larger deviations for the cyclic - chair isomer of the water hexamer , supporting early observations that many - body effects in water are relatively more important for ring structures.@xcite decomposition of the ir spectrum obtained from cmd trajectories with the mb - pol pes in terms of the many - body components of the mb-@xmath22 dms .
1b - dip indicates that the one - body ( gas - phase monomer ) dipoles were used to calculate the dipole of the molecules sampled along the mb - pol cmd trajectories , from which the ir spectrum was calculated .
( 1b+2b)-dip indicates that short - ranged two - body dipoles were used in addition to the one - body dipoles .
mb - dip is the full mb-@xmath22 many - body dipole .
the spectra were smoothed to facilitate the comparison between the line shapes obtained using the different approximations .
the experimental ir spectrum was taken from ref . .
] while the differences in the pess clearly affect the underlying vibrational structure , as shown in figure 2 , the infrared activity of those vibrations is ultimately dictated by the associated dipole moment surfaces ( dmss ) .
many - body representations of higher - order electric properties for molecular systems were characterized beginning in the early 1980s.@xcite the first analysis of many - body effects on the dipole moment of polar molecules was reported by skwara et al.@xcite as noted in 1996 by schwenke,@xcite molecular dipole moments can efficiently be represented in terms of geometry - dependent effective charges multiplied by their cartesian positions . in line with these ideas ,
the first representation of the 2b dipole of water was reported as part of the whbb suite.@xcite a refined version of this 1b + 2b dms@xcite has been used to calculate the transition dipole moments which were then used to modulate the whbb5 frequency distributions calculated within the local monomer approximation from configurations of liquid water extracted from md simulations with the rigid e3b.@xcite neglecting both 3b and higher - order contributions to the dipole moments and dynamical effects ( e.g. , motional narrowing ) , good agreement in the relative intensities was obtained with the measured ir spectrum of liquid water.@xcite recently , a complete many - body representation of the dms of water ( mb-@xmath22 ) was reported,@xcite consisting of the one - body dms of lodi et al.@xcite , an explicit two - body term , and a slightly modified version of ttm4-f polarization for long - range two - body and all higher - order many - body induced dipole moments . the particular functional form of mb-@xmath22 was derived from a systematic analysis of the many - body convergence of the electrostatic properties of water.@xcite c c c model & description & cost per md step relative to q - tip4p / f + q - tip4p / f & point charge & 1.0x + ttm3-f & 1 polarizable site / molecule & 7.3x + ttm4-f & 3 polarizable sites / molecule & 8.5x + whbb & ttm3-f electrostatics + & 29,000x + & empirical 2b dispersion + & + & 2b short - range 7@xmath23-degree polynomial + & + & 3b short - range 5@xmath23-degree polynomial & + mb - pol & ttm4-f electrostatics + & 47x + & _ ab initio
_ 2b dispersion + & + & 2b short - range 4@xmath23-degree polynomial + & + & 3b short - range 4@xmath23-degree polynomial & + to investigate the effects of many - body dipole moments on the ir spectrum of liquid water , many - body molecular dynamics ( mb - md ) simulations,@xcite within the centroid molecular dynamics ( cmd ) formalism , were carried out with the mb - pol pes in combination with the mb-@xmath22 dms . as shown in figure 3 ,
while the overall shape of the ir spectrum is captured in the 1b + 2b representation of the dms , 3b and higher - order contributions to the dipole moment significantly affect the ir intensities .
2b contributions to the dms lower the bend intensity while they contribute to @xmath970% of the intensity of the oh stretch band .
importantly , while 2b contributions are critical to appearance of the shoulder at @xmath9180 @xmath21 corresponding to the hydrogen - bonding stretch , the 1b contributions are non - negligible .
this indicates that some rotational motion is also involved in the hydrogen - bonding stretch .
nevertheless , the absolute intensities of the librational , bending , and stretching bands are only recovered when all many - body effects are included in the calculation of the dipole moment.@xcite these results support early observations based on molecular dynamics simulations with the spc potential supplemented with a 3b dipole - induced dipole term showing that the calculated far infrared spectrum of liquid water was in better agreement with experiment relative to results obtained including only a 2b description of the dipole moment.@xcite while using a highly accurate pes is a prerequisite for a physically correct description of the water properties at the molecular level , an appropriate balance between accuracy and efficiency is often critical when deciding which potential to employ in computer simulations since the computational cost directly affects the ability to calculate statistically converged quantities . as shown in table [ tab : perform ] , a performance analysis carried out on a single intel xeon e5 - 2640 processor for a system consisting of 256 water molecules in periodic boundary conditions indicates that mb - pol is able to achieve a high level of accuracy at a cost of 47x that of the fixed point charge q - tip4p / f model@xcite and @xmath95.5x that of the ttm4-f potential .
whbb , on the other hand , is @xmath929,000 more expensive than q - tip4p / f .
a reference implementation of mb - pol is available as an independent plugin@xcite for the openmm toolkit for molecular simulations.@xcite
the role of short- and long - range contributions to the total energy of water systems ranging from the gas - phase dimer to the liquid phase was investigated by considering four water potentials that treat the leading terms of the many - body expansion through implicit ( i.e. , ttm3-f and ttm4-f pess ) and explicit ( i.e. , whbb and mb - pol pess ) representations . the analysis of the energetics , vibrational frequencies , and infrared intensity indicates that explicit short - range representations of 2b and 3b interactions along with a physically correct incorporation of short- and long - range contributions are necessary for an accurate representation of the water interactions , independently of the system size .
these results thus suggest that atomistic water potentials built upon the many - body expansion of the interaction energy derived from `` first principles '' hold great promise to achieve the long - sought - after goal of describing the macroscopic properties of water across different phases from a rigorous microscopic viewpoint . a question that remains to be addressed
is the extent to which these `` first principles '' many - body water potentials can be extended to the modeling of complex aqueous solutions .
this research was supported by the national science foundation ( grant number che-1453204 ) and used the extreme science and engineering discovery environment ( xsede ) , which is supported by the national science foundation ( grant number aci-1053575 , allocation tg - che110009 ) .
awg acknowledges support by the national science foundation ( grant che-1416571 ) .
grm acknowledges the department of education for support through the gaann fellowship program .
mam was supported through the predictive theory and modeling for materials and chemical science program by the office of basic energy sciences ( bes ) , department of energy ( doe ) and under the auspices of the us doe by llnl under contract de - ac52 - 07na27344 .
we thank prof .
rich saykally for helpful discussions about the origin of the low frequency portion of the ir spectrum of liquid water .
74ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase
10.1021/jp056416 m [ * * , ( ) ] link:\doibase 10.1063/1.3664730 [ * * , ( ) ] link:\doibase 10.1039/c2cp42522j [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1669604 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1672509 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/0009-2614(69)85004-9 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1673723 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/0009-2614(70)87052-x [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1679143 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1680187 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1681930 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.432539 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.432402 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/0009-2614(84)85562-1 [ * * , ( ) ] link:\doibase 10.1002/qua.560360845 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) \doibase http://dx.doi.org/10.1063/1.1423940 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1423941 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.1423942 [ * * , ( ) ] link:\doibase 10.1063/1.1447904 [ * * , ( ) ] link:\doibase 10.1063/1.2837299 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2895750 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2968555 [ * * , ( ) ] link:\doibase 10.1021/jp907648y [ * * , ( ) ] link:\doibase 10.1063/1.4916629 [ * * , ( ) ] link:\doibase 10.1063/1.1311289 [ * * , ( ) ] link:\doibase 10.1063/1.1542871 [ * * , ( ) ] link:\doibase 10.1126/science.1136371 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4722338 [ * * , ( ) ] link:\doibase 10.1021/jp512847z [ * * , ( ) ] link:\doibase 10.1063/1.3579995 [ * * , ( ) ] link:\doibase 10.1021/ct400863 t [ * * , ( ) ] link:\doibase 10.1021/ct500079y [ * * , ( ) ] link:\doibase 10.1021/ct5004115 [ * * , ( ) ] link:\doibase 10.1063/1.473987 [ * * , ( ) ] @noop _ _ ( , ) link:\doibase 10.1021/jz3016777 [ * * , ( ) ] link:\doibase 10.1002/qua.24927 [ ( ) ] link:\doibase 10.1103/physrevb.88.054104 [ * * , ( ) ] link:\doibase 10.1063/1.1408302 [ * * , ( ) ] link:\doibase 10.1021/ct300913 g [ * * , ( ) ] link:\doibase 10.1021/jp8105919 [ * * , ( ) ] link:\doibase 10.1021/ct500129p [ * * , ( ) ] @noop `` '' link:\doibase 10.1063/1.3054300 [ * * , ( ) ] link:\doibase 10.1063/1.2817618 [ * * , ( ) ] link:\doibase 10.1063/1.2831537 [ * * , ( ) ] link:\doibase 10.1039/b617230j [ * * , ( ) ] link:\doibase 10.1080/00268970802708942 [ * * , ( ) ] link:\doibase 10.1016/s0301 - 0104(00)00189 - 0 [ * * , ( ) ] @noop * * , ( ) @noop `` , '' ( ) link:\doibase 10.1063/1.3012573 [ * * , ( ) ] link:\doibase 10.1063/1.4730035 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4885440 [ * * , ( ) ] link:\doibase 10.1021/acs.jctc.5b00225 [ * * , ( ) ] link:\doibase 10.1063/1.4820448 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1016/j.cplett.2007.01.032 [ * * , ( ) ] link:\doibase 10.1021/jp9525447 [ * * , ( ) ] link:\doibase 10.1063/1.3587053 [ * * , ( ) ] link:\doibase 10.1021/ct501131j [ * * , ( ) ] link:\doibase 10.1063/1.3604934 [ * * , ( ) ] link:\doibase 10.1063/1.3554905 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.461069 [ * * , ( ) ] link:\doibase 10.1063/1.3167790 [ * * , ( ) ] @noop `` , '' , @noop * * , ( ) | recent work has shown that the many - body expansion of the interaction energy can be used to develop analytical representations of global potential energy surfaces ( pess ) for water . in this study , the role of short- and long - range interactions at different orders is investigated by analyzing water potentials that treat the leading terms of the many - body expansion through implicit ( i.e. , ttm3-f and ttm4-f pess ) and explicit ( i.e. , whbb and mb - pol pess ) representations .
it is found that explicit short - range representations of 2-body and 3-body interactions along with a physically correct incorporation of short- and long - range contributions are necessary for an accurate representation of the water interactions from the gas to the condensed phase .
similarly , a complete many - body representation of the dipole moment surface is found to be crucial to reproducing the correct intensities of the infrared spectrum of liquid water . |
infrared dark clouds ( irdcs ) are extinction features against the background mid - ir emission ( e.g. simon et al .
2006 ) , which were recognized by the observations with the mid - infrared satellites such as iso and msx ( prault et al . 1996
; egan et al . 1998 ) . mapping observations of the mm and sub - mm continuum emissions ( carey et al . 1998 ; beuther et al . 2002 ; rathborne et al . 2006 ) revealed that massive clumps are associated with several irdcs .
although such massive clumps have a similar mass to the clumps with ongoing high - mass star formation , they are less active .
the gas kinetic temperature of irdcs is as low as 1020 k ( carey et al .
1998 , teyssier et al .
2002 , pillai et al .
therefore , massive clumps in irdcs are thought to be starless or at least in a very early stage of high - mass star formation .
recently , evidences of high - mass star formation were found in several massive clumps associated with irdcs .
rathborne et al .
( 2005 ) found luminous sources ( 9,000@xmath1732,000 @xmath18 ) in massive clumps of irdc g034.43 + 00.24 , which are recognized as high - mass protostars ( @xmath17 10 @xmath19 ) .
rathborne et al .
( 2007 ) also reported that a high - mass protostar ( @xmath20 8 @xmath19 ) is associated with a massive clump in irdc g024.33 + 00.11 on the basis of their high angular resolution mm - wave observation .
beuther et al .
( 2005b ) and beuther & steinacker ( 2007 ) found a protostar in irdc 18223 - 3 using the spitzer and the plateau de bure data , which would grow up to a high - mass star at the end of its formation process .
in addition to these , indirect evidences of high - mass star formation , such as the ch@xmath3oh masers , are found in some irdcs ( ellingsen 2006 ; purcell et al .
2006 ) .
as mentioned above , irdcs are good targets for studies on high - mass star formation . but little attention has been paid for their chemical compositions , although it would be useful to understand the evolutionary stages of irdcs . since binary chemical reactions , such as neutral - neutral and ion - neutral reactions ,
are responsible for chemical processes in the deep inside of molecular clumps , the chemical composition changes slowly toward chemical equilibrium with a time scale of @xmath21 yr .
this time scale is comparable with the dynamical time scale of molecular cloud clumps .
therefore , chemical composition would reflect evolutionary stages of clumps .
for example , carbon - chain molecules , such as ccs and hc@xmath3n , are abundant in the early stage of chemical evolution , while n@xmath0h@xmath1 and nh@xmath3 are abundant in the late stage ( e.g. lee et al . 1996 ) .
furthermore , carbon - chain molecules suffer the great effect from depletion , whereas n@xmath0h@xmath1 and nh@xmath3 are less affected by depletion .
when co , a major destroyer of n@xmath0h@xmath1 , is depleted , n@xmath0h@xmath1 can be abundant in dense parts .
since nh@xmath3 is produced from n@xmath0 , nh@xmath3 can also be abundant ( aikawa et al . 2001 , 2005 ; bergin & tafalla 2007 ) .
thus , we can investigate the evolutionary stages of clumps from the ratios between carbon - chain molecules and n@xmath0h@xmath1 or nh@xmath3 .
this method has been established in low - mass star forming regions .
suzuki et al .
( 1992 ) and benson et al .
( 1998 ) surveyed the carbon - chain molecule toward several low - mass cores , and they found that carbon - chain molecules are more abundant in starless cores than star - forming cores , whereas nh@xmath3 or n@xmath0h@xmath1 is more abundant in star - forming cores .
in addition , hirota et al .
( 2002 , 2004 ) and hirota & yamamoto ( 2006 ) found that carbon - chain molecules are very abundant in several low - mass starless cores .
recently , sakai et al .
( 2007 ) suggested that this method can also be applied to high - mass star forming regions .
sakai et al . ( 2006 , 2007 ) observed the w 3 giant molecular cloud ( gmc ) with the ccs @xmath4=@xmath5@xmath6 and n@xmath0h@xmath1 @xmath2=10 lines , and found that the [ ccs]/[n@xmath0h@xmath1 ] ratio is higher toward a few massive starless cores in the afgl 333 cloud than toward the other massive star forming cores in the w 3 gmc .
sakai et al . ( 2007 ) proposed that the massive starless cores in the afgl 333 cloud are novel candidates which will form high- or intermediate - mass stars , because the ccs emission has hardly been detected in high - mass star - forming regions ( lai & crutcher , 2000 ) .
note that definitions of `` core '' and `` clump '' are described in sakai et al .
( 2007 ) .
chemical composition of clumps would be changed by various activities of newly born stars .
saturated organic molecules like ch@xmath3oh become abundant due to evaporation of grain mantles .
the high excitation lines of ch@xmath3oh are therefore employed as a tracer of hot cores and shocked regions ( e.g. bachiller & prez gutirrez 1997 ) . in this paper , we report a systematic survey of the n@xmath0h@xmath1 @xmath2=10 , ccs @xmath4=@xmath5@xmath6 , hc@xmath3n @xmath2=54 , ch@xmath3oh @xmath2=76 lines , as well as the nh@xmath3 ( @xmath2 , @xmath7)=(1 , 1 ) , ( 2 , 2 ) and ( 3 , 3 ) lines toward massive clumps associated with irdcs .
although observations of several molecular lines were carried out toward massive clumps in the irdcs ( pillai et al . 2006b , 2007 ; leurini et al .
2007 ; purcell et al . 2006
; ragan et al . 2006 ; beuther & sridharan 2007 ) , a systematic survey of molecular lines focusing on chemical composition has been limited . in this paper , we present chemical compositions and physical properties of fifty - five nearby massive clumps associated with the irdcs , and investigate their evolutionary stages .
sridharan et al . ( 2005 ) identified the candidates of massive starless clumps from the objects found with mambo by beuther et al .
their candidates involve clumps associated with irdcs .
rathborne et al .
( 2006 ) also observed 38 irdcs with known kinematic distance by using mambo , and listed the positions of clumps .
we prepared a sample of massive clumps associated with irdcs by use of these two lists on the basis of the following criteria : \1 .
distance is less than 4.5 kpc .
mass is larger than 100 @xmath19 .
galactic longitude is less than 34.5 degree .
in addition to these , we also involved the massive clumps located near the clumps listed by sridharan et al .
their masses are calculated from the data by beuther et al .
( 2002 ) , where the revised equation by beuther et al .
( 2005 ) is employed .
finally , the sample consists of 55 objects , as listed in table [ tab : t1 ] .
in the course of this study , 15 objects are eventually found to have the different lsr velocity from those listed by rathborne et al .
( 2006 ) . as the result
, such objects do not satisfy the above criteria .
we investigated whether the observed objects are associated with point - like sources or extinction features in the msx 8 @xmath9 m data .
then , 12 objects are found to be associated with the msx 8 @xmath9 m sources , and 26 objects to be associated with extinction features .
we call an object with an msx 8 @xmath9 m extinction feature as an msx dark object . in the archival spitzer 24 @xmath9 m data
, we also inspected whether the observed objects are associated with point - like sources or extinction features of 24 @xmath9 m . as a result
, the 31 objects are found to be associated with the point - like 24 @xmath9 m sources .
note that there are no available spitzer data for i18151 - 1208 .
there are 11 dark objects in the spitzer 24 @xmath9 m data , which are all msx 8 @xmath9 m dark objects .
these results are summarized in table [ tab : t1 ] .
the n@xmath0h@xmath1 @xmath2=10 , ccs @xmath4=@xmath5@xmath6 , hc@xmath3n @xmath2=54 , nh@xmath3 ( @xmath2 , @xmath7 ) = ( 1 , 1 ) , ( 2 , 2 ) and ( 3 , 3 ) lines ( table [ tab : t2 ] ) were observed by using the nobeyama radio observatory ( nro ) 45 m telescope from january to march in 2007 .
we used the two sis receivers , s100 for n@xmath0h@xmath1 and s40 for ccs and hc@xmath3n .
the n@xmath0h@xmath1 , ccs , and hc@xmath3n lines were simultaneously observed .
the half - power beam width is about @xmath22 at 45 ghz and about @xmath23 at 93 ghz , and the main beam efficiency is 0.75 at 45 ghz and 0.53 at 93 ghz .
the nh@xmath3 ( 1,1 ) , ( 2,2 ) and ( 3 , 3 ) lines were observed with the hemt receiver .
the half - power beam width is about @xmath24 at 22 ghz , and the main beam efficiency is 0.81 at 22 ghz .
for all the observations , acousto - optical radiospectrometers were employed as a backend , each of which has bandwidth and frequency resolution of 40 mhz and 37 khz , respectively .
the velocity resolutions are 0.50 , 0.25 , and 0.12 km s@xmath25 at 22 ghz , 45 ghz , and 93 ghz , respectively .
the telescope pointing was checked by observing the nearby sio maser sources every 2 hours , and was maintained to be better than @xmath26 .
the line intensities were calibrated by the chopper wheel method .
the system noise temperature was 200300 k for the n@xmath0h@xmath1 , ccs , and hc@xmath3n observations , whereas it was 100150 k for the nh@xmath3 observation .
all the observations were carried out with the position switching mode .
the emission free regions in the galactic ring survey @xmath27co @xmath2=10 data ( jackson et al .
2006 ) were employed as the off positions .
the data reduction was carried out by use of the aips - based software package newstar developed at nro .
the ch@xmath3oh @xmath2=76 lines ( table [ tab : t2 ] ) were observed with the atacama submillimeter telescope experiment ( aste ) 10 m telescope ( ezawa et al .
2004 ; kohno 2005 ) from september to october in 2006 .
we also observed the ch@xmath3oh @xmath10=@xmath28@xmath29 @xmath13 and @xmath10=@xmath30@xmath31 @xmath16 lines simultaneously .
the beam size of the aste telescope is about 22@xmath32 at 345 ghz .
we used the sis receiver , sc345 , operated in the double - side - band mode .
the backends were autocorrelators , whose band width and resolution each are 512 mhz and 0.5 mhz , respectively . by using three autocorrelators , we covered the frequency range from 337.9 to 338.8 ghz .
the telescope pointing was calibrated by observing the thermal continuum emission from the jupiter .
the pointing observation was carried out every observing day , and the pointing accuracy was maintained within 3@xmath32 .
all the observations were performed with the position switching mode .
antenna temperature was calibrated by the single load chopper - wheel method .
the intensity scale was checked by observing the ch@xmath3oh lines toward the ngc 6334 hot core every 0.51 hr in order to correct a small variation of the telescope gain due to the temperature and elevation change . as a result , the relative error in intensity calibration can be maintained to be less than 9 % .
the main beam efficiency of the aste telescope is estimated to be 0.60 , although this includes an uncertainty of 20 % .
the data reduction was carried out by use of the newstar .
table [ tab : t3 ] shows the 3@xmath33 detection rates of the observed lines .
the n@xmath0h@xmath1 @xmath2=10 emission is detected toward 54 out of 55 objects , the hc@xmath3n @xmath2=54 emission toward 43 objects , and the nh@xmath3 ( @xmath2 , @xmath7 ) = ( 1 , 1 ) emission toward all the 55 objects .
on the other hand , the ccs @xmath4=@xmath5@xmath6 emission is not detected at all .
this survey indicates that the ccs emission is hardly detected in irdcs , as in the case of high - mass star - forming regions .
this is not a distance effect .
if the afgl 333 clump were located on the distance of 4.5 kpc , the peak temperature would be 0.29 k. this value is higher than the upper limits to the ccs peak intensity of all the observed objects , except for i18306 - 0835 mm1 , mm2 , and mm3 .
thus , we can detect the ccs line , if a chemically young massive clump like the afgl 333 clump would exist within the distance of 4.5 kpc .
the ch@xmath3oh @xmath2=76 emissions are detected toward 18 objects .
although the upper state energy of the ch@xmath3oh @xmath10=@xmath11@xmath12 @xmath13 line is as high as 65 k , the line is detected even in the objects without luminous heating sources seen in the msx 8 @xmath9 m image ( msx dark objects ) .
in contrast , the spitzer 24 @xmath9 m sources are associated with all the ch@xmath3oh - detected objects except for i18151 - 1208 mm2 , for which there are no available spitzer data .
thus , it is likely that all the objects with the ch@xmath3oh @xmath2=@xmath11@xmath12 @xmath13 emission harbor hot regions .
figure [ fig : f1]a shows the spectra of n@xmath0h@xmath1 @xmath2=10 , hc@xmath3n @xmath2=54 , ch@xmath3oh @xmath10=@xmath11@xmath12 @xmath13 , and ccs @xmath4=@xmath5@xmath6 for the selected objects , whereas figure [ fig : f2]a shows those of nh@xmath3 ( @xmath2 , @xmath7 ) = ( 1 , 1 ) , ( 2 , 2 ) and ( 3 , 3 ) . the spectra in figures [ fig : f1]a and [ fig : f2]a are `` typical ones '' taken arbitrarily from the spectra of our sample .
the observed spectra toward all the objects are shown in figures [ fig : f1]b e and [ fig : f2]b e , which are available in the electric edition of the journal .
note that the observed lines toward g024.08 + 00.04 mm2 have two velocity components ( 52 km s@xmath25 and 114 km s@xmath25 ) .
the n@xmath0h@xmath1 @xmath2=10 line shows complicated hyperfine splittings due to the nuclear spin ( @xmath34=1 ) of the nitrogen nuclei .
the observed line shows triplet structure , where the center and higher velocity components consist of three hyperfine lines each unresolved due to the velocity width .
the velocity width , optical depth , and excitation temperature of n@xmath0h@xmath1 are derived by fitting the observed spectrum pattern to a multi gaussian function ( see appendix b in sakai et al .
the results are given in table [ tab : t4 ] .
we fit the spectra of the hc@xmath3n and ch@xmath3oh lines to a single gaussian function , and these results are summarized in tables [ tab : t5 ] and [ tab : t6 ] , respectively .
we represent the upper limits to peak temperature and integrated intensity for the non - detected lines in the tables .
the upper limits to peak temperature for the non - detected lines are a 3@xmath33 value at the velocity resolution of one channel width .
the upper limits to integrated intensity for the non - detected lines are derived by using the following equation ; @xmath35 , where @xmath36 is a channel width in km s@xmath25 , and @xmath37 is a rms noise level at the velocity resolution of @xmath36 . as for the nh@xmath3 ( 1 , 1 ) line , five hyperfine components are fitted to the single gaussian each , and the parameters of the main hyperfine components are listed in table [ tab : t7 ] . on the other hand , the nh@xmath3
( 2 , 2 ) and ( 3 , 3 ) lines appear as a single line without hyperfine structure , and hence , they are fitted to a single gaussian function each ( tables [ tab : t8 ] and [ tab : t9 ] ) .
note that the nh@xmath3 ( 1 , 1 ) and ( 2 , 2 ) lines show a double - peak feature in g024.33 + 00.11 mm3 and mm9 . since the double peak structure is seen even in the weak hyperfine components , it seems to be a velocity structure .
then , we fit the spectra to a double gaussian function .
the parameters for each peak are listed in tables [ tab : t7 ] and [ tab : t8 ] .
as mentioned earlier , we found that the lsr velocities of 15 objects including g024.33 + 00.11 are much different from the values in the lists of rathborne et al .
this is because they estimated the systemic velocity of the objects from the @xmath27co @xmath2=10 line .
the critical density of n@xmath0h@xmath1 @xmath2=10 is much higher than that of @xmath27co @xmath2=10 , and hence , the n@xmath0h@xmath1 @xmath2=10 emission well traces the dense clumps .
we re - evaluate the kinematic distance of the dense clumps from the lsr velocity of the n@xmath0h@xmath1 or nh@xmath3 lines for all the observed objects by using the rotation curve obtained by clemens ( 1985 ) with ( @xmath38 , @xmath39 ) = ( 220 km s@xmath25 , 8.5 kpc ) .
the derived distances are listed in table [ tab : t4 ] .
figure [ fig : f3 ] shows the spectra covering the ch@xmath3oh @xmath10=@xmath40@xmath41 transitions toward the ch@xmath3oh - detected objects .
two strong lines are seen around 338.4 ghz ; these are the @xmath10=@xmath14@xmath15 @xmath16 and @xmath10=@xmath11@xmath12 @xmath13 lines .
in addition , higher excitation lines ( @xmath11@xmath12 @xmath16 and @xmath42@xmath43 @xmath16 ) are seen toward most of the sources .
the ch@xmath3oh @xmath10=@xmath28@xmath29 @xmath13 and @xmath10=@xmath30@xmath31 @xmath16 lines from the other side band are also observed .
although these two lines are lower-@xmath2 lines , the intensities of these lines are generally lower than those of the @xmath10=@xmath11@xmath12 line because of their lower line strength . therefore , the two lower-@xmath2 ch@xmath3oh lines are not detected toward the objects without ch@xmath3oh @xmath10=@xmath11@xmath12 @xmath13 emission
the strongest ch@xmath3oh source is g034.43 + 00.24 mm1 , where even the @xmath10=@xmath44@xmath45 lines ( @xmath46 = 112127 k ) are detected . the h@xmath0cs 10@xmath479@xmath48 line ( @xmath46 = 102 k ) is detected toward three objects , g024.33 + 00.11 mm1 , g034.43 + 00.24 mm1 , and i18182 - 1433 mm1 , which seem to contain hot core activities .
in fact , rathborne et al .
( 2007 ) reported the detection of many molecular lines , such as hcooch@xmath3 and ch@xmath3ch@xmath0cn , toward g024.33 + 00.11 mm1 .
since large organic molecules are thought to be formed from molecules evaporated from dust grains , such as c@xmath0h@xmath0 , hcn , and ch@xmath3oh , in the hot gas phase ( e.g. nomura & millar , 2004 ) , the detections of hcooch@xmath3 and ch@xmath3ch@xmath0cn definitively indicate the hot core activity . in figure [
fig : f4 ] , we plot the integrated intensities of a few observed lines against the peak flux of the 1.2 mm continuum , @xmath49 , reported by beuther et al . ( 2002 ) and rathborne et al .
we plot only the values of the objects with @xmath50 @xmath51 4.5 kpc .
the mm - wave dust continuum emission is optically thin in general , and it is a good tracer of column density of a dense region . a correlation between the n@xmath0h@xmath1 @xmath2=10 emission and the dust continuum emission was suggested in several regions ( caselli et al .
1999 , 2002a ; bergin et al .
2001 ; sakai et al .
2007 ) , although it was also reported that the n@xmath0h@xmath1 emission does not always trace the structures by dust continuum emission in very active star forming regions , such as orion kl ( ungerechts et al . 1998 ) . in irdcs
, the n@xmath0h@xmath1 emission is correlated with the 1.2 mm peak flux , as shown in figure [ fig : f4 ] , confirming that the n@xmath0h@xmath1 emission well traces the column density of a dense regions .
the hc@xmath3n integrated intensity is also roughly correlated with the 1.2 mm peak flux , suggesting that the hc@xmath3n emission also comes from a dense region , as well as the n@xmath0h@xmath1 emission .
the ch@xmath3oh @xmath10=@xmath11@xmath12 @xmath13 line is only detected toward the objects where @xmath49 is higher than 100 mjy . by using the relation reported by beuther et al .
( 2005a ) , @xmath49 of 100 mjy corresponds to @xmath52 of 49 , 68 , and 114 at @xmath53 of 40 , 30 , and 20 k , respectively .
furthermore , all the objects with @xmath49 @xmath20 130 mjy are the spitzer sources .
hence , the appearance of the ch@xmath3oh @xmath2=76 emission is related to the star formation activity .
a marginal correlation can be seen between the nh@xmath3 integrated intensities and the 1.2 mm peak flux .
the nh@xmath3 integrated intensities are scattered even at the low @xmath49 values , probably because the nh@xmath3 emission traces less dense gas due to its lower critical density ( @xmath17 10@xmath54 @xmath55 ) , and also because the beam size of the nh@xmath3 observation ( 73@xmath32 ) is larger than those of the other observations ( 18@xmath3237@xmath32 ) .
we derive the nh@xmath3 rotation temperature , @xmath56(nh@xmath3 ) , from the nh@xmath3 ( 1 , 1 ) and ( 2 , 2 ) lines by using the method presented by ho & townes ( 1983 ) and li et al .
we first evaluate the optical depth of the nh@xmath3 ( 1 , 1 ) line from the peak intensity ratio of the main and satellite hyperfine components .
then we calculate the rotation temperature from the peak intensity ratio of the main hyperfine components of the ( 1 , 1 ) and ( 2 , 2 ) lines considering the optical depth of the ( 1 , 1 ) line obtained above .
the derived @xmath56(nh@xmath3 ) values are listed in table [ tab : t10 ] .
a range of @xmath56(nh@xmath3 ) is from 10.3 to 20.8 k with the median of 1416 k. this is higher than the corresponding values of @xmath56(nh@xmath3 ) reported for dark clouds ( 1015 k ; benson & myers , 1989 ) .
figure [ fig : f5 ] represents a histogram of @xmath56(nh@xmath3 ) for the objects with @xmath50 @xmath51 4.5 kpc .
figure [ fig : f5 ] shows that @xmath56(nh@xmath3 ) for the msx sources tends to be higher than that for the msx dark objects .
this confirms that the emitting regions are associated with the msx sources .
the @xmath56(nh@xmath3 ) values of the msx and spitzer dark objects tend to be lower than those of the other objects .
the average @xmath56(nh@xmath3 ) value of the msx and spitzer dark objects is [email protected] k , which is comparable to that for the other samples of the msx - dark irdcs ( 13.9 k : pillai et al .
2006b ; 15.3 k : sridharan et al .
2005 ) , but is lower than that for the high - mass protostellar objects ( 22.5 k : sridharan et al .
2002 ) .
the rotation temperatures of ch@xmath3oh , @xmath56(ch@xmath3oh ) , are derived by using the rotation diagram method ( e.g. turner 1991 ) under the assumption of local thermodynamic equilibrium ( lte ) .
since several e type lines are detected in many objects ( figure [ fig : f3 ] ) , we derive the rotation temperature from them .
the results are listed in table [ tab : t11 ] .
the derived @xmath56(ch@xmath3oh ) values range from 7 to 12 k. the range is lower than that for the high - mass star forming cores ( 24203 k reported by van der tak et al .
( 2000b ) , if we exclude one exception of gl 7009s ( 8 k ) ) .
the average value of @xmath56(ch@xmath3oh ) in the msx dark objects is [email protected] k , which seems to be slightly lower than that of the msx sources ( 10@xmath572 k ) .
although the upper state energy of the ch@xmath3oh @xmath2=@xmath14@xmath15 @xmath16 line is 70 k , the derived temperatures are less than 12 k , which are even lower than @xmath56(nh@xmath3 ) .
this may suggest that the rotational levels of ch@xmath3oh are not thermalized to the gas kinetic temperature .
leurini et al .
( 2007 ) modeled the ch@xmath3oh emission toward several irdcs , and they derived the gas kinetic temperatures for an extended component , an outflow , and a core to be 24@xmath577 , 15@xmath575 , and 47@xmath579 k , respectively .
the @xmath56(ch@xmath3oh ) values derived in the present study are lower than their values , implying that the ch@xmath3oh lines are subthermally excited .
we derive the column densities of the observed molecules . for simplicity , we assume the lte condition for all the molecules . as for n@xmath0h@xmath1 , we derive the column density by assuming that the excitation temperature is equal to the rotation temperature of nh@xmath3 , and the lines are optically thin
. this can be justified , because @xmath58(n@xmath0h@xmath1 ) is similar to @xmath58(nh@xmath3 ) , as mentioned in the next section .
the derived column densities are listed in table [ tab : t10 ] . in the case of g034.43
+ 00.24 mm1 , a change in @xmath59 by @xmath571 k results in a change in the derived column density by @xmath574 % .
most of the n@xmath0h@xmath1 column densities are found to be larger than 1@xmath6010@xmath27 @xmath61 .
this is higher than a typical value found in low mass cores , which is several times 10@xmath62 @xmath61 ( benson et al .
