article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
in these thirty years , it seems that phase field equations describe the novel approach to the stefan problems . for example
, it has been proved that the phase field equations converge to the equations of the stefan problem as a singular limit .
caginalp and a co - researcher [ 1 - 4 ] have studied the reduction of these equations in the singular limit of stefan problems and hele - shaw problem .
the solution of the stefan problem is obtained from the phase - field equations which consist the heat equation and the ginzburg - landau equation by soner [ 10 ] .
hence , by the phase - field equations , we determine the sharp profile of the phase , the temperature distribution and the velocity of the moving interface between a solid and a liquid .
+ in this paper , using the landau - ginzburg theory of phase transition , physical properties show the phase field equations we consider , which is reduced to by the free energy @xmath1.\ ] ] first we use the phase field equations of `` caginalp type '' due to [ 2 ] , @xmath2 directly and numerically . here
@xmath3 is a derivative of a symmetric double - well potential with minima at @xmath4 , @xmath5 a latent heat and @xmath0 @xmath6 liquid ( e.g. water ) , @xmath7 solid ( e.g. ice ) .
+ second we investigate the following equations @xmath8\ ] ] in two space dimensions , where @xmath9 is a mobility of @xmath0 and these equations are called the modified allen - cahn equations .
+ third we consider the following equations due to [ 6 ] , @xmath10\left[1-\varphi^2\right],\\\\ \displaystyle\frac{\partial u}{\partial t}=d\frac{\partial^2 u}{\partial x^2}+\frac{1}{2}\frac{\partial\varphi}{\partial t } \end{array}\ ] ] then applying the moving frame to these equations yields @xmath11\left[1-\varphi^2\right]=0\\\\ \displaystyle v\frac{\partial u}{\partial x}+d\frac{\partial^2 u}{\partial x^2}-\frac{v}{2}\frac{\partial\varphi}{\partial x}=0 \end{array}\ ] ] finally , we also use the following relations due to [ 8 ] , @xmath12\\\\ \displaystyle\frac{\partial c}{\partial t}=\nabla^2 c+\alpha\frac{\partial\varphi}{\partial t}+\left(\nabla^2\varphi-\frac{\partial\varphi}{\partial t}\right)\frac{\alpha\cdot\displaystyle\frac{\partial\varphi}{\partial t}}{da|\nabla\varphi| } \end{array}\ ] ] to simulate the symmetric circular domain of solid . here
@xmath13 is pclet number , @xmath14 a positive parameter , @xmath15 a mean curvature , @xmath16 also a positive parameter , @xmath17 a damkhler number .
+ the so - called one phase stefan problem has no discontinuous - gradient model for the temperature distribution .
however , in the two phase stefan problem , this phenomena occurs .
thus one can obtain the sharp - interface numerically by investigating what control it .
+ our first purpose of this paper is to find out what determine the sharpness of the interface between two different phases ( e.g. ice and water ) .
the second is to investigate asymptotic behavior of the volume of the solidification and dissolution and/or precipitation . + we calculate these equations numerically performed in two - dimensions .
we take full advantage of the phase field parameter @xmath0 to track the interface on which complicated movements or changes are needed .
more precisely , we analyze the dissolution and/or precipitation without the unknown velocity functions like the navier - stokes equations numerically .
in our survey , we use the finite differential method for the time evolution equations numerically . + for eqs .
( 1 ) , we use the 5-point finite difference stencil @xmath18 and we use the forward scheme in time @xmath19 for eq .
( 2 ) , we use the forward scheme in time @xmath19 on the other hand , the second order center - ward scheme in space @xmath20 for eqs .
( 3 ) , the gradient of @xmath0 with respect to @xmath21 , @xmath22 , we use the central difference approximation @xmath23 respectively , and @xmath24 if we do not have any confusion , we abbreviate the `` @xmath25 '' .
this result is well - known
. however we consider it to show an asymptotic behavior of the solutions for the stefan problems . + in general
, the parabolic partial differential equations often have initial - boundary value problems in a certain sense and the exact solutions are difficult to obtain by the direct calculations .
therefore the fact that we investigate the asymptotic behavior of solutions of general case plays an important rle in this sense .
thus we get @xmath26 where @xmath27 corresponds to the integral formula `` gaussian distribution '' ( see friedman , 1964 , 2008 ) .
table i. numerical volumes of the solidification and dissolution and/or precipitation for the time step . from the above , eq .
( 1 ) v.s .
volume 1 and eq .
( 2 ) v.s .
volume 2 . + we obtain the fact that the interface between two regions in the solidification and dissolution and/or precipitation are determined by the `` coupling parameters '' and coefficients , and sharp in our numerical simulation .
+ the volume controlled by the solidification and dissolution and/or precipitation is the dimensionless versions .
we report that the volume of them is proportional to the time step ( nstep in table i ) . | phase field equations describe the novel approach to the stefan problems .
we calculate these equations numerically performed in two - dimensions . we take full advantage of the phase field parameter @xmath0 to track the interface on which complicated movements or changes are needed .
more precisely , we analyze the dissolution and/or precipitation without the unknown velocity functions like the navier - stokes equations numerically .