1998 ; caselli et al .
2002b ) .
the upper limit to the ccs column density is derived by assuming the optically thin condition .
the rotation temperature of nh@xmath3 is employed as the excitation temperature .
the derived upper limits are listed in table [ tab : t10 ] .
they are found to be lower than 7@xmath6010@xmath62 @xmath61 for all the objects . these are lower than the ccs column densities of some low - mass starless cores in the dark clouds , typically a few times 10@xmath27 @xmath61 ( suzuki et al . 1992 ; benson et al . 1998 ) , and are also lower than the ccs column density of the massive clump in the afgl 333 cloud , which is ( 1.23.0)@xmath6010@xmath27 @xmath61 ( sakai et al . 2007 ) .
table [ tab : t10 ] lists the upper limit to the [ ccs]/[n@xmath0h@xmath1 ] ratio .
most of the ratios are lower than 1 even in the spitzer dark objects , whereas the corresponding ratio is as high as 2.63.2 for some low - mass starless cores
( benson et al .
since we use the [ ccs]/[n@xmath0h@xmath1 ] ratio to discuss the chemical evolution , the distance effect would be cancelled .
it is therefore suggested that the massive clumps associated with the irdcs are chemically evolved than low mass starless cores .
this survey demonstrates that the afgl 333 clump could be a rare case of a chemically young massive clump .
the hc@xmath3n column density is evaluated by assuming the optically thin condition , where the excitation temperature is assumed to be equal to @xmath56(nh@xmath3 ) . in the case of g034.43
+ 00.24 mm1 , a change in @xmath59 by @xmath571 k results in a change in the derived column density by @xmath573 % .
in contrast to ccs , the derived hc@xmath3n column densities ( table [ tab : t10 ] ) are comparable to those for the low - mass starless cores , @xmath1710@xmath63 @xmath61(suzuki et al .
1992 ) , even in the objects with msx sources , probably because hc@xmath3n is less sensitive to chemical evolution than ccs .
in addition , hc@xmath3n could also be formed in a hot gas .
in fact , strong hc@xmath3n emission is detected toward hot core sources like orion kl ( ungerechts et al .
1998 ) .
the nh@xmath3 column density is derived from the optical depth and rotation temperature ( section [ sec : temp ] ) .
the derived column densities are listed in table [ tab : t10 ] .
the nh@xmath3 column densities ( a few times 10@xmath64 @xmath61 ) are higher than those of the low mass cores , @xmath1710@xmath65 @xmath61 ( suzuki et al .
1992 ) , and are comparable to those of the irdcs observed by pillai et al .
( 2006b ) . the @xmath66(ccs)/@xmath66(nh@xmath3 ) ratios are found to be lower than 0.007 for most cases , which are lower than those in some low - mass dark cloud cores ( @xmath200.01 ) ( suzuki et al .
along with the results of the @xmath66(ccs)/@xmath66(n@xmath0h@xmath1 ) ratio , this further supports that the observed irdcs are chemically evolved .
we derive the ch@xmath3oh column density from the integrated intensity of the ch@xmath3oh @xmath10=@xmath11@xmath12 @xmath13 line . in this calculation , @xmath56(ch@xmath3oh )
is not used as the excitation temperature , because the rotational levels of ch@xmath3oh are not thermalized as mentioned before .
@xmath56(ch@xmath3oh ) is an effective temperature determined by the population difference between the different @xmath7 ladders , and does not always represent the excitation temperature of the @xmath11@xmath12 @xmath13 line .
therefore , we assume the excitation temperature ranging from 20 k to 50 k by referring the result by leurini et al .
we also estimate the upper limit to the ch@xmath3oh column density for the non - detected objects with adopting the rotation temperature of 20 k. the derived ch@xmath3oh column densities ranges from 2.9@xmath6010@xmath67 to 4.9@xmath6010@xmath64 @xmath61 toward the ch@xmath3oh - detected objects , as shown in table [ tab : t10 ] .
they are comparable to those in the irdcs obtained by leurini et al .
( 2007 ) , although they tend to be lower than those in several irdcs obtained by beuther & sridharan ( 2007 ) ( 1.7@[email protected]@xmath6010@xmath67 @xmath61 ) .
in addition , they are also comparable to those in the massive star forming regions obtained by van der tak et al .
( 2000b ) . on the other hand , maret et al .
( 2005 ) reported that the ch@xmath3oh column densities ranges from 5.3@xmath6010@xmath27 to 8.1@xmath6010@xmath67 @xmath61 in the low - mass star forming regions , ngc 1333 , l1448 , l1157 , and iras16293 - 2422 .
thus , the ch@xmath3oh column densities tend to be higher than those in the low - mass star forming regions .
velocity widths give information about turbulence of clumps including effects of star formation activity . in figure
[ fig : f6 ] , we plot the velocity widths of the hc@xmath3n @xmath2=54 , ch@xmath3oh @xmath10=@xmath11@xmath12 @xmath13 , nh@xmath3 ( @xmath2 , @xmath7 ) = ( 1 , 1 ) , ( 2 , 2 ) and ( 3 , 3 ) lines against that of the n@xmath0h@xmath1 @xmath2=10 line for the objects with @xmath50 @xmath51 4.5 kpc . the velocity widths of the hc@xmath3n , nh@xmath3 ( 1 , 1 ) and ( 2 , 2 ) lines are correlated with that of n@xmath0h@xmath1 , indicating that they come from almost the same region . on the other hand ,
the correlation is poor for high excitation lines such as ch@xmath3oh ( @xmath10=@xmath11@xmath12 @xmath13 and @xmath14@xmath15 @xmath16 ) and nh@xmath3 ( 3 , 3 ) . since the upper state energies of ch@xmath3oh @xmath2=@xmath40@xmath41 ( 65 k ) is higher than that of n@xmath0h@xmath1 @xmath2=10 ( 4 k ) and that of nh@xmath3 ( 3 , 3 ) is as high as 124.5 k , the ch@xmath3oh @xmath2=76 and nh@xmath3 ( 3 , 3 ) lines would trace the denser and hotter region near the protostar . in fact , the velocity widths of the ch@xmath3oh @xmath2=76 and nh@xmath3 ( 3 , 3 ) lines are correlated with each other ( figure [ fig : f7 ] ) .
pillai et al .
( 2006a ) also found that the nh@xmath3 ( 3 , 3 ) velocity width is larger than the nh@xmath3 ( 1 , 1 ) and ( 2 , 2 ) velocity widths toward the irdc , g11.11 - 0.12 .
beuther et al .
( 2002 ) reported a similar trend in the c@xmath68s and nh@xmath3 velocity widths , where the c@xmath68s @xmath2=21 velocity widths are about 1.5 times larger than the nh@xmath3 ( 1 , 1 ) velocity widths . since the c@xmath68s emission is optically thin and the critical density of c@xmath68s is as high as @xmath69 @xmath55 , it also traces dense regions .
furthermore , cs could be abundant in the hot gas phase in contrast to n@xmath0h@xmath1 ( e.g. nomura & millar 2004 ) .
thus it is likely that the c@xmath68s emission comes from regions close to the newly born stars , as well as the ch@xmath3oh and nh@xmath3 ( 3 , 3 ) emissions .
figure [ fig : f8 ] shows histograms of the velocity width of n@xmath0h@xmath1 for the msx dark objects and the objects with msx sources .
the mean n@xmath0h@xmath1 velocity width of the msx dark objects ( 2.3 km s@xmath25 ) is almost comparable to that of the objects with msx sources ( 2.5 km s@xmath25 ) .
these values are also comparable to that of the high - mass star forming regions observed by pirogov et al .
( 2003 ; 2.4 km s@xmath25 ) .
the histogram shapes are also similar to each other . since the n@xmath0h@xmath1 emission arises from the whole clump , it is not sensitive to the star formation activities .
the mean n@xmath0h@xmath1 velocity widths of the msx and spitzer dark objects ( 2.1 km s@xmath25 ) are significantly larger than those of the low - mass starless cores ( @xmath170.5 km s@xmath25 ) ( benson et al .
1998 ) , and are also larger than those of the cores in a giant molecular cloud , orion a ( tatematsu et al . 1993 ) .
the massive clumps associated with the irdcs seem to be initially turbulent .
this gives an important constraint when one would construct a model of star - formation processes in irdcs .
for instance , mckee & tan ( 2003 ) argued that massive stars can form in turbulence - supported cores with a high accretion rate .
thus , the large velocity widths observed in irdcs are consistent with the theoretical prediction , if the irdcs are birthplaces of high - mass stars .
as shown in figure [ fig : f6 ] , the ch@xmath3oh @xmath2=76 velocity widths for the msx dark objects tend to be broader than those for the objects with msx sources at a given n@xmath0h@xmath1 velocity width .
this trend is clear in the case of g034.43 + 00.24 .
figure [ fig : f9 ] shows the spitzer 24 @xmath9 m image toward the g034.43 + 00.24 region and the spectra of the ch@xmath3oh @xmath10=@xmath11@xmath12 @xmath13 and n@xmath0h@xmath1 @xmath2=10 @xmath70=01 @xmath71=12 lines toward g034.43 + 00.24 mm1 , mm2 and mm3 . in the spitzer image , g034.43 + 00.24 mm3 , which is a msx dark object , is less luminous than mm1 and mm2 .
it is known that a ultracompact hii region is associated with mm2 ( miralles et al .
1994 ; shepherd et al .
thus , mm2 is thought to be most evolved among three objects , whereas mm3 is thought to be youngest . in figure
[ fig : f9 ] , the ch@xmath3oh spectrum of g034.43 + 00.24 mm3 is clearly broader than the n@xmath0h@xmath1 spectrum for g034.43 + 00.24 mm3 . on the other hand ,
the ch@xmath3oh velocity width is comparable to the n@xmath0h@xmath1 velocity width toward g034.43 + 00.24 mm1 and mm2 .
the ch@xmath3oh linewidths toward these two sources are relatively narrower than that toward mm3 .
as seen in figure [ fig : f9 ] , the line shape of the ch@xmath3oh emission sometimes differs from the gaussian shape .
it may consist of the broad wing component and the narrow component .
in the present analysis , we fit all the ch@xmath3oh spectra with a single gaussian , so that the derived velocity widths have to be recognized as `` effective '' values for some sources where the line shape is much deviated from the single gaussian shape .
we employ the single gaussian fitting rather than the double gaussian fitting , because the latter fitting is difficult for most of the objects due to the limited s / n ratio .
nevertheless , the `` effective '' line width reflects the presence of the broad component , and hence , our discussion on the basis of the line width would be justified at least qualitatively .
van der tak et al .
( 2000a , 2000b ) carried out multi - line observations of ch@xmath3oh , c@xmath68s , and c@xmath72o toward several massive star - forming regions , and found that the velocity widths of the ch@xmath3oh lines are broader than those of the c@xmath68s and c@xmath72o lines for the less luminous objects , whereas they are comparable in the luminous objects . van der tak et al .
( 2000b ) argued that shock by the interaction between an outflow and an ambient cloud is important as the origin of the gas - phase ch@xmath3oh enhancement in low - luminosity objects , while radiation heating is important in high - luminosity objects .
in fact , the broad ch@xmath3oh emission are reported for such an interacting region in several low - mass star forming regions , including the famous case of l1157 , where a clear interaction between an outflow and an ambient cloud is observed as broad lines of ch@xmath3oh , nh@xmath3 , and hc@xmath3n ( umemoto et al .
1992 ; mikami et al . 1992 ; bachiller & prez gutirrez 1997 ; umemoto et al .
in contrast , the n@xmath0h@xmath1 line is not detected in the interaction region .
therefore , the ch@xmath3oh @xmath10=@xmath11@xmath12 @xmath13 emission would well trace the shocked region caused by the interaction between an outflow and an ambient cloud in the case of the msx dark objects . furthermore
, the broader line width can be seen in the observed lower-@xmath2 lines , @xmath10=@xmath28@xmath29 @xmath13 and @xmath30@xmath73 @xmath16 .
this may imply that these ch@xmath3oh emissions do not come from the quiescent and extended regions . to understand the properties of ch@xmath3oh in the quiescent gas
, we should observe the lower excitation transitions at the millimeter - wave region , since the critical density is proportional to @xmath74 .
our result that the line width of the ch@xmath3oh line is broader in the msx dark object means that the interaction between an outflow and an ambient cloud is important in the early stage of high - mass star formation .
* we have surveyed the n@xmath0h@xmath1 @xmath2=10 , hc@xmath3n @xmath2=54 , ccs @xmath4=@xmath5@xmath6 , ch@xmath3oh @xmath2=76 , nh@xmath3 ( j , k ) = ( 1 , 1 ) , ( 2 , 2 ) , and ( 3 , 3 ) lines toward the massive clumps associated with the irdcs .
the n@xmath0h@xmath1 @xmath2=10 emission is detected toward 54 out of 55 objects , the hc@xmath3n @xmath2=54 emission toward 43 objects , the nh@xmath3 ( @xmath2 , @xmath7 ) = ( 1 , 1 ) emission toward all the 55 objects , the nh@xmath3 ( @xmath2 , @xmath7 ) = ( 2 , 2 ) emission toward 52 objects , and the nh@xmath3 ( @xmath2 , @xmath7 ) = ( 3 , 3 ) emission toward 34 objects .
on the other hand , the ccs emission is not detected at all .
the ch@xmath3oh @xmath2=76 emission is detected toward 18 out of 55 objects .
all the ch@xmath3oh detected objects are associated with the spitzer 24 @xmath9 m sources . * the mean nh@xmath3 rotation temperature of the msx and spitzer dark objects is 13.9 k , which tends to be lower than that of the objects with those of the msx sources ( 17.9 k ) . * the ccs column densities and the [ ccs]/[n@xmath0h@xmath1 ] and [ ccs]/[nh@xmath3 ] ratios in the massive clumps associated with the irdcs are lower than those in the low - mass starless cores .
thus , most of the massive clumps would be more chemically evolved than the low - mass starless cores .
this survey demonstrates that the massive clump in the afgl 333 region , which was found by sakai et al .
( 2006 , 2007 ) , could be very novel . * the mean n@xmath0h@xmath1 velocity widths are significantly broader than those of the low - mass starless cores even in the msx and spitzer dark objects , indicating that massive cores associated with the irdcs are initially turbulent .
this along with the chemical ages inferred from the [ ccs]/[n@xmath0h@xmath1 ] ratio would be useful to constrain a models of star formation processes in irdcs . *
the velocity widths of the ch@xmath3oh @xmath2=76 and nh@xmath3 ( 3 , 3 ) lines are broader than those of the other lines .
this seems to originate from the interaction between an outflow and an ambient cloud .
the ch@xmath3oh velocity widths of the msx dark objects are larger than those of the objects with msx sources .
this suggests that outflow interaction plays an important role in production of the gas phase ch@xmath3oh in the early stage of high - mass star formation .
we are grateful to the nro staff for excellent support in the 45 m observations .
we would like to thank the members of the aste team for supporting the aste observations .
the 45 m radio telescope is operated by nobeyama radio observatory , a branch of national astronomical observatory of japan .
the aste project is driven by nobeyama radio observatory , a branch of national astronomical observatory of japan , in collaboration with university of chile , and japanese institutes including university of tokyo , nagoya university , osaka prefecture university , ibaraki university , and kobe university .
aikawa , y. , ohashi , n. , inutsuka , s. , herbst , e. , takakuwa , s. 2001 , , 552 , 639 aikawa , y. , herbst , e. , roberts , h. , caselli , p. 2005
, , 620 , 330 anderson , t. , herbst , e. , de lucia , f. c. 1990 , , 72 , 797 bachiller , r. , prez gutirrez , m. 1997 , , 487 , l93 benson , p. j. , myers , p. c. 1989 , , 71 , 89 benson , p. j. , caselli , p. , myers , p. c. 1998 , , 506 , 743 bergin , e. a. , ciardi , d. r. , lada , c. j. , alves , j. , lada , e. a. 2001 , , 557 , 209 bergin , e. a. , tafalla , m. 2007 , , 45 , 339 beuther , h. , schilke , p. , menten , k. m. , motte , f. , sridharan , t. k. , wyrowski , f. 2002 , , 566 , 945 beuther , h. , schilke , p. , menten , k. m. , motte , f. , sridharan , t. k. , wyrowski , f. 2005a , , 633 , 535 beuther , h. , sridharan , t. k. , saito , m. 2005b , , 634 , l185 beuther , h. , steinacker , j. 2007 , , 656 , l85 beuther , h. , sridharan , t. k. 2007 , , 668 , 348 carey , s. j. , clark , f. o. , egan , m. p. , price , s. d. , shipman , r. f. , kuchar , t. a. 1998 , , 508 , 721 caselli , p. , myers , p. c. , thaddeus , p. 1995
, , 455 , l77 caselli , p. , walmsley , c. m. , tafalla , m. , dore , l. , myers , p. c. 1999 , , 523 , l165 caselli , p. , walmsley , c. m. , zucconi , a. , tafalla , m. , dore , l. , myers , p. c. 2002a , , 565 , 344 caselli , p. , benson , p. j. , myers , p. c. , tafalla , m. 2002b , , 572 , 238 clemens , d. p. 1985
, , 295 , 422 egan , m. p. , shipman , r. f. , price , s. d. , carey , s. j. , clark , f. o. , cohen , m. 1998 , , 494 , l199 ellingsen , s. p. 2006
, , 638 , 241 ezawa , h. , kawabe , r. , kohno , k. , yamamoto , s. 2004 , proc .
spie , 5489 , 763 jackson , j. m. , rathborne , j. m. , shah , r. y. , simon , r. , bania , t. m. , clemens , d. p. , chambers , e. t. , johnson , a. m. , dormody , m. , lavoie , r. , heyer , m. h. 2006 , , 163 , 145 hirota , t. , ito , t. , yamamoto , s. 2002 , , 565 , 359 hirota , t. , maezawa , h. , yamamoto , s. 2004 , , 617 , 399 hirota , t. , yamamoto , s. 2006 , , 646 , 258 ho , p. t. p. , townes , c. h. 1983 , , 21 , 239 kohno , k. 2005 , asp conf .
, 344 , 242 lai , s .-
p . , crutcher , r. m. 2000 , , 128 , 271 lee , h .- h . , herbst , e. , pineau des forts , g. , roueff , e. , le bourlot , j. 1996 , , 311 , 690 leurini , s. , schilke , p. , wyrowski , f. , menten , k. m. 2007 , , 466 , 215 li , d. , goldsmith , p. f. , menten , k. 2003 , , 587 , 262 maret , s. , ceccarelli , c. , tielens , a. g. g. m. , caux , e. , lefloch , b. , faure , a. , castets , a. , flower , d. r. 2005 , , 442 , 527 mckee , c. f. , tan , j. c. 2003 , , 585 , 850 mikami , h. , umemoto , t. , yamamoto , s. , saito , s. 1992 , , 392 , l87 miralles , m. p. , rodrguez , l. f. , scalise , e. 1992 , , 92 , 173 nomura , h. , millar , t. j. 2004 , , 414 , 409 prault , m. , omont , a. , simon , g. , sguin , p. , ojha , d. , blommaert , j. , felli , m. , gilmore , g. , et al .
1996 , , 315 , l165 pillai , t. , wyrowski , f. , menten , k. m. , krgel , e. 2006a , , 447 , 929 pillai , t. , wyrowski , f. , carey , s. j. , menten , k. m. 2006b , , 450 , 569 pillai , t. , wyrowski , f. , hatchell , j. , gibb , a. g. , thompson , m. a. 2007 , , 467 , 207 pirogov , l. , zinchenko , i. , caselli , p. , johansson , l. e. b. , myers , p. c. 2003 , , 405 , 639 purcell , c. r. , balasubramanyam , r. , burton , m. g. , walsh , a. j. , minier , v. , hunt - cunningham , m. r. , kedziora - chudczer , l. l. , longmore , s. n. , hill , t. , et al .
2006 , , 367 , 553 ragan , s. e. , bergin , e. a. , plume , r. , gibson , d. l. , wilner , d. j. , obrien , s. , hails , e. 2006 , 166 , 567 rathborne , j. m. , jackson , j. m. , chambers , e. t. , simon , r. , shipman , r. , frieswijk , w. 2005 , , 630 , l181 rathborne , j. m. , jackson , j. m. , simon , r. 2006 , , 641 , 389 rathborne , j. m. , simon , r. , jackson , j. m. 2007 , , 662 , 1082 sakai , t. , oka , t. , yamamoto , s. 2006 , , 649 , 268 sakai , t. , oka , t. , yamamoto , s. 2007 , , 662 , 1043 shepherd , d. s. , povich , m. s. , whitney , b. a. , robitaille , t. p. , nnberger , d. e. a. , bronfman , l. , stark , d. p. , indebetouw , r. , meade , m. r. , babler , b. l. 2007 , , 669 , 464 simon , r. , jackson , j. m. , rathborne , j. m. , chambers , e. t. 2006 , , 639 , 227 sridharan , t. k. , beuther , h. , schilke , p. , menten , k. m. , wyrowski , f. 2002 , , 566 , 931 sridharan , t. k. , beuther , h. , saito , m. , wyrowski , f. , schilke , p. 2005
, , 534 , l57 suzuki , h. , yamamoto , s. , ohishi , m. , kaifu , n. , ishikawa , s. , hirahara , y. , takano , s. 1992 , , 392 , 551 tatematsu , k. , umemoto , t. , kameya , o. , hirano , n. , hasegawa , t. , hayashi , m. , iwata , t. , kaifu , n. , mikami , h. , murata , y. , nakano , m. , nakano , t. , ohashi , n. , sunada , k. , takaba , h. , yamamoto , s. 1993 , , 404 , 643 teyssier , d. , hennebelle , p. , prault , m. 2002 , , 382 , 624 turner , b. e. 1991 , , 76 , 617 umemoto , t. , iwata , t. , fukui , y. , mikami , h. , yamamoto , s. , kameya , o. , hirano , n. 1992 , , 392 , l83 umemoto , t. , mikami , h. , yamamoto , s. , hirano , n. 1999 , , 525 , l105 ungerechts , h. , bergin , e. a. , goldsmith , p. f. , irvine , w. m. , schloerb , f. p. , snell , r. l. 1997 , , 482 , 245 van der tak , f. f. s. , van dishoeck , e. f. , evans , n. j. , ii ; blake , g. a. 2000a , , 537 , 283 van der tak , f. f. s. , van dishoeck , e. f. , caselli , p. 2000b
, , 361 , 327 yamamoto , s. , saito , s. , kawaguchi , k. , chikada , y. , suzuki , h. , kaifu , n. , ishikawa , s. , ohishi , m. 1990 , , 361 , 318 lrrrrrr g015.05 + 00.07 mm1 & 18 17 50.4 & -15 53 38 & d & d & 115 & 1 + g015.31 - 00.16 mm2 & 18 18 50.4 & -15 43 19 & d & s & 39 & 1 + g015.31 - 00.16 mm3 & 18 18 45.3 & -15 41 58 & d & d & 37 & 1 + g019.27 + 00.07 mm1 & 18 25 58.5 & -12 03 59 & d & s & 150 & 1 + g019.27 + 00.07 mm2 & 18 25 52.6 & -12 04 48 & d & s & 89 & 1 + g022.35 + 00.41 mm1 & 18 30 24.4 & -09 10 34 & d & s & 349 & 1 + g022.35 + 00.41 mm2 & 18 30 24.2 & -09 12 44 & s & s & 67 & 1 + g022.35 + 00.41 mm3 & 18 30 38.1 & -09 12 44 & & & 53 & 1 + g023.60 + 00.00 mm1 & 18 34 11.6 & -08 19 06 & & s & 375 & 1 + g023.60 + 00.00 mm2 & 18 34 21.1 & -08 18 07 & d & s & 272 & 1 + g023.60 + 00.00 mm3 & 18 34 10.0 & -08 18 28 & & & 81 & 1 + g023.60 + 00.00 mm4 & 18 34 23.0 & -08 18 21 & & & 68 & 1 + g023.60 + 00.00 mm7 & 18 34 21.1 & -08 17 11 & d & s & 58 & 1 + g024.08 + 00.04 mm1 & 18 34 57.0 & -07 43 26 & s & s & 219 & 1 + g024.08 + 00.04 mm2 & 18 34 51.1 & -07 45 32 & d & & 68 & 1 + g024.08 + 00.04 mm3 & 18 35 02.2 & -07 45 25 & d & d & 50 & 1 + g024.08 + 00.04 mm4 & 18 35 02.6 & -07 45 56 & d & d & 48 & 1 + g024.33 + 00.11 mm1 & 18 35 07.9 & -07 35 04 & s & s & 1199 & 1 + g024.33 + 00.11 mm2 & 18 35 34.5 & -07 37 28 & & & 117 & 1 + g024.33 + 00.11 mm3 & 18 35 27.9 & -07 36 18 & & & 96 & 1 + g024.33 + 00.11 mm4 & 18 35 19.4 & -07 37 17 & d & d & 90 & 1 + g024.33 + 00.11 mm5 & 18 35 33.8 & -07 36 42 & & & 79 & 1 + g024.33 + 00.11 mm6 & 18 35 07.7 & -07 34 33 & d & & 77 & 1 + g024.33 + 00.11 mm8 & 18 35 23.4 & -07 37 21 & & & 72 & 1 + g024.33 + 00.11 mm9 & 18 35 26.5 & -07 36 56 & s & s & 66 & 1 + g024.33 + 00.11 mm11 & 18 35 05.1 & -07 35 58 & d & d & 48 & 1 + g024.60 + 00.08 mm1 & 18 35 40.2 & -07 18 37 & & s & 279 & 1 + g024.60 + 00.08 mm2 & 18 35 35.7 & -07 18 09 & d & s & 230 & 1 + g025.04 - 00.20 mm1 & 18 38 10.2 & -07 02 34 & & s & 203 & 1 + g025.04 - 00.20 mm2 & 18 38 17.7 & -07 02 51 & d & s & 92 & 1 + g025.04 - 00.20 mm4 & 18 38 13.7 & -07 03 12 & d & d & 82 & 1 + g034.43 + 00.24 mm1 & 18 53 18.0 & 01 25 24 & & s & 2228 & 1 + g034.43 + 00.24 mm2 & 18 53 18.6 & 01 24 40 & s & s & 964 & 1 + g034.43 + 00.24 mm3 & 18 53 20.4 & 01 28 23 & d & s & 244 & 1 + g034.43 + 00.24 mm4 & 18 53 19.0 & 01 24 08 & & s & 221 & 1 + g034.43 + 00.24 mm5 & 18 53 19.8 & 01 23 30 & & & 122 & 1 + g034.43 + 00.24 mm6 & 18 53 18.6 & 01 27 48 & d & d & 57 & 1 + g034.43 + 00.24 mm8 & 18 53 16.4 & 01 26 20 & d & d & 51 & 1 + g034.43 + 00.24 mm9 & 18 53 18.4 & 01 28 14 & d & d & 50 & 1 + i18102 - 1800 mm1 & 18 13 11.0 & -17 59 59 & & s & 316 & 2,3,4 + i18151 - 1208 mm1 & 18 17 58.0 & -12 07 27 & s & no data & 672 & 2,3,4 + i18151 - 1208 mm2 & 18 17 50.4 & -12 07 55 & d & no data & 424 & 2,3,4 + i18151 - 1208 mm3 & 18 17 52.2 & -12 06 56 & s & no data & 149 & 2,3,4 + i18182 - 1433 mm1 & 18 21 09.2 & -14 31 57 & s & s & 1303 & 2,3,4 + i18182 - 1433 mm2 & 18 21 14.9 & -14 33 06 & & s & 100 & 2,3,4 + i18223 - 1243 mm1 & 18 25 10.5 & -12 42 26 & s & s & 328 & 2,3,4 + i18223 - 1243 mm2 & 18 25 09.5 & -12 44 15 & & s & 124 & 2,3,4 + i18223 - 1243 mm3 & 18 25 08.3 & -12 45 28 & d & s & 205 & 2,3,4 + i18223 - 1243 mm4 & 18 25 07.2 & -12 47 54 & d & s & 52 & 2,3,4 + i18306 - 0835 mm1 & 18 33 24.0 & -08 33 31 & s & s & 731 & 2,3,4 + i18306 - 0835 mm2 & 18 33 17.2 & -08 33 26 & & s & 212 & 2,3,4 + i18306 - 0835 mm3 & 18 33 32.1 & -08 32 29 & d & d & 103 & 2,3,4 + i18337 - 0743 mm1 & 18 36 41.0 & -07 39 20 & s & s & 485 & 2,3,4 + i18337 - 0743 mm2 & 18 36 27.7 & -07 40 28 & s & s & 180 & 2,3,4 + i18337 - 0743 mm3 & 18 36 18.2 & -07 41 01 & d & s & 110 & 2,3,4 + [ tab : t1 ] llrrrrr ccs & @xmath4=@xmath5@xmath6 & 45.379033 & 31.3 & 5.4 & 1 & nro
45 m + hc@xmath3n & @xmath2=54 & 45.490302 & 69.2 & 6.5 & 2 & nro 45 m + n@xmath0h@xmath1 & @xmath2=10 & 93.173777 & 11.6 & 4.47 & 3 & nro 45 m + nh@xmath3 & ( @xmath2 , @xmath7 ) = ( 1 , 1 ) & 23.694495 & 3.24 & 23.4 & 4 & nro 45 m + nh@xmath3 & ( @xmath2 , @xmath7 ) = ( 2 , 2 ) & 23.722633 & 7.20 & 64.9 & 4 & nro 45 m + nh@xmath3 & ( @xmath2 , @xmath7 ) = ( 3 , 3 ) & 23.870129 & 11.3 & 124.5 & 4 & nro 45 m + ch@xmath3oh & @xmath10 = @xmath11@xmath12 @xmath16 & 338.124502 & 5.480 & 77.1 & 5 & aste + ch@xmath3oh & @xmath10 = @xmath14@xmath15 @xmath16 & 338.344628 & 5.375 & 69.6 & 5 & aste + ch@xmath3oh & @xmath10 = @xmath76@xmath77 @xmath13 & 338.408681 & 5.484 & 65.0 & 5 & aste + ch@xmath3oh & @xmath10 = @xmath78@xmath79 @xmath80 & 338.512762 & 5.059 & 86.4 & 5 & aste + ch@xmath3oh & @xmath10 = @xmath81@xmath82 @xmath13 & 338.540795 & 4.457 & 114.8 & 5 & aste + ch@xmath3oh & @xmath10 = @xmath81@xmath82 @xmath80 & 338.543204 & 4.457 & 114.8 & 5 & aste + ch@xmath3oh & @xmath10 = @xmath83@xmath84 @xmath16 & 338.559928 & 4.503 & 126.8 & 5 & aste + ch@xmath3oh & @xmath10 = @xmath81@xmath82 @xmath16 & 338.583195 & 4.486 & 111.8 & 5 & aste + ch@xmath3oh & @xmath10 = @xmath85@xmath86 @xmath16 & 338.614999 & 5.508 & 85.1 & 5 & aste + ch@xmath3oh & @xmath10 = @xmath78@xmath79 @xmath13 & 338.639939 & 5.509 & 102.7 & 5 & aste + ch@xmath3oh & @xmath10 = @xmath78@xmath79 @xmath16 & 338.721630 & 4.980 & 86.3 & 5 & aste + ch@xmath3oh & @xmath10 = @xmath87@xmath88 @xmath16 & 338.722940 & 5.046 & 90.0 & 5 & aste + ch@xmath3oh & @xmath10 = @xmath89@xmath31 @xmath16 & 350.687730 & 1.553 & 35.4 & 5 & aste + ch@xmath3oh & @xmath10 = @xmath90@xmath91 @xmath13 & 350.905119 & 1.977 & 16.8 & 5 & aste + [ tab : t2 ] llrrrrrrr dark&dark & 11/11&7/11 & 0/11 & 0/11 & 11/11 & 10/11&3/11 + dark&source&12/12&12/12 & 0/12&6/12&12/12&11/12&8/12 + dark&others&3/3&2/3&0/3&1/3&3/3&3/3&2/3 + source&source&12/12&9/12&0/12&6/12&12/12&11/12&9/12 + others&source&9/9&7/9&0/9&5/9&9/9&9/9&7/9 + others&others&7/8&6/8&0/8&0/8&8/8&8/8&5/8 + total&&54/55&43/55&0/55&18/55&55/55&52/55&34/55 + [ tab : t3 ] lrrrrrr g015.05 + 00.07 mm1 & 5.1 ( 0.2 ) & 2.5 ( 0.4 ) & 24.73 ( 0.03 ) & 2.79 ( 0.08 ) & 6.0 ( 0.1 ) & 2.5 + g015.31 - 00.16 mm2 & 4.0 ( 0.2 ) & 6.9 ( 2.0 ) & 31.12 ( 0.02 ) & 0.81 ( 0.07 ) & 2.3 ( 0.1 ) & 3.1 + g015.31 - 00.16 mm3 & 4.0 ( 0.1 ) & 11.4 ( 2.6 ) & 30.81 ( 0.02 ) & 0.62 ( 0.04 ) & 2.6 ( 0.1 ) & 3.0 + g019.27 + 0.07 mm1 & 15 ( 1 ) & 1.1 ( 0.1 ) & 26.76 ( 0.01 ) & 3.15 ( 0.03 ) & 17.7 ( 0.1 ) & 2.3 + g019.27 + 0.07 mm2 & 9.0 ( 0.6 ) & 1.5 ( 0.2 ) & 27.01 ( 0.02 ) & 2.62 ( 0.04 ) & 9.8 ( 0.1 ) & 2.3 + g022.35 + 00.41 mm1 & 5.9 ( 0.1 ) & 3.1 ( 0.3 ) & 52.73 ( 0.02 ) & 2.74 ( 0.05 ) & 8.9 ( 0.1 ) & 3.7 + g022.35 + 00.41 mm2 & 6 ( 2 ) & 1.50 ( 1.48 ) & 60.22 ( 0.02 ) & 0.83 ( 0.07 ) & 1.9 ( 0.1 ) & 4.1 + g022.35 + 00.41 mm3 & 5.0 ( 0.2 ) & 3.4 ( 0.5 ) & 83.97 ( 0.03 ) & 2.23 ( 0.07 ) & 5.4 ( 0.1 ) & 5.3 + g023.60 + 00.00 mm1 & 7.9 ( 0.5 ) & 1.5 ( 0.2 ) & 106.89 ( 0.06 ) & 3.82 ( 0.07 ) & 11.8 ( 0.1 ) & 6.7 + g023.60 + 00.00 mm2 & 5.8 ( 0.3 ) & 2.2 ( 0.3 ) & 53.34 ( 0.03 ) & 3.03 ( 0.07 ) & 7.6 ( 0.1 ) & 3.7 + g023.60 + 00.00 mm3 & 8 ( 3 ) & 0.7 ( 0.4 ) & 105.51 ( 0.04 ) & 3.4 ( 0.1 ) & 5.8 ( 0.1 ) & 6.6 + g023.60 + 00.00 mm4 & 13 ( 24 ) & 0.2 ( 0.6 ) & 53.87 ( 0.05 ) & 2.6 ( 0.1 ) & 3.0 ( 0.1 ) & 3.7 + g023.60 + 00.00 mm7 & 8 ( 1 ) & 1.3 ( 0.3 ) & 54.00 ( 0.05 ) & 1.90 ( 0.06 ) & 6.0 ( 0.1 ) & 3.7 + g024.08 + 00.04 mm1 & 49 ( 26 ) & 0.21 ( 0.13 ) & 113.63 ( 0.01 ) & 2.67 ( 0.03 ) & 12.8 ( 0.1 ) & 7.8 + g024.08 + 00.04 mm2(1 ) & 9.4 ( 8.5 ) & 0.6 ( 0.9 ) & 52.08 ( 0.03 ) & 1.5 ( 0.1 ) & 2.8 ( 0.1 ) & 3.6 + g024.08 + 00.04 mm2(2 ) & 7 ( 3 ) & 0.8 ( 0.7 ) & 114.45 ( 0.04 ) & 2.1 ( 0.1 ) & 3.6 ( 0.1 ) & 7.8 + g024.08 + 00.04 mm3 & 3.5 ( 0.1 ) & 3.5 ( 1 ) & 51.86 ( 0.06 ) & 2.2 ( 0.2 ) & 2.0 ( 0.1 ) & 3.6 + g024.08 + 00.04 mm4 & 3.32 ( 0.04 ) & 19 ( 6 ) & 51.95 ( 0.03 ) & 0.85 ( 0.08 ) & 1.7 ( 0.1 ) & 3.6 + g024.33 + 00.11 mm1 & 7.6 ( 0.1 ) & 3.4 ( 0.2 ) & 113.76 ( 0.02 ) & 3.53 ( 0.04 ) & 19.0 ( 0.1 ) & 7.7 + g024.33 + 00.11 mm2 & 5.4 ( 0.3 ) & 2.3 ( 0.4 ) & 118.22 ( 0.04 ) & 3.23 ( 0.09 ) & 6.8 ( 0.1 ) & 7.7 + g024.33 + 00.11 mm3 & 8 ( 2 ) & 0.8 ( 0.3 ) & 117.78 ( 0.05 ) & 4.3 ( 0.1 ) & 7.9 ( 0.1 ) & 7.7 + g024.33 + 00.11 mm4 & 6 ( 2 ) & 0.9 ( 0.6 ) & 115.19 ( 0.07 ) & 3.6 ( 0.2 ) & 4.2 ( 0.1 ) & 7.7 + g024.33 + 00.11 mm5 & 4.69 ( 0.09 ) & 4.6 ( 0.5 ) & 117.26 ( 0.04 ) & 3.00 ( 0.08 ) & 7.6 ( 0.1 ) & 7.7 + g024.33 + 00.11 mm6 & 7 ( 2 ) & 0.9 ( 0.5 ) & 114.43 ( 0.04 ) & 2.97 ( 0.10 ) & 4.5 ( 0.1 ) & 7.7 + g024.33 + 00.11 mm8 & & & & & 0.9 ( 0.1 ) & 7.7 + g024.33 + 00.11 mm9 & 3.8 ( 0.2 ) & 4.2 ( 1.5 ) & 119.78 ( 0.05 ) & 1.7 ( 0.2 ) & 2.7 ( 0.1 ) & 7.7 + g024.33 + 00.11 mm11 & 4.3 ( 0.3 ) & 2.4 ( 0.9 ) & 112.60 ( 0.05 ) & 2.18 ( 0.05 ) & 3.0 ( 0.1 ) & 7.7 + g024.60 + 00.08 mm1 & 5.4 ( 0.1 ) & 3.9 ( 0.3 ) & 53.14 ( 0.03 ) & 3.10 ( 0.06 ) & 9.8 ( 0.1 ) & 3.6 + g024.60 + 00.08 mm2 & 10 ( 2 ) & 1.2 ( 0.2 ) & 115.16 ( 0.01 ) & 2.37 ( 0.04 ) & 9.3 ( 0.1 ) & 7.7 + g025.04 - 00.20 mm1 & 11 ( 1 ) & 1.1 ( 0.2 ) & 64.00 ( 0.02 ) & 2.36 ( 0.05 ) & 9.8 ( 0.1 ) & 4.2 + g025.04 - 00.20 mm2 & 6.5 ( 0.4 ) & 2.0 ( 0.3 ) & 63.83 ( 0.03 ) & 2.77 ( 0.07 ) & 8.1 ( 0.1 ) & 4.2 + g025.04 - 00.20 mm4 & 6.2 ( 0.5 ) & 1.9 ( 0.4 ) & 63.69 ( 0.02 ) & 2.18 ( 0.08 ) & 6.0 ( 0.1 ) & 4.1 + g034.43 + 00.24 mm1 & 39 ( 2 ) & 0.75 ( 0.05 ) & 57.77 ( 0.05 ) & 3.15 ( 0.01 ) & 40.5 ( 0.1 ) & 3.5 + g034.43 + 00.24 mm2 & 29 ( 2 ) & 0.67 ( 0.06 ) & 57.832 ( 0.009 ) & 4.04 ( 0.02 ) & 34.2 ( 0.1 ) & 3.5 + g034.43 + 00.24 mm3 & 36 ( 7 ) & 0.42 ( 0.09 ) & 59.686 ( 0.008 ) & 2.90 ( 0.02 ) & 20.3 ( 0.1 ) & 3.6 + g034.43 + 00.24 mm4 & 14.2 ( 0.5 ) & 1.5 ( 0.1 ) & 57.689 ( 0.008 ) & 2.81 ( 0.02 ) & 20.8 ( 0.1 ) & 3.5 + g034.43 + 00.24 mm5 & 9 ( 2 ) & 0.9 ( 0.3 ) & 58.12 ( 0.03 ) & 2.58 ( 0.08 ) & 6.0 ( 0.1 ) & 3.5 + g034.43 + 00.24 mm6 & 10 ( 2 ) & 0.8 ( 0.4 ) & 58.49 ( 0.02 ) & 2.07 ( 0.07 ) & 5.6 ( 0.1 ) & 3.6 + g034.43 + 00.24 mm8 & 6 ( 1 ) & 1.1 ( 0.5 ) & 57.63 ( 0.05 ) & 3.2 ( 0.1 ) & 4.5 ( 0.1 ) & 3.5 + g034.43 + 00.24 mm9 & 12 ( 3 ) & 0.7 ( 0.3 ) & 59.00 ( 0.02 ) & 2.45 ( 0.06 ) & 7.3 ( 0.1 ) & 3.6 + i18102 - 1800 mm1 & 9.3 ( 0.5 ) & 1.7 ( 0.2 ) & 22.36 ( 0.02 ) & 3.49 ( 0.05 ) & 15.9 ( 0.1 ) & 2.7 + i18151 - 1208 mm1 & 13 ( 2 ) & 0.8 ( 0.2 ) & 33.19 ( 0.01 ) & 2.10 ( 0.04 ) & 9.8 ( 0.1 ) & 2.8 + i18151 - 1208 mm2 & 14 ( 2 ) & 0.9 ( 0.2 ) & 29.74 ( 0.02 ) & 3.09 ( 0.04 ) & 14.2 ( 0.1 ) & 2.6 + i18151 - 1208 mm3 & 5.7 ( 0.2 ) & 4.9 ( 0.9 ) & 30.65 ( 0.01 ) & 0.87 ( 0.04 ) & 4.4 ( 0.1 ) & 2.7 + i18182 - 1433 mm1 & 30 ( 8) & 0.4 ( 0.1 ) & 60.00 ( 0.01 ) & 3.06 ( 0.04 ) & 16.6 ( 0.1 ) & 4.6 + i18182 - 1433 mm2 & 5.8 ( 0.4 ) & 2.3 ( 0.4 ) & 41.10 ( 0.03 ) & 2.28 ( 0.08 ) & 6.2 ( 0.1 ) & 3.6 + i18223 - 1243 mm1 & 14 ( 1 ) & 1.2 ( 0.2 ) & 45.32 ( 0.01 ) & 2.16 ( 0.03 ) & 13.4 ( 0.1 ) & 3.6 + i18223 - 1243 mm2 & 8.1 ( 0.7 ) & 1.3 ( 0.2 ) & 45.29 ( 0.02 ) & 2.99 ( 0.06 ) & 8.7 ( 0.1 ) & 3.6 + i18223 - 1243 mm3 & 9.4 ( 0.3 ) & 2.2 ( 0.2 ) & 45.73 ( 0.01 ) & 2.64 ( 0.03 ) & 14.5 ( 0.1 ) & 3.7 + i18223 - 1243 mm4 & 5.3 ( 0.3 ) & 2.4 ( 0.5 ) & 45.83 ( 0.03 ) & 1.97 ( 0.07 ) & 5.0 ( 0.1 ) & 3.7 + i18306 - 0835 mm1 & 27 ( 31 ) & 0.2 ( 0.3 ) & 78.02 ( 0.03 ) & 2.81 ( 0.07 ) & 7.2 ( 0.1 ) & 4.9 + i18306 - 0835 mm2 & 10 ( 11 ) & 0.4 ( 0.6 ) & 76.67 ( 0.05 ) & 2.6 ( 0.1 ) & 3.4 ( 0.1 ) & 4.9 + i18306 - 0835 mm3 & 5 ( 2 ) & 0.95 ( 1.5 ) & 53.98 ( 0.09 ) & 2.1 ( 0.3 ) & 1.6 ( 0.1 ) & 3.7 + i18337 - 0743 mm1 & 7.03 ( 0.08 ) & 5.9 ( 0.3 ) & 58.37 ( 0.02 ) & 3.35 ( 0.04 ) & 21.5 ( 0.1 ) & 3.9 + i18337 - 0743 mm2 & 8 ( 1 ) & 1.3 ( 0.4 ) & 58.60 ( 0.03 ) & 2.71 ( 0.08 ) & 6.9 ( 0.1 ) & 3.9 + i18337 - 0743 mm3 & 6.7 ( 0.6 ) & 1.7 ( 0.4 ) & 56.37 ( 0.03 ) & 2.44 ( 0.08 ) & 6.6 ( 0.1 ) & 3.8 + [ tab : t4 ] lrrrrrrr g015.05 + 00.07 mm1 & 0.25 ( 0.02 ) & 24.4 ( 0.1 ) & 3.5 ( 0.3 ) & 1.05 ( 0.09 ) & & @xmath920.22 & @xmath920.25 + g015.31 - 00.16 mm2 & 0.28 ( 0.03 ) & 30.92 ( 0.08 ) & 1.4 ( 0.2 ) & 0.61 ( 0.09 ) & & @xmath920.24 & @xmath920.28 + g015.31 - 00.16 mm3 & @xmath920.23 & & & 0.51 ( 0.09 ) & & @xmath920.25 & @xmath920.29 + g019.27 + 0.07 mm1 & 0.60 ( 0.02 ) & 26.39 ( 0.05 ) & 3.7 ( 0.1 ) & 2.60 ( 0.07 ) & & @xmath920.19 & @xmath920.23 + g019.27 + 0.07 mm2 & 0.64 ( 0.02 ) & 27.02 ( 0.04 ) & 2.9 ( 0.1 ) & 2.35 ( 0.07 ) & & @xmath920.20 & @xmath920.23 + g022.35 + 00.41 mm1 & 0.46 ( 0.02 ) & 53.00 ( 0.07 ) & 3.6 ( 0.2 ) & 1.94 ( 0.08 ) & & @xmath920.20 & @xmath920.23 + g022.35 + 00.41 mm2 & @xmath920.20 & & & @xmath920.24 & & @xmath920.21 & @xmath920.24 + g022.35 + 00.41 mm3 & 0.34 ( 0.03 ) & 84.36 ( 0.06 ) & 1.6 ( 0.1 ) & 0.59 ( 0.07 ) & & @xmath920.19 & @xmath920.22 + g023.60 + 00.00 mm1 & 0.57 ( 0.02 ) & 106.7 ( 0.06 ) & 3.6 ( 0.1 ) & 2.28 ( 0.08 ) & & @xmath920.23 & @xmath920.26 + g023.60 + 00.00 mm2 & 0.81 ( 0.02 ) & 53.37 ( 0.04 ) & 3.4 ( 0.1 ) & 3.35 ( 0.08 ) & & @xmath920.20 & @xmath920.23 + g023.60 + 00.00 mm3 & @xmath920.21 & & & 0.41 ( 0.08 ) & & @xmath920.23 & @xmath920.27 + g023.60 + 00.00 mm4 & 0.33 ( 0.02 ) & 53.91 ( 0.07 ) & 2.3 ( 0.2 ) & 0.76 ( 0.07 ) & & @xmath920.21 & @xmath920.25 + g023.60 + 00.00 mm7 & 0.55 ( 0.02 ) & 54.08 ( 0.05 ) & 2.1 ( 0.1 ) & 1.56 ( 0.08 ) & & @xmath920.21 & @xmath920.24 + g024.08 + 00.04 mm1 & 0.30 ( 0.02 ) & 113.68 ( 0.08 ) & 2.8 ( 0.2 ) & 0.89 ( 0.07 ) & & @xmath920.21 & @xmath920.24 + g024.08 + 00.04 mm2(1 ) & 0.22 ( 0.03 ) & 52.18 ( 0.07 ) & 0.9 ( 0.2 ) & 0.44 ( 0.07 ) & & @xmath920.21 & @xmath920.24 + g024.08 + 00.04 mm2(2 ) & @xmath920.19 & & & 0.38 ( 0.07 ) & & @xmath920.21 & @xmath920.24 + g024.08 + 00.04 mm3 & 0.39 ( 0.02 ) & 51.67 ( 0.04 ) & 1.6 ( 0.1 ) & 0.78 ( 0.07 ) & & @xmath920.24 & @xmath920.28 + g024.08 + 00.04 mm4 & 0.56 ( 0.03 ) & 51.90 ( 0.02 ) & 1.08 ( 0.07 ) & 0.83 ( 0.07 ) & & @xmath920.21 & @xmath920.24 + g024.33 + 00.11 mm1 & 0.71 ( 0.02 ) & 113.69 ( 0.06 ) & 4.0 ( 0.1 ) & 3.22 ( 0.09 ) & & @xmath920.28 & @xmath920.32 + g024.33 + 00.11 mm2 & 0.34 ( 0.02 ) & 118.1 ( 0.1 ) & 3.4 ( 0.2 ) & 1.27 ( 0.09 ) & & @xmath920.25 & @xmath920.29 + g024.33 + 00.11 mm3 & 0.24 ( 0.02 ) & 117.2 ( 0.2 ) & 5.1 ( 0.4 ) & 1.31 ( 0.09 ) & & @xmath920.27 & @xmath920.31 + g024.33 + 00.11 mm4 & @xmath920.23 & & & 1.11 ( 0.09 ) & & @xmath920.26 & @xmath920.30 + g024.33 + 00.11 mm5 & 0.34 ( 0.02 ) & 117.4 ( 0.1 ) & 4.0 ( 0.3 ) & 1.48 ( 0.09 ) & & @xmath920.23 & @xmath920.27 + g024.33 + 00.11 mm6 & @xmath920.24 & & & 0.73 ( 0.09 ) & & @xmath920.28 & @xmath920.32 + g024.33 + 00.11 mm8 & @xmath920.23 & & & @xmath920.27 & & @xmath920.29 & @xmath920.34 + g024.33 + 00.11 mm9 & 0.21 ( 0.03 ) & 119.4 ( 0.1 ) & 1.3 ( 0.2 ) & 0.50 ( 0.09 ) & & @xmath920.29 & @xmath920.33 + g024.33 + 00.11 mm11 & @xmath920.23 & & & @xmath920.27 & & @xmath920.28 & @xmath920.32 + g024.60 + 00.08 mm1 & 0.63 ( 0.02 ) & 53.22 ( 0.05 ) & 3.5 ( 0.1 ) & 2.59 ( 0.08 ) & & @xmath920.17 & @xmath920.20 + g024.60 + 00.08 mm2 & 0.29 ( 0.02 ) & 115.2 ( 0.1 ) & 3.0 ( 0.2 ) & 1.23 ( 0.09 ) & & @xmath920.28 & @xmath920.32 + g025.04 - 00.20 mm1 & 0.46 ( 0.02 ) & 63.67 ( 0.07 ) & 3.0 ( 0.2 ) & 1.67 ( 0.09 ) & & @xmath920.23 & @xmath920.26 + g025.04 - 00.20 mm2 & 0.46 ( 0.03 ) & 63.63 ( 0.07 ) & 2.4 ( 0.2 ) & 1.21 ( 0.09 ) & & @xmath920.29 & @xmath920.34 + g025.04 - 00.20 mm4 & 0.35 ( 0.03 ) & 63.76 ( 0.08 ) & 2.0 ( 0.2 ) & 0.67 ( 0.09 ) & & @xmath920.23 & @xmath920.27 + g034.43 + 00.24 mm1 & 1.22 ( 0.02 ) & 58.04 ( 0.03 ) & 3.85 ( 0.06 ) & 5.35 ( 0.06 ) & & @xmath920.18 & 0.35 ( 0.07 ) + g034.43 + 00.24 mm2 & 1.10 ( 0.01 ) & 57.86 ( 0.03 ) & 4.22 ( 0.06 ) & 5.09 ( 0.06 ) & & @xmath920.18 & 0.43 ( 0.07 ) + g034.43 + 00.24 mm3 & 0.57 ( 0.01 ) & 60.05 ( 0.07 ) & 5.8 ( 0.2 ) & 3.29 ( 0.07 ) & & @xmath920.18 & 0.41 ( 0.07 ) + g034.43 + 00.24 mm4 & 0.80 ( 0.02 ) & 57.50 ( 0.03 ) & 3.22 ( 0.08 ) & 3.03 ( 0.07 ) & & @xmath920.17 & 0.30 ( 0.07 ) + g034.43 + 00.24 mm5 & 0.31 ( 0.03 ) & 57.81 ( 0.08 ) & 1.8 ( 0.2 ) & 0.67 ( 0.08 ) & & @xmath920.22 & 0.38 ( 0.08 ) + g034.43 + 00.24 mm6 & 0.45 ( 0.02 ) & 58.34 ( 0.06 ) & 2.5 ( 0.1 ) & 1.25 ( 0.08 ) & & @xmath920.23 & @xmath920.26 + g034.43 + 00.24 mm8 & 0.23 ( 0.02 ) & 57.7 ( 0.2 ) & 4.4 ( 0.4 ) & 1.1 ( 0.08 ) & & @xmath920.21 & @xmath920.24 + g034.43 + 00.24 mm9 & 0.50 ( 0.02 ) & 58.40 ( 0.05 ) & 2.4 ( 0.1 ) & 1.31 ( 0.08 ) & & @xmath920.21 & @xmath920.24 + i18102 - 1800 mm1 & 0.83 ( 0.02 ) & 21.53 ( 0.05 ) & 4.4 ( 0.1 ) & 4.14 ( 0.09 ) & & @xmath920.29 & @xmath920.34 + i18151 - 1208 mm1 & 0.53 ( 0.03 ) & 33.26 ( 0.05 ) & 2.1 ( 0.1 ) & 1.39 ( 0.09 ) & & @xmath920.26 & @xmath920.3 + i18151 - 1208 mm2 & 0.25 ( 0.02 ) & 30.3 ( 0.1 ) & 3.2 ( 0.3 ) & 0.93 ( 0.10 ) & & @xmath920.27 & @xmath920.32 + i18151 - 1208 mm3 & @xmath920.24 & & & 0.39 ( 0.09 ) & & @xmath920.28 & @xmath920.32 + i18182 - 1433 mm1 & 1.04 ( 0.03 ) & 59.86 ( 0.03 ) & 2.58 ( 0.07 ) & 3.07 ( 0.09 ) & & @xmath920.26 & @xmath920.3 + i18182 - 1433 mm2 & @xmath920.25 & & & 0.39 ( 0.10 ) & & @xmath920.26 & @xmath920.3 + i18223 - 1243 mm1 & 0.44 ( 0.03 ) & 45.38 ( 0.06 ) & 2.1 ( 0.1 ) & 1.09 ( 0.08 ) & & @xmath920.22 & @xmath920.25 + i18223 - 1243 mm2 & 0.24 ( 0.03 ) & 45.5 ( 0.1 ) & 2.1 ( 0.3 ) & 0.68 ( 0.09 ) & & @xmath920.21 & @xmath920.25 + i18223 - 1243 mm3 & 0.44 ( 0.02 ) & 45.65 ( 0.07 ) & 3.2 ( 0.2 ) & 1.59 ( 0.08 ) & & @xmath920.24 & @xmath920.28 + i18223 - 1243 mm4 & @xmath920.22 & & & 0.71 ( 0.08 ) & & @xmath920.22 & @xmath920.25 + i18306 - 0835 mm1 & 0.39 ( 0.02 ) & 78.5 ( 0.1 ) & 4.0 ( 0.2 ) & 1.7 ( 0.09 ) & & @xmath920.30 & @xmath920.34 + i18306 - 0835 mm2 & @xmath920.25 & & & 0.41 ( 0.10 ) & & @xmath920.31 & @xmath920.36 + i18306 - 0835 mm3 & @xmath920.25 & & & 0.32 ( 0.10 ) & & @xmath920.31 & @xmath920.36 + i18337 - 0743 mm1 & 1.06 ( 0.02 ) & 58.31 ( 0.04 ) & 3.51 ( 0.08 ) & 4.21 ( 0.09 ) & & @xmath920.22 & @xmath920.26 + i18337 - 0743 mm2 & 0.50 ( 0.03 ) & 58.13 ( 0.06 ) & 2.3 ( 0.1 ) & 1.22 ( 0.09 ) & & @xmath920.24 & @xmath920.28 + i18337 - 0743 mm3 & 0.25 ( 0.02 ) & 55.6 ( 0.2 ) & 3.7 ( 0.4 ) & 1.07 ( 0.09 ) & & @xmath920.28 & @xmath920.33 + [ tab : t5 ] lrrrr g015.05 + 00.07 mm1 & @xmath920.29 & & & @xmath920.6 + g015.31 - 00.16 mm2 & @xmath920.30 & & & @xmath920.6 + g015.31 - 00.16 mm3 & @xmath920.40 & & & @xmath920.8 + g019.27 + 0.07 mm1 & 0.57 ( 0.04 ) & 26.7 ( 0.2 ) & 5.2 ( 0.4 ) & 3.1 ( 0.2 ) + g019.27 + 0.07 mm2 & @xmath920.28 & & & @xmath920.6 + g022.35 + 00.41 mm1 & 0.64 ( 0.04 ) & 52.6 ( 0.2 ) & 5.4 ( 0.4 ) & 3.6 ( 0.2 ) + g022.35 + 00.41 mm2 & @xmath920.34 & & & @xmath920.7 + g022.35 + 00.41 mm3 & @xmath920.28 & & & @xmath920.6 + g023.60 + 00.00 mm1 & 0.67 ( 0.04 ) & 107.1 ( 0.2 ) & 4.9 ( 0.4 ) & 3.5 ( 0.2 ) + g023.60 + 00.00 mm2 & 0.34 ( 0.03 ) & 53.0 ( 0.3 ) & 6.7 ( 0.7 ) & 2.2 ( 0.2 ) + g023.60 + 00.00 mm3 & @xmath920.25 & & & @xmath920.5 + g023.60 + 00.00 mm4 & @xmath920.25 & & & @xmath920.5 + g023.60 + 00.00 mm7 & @xmath920.39 & & & @xmath920.8 + g024.08 + 00.04 mm1 & @xmath920.34 & & & @xmath920.7 + g024.08 + 00.04 mm2(1 ) & @xmath920.46 & & & @xmath921.0 + g024.08 + 00.04 mm2(2 ) & @xmath920.46 & & & @xmath921.0 + g024.08 + 00.04 mm3 & @xmath920.31 & & & @xmath920.7 + g024.08 + 00.04 mm4 & @xmath920.39 & & & @xmath920.8 + g024.33 + 00.11 mm1 & 0.48 ( 0.04 ) & 114.2 ( 0.3 ) & 6.6 ( 0.6 ) & 3.0 ( 0.2 ) + g024.33 + 00.11 mm2 & @xmath920.29 & & & @xmath920.6 + g024.33 + 00.11 mm3 & @xmath920.19 & & & @xmath920.4 + g024.33 + 00.11 mm4 & @xmath920.28 & & & @xmath920.6 + g024.33 + 00.11 mm5 & @xmath920.28 & & & @xmath920.6 + g024.33 + 00.11 mm6 & @xmath920.36 & & & @xmath920.8 + g024.33 + 00.11 mm8 & @xmath920.24 & & & @xmath920.5 + g024.33 + 00.11 mm9 & @xmath920.30 & & & @xmath920.6 + g024.33 + 00.11 mm11 & @xmath920.26 & & & @xmath920.5 + g024.60 + 00.08 mm1 & 0.32 ( 0.03 ) & 52.1 ( 0.3 ) & 5.9 ( 0.7 ) & 1.9 ( 0.2 ) + g024.60 + 00.08 mm2 & @xmath920.32 & & & @xmath920.7 + g025.04 - 00.20 mm1 & @xmath920.28 & & & @xmath920.6 + g025.04 - 00.20 mm2 & @xmath920.27 & & & @xmath920.6 + g025.04 - 00.20 mm4 & @xmath920.33 & & & @xmath920.7 + g034.43 + 00.24 mm1 & 1.79 ( 0.09 ) & 57.6 ( 0.1 ) & 4.8 ( 0.3 ) & 9.5 ( 0.1 ) + g034.43 + 00.24 mm2 & 1.26 ( 0.06 ) & 57.7 ( 0.1 ) & 5.7 ( 0.3 ) & 7.3 ( 0.2 ) + g034.43 + 00.24 mm3 & 0.73 ( 0.03 ) & 58.8 ( 0.2 ) & 8.6 ( 0.5 ) & 5.4 ( 0.1 ) + g034.43 + 00.24 mm4 & 0.57 ( 0.05 ) & 57.2 ( 0.2 ) & 4.1 ( 0.4 ) & 2.5 ( 0.2 ) + g034.43 + 00.24 mm5 & @xmath920.30 & & & @xmath920.6 + g034.43 + 00.24 mm6 & @xmath920.31 & & & @xmath920.7 + g034.43 + 00.24 mm8 & @xmath920.28 & & & @xmath920.6 + g034.43 + 00.24 mm9 & @xmath920.32 & & & @xmath920.7 + i18102 - 1800 mm1 & 0.67 ( 0.03 ) & 21.7 ( 0.2 ) & 6.2 ( 0.4 ) & 4.2 ( 0.1 ) + i18151 - 1208 mm1 & 0.30 ( 0.05 ) & 33.1 ( 0.3 ) & 3.6 ( 0.6 ) & 1.1 ( 0.2 ) + i18151 - 1208 mm2 & 0.51 ( 0.04 ) & 30.4 ( 0.3 ) & 7.1 ( 0.6 ) & 3.3 ( 0.2 ) + i18151 - 1208 mm3 & @xmath920.27 & & & @xmath920.6 + i18182 - 1433 mm1 & 0.43 ( 0.04 ) & 60.0 ( 0.2 ) & 4.4 ( 0.5 ) & 2.0 ( 0.2 ) + i18182 - 1433 mm2 & @xmath920.39 & & & @xmath920.8 + i18223 - 1243 mm1 & @xmath920.37 & & & @xmath920.8 + i18223 - 1243 mm2 & @xmath920.42 & & & @xmath920.9 + i18223 - 1243 mm3 & 0.27 ( 0.03 ) & 44.9 ( 0.3 ) & 5.4 ( 0.6 ) & 1.5 ( 0.1 ) + i18223 - 1243 mm4 & @xmath920.29 & & & @xmath920.6 + i18306 - 0835 mm1 & 0.50 ( 0.04 ) & 78.0 ( 0.2 ) & 3.6 ( 0.4 ) & 1.6 ( 0.1 ) + i18306 - 0835 mm2 & @xmath920.45 & & & @xmath920.9 + i18306 - 0835 mm3 & @xmath920.43 & & & @xmath920.9 + i18337 - 0743 mm1 & 0.34 ( 0.03 ) & 58.1 ( 0.3 ) & 5.7 ( 0.6 ) & 1.9 ( 0.1 ) + i18337 - 0743 mm2 & @xmath920.25 & & & @xmath920.5 + i18337 - 0743 mm3 & 0.23 ( 0.03 ) & 56.1 ( 0.3 ) & 5.0 ( 0.8 ) & 1.3 ( 0.1 ) + [ tab : t6 ] lrrrr g015.05 + 00.07 mm1 & 0.41 ( 0.02 ) & 24.72 ( 0.04 ) & 2.3 ( 0.1 ) & 1.16 ( 0.04 ) + g015.31 - 00.16 mm2 & 0.45 ( 0.02 ) & 31.05 ( 0.03 ) & 1.34 ( 0.07 ) & 0.72 ( 0.04 ) + g015.31 - 00.16 mm3 & 0.38 ( 0.02 ) & 30.93 ( 0.03 ) & 1.27 ( 0.07 ) & 0.50 ( 0.04 ) + g019.27 + 0.07 mm1 & 0.93 ( 0.03 ) & 26.37 ( 0.03 ) & 2.57 ( 0.08 ) & 2.85 ( 0.04 ) + g019.27 + 0.07 mm2 & 0.86 ( 0.02 ) & 26.79 ( 0.04 ) & 2.71 ( 0.09 ) & 2.68 ( 0.04 ) + g022.35 + 00.41 mm1 & 0.36 ( 0.01 ) & 52.67 ( 0.05 ) & 3.0 ( 0.1 ) & 1.31 ( 0.03 ) + g022.35 + 00.41 mm2 & 0.29 ( 0.02 ) & 60.22 ( 0.04 ) & 1.5 ( 0.1 ) & 0.53 ( 0.04 ) + g022.35 + 00.41 mm3 & 0.78 ( 0.02 ) & 84.29 ( 0.03 ) & 2.45 ( 0.07 ) & 2.06 ( 0.04 ) + g023.60 + 00.00 mm1 & 0.60 ( 0.02 ) & 106.10 ( 0.05 ) & 4.1 ( 0.1 ) & 2.72 ( 0.03 ) + g023.60 + 00.00 mm2 & 1.02 ( 0.02 ) & 53.49 ( 0.03 ) & 2.47 ( 0.07 ) & 2.93 ( 0.04 ) + g023.60 + 00.00 mm3 & 0.76 ( 0.02 ) & 105.45 ( 0.05 ) & 3.6 ( 0.1 ) & 3.06 ( 0.04 ) + g023.60 + 00.00 mm4 & 0.78 ( 0.02 ) & 53.65 ( 0.03 ) & 2.2 ( 0.07 ) & 2.06 ( 0.04 ) + g023.60 + 00.00 mm7 & 1.04 ( 0.03 ) & 53.74 ( 0.03 ) & 2.22 ( 0.07 ) & 2.74 ( 0.04 ) + g024.08 + 00.04 mm1 & 0.55 ( 0.02 ) & 113.61 ( 0.04 ) & 2.9 ( 0.1 ) & 1.92 ( 0.04 ) + g024.08 + 00.04 mm2(1 ) & 0.31 ( 0.02 ) & 51.80 ( 0.06 ) & 1.6 ( 0.1 ) & 0.69 ( 0.04 ) + g024.08 + 00.04 mm2(2 ) & 0.39 ( 0.02 ) & 114.02 ( 0.05 ) & 2.4 ( 0.1 ) & 1.14 ( 0.04 ) + g024.08 + 00.04 mm3 & 0.52 ( 0.02 ) & 51.59 ( 0.03 ) & 1.89 ( 0.07 ) & 1.21 ( 0.04 ) + g024.08 + 00.04 mm4 & 0.55 ( 0.02 ) & 51.59 ( 0.03 ) & 1.67 ( 0.07 ) & 1.18 ( 0.04 ) + g024.33 + 00.11 mm1 & 1.25 ( 0.04 ) & 113.61 ( 0.05 ) & 3.4 ( 0.1 ) & 4.88 ( 0.06 ) + g024.33 + 00.11 mm2 & 0.71 ( 0.02 ) & 118.11 ( 0.05 ) & 3.5 ( 0.1 ) & 2.50 ( 0.05 ) + g024.33 + 00.11 mm3 p1 & 0.58 ( 0.01 ) & 116.63 ( 0.04 ) & 2.16 ( 0.09 ) & 2.26 ( 0.05 ) + g024.33 + 00.11 mm3 p2 & 0.47 ( 0.01 ) & 119.09 ( 0.05 ) & 1.9 ( 0.1 ) & + g024.33 + 00.11 mm4 & 0.55 ( 0.02 ) & 114.89 ( 0.07 ) & 4.6 ( 0.2 ) & 2.84 ( 0.05 ) + g024.33 + 00.11 mm5 & 0.97 ( 0.01 ) & 117.47 ( 0.03 ) & 3.06 ( 0.06 ) & 3.18 ( 0.05 ) + g024.33 + 00.11 mm6 & 1.02 ( 0.03 ) & 113.85 ( 0.05 ) & 3.2 ( 0.1 ) & 3.77 ( 0.06 ) + g024.33 + 00.11 mm8 & 0.29 ( 0.02 ) & 119.93 ( 0.09 ) & 2.8 ( 0.2 ) & 1.65 ( 0.03 ) + g024.33 + 00.11 mm9 p1 & 0.38 ( 0.01 ) & 116.18 ( 0.04 ) & 1.98 ( 0.09 ) & 1.99 ( 0.03 ) + g024.33 + 00.11 mm9 p2 & 0.45 ( 0.01 ) & 119.28 ( 0.03 ) & 1.90 ( 0.08 ) & + g024.33 + 00.11 mm11 & 0.59 ( 0.02 ) & 112.99 ( 0.04 ) & 2.4 ( 0.1 ) & 1.60 ( 0.06 ) + g024.60 + 00.08 mm1 & 0.93 ( 0.02 ) & 52.99 ( 0.03 ) & 2.37 ( 0.07 ) & 2.61 ( 0.04 ) + g024.60 + 00.08 mm2 & 0.34 ( 0.01 ) & 115.11 ( 0.04 ) & 2.3 ( 0.1 ) & 0.92 ( 0.04 ) + g025.04 - 00.20 mm1 & 1.18 ( 0.03 ) & 63.53 ( 0.02 ) & 2.08 ( 0.06 ) & 2.87 ( 0.04 ) + g025.04 - 00.20 mm2 & 0.86 ( 0.02 ) & 63.44 ( 0.03 ) & 2.56 ( 0.08 ) & 2.48 ( 0.04 ) + g025.04 - 00.20 mm4 & 1.00 ( 0.03 ) & 63.55 ( 0.03 ) & 2.03 ( 0.06 ) & 2.31 ( 0.04 ) + g034.43 + 00.24 mm1 & 1.77 ( 0.04 ) & 57.79 ( 0.04 ) & 3.39 ( 0.09 ) & 6.83 ( 0.03 ) + g034.43 + 00.24 mm2 & 1.58 ( 0.04 ) & 57.50 ( 0.04 ) & 3.8 ( 0.1 ) & 6.70 ( 0.03 ) + g034.43 + 00.24 mm3 & 1.16 ( 0.03 ) & 58.98 ( 0.03 ) & 2.56 ( 0.07 ) & 3.62 ( 0.04 ) + g034.43 + 00.24 mm4 & 1.43 ( 0.03 ) & 57.50 ( 0.04 ) & 3.28 ( 0.09 ) & 5.32 ( 0.04 ) + g034.43 + 00.24 mm5 & 0.73 ( 0.03 ) & 57.57 ( 0.04 ) & 2.3 ( 0.1 ) & 1.97 ( 0.06 ) + g034.43 + 00.24 mm6 & 0.85 ( 0.03 ) & 58.31 ( 0.04 ) & 2.59 ( 0.09 ) & 2.55 ( 0.06 ) + g034.43 + 00.24 mm8 & 1.29 ( 0.03 ) & 57.84 ( 0.04 ) & 3.14 ( 0.09 ) & 4.62 ( 0.06 ) + g034.43 + 00.24 mm9 & 1.10 ( 0.03 ) & 58.53 ( 0.04 ) & 2.62 ( 0.09 ) & 3.28 ( 0.06 ) + i18102 - 1800 mm1 & 1.63 ( 0.04 ) & 21.53 ( 0.04 ) & 2.88 ( 0.08 ) & 5.61 ( 0.04 ) + i18151 - 1208 mm1 & 0.58 ( 0.02 ) & 32.84 ( 0.03 ) & 2.37 ( 0.08 ) & 1.59 ( 0.04 ) + i18151 - 1208 mm2 & 0.45 ( 0.01 ) & 29.93 ( 0.04 ) & 2.9 ( 0.1 ) & 1.48 ( 0.04 ) + i18151 - 1208 mm3 & 0.23 ( 0.01 ) & 30.86 ( 0.06 ) & 2.1 ( 0.1 ) & 0.42 ( 0.04 ) + i18182 - 1433 mm1 & 0.75 ( 0.02 ) & 59.72 ( 0.04 ) & 3.18 ( 0.09 ) & 2.59 ( 0.04 ) + i18182 - 1433 mm2 & 0.53 ( 0.02 ) & 40.93 ( 0.04 ) & 2.08 ( 0.09 ) & 1.22 ( 0.05 ) + i18223 - 1243 mm1 & 0.93 ( 0.03 ) & 45.28 ( 0.03 ) & 1.87 ( 0.07 ) & 3.03 ( 0.03 ) + i18223 - 1243 mm2 & 0.82 ( 0.03 ) & 45.84 ( 0.05 ) & 2.9 ( 0.1 ) & 2.16 ( 0.06 ) + i18223 - 1243 mm3 & 0.92 ( 0.02 ) & 45.66 ( 0.04 ) & 2.95 ( 0.09 ) & 2.62 ( 0.06 ) + i18223 - 1243 mm4 & 0.62 ( 0.02 ) & 45.78 ( 0.03 ) & 1.73 ( 0.08 ) & 1.34 ( 0.06 ) + i18306 - 0835 mm1 & 0.91 ( 0.02 ) & 77.88 ( 0.03 ) & 2.57 ( 0.08 ) & 2.75 ( 0.04 ) + i18306 - 0835 mm2 & 0.29 ( 0.02 ) & 76.58 ( 0.07 ) & 2.3 ( 0.2 ) & 0.82 ( 0.06 ) + i18306 - 0835 mm3 & 0.37 ( 0.02 ) & 54.45 ( 0.06 ) & 2.1 ( 0.1 ) & 0.73 ( 0.06 ) + i18337 - 0743 mm1 & 1.38 ( 0.04 ) & 58.37 ( 0.05 ) & 3.7 ( 0.1 ) & 5.68 ( 0.04 ) + i18337 - 0743 mm2 & 0.32 ( 0.01 ) & 58.53 ( 0.07 ) & 3.2 ( 0.2 ) & 1.15 ( 0.04 ) + i18337 - 0743 mm3 & 0.45 ( 0.02 ) & 55.85 ( 0.05 ) & 3.2 ( 0.1 ) & 1.67 ( 0.05 ) + [ tab : t7 ] lrrrr g015.05 + 00.07 mm1 & 0.18 ( 0.01 ) & 24.60 ( 0.08 ) & 2.1 ( 0.2 ) & 0.48 ( 0.05 ) + g015.31 - 00.16 mm2 & @xmath920.09 & & & @xmath920.14 + g015.31 - 00.16 mm3 & @xmath920.08 & & & @xmath920.13 + g019.27 + 0.07 mm1 & 0.43 ( 0.01 ) & 26.37 ( 0.03 ) & 2.24 ( 0.07 ) & 1.08 ( 0.04 ) + g019.27 + 0.07 mm2 & 0.40 ( 0.01 ) & 26.92 ( 0.04 ) & 2.44 ( 0.09 ) & 1.19 ( 0.04 ) + g022.35 + 00.41 mm1 & 0.17 ( 0.01 ) & 52.57 ( 0.07 ) & 2.0 ( 0.2 ) & 0.43 ( 0.04 ) + g022.35 + 00.41 mm2 & @xmath920.08 & & & 0.27 ( 0.05 ) + g022.35 + 00.41 mm3 & 0.30 ( 0.01 ) & 84.32 ( 0.03 ) & 2.1 ( 0.1 ) & 0.81 ( 0.05 ) + g023.60 + 00.00 mm1 & 0.31 ( 0.01 ) & 106.37 ( 0.04 ) & 3.3 ( 0.1 ) & 1.11 ( 0.03 ) + g023.60 + 00.00 mm2 & 0.46 ( 0.01 ) & 53.47 ( 0.03 ) & 2.62 ( 0.07 ) & 1.47 ( 0.04 ) + g023.60 + 00.00 mm3 & 0.32 ( 0.01 ) & 105.73 ( 0.06 ) & 3.8 ( 0.1 ) & 1.29 ( 0.04 ) + g023.60 + 00.00 mm4 & 0.30 ( 0.01 ) & 53.44 ( 0.04 ) & 2.2 ( 0.1 ) & 0.71 ( 0.04 ) + g023.60 + 00.00 mm7 & 0.33 ( 0.01 ) & 53.6 ( 0.04 ) & 2.0 ( 0.1 ) & 0.73 ( 0.05 ) + g024.08 + 00.04 mm1 & 0.30 ( 0.01 ) & 113.66 ( 0.05 ) & 2.5 ( 0.1 ) & 0.87 ( 0.04 ) + g024.08 + 00.04 mm2(1 ) & 0.12 ( 0.02 ) & 51.75 ( 0.08 ) & 1.1 ( 0.2 ) & 0.12 ( 0.04 ) + g024.08 + 00.04 mm2(2 ) & 0.17 ( 0.01 ) & 114.17 ( 0.07 ) & 2.1 ( 0.2 ) & 0.38 ( 0.04 ) + g024.08 + 00.04 mm3 & 0.12 ( 0.01 ) & 51.62 ( 0.09 ) & 1.7 ( 0.2 ) & 0.16 ( 0.04 ) + g024.08 + 00.04 mm4 & 0.15 ( 0.01 ) & 51.63 ( 0.07 ) & 1.5 ( 0.2 ) & 0.23 ( 0.04 ) + g024.33 + 00.11 mm1 & 0.84 ( 0.02 ) & 113.46 ( 0.03 ) & 3.04 ( 0.07 ) & 2.84 ( 0.06 ) + g024.33 + 00.11 mm2 & 0.34 ( 0.01 ) & 118.00 ( 0.06 ) & 3.1 ( 0.2 ) & 1.05 ( 0.06 ) + g024.33 + 00.11 mm3 p1 & 0.18 ( 0.01 ) & 116.9 ( 0.1 ) & 2.4 ( 0.3 ) & 0.62 ( 0.06 ) + g024.33 + 00.11 mm3 p2 & 0.17 ( 0.02 ) & 119.24 ( 0.08 ) & 1.2 ( 0.2 ) & + g024.33 + 00.11 mm4 & 0.16 ( 0.01 ) & 114.61 ( 0.2 ) & 4.8 ( 0.4 ) & 0.80 ( 0.05 ) + g024.33 + 00.11 mm5 & 0.47 ( 0.01 ) & 117.4 ( 0.05 ) & 2.9 ( 0.1 ) & 1.28 ( 0.05 ) + g024.33 + 00.11 mm6 & 0.62 ( 0.02 ) & 113.76 ( 0.04 ) & 3.02 ( 0.09 ) & 2.08 ( 0.06 ) + g024.33 + 00.11 mm8 & 0.10 ( 0.01 ) & 119.4 ( 0.1 ) & 2.0 ( 0.3 ) & 0.41 ( 0.03 ) + g024.33 + 00.11 mm9 p1 & 0.13 ( 0.01 ) & 116.26 ( 0.07 ) & 2.2 ( 0.2 ) & 0.62 ( 0.03 ) + g024.33 + 00.11 mm9 p2 & 0.19 ( 0.01 ) & 119.22 ( 0.04 ) & 1.4 ( 0.1 ) & + g024.33 + 00.11 mm11 & 0.24 ( 0.02 ) & 112.91 ( 0.09 ) & 2.5 ( 0.2 ) & 0.67 ( 0.06 ) + g024.60 + 00.08 mm1 & 0.37 ( 0.01 ) & 52.98 ( 0.04 ) & 2.4 ( 0.1 ) & 1.10 ( 0.04 ) + g024.60 + 00.08 mm2 & 0.14 ( 0.01 ) & 114.7 ( 0.1 ) & 3.4 ( 0.3 ) & 0.55 ( 0.04 ) + g025.04 - 00.20 mm1 & 0.44 ( 0.01 ) & 63.6 ( 0.03 ) & 2.18 ( 0.08 ) & 1.12 ( 0.04 ) + g025.04 - 00.20 mm2 & 0.25 ( 0.01 ) & 63.52 ( 0.07 ) & 2.9 ( 0.2 ) & 0.88 ( 0.05 ) + g025.04 - 00.20 mm4 & 0.28 ( 0.01 ) & 63.61 ( 0.05 ) & 2.1 ( 0.1 ) & 0.64 ( 0.05 ) + g034.43 + 00.24 mm1 & 1.02 ( 0.01 ) & 57.71 ( 0.02 ) & 3.54 ( 0.04 ) & 4.06 ( 0.04 ) + g034.43 + 00.24 mm2 & 0.93 ( 0.01 ) & 57.53 ( 0.02 ) & 3.66 ( 0.05 ) & 3.73 ( 0.04 ) + g034.43 + 00.24 mm3 & 0.46 ( 0.01 ) & 58.97 ( 0.03 ) & 2.91 ( 0.08 ) & 1.60 ( 0.04 ) + g034.43 + 00.24 mm4 & 0.76 ( 0.01 ) & 57.49 ( 0.02 ) & 3.17 ( 0.05 ) & 2.64 ( 0.04 ) + g034.43 + 00.24 mm5 & 0.28 ( 0.02 ) & 57.59 ( 0.07 ) & 1.9 ( 0.2 ) & 0.57 ( 0.06 ) + g034.43 + 00.24 mm6 & 0.27 ( 0.02 ) & 58.62 ( 0.08 ) & 2.5 ( 0.2 ) & 0.85 ( 0.07 ) + g034.43 + 00.24 mm8 & 0.63 ( 0.02 ) & 57.94 ( 0.04 ) & 3.3 ( 0.1 ) & 2.37 ( 0.07 ) + g034.43 + 00.24 mm9 & 0.36 ( 0.02 ) & 58.99 ( 0.06 ) & 2.3 ( 0.1 ) & 0.82 ( 0.07 ) + i18102 - 1800 mm1 & 0.93 ( 0.01 ) & 21.57 ( 0.02 ) & 3.04 ( 0.05 ) & 3.28 ( 0.05 ) + i18151 - 1208 mm1 & 0.34 ( 0.01 ) & 32.92 ( 0.04 ) & 2.2 ( 0.1 ) & 0.90 ( 0.05 ) + i18151 - 1208 mm2 & 0.20 ( 0.01 ) & 30.0 ( 0.1 ) & 4.2 ( 0.2 ) & 0.97 ( 0.05 ) + i18151 - 1208 mm3 & 0.11 ( 0.01 ) & 31.2 ( 0.2 ) & 3.6 ( 0.4 ) & 0.45 ( 0.04 ) + i18182 - 1433 mm1 & 0.39 ( 0.01 ) & 59.69 ( 0.04 ) & 3.0 ( 0.1 ) & 1.15 ( 0.04 ) + i18182 - 1433 mm2 & 0.17 ( 0.01 ) & 40.9 ( 0.1 ) & 2.6 ( 0.2 ) & 0.42 ( 0.05 ) + i18223 - 1243 mm1 & 0.48 ( 0.02 ) & 45.20 ( 0.04 ) & 2.2 ( 0.1 ) & 1.21 ( 0.04 ) + i18223 - 1243 mm2 & 0.33 ( 0.02 ) & 45.72 ( 0.07 ) & 2.8 ( 0.2 ) & 1.28 ( 0.06 ) + i18223 - 1243 mm3 & 0.42 ( 0.01 ) & 45.69 ( 0.03 ) & 2.57 ( 0.07 ) & 0.93 ( 0.06 ) + i18223 - 1243 mm4 & 0.23 ( 0.02 ) & 45.75 ( 0.08 ) & 1.8 ( 0.2 ) & 0.56 ( 0.06 ) + i18306 - 0835 mm1 & 0.49 ( 0.01 ) & 77.87 ( 0.04 ) & 2.82 ( 0.08 ) & 1.56 ( 0.04 ) + i18306 - 0835 mm2 & 0.12 ( 0.02 ) & 76.7 ( 0.2 ) & 2.5 ( 0.4 ) & 0.23 ( 0.06 ) + i18306 - 0835 mm3 & 0.13 ( 0.02 ) & 55.2 ( 0.2 ) & 2.3 ( 0.4 ) & 0.30 ( 0.06 ) + i18337 - 0743 mm1 & 0.77 ( 0.01 ) & 58.37 ( 0.03 ) & 3.52 ( 0.06 ) & 2.92 ( 0.05 ) + i18337 - 0743 mm2 & 0.18 ( 0.01 ) & 59.99 ( 0.09 ) & 2.4 ( 0.2 ) & 0.37 ( 0.05 ) + i18337 - 0743 mm3 & 0.16 ( 0.01 ) & 56.3 ( 0.1 ) & 2.9 ( 0.3 ) & 0.59 ( 0.05 ) + [ tab : t8 ] lrrrr g015.05 + 00.07 mm1 & @xmath920.08 & & & 0.34 ( 0.04 ) + g015.31 - 00.16 mm2 & @xmath920.08 & & & @xmath920.13 + g015.31 - 00.16 mm3 & @xmath920.08 & & & @xmath920.12 + g019.27 + 0.07 mm1 & 0.16 ( 0.01 ) & 26.6 ( 0.1 ) & 4.0 ( 0.3 ) & 0.69 ( 0.04 ) + g019.27 + 0.07 mm2 & 0.16 ( 0.01 ) & 27.0 ( 0.1 ) & 3.9 ( 0.3 ) & 0.77 ( 0.04 ) + g022.35 + 00.41 mm1 & 0.15 ( 0.01 ) & 52.97 ( 0.07 ) & 4.5 ( 0.2 ) & 0.86 ( 0.04 ) + g022.35 + 00.41 mm2 & @xmath920.08 & & & @xmath920.12 + g022.35 + 00.41 mm3 & @xmath920.08 & & & 0.24 ( 0.04 ) + g023.60 + 00.00 mm1 & 0.19 ( 0.01 ) & 106.56 ( 0.09 ) & 5.0 ( 0.2 ) & 0.99 ( 0.03 ) + g023.60 + 00.00 mm2 & 0.23 ( 0.01 ) & 53.51 ( 0.08 ) & 5.2 ( 0.2 ) & 1.29 ( 0.03 ) + g023.60 + 00.00 mm3 & 0.17 ( 0.01 ) & 106.3 ( 0.1 ) & 4.3 ( 0.3 ) & 0.82 ( 0.04 ) + g023.60 + 00.00 mm4 & 0.19 ( 0.01 ) & 53.46 ( 0.09 ) & 2.9 ( 0.2 ) & 0.72 ( 0.05 ) + g023.60 + 00.00 mm7 & 0.14 ( 0.01 ) & 53.3 ( 0.1 ) & 4.2 ( 0.3 ) & 0.72 ( 0.04 ) + g024.08 + 00.04 mm1 & 0.11 ( 0.01 ) & 113.9 ( 0.1 ) & 2.7 ( 0.3 ) & 0.28 ( 0.03 ) + g024.08 + 00.04 mm2(1 ) & @xmath920.07 & & & @xmath920.11 + g024.08 + 00.04 mm2(2 ) & @xmath920.07 & & & @xmath920.11 + g024.08 + 00.04 mm3 & @xmath920.07 & & & 0.13 ( 0.04 ) + g024.08 + 00.04 mm4 & @xmath920.07 & & & @xmath920.11 + g024.33 + 00.11 mm1 & 0.70 ( 0.01 ) & 113.81 ( 0.04 ) & 4.2 ( 0.1 ) & 3.30 ( 0.06 ) + g024.33 + 00.11 mm2 & 0.18 ( 0.01 ) & 117.6 ( 0.2 ) & 5.6 ( 0.4 ) & 1.01 ( 0.05 ) + g024.33 + 00.11 mm3 & 0.12 ( 0.01 ) & 118.2 ( 0.2 ) & 5.7 ( 0.5 ) & 0.76 ( 0.05 ) + g024.33 + 00.11 mm4 & 0.14 ( 0.02 ) & 114.9 ( 0.1 ) & 2.1 ( 0.3 ) & 0.31 ( 0.05 ) + g024.33 + 00.11 mm5 & 0.19 ( 0.01 ) & 119.1 ( 0.1 ) & 5.0 ( 0.3 ) & 1.02 ( 0.05 ) + g024.33 + 00.11 mm6 & 0.48 ( 0.01 ) & 113.98 ( 0.05 ) & 3.7 ( 0.1 ) & 2.07 ( 0.06 ) + g024.33 + 00.11 mm8 & @xmath920.07 & & & 0.19 ( 0.03 ) + g024.33 + 00.11 mm9 & @xmath920.07 & & & 0.26 ( 0.03 ) + g024.33 + 00.11 mm11 & @xmath920.12 & & & 0.36 ( 0.06 ) + g024.60 + 00.08 mm1 & 0.11 ( 0.01 ) & 52.2 ( 0.2 ) & 7.6 ( 0.5 ) & 0.78 ( 0.04 ) + g024.60 + 00.08 mm2 & @xmath920.08 & & & 0.41 ( 0.04 ) + g025.04 - 00.20 mm1 & 0.13 ( 0.01 ) & 63.2 ( 0.1 ) & 3.0 ( 0.3 ) & 0.44 ( 0.04 ) + g025.04 - 00.20 mm2 & 0.10 ( 0.01 ) & 63.7 ( 0.2 ) & 3.4 ( 0.4 ) & 0.41 ( 0.04 ) + g025.04 - 00.20 mm4 & @xmath920.08 & & & 0.18 ( 0.04 ) + g034.43 + 00.24 mm1 & 0.66 ( 0.01 ) & 57.9 ( 0.03 ) & 5.17 ( 0.08 ) & 3.58 ( 0.04 ) + g034.43 + 00.24 mm2 & 0.59 ( 0.01 ) & 57.65 ( 0.04 ) & 5.76 ( 0.09 ) & 3.45 ( 0.04 ) + g034.43 + 00.24 mm3 & 0.215 ( 0.006 ) & 59.4 ( 0.1 ) & 7.2 ( 0.2 ) & 1.42 ( 0.04 ) + g034.43 + 00.24 mm4 & 0.399 ( 0.007 ) & 57.55 ( 0.05 ) & 5.2 ( 0.1 ) & 2.15 ( 0.04 ) + g034.43 + 00.24 mm5 & @xmath920.12 & & & 0.39 ( 0.06 ) + g034.43 + 00.24 mm6 & @xmath920.12 & & & 0.33 ( 0.06 ) + g034.43 + 00.24 mm8 & 0.33 ( 0.02 ) & 57.8 ( 0.1 ) & 4.3 ( 0.2 ) & 1.53 ( 0.07 ) + g034.43 + 00.24 mm9 & 0.19 ( 0.02 ) & 59.2 ( 0.1 ) & 2.9 ( 0.3 ) & 0.64 ( 0.07 ) + i18102 - 1800 mm1 & 0.61 ( 0.01 ) & 21.47 ( 0.03 ) & 4.71 ( 0.08 ) & 3.01 ( 0.04 ) + i18151 - 1208 mm1 & 0.14 ( 0.01 ) & 32.8 ( 0.1 ) & 3.4 ( 0.3 ) & 0.58 ( 0.04 ) + i18151 - 1208 mm2 & 0.11 ( 0.01 ) & 29.4 ( 0.2 ) & 6.8 ( 0.5 ) & 0.72 ( 0.05 ) + i18151 - 1208 mm3 & @xmath920.08 & & & 0.29 ( 0.04 ) + i18182 - 1433 mm1 & 0.17 ( 0.01 ) & 59.4 ( 0.1 ) & 6 ( 0.3 ) & 1.02 ( 0.04 ) + i18182 - 1433 mm2 & @xmath920.09 & & & 0.19 ( 0.05 ) + i18223 - 1243 mm1 & 0.16 ( 0.02 ) & 45.2 ( 0.1 ) & 1.9 ( 0.3 ) & 0.36 ( 0.06 ) + i18223 - 1243 mm2 & @xmath920.12 & & & 0.32 ( 0.06 ) + i18223 - 1243 mm3 & 0.15 ( 0.01 ) & 45.1 ( 0.1 ) & 5.3 ( 0.3 ) & 0.8 ( 0.03 ) + i18223 - 1243 mm4 & @xmath920.12 & & & 0.24 ( 0.07 ) + i18306 - 0835 mm1 & 0.31 ( 0.01 ) & 78.04 ( 0.06 ) & 3.5 ( 0.1 ) & 1.22 ( 0.04 ) + i18306 - 0835 mm2 & 0.09 ( 0.02 ) & 77.6 ( 0.2 ) & 2.8 ( 0.5 ) & 0.3 ( 0.06 ) + i18306 - 0835 mm3 & @xmath920.11 & & & @xmath920.18 + i18337 - 0743 mm1 & 0.39 ( 0.01 ) & 58.29 ( 0.06 ) & 4.8 ( 0.1 ) & 1.99 ( 0.05 ) + i18337 - 0743 mm2 & 0.13 ( 0.01 ) & 58.3 ( 0.1 ) & 2.5 ( 0.3 ) & 0.32 ( 0.04 ) + i18337 - 0743 mm3 & @xmath920.08 & & & 0.67 ( 0.04 ) + [ tab : t9 ] lrrrrrrrr g015.05 + 00.07 mm1 & 15.2@xmath93 & 17.4@xmath94&@xmath923.8 & 9.7@xmath95 & 2.7@xmath96 & @xmath922.2 & @xmath920.24 & @xmath920.20 + g015.31 - 00.16 mm2 & 10.3@xmath97 & 5.4@xmath98&@xmath921.6 & 5.0@xmath98 & 1.1@xmath99 & @xmath922.3 & @xmath920.30 & @xmath920.32 + g015.31 - 00.16 mm3 & 12.6@xmath100 & 6.9@xmath101&@xmath921.7 & 4.4@xmath102 & 0.6@xmath103 & @xmath923.1 & @xmath920.27 & @xmath921.74 + g019.27 + 0.07 mm1 & 15.4@xmath104 & 52.0@xmath105&@xmath923.7 & 24.2@xmath106 & 3.1@xmath107 & 7.916 & @xmath920.07 & @xmath920.15 + g019.27 + 0.07 mm2 & 15.8@xmath108 & 29.2@xmath109&@xmath923.5 & 22.1@xmath97 & 3.0@xmath99 & @xmath922.2 & @xmath920.12 & @xmath920.15 + g022.35 + 00.41 mm1 & 16.0@xmath110 & 26.7@xmath111&@xmath923.6 & 18.3@xmath106 & 3.3@xmath112 & 9.218.7 & @xmath920.14 & @xmath920.15 + g022.35 + 00.41 mm2 & 12.8@xmath113 & 4.9@xmath114&@xmath921.7 & @xmath922.2 & 1.3@xmath115 & @xmath922.6 & @xmath920.37 & @xmath920.42 + g022.35 + 00.41 mm3 & 15.2@xmath104 & 15.8@xmath97&@xmath923.0 & 5.5@xmath116 & 2.2@xmath117 & @xmath922.1 & @xmath920.20 & @xmath920.18 + g023.60 + 00.00 mm1 & 17.0@xmath118 & 36.7@xmath119&@xmath925.2 & 22.2@xmath106 & 4.3@xmath120 & 8.918 & @xmath920.15 & @xmath920.14 + g023.60 + 00.00 mm2 & 16.0@xmath108 & 22.8@xmath121&@xmath923.8 & 31.6@xmath121 & 2.5@xmath117 & 5.511.2 & @xmath920.17 & @xmath920.19 + g023.60 + 00.00 mm3 & 16.4@xmath109 & 17.5@xmath106&@xmath924.8 & 3.9@xmath98 & 2.6@xmath122 & @xmath921.9 & @xmath920.28 & @xmath920.25 + g023.60 + 00.00 mm4 & 14.9@xmath123 & 8.6@xmath114&@xmath923.5 & 7.0@xmath124 & 2.1@xmath117 & @xmath922.0 & @xmath920.42 & @xmath920.22 + g023.60 + 00.00 mm7 & 13.1@xmath106 & 16.0@xmath117&@xmath922.6 & 13.6@xmath102 & 2.7@xmath99 & @xmath923.0 & @xmath920.17 & @xmath920.13 + g024.08 + 00.04 mm1 & 16.4@xmath110 & 38.9@xmath105&@xmath923.8 & 8.5@xmath114 & 4.0@xmath125 & @xmath923.6 & @xmath920.10 & @xmath920.12 + g024.08 + 00.04 mm2(1 ) & 14.9@xmath126 & 8.1@xmath127&@xmath922.6 & 4.0@xmath114 & 1.3@xmath125 & @xmath923.6 & @xmath920.36 & @xmath920.65 + g024.08 + 00.04 mm2(2 ) & 15.1@xmath128 & 10.5@xmath106&@xmath923.1 & 3.5@xmath116 & 2.8@xmath125 & @xmath923.6 & @xmath920.32 & @xmath920.16 + g024.08 + 00.04 mm3 & 12.1@xmath108 & 5.1@xmath124&@xmath923.1 & 6.7@xmath116 & 2.0@xmath99 & @xmath922.4 & @xmath920.63 & @xmath920.22 + g024.08 + 00.04 mm4 & 12.9@xmath123 & 4.6@xmath124&@xmath921.7 & 7.2@xmath116 & 1.7@xmath103 & @xmath923.0 & @xmath920.39 & @xmath920.14 + g024.33 + 00.11 mm1 & 18.8@xmath129 & 63.3@xmath130&@xmath926.6 & 33.0@xmath111 & 5.0@xmath112 & 7.715.6 & @xmath920.11 & @xmath920.16 + g024.33 + 00.11 mm2 & 16.6@xmath118 & 20.8@xmath123&@xmath925.1 & 12.2@xmath95 & 3.4@xmath112 & @xmath922.6 & @xmath920.26 & @xmath920.20 + g024.33 + 00.11 mm3 p1 & 13.9@xmath120 & 21.9@xmath107&@xmath925.5 & 11.7@xmath124 & 2.1@xmath131 & @xmath921.5 & @xmath920.26 & @xmath920.32 + g024.33 + 00.11 mm3 p2 & 14.5@xmath132 & & & & 1.8@xmath131 & & & + g024.33 + 00.11 mm4 & 13.4@xmath108 & 11.5@xmath95&@xmath924.6 & 9.8@xmath102 & 4.2@xmath133 & @xmath922.1 & @xmath920.42 & @xmath920.15 + g024.33 + 00.11 mm5 & 17.1@xmath117 & 23.7@xmath117&@xmath924.7 & 14.4@xmath124 & 2.7@xmath124 & @xmath922.2 & @xmath920.20 & @xmath920.19 + g024.33 + 00.11 mm6 & 17.3@xmath110 & 14.3@xmath121&@xmath925.7 & 7.2@xmath134 & 4.7@xmath112 & @xmath922.8 & @xmath920.42 & @xmath920.15 + g024.33 + 00.11 mm8 & &&& & & @xmath921.9 & & + g024.33 + 00.11 mm9 p1 & 15.1@xmath109 & & & & & & + g024.33 + 00.11 mm9 p2 & 15.2@xmath108 & 7.7@xmath134&@xmath923.9 & 4.6@xmath98 & 2.2@xmath131 & @xmath922.3 & @xmath920.52 & @xmath920.21 + g024.33 + 00.11 mm11 & 14.8@xmath111 & 8.6@xmath135&@xmath924.2 & @xmath922.5 & 2.7@xmath96 & @xmath922.0 & @xmath920.51 & @xmath920.22 + g024.60 + 00.08 mm1 & 15.3@xmath121 & 28.5@xmath108&@xmath923.2 & 24.0@xmath135 & 2.1@xmath103 & 4.99.9 & @xmath920.12 & @xmath920.20 + g024.60 + 00.08 mm2 & 14.2@xmath109 & 25.9@xmath136&@xmath924.2 & 11.1@xmath114 & 3.1@xmath99 & @xmath923.0 & @xmath920.17 & @xmath920.17 + g025.04 - 00.20 mm1 & 14.7@xmath121 & 27.9@xmath108&@xmath923.6 & 15.2@xmath114 & 2.0@xmath137 & @xmath922.5 & @xmath920.13 & @xmath920.22 + g025.04 - 00.20 mm2 & 13.1@xmath106 & 21.5@xmath97&@xmath924.5 & 10.5@xmath124 & 2.7@xmath99 & @xmath922.1 & @xmath920.21 & @xmath920.21 + g025.04 - 00.20 mm4 & 13.5@xmath121 & 16.2@xmath135&@xmath923.3 & 5.9@xmath98 & 1.7@xmath103 & @xmath922.5 & @xmath920.21 & @xmath920.27 + g034.43 + 00.24 mm1 & 18.5@xmath104 & 133.6@xmath138&@xmath923.9 & 54.3@xmath139 & 3.4@xmath107 & 24.249.2 & @xmath920.03 & @xmath920.14 + g034.43 + 00.24 mm2 & 18.8@xmath104 & 113.9@xmath140&@xmath924.5 & 52.0@xmath141 & 3.8@xmath107 & 18.737.9 & @xmath920.04 & @xmath920.14 + g034.43 + 00.24 mm3 & 15.5@xmath142 & 59.6@xmath143&@xmath923.2 & 30.6@xmath106 & 2.2@xmath103 & 13.928.1 & @xmath920.06 & @xmath920.19 + g034.43 + 00.24 mm4 & 17.6@xmath104 & 66.0@xmath144&@xmath923.4 & 29.9@xmath121 & 3.3@xmath107 & 6.413 & @xmath920.05 & @xmath920.13 + g034.43 + 00.24 mm5 & 14.3@xmath145 & 16.8@xmath108&@xmath923.5 & 6.1@xmath102 & 2.7@xmath96 & @xmath922.3 & @xmath920.22 & @xmath920.18 + g034.43 + 00.24 mm6 & 14.0@xmath136 & 15.4@xmath106&@xmath923.2 & 11.2@xmath114 & 2.3@xmath146 & @xmath922.4 & @xmath920.22 & @xmath920.20 + g034.43 + 00.24 mm8 & 17.2@xmath109 & 14.0@xmath107&@xmath924.4 & 10.7@xmath114 & 2.8@xmath99 & @xmath922.1 & @xmath920.33 & @xmath920.20 + g034.43 + 00.24 mm9 & 13.9@xmath123 & 20.1@xmath121&@xmath923.2 & 11.7@xmath134 & 2.5@xmath99 & @xmath922.4 & @xmath920.16 & @xmath920.17 + i18102 - 1800 mm1 & 17.9@xmath109 & 51.2@xmath147&@xmath926.6 & 41.2@xmath148 & 3.2@xmath99 & 10.721.7 & @xmath920.13 & @xmath920.25 + i18151 - 1208 mm1 & 20.8@xmath149 & 34.9@xmath150&@xmath925.3 & 15.0@xmath106 & 1.3@xmath117 & 2.95.9 & @xmath920.16 & @xmath920.66 + i18151 - 1208 mm2 & 17.8@xmath151 & 45.7@xmath152&@xmath925.8 & 9.3@xmath137 & 1.5@xmath107 & 8.417 & @xmath920.13 & @xmath920.65 + i18151 - 1208 mm3 & 16.0@xmath153 & 13.2@xmath154&@xmath922.8 & 3.6@xmath102 & 2.4@xmath155 & @xmath922.1 & @xmath920.23 & @xmath920.20 + i18182 - 1433 mm1 & 19.0@xmath151 & 55.8@xmath156&@xmath925.8 & 31.6@xmath118 & 1.9@xmath107 & 5.110.3 & @xmath920.11 & @xmath920.45 + i18182 - 1433 mm2 & 13.8@xmath109 & 17.0@xmath106&@xmath923.8 & 3.4@xmath98 & 2.0@xmath99 & @xmath923.0 & @xmath920.23 & @xmath920.27 + i18223 - 1243 mm1 & 17.5@xmath141 & 42.4@xmath157&@xmath923.8 & 10.7@xmath95 & 1.8@xmath137 & @xmath923.3 & @xmath920.09 & @xmath920.27 + i18223 - 1243 mm2 & 15.1@xmath132 & 25.2@xmath158&@xmath923.8 & 6.3@xmath102 & 2.9@xmath122 & @xmath923.3 & @xmath920.16 & @xmath920.17 + i18223 - 1243 mm3 & 16.2@xmath108 & 43.9@xmath139&@xmath924.3 & 15.1@xmath137 & 2.9@xmath99 & 3.87.7 & @xmath920.10 & @xmath920.19 + i18223 - 1243 mm4 & 15.5@xmath93 & 14.6@xmath108&@xmath923.3 & 6.6@xmath134 & 1.2@xmath103 & @xmath922.2 & @xmath920.24 & @xmath920.47 + i18306 - 0835 mm1 & 17.3@xmath109 & 22.6@xmath108&@xmath925.9 & 16.6@xmath135 & 2.8@xmath117 & 4.18.4 & @xmath920.27 & @xmath920.26 + i18306 - 0835 mm2 & 16.4@xmath159 & 10.4@xmath160&@xmath925.6 & 4.0@xmath114 & 1.6@xmath161 & @xmath923.5 & @xmath920.60 & @xmath921.39 + i18306 - 0835 mm3 & 13.4@xmath113 & 4.3@xmath114&@xmath924.2 & 2.8@xmath116 & 2.7@xmath162 & @xmath923.3 & @xmath921.05 & @xmath920.26 + i18337 - 0743 mm1 & 17.1@xmath108 & 67.2@xmath152&@xmath924.7 & 41.1@xmath109 & 4.6@xmath122 & 4.89.8 & @xmath920.07 & @xmath920.12 + i18337 - 0743 mm2 & 17.1@xmath163 & 21.5@xmath164&@xmath924.6 & 11.9@xmath165 & 4.1@xmath166 & @xmath921.9 & @xmath920.23 & @xmath920.16 + i18337 - 0743 mm3 & 15.0@xmath110 & 19.1@xmath136&@xmath924.6 & 9.8@xmath95 & 2.5@xmath167 & 3.26.5 & @xmath920.25 & @xmath920.29 + [ tab : t10 ] lrrrrr g019.27 + 0.07 mm1 & 3.2 ( 0.2 ) & @xmath920.4 & 0.7 ( 0.1 ) & 1.7 ( 0.2 ) & 7.3 ( 0.8 ) + g022.35 + 00.41 mm1 & 2.9 ( 0.2 ) & 0.5 ( 0.1 ) & 0.7 ( 0.1 ) & 1.1 ( 0.2 ) & 8.4 ( 0.3 ) + g023.60 + 00.00 mm1 & 3.6 ( 0.2 ) & @xmath920.4 & 0.5 ( 0.1 ) & 0.9 ( 0.2 ) & 6.9 ( 0.6 ) + g023.60 + 00.01 mm2 & 2.4 ( 0.2 ) & @xmath920.3 & @xmath920.3 & @xmath920.5 & @xmath927.4 + g024.33 + 00.11 mm1 & 2.9 ( 0.2 ) & @xmath920.4 & 1.0 ( 0.1 ) & 2.3 ( 0.2 ) & 8.3 ( 2.4 ) + g024.60 + 00.08 mm1 & 1.9 ( 0.2 ) & @xmath920.4 & 0.5 ( 0.1 ) & 0.6 ( 0.2 ) & 9.2 ( 1.1 ) + g034.43 + 00.24 mm1 & 8.7 ( 0.1 ) & 2.5 ( 0.1 ) & 3.4 ( 0.1 ) & 5.4 ( 0.1 ) & 10.9 ( 1.2 ) + g034.43 + 00.25 mm2 & 6.8 ( 0.2 ) & 1.2 ( 0.1 ) & 1.6 ( 0.1 ) & 3.1 ( 0.2 ) & 8.3 ( 0.5 ) + g034.43 + 00.26 mm3 & 6.3 ( 0.1 ) & 0.7 ( 0.1 ) & 1.2 ( 0.1 ) & 2.7 ( 0.1 ) & 7.0 ( 0.5 ) + g034.43 + 00.27 mm4 & 2.8 ( 0.2 ) & @xmath920.4 & 0.7 ( 0.1 ) & @xmath920.5 & 9.3 ( 3.1 ) + i18102 - 1800 mm1 & 4.5 ( 0.1 ) & 0.4 ( 0.1 ) & 1.0 ( 0.1 ) & 1.5 ( 0.1 ) & 7.3 ( 1.0 ) + i18151 - 1208 mm1 & 1.6 ( 0.2 ) & @xmath920.4 & 0.8 ( 0.1 ) & 0.9 ( 0.2 ) & 12.1 ( 0.8 ) + i18151 - 1209 mm2 & 4.3 ( 0.2 ) & @xmath920.4 & 1.8 ( 0.1 ) & 2.7 ( 0.2 ) & 8.1 ( 2.8 ) + i18182 - 1433 mm1 & 2.0 ( 0.2 ) & 0.8 ( 0.1 ) & 0.8 ( 0.1 ) & 1.5 ( 0.2 ) & 12.4 ( 3.5 ) + i18223 - 1245 mm3 & 1.3 ( 0.1 ) & 0.3 ( 0.1 ) & @xmath920.3 & 0.4 ( 0.1 ) & 9.5 ( 2.5 ) + i18306 - 0835 mm1 & 1.8 ( 0.2 ) & 0.5 ( 0.1 ) & 0.6 ( 0.1 ) & 0.9 ( 0.2 ) & 10.2 ( 1.1 ) + i18337 - 0743 mm1 & 1.8 ( 0.1 ) & @xmath920.4 & @xmath920.4 & 0.5 ( 0.1 ) & 8.6 ( 1.9 ) + i18337 - 0745 mm3 & 1.7 ( 0.1 ) & 0.3 ( 0.1 ) & @xmath920.3 & 0.6 ( 0.1 ) & 8.2 ( 1.2 ) + [ tab : t11 ] | we have surveyed the n@xmath0h@xmath1 @xmath2=10 , hc@xmath3n @xmath2=54 , ccs @xmath4=@xmath5@xmath6 , nh@xmath3 ( @xmath2 , @xmath7 ) = ( 1 , 1 ) , ( 2 , 2 ) , ( 3 , 3 ) , and ch@xmath3oh @xmath2=76 lines toward the 55 massive clumps associated with infrared dark clouds by using the nobeyama radio observatory 45 m telescope and the atacama submillimeter telescope experiment 10 m telescope .