+ + pacs numbers : v79b 68 . 70 . w+ . |
one of the testing ground of grand unified theories is the strict unification of the three coupling constants @xmath1,@xmath2 and @xmath3 of the standard model @xmath4 at some high scale ( @xmath5 ) , using as inputs their experimental values at the z - pole mass . the coupling constants @xmath1 and @xmath2 can be determined from the accurate experimental measurement of @xmath6 and @xmath7 at @xmath8 pole mass .
the world average values for @xmath8,@xmath6 and @xmath7 are @xcite , [ 1 ] @xmath9 the coupling constants @xmath1 and @xmath2 of the @xmath10 and @xmath11 gauge are related to @xmath6 and @xmath7 as , @xmath12 the strong coupling constant ( @xmath3 ) has also been measured , its world average value is ; @xmath13 @xcite .
recently , it has been pointed out that qcd can not tolerate such a large @xmath3 @xcite .
several low energy experiments also indicate that @xmath3 must be close to 0.11 @xcite , @xmath14 below the z - peak value measured at collider experiments .
theoretically clean deep inelastic scattering experiments give @xmath15 @xcite . a new analysis of the @xmath16 sum rule yield @xmath17 @xcite .
similar values are obtained in lattice qcd : @xmath18 ( @xmath19 spectrum ) @xcite and @xmath20 ( @xmath21 spectrum ) @xcite .
the apparent conflict between the low energy and the collider energy z - peak determination of @xmath3 may be due to the higher order corrections to the lep values .
preliminary analysis suggests that these higher order correction may bring down the collider values to be consistent with the low energy measurements @xcite .
indeed it has also been argued that the systematic error usually quoted in lep number is grossly underestimated , and the lep can only claim to determine the strong coupling constant within the limit , @xmath22 @xcite .
it is now established that the standard model is inconsistent with gauge unification , with the experimentally measured coupling constants @xcite . at early 90 ,
the measurements were consistent with the gauge unification in the minimal supersymmetric standard model ( mssm ) @xcite .
recently langacker and polonsky @xcite using the recent electroweak data found that in mssm , @xmath23 is required for strict gauge unification .
this value is considerably higher than the world average value of @xmath24 .
also it is much above the low energy measurements and qcd motivated value of @xmath25 . in the context of mssm
the lower @xmath3 require raising the susy particle masses considerably higher than the 1 tev scale .
however , susy mass spectra considerably higher than 1 tev scale will have problem of diverging radiative correction . also , heavy susy masses will be inconsistent with our expectation that the lightest supersymmetric particle ( lsp ) will be neutral and the candidate for the dark matter . the other possibility , to reconcile @xmath26 with gauge unification , is to assume very large negative heavy threshold correction at gut scale and nro s which could decrease @xmath27 by 10% @xcite .
recently , roszkowski and shifman @xcite argued that in the simple supersymmetric extension of the standard model , one need not have the constraint of universal gaugino masses at the gut scale , which is , in general assumed in mssm . to fully specify the mssm , one needs to invoke the supersymmetry breaking pattern . the soft breaking term , which in essence couple the mssm with the n=1 supergravity give rise to the constraint of universal gaugino masses at gut scale .
thus in a pure phenomenological approach , one can relax the condition of universal gaugino masses .
they showed that by relaxing the constraint small @xmath3 compatible with low energy measurements can be obtained@xcite . in that case , gluino becomes lighter than the wino and can be well below 200 gev @xcite .
however , the approch is unsatisfactory , as there is no theoretical scheme for the symmetry breaking term . in the present paper
, we show that consitent gauge unification with small @xmath3 is not possible within the framework of mssm .
our approah is phenomenological .
we keep the constraint of universal gaugino masses .
we treat the higgsino masses as a parameter of unification , and tune it to obtain unification for a given value of @xmath3
. gauge unification with @xmath28 0.11 , require higgsino masses in the range of @xmath29 gev , three order of maginitude larger than other sparticle masses .
higgsino masses in the range of @xmath29 gev will require that the supersymmetry break at that scale only and the model will face the unsatisfactory aspect of gauge hierarchy problem .
the paper is organised as follows . in section 2 , we describe the model briefly , in section 3 , the results obtained will be discussed .
summary and conclusions will be given in section 4 .
the renormalisation group equations for the gauge couplings are given by ( neglecting the small yukawa couplings ) , @xmath30 where @xmath31 , with q the running scale and @xmath5 the unification scale mass .
the one loop coefficients @xmath32 of the @xmath33 functions for the gauge couplings change across each new running mass threshold . in the mssm
they can be parameterized as @xcite , @xmath34\\ b_2 = & & -\frac{19}{6}+ \frac{4}{3 } \theta_{\tilde{w } } + \frac{2}{3}\theta_{\tilde{h } } + \frac{1}{6}\theta_{h_2 } + \frac{1}{2}\sum^3_{i=1 } [ \theta_{\tilde{u}_{l_i } } \theta_{\tilde{d}_{l_i } } + \frac{1}{3 } \theta_{\tilde{e}_{l_i } } \theta_{\tilde{\nu}_{l_i}}]\\ b_3 = & & -7 + 2\theta_{\tilde{g}}+ \frac{1}{6}\sum^3_{i=1 } [ \theta_{\tilde{u}_{l_i}}+ \theta_{\tilde{d}_{l_i } } + \theta_{\tilde{u}_{r_i}}+ \theta_{\tilde{d}_{r_i}}]\end{aligned}\ ] ] where @xmath35 . in
the above equations @xmath36 stands for the mass degenerate higgsino fields , @xmath37 for the winos , the partner of the su(2 ) gauge bosons ( @xmath38 ) , @xmath39 is the gluino , the partner of the gluon , all are assumed to be mass eigen states in this approximation .
the @xmath40 stands for the heavy higgs boson doublet @xcite .
the effect of low mass threshold on two loop beta function is expected to be small . in the weak scale ,
we thus use the sm values : @xmath41 and from the weak to the gut scale , we use , @xmath42 the rge equations can be integrated in a step function approximation to obtain @xmath43 at a scale @xmath44 for a given @xmath45 , @xmath46 with @xmath47 the eq.[2 ] can be solved iteratively to obtain the coupling constants at any arbitrary energy , knowing their value at a given energy .
in addition to the above equations , we consider the evolution of gaugino masses
. there evolution equations are simple , @xmath48 with the boundary condition at @xmath5 : @xmath49 .