the n@xmath0h@xmath1 , hc@xmath3n , and nh@xmath3 lines are detected toward most of the objects . on the other hand , the ccs emission is detected toward none of the objects .
the [ ccs]/[n@xmath0h@xmath1 ] ratios are found to be mostly lower than unity even in the spitzer 24 @xmath9 m dark objects .
this suggests that most of the massive clumps are chemically more evolved than the low - mass starless cores .
the ch@xmath3oh emission is detected toward 18 out of 55 objects .
all the ch@xmath3oh - detected objects are associated with the spitzer 24 @xmath9 m sources , suggesting that star formation has already started in all the ch@xmath3oh - detected objects .
the velocity widths of the ch@xmath3oh @xmath10=@xmath11@xmath12 @xmath13 and @xmath14@xmath15 @xmath16 lines are broader than those of n@xmath0h@xmath1 @xmath2=10 .
the ch@xmath3oh @xmath10=@xmath11@xmath12 @xmath13 and @xmath14@xmath15 @xmath16 lines tend to have broader linewidth in the msx dark objects than in the others , the former being younger or less luminous than the latter .
the origin of the broad emission is discussed in terms of the interaction between an outflow and an ambient cloud . |
the carmeli cosmology @xcite is as revolutionary in its implementation as it is in its interpretation .
the metric used by carmeli is unique in that it extends the number of dimensions of the universe by , either one dimension if we consider only the radial velocity of the galaxies in the hubble flow or by three if we consider all three velocity components .
we will confine the discussion in this paper to only one extra dimension as does carmeli . in that case
the line element in five dimensions becomes @xmath2 where @xmath3 and @xmath4 are @xmath5 potential functions to be determined .
the time ( @xmath6 ) is measured in the observer s frame .
the new dimension ( @xmath7 ) is the radial velocity of the galaxies in the expanding universe , in accordance with hubble flow .
the parameter @xmath8 , the hubble - carmeli constant , is a constant at any epoch and its reciprocal ( designated @xmath9 ) is approximately the hubble constant @xmath10 .
the line element represents a spherically symmetric isotropic universe , and the expansion is the result of _ spacevelocity _ expansion .
the expansion is observed at a definite time and thus @xmath11 .
taking into account @xmath12 ( isotropy condition ) and equation ( [ eqn : metric ] ) becomes @xmath13
the solution of ( [ eqn : phasespacemetric ] ) ( given by equation b.38 and solved in section b.10 in @xcite ) is reproduced here .
@xmath14 + the parameter @xmath15 is the mass / energy density of the universe expressed as a fraction of the critical or `` closure '' density , i.e. @xmath16 , where @xmath17 is the averaged matter / energy density of the universe . in this model , @xmath18
+ then ( [ eqn : phasespacederiv ] ) may be integrated exactly to get @xmath19 + carmeli expanded ( [ eqn : phasespacesoln ] ) in the limit of small @xmath20 and small @xmath15 to get @xmath21 @xmath22 thus we can write the expansion in terms of normalized or natural units @xmath23 . equation ( [ eqn : phasespacesolncz ] ) is plotted in fig . [
fig : fig1 ] for various values of @xmath24 and @xmath25 .
let us now re - write ( [ eqn : phasespacesoln ] ) in terms of natural units and for small @xmath26 but arbitrary @xmath15 , and we get @xmath27 + equation ( [ eqn : phasespacesolnnatural ] ) produces curves almost indistinguishable from ( [ eqn : phasespacesolncz ] ) so this verifies that the approximations work well for @xmath28 .
now let us consider what happens to the density of matter as we look back in the cosmos with redshift , @xmath26 .
it was assumed in fig . [
fig : fig1 ] that the value of @xmath15 is fixed for each curve .
carmeli does this also in figure a4 , page 134 in @xcite .
however , more correctly @xmath15 varies as a function of @xmath26 . for flat space
we assume the following relation to hold , @xmath29 where @xmath30 is a function of the redshift @xmath26 , and @xmath31 is the averaged matter density of the universe locally ( @xmath32 ) .
the parameter @xmath33 is then the local averaged matter density expressed as a fraction of `` closure '' density .
equation ( [ eqn : densityeqn ] ) results from the fact that as the redshift increases the volume decreases as @xmath34 .
notice at @xmath35 that the universe is 8 times smaller in volume and therefore 8 times more dense , that is , at @xmath36 , @xmath37 . substituting ( [ eqn : densityeqn ] ) into ( [ eqn : phasespacesolnnatural ] ) we get @xmath38
+ carmeli was able to simulate the form of the @xmath39 redshift data of @xcite published in 1998 which announced an accelerating universe following the observations of @xcite .
see figure a4 , page 134 in @xcite .
but in fact he had predicted this in 1996 @xcite .
so this means that carmeli assumed a value of total matter ( normal + dark matter ) density @xmath40 , which was the accepted value in 1998 .
now let s plot ( [ eqn : phasespacesolnnatural ] ) with @xmath40 and ( [ eqn : phasespacesolnnaturald ] ) with @xmath41 .
[ fig : fig2 ] .
this means that my modified equation ( [ eqn : phasespacesolnnaturald ] ) with @xmath41 gives the same result as carmeli s unapproximated equation ( [ eqn : phasespacesolnnatural ] ) with his assumed value of @xmath40 , but that included dark matter .
in fact , comparing ( [ eqn : phasespacesolnnatural ] ) and ( [ eqn : phasespacesolnnaturald ] ) , a local matter density of only @xmath42 0.030.04 is necessary to have agreement .
this effectively _ eliminates the need for the existence of dark matter _ on the cosmic scale .
table i shows the critical data from the comparison at redshifts between @xmath43 and @xmath35 .
it can be seen that the difference between the two equations over the domain of the measurements is much less significant than the fit to the data .
if we assume @xmath44 instead of @xmath45 , since both are within measured parameters , we get closer agreement at smaller redshifts but worse near @xmath35 .
+ in any case ( [ eqn : phasespacesolnnatural ] ) and ( [ eqn : phasespacesolnnaturald ] ) must be modified as @xmath46 to allow for relativistic effects , by replacing @xmath47 with the relativistic form @xmath48 . therefore we can re - write ( [ eqn : phasespacesolnnatural ] ) and ( [ eqn : phasespacesolnnaturald ] ) respectively as @xmath49 and @xmath50 where the varying matter density has been taken into account . in fig .
[ fig : fig3 ] , ( [ eqn : phasespacesolnnaturalr ] ) and ( [ eqn : phasespacesolnnaturalrd ] ) are compared .
the density approximation may be no longer valid past @xmath35 , because it is shown below that the vacuum energy term dominates and the universe is far from spatially flat . [
cols="^,^,^,^,^",options="header " , ]
based on the above analysis we can rewrite equation a.54 from @xcite for @xmath10 as @xmath51 which according to equation a.51 from @xcite may be further generalized without approximation , and using the relativistic form of the redshift .
still that equation may only be approximate for @xmath52 because of the density assumptions .
however it becomes @xmath53 where @xmath54 . both ( [ eqn : hubbleparam ] ) and ( [ eqn : hubbleparamgen ] ) have been plotted in fig .
[ fig : fig4 ] , and for carmeli s chosen value of @xmath55 at @xmath35 in ( [ eqn : hubbleparam ] ) yields @xmath56 ( very close to carmeli s value ) but ( [ eqn : hubbleparamgen ] ) yields @xmath57 .
this means without the small z approximation the value of @xmath9 is reduced when compared to that in @xcite .
the vacuum or dark energy parameter @xmath58 does not appear explicitly in carmeli s model .
it is only by a comparison with f - l models can an assignment be made .
on page 138 of @xcite by comparing with the standard model it is shown that @xmath59 , therefore we can write @xmath60 from ( [ eqn : darkenergy ] ) it is expected that using the unapproximated equation ( [ eqn : hubbleparamgen ] ) for @xmath10 the value of @xmath58 will be larger than carmeli s value using the form of ( [ eqn : phasespacesolncz ] ) .
[ fig : fig5 ] shows the values for the vacuum energy density @xmath58 ( broken curve ) and for the total energy density @xmath61 ( solid curve ) as a function of redshift , @xmath26 . from ( [ eqn : darkenergy ] ) it follows that as the universe expands the total density tends to the vacuum energy density @xmath62 ( since @xmath63 ) .
this means a totally 3d spatially flat universe in a totally relaxed state . for small @xmath26
the total density becomes @xmath64 it follows from ( [ eqn : darkenergysmz ] ) that for @xmath41 at z = 0 the total density @xmath65 .
this value is consistent with carmeli s result of @xmath66 .
however , it follows from ( [ eqn : densityeqn ] ) and ( [ eqn : darkenergysmz ] ) that the universe will always be open , @xmath67 as it expands . from fig .
[ fig : fig5 ] the total density @xmath61 is always greater than unity and as the universe expands , it asymptotically approaches unity therefore a spatially flat universe devoid of dark matter .
the 5d brane world of moshe carmeli has been has been applied to the expanding accelerating universe and the redshift distance relation has been generalised for redshifts up to at least @xmath68 .
it has been found that if a certain form of the dependence of baryonic matter density on redshift is assumed then the resulting distance - redshift relation will approximate the form of the data of the high-@xmath26 supernova teams without the need for dark matter .
even though it does not explicitly appear in the carmeli _ spacevelocity _ metric , the vacuum energy contribution to gravity , @xmath58 tends to unity as a function of decreasing redshift . also since the baryonic matter density @xmath63 as the universe expands , the total mass / energy density @xmath69
this indicates that the universe , though always open because @xmath67 , is asymptotically expanding towards a spatially flat state .
i would like to thank prof .
moshe carmeli for many valuable discussions
. + 99 carmeli , m. ( 1996 ) . _ cosmological general relativity _ , commun .
. phys . * 5*:159 behar , s. carmeli , m. ( 2000 ) _ cosmological relativity : a new theory of cosmology _ int . j. theor . phys . *
39*(5 ) carmeli , m. ( 2002 ) .
_ cosmological special relativity_. singapore , world scientific .
garnavich , p. m. , _ et al_. ( 1997 ) . _
constraints on cosmological models from hubble space telescope observations of high-@xmath26 supernovae _ bulletin of the american astronomical society * 29*(7 ) : 1350 perlmutter , s. , _ et al_. ( 1997 ) .
_ cosmology from type ia supernovae : measurements , calibration techniques , and implications .
_ bulletin of the american astronomical society * 29*(5 ) : 1351 riess , a. g. , _ et al_. ( 1998 ) . _ observational evidence from supernovae for an accelerating universe and a cosmological constant _ astron .
j. * 116*(sept ) : 1009 - 1038 | carmeli s 5d brane cosmology has been applied to the expanding accelerating universe and it has been found that the distance redshift relation will fit the data of the high - z supernova teams without the need for dark matter . also the vacuum energy contribution to gravity , @xmath0 indicates that the universe is asymptotically expanding towards a spatially flat state , where the total mass energy density @xmath1 . |
rigorous mathematical analysis of random operators has been a vital field of mathematical physics for the last few decades . for one - particle systems localization in the spectral and dynamical sense
has been established rigorously in various regimes of energy and disorder both for lattice and continuum models @xcite . despite of its physical relevance , the question of localization in disordered many - particle systems ,
however , is much less understood .
so far , the mathematical analysis of interacting disordered systems with a macroscopic number of particles has been mostly restricted to questions of the existence and properties of thermodynamic quantities @xcite with the exception of a proof of bose - einstein condensation for the disordered lieb - liniger model in one dimension @xcite .
in the course of the last years , several localization results have been obtained for systems of a fixed number of interacting particles .
persistence of localization has first been established for lattice systems , using multiscale @xcite or fractional moment @xcite techniques .
the multiscale analysis has been adapted to continuum multiparticle systems @xcite .
in addition , an improved version of the multiscale analysis , commonly referred to as `` bootstrap multiscale analysis '' has recently been extended to multiparticle systems as well , both for lattice @xcite and continuum models @xcite .
the aim of the present work is to complement these results and present an adaption of the multiparticle fractional - moment method developed in @xcite to the continuum .
the corresponding multiparticle random schrdinger operator has an alloy - type external potential and includes an interaction potential with rapidly decaying two - body interaction terms .
we focus on localization in energy regimes at the bottom of the spectrum and pay particular attention to how the strength and the decay of the interaction between different particles affects the localization estimates .
our main results are divided into two different types of decay of the interaction . in the first case , we assume that the interaction between particles located at points @xmath2 and @xmath3 decays ( sub)exponentially , i.e. , as @xmath4 with @xmath5 and @xmath6 .
this yields strong dynamical localization with decay estimates of the same order , albeit with respect to the `` hausdorff distance '' @xmath7 and not the regular euclidean metric on the configuration space of the particles . in the second case , we only assume fast polynomial decay which allows us to establish localization for systems consisting of at most a maximal number of particles .
although the interaction decays only polynomially , the localization estimates decay as @xmath8 , with @xmath5 and @xmath9 $ ] getting smaller as the particle number increases . in our analysis
it is essential to switch between different notions of localization , namely fractional moment localization and eigenfunction correlator localization .
section [ section : fmec ] discusses these two terms and their relation in a fairly general way that is of interest in itself and might be of use in other settings as well .
the main goal in that section is a proof that fractional moment localization and eigenfunction correlator localization are equivalent up to the localization length and a small change in the interval of validity .
the main part of our analysis is placed in section [ section : mploc ] .
we prove the inductive step that localization for up to @xmath10 particles implies localization for @xmath11 particles .
we proceed in several steps , first considering partially interactive systems and then using this to prove a `` rescaling inequality '' for the fully interacting systems .
sufficiently good estimates on an initial length scale will then yield localization for the @xmath11-particle system .
finally , in section [ section : proofofmainresults ] , we will use these results to prove the main theorems that are presented in section [ section : mainresults ] .
we start with a one - particle hamiltonian @xmath12 acting in @xmath13 , which is the sum of a deterministic term @xmath14 , consisting of the @xmath15-dimensional laplacian and a background potential @xmath16 , and an alloy - type random potential @xmath17 the @xmath11-particle operator @xmath18 , acting on the tensor product space @xmath19 , has the form @xmath20 where @xmath21 with @xmath22 acting on the variable @xmath23 .
the same holds for @xmath24 . denoting the @xmath11-particle configuration by @xmath25 ,
the random potential reads @xmath26 the additional operator @xmath27 denotes the potential corresponding to interactions between different particles .
its strength is controlled by @xmath28 .
we make the following assumptions : 1 . the background potential @xmath29 is bounded and @xmath30-periodic .
the single site potential @xmath31 is non - negative , bounded and supported in a ball of radius @xmath32 around @xmath33 .
furthermore , the added single - site potentials satisfy the `` covering - condition '' @xmath34 3 .
the random variables @xmath35 , @xmath36 , are independent and identically distributed with an absolutely continuous marginal distribution . the corresponding density @xmath37 is bounded and has compact suppport . without loss of generality , @xmath38 .
the interaction potential @xmath27 has the form @xmath39 a bound on the two - particle interaction potential @xmath40 is given by @xmath41 with a monotonely decreasing function @xmath42 satisfying @xmath43 for some @xmath44 . as the variables @xmath35 , @xmath36 , are independent and identically distributed , the hamiltonians @xmath45 form an ergodic family of operators with respect to the translations @xmath46 , @xmath47 .
the non - randomness of the spectrum is a standard consequence ( cf .
@xcite ) .
the spectrum of @xmath18 is almost surely a non - random set @xmath48 .
we denote the almost sure infimum of the spectrum by @xmath49 for later purpose , we also note the following : [ prop : gs ] if @xmath50 , then @xmath51 in the case of a repulsive interaction potential @xmath52 , this is in fact an equality .
we give an easily generalizable sketch of the proof of the inequality @xmath53 : choose @xmath54 such @xmath55 and @xmath56 .
now let @xmath57 be a spatially translated version of @xmath58 with a support that is far away from the support of @xmath58 , so that @xmath59 .
due to the assumptions on @xmath60 , one can almost surely choose this translation in such a way that @xmath61 .
consequently , @xmath62 with probability one , which implies the assertion .
+ the results we prove are expressed in estimates on the spatial decay of certain quantities .
this decay is measured with respect to the pseudo - metric @xmath63 for configurations @xmath64 , which is in fact the hausdorff distance of the sets @xmath65 and @xmath66 in @xmath67 .
the norm @xmath68 we use on @xmath67 and on @xmath69 is the @xmath70-norm .
balls around a point @xmath71 with a radius @xmath72 are denoted by @xmath73 , whereas balls around configurations @xmath74 are denoted by @xmath75 . the characteristic function @xmath76 of @xmath77
is denoted by @xmath78 .
most of the analysis in the following sections deals with restrictions of @xmath18 to finite volumes .
we will always impose dirichlet boundary conditions , which will be denoted by @xmath79 if the hamiltonian is restricted to @xmath80 .
the central tool in our analysis is the concept of fractional moment localization , which we define as follows : a bounded interval @xmath81 is a regime of @xmath11-particle fractional moment localization of order @xmath82 $ ] if there exist @xmath83 and @xmath84 such that @xmath85 \leq ce^{-\mu\operatorname{dist}_h({\mathbf{x}},{\mathbf{y}})^\gamma}\ ] ] for all @xmath64 . as we will see at the end of section [ section : fmec ]
, this implies dynamical localization in @xmath81 in the sense that @xmath86 \leq ae^{-\nu\operatorname{dist}_h({\mathbf{x}},{\mathbf{y}})^\gamma}\ ] ] for some @xmath87 and all @xmath64 , where we denote by @xmath88 the spectral projection of @xmath18 to the interval @xmath81 . in particular
this implies absence of continuous spectrum in @xmath81 and ( sub)exponential decay of the eigenfunctions with eigenvalues in @xmath81 , cf .
theorem [ thm : locinfvolume ] . in the following , we state our main results .
they are divided into two different cases depending on the decay of the two - particle interaction .
we provide criteria for fractional - moment localization near the bottom of the almost - sure spectrum in the regime of weak interaction and lifshitz tails .
[ thm : main : case1 ] assume that @xmath18 satisfies ( i)-(iv ) and that @xmath89 for some @xmath90 , @xmath91 $ ] and all @xmath92 . 1 .
let @xmath93 $ ] be a bounded interval of length @xmath94 that is a regime of one - particle fractional moment localization of order @xmath91 $ ] .
for all @xmath95 there is @xmath96 such that for any @xmath97 $ ] the interval @xmath98 $ ] is a regime of @xmath11-particle fractional moment localization of order @xmath99 .
2 . suppose that for any @xmath100 there exist @xmath101 and @xmath102 such that @xmath103 for all @xmath104 and @xmath105 .
then there is @xmath106 such that @xmath98 $ ] is a regime of @xmath11-particle fractional moment localization of order @xmath99 .
one - particle fractional moment localization , which constitutes the assumption of the first part of the theorem , has been pioneered in @xcite . in particular , the authors show under an additional assumption on the single - site distribution that any bounded interval @xmath107 is a regime of one - particle fractional moment localization of order @xmath108 provided the disorder strength is sufficiently large .
they also provide criteria on the density of states which imply this token .
this criterion has been refined and simplified for the bottom of the spectrum in @xcite . in one dimension @xmath109 ,
one - particle fractional moment localization up to arbitrary energies has been established in @xcite . in the multi - dimensional case @xmath110 ,
even if one - particle localization is established at higher energies ( e.g. at band edges ) , one can not generally expect multi - particle localization in energy regimes separated from the bottom the spectrum .
this is due to the possible appearance of extended states in the one - particle model at lower energies . as for the assumption in the second part of the theorem , finite - volume estimates of the form are well - known in the case @xmath111 , as they appear in the analysis of lifshitz tails , cf .
@xcite , and are used in the analysis in @xcite . in an @xmath11-particle system with repulsive interaction ( @xmath112 ) ,
the estimate follows directly from the one - particle case due to the fact that @xmath113 and @xmath114 .
( this fact has already been used in @xcite . ) however , if the interaction is ( partially ) repulsive , a proof becomes more involved .
one way to achieve this is to employ the techniques that are used in the proof of the inequality in the one - particle case , which is possible at least under certain assumptions on the interaction potential .
the analogue of theorem [ thm : main : case1 ] in case of polynomial decay is [ thm : main : case2 ] assume that @xmath18 satisfies ( i)-(iv ) and that @xmath115 for some @xmath116 and all @xmath92 . 1 .
let @xmath117 and @xmath93 $ ] be a bounded interval of length @xmath94 that is a regime of one - particle fractional moment localization of order @xmath118 $ ] .
for all @xmath95 there is @xmath96 auch that , if @xmath97 $ ] , then @xmath98 $ ] is a regime of @xmath11-particle fractional moment localization of order @xmath119 .
2 . suppose @xmath117 and that for any @xmath100 there exist @xmath101 and @xmath102 such that @xmath120 for all @xmath104 and @xmath105 .
then there exists @xmath95 such that @xmath98 $ ] is a regime of @xmath11-particle fractional moment localization of order @xmath119 .
in contrast to theorem [ thm : main : case1 ] , this theorem yields a bound on fractional moments that decays significantly faster than the interaction potential .
the price we pay , however , is that we can prove this only up to a maximal number of particles and that the decay becomes slower as @xmath11 increases .
the theorem does not exclude the possibility that there is no localization for large numbers of particles .
+ most results on localization for multi - particle systems deal with lattice models . aside from the pioneering works
@xcite , progress has been made in the lattice case on the energy regime in which localization is established and the regularity of the distribution of the random variables @xcite , see also the recent monograph @xcite and references therein . as in the one - particle case , the multi - scale approach @xcite has the advantage of being able to accommodate more singular distributions of the random variables whereas the fractional moment method @xcite allows one to establish exponential dynamical localization and not only polynomial decay or sub - exponential decay .
the latter is obtained through the bootstrap multi - scale method which has only recently been generalised to the multi - particle setup @xcite . for systems on continuous space , only two results are available so far .
the paper @xcite uses the multi - scale method to establish polynomial dynamical localization for a multi - particle model with alloy - type random potential similar to the one studied in this paper . aside from the difference in the decay rate
, they establish the analogue of theorem [ thm : main : case1](ii ) , i.e. , localization in a sufficiently small interval above the bottom of the spectrum .
as mentioned above , their method allows them to treat more singular distributions of the random variables . in the recent work @xcite the polynomial decay
has been improved to sub - exponential decay using the bootstrap method .
in contrast to @xcite , which allow for non - negative interactions of finite - range only , we investigate interactions which may have infinite range . for lattice models , ( complete )
localization in the presence of subexponentially decaying interactions has been considered before @xcite .
our novel point here is to study also the case of ( fast ) polynomial decay .
the focus of this section is the notion of _ fractional moment localization _ and its relation to other forms of localization .
the main results in theorems [ thm : main : case1 ] and [ thm : main : case2 ] are bounds on the spatial decay of fractional moments of the resolvent .
it is well known from the one - particle case @xcite that estimates of this type imply pure point spectrum with exponentially decaying eigenfunctions as well as a strong form of dynamical localization .
these results have been proved to carry over to multiparticle systems on the lattice in @xcite .
an important term in this context are eigenfunctions correlators , which will be used in our analysis as well .
we define _ eigenfunction correlator localization _ as follows : an interval @xmath121 is a regime of @xmath11-particle eigenfunction correlator localization of order @xmath82 $ ] if there exist @xmath122 such that @xmath123 \leq ce^{-\mu\operatorname{dist}_h({\mathbf{x}},{\mathbf{y}})^\gamma}\ ] ] for all bounded , open sets @xmath124 .
as we will see , fractional moment localization indeed implies eigenfunction correlator localization , a fact that is well - known from the one - particle theory . in the multiparticle setting , however , the right choice of the metric is crucial @xcite .
we will address this problem in our discussion below .
an essential element in the multiparticle analysis of @xcite was that exponential decay of the eigenfunction correlator implies exponential decay of fractional moments as well , so that the two terms are in a sense equivalent .
in contrast to their analysis , we do not consider complete localization , but localization in a bounded interval i. as it turns out , we can not prove that ( sub)exponential decay of the eigenfunction correlator in a bounded interval implies ( sub)exponential decay of fractional moments of the resolvent with energies in the interval @xmath81 . instead , one needs to restrict the energies to a slightly smaller interval @xmath125 that does not extend to the endpoints of @xmath81 .
+ the analysis of different notions of localization and their relation to each other is not an issue specific to the fractional moment method , see @xcite and references therein . since this is a topic of interest beyond the question of multi - particle localization and we will present a new technique for relating the eigenfunction correlator to the resolvent , we state our results in a fairly general way . the basic setup we impose for this part of the analysis deals with the following three basic objects : 1 .
the operators @xmath126 with @xmath127 and @xmath128 some probability space form a measurable family of self - adjoint operators in @xmath129 ( cf . @xcite*sec .
@xmath130 is a family of open subsets of @xmath131 and for any @xmath132 there is a self - adjoint measurable restriction @xmath133 of @xmath134 with domain @xmath135 .
furthermore , the spectrum of @xmath136 is only pure point for all @xmath137 .
3 . @xmath138 is a pseudo - metric on @xmath131 satisfying @xmath139 for all @xmath140 and @xmath82 $ ] . for our @xmath0-particle model presented in the previous section
this applies to the following choices : ad ( h ) : : : @xmath141 and @xmath142 .
ad ( o ) : : : the family @xmath130 consists of all sets of the form @xmath80 with an open , bounded set @xmath143 .
the restriction to sets of this form is physically reasonable and furthermore necessary in our analysis , as we make use of the tensor - product structure of @xmath144 and the corresponding identity for the @xmath11-particle hamiltonian without interaction .
ad ( m ) : : : the hausdorff distance @xmath7 is the pseudo - metric @xmath138 and it is easily checked that the summability condition is fulfilled ( cf .
a ) . the choice of this particular metric is also tied to assumption ( gw ) below .
e.g. for the euclidean metric the latter does not hold . in case
@xmath111 , of course , @xmath7 is just the metric induced induced by the max - norm .
further assumptions on these basic objects , which we need in order to establish a relation of the eigenfunction correlator to the fractional moment of the resolvent , are : 1 . for any @xmath145 with @xmath146
there is a cutoff function @xmath147 , with @xmath148 and @xmath149 where @xmath150 and @xmath151 denotes the maximum distance on @xmath152 , such that for almost all @xmath137 : 1 . for any @xmath153 : + @xmath154 and @xmath155 , 2 .
@xmath156=0 $ ] , 3 . for any bounded interval @xmath81
there is @xmath157 , which is independent of @xmath158 , @xmath159 and @xmath160 , such that for any eigenvalue @xmath161
@xmath162p_e(h_\omega(\omega))\| \leq c_i.\ ] ] 2 . for any @xmath163 with a minimum distance with respect to @xmath138 and for all @xmath132 ,
there is a bounded set @xmath164 , @xmath165 , and a sub-@xmath166-algebra @xmath167 such that for either @xmath168 or @xmath169 : 1 .
@xmath170 is a subset of @xmath171 and for all @xmath172 we have @xmath173 furthermore the number of unit balls in the @xmath70-norm needed to cover @xmath174 is polynomially bounded with respect to @xmath175 .
the operator @xmath176 is measurable with respect to the @xmath166-algebra @xmath177 .
3 . for any bounded interval @xmath121
there is @xmath178 such that @xmath179 \leq c_i|j|,\ ] ] for any interval @xmath125 .
the constant @xmath178 is independent of @xmath180 , @xmath181 and @xmath158 .
as these assumptions are not that intuitive , let us immediately add a few comments on their relation to our @xmath11-particle model .
ad ( l ) : : : this is an assumption of locality of the operator @xmath134 . in the case of the @xmath11-particle system we choose @xmath182 such that it satisfies @xmath183 with this choice , ( i ) is satisfied for the operator @xmath142 .
the commutator appearing in ( ii ) equals zero outside of the support of @xmath184 . as this support , however , is a subset of @xmath185 , the identity ( ii ) follows .
the inequality in ( iii ) is basically a consequence of the boundedness of the operator @xmath186 . in order to make the constant @xmath187 independent of @xmath158 , @xmath159 and @xmath188 , the chosen cutoff functions @xmath182 need to have @xmath189-bounds on their first and second derivatives that are independent of @xmath158 and @xmath159 . for later purpose
, we also note that ( l ) implies a geometric resolvent equation ( cf .
@xcite*lemma 4.2 ) , i.e. @xmath190(h_\omega - z)^{-1 } \label{eq : geomres}\ ] ] for any @xmath191 .
ad ( gw ) : : : the geometric assumption ( i ) is intertwined with the probabilistic assumptions ( ii ) and ( iii ) . for the @xmath11-particle model , we choose the appearing terms as follows .
if @xmath192 , then there is a particle @xmath193 such that @xmath194 with @xmath195 .
the set @xmath196 is then chosen as @xmath197 , cf .
figure [ fig : glw ] .
provided @xmath198 is sufficiently large , ( i ) is satisfied .
the sub-@xmath166-algebra @xmath177 is generated by the random variables @xmath35 , @xmath199 .
due to this choice , the potential on @xmath196 is measurable with respect to @xmath177 .
lemma [ lemma : exptrproj ] then ensures the validity of ( iii ) . and @xmath200 .
for the sketched configurations @xmath201 and
@xmath202 in the configuration space @xmath203 the hausdorff distance equals @xmath204 . in this example , it is necessary to choose @xmath205 and @xmath206 , since we need to find random variables @xmath207 such that i ) the corresponding potential @xmath208 satisfies the covering condition on @xmath209 or @xmath210 and ii ) the support of @xmath211 does not intersect @xmath212 .
this is achieved here by random variables @xmath213 in the vicinity of @xmath214 .
the shaded area indicates the support of @xmath215 . ] finally , for one part of our result we only need the following assumptions : 1 . for any bounded interval @xmath121
, there exists @xmath178 such that @xmath216 \leq c_i|j|\ ] ] for any interval @xmath125 .
2 . for any @xmath163 ,
all @xmath132 , almost all @xmath217 and all @xmath218 @xmath219 with uniform constants @xmath220 . to conclude ,
let us also comment on these assumption in relation to our @xmath221-particle model .
ad ( w ) : : : this assumption is a standard wegner estimate ( cf .
@xcite ) , which is proven here in [ lemma : exptrproj ] .
we note that it in particular implies the bound @xmath222 \leq c_i\|f\|_{l^1(i)}\ ] ] for any function @xmath223 , a fact that we will also use for the similar estimate ( gw)(iii ) . ad ( ct ) : : this is a combes - thomas estimate in a specific form adjusted to our purposes . as @xmath45 is a schrdinger operator with a bounded potential ,
this assumption is valid thanks to the results of , e.g. , @xcite .
our first main result in this section is : [ thm : generalecimpliesfm ] in the setting ( h ) , ( o ) , ( m ) , assume that ( w ) and ( ct ) hold . then for any bounded interval @xmath121 and any interval @xmath125 with @xmath224 if there exist @xmath225 and @xmath82 $ ] such that @xmath226 \leq ce^{-\mu\operatorname{d}({\mathbf{x}},{\mathbf{y}})^\gamma}\ ] ] for all @xmath163 , then for any @xmath84 there exist @xmath227 and such that @xmath228 \leq ae^{-\nu\operatorname{d}({\mathbf{x}},{\mathbf{y}})^\gamma}\ ] ] for all @xmath229 . as the eigenfunction correlator contains only information about @xmath133 restricted to the spectral subspace associated to an interval @xmath81 ,
it can not be straightforwardly used to bound a fractional moment of the full resolvent .
accordingly , we need to deal with contributions from energies outside of the interval in a different way .
as we will see , it is convenient to introduce a smooth cutoff function @xmath230 that satisfies @xmath231 for all @xmath232 , where @xmath125 is a slightly smaller interval .
we then split the resolvent into two parts , @xmath233 the first of which is dealt with by using the eigenfunction correlator bound and the second of which is bounded by means of the following lemma .
its proof requires assumption ( ct ) and is deferred to [ appendix : hsct ] .
assume that ( h ) , ( o ) and ( ct ) hold and let @xmath234 be a bounded interval , @xmath132 , @xmath217 and @xmath235 .
for any @xmath163 there is a function @xmath236 with @xmath237 such that @xmath238 if @xmath239 , @xmath240 if @xmath241 and that for all @xmath218 @xmath242 holds with constants @xmath243 that can be chosen independently of @xmath180 , @xmath181 , @xmath244 , @xmath158 and @xmath245 . in our setting , we choose @xmath235 such that the eigenfunction correlator bound holds on the larger interval @xmath246 $ ] .
thus , as long as @xmath247 , the second summand in can be bounded deterministically ( and hence in a fractional moment ) by .
we note that the cutoff function depends on @xmath180 and @xmath181 , the reason of which can be seen in the proof of the lemma .
however , in our setting , this peculiarity does not pose a problem .
it remains to find a bound on the first term in , @xmath248 as we will see , this term is easier to handle if @xmath249 is replaced by its square .
this modification can be justified as follows .
the inequalities @xmath250 \notag \\ & \leq \sum_{{\mathbf{w}}\in{\mathbb{z}^d}}{\mathbb{e}}\bigg[\bigg(\sum_{e\in\sigma(h_\omega)\cap i } \mkern-12mu\frac{\|\chi_{\mathbf{x}}p_e(h_\omega)\chi_{\mathbf{w}}\|^2}{|e - z|}\bigg)^s\bigg]^{\frac{1}{2 } } { \mathbb{e}}\bigg[\bigg(\sum_{e\in\sigma(h_\omega)\cap i } \mkern-12mu\frac{\|\chi_{\mathbf{w}}p_e(h_\omega)\chi_{\mathbf{y}}\|^2}{|e - z|}\bigg)^s\bigg]^{\frac{1}{2}}\end{aligned}\ ] ] together with assumption ( m ) imply that the left side decays ( sub)exponentially provided both two terms on the right side do so for some @xmath84 . in order to prove a bound on the terms on the right side , we choose @xmath251 such that @xmath252 and let @xmath253 be the conjugate hlder exponent , i.e. @xmath254
. then a two - fold application of hlder s inequality yields @xmath255 \notag\\ & \leq { \mathbb{e}}\bigg[\bigg(\sum_{e\in\sigma(h_\omega)\cap i }
\mkern-12mu\|\chi_{\mathbf{x}}p_e(h_\omega)\chi_{\mathbf{y}}\|^{\big(2-\frac{1}{sq}\big)p}\bigg)^s\bigg]^{\frac{1}{p}}{\mathbb{e}}\bigg[\bigg(\sum_{e\in\sigma(h_\omega)\cap i } \mkern-12mu \frac{\|\chi_{\mathbf{x}}p_e(h_\omega)\chi_{\mathbf{y}}\|^{\frac{q}{sq}}}{|e - z|^q}\bigg)^s\bigg]^{\frac{1}{q } } \notag \\ & \leq { \mathbb{e}}\bigg[\sum_{e\in\sigma(h_\omega)\cap i}\|\chi_{\mathbf{x}}p_e(h_\omega)\chi_{\mathbf{y}}\|\bigg]^{\frac{s}{p}}{\mathbb{e}}\bigg[\sum_{e\in\sigma(h_\omega)\cap i}\frac{\|\chi_{\mathbf{x}}p_e(h_\omega)\chi_{\mathbf{y}}\|}{|e - z|^{sq}}\bigg]^{\frac{1}{q}},\end{aligned}\ ] ] where we used that @xmath256 and @xmath257 .
the first factor contains the eigenfunction correlator and decays ( sub)exponentially in the distance of @xmath180 and @xmath181 . as for the second term
, we first estimate @xmath258 and then use the cauchy - schwarz inequality with respect to the summation and expectation . it hence remains to estimate @xmath259 \leq { \mathbb{e}}\bigg[\sum_{e\in\sigma(h_\omega)\cap i}\frac{\operatorname{tr}(\chi_{\mathbf{x}}p_e(h_\omega)\chi_{\mathbf{x}})}{|e - z|^{sq}}\bigg ] \notag \\ & \qquad \leq 2\liminf_{\eta\downarrow0}\int_i |e - z|^{-sq}{\mathbb{e}}\big[\operatorname{tr}(\chi_{\mathbf{x}}f_\eta(h_\omega - e)\chi_{\mathbf{x}})\big]{\mathrm{d}e}\ , , \end{aligned}\ ] ] and likewise for @xmath260 . in the last step @xmath261
denotes a suitable approximating @xmath262-function with compact support .
an upper bound now follows from assumption ( w ) ( using that , for small @xmath263 , @xmath264 with a slightly enlarged interval @xmath265 ) and the fact that @xmath266 . in order to establish eigenfunction correlator decay using that of the fractional moment we have [ thm : generalfmimpliesec ] in the setting ( h ) , ( o ) , ( m ) , assume that ( l ) and ( gw ) hold . then for any bounded interval @xmath121
if there exist @xmath267 , @xmath84 and @xmath82 $ ] such that @xmath268 \leq ae^{-\nu\operatorname{d}({\mathbf{x}},{\mathbf{y}})^\gamma}\ ] ] for all @xmath163 , then there exist @xmath243 such that @xmath269 \leq ce^{-\mu\operatorname{d}({\mathbf{x}},{\mathbf{y}})^\gamma}\ ] ] for all @xmath163 .
this theorem is also of interest in the one - particle case for which it constitutes an alternative to the steps taken in @xcite . in this case , the locality requirement ( l ) is clearly satisfied with @xmath270 the set of all bounded , open sets in @xmath271 .
assumption ( gw ) with @xmath272 the euclidean distance boils down to the `` independence at a distance '' ( cf .
@xcite ) of the basic random variables together with a wegner estimate involving local averages only .
similarly as in the proof of theorem [ thm : generalecimpliesfm ] , it is more convenient prove a bound on @xmath273 .
\label{efcorrsq}\ ] ] the additional exponent of @xmath274 can be justified as before .
we apply assumption ( l ) and ( gw ) with @xmath275 and infer for any eigenvalue @xmath276 of @xmath133 and a fixed @xmath277 @xmath278 - i { \varepsilon})\ , \xi_{\omega,\omega_{{\mathbf{x}},{\mathbf{y}}}}p_e(h_\omega)\end{aligned}\ ] ] and hence @xmath279 we insert this into and use that @xmath280 , such that @xmath281 \\
c_i^s \ , { \mathbb{e}}\bigg[\sum_{e\in\sigma(h_\omega)\cap i } \|\chi_{\mathbf{u}}(h_{\omega_{{\mathbf{x}},{\mathbf{y}}}}-e - i{\varepsilon})^{-1}{\boldsymbol{1}}_{\gamma_{\omega,\omega_{{\mathbf{x}},{\mathbf{y}}}}}\|^s\|\chi_{\mathbf{w}}p_e(h_\omega)\|^2\bigg ] \\
+ { \varepsilon}^s\ , { \mathbb{e}}\bigg[\sum_{e\in\sigma(h_\omega)\cap i}\|\chi_{\mathbf{u}}(h_{\omega_{{\mathbf{x}},{\mathbf{y}}}}-e - i{\varepsilon})^{-1}\|^s\|\chi_{\mathbf{w}}p_e(h_\omega)\|^2\bigg ] .
\label{efdecayinefcorrsq}\end{gathered}\ ] ] now let @xmath261 be a compactly supported approximating @xmath262-function .
then @xmath282 \\
\leq 2\liminf_{\eta\downarrow0}\int_i { \mathbb{e}}\big[\|\chi_{\mathbf{u}}(h_{\omega_{{\mathbf{x}},{\mathbf{y}}}}-e - i{\varepsilon})^{-1}{\boldsymbol{1}}_{\gamma_{\omega,\omega_{{\mathbf{x}},{\mathbf{y}}}}}\|^s\operatorname{tr}(\chi_{\mathbf{w}}f_\eta(h_\omega - e))\big]{\mathrm{d}e}\ , . \end{gathered}\ ] ] we use that @xmath176 is measurable with respect to @xmath177 due to assumption ( gw)(ii ) and conclude @xmath283 \\ = { \mathbb{e}}\big[\|\chi_{\mathbf{u}}(h_{\omega_{{\mathbf{x}},{\mathbf{y}}}}-e - i{\varepsilon})^{-1}{\boldsymbol{1}}_{\gamma_{\omega,\omega_{{\mathbf{x}},{\mathbf{y}}}}}\|^s{\mathbb{e}}\big[\operatorname{tr}(\chi_{\mathbf{w}}f_\eta(h_\omega - e))\big|{\mathcal{f}}_{{\mathbf{x}},{\mathbf{y}}}\big]\big ] \\
\leq c_i\ , { \mathbb{e}}\big[\|\chi_{\mathbf{u}}(h_{\omega_{{\mathbf{x}},{\mathbf{y}}}}-e - i{\varepsilon})^{-1}{\boldsymbol{1}}_{\gamma_{\omega,\omega_{{\mathbf{x}},{\mathbf{y}}}}}\|^s\big].\end{gathered}\ ] ] we finally use the decay estimate on fractional moments and infer @xmath284 \leq ce^{-\mu\operatorname{d}({\mathbf{x}},{\mathbf{y}})^\gamma}.\ ] ] this is possible since we assumed in ( gw)(i ) that the set @xmath174 can be covered by unit balls around points @xmath285 satifying @xmath286 and the number of which is bounded polynomially in @xmath175 .
this yields a ( sub)exponential bound on the first term in .
the second term vanishes in the limit @xmath287 .
the preceding theorems do not contain any localization results for the infinite volume operator @xmath134 ( unless @xmath288 ) . in order to conclude such results from fractional moment bounds as in
, we need an additional assumption : 1 .
there is a sequence of sets @xmath289 , @xmath290 , which exhaust @xmath131 in the sense that for all @xmath291 there exists @xmath292 such that @xmath293 for all @xmath294 , and almost surely : 1 .
@xmath295 converges to @xmath134 in strong resolvent sense .
@xmath296 is compact for all @xmath198 and all bounded intervals @xmath234 .
this assumption is evidently satisfied in the case of the operator @xmath18 , for which on may choose @xmath297 .
[ thm : locinfvolume ] in the setting ( h ) , ( o ) and ( m ) , assume ( e ) and that ( l ) and ( gw ) also hold for @xmath298 .
suppose @xmath121 is a bounded interval and that there exist @xmath267 , @xmath84 and @xmath82 $ ] such that @xmath299 \leq ae^{-\nu\operatorname{d}({\mathbf{x}},{\mathbf{y}})^\gamma}\ ] ] for all @xmath163 .
then the following holds : 1 .
there exist @xmath243 such that for all @xmath163 @xmath300 \leq ce^{-\mu\operatorname{d}({\mathbf{x}},{\mathbf{y}})^\gamma},\ ] ] where the supremum is taken over all measurable functions on @xmath81 that are in modulus uniformly bounded by one .
the operator @xmath134 has almost surely only pure point spectrum in @xmath81 and @xmath301 \leq ce^{-\mu\operatorname{d}({\mathbf{x}},{\mathbf{y}})^\gamma } \label{eq : efcorrrd}\ ] ] for some @xmath243 and all @xmath163 .
3 . let @xmath302 be a function with @xmath303 . then all normalized eigenfunctions @xmath304 of @xmath134 with eigenvalue @xmath276 satisfy @xmath305 for all @xmath306 , all @xmath307 and some @xmath308 . here
@xmath309 and @xmath310 is chosen such that @xmath311 . 1 .
it follows directly from the estimate on the eigenfunction correlator in theorem [ thm : generalfmimpliesec ] that for all sets @xmath132 @xmath312 \leq ce^{-\mu\operatorname{d}({\mathbf{x}},{\mathbf{y}})^\gamma}\ , .\ ] ] this estimate can be extended to the whole space @xmath298 . in the case of continuous functions
@xmath313 with compact support , one can use that @xmath314 converges in strong resovent sense to @xmath134 .
an approximation argument extends the result to general measurable functions , cf .
@xcite*section 2.5 . 2 . from ( m ) , ( e ) and
( i ) it follows ( with @xmath315 ) that for all @xmath291 @xmath316 \leq \lim_{l\rightarrow\infty } \mkern-10mu \sum_{\substack{{\mathbf{u}}\in{\mathbb{z}^d}\\ b_{\frac{1}{2}}({\mathbf{u } } ) \cap \omega_l^c \neq \emptyset } } \mkern-10mu { \mathbb{e}}\big[\|\chi_{\mathbf{u}}e^{-ith}p_i(h)\chi_{\mathbf{x}}\|\big ] = 0.\ ] ] using the rage theorem ( cf .
@xcite*theorem 4.1.20 ) , we conclude that the spectrum of @xmath134 in @xmath81 is almost surely only pure point . now that we have established pure point spectrum also for the infinite volume operator @xmath134 , we conclude inequality using assumptions ( l ) and ( gw ) for @xmath298 along the same lines of reasoning as the proof of theorem [ thm : generalfmimpliesec ] .
3 . defining @xmath317 we have @xmath318<\infty$ ] by ( ii ) and ( m ) .
therefore , for any normalized eigenfunction @xmath304 with eigenvalue @xmath276 @xmath319 for all @xmath320 .
inequality follows by an adaption of @xcite*lemma 3.5 to our setting .
going back to our @xmath11-particle random schrdinger , theorems [ thm : generalecimpliesfm ] and [ thm : generalfmimpliesec ] imply the following : [ thm : equivalenceoffmandec ] let @xmath121 be a bounded interval . 1 .
if @xmath81 is a regime of @xmath11-particle fractional moment localization of order @xmath82 $ ] , then it is also a regime of @xmath11-particle eigenfunction correlator localization of order @xmath82 $ ] .
if @xmath81 is a regime of @xmath11-particle eigenfunction correlator localization of order @xmath82 $ ] , then any interval @xmath125 satisfying @xmath224 is a regime of @xmath11-particle fractional moment localization of order @xmath82 $ ] .
as a first step , we consider @xmath11-particle systems which can be divided into two non - interacting subclusters .
suppose we have a partition @xmath321 with non - empty sets @xmath322 and @xmath323 .
if there is no interaction between the particles with numbers in @xmath322 and particles with numbers in @xmath323 , respectively , the corresponding hamiltonian takes the form @xmath324 where the tensor product is to be understood in the sense that @xmath325 acts on the variables @xmath23 , @xmath326 , and @xmath327 acts on the variables @xmath328 , @xmath329 .
now assume that for any @xmath330 , @xmath331-particle fractional moment localization has been proven in the interval @xmath332 $ ] .
then it is natural to assume that we can also prove localization of @xmath333 for energies near @xmath334 , as both its subsystems are localized at the bottom of the spectrum .
however , we can not directly conclude fractional moment localization for @xmath333 , but instead take a detour via eigenfunction correlator localization .
the drawback of this method is that the length of the interval of localization becomes slightly smaller .
the decay of fractional moments can be proven not only with respect to the `` usual '' @xmath11-particle hausdorff distance @xmath7 but with respect to a distance @xmath335 that takes into account the decomposition into subclusters .
it is defined by @xmath336 where @xmath337 , @xmath338 etc .
the decay of fractional moments with respect to this distance will be necessary for the analysis in the following section .
[ thm : fmboundnonint ] assume that there exists @xmath339 such that for any @xmath340 the operator @xmath341 exhibits fractional moment localization of order @xmath9 $ ] in the interval @xmath342 $ ] . then for any @xmath343 , any @xmath84 and any partition @xmath344 @xmath345 \leq ce^{-\mu\operatorname{dist}_h^{(j , k)}({\mathbf{x}},{\mathbf{y}})^\zeta } \label{eq : thm : fmboundnonint}\ ] ] for some @xmath122 and all @xmath64 , where @xmath98 $ ] . due to theorem [ thm : generalfmimpliesec ] ( more specifically corollary [ thm : equivalenceoffmandec ] ) , we have @xmath346 \leq c_je^{-\nu_j\operatorname{dist}_h({\mathbf{x}}_j,{\mathbf{y}}_j)^\zeta } \label{nointec1}\ ] ] for all bounded open sets @xmath143 and all @xmath347 .
an analogous inequality holds for @xmath348 . for the tensor product sum @xmath333 of these operators
, we have @xmath349 where @xmath350 $ ] . as @xmath351 by proposition [ prop : gs ] , all eigenvalues @xmath352 and
@xmath353 satisfying @xmath354 belong to the intervals @xmath355 and @xmath356 , respectively .
the estimate and its analogue for @xmath327 thus allow us to infer @xmath357 \notag\\ & \leq { \mathbb{e}}\bigg[\sum_{e_j\in\sigma(h_{\omega^{\ # j}}^{(\ # j)})\cap i^{(\ # j)}}\mkern-25mu\|\chi_{{\mathbf{x}}_j } p_{e_j}(h_{\omega^{\ # j}}^{(\ # j)})\chi_{{\mathbf{y}}_j}\| \mkern-10mu\sum_{e_k\in\sigma(h_{\omega^{\ # k}}^{(\ # k)})\cap i^{(\ # k)}}\mkern-25mu\|\chi_{{\mathbf{x}}_k } p_{e_k}(h_{\omega^n}^{(\ # k)})\chi_{{\mathbf{y}}_k}\|\bigg ] \notag \\ & \leq c \ , e^{-\nu\max\{\operatorname{dist}_h({\mathbf{x}}_j,{\mathbf{y}}_j),\operatorname{dist}_h({\mathbf{x}}_k,{\mathbf{y}}_k)\}^\zeta}.\end{aligned}\ ] ] in the last step we used that each of the sums can be bounded deterministically , i.e. , for all @xmath358 with suitably large @xmath359 : @xmath360 theorem [ thm : generalecimpliesfm ] with @xmath361 as the pseudo - metric then shows that holds . in this case
, we do not have to cut off a part of @xmath362 at the lower endpoint as it is located at the bottom of the spectrum .
building on the result of the previous section , we will now deal with the fully - interacting @xmath11-particle system . for this purpose , we define @xmath363\ ] ] for any @xmath84 , @xmath364 and any bounded interval @xmath121 .
clearly , ( sub)exponential decay of this quantity with respect to @xmath198 is just another way to express fractional moment localization in @xmath81 ( with the same rate of decay ) .
the conclusion of theorem [ thm : fmboundnonint ] can then be formulated as @xmath365 where @xmath366 is defined in the same way as @xmath367 , but with @xmath18 and @xmath7 replaced by @xmath333 and @xmath368 , respectively . for the analysis in this section ,
it is not of importance if the interval @xmath81 is located at the bottom of the spectrum of the hamiltonian .
we will thus simply assume that holds for some bounded interval @xmath81 , some constants @xmath369 , @xmath370 $ ] and all partitions @xmath344 .
the key theorem is the following rescaling inequality . for technical purposes ,
we introduce a " safety distance @xmath371 which will be needed in some of our arguments .
[ thm : rescalingineq ] let @xmath81 be a bounded interval such that holds for some @xmath372 , some @xmath369 , @xmath370 $ ] , all @xmath373 and all partitions @xmath344
. then there are constants @xmath374 such that for any @xmath375 $ ] and all sufficiently large @xmath198 @xmath376 the constant @xmath377 is independent of @xmath198 . in the case of @xmath111 ,
the assumption that holds is meaningless .
the theorem is then understood to hold for any bounded interval @xmath81 . as will become clear in the proof , the second and
the third summand in the bound do not appear in the case @xmath111 .
once the theorem is proved , one just needs show that @xmath378 is sufficiently small on suitable initial lengths @xmath379 $ ] .
the following corollary then yields ( sub)exponential decay of @xmath367 .
the initial estimates for different regimes of energies and interaction are the subject of section [ section : initialestimates ] .
[ cor : rescalingimpliesdecay ] in the situation of theorem [ thm : rescalingineq ] , the following holds : 1 . assume that @xmath380 , where @xmath381 satisfies @xmath382 with @xmath383 .
furthermore , let @xmath384 , @xmath385 .
then for all sufficiently large @xmath386 the following holds : if @xmath387 for all @xmath388 $ ] , then @xmath389 for some @xmath390 and all @xmath391 with @xmath392 .
assume that @xmath393 and let @xmath384 , @xmath385 .
then for all sufficiently large @xmath386 the following holds : if @xmath387 for all @xmath394 $ ] , then @xmath395 for some @xmath390 and all @xmath391 .
* we abbreviate @xmath378 by @xmath396 .
+ let @xmath392 and @xmath397 .
our assumption implies that for all @xmath388 $ ] : @xmath398 assuming that @xmath399 is sufficiently large we will prove that implies that holds for all @xmath198 in the larger interval @xmath400 $ ] .
iteration of this step then yields the result for all @xmath391 .
+ assume that @xmath401 $ ] .
then and imply @xmath402 if @xmath386 was chosen sufficiently large .
note that this choice can be made independently of @xmath403 ( which itself depends on @xmath386 ) .
we conclude that holds for @xmath404 $ ] and hence the inequality is valid in a larger interval as mentioned above .
* we proceed very similarly to ( i ) .
choosing @xmath405 and @xmath406 we have for all @xmath394 $ ] : @xmath407 then for sufficiently large @xmath399 and all @xmath401 $ ] we have @xmath408 hence , holds for @xmath409 $ ] .
iteration yields the result for all @xmath391 .
we divide the proof of theorem [ thm : rescalingineq ] into several lemmas . throughout this section
we will use the letter `` @xmath410 '' for a generic positive constant which may change from line to line , but is independent of @xmath198 . as a first step ,
we show that it suffices to prove a bound on the quantity @xmath411 , which differs from @xmath367 by an additional restriction on the spatial domain @xmath158 .
more specifically , @xmath412,\ ] ] which ensures that a volume factor of @xmath413 in the estimates only yields an additional polynomial growth with respect to @xmath198 .
[ lemma : rescalingstepone ] let @xmath81 and @xmath414 be as in theorem [ thm : rescalingineq ] .
then for sufficiently large @xmath198 @xmath415 let @xmath180 , @xmath181 , @xmath158 and @xmath244 as in the defnition of @xmath416 .
we assume without loss of generality that @xmath417 and define @xmath418 since @xmath419 and hence @xmath420 for the cutoff function @xmath421 defined in , the geometric resolvent equation yields @xmath422\theta_{\mathbf{w}}(h_{\omega^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\| \notag \\ & \leq \sum_{\substack{{\mathbf{w}}\in{\mathbb{z}^{dn}}\\ \theta_{\mathbf{w}}\nabla\xi\neq0 } } \|\chi_{\mathbf{x}}(h_{\tilde{\omega}^n}^{(n)}-z)^{-1}\tilde{\theta}_{\mathbf{w}}(h_{0,\tilde{\omega}^n}^{(n)}+a)^{\frac{1}{2}}\| \notag\\ & \quad\quad\quad\times\|(h_{0,\tilde{\omega}^n}^{(n)}+a)^{-\frac{1}{2}}[h_{\omega^n}^{(n)},\xi]\| \cdot \|{\boldsymbol{1}}_{b_1({\mathbf{w}})}(h_{\omega^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\| \label{eq : leminserta}\end{aligned}\ ] ] where @xmath423 is a smooth partition of unity with @xmath424 , @xmath425 is a smooth function with @xmath426 and @xmath427 .
the second factor is non - random and bounded . taking the @xmath428th moment of a term and conditioning
, we arrive at @xmath429 \notag \\ & \leq { \mathbb{e}}\big[{\mathbb{e}}\big[\|\chi_{\mathbf{x}}(h_{\tilde{\omega}^n}^{(n)}-z)^{-1}\tilde{\theta}_{\mathbf{w}}(h_{0,\tilde{\omega}^n}^{(n)}+a)^{\frac{1}{2}}\|^{2s}\big|{\mathcal{f}}_{{\mathcal{i}}_{\mathbf{w}}}\big]^{\frac{1}{2 } } \notag \\ & \mkern100mu \times { \mathbb{e}}\big[\|{\boldsymbol{1}}_{b_1({\mathbf{w}})}(h_{\omega^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\|^{2s}\big|{\mathcal{f}}_{{\mathcal{i}}_{\mathbf{w}}}\big]^{\frac{1}{2}}\big ] , \label{proof : resamplingstepone:1}\end{aligned}\ ] ] where @xmath430 is the @xmath431-algebra generated by the random variables @xmath432 and the set @xmath433 is defined as follows . as @xmath434 , there is @xmath193 such that @xmath435 .
we then set @xmath436 the crucial point here ist that firstly , @xmath437 on @xmath438 and @xmath439 and secondly , @xmath440 on @xmath441 .
lemmas [ apriori ] and [ resampling ] hence become applicable and , using @xmath442 , we bound the right hand side of up to a multiplicative constant by @xmath443\end{aligned}\ ] ] where @xmath444 .
the set @xmath445 can be covered by at most @xmath446 balls of radius @xmath447 around configurations @xmath448 .
furthermore , each of these configurations satisfies @xmath449 . hence , @xmath450 \\ = { \mathbb{e}}\big[\|\chi_{\mathbf{x}}(h_{\tilde{\omega}^n}^{(n)}-z)^{-1}{\boldsymbol{1}}_{\xi_{\mathbf{w}}}\|^s\big ] \leq l^{dn}\tilde{b}_s^{(n)}(i,2l+2l^\alpha+6r ) .\end{gathered}\ ] ] combining all of the above , we arrive at @xmath451 & = \hat{{\mathbb{e}}}{\mathbb{e}}\big[\|\chi_{\mathbf{x}}(h_{\tilde{\omega}^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\|^s\big ] \notag \\ & \leq \sum_{\substack{{\mathbf{w}}\in{\mathbb{z}^{dn}}\\ \theta_{\mathbf{w}}\theta\neq 0 } } c l^{2dn}\tilde{b}_s^{(n)}(i,2l+2l^\alpha+6r ) \notag \\ & \leq cl^{3dn } \tilde{b}_s^{(n)}(i,2l+2l^\alpha+6r ) , \end{aligned}\ ] ] which concludes the proof of the lemma
. we can now proceed with a bound on @xmath411 , which yields theorem [ thm : rescalingineq ] if it is combined with lemma [ lemma : rescalingstepone ] .
[ lemma : rescalingsteptwo ] let @xmath81 and @xmath414 be as in theorem [ thm : rescalingineq ] .
then @xmath452 we split the proof of this lemma into two cases , depending on the diameters of the configurations @xmath180 and @xmath181 in @xmath453 $ ] , with @xmath180 , @xmath181 , @xmath244 and @xmath158 as in the definition of @xmath454 .
the diameter is defined as @xmath455 . in the case of one particle
, we trivially have @xmath456 and thus lemma [ lemma : rescalingsteptwoparttwo ] does not apply . as a consequence
, only the first summand appears in the bound .