the solution for the gluino and the winos can be written as , @xmath50 where @xmath51 is the unified coupling constant at the gut scale . combining the two equation we obtain , @xmath52 for a given @xmath53 , @xmath54
can be obtain iteratively from eq.[3 ] the simple step - function approximation used to integrate the rg equations is justified only in @xmath55 scheme .
however , the experimental @xmath6 and @xmath7 were obtained in the @xmath56 scheme .
we therefore convert the coupling constants into the @xmath55 scheme by , @xmath57 where @xmath58 and @xmath59 .
we note that apart from the winos and the gluino , the beta function coefficients are most affected by the higgsinos .
not only they appear in two beta function @xmath60 and @xmath61 , compared to other sparticles , their coefficients are also larger .
we therefore choose the higgsino mass as the parameter which will be tuned to obtain strict gauge unification with recent electroweak data .
the wino and the gluino will be treated separately as they are connected by the universal gaugino mass at gut scale .
all the other susy masses will be assumed to be degenerate at a common mass ( @xmath62 ) .
this is certainly an assumption , however , it helps to identify the role of higgsinos in the unification process .
we define the gut scale @xmath63 as the scale where all the three coupling constants unify at some value @xmath51 .
the rg equations for @xmath1 , @xmath2 and @xmath64 were run simultaneously from @xmath8 to @xmath5 .
the input @xmath1 and @xmath2 were calculated from the experimental @xmath6 and @xmath7 ( eq.[1 ] ) .
we have used the central value for @xmath6 .
unification of three forces is sensitive to the input value of @xmath7 .
this sensitivity is due to the fact that @xmath2 does not change much between @xmath8 and @xmath5 , as do the other two couplings .
thus a small change in @xmath7 has an enhanced effect on the unification .
we therefore consider two input values of @xmath7 : @xmath65 and @xmath66 , covering the @xmath67 variance .
for the input strong coupling @xmath3 , we choose a value between 0.10 - 0.13 . for a given wino mass ( @xmath53 ) and the common susy mass @xmath68
we vary the higgsino mass ( @xmath69 ) 100 gev onwards to find the minimum @xmath69 required for the strict unification of the three couplings .
the gluino mass was obtained iteratively from eq.[3 ] . in fig.1a , we have shown the higgsino mass @xmath69 required to obtain strict gauge unification , as a function of the input @xmath3 .
the common susy particle mass(@xmath62 ) and the wino mass were fixed at 1000 gev .
the black dots and the open triangles corresponds to @xmath70 and @xmath71 respectively .
the higgsino mass required for unification shows a sensitive dependence on the input @xmath3 .
it also depends on the input @xmath7 . for @xmath7=0.231 ,
the higgsino mass varies between @xmath72 gev to @xmath73gev , as @xmath3 changes from 0.10 to 0.13 .
the higgsino mass is lowered approximately by a factor of 10 , if @xmath66 .
it then varies between @xmath74 gev to @xmath75gev . in fig.1b and 1c
, we have shown the unification scale ( @xmath5 ) and the inverse of the unified coupling ( @xmath76 ) .
we note that they are anti - correlated .
while @xmath5 increases , @xmath76 decreases with the input @xmath3 . in comparison to the higgsino mass
, @xmath5 is quite insensitive to the input @xmath3 .
as @xmath3 changes from 0.10 to 0.13 , the gut scale is increased by a factor of 2 only .
similarly , @xmath76 also shows weak dependence on the input @xmath3 .
it is changed by less than 10% .
dependence of the gut scale on the input @xmath7 is also manifest . within @xmath67 variation of @xmath7
, it is changed approximately by 10% .
interestingly , the unified coupling do not show appreciable dependence on the input @xmath7 .
in fig.1d , the ratio @xmath77 is shown . within @xmath67 variation of @xmath7 the gluino mass also do not show appreciable dependence on the input value of the weinberg angle . however , as input @xmath3 varies from 0.10 to 0.13 , it increases from 2300 gev to 2800 gev , showing weak dependence on the input @xmath3 . in fig.1e , we have depicted the variation of the universal gaugino mass @xmath78 .
@xmath78 also do not show appreciable dependence on the input @xmath7 .
it shows weak dependence on the input @xmath3 , increasing from @xmath79 gev to @xmath80 gev , as @xmath3 changes from 0.10 to 0.13 .
the results discussed so far indicate that while the higgsino mass shows a sensitive dependence to the input @xmath3 for strict unification , ( it changes by 6 order of magnitude as @xmath3 changes from 0.10 to 0.13 ) other parameters of the model , e.g. @xmath5 , @xmath51 , @xmath54 and @xmath78 shows a weak dependence on the input @xmath3 . also , unlike the higgsino mass , the input dependence of @xmath7 is also less in those parameters . with the world average value of @xmath3=0.123 ,
strict unification is possible with higgsino mass as well as all the other sparticle masses in the range of tev scale .
however , if we consider the qcd motivated small @xmath810.11 , strict unification is obtained with higgsino masses in the range of @xmath82 gev .
all the other susy masses can be in the tev scale . to be specific , for @xmath83
, strict gauge unification is obtained with @xmath84 ) gev for @xmath7=0.231(0.232 ) . in the above calculations ,
the mass of the winos ( @xmath53 ) were fixed at 1000 gev . to observe the effect of wino mass on the unification , with small @xmath3 , we now vary @xmath53 from 100 gev to 2000 gev .
we fix the strong coupling at @xmath83 .
the common susy mass @xmath62 was fixed at 1000 gev .
as before , while the central value was used for @xmath6 , for @xmath7 , we consider variation within @xmath67 . in fig.2 ,
the results are shown . in fig.2a , the higgsino mass required for strict unification
is shown as a function of the wino mass . as observed earlier ,
the higgsino mass depends on the input @xmath7 . within its @xmath67 variation
, it is changed approximately by a factor of 10 .
however , the higgsino mass required for unification shows a weak dependence on the wino mass . as the wino mass is varied from 100 to 2000 gev , the higgsino mass is decreased by a factor of 5 only .
interestingly , we find that after 1000 gev , the the higgsino mass remains nearly same .
variation of the gut scale and inverse of the unified coupling with the wino mass is shown in fig.2b and 2c .
as before , they are anti - correlated . while @xmath5 shows weak dependence on the input @xmath7 , @xmath76 do not exhibit any such dependence .
there dependence on the wino mass is also weak .
here also , after @xmath53=1000 gev , rate of variation of @xmath5 and @xmath76 slows down .
the variation of the ratio @xmath77 with the wino mass is shown in fig.2d .
the ratio do not depend on the input @xmath7 .
it decreases with the wino mass . again after 1000 gev , the rate of decrease is slowed down and the relation @xmath85 become valid . in fig.2e , the universal gaugino mass @xmath78 and its variation with the wino mass is shown .
@xmath78 do not show any dependence on the input @xmath7 .
as expected , it increases linearly with the wino mass .
the results indicate that the unification of three forces with @xmath83 depend weakly upon the wino mass . increasing the wino mass beyond 1000 gev , does not affect the unification appreciably .
we have also studied the dependence of the common susy mass @xmath62 on the unification process with @xmath3=0.11 .
the wino mass was fixed at 1000 gev . in fig.3a , the variation of the higgsino mass with @xmath62
the input dependence of @xmath7 on the higgsino mass is again manifest .
the higgsino mass shows a very weak dependence on the common susy mass @xmath62 .
similarly , the gut scale ( fig.3b ) , the unified coupling ( fig.3c ) , the ratio @xmath77 ( fig.3d ) and @xmath78 ( fig.3e ) shows very weak dependence on the input @xmath62 .
thus the unification of three forces with @xmath83 is insensitive to the common susy mass @xmath62 within the range 500 - 2500 gev .
thus a more accurate susy mass spectra will not alter the present results significantly .
the present analysis indicate that low energy qcd motivated strong coupling constant @xmath86 is consistent with gauge unification in mssm , if the higgsino masses are in the range of @xmath82 gev and other susy masses are in the tev range . in the following
we will examine its effect on some other features of the model .
any gut theory must consider the nucleon decay rate in the model .
as the gut scale obtained presently exceeds @xmath87 gev , we do not expect problem from the direct proton decay @xmath88 .
however , nucleons can decay via exchange of color - triplet higgs multiplet @xcite , through the dimension five operator .
the color triplet higgs mass @xmath89 , then has significant observable effect .
the lower bound on @xmath89 from nucleon - decay experiments can be obtained as @xcite , @xmath90 we have calculated the color triplet higgs mass in our model . following @xcite , we write at the gut scale , @xmath91 the first term of the rhs corresponds to heavy threshold correction at gut scale as a function of the color triplet higgs mass @xmath89 @xcite .
the second term is the light threshold correction at the weak scale ( @xmath92 ) .
( we note that in the calculations presented at earlier sections , we have not taken into account the heavy threshold effects .
however , if the heavy thresholds are much heavier than the gut scale , then the results obtained previously will not be altered ) . for strict unification at the gut scale ,
the lhs of the above equation is zero and we can find the color triplet higgs mass in terms of the susy masses . in the above equation all the susy particles
are assumed to be degenerate at a common mass @xmath93 .
however , as we have treated the winos , gluino and the higgsinos separately , the the second term is to be replaced by @xmath94 in fig.4 , we have shown the variation of the color triplet higgs mass @xmath89 with the wino mass @xmath53 as obtained from strict gauge unification .
the strong coupling was fixed at 0.11 .
the common susy mass @xmath62 was fixed at 1000 gev .
as discussed earlier , for a given wino mass the unification was obtained by varying the higgsino mass only .
color triplet higgs mass depend on the input value of @xmath7 , lower @xmath7 results in a higher value for the color triplet higgs mass .
within @xmath67 , the value is changed approximately by a factor of 10 .
@xmath89 also depend on the wino mass .
as the wino mass is increased its value is decreased .
however , 1000 gev onwards , the rate of decrease is slowed down .
we observe that the color triplet higgs mass @xmath89 as obtained presently is much above the lower limit ( @xmath95 gev ) required to respect the experimental nucleon decay rate .
the result suggest that large higgsino mass will not cause rapid proton decay through the dimension five operator . about the lsp being the candidate dark matter
, we observe that the higgsinos mixes with gauginos ( both are spin 1/2 particle ) to give the neutralino s . in the limiting case when @xmath96,@xmath97 and @xmath69 >
> @xmath8 , the bino and the neutral wino do not mix with each other nor they mix with the higgsinos .
thus binos can still be the lsp and the candidate for the dark matter .
thus if only the higgsinos are in the mass range @xmath82 gev , and all the other s - particles are in the tev range , the model will not suffer from radiative correction at the electroweak scale .
also the expectation that bino s are the dark matter candidate need not be altered .
one of the motivation of introducing the supersymmetry is the gauge hierarchy problem , i.e. existence of very small and large mass scale in nature . in su(5 ) , as well known , the tree level parameters needed to be fine tuned to an accuracy of @xmath98 or so , in order to obtain the mass ratio @xmath99 .
however , tree level fine tunings are upset at higher order due to quadratic radiative corrections and su(5 ) is in problem .
this need not happen in supersymmetric theories due to the nonrenormalisation theorem of grisaru , rocek and siegel @xcite . according to this theory , which is valid for exact or softly broken supersymmetry
, the parameters of the super potential do not only receive infinite renormalisation but they also do not receive finite renormalisation in higher orders .
thus the parameters , once fine tuned at tree level to obtain hierarchy , the radiative correction donot disturb the hierarchy at higher order . thus problem of gauge hierarchy in supersymmetric model is not as severe as in su(5 ) .