[ lemma : rescalingsteptwopartone ] if @xmath457 , then @xmath458 \leq c l^{5dn}b_s^{(n)}(i , l+l^\alpha)^2.\ ] ]
let @xmath459 such that @xmath460 and define @xmath461 , @xmath462 .
an application of the geometric resolvent inequality with @xmath463 and @xmath464 then yields @xmath465(h_{\omega^n}^{(n)}-z)^{-1}[\psi_{\mathbf{y}},h_{\omega^n}^{(n)}](h_{\omega_{\mathbf{y}}^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\ ] ] and hence , proceeding as in by inserting the smooth partition of unity @xmath423 with @xmath425 denoting its enlarged version , @xmath466 \notag\\ & \leq \sum_{\substack{{\mathbf{w}},{\mathbf{v}}\in{\mathbb{z}^{dn}}\\ \theta_{\mathbf{w}}\nabla\psi_{\mathbf{x}}\neq 0 \\ \theta_{\mathbf{v}}\nabla\psi_{\mathbf{y}}\neq 0 } } { \mathbb{e}}\big[\|\chi_{\mathbf{x}}(h_{\omega_{\mathbf{x}}^n}^{(n)}-z)^{-1}\tilde{\theta}_{\mathbf{w}}(h_{0,\omega_{\mathbf{x}}^n}^{(n)}+a)^{\frac{1}{2}}||^s \|(h_{0,\omega_{\mathbf{x}}^n}^{(n)}+a)^{-\frac{1}{2}}[h_{\omega^n}^{(n)},\psi_{\mathbf{x}}]\|^s \notag\\ & \hphantom{\leq \sum_{\substack{{\mathbf{w}},{\mathbf{v}}\in{\mathbb{z}^{dn}}\\ \theta_{\mathbf{w}}\nabla\psi_{\mathbf{x}}\neq 0 \\ \theta_{\mathbf{v}}\nabla\psi_{\mathbf{y}}\neq 0 } } { \mathbb{e}}\ } \times\|\theta_{{\mathbf{w}}}(h_{\omega^n}^{(n)}-z)^{-1}\theta_{\mathbf{v}}\|^s \|[\psi_{\mathbf{y}},h_{\omega^n}^{(n)}](h_{0,\omega_{\mathbf{y}}^n}^{(n)}+a)^{-\frac{1}{2}}\|^s \notag\\ & \hphantom{\leq \sum_{\substack{{\mathbf{w}},{\mathbf{v}}\in{\mathbb{z}^{dn}}\\ \theta_{\mathbf{w}}\nabla\psi_{\mathbf{x}}\neq 0 \\ \theta_{\mathbf{v}}\nabla\psi_{\mathbf{y}}\neq 0 } } { \mathbb{e}}\ } \times\|(h_{0,\omega_{\mathbf{y}}^n}^{(n)}+a)^{\frac{1}{2}}\tilde{\theta}_{\mathbf{v}}(h_{\omega_{\mathbf{y}}^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\|^s\big ] \notag\\ & \leq c\sum_{\substack{{\mathbf{w}},{\mathbf{v}}\in{\mathbb{z}^{dn}}\\ \theta_{\mathbf{w}}\nabla\psi_{\mathbf{x}}\neq 0 \\ \theta_{\mathbf{v}}\nabla\psi_{\mathbf{y}}\neq 0 } } { \mathbb{e}}\big[{\mathbb{e}}\big[\|\chi_{\mathbf{x}}(h_{\omega_{\mathbf{x}}^n}^{(n)}-z)^{-1}\tilde{\theta}_{\mathbf{w}}(h_{0,\omega_{\mathbf{x}}^n}^{(n)}+a)^{\frac{1}{2}}||^{3s } \big| { \mathcal{f}}_{{\mathcal{i}}_{\mathbf{w}}\cup{\mathcal{i}}_{\mathbf{v}}}\big]^{\frac{1}{3 } } \notag\\ & \hphantom{\leq c\sum_{\substack{{\mathbf{w}},{\mathbf{v}}\in{\mathbb{z}^{dn}}\\ \theta_{\mathbf{w}}\nabla\psi_{\mathbf{x}}\neq 0 \\ \theta_{\mathbf{v}}\nabla\psi_{\mathbf{y}}\neq 0 } } { \mathbb{e}}\ } \times { \mathbb{e}}\big[\|\theta_{{\mathbf{w}}}(h_{\omega^n}^{(n)}-z)^{-1}\theta_{\mathbf{v}}\|^{3s } \big| { \mathcal{f}}_{{\mathcal{i}}_{\mathbf{w}}\cup{\mathcal{i}}_{\mathbf{v}}}\big]^{\frac{1}{3 } } \notag\\ & \hphantom{\leq c\sum_{\substack { { \mathbf{w}},{\mathbf{v}}\in{\mathbb{z}^{dn}}\\ \theta_{\mathbf{w}}\nabla\psi_{\mathbf{x}}\neq 0 \\ \theta_{\mathbf{v}}\nabla\psi_{\mathbf{y}}\neq 0 } } { \mathbb{e}}\ } \times { \mathbb{e}}\big[\|(h_{0,\omega_{\mathbf{y}}^n}^{(n)}+a)^{\frac{1}{2}}\tilde{\theta}_{\mathbf{v}}(h_{\omega_{\mathbf{y}}^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\|^{3s } \big| { \mathcal{f}}_{{\mathcal{i}}_{\mathbf{w}}\cup{\mathcal{i}}_{\mathbf{v}}}\big]^{\frac{1}{3}}\big ] , \label{proof : rescalingsteptwopartone:1}\end{aligned}\ ] ] where @xmath467 with @xmath468 chosen such that @xmath469 and @xmath470 . by an application of lemma [ apriori ] ( bound on fractional moments ) and lemma [ resampling ] ( resampling ) , each of the summands above can be bounded up to a multiplicative constant by @xmath471^{\frac{1}{3 } } \notag\\ & \qquad\times { \mathbb{e}}\big[\|(h_{0,\omega_{\mathbf{y}}^n}^{(n)}+a)^{\frac{1}{2 } } \tilde{\theta}_{\mathbf{v}}(h_{\omega_{\mathbf{y}}^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\|^{3s } \big| { \mathcal{f}}_{{\mathcal{i}}_{\mathbf{v}}}\big]^{\frac{1}{3}}\big ] \notag\\ & \leq cl^{dn}\ , { \mathbb{e}}\bigg[\mkern-10mu\sum_{\substack{{\mathbf{a}}\in{\mathbb{z}^{dn}}\\ { \boldsymbol{1}}_{b_3({\mathbf{a } } ) } \sum_{\zeta\in{\mathcal{i}}_{\mathbf{w}}}n_\zeta\neq0 } } \mkern-25mu\|\chi_{\mathbf{x}}(h_{\omega_{\mathbf{x}}^n}^{(n)}-z)^{-1}\chi_{\mathbf{a}}\|^s \mkern-30mu\sum_{\substack{{\mathbf{b}}\in{\mathbb{z}^{dn}}\\ { \boldsymbol{1}}_{b_3({\mathbf{b}})}\sum_{\zeta\in{\mathcal{i}}_{\mathbf{v}}}n_\zeta\neq0 } } \mkern-30mu\|\chi_{\mathbf{b}}(h_{\omega_{\mathbf{y}}^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\|^s\bigg ] \notag\\ & = cl^{dn } \sum_{\substack { { \mathbf{a}},{\mathbf{b}}\in{\mathbb{z}^{dn}}\\ { \boldsymbol{1}}_{b_3({\mathbf{a } } ) } \sum_{\zeta\in{\mathcal{i}}_{\mathbf{w } } } n_\zeta\neq0 \\ { \boldsymbol{1}}_{b_3({\mathbf{b } } ) } \sum_{\zeta\in{\mathcal{i}}_{\mathbf{v } } } n_\zeta\neq0}}\mkern-20mu { \mathbb{e}}\big[\|\chi_{\mathbf{x}}(h_{\omega_{\mathbf{x}}^n}^{(n)}-z)^{-1}\chi_{\mathbf{a}}\|^s ] \,{\mathbb{e}}\big[\|\chi_{\mathbf{b}}(h_{\omega_{\mathbf{y}}^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\|^s\big].\end{aligned}\ ] ] note that the potential on @xmath472 and the potential on @xmath473 are stochastically independent , which justifies the last equality .
it is easy to see that @xmath474 and @xmath475 for any @xmath476 and @xmath477 appearing in the above sum and hence @xmath478\ , { \mathbb{e}}\big[\|\chi_{\mathbf{b}}(h_{\omega_{\mathbf{x}}^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\|^s\big ] \leq b_s^{(n)}(i , l+l^\alpha)^2.\ ] ] taking into account the number of terms in , we arrive at @xmath458\leq cl^{5dn}b_s^{(n)}(i , l+l^\alpha)^2\ ] ] which completes the proof .
[ lemma : rescalingsteptwoparttwo ] if @xmath479 or @xmath480 , then @xmath458 \leq c\bigg(e^{-\nu l^{\alpha\gamma^*}}+ l^{(2+\alpha)nd}\bigg(w_b\bigg(\frac{l^\alpha}{4n}\bigg)\bigg)^s b_s^{(n)}(i,2l)\bigg).\ ] ] without loss of generality , we can assume that @xmath481 .
( here it does not matter whether @xmath482 or @xmath483 in the definition of @xmath484 . ) additionally , we choose a partition @xmath344 such that @xmath485 which is possible since @xmath481 ( cf .
@xcite*lemma a.1 ) , and assume without loss of generality that @xmath486 the latter can be justified as follows : in case there is @xmath487 such that @xmath488 the claim is obvious ( for @xmath489 playing the role of @xmath490 ) .
else there is @xmath487 such that @xmath491 and we have @xmath492 for all @xmath493 .
this yields the claim if @xmath494 .
if @xmath495 , one may simply interchange the roles of @xmath180 and @xmath181 in this case .
an application of the resolvent equation with @xmath496 ( cf . ) and yields @xmath497 & \leq { \mathbb{e}}\big[\|\chi_{\mathbf{x}}(h_{\omega^n}^{(j , k)}-z)^{-1}\chi_{\mathbf{y}}\|^s\big ] \notag\\ & \quad+ { \mathbb{e}}\big[\|\chi_{\mathbf{x}}(h_{\omega^n}^{(j , k)}-z)^{-1}w^{(j , k)}_{\omega^n}(h_{\omega^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\|^s\big ] \notag \\ & \leq a^*e^{-\nu^*l^{\gamma^ * } } + ( i ) + ( ii)\end{aligned}\ ] ] where @xmath498^{\frac{1}{2 } } \|w^{(j , k)}_{\omega^n}\chi_{\mathbf{w}}\|_\infty^s { \mathbb{e}}\big[\|\chi_{\mathbf{w}}(h_{\omega^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\|^{2s}\big]^{\frac{1}{2}}\end{aligned}\ ] ] and @xmath499.\end{aligned}\ ] ] as for @xmath500 , we use the boundedness of the last two factors and . more precisely , the `` one - for - all principle '' ( cf .
@xcite*lemma 4.2 ) allows one to conclude also for @xmath501 ( with an altered decay exponent ) such that @xmath502 \leq ce^{-\mu\operatorname{dist}_h^{(j , k)}({\mathbf{x}},{\mathbf{w}})^{\gamma * } } \leq ce^{-\tilde{\mu}l^{\alpha\gamma * } } \ , .\ ] ] the last step follows from @xmath503 and is used for part of the decay only .
the other part controls the sum , and we arrive at @xmath504 as for @xmath505 , we proceed differently .
let @xmath506 such that @xmath507 .
assuming without loss of generality that @xmath508 , we can infer that there exists @xmath326 such that @xmath509 . as a consequence ,
@xmath510 now define @xmath511 due to the a priori bound on fractional moments in lemma [ apriori ] and the resampling estimate in lemma [ resampling ] , we infer that @xmath512 \notag\\ & \leq { \mathbb{e}}\big[{\mathbb{e}}\big[\|\chi_{\mathbf{x}}(h_{\omega^n}^{(j , k)}-z)^{-1}\chi_{\mathbf{w}}\|^{2s}\big|{\mathcal{f}}_{{\mathcal{i}}_{x_1,w_j}}\big]^{\frac{1}{2}}{\mathbb{e}}\big[\|\chi_{\mathbf{w}}(h_{\omega^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\|^{2s}\big|{\mathcal{f}}_{{\mathcal{i}}_{x_1,w_j}}\big]^{\frac{1}{2}}\big ] \notag \\ & \leq c\ , { \mathbb{e}}\big[{\mathbb{e}}\big[\|\chi_{\mathbf{w}}(h_{\omega^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\|^{2s}\big|{\mathcal{f}}_{{\mathcal{i}}_{x_1,w_j}}\big]^{\frac{1}{2}}\big ] \notag \\ & \leq cl^{dn}\sum_{\substack{{\mathbf{u}}\in{\mathbb{z}^{dn}}\\ { \boldsymbol{1}}_{b_3({\mathbf{u}})}\sum_{\zeta\in{\mathcal{i}}_{x_1,w_j}}n_\zeta\neq0 } } { \mathbb{e}}\big[\|\chi_{\mathbf{u}}(h_{\omega^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\|^s\big ] \notag \\ & \leq cl^{2nd } b_s^{(n)}(i,2l).\end{aligned}\ ] ] in the last step we used that the number of contributing summands is of order @xmath446 as @xmath158 is a subset of @xmath513 .
we conclude that @xmath514 for any @xmath515 in the sum above , there exist @xmath516 and @xmath517 such that @xmath518 .
additionally , from @xmath519 , it follows that there exist @xmath520 and @xmath521 such that @xmath522 . as a consequence , using @xmath523 and hence @xmath524 we conclude @xmath525 this finishes the proof . combining the lemmas in this section
finally yields the claim of theorem [ thm : rescalingineq ] .
we proceed with the proof of the initial length scale estimate that is assumed in corollary [ cor : rescalingimpliesdecay ] , i.e. , the estimate @xmath387 .
we state two possible situations in the following theorem : [ thm : initialestimates ] 1 .
let @xmath384 and @xmath385 .
suppose @xmath81 is an interval such that for sufficiently large @xmath198 @xmath526 for the non - interacting @xmath11-particle system , i.e. , @xmath527 .
then for any sufficiently large @xmath386 there exists @xmath528 such that , if @xmath529 $ ] , then @xmath530 for all @xmath394 $ ] .
2 . let @xmath384 and @xmath385 and suppose that for some @xmath531 there exists @xmath532 such that @xmath533 for @xmath534 and any @xmath74 .
then for any sufficiently large @xmath386 there exists @xmath106 such that @xmath530 for all @xmath394 $ ] , where @xmath535 $ ] .
the proof of ( i ) is a perturbation argument .
we apply the resolvent equation , use and obtain @xmath497 & \leq { \mathbb{e}}\big[\|\chi_{\mathbf{x}}(h_{\omega^n,\alpha_w=0}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\|^s\big ] \notag\\ & \quad+ { \mathbb{e}}\big[\|\chi_{\mathbf{x}}(h_{\omega^n,\alpha_w=0}^{(n)}-z)^{-1}\alpha_w w(h_{\omega^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\|^s\big ] \notag\\ & \leq \frac{1}{2}c'l^{-q ' } + \alpha_w^s c\end{aligned}\ ] ] for @xmath536 with @xmath537 sufficiently large and @xmath143 open , bounded , @xmath538 , @xmath539 .
given @xmath399 sufficiently large and @xmath540 we can hence conclude @xmath541 for all @xmath394 $ ] .
we proceed with the proof of ( ii ) .
let @xmath74 and @xmath542 , @xmath543 , be the connected components of @xmath544 . clearly , @xmath545 for some @xmath546 .
then for any open set @xmath547 , we have @xmath548 where the first inequality follows from the monotonicity of the ground state energy in the domain and the last inequality follows from for sufficiently large @xmath198 . pick @xmath549 such that @xmath536 and @xmath550 with @xmath551 $ ] , @xmath539 .
then @xmath552 \notag \\ & \leq { \mathbb{e}}\big[\|\chi_{\mathbf{x}}(h_{\omega^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\|^s { \boldsymbol{1}}_{\mathbf{\omega}_b}\big ] + { \mathbb{e}}\big[\|\chi_{\mathbf{x}}(h_{\omega^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\|^s { \boldsymbol{1}}_{\mathbf{\omega_b^c}}\big ] \notag\\ & \leq { \mathbb{e}}\big[\|\chi_{\mathbf{x}}(h_{\omega^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\|^{2s}\big]^{\frac{1}{2 } } { \mathbb{p}}\big(\mathbf{\omega}_b\big)^{\frac{1}{2 } } + { \mathbb{e}}\big[\|\chi_{\mathbf{x}}(h_{\omega^n}^{(n)}-z)^{-1}\chi_{\mathbf{y}}\|^s { \boldsymbol{1}}_{\mathbf{\omega}_b^c}\big],\end{aligned}\ ] ] where @xmath553 . the first term is bounded by @xmath554 , whereas the second term can be bounded with the help of the combes - thomas estimate ( cf .
@xcite*theorem 1 ) , since for @xmath555 we have @xmath556 and hence @xmath557 as a consequence , we have @xmath558,l ) \leq \tilde{c}l^{-\frac{\xi}{2}}\ ] ] for sufficiently large @xmath198 .
similarly to lemma [ lemma : rescalingstepone ] , we can conclude @xmath559,l ) \leq cl^{3dn-\frac{\xi}{2}}\ ] ] for large @xmath198 . as @xmath531 , this yields the claim of ( ii ) . 1 .
the assumption of theorem [ thm : main : case1](i ) yields the basis of an induction on @xmath11 , i.e. , @xmath560,l ) \leq c^{(1)}e^{-\mu^{(1 ) } l^{\gamma_w}}\ ] ] for all @xmath561 . given any @xmath95 , the induction step proceeds as follows : assuming that we have @xmath562,l ) \leq c^{(m)}e^{-\mu^{(m)}l^{\gamma_w}}\ ] ] for some @xmath563 and
all @xmath330 , theorem [ thm : fmboundnonint ] guarantees the applicability of theorem [ thm : rescalingineq ] with @xmath564 $ ] .
theorem [ thm : fmboundnonint ] also implies the ( sub)exponential decay of the fractional moments of the resolvent of the non - interacting system .
in particular , the assumption of theorem [ thm : initialestimates](i ) is satisfied and we apply corollary [ cor : rescalingimpliesdecay](ii ) to conclude that @xmath565,l ) \leq c^{(n)}e^{-\mu^{(n)}l^{\gamma_w}}\ ] ] if @xmath566 is sufficiently small .
this yields the claim of theorem [ thm : main : case1](i ) .
2 . in the case
@xmath111 and as a basis of the induction , the assertion follows by combining theorem [ thm : initialestimates](ii ) and corollary [ cor : rescalingimpliesdecay](ii ) , where we note that the assumption of theorem [ thm : rescalingineq ] is trivially satisfied for @xmath567 . as in the proof of ( i ) , we proceed with the induction step and assume a bound on @xmath568,l)$ ] for some @xmath569 and all @xmath330 . as a consequence
, we infer the condition of theorem [ thm : rescalingineq ] with @xmath570 $ ] with some @xmath571 . if we choose @xmath572 and @xmath385 such that @xmath573 ( which is possible due to the assumption on @xmath574 ) , theorem [ thm : initialestimates](ii ) ensures the existence of an @xmath575 such that corollary [ cor : rescalingimpliesdecay](ii ) can be applied with @xmath564 $ ] , which concludes the proof .
the proof proceeds in the same fashion as the proof of theorem [ thm : main : case1 ] .
therefore , we focus on the differences only .
we conduct the induction step as long as @xmath576 with @xmath577 .
the @xmath578 in the rescaling inequality in theorem [ thm : rescalingineq ] is chosen such that the resulting @xmath579 with @xmath580 in corollary [ cor : rescalingimpliesdecay](i ) is maximal .
this is the case if @xmath581 , i.e. , @xmath582 and hence @xmath583 .
the induction basis holds with @xmath118 $ ] and consequently , for @xmath584 we have @xmath585 and @xmath119 .
in particular , @xmath586 , so our condition on @xmath11 ensures the applicability of corollary [ cor : rescalingimpliesdecay](i ) .
apart from these considerations , the proof proceeds analogously to the proof of theorem [ thm : main : case1 ] .
in this section , we present some auxiliary results that are essential for our analysis of the multiparticle system .
most of these are familiar from the one - particle case ( cf .
we state the results in a more general setting : * the random operator @xmath587 acts on @xmath129 , where @xmath588 is arbitrary . *
the potential @xmath589 is not necessarily an interaction potential , but can be an arbitrary bounded background potential . *
the random potential @xmath590 takes the form @xmath591 where @xmath592 is an arbitrary countable index set , the functions @xmath593 are measurable , non - negative and satisfy @xmath594 and the random variables @xmath35 , @xmath595 , are independent and identically distributed .
the distribution of @xmath35 has a bounded density @xmath37 with a compact support that is a subset of the non - negative half - line @xmath596 .
in addition , we introduce some notation . for any set @xmath597
, @xmath598 is the @xmath166-algebra generated by the random variables @xmath599 .
the infimum of the spectrum of @xmath600 is denoted by @xmath601 .
the first lemma is basically a reformulation of @xcite*lemma 3.3 in the above setting .
its message is the following : suppose @xmath602 are bounded open sets and @xmath597 such that the sum @xmath603 is bounded from below by a positive number on a neighborhood of these sets .
then it suffices to average over the `` local '' random variables @xmath35 , @xmath604 , in order that a fractional moment of @xmath605 , @xmath191 , to be bounded uniformly with respect to @xmath606 .
[ apriori ] let @xmath607 be open and bounded sets and let @xmath608 be finite sets and @xmath609 such that @xmath610 for both @xmath611 and @xmath612 .
suppose @xmath613 such that @xmath614
. then there exists @xmath377 such that for any open set @xmath615 , @xmath84 and @xmath616 @xmath617 \\
\leq \frac{c^s}{1-s}(1+|\lambda_1|^{\frac{1}{2}}|\lambda_2|^{\frac{1}{2}})^s ( \#({\mathcal{i}}_1\cup{\mathcal{i}}_2))^{(2d+8)s } \bigg(1+|z - e_0|+\delta^{-2}\bigg)^{(d+3)s}.\end{gathered}\ ] ] we only give a sketch of the proof , as it does not differ much from the one found in @xcite .
we pick the least integer @xmath618 and choose a family of smooth cutoff functions @xmath619 satisfying @xmath620 , @xmath621 on @xmath622 for @xmath623 , @xmath624 on @xmath625 and @xmath626 .
define the operator @xmath627 and let @xmath628 . by a repeated application of the identity @xmath629)\theta_{j+1}^2(h - z)^{-1 } , \end{aligned}\ ] ]
one can see that @xmath630 where @xmath631 is a bounded operator and @xmath632 is a hilbert - schmidt operator . proceeding similarly to the above , we can show that @xmath633 where @xmath634 , @xmath635 , @xmath636 and @xmath637 are analogous to @xmath638 , @xmath639 , @xmath631 and @xmath632 , respectively , but with @xmath640 and @xmath641 playing the roles of @xmath625 and @xmath642 , respectively . combining this with yields @xmath643 with @xmath644 .
the operators @xmath645 , @xmath646 are polynomials in the random variables @xmath647 .
they can be written as @xmath648 where @xmath649 is the absolute value of the multiindex @xmath414 and each operator @xmath650 is hilbert - schmidt and independent of @xmath647 .
define @xmath651 for @xmath646 . as
@xmath652 is strictly positve on @xmath653 , we can insert @xmath654 and @xmath655 next to @xmath639 and @xmath635 , respectively , into and obtain @xmath656 using this and , we estimate @xmath657 where @xmath658 is the total number of terms in . as @xmath659 is uniformly bounded , the first term can be bounded by @xmath660 for a suitable @xmath661 . as for the other terms in
, we employ proposition [ weakl1estimate ] below . in order to estimate a term from above , we first integrate over the random variables @xmath35 and @xmath662 . for this purpose ,
we employ or , depending on whether @xmath230 is equal to @xmath663 .
this yields a weak @xmath664-bound as above .
we can therefore conclude @xmath665 for suitable constants @xmath666 .
this implies the claim of the lemma , since for a random variable @xmath667 the weak @xmath664-estimate @xmath668 implies @xmath669 \leq c^s/(1-s)$ ] .
the specific form of the claimed bound can be obtained by a careful analysis of the proof .
we omit this analysis as it is mostly just a repetition of arguments in the proof of @xcite*lemma 3.3 .
the proof above used the following proposition , which was proved in @xcite .
( actually , only the second inequality of the proposition was stated in @xcite .
however , the proof of the first one is more or less contained in the proof of the second one . )
[ weakl1estimate ] suppose @xmath670 is a separable hilbert space and @xmath671 a maximally dissipative operator on @xmath670 with strictly positive imaginary part . if @xmath672 are hilbert - schmidt operators and @xmath673 non - negative operators , then for all @xmath674 @xmath675\bigg|\,\bigg\|m_1u_1^{\frac{1}{2}}(a - vu_1)^{-1}u_1^{\frac{1}{2}}m_2\bigg\|_{\mathrm{hs}}>t\bigg\}\bigg| \leq c_w\|m_1\|_{\mathrm{hs}}\|m_2\|_{\mathrm{hs}}\frac{1}{t } \label{weakl1estimate1}\ ] ] and @xmath676 ^ 2\bigg|\,\bigg\|m_1u_1^{\frac{1}{2}}(a - v_1u_1-v_2u_2)^{-1}u_2^{\frac{1}{2}}m_2\bigg\|_{\mathrm{hs}}>t\bigg\}\bigg| \\
\leq 2c_w\lambda\|m_1\|_{\mathrm{hs}}\|m_2\|_{\mathrm{hs}}\frac{1}{t}. \label{weakl1estimate2}\end{gathered}\ ] ] the next lemma is a reformulation of @xcite*lemma 3.4 .
the setting is similar to lemma [ apriori ] .
a resolvent @xmath677 is again bracketed between two functions supported on bounded sets .
however , in this case , we require that @xmath603 covers only a neighborhood of one of these sets . as a result , averaging over the random variables @xmath35 , @xmath604 , yields a bound that includes the resolvent of an operator @xmath678 , which is the hamiltonian @xmath134 with `` resampled '' random variables @xmath679 , @xmath604 . in order to state this more rigorously , we introduce some new notation : let @xmath679 , @xmath595 , be a collection of random variables with the same distribution as ( but independent of ) the random variables @xmath35 , @xmath595 . for any @xmath597 , we denote by @xmath680 the resampled operator @xmath681 . [ resampling ]
let @xmath607 be disjoint , open and bounded sets , and let @xmath597 be a finite set and @xmath609 such that @xmath682 on @xmath641 and @xmath683 on @xmath684 .
suppose @xmath685 is a smooth function supported in an open set @xmath686 satisfying @xmath687 for some @xmath235
. then for all @xmath84 , there exists @xmath377 such that for all open and bounded sets @xmath615 @xmath688,{\mathbb{e}}\big[\|{\boldsymbol{1}}_{\lambda_1}(h_\omega - z)^{-1}\psi(h_{w,\omega}+a)^{1/2}\|^s\big|{\mathcal{f}}_{{\mathcal{i}}'}\big]\right\ } \notag\\ & \leq \frac{1}{1-s}c^s(\#{\mathcal{i}}')^{(3d+10)s}(1+|b_{2\delta}(\omega\cap\operatorname{supp}{\textstyle \sum_{\zeta\in{\mathcal{i}}'}n_\zeta})|)^s(1+|z - e_0|+\delta^{-2})^{(d+3)s } \notag\\ & \qquad\times\|{\boldsymbol{1}}_{\lambda_1}({\widehat{h}}_{{\mathcal{i}}',\omega}-z)^{-1}{\boldsymbol{1}}_\xi\|^s , \label{eq : resampling}\end{aligned}\ ] ] where @xmath628 and @xmath689 .
the proof follows closely that of @xcite*lemma 3.4 . due to the assumption that @xmath690 , we have @xmath691 and thus @xmath692)(h_w+a)^{-1/2 } \ , \ ] ] where @xmath693 is a suitable smooth function with @xmath694 , @xmath695 on @xmath640 and @xmath696
this yields the estimate @xmath697 with a suitable constant @xmath410 . in order to prove , we thus need to find a bound on @xmath698 $ ] .
an application of the resolvent equality yields @xmath699 applying an identity of the same form as , we get @xmath700 where @xmath701 is bounded , @xmath702 is hilbert - schmidt and @xmath703 is a smooth function such that @xmath704 , @xmath705 on @xmath706 and @xmath707 .
therefore , @xmath708 where @xmath709 is a smooth cutoff function satisfying @xmath710 on @xmath711 and @xmath712 if @xmath713 .
next we apply the commutator argument to @xmath714 and obtain @xmath715 with a hilbert - schmidt operator @xmath716 , a smooth function @xmath717 with @xmath718 and a bounded operator @xmath719 . combining these identities and using that @xmath720 , we conclude @xmath721 and hence @xmath722 where @xmath723 . taking the @xmath724 power and averaging over the random variables @xmath35 , @xmath604 , as in the proof of lemma [ apriori ] now yields @xmath725 \notag\\ & \leq c_0^s(\#{\mathcal{i}}')^{(d/2 + 1)s}(1+|z - e_0|+\delta^{-2})^{(d/2 + 1)s}\| { \boldsymbol{1}}_{\lambda_1}({\widehat{h}}_{{\mathcal{i}}'}-z)^{-1}\widehat{\psi}^2\|^s \notag\\ & \quad+\frac{1}{1-s}c_1^s(\#{\mathcal{i}}')^{(3d+10)s}(1+|b_{2\delta}(\omega\cap\operatorname{supp}{\textstyle
\sum_{\zeta\in{\mathcal{i}}'}n_\zeta})|)^s(1+|z - e_0|+\delta^{-2})^{(d+2)s } \notag\\ & \qquad\times\|{\boldsymbol{1}}_{\lambda_1}({\widehat{h}}_{{\mathcal{i}}'}-z)^{-1}\tilde{\theta}\|^s\end{aligned}\ ] ] in our analysis in section [ section : fmec ] the assumptions ( w ) and ( gw)(iii ) take the form of a wegner estimate .
the following lemma ensures that these assumptions are indeed valid in our model . in comparison to the wegner estimates in @xcite , the important feature of the lemma
is that it requires only a `` local '' average , similarly to the apriori bound on fractional moments of the resolvent in lemma [ apriori ] .
more specifically , given a bounded set @xmath726 , we average only over a collection of random variables @xmath35 with the property that the corresponding potentials @xmath593 cover a neighborhood of @xmath727 .
the lemma shows that in this case the conditional expectation of the trace of @xmath728 is bounded by a constant times the lebesgue measure of @xmath81 , for any borel set @xmath121 whose supremum does not exceed a fixed bound .
[ lemma : exptrproj ] let @xmath729 be open and bounded and let @xmath597 be a finite set such that @xmath730 for some @xmath609 .
suppose @xmath731 is an open set such that @xmath732 .
then there exists @xmath733 such that for any open set @xmath615 @xmath734\leq c(\#{\mathcal{i}}')^{d+4 } \sup_{e\in i}|e+a|^{d+2}(1+\delta^{-2})^{d+2}|\lambda||i| .
\label{exptrprojineq}\ ] ] choose a sequence of functions @xmath735 as in the proof of lemma [ apriori ] , satisfying @xmath736 on @xmath727 and @xmath737 .
let @xmath628 and define @xmath738 .
then for all @xmath739 , we have @xmath740\frac{1}{h'+a}\theta_j \notag\\ & = ( h+a)p_i(h)\theta_{j+1}^2\theta_j\frac{1}{h'+a}\theta_j -\sum_{\zeta\in{\mathcal{i } } ' } p_i(h)\theta_{j+1}^2 \eta_\zeta n_\zeta\theta_j\frac{1}{h'+a}\theta_j \notag\\ & \quad + p_i(h)\theta_{j+1}^2[\theta_j , h']\frac{1}{h'+a}\theta_j .\end{aligned}\ ] ] by induction , we obtain @xmath741 where @xmath742 , @xmath743 , @xmath744 , and @xmath745 is a product of @xmath331 terms of the form @xmath746\frac{1}{h'+a}\theta_j \ .\ ] ] each of these operators is in the schatten class @xmath747 for any @xmath748 .
since @xmath618 , @xmath745 is a hilbert - schmidt operator . as a consequence
, we have @xmath749 due to the assumption on @xmath750 , we can insert @xmath751 with @xmath752 next to @xmath753 and obtain @xmath754 taking the expected value , conditioned on @xmath598 , and using hlder s inequality , we get @xmath755 \leq c\sup_{e\in i}|e+a|^m \sum_l\sum_{\zeta\in{\mathcal{i}}'}\sqrt{{\mathbb{e}}\big[\operatorname{tr}({\boldsymbol{1}}_{\lambda ' } p_i(h))\big| { \mathcal{f}}_{{\mathcal{i}}'}\big ] } \\ \times\sqrt{{\mathbb{e}}\big[\operatorname{tr}\big(t^{(l)^*}\theta_{m+1}^2\frac{n_\zeta}{f}p_i(h)\frac{n_\zeta}{f}\theta_{m+1}^2 t^{(l)}\big)\big| { \mathcal{f}}_{{\mathcal{i}}'}\big ] } \ .\end{gathered}\ ] ] let @xmath756 be the singular value decomposition of @xmath757
. then it follows that @xmath758 \\ = \sum_j \kappa_j^2 { \mathbb{e}}\big [ \langle\psi_j,\sqrt{n_\zeta}p_i(h)\sqrt{n_\zeta}\psi_j\rangle \big|{\mathcal{f}}_{{\mathcal{i}}'}\big ] \leq \|\rho\|_\infty\|t^{(l)}\|_\mathrm{hs}^2|i| \ .\end{gathered}\ ] ] the last step is obtained by first integrating out the random variable @xmath35 , where we apply the spectral averaging lemma from @xcite .
the estimates above now allow us to conclude the claim of the lemma .
the specific form of the bound is obtained similarly to lemmas [ apriori ] and [ resampling ] by explicit estimates on the hilbert - schmidt norms .
in this appendix , we study the decay of `` operator kernels '' of a certain type of functions of a self - adjoint operator @xmath134 that satisfies a combes - thomas estimate . more specifically
, we analyze how fast a bound on a term of the form @xmath759 decays with respect to the distance of @xmath180 and @xmath181 . in @xcite the helffer - sjstrand formula ( cf .
@xcite ) and a combes - thomas estimate ( cf .
@xcite ) were used to study the decay of operator kernels in the case of _
_ @xmath760 .
among other things , the results obtained there yield sub - exponential decay of @xmath761 with respect to the distance of @xmath180 and @xmath181 , where @xmath230 is a suitably chosen smooth cutoff function which vanishes in a neighborhood of @xmath762 .
this , however , is not sufficient for the purposes of this paper , as we need an exponentially decaying bound .
we circumvent this problem by choosing a cutoff function that depends on @xmath180 and @xmath181 , but nevertheless satisfies adequate estimates . in the following , we assume that @xmath134 is a self - adjoint operator @xmath133 on @xmath763 ( @xmath764 open ) that satisfies a combes - thomas bound of the following form with respect to a pseudo - metric @xmath138 on @xmath131 : @xmath765 for all @xmath163 and all @xmath766
. estimates of this type hold for a very general class of schrdinger operators ( cf .
in particular , we note that the random schrdinger operator @xmath767 as defined in section [ section : model ] satisfies such an estimate with constants @xmath220 that are independent of the realization @xmath768 of the random potential .