also , in supersymmetric theories , logarithmic dependence of radiative corrections makes the emergence of high mass scale from a low mass scale quite natural .
however , with the higgsino mass in the scale of @xmath29 gev , effective breaking of supersymmetry will occur at that scale only .
the effective theory below that scale will not be protected by the non - renormalisation theory .
quadratic divergences will be generated in higgs mass due to absence of higgs - higgsino - gaugino coupling at low energy .
with higgsino mass in the range of @xmath29 gev , the corrections to the higgs mass will be of that scale and correct electroweak scale can not be achieved without fine tuning .
thus we find that if the strong coupling constant is small as obtained in qcd experiments , gauge unification with recent electroweak data can be achieved , alongwith universal gaugino masses , only if we abandon one of the primary motivation of supersymmetry , namely the resolution to the gauge hierarchy problem .
also generating three order of hierarchy among the sparticles will be difficult to obtain in any theoretical framework .
the discussion suggests that consistent gauge unification with @xmath100 is not possible within the frame work of mssm .
we have studied the gauge unification in the mssm with respect to the higgsino masses .
the purpose of our study was to find whether with @xmath100 gauge unification is possible in the mssm with universal gaugino mass at gut scale .
our approach is phenomenological .
we choose the higgsino mass as a parameter . for a given wino mass ,
it was tuned to obtain gauge unification .
the gluino mass was obtained iteratively from the condition that at gut scale the winos and the gluino have a common mass @xmath78 .
all the other sparticle masses were assumed to be degenerate at a common mass @xmath68 .
the rg equations were run from @xmath8 to the gut scale , with the beta function coefficients changing at appropriate masses .
we find that if @xmath3 ranges between 0.10 to 0.13 , with the recent electroweak data , it is possible to unify all the three forces by varying the higgsino mass between @xmath101 gev .
all the other sparticles can be in the tev scale .
the result is insensitive to the wino mass or the common mass @xmath62 .
thus a more accurate susy spectra will not alter the results significantly .
it was found that if @xmath102 , as measured in qcd experiments , gauge unification require higgsino masses in the range of @xmath29 gev , three order of magnitude higher than other sparticle masses .
colour triplet higgs were also calculated and found to be heavy enough to forbid rapid proton decay .
however , with higgsino masses in the scale of @xmath29 gev , supersymmetry breaks at that scale only and problem of gauge hierarchy resurface .
one of the fundamental motivation of introducing supersymmetry is then lost . to conclude , if @xmath3 is indeed small , as measured in qcd experiments , then consistent gauge unification within mssm is not possible . however , mssm is merely the simplest extension of standard model with supersymmetry .
it is possible that with a more complicated model a consistent picture will emerge . [email protected] review of particle physics , particle data group , , 1 ( 1996 ) .
m. shifman , mod .
phys . lett.a
* 10 * , 605(1995 ) .
g. altarelli , in qcd- 20 years later , proceedings of the workshop , aachen , germany,1992 , edited by p.zerwas and h. kastrup ( world scientific .
singapore , 1993 ) , vol.1,p.172 ; s. bethke , in qcd 94 , proceedings of the international conference , montpellier , france , 1994 , edited by s. narison [ nucl .
b(proc . suppl.)*39b * ( 1995 ) .
m. virchaux and a. milsztajn , b * 274 * , 221 ( 1992 ) .
a. martin , w. stirling and r. roberts , phys .
b*354 * , 155 ( 1995 ) .
m. voloshin , int .
phys.a*10 * , 2865 ( 1995 ) .
el - khadra , g. hockney , a. kronfeld and p. mackenzie , , * 69 * , 729 ( 1992 ) .
c. davies , k. hornbostel , g.p.lepage , a. lidsey , j. shigemitsu and j. sloan , b*345 * , 42(1995 ) .
j. ellis , d.v .
nanopoulos and d. ross , b * 267 * , 132 ( 1991 ) .
m. consoli and f. ferroni , b*349 * , 375 ( 1995 ) .
u. amaldi , w. de boer and h. furstenau , b * 260 * , 447 ( 1991 ) .
j. ellis , s. kelly and d.v .
nanopoulos , b * 260 * , 131 ( 1991 ) .
p. langacker and m.x .
luo , 817 ( 1991 ) .
p. langacker and n. polonsky , , 3081 ( 1995 ) .
d. ring , s. urano and r. arnowitt , , 6623 ( 1995 ) .
t. dasgupta , p. mamales and p. nath , , 5366 ( 1995 ) .
l. clavelli and p. w. coulter , , 3913 ( 1995 ) .
l.roszkowski and m. shifman , , 404 ( 1996 ) .
roberts and g.g .
ross , nucl .
phys.b*377 * , 571 ( 1992 ) .
j.ellis , s. kelly and d.v .
nanopoulos , nucl . phys .
* b373 * , 55 ( 1992 ) ; b*287 * , 95 ( 1992 ) ; j. hisano , h. murayama and t. yanagida , , 1014 ( 1992 ) . g. kane , c. kolda , l. roszowski and j. wells , , 6173 ( 1994 ) . n. sakai and t. yanagida , nucl .
phys.*b197 * , 533 ( 1982 ) .
s. weinberg , , 287 ( 1982 ) .
j. hisano , h. murayama and t. yanagida , nucl . phys .
* b402 * , 46 ( 1993 ) . c. d. carone , h. murayama , , 1658 ( 1996 ) . m. t. grisara , m. rocek and w. siegel , nucl
b159 * , 429 ( 1979 ) . | we have studied the gauge unification with the recent electroweak data as a function of the higgsino mass .
it was shown that if the strong coupling constant is small @xmath0 , consistent picture of gauge unification is not possible in the minimal supersymmetric standard model . |
genius(germanium in liquid nitrogen underground setup ) @xcite is a proposal for a large scale detector of rare events like double beta decay , cold dark matter and low - energy solar neutrinos in real time .
the idea of genius is to operate a large amount of `` naked '' ge detectors in liquid nitrogen , which has been shown to work already in @xcite . + the proposed scale of the genius experiment is a nitrogen tank of about 12 m diameter and 12 m height with 100 kg of natural ge or 100 kg of enriched @xmath3ge , in the dark matter and double beta decay versions , respectively , suspended in its center . with this configuration , which removes most of the material in the vicinity of the detectors ,
the background can be strongly reduced with respect to conventionally operated detectors . moreover
, the liquid nitrogen acts at the same time as cooling medium and as a shield against external radioactivity .
+ it is clear that residual impurities in the liquid nitrogen , like @xmath4u and @xmath5th contaminations , @xmath6rn , @xmath7k and @xmath8co , can be a source of background for the genius experiment and for this reason their contribution has to be deeply investigated .
this has been done in @xcite , where we show , for example , that the contribution of the mentioned isotopes to the background of genius , in the low - energy region , is of the order of @xmath9 counts / kg kev y for @xmath6rn and @xmath10 counts / kg kev y for the remaining isotopes . for these calculations ,
values of liquid nitrogen purity measured by the borexino collaborations were used @xcite . for the borexino @xcite
experiment the nitrogen purity is a very crucial point ; they realized only recently @xcite that they have to address the problem of argon and krypton background at low energies more seriously .
this is true as well for the kamland @xcite experiment , as long as they aim to measure solar neutrinos , after the first successful reactor neutrino phase .
+ this motivated us to study in detail the effect of argon and krypton on the genius background . in this note
we investigate the background contribution coming from the two long - lived isotopes @xmath2kr ( t@xmath11=10.76 y ) and @xmath1ar ( t@xmath11=269 y ) and we show that the potential of genius to look for dark matter at low energies is not affected , while , for solar neutrinos , a reduction of a factor 10 of the krypton contamination with respect to best - quality commercially available nitrogen is needed ( from which a factor 8 has already been achieved , see below ) .
the two isotopes @xmath2kr and @xmath1ar decay both via @xmath12 decay with end point energy of 565.0 kev and 687.1 kev respectively ; in addition @xmath2kr emits four gammas , with energies and intensities listed in table [ tab1 ] . +
both the electrons coming from the @xmath13 decay and the photons from the @xmath14-lines can be a source of background at low energies ( @xmath15500 kev ) . since the mean free path of the electrons
is expected to be very small , only the electrons produced very close to the detectors can reach them . in the case of photons , even if produced far from the detectors , we expect them to be slowed down and to produce a diffuse background at low energies .
+ we calculated the amount of this background by performing a simulation of the genius setup with the geant4 simulation tool . the configuration of the detectors used for the simulation is shown in fig . [
fig : simul ] : we assume 10 ge detectors , each one with a radius of 6.7 cm and 13.4 cm height , arranged in two layers .
the total mass is 100.534 kg and reproduces the mass of natural germanium foreseen for the application of genius to dark matter search .
the detectors are suspended in the middle of a cylindrical tank of 12 m diameter and 12 m height . according to the measurements performed by the heidelberg group of the borexino collaboration @xcite , the purest nitrogen available on the market has a contamination of 0.03 - 0.2 ppm for argon and about 10 ppt for krypton . in the following
we will assume a contamination of 0.2 ppm ( 2.4 @xmath16bq / m@xmath17 ) from argon and about 10 ppt ( 10 @xmath16bq / m@xmath17 ) from krypton .
these values refer to gaseous nitrogen and have to be multiplied by a factor 650 when considering liquid nitrogen . as a first step we simulated the background coming from @xmath1ar .
we started assuming that only a cylindrical region of 1 m diameter and 1 m height around the center of the tank is filled with liquid nitrogen , that means that each detector is surrounded by at least 35 cm of liquid nitrogen in each direction .
the total number of simulated events is @xmath18 . in fig .
[ fig : argon ] we see the output of this first simulation for two of the detectors ( the central one and one of the external ones of the upper layer ) , fig .
[ fig : argon_sum ] ( left ) shows the sum spectrum of all 10 detectors .
+ we can express the background in terms of the usual units ( counts / kg kev y ) : the result is shown in fig . [
fig : argon_sum ] ( right ) . increasing
the amount of liquid nitrogen around the detectors will not change the number of detected counts .
this is because the electrons coming from @xmath1ar - decay have a very short range and so only those produced very close to the detectors can contribute to the background . to verify this
, we performed another simulation , this time assuming that a cylindrical region of 2 m diameter and 2 m height around the center of the tank is filled with liquid nitrogen ( a volume of liquid 8 times bigger than the previous case ) ; the total number of simulated events was @xmath19 . as we see again from fig .
[ fig : argon_sum ] , the resulting background does not change . in the case of @xmath2kr , a contribution to the background will come both from electrons and from the 514 kev photons
; photons can travel in liquid nitrogen much longer than electrons and they can reach the detectors even if produced far from them .
+ we can have an idea of how far they can travel calculating the mean free path of photons , with the appropriate energy , in liquid nitrogen .
the total mass attenuation coefficient in nitrogen for @xmath14-rays with energy of 500 kev is @xcite : @xmath20 multiplying by the density of liquid nitrogen ( 0.808 g / cm@xmath17 ) we obtain : @xmath22 where @xmath23 is the mean free path of a 500 kev - photon in liquid nitrogen . + from a previous simulation performed with a simpler geometry we obtained some results on the propagation of the photons coming from @xmath2kr .
we considered only four detectors in a single layer , for a total mass of 13.4 kg , suspended in the middle of a cubic copper box , filled with liquid nitrogen .
this simulation has been performed in five steps , the details of each step are described in table [ tab : simul ] .
the number of simulated events was increased at each step in order to have approximately the same statistics in all five cases . in fig .