[ rescutoffdecay ] let @xmath234 be a bounded interval and @xmath72 . for any @xmath163 , there exists a smooth function @xmath769 $ ] such that @xmath770 if @xmath771 , @xmath772 if @xmath773 and such that for all @xmath616 with @xmath247 the inequality @xmath774 holds with positive constants @xmath410 and @xmath775 that depend only on @xmath322 , @xmath776 and the constants in the combes - thomas estimate . as mentioned before
, the choice of the cutoff function @xmath777 depends on the distance of @xmath180 and @xmath181 . in the following
, we present a lemma that specifies the conditions the cutoff function ( multiplied by @xmath778 ) needs to satisfy . here and in the following ,
we define @xmath779 for any @xmath780 .
[ lemma : l1gevreyupton ] suppose @xmath781 is an open interval , @xmath782 and @xmath783 such that @xmath784 for some @xmath785 and all @xmath786 .
there exist @xmath787 depending only on @xmath81 and on the constants in the combes - thomas estimate such that @xmath788 for all @xmath163 satisfying @xmath789 , where @xmath790 .
the assumption differs from the definition of _ @xmath664-gevrey functions of class @xmath108 _ in @xcite in two ways : we require the bound to hold only for @xmath791 and we allow an additional factor of @xmath792 . as a consequence ,
the bound on the norm of @xmath793 holds only in case @xmath175 lies in an @xmath794-dependent interval .
however , the proof , which we postpone to the end of this section , proceeds like the analysis in @xcite and needs only a minor modifcation .
the lemma above is of use only if for any @xmath794 , there is a cutoff function @xmath777 for which @xmath795 satisfies with a uniform constant @xmath796 .
the existence of such a family of functions is ensured by the following proposition , which is a direct consequence of @xcite*proposition 1.4.10 .
[ cutoffbound ] let @xmath797 , @xmath72 and @xmath798 be a bounded interval .
then there exists a function @xmath799 with @xmath800 , @xmath772 if @xmath801 and @xmath770 if @xmath802 such that for all @xmath780 and @xmath803 with @xmath804 @xmath805 where the constants @xmath806 are independent of @xmath807 , @xmath794 , @xmath808 and @xmath776
. we can now combine lemma [ lemma : l1gevreyupton ] and proposition [ cutoffbound ] in order to prove theorem [ rescutoffdecay ] .
let @xmath163 be given .
we define three auxiliary constants @xmath809 where @xmath810 $ ] and @xmath671 is chosen as in .
now let @xmath811 as in lemma [ lemma : l1gevreyupton ] with @xmath812 ( @xmath813 as in ) and @xmath814 and let @xmath797 such that @xmath789 .
choose a cutoff function @xmath777 according to proposition [ cutoffbound ] .
the function @xmath777 thus satisfies the bound @xmath815 for all @xmath780 and @xmath804 .
we define @xmath816 and @xmath817 , and we conclude @xmath818 and hence @xmath819 as a consequence , for any @xmath786 , we have @xmath820 lemma [ lemma : l1gevreyupton ] now implies @xmath821 where @xmath187 , @xmath822 and @xmath262 depend only on @xmath322 , @xmath776 and the constants in the combes - thomas estimate .
this implies the claim of the theorem . it remains to prove lemma [ lemma : l1gevreyupton ]
: the analysis in @xcite shows that , up to a factor depending only on @xmath81 and the constants in the combes - thomas estimate , the norm of @xmath823 can be bounded by the sum of three terms : @xmath824 here @xmath825 depends only on @xmath81 and the constants in the combes - thomas estimate and both @xmath11 and @xmath826 can be chosen arbitrarily , as long as they satisfy @xmath827 and @xmath828 .
now suppose @xmath789 with @xmath829 .
then , choosing @xmath830 and @xmath831 , we can estimate @xmath832 and @xmath833 as @xmath834 , this implies the claimed bound .
michael fauser was partially supported by topmath , the graduate program of the elite network of bavaria and the graduate center of tum graduate school . | we investigate spectral and dynamical localization of a quantum system of @xmath0 particles on @xmath1 which are subject to a random potential and interact through a pair potential which may have infinite range .
we establish two conditions which ensure spectral and dynamical localization near the bottom of the spectrum of the @xmath0-particle system : i ) localization is established in the regime of weak interactions supposing one - particle localization , and ii ) localization is also established under a lifshitz - tail type condition on the sparsity of the spectrum . in case of polynomially decaying interactions , we provide an upper bound on the number of particles up to which these conditions apply . |
much of the understanding of atomic nuclei is based on the assumption that nucleons move , independently from each other , in the average potential created by the interaction with all other nucleons . a more fundamental approach to the understanding of nuclei
has to start from the underlying nucleon - nucleon ( n - n ) interaction .
realistic models of the n - n interaction exhibit a strongly repulsive central interaction at small inter - nucleon distances and a strong tensor component .
these features lead to properties of nuclear wave functions that are beyond what is describable in terms of independent particle ( ip ) motion .
in particular , strong short - range correlations ( src ) are expected to occur .
the effects of the short - range correlations are known for systems where an accurate solution of the schrdinger equation for a realistic n - n interaction can be obtained @xcite
. very light nuclei ( today up to a@xmath010 ) and infinite nuclear matter are amongst the systems where this is feasible @xcite .
the corresponding calculations show that in a microscopic description of nuclear systems the short - range and tensor parts of the n - n interaction have a very important , not to say dominating , influence without which not even nuclear binding can be explained . due to these short - range correlations
the momentum distributions of nucleons acquire a tail extending to very high momenta @xmath1 and part of the strength , located in ip descriptions at low excitation energy @xmath2 , is moved to very high excitation energies . in the past
, most experimental investigations were confined to rather low momenta and energies , _
i.e. _ to the region where the strength is dominated ( but not entirely given ) by the ip properties . in this region
, the consequences of short - range correlations are indicated primarily by a _ depopulation _ of states in comparison to the predictions of ip models ( including the long - range correlations which can be described by configuration mixing ) . according to the calculations mentioned above
, a depopulation of the order of 20% is expected .
> from the experimental information available up to now , the depopulation of ip strength at low @xmath3 is established @xcite ( for a caveat see below ) .
much less is known from a _ direct _ measurement of the strength of the spectral function @xmath4 at large @xmath1 and @xmath2 .
a recent ( e , ep ) experiment , performed at high momentum transfer @xmath5 in parallel kinematics by rohe _
et al . _ , provides the first direct measurement @xcite .
this correlated strength has always been discussed in @xmath3-space where ( part of ) it can be separated from the ip strength . in this rapid communication , we take an orthogonal look at the correlated strength and discuss it in coordinate space ( r - space ) .
we address this question from both the theory and experiment side .
the trigger for this study lies in difficulties experienced in the past in interpreting data in terms of ip models .
for example , fits with ip wavefunctions of the nuclear charge density often yielded form factors ( of ip dominated transitions ) with incorrect @xmath5-dependence .
fits with ip wave functions also have difficulties to reproduce the total densities in the nuclear interior .
the origin of these difficulties : total densities have contributions from correlated nucleons that do not appear in observables dominated by quasi - particle properties .
the correlated nucleons presumably have a different radial distribution .
the goal of this paper is to derive quasi - particle ( qp ) and correlated distributions in r - space using greens - function theory .
we compare the results to the correlated density in r - space which we obtain from the difference of the density known from elastic electron scattering and the qp contributions known from ( e , ep ) reactions . as a side - product
, this study also sheds light on a recently uncovered problem with qp occupation numbers derived from ( e , ep ) experiments with low and high @xmath5 , respectively @xcite .
the evaluation of the single - particle spectral functions for @xmath6 , the nucleus we use for our study , has been performed within the framework of the green s function method @xcite using the techniques described in @xcite .
the nucleon self - energy @xmath7 is determined in a discrete basis of bessel functions @xmath8 with appropriate boundary conditions at the surface of a spherical box with radius @xmath9 .
these basis states are identified by the angular momentum quantum numbers @xmath10 and @xmath11 and a radial quantum number ( @xmath12 , @xmath13 ) referring to the momentum . for a box radius @xmath9 of typically 20 fm
, it turns out to be sufficient to include around 60 basis states for each partial wave .
the self - energy contains a hartree - fock contribution @xmath14 calculated in terms of a nuclear matter @xmath15-matrix plus complex correction terms , @xmath16 , which account for the inclusion of two - particle one - hole and two - hole one - particle contributions .
these correction terms are calculated directly for the finite nucleus @xmath6 , describing the intermediate particle states by plane waves orthogonalized with respect to the occupied hole states .
this is a good approximation to describe the effects of src , however , it tends to underestimate the spectral strength due to long - range correlations at missing energies slightly above the two - hole one - particle threshold .
the single - particle green s function is determined from this self - energy @xmath17 by solving the dyson equation in the box basis described above .
from the imaginary part of this green s function one can calculate the spectral function in this basis @xcite or determine it in configuration space by the transformation s_lj(r , r;e ) = _ m , n _
p_mlj(r ) s_lj(p_m , p_n;e)^*_p_nlj(r),[eq : spectr1 ] using the bessel functions @xmath8 described above .
the spectral function can be split into the qp contribution @xmath18 , which only occurs in the @xmath19 and @xmath20 partial waves , and in the continuum contribution , @xmath21 , which originates from the imaginary components in the self - energy .
this leads to the single - particle density ( r ) & = & _ lj s^qp_lj(r , r ) + _ lj__2h1p^ de s^cont_lj(r , r;e ) + & = & _ qp(r ) + _ corr(r),[eq : densit ] where the integration over missing energies @xmath2 starts at the threshold of two - hole one - particle configurations .
we have assigned the label @xmath22 to the part of the single - particle density , which originates from the continuum part of the spectral function to indicate that this correlated density is absent in the ip model .
results for these contributions to the point density of protons in @xmath6 are displayed in fig .
[ rhomuth ] .
the calculation of these densities have been performed using the cd - bonn potential for the n - n interaction @xcite .
a fraction of the proton density , which accounts for around 5 protons , is described by the qp part and the rest is covered by the correlated density . in order to allow for a better comparison of the radial shape of the density contributions , fig .
[ rhomuth ] also contains the correlated density @xmath23 multiplied by a factor of 3 .
the comparison shows quite clearly that the correlated density is located much more in the center of the nucleus than the qp contribution .
the correlated single - particle density is distributed over partial waves @xmath24 including those which are unoccupied in the ip model . a large fraction of the correlated strength , however , is contained in partial waves with @xmath25 ( around 31% ) and @xmath26 ( around 37% ) .
one also should note that , contrary to what one naively could expect , the strength in the higher @xmath10 states does not contribute at large @xmath27 ; the corresponding large values of @xmath2 pull the radial wave functions to lower @xmath27 .
, width=321,height=283 ]
for @xmath6 an extensive set of elastic electron scattering data is available @xcite covering the range of momentum transfer between 0.13 and 3.7 @xmath29 .
the carbon @xmath30-radius is precisely known from a @xmath31-x - ray experiment @xcite .
these data have been used to determine a model - independent charge density using the sog - approach of @xcite . the procedure employed has been described in @xcite .
this yields the charge density as a function of @xmath27 , together with an error bar that covers both the statistical and systematic uncertainties of the data , as well as the model - error . in order to obtain the point density ,
we have unfolded the effect of the finite size of the proton and neutron charge density .
this has been done by parameterizing the point density , folding it with @xmath32 and fitting the resulting density to the charge density as determined above .
the contribution of the electro - magnetic spin - orbit term to the charge density turned out to be negligible .
the resulting folded density agrees within @xmath331% with the one given in @xcite .
the resulting point density is shown in figs .
[ rosum1],[rosum2 ] .
the error bars in general are too small to be seen .
for carbon , quite an extensive set of ( e , ep ) data is available @xcite ; a compilation is discussed in @xcite .
part of this data has been taken at low @xmath5 with the goal to determine the 1p and 1s quasi - particle momentum distributions and occupation numbers .
some data have been taken at large @xmath5 mainly in connection with the determination of nuclear transparencies for high - energy protons .
the low-@xmath5 data , taken with good energy- and momentum resolution mainly at nikhef and saclay , have been analyzed in the framework of dwba using optical potentials known from proton - carbon scattering .
the qp radial wave functions have been parameterized using woods - saxon ( ws ) potentials .
_ @xcite have made a coherent analysis of the entire data set .
the occupation of the qp orbits , obtained by summing the experimental spectroscopic factors , turns out to be rather low in comparison to what is known for other nuclei @xcite ; the summed 1p plus 1s strength amounts to 3.4 protons only ( 56% occupation ) .
the high-@xmath5 data , determined in part with moderate energy- and momentum resolution , were taken mainly at slac and jlab .
the data were summed over a large region of initial momentum @xmath1 and removal energy @xmath2 , and fitted using ws radial wave functions and theoretical transparencies . when using the most reliable transparencies , confirmed by other experiments , lapikas _ et al .
_ found a much higher occupancy of the qp orbits , 5.0 - 5.6 protons ( @xmath3387% occupation ) .
this discrepancy which is very embarrassing to the practitioners of ( e , ep ) as it sheds serious doubts on the quantitative interpretation of ( e , ep ) data obviously needs to be better understood , and is discussed in more detail below .
this difference also has led to speculations about @xmath5-dependent qp - occupations , for which we see no physical basis .
a partial reason for the difference between the low-@xmath5 and high-@xmath5 results is immediately clear : the low-@xmath5 data cover the region of missing momenta of typically @xmath35180mev / c and missing energy @xmath2 @xmath3550mev , the high-@xmath5 data extend to 300mev / c .
the high-@xmath5 data also cover a larger range in missing energy , they are integrated up to typically 80mev . in this larger @xmath3-range ,
there is not only qp strength , but also a fraction of the correlated strength is integrated over . before making a valid comparison
, this correlated strength needs to be removed . in order to correct for this effect , we start from the high-@xmath5 ( e , ep ) data taken in a recent jlab experiment @xcite in quasi - elastic kinematics , which minimize final state interactions fsi and meson exchange current contributions mec . this experiment yields , in agreement with the previous jlab and slac experiments , 5.2 protons in the integration region e@xmath3580mev , k@xmath35300mev / c . this number we correct for the continuum contribution using the calculated spectral function discussed above . with this correction
the discrepancy between the low-@xmath5 and high-@xmath5 results is significantly reduced ; the qp occupations now are 3.4 _ vs _ 4.5 protons for the low-@xmath5/high-@xmath5 data , respectively . given the uncertainties of these numbers believed to perhaps 10% there still is a worrisome incompatibility . in order to proceed
, we have to choose .
we have decided to use the occupation coming from the high-@xmath5 measurements , as we judge the interpretation of these data to be more safe .
the low-@xmath5 data suffer from uncertainties in the treatment of the final state interaction .
due to the low energies of the outgoing proton ( 70mev for the nikhef data @xcite ) coupled channel effects not treated in the usual dwba analysis should be relevant .
van der steenhoven _
et al _ @xcite have shown that inclusion of these effects would increase , for the rather soft nucleus @xmath6 , the qp occupation by @xmath3320% .
for the kinematics of the low-@xmath5 experiments , the calculations of boffi _ et al .
_ @xcite also predict significant mec effects that would lead to a further increase of the qp occupation .
the value for the qp occupation deduced from the high-@xmath5 measurement is also compatible with the correlated strength measured directly in the recent experiment by rohe _
this measurement agrees with theoretical predictions for the correlated strength of @xmath3320% . the summed qp strength from the low-@xmath5 ( e , ep ) data ( which thus includes the fragmentation due to long - range correlations ) ,
on the other hand , would correspond to @xmath3640% correlated strength , _
i.e. _ be unrealistically high .
furthermore , as we will see below , our choice of the qp occupation is confirmed by a consistency check in our analysis . before proceeding to the calculation of the qp density , one more effect of correlations
must be removed from the results of lapikas _
et al_. while the p - strength is located in discrete states where no ambiguity occurs , the s - strength is located in the continuum between 20 and @xmath3350 mev removal energy . in this region ,
also the correlated strength contributes , and affects the _ shape _ of the fitted ws momentum distribution .
the correlated strength has a momentum distribution that falls much more slowly with increasing @xmath1 than the 1s qp strength ; when fitting the sum with a ws parameterization , the resulting ws momentum distribution extends somewhat too far in momentum , _ i.e. _ it would have too small a radial extension in @xmath27-space .
we have used the theoretical 1s qp and correlated strength in the region used in @xcite for the determination of the 1s momentum distribution to calculate a correction to the ws parameterization fitted to the sum .
we find that the extent of the wave function in k - space needs to be reduced by 11% .
this modified ws - shape has been confirmed @xcite by an independent analysis employing the recent jlab data of @xcite and the correlated spectral function of @xcite . with these qp radial wave functions , and occupations renormalized to the one
derived above from the high-@xmath5 experiments , we can compute the qp density in @xmath27-space . as the radial wave functions fitted by lapikas _
et al . _ refer to relative coordinate between the proton and the cm of the ( a1 ) system , we need to rescale the radial coordinate by a factor 11/12 , with the corresponding adjustment in height to conserve the normalization . ,
width=321,height=283 ] qp occupation .
[ rosum2],width=321,height=359 ] the resulting qp density is shown in figs .
[ rosum1],[rosum2 ] .
the difference of the @xmath6 point density and this qp density yields the correlated density also shown in figs . [ rosum1],[rosum2 ] .
the error bar on the correlated density is not straightforward to calculate due to the various adjustments that had to made in the analysis ; @xmath3720% for @xmath38 is probably a realistic estimate .
the first observation one can make from figs . [ rosum1],[rosum2 ] concerns the fact that like in the theoretical analysis of fig .
[ rhomuth ] also the correlated density deduced from experimental data is significantly more concentrated towards the nuclear interior than the qp density .
one also observes that the correlated density in the nuclear interior gives a very significant contribution , of order 30% , of the central density , _
i.e. _ larger than one could have expected from the number of 20% or so of correlated nucleons .
this explains why attempts to explain the total densities in terms of qp orbitals can not be very successful .
figure [ rosum2 ] shows another important feature : the experimental qp density at large @xmath27 agrees perfectly with the point density measured via elastic electron scattering .
such an agreement should occur , as at large @xmath27 outside the range of the nuclear potential the density is entirely given by the tail of the least - bound qp orbit , the 1p@xmath39 state in @xmath6 .
more deeply bound states , or correlated nucleons with large removal energy , can not contribute .
the good agreement between qp density and point density at large @xmath27 also confirms the correctness of our choice of qp occupations .
had we used the occupation derived from the low-@xmath5 experiments and , as above , the shape of r(r ) determined in @xcite from the fit to the world ( e , ep ) data , we would have obtained the tail indicated in fig .
[ rosum2 ] by the crosses .
these are obviously significantly too low .
this comparison thus provides an _ a posteriori _ justification of our procedure . in figs .
[ rosum1],[rosum2 ] we also show the correlated density obtained by theory .
considering the above - mentioned uncertainty of the experimental result , we consider the agreement between theory and experiment as a good one .
the size of the correlated contribution in the nuclear interior is very similar , the rapid fall - off of the correlated density at large @xmath27 also agrees within our estimated uncertainty .
the correlated density deduced from experimental data seems to contain more spectral strength in partial waves with @xmath40 than the theoretical one .
the large contribution of the correlated density in the nuclear interior shows that the neglect of this correlated contribution in the standard ip calculations ( _ e.g. _ all the shell - model descriptions ) is not justified .
starting from ( e , ep ) data we have constructed the qp density for @xmath6 in coordinate space .
the difference to the total density , obtained from elastic electron scattering , provides the density distribution of the correlated nucleons .
we find that it is significantly more concentrated towards the nuclear interior .
we also find good agreement with the theoretical calculation of the correlated density distribution .
the large contribution of the density related to short - range n - n correlations , @xmath3330% in the nuclear interior , together with the fact that the _ shape _ of the correlated density differs strongly from the qp density , explains the poor performance of qp wave functions in explaining many observables . due to the shape difference
, the shortcoming of the neglect of the correlated contribution also can not satisfactorily be compensated by using effective quantities like effective charges , _
etc_. as a side product , our analysis provides a solution to the puzzle raised in @xcite , the pronounced disagreement between qp occupations derived from the low-@xmath5 and high-@xmath5 ( e , ep ) experiments .
we find that only the high-@xmath5 occupation ( suitably corrected for the correlated contribution not considered in @xcite ) is compatible with the independent information from elastic electron scattering .
we would like to thank louk lapikas for providing the qp wave functions , kai hencken and dirk trautmann for elucidation of cm problems and daniela rohe for an independent determination of the 1s qp momentum distribution .
this work has been supported by the dfg ( graduiertenkolleg basel - tbingen ) and the schweizerische nationalfonds . | several recent studies have dealt with the effects of short range correlations on the momentum distribution of nucleons in nuclei . here
we investigate the correlation effects on the density and spectral distribution in coordinate space .
a combination of experimental data and spectral functions calculated from realistic n - n interactions allows us to resolve a recently uncovered discrepancy with occupation of quasi - particle states derived from ( e , ep ) data . |
the relativistic heavy ion collider ( rhic ) was constructed to explore the physics of the quark - gluon plasma , a novel state of nuclear matter in which quarks and gluons are not locally confined to color singlet bound states @xcite .
the experiments performed at the rhic facility since the year 2000 have confirmed @xcite that thermalized matter with an initial energy density exceeding 5 gev/@xmath0 ( more than 30 times normal nuclear matter density ) is created in collisions of two heavy nuclei at a center - of - mass energy per nucleon pair @xmath1 gev . the final state is well described in its chemical composition as a hadron gas with temperature @xmath2 mev @xcite , expanding with a collective velocity exceeding half the speed of light .
two primary probes of the properties of this matter before its dissociation into individual hadrons have emerged from the rhic experiments : 1 .
the axial anisotropy of the collective flow pattern ( called `` elliptic flow '' ) of the final - state hadrons with respect to the beam axis in off - central collisions ; 2 . the suppression of the emission of hadrons with a high transverse momentum @xmath3 , compared with the yield expected from a superposition of independent nucleon - nucleon interactions .
theoretical studies , discussed below , have related each of these phenomena to a transport coefficient of the produced matter : the elliptic flow has been shown to be highly sensitive to the value of the shear viscosity @xmath4 ; the energy loss responsible for the high-@xmath3 hadron suppression has been shown to be governed a parameter @xmath5 , which specifies the average square of the transverse momentum kicks per unit path length experienced by a parton traversing the matter . remarkably , it turns out that these two parameters are related to each other , if the matter is a quasi - particle medium composed of quasi - particle excitations with the quantum numbers of quarks and gluons , i. e. a _ quark - gluon plasma _ in the literal sense of the word @xcite . on the other hand , if the medium is so strongly coupled that no quasi - particle description applies , or if the quasi - particles have different quantum numbers than quarks and gluons , the relation fails to hold .
a reliable measurement of both parameters , @xmath4 and @xmath5 , would enable us to differentiate between such different scenarios .
when two heavy nuclei collide off - centrally , i. e. with a nonzero impact parameter @xmath6 , the overlap region is shaped roughly like an american football . in a plane transverse to the beam axis , the region is almond - shaped and can be characterized by a spatial eccentrity @xmath7 .
when such a region is filled with equilibrated matter endowed with a large internal pressure , its geometrical shape ensures that the pressure gradient in the direction of the collision plane is larger than the gradient perpendicular to that plane .
this pressure anisotropy leads to a more violent expansion along the collision plane and thus results in an axially anisotropic transverse ( with respect to the beam axis ) flow pattern , which is commonly characterized by the second fourier moment @xmath8 of the angular distribution of hadrons for a given @xmath3 .
the elliptic flow parameter @xmath8 can be measured as a function of @xmath3 and also for individual hadron species . .
the large pressure gradient in the horizontal direction ( the collision plane ) causes a faster expansion than in the vertical direction .
the momentum anisotropy is characterised by the elliptic flow parameter @xmath8 .
[ fig1 ] ] 0.05 it is not surprising that the elliptic flow produced by a given geometric eccentricity is related by the ability of the matter to flow freely locally . in fluid dynamics
, this ability is parametrized by the _ shear viscosity _
the shear viscosity describes the response of the medium to flow gradients .
a low value of @xmath4 implies that the matter is insensitive to flow gradients .
relevant scales determining the importance of shear viscosity are : the _ mean free path _
@xmath9 , the _ homogeneity length _ @xmath10 of the flow pattern , and possibly the thermal wave length @xmath11
. the ratio between the mean free path and the homogeneity scale is called the _ knudsen number _ : @xmath12 . when @xmath13 , the fluid can be described by viscous hydrodynamics
; when @xmath14 such a macroscopic description fails , and the fluid has to be described microscopically , e. g. by the boltzmann equation .
is caused by the exchange of momentum ( vertical [ blue ] arrow ) between regions with different flow velocity ( horizontal arrows ) .
a small shear viscosity implies a short mean free path for momentum transport in the fluid .
[ fig2 ] ] 0.05 when there is a flow shear , i. e. when the flow velocity changes in a direction transverse to the direction of the flow ( see fig . [ fig2 ] ) , the shear viscosity is proportional to the distance over which longitudinal momentum can be readily transported transverse to the flow . kinetic theory for a relativistic fluid yields the relation ( @xmath15 is the particle density , @xmath16 the entropy density , @xmath17 the average momentum of a fluid particle ) @xmath18 in other words , a short mean free path or , equivalently , a large transport cross section @xmath19 imply a small shear viscosity .
somewhat counterintuitively , a low shear viscosity thus requires a strongly coupled medium !
on the other hand , a strongly coupled medium can , and often does , have a large shear viscosity , because its degrees of freedom rearrange themselves under the strong interaction , and momentum can be transferred by mechanisms other than particle transport , e. g. in solids by phonons , or in polymers by momentum transport along molecular chains . as a result ,
one finds in most fluids that the dimensionless ratio of shear viscosity to entropy density ( a generalization of the concept of _ kinematic viscosity _ ) first decreases rapidly with temperature , corresponding to an increase of the ratio between potential and kinetic energy , and then increases again below a critical temperature , often coinciding with a phase transition @xcite , due to the emergence of new momentum transport mechanisms .
an exception are scale invariant theories , such as the celebrated supersymmetric su(@xmath20 ) gauge theory at large @xmath20 , for which the ratio @xmath21 continually decreases with increasing interaction strength and asymptotically reaches the value @xmath22 @xcite , which is widely believed to be the lower bound allowed by quantum mechanics and unitarity @xcite .
the question is sometimes raised whether hydrodynamic flow implies local equilibration of the medium .
it is certainly true that a hydrodynamical description can be applied whenever a medium is locally thermalized and the knudsen number is much less than unity .
however , the reverse is not true ! for example , collective motion in a plasma can be enforced by the action of magnetic fields , which effectively limit the ability of particle to move freely and transport momentum over large distances .
this effect is encoded in the theory of _ magneto - hydrodynamics_. another role of magnetic fields in a plasma is their ability to reduce the shear viscosity by a large factor .
plasmas containing spontaneously generated , random magnetic fields are called `` turbulent '' plasmas .
similar mechanisms have been theoretically shown to occur in locally anisotropic quark - gluon plasmas @xcite , and these may give rise to an anomalously low shear viscosity @xcite .
strong gluon fields are also thought to be present in the pre - thermal `` glasma '' phase and may be responsible for a small shear viscosity at early times .
as mentioned before , the shear viscosity of the quark - gluon plasma produced in nuclear collisions at rhic can be constrained by a comparison of the measured elliptic flow parameter @xmath23 with the results obtained by simulations of the transverse expansion in viscous hydrodynamics .
the result of such comparisons @xcite is that the shear viscosity must be quite small , most likely in the range @xmath24 .
the extracted value depends somewhat on the initial conditions for the geometric anisotropy of the fireball used in the hydrodynamics calculations , with the glauber model giving a value near the lower bound and parton saturation models giving a value near the upper bound .
a slightly different way of extracting the value of @xmath21 is by comparing the hydrodynamics simulations with the dependence of the momentum averaged @xmath8 with the data for the average @xmath8 as a function of the charged particle multiplicity @xmath25 @xcite .
this comparison yields the range @xmath26 .
the dominant source of energy loss of an energetic parton traversing a quark - gluon plasma is believed to be gluon bremsstrahlung .
the parton radiates away part of its energy after scattering off mass shell on one of the medium constituents ( also partons if the quark - gluon plasma can be described as a quasi - particulate medium ) . in qcd
, the energy loss can be expressed in terms of a parameter @xmath5 , which describes the rate at which the medium transfers transverse momentum to the fast parton .
if the differential cross section for elastic scattering in the medium is given by @xmath27 , the energy loss parameter is given by @xcite @xmath28 and the energy loss per unit path length is @xmath29 , where @xmath10 denotes the total length of the path in the medium .
the reason for the additional length dependence on the right - hand side of this equation is that also the radiated gluon re - scatters in the medium ( see fig .
[ fig3 ] ) .
the quadratic path length dependence of the energy loss is characteristic of qcd . controlling
the transverse momentum exchange determines the rate of radiative energy loss of the parton .
[ fig3 ] ] 0.05 the fact that the radiative energy loss depends on the rate of momentum exchange between the fast parton and the medium allows us to relate the coefficient @xmath5 to the shear viscosity , if the average thermal parton can be described as a quasi - particle with the same quantum numbers as a hard parton , i.e. , a quark or gluon .
we can then apply kinetic theory as discussed in the previous section and relate @xmath4 to the transport cross section @xmath30 of a thermal parton in the medium . in qcd , like in qed , the interaction is , apart from debye screening , long - rangedand thus the transport cross section is dominated by small angle scattering .
hence we find : @xmath31 where @xmath32 denotes the average center - of - mass energy in a collision between two thermal partons .
if we also make use of the relation @xmath33 for the entropy per particle in a gas of massless particles , we obtain the remarkable relation @xcite : @xmath34 this relation holds under two general conditions , which one may call the `` weak '' coupling scenario ( see fig .
[ fig4 ] ) : 1 .
the medium is described by nearly massless quasi - particles with the same properties ( quantum numbers ) as high energy partons .
the thermal interactions among medium partons are dominated by small - angle scattering .
these conditions are fulfilled in a perturbative quark - gluon plasma , which can be described by the so - called hard - thermal loop effective theory ; they also hold when the shear viscosity is dominated by anomalous processes in a turbulent quark - gluon plasma .
they do not hold , and the relation is violated , at strong coupling , e.g. in the strongly coupled @xmath35 supersymmetric yang - mills theory , where @xmath36 ( upper [ red ] line ) and @xmath37 ( lower [ green ] line ) as function of the inverse coupling strength .
the two quantities approximately agree in the weak coupling domain , when a quasi - particle picture applies , but not in the strong coupling domain , where @xmath38 .
[ fig4 ] ] 0.05 0.5 cm
what do the rhic experiments tell us about the validity of this relation for the quark - gluon plasma formed in nuclear collisions at rhic ? for this comparison we have to extract the value of the energy loss parameter @xmath5 from the jet quenching data , i.e. from the measured suppression of the yield of energetic hadrons .
a comprehensive analysis of the phenix data on @xmath39 suppression in au+au collisions has recently been performed by bass _
@xcite in three different schemes of radiative jet quenching ( asw @xcite , higher twist ( ht ) @xcite , and amy @xcite ) .
all calculations used the same evolution model for the medium , _ viz . _ three - dimensional ideal hydrodynamics .
the table shows the results for all three schemes , always normalized to @xmath40 mev and for a gluon - induced jet .
[ tab : qhat ] + .values of the energy loss parameter @xmath5 in gev@xmath41/fm deduced @xcite from the @xmath39 yield in @xmath42 gev au+au collisions at rhic , normalized to gluons and @xmath40 mev .
the scaling of @xmath5 with medium density is either assumed to be proportional to @xmath43 or to @xmath44 , where @xmath7 denotes the energy density of the medium . [ cols="^,^,^,^ " , ] as the table shows , there is no consensus yet between the different schemes . for our purposes here , let us adopt the amy scheme , because a more complete analysis has been performed in this scheme , which also includes collisional energy loss in the same framework @xcite .
the additional energy loss mechanism leads to a somewhat lower value of @xmath45 gev@xmath41/fm , corresponding to @xmath46 which nicely falls into the range @xmath47 preferred by the rhic experiments . even more impressive
is the coincidence with the value @xmath48 obtained by meyer @xcite for @xmath49 mev in quenched lattice qcd .
this suggests that the matter produced at rhic is a partonic quasi - particle plasma . [ if
, on the other hand , the value of @xmath5 shown in the table for the asw approach is correct , this would imply a value @xmath50 , which lies far below the kss bound for @xmath51 and thus would prove that the matter is a strongly coupled plasma in the true sense of the word . ]
the shear viscosity @xmath4 and the radiative energy loss parameter @xmath5 , both probe the momentum transport properties of the medium .
a small value of the ( kinematic ) shear viscosity @xmath21 corresponds to a large energy loss parameter . in other words ,
an almost perfect qcd fluid is extremely opaque to particles carrying color charge .
a simple inverse relation holds in all thermal gauge theories , including qcd , at weak coupling . at strong coupling ,
the ratio @xmath21 is limited by the kss bound , but @xmath5 can become arbitrarily large .
the existing approaches to jet quenching do not agree in their conclusions about the physical structure of the quark - gluon plasma produced in nuclear collisions at rhic , some hinting at a `` weakly '' coupled plasma with quasi - particle excitations , and some suggesting a strongly coupled medium .
reliable determinations of @xmath21 and @xmath5 from rhic ( and soon lhc ) data thus have great importance .
i thank the organizers for the invitation to the international school of nuclear physics in erice , sicily and support . this work was supported in part by a grant from the u.s .
department of energy ( de - fg02 - 05er41367 ) . | the qcd matter produced in nuclear collisions at the relativistic heavy ion collider ( rhic ) has been found to have a very low shear viscosity , which is close to the lower bound allowed by unitarity .
the matter has also been found to strongly suppress the emission of energetic hadrons .
this phenomenon , called `` jet quenching '' is interpreted to be the result of a large energy loss by the precursor parton on its path through the dense matter , primarily due to gluon radiation .
i discuss how the two phenomena are related .
the rhic data suggest , in some scenarios of jet quenching , that the quark - gluon plasma created in nuclear collisions is characterized by strong coupling , but still admits a quasi - particle description . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.