[ fig3 ] we compare the result of the five simulations ( after converting each sum spectrum to the usual units ) : as we can see , increasing the dimensions of the copper box the number of decaying atoms increases and , consequently , we get a higher count rate in the detectors .
this behaviour goes on until we reach a situation were , even adding more liquid nitrogen to the setup ( and therefore more impurities ) we do not see an increase in the detected backround : this is due to the fact that the photons coming from the decay of @xmath2kr atoms are absorbed by the liquid nitrogen itself . according to fig .
[ fig3 ] , 100 cm of liquid nitrogen are enough to absorb the photons coming from the rest of the liquid .
+ we can use this information for a simulation with the more realistic geometry described before ( see fig .
[ fig : simul ] ) .
we simulated 1.5 @xmath24 @xmath2kr decays in a cylindrical volume with a radius of 2 m and 2 m height around the detectors , to be sure that they are surrounded by at least 150 cm of liquid nitrogen .
we obtained the results shown in fig .
[ krypton_norm ] , which represents the total background expected in genius from the decay of @xmath2kr .
+ we started also another simulation in a smaller volume ( 1.5 m radius and 3 m height ) and we show in the same fig .
[ krypton_norm ] that the resulting background ( red histogram ) is the same , apart from statistical fluctuations .
+ finally , fig .
[ total ] shows the total background coming from the contribution of both krypton and argon in genius .
the results on the background indexes are summarized in table [ tab : results ] .
+ the final goal of genius is to obtain a level of background of the order of 10@xmath0 counts / kg kev y in the low - energy region .
this low background is requested for real - time detection of solar pp neutrinos @xcite . already with a background 10 times higher , genius
would obtain the planned results in the field of dark matter searches .
it would be able to test the wimp - parameter space down to cross sections of the order of 10@xmath25pb @xcite . +
as shown in table [ tab : results ] , the argon contribution to the background will remain well under the level of 10@xmath0 counts / kg kev y , while the background coming from krypton is expected to be of the level of 10@xmath26 counts / kg kev y , which is still acceptable for dark matter detection .
a further reduction of the krypton contamination by a factor of 10 will allow genius to fulfill its background requirement also for solar neutrino detection : it must be noted that a krypton purification of a factor 8 has already been reached from the heidelberg group of the borexino collaboration @xcite .
+ the genius experiment will also be equipped with a nitrogen recycling device ( through condensation ) @xcite , in order to avoid accumulation of argon and krypton inside the liquid : this would happen if the evaporated liquid would be replaced by fresh nitrogen .
a test facility for the genius experiment , genius - tf @xcite , is under construction at the gran sasso national laboratory , with at present six natural ge detectors ( about 15 kg ) operated in a 50 cm @xmath27 50 cm @xmath27 50 cm liquid nitrogen container .
+ the goal of genius - tf is to reach a background of 2 - 4 counts / kg kev y in the low - energy region .
the estimated contribution of argon and krypton contamination to the genius - tf background at low energies , under the same assumptions as made for genius , is less than 10@xmath26 counts / kg kev y .
even a contamination 100 times higher would not affect the background in a noticeable way .
+ the problem of the accumulation of argon and krypton in the liquid nitrogen because of the periodic refilling procedure has been considered .
assuming that argon and krypton do not evaporate and that we fill the container when the liquid nitrogen is half of the total volume , we would have to refill 200 times in order to increase the contamination by a factor 100 .
in the genius detector , a big amount of liquid nitrogen will be used both as cooling medium and shielding material and it will be in direct contact with the detectors @xcite .
for this reason the purity of the liquid nitrogen not only in terms of radon , but also of argon and krypton contamination , is a very crucial point .
the same , and even much more seriously , will be true for the borexino experiment , which foresees the use of gaseous nitrogen to purify the liquid scintillator , and for the kamland experiment when coming to the solar neutrino detection .
for the borexino experiment , only a level of nitrogen contamination of 0.36 ppm for argon and 0.16 ppt is acceptable @xcite , while the best - quality nitrogen commercially available has typically 0.03 - 0.2 ppm of argon and about 10 ppt of krypton .
a reduction on the krypton contamination by at least a factor 60 is then needed .
+ in this work we have shown that the above mentioned commercially available purity of 0.03 - 0.2 ppm of argon and 10 ppt of krypton would allow the genius experiment to look for dark matter at low energies ( @xmath15 100 kev ) , with a background of the order of 10@xmath26 counts / kg kev y , which corresponds to the planned sensitivity
. a further reduction of a factor 10 ( of which a factor of 8 has been already obtained , see @xcite ) will bring the background to the level of 10@xmath0 counts / kg kev y , which is the requirement of genius for its use as a solar neutrino detector .
one of us ( c.t . ) thanks the max planck institut fr kernphysik for the ospitality during her stay in heidelberg and would like to thank dr .
g.zuzel and b.freudiger for useful discussions .
kleingrothaus : in _ proc . of beyond the desert 1997
_ , ed . by h.v .
klapdor kleingrothaus , h. ps , iop bristol ( 1998 ) , pp .
485531 , h.v .
klapdor kleingrothaus , int .
a * 13 * , 3953 ( 1998 ) , h.v .
klapdor - kleingrothaus , m. hirsch , z. phys .
a * 359 * , 351 ( 1997 ) , h.v .
klapdor - kleingrothaus , j. hellmig , m. hirsch , j. phys .
g * 24 * , 438 ( 1998 ) .
klapdor - kleingrothaus , l. baudis , g. heusser , b. majorovits , h. ps , ` genius : a supersensitive germanium detector system for rare events ' , mpi - h - v26 - 1999 , hep - ph/9910205 ( 1999 ) and in _ proc .
of beyond the desert 1999 _ , ed . by h.v .
klapdor kleingrothaus and i. krivosheina , iop bristol ( 2000 ) , p. 915 - 1024 .
klapdor - kleingrothaus , in _ proc . of the international workshop on low energy solar neutrinos ( lownu2 )
_ , december 4 and 5 , 2000 , tokyo , japan , ed by y.suzuki et al . , world scientific ( 2001 ) 116 - 131 .
ar simulation : sum spectrum of all 10 detectors ( left ) and the same spectrum in units of counts / kg kev y ( right ) .
the dashed histogram represents the expected @xmath1ar background for a volume of liquid nitrogen 8 times bigger ( assuming the same argon contamination).,title="fig:",width=264,height=207 ] ar simulation : sum spectrum of all 10 detectors ( left ) and the same spectrum in units of counts / kg kev y ( right ) .
the dashed histogram represents the expected @xmath1ar background for a volume of liquid nitrogen 8 times bigger ( assuming the same argon contamination).,title="fig:",width=245,height=207 ] kr simulations described in table [ tab : simul ] .
each of the five histograms corresponds to a different step of the simulation and then to a different volume of liquid nitrogen around the detectors ( see table [ tab : simul]).,width=340,height=302 ] kr background simulated with the geometry of fig .
[ fig : simul ] , from the first simulation ( dashed histogram ) to the second ( solid histogram ) we increased the volume of liquid nitrogen , keeping the same @xmath2kr concentration.,width=340,height=302 ] | genius is a proposal for a large scale detector of rare events like double beta decay , cold dark matter and low - energy solar neutrinos in real time .
the idea of genius is to operate a large amount of `` naked '' ge detectors in liquid nitrogen , with the aim of reducing the background down to a level of 10@xmath0 counts / kg kev y. in this work we investigate the contribution to the background of genius coming from argon ( @xmath1ar ) and krypton ( @xmath2kr ) contamination in the liquid nitrogen .
dark matter , solar neutrinos , low - background experiments , high - purity ge detectors 07.05.fb , 95.35.d , 95.55.vj , 23.40.s , 14.60.pq , 23.40.bw , 14.80.ly |
the development over the last two decades of efficient and reliable wide - field multi - object spectrographs has resulted in enormous advances in the ability to compile large samples of high - quality astronomical spectra . the sloan digital sky survey ( sdss ; * ? ? ? * ) represents the most impressive application yet of such an instrument to provide spectroscopic samples of quasars , galaxies and stars of unprecedented size . the third data release ( dr3 ; abazajian et al .
2004b ) provides the astronomical community with nearly 500,000 object spectra and further releases will take the number to 750,000 within eighteen months .
the sdss spectra present a dramatic improvement in quality over data from previous surveys : extended wavelength coverage , @xmath0 ; intermediate resolution , @xmath3 ; high - quality relative and absolute flux - calibration ; typical signal - to - noise ratios ( s / n ) of @xmath4 ; availability of accurate noise and mask arrays . coupled with the overall homogeneity of the dataset this makes them suitable for an almost limitless number of quantitative investigations .
a particularly impressive aspect of the instrumental design , observing strategy and data reduction pipeline @xcite has been the quality of the sky - subtraction achieved considering the relatively large , 3 arcsecond diameter fibres and fully automated reduction .
subsets of the sdss spectra , such as the fainter quasars , include objects with magnitudes in red passbands equivalent to the sky - brightness per square arcsecond . despite the demonstrable quality of the sdss spectra
, visual inspection reveals significant systematic sky - subtraction residuals longward of @xmath1 in many spectra .
for fainter objects in particular , the sky - subtraction residuals are the dominant source of uncertainty over a wavelength interval of some @xmath5 . the spectroscopic region involved is extensive and includes features of significant astrophysical interest ; examples include the calcium triplet ( 8500 , 8545 , @xmath6 ) , a powerful diagnostic of stellar populations in low - redshift galaxies , and the h@xmath7 + [ oiii ] 4959 , @xmath8 emission region in quasars and active galactic nuclei ( agn ) in the redshift interval @xmath9 . with sufficient care ( e.g. * ? ? ?
* ) it is possible to quantify the noise properties of the sdss spectra in the red , allowing the statistical significance of features to be estimated reliably .
however , the decrease in the s / n relative to that expected from counting statistics can be factors of 2 - 3 . coupled with the substantial fraction of the spectral range affected ,
many potential scientific investigations are compromised .
the sky - subtraction procedures incorporated in the sdss spectroscopic pipeline are very effective , overcoming a number of the longstanding problems associated with wide - field multi - fibre observations .
brief details can be found in the early data release paper ( * ? ? ?
* sections 4.8.5 and 4.10.1 ) and information on calibration improvements for the second data release ( dr2 ) are given in abazajian et al .
( 2004a ) .
while the large area of sky , @xmath10 square arcseconds , observed through the fibre aperture exacerbates the problem of achieving accurate sky - subtraction , the origin of the dominant systematic residuals present in the sdss spectra is not a preserve of fibre spectroscopy alone .
rather , the fundamental issue lies in the ability to remove oh emission features from spectra in which the line profiles are barely resolved . combined with sub - pixel changes in the pixel - to - wavelength calibration between spectra
, sharp residuals often remain .
@xcite provides a recent summary of the difficulties associated with accurate sky - subtraction in long - slit observations , presenting a highly effective procedure for removing the signature of even rapidly varying and poorly sampled emission line features .
the technique described by kelson is certainly not applicable to the extracted fibre spectra available as part of the official sdss data releases and any improvement must rely on an alternative procedure .
the sky - subtraction residuals arise from the subtraction of two essentially identical tooth - comb signatures that have been very slightly misaligned relative to one another leading to well - defined patterns ( fig .
[ fig_skyegs ] ) .
this suggests a correction procedure that takes advantage of the correlations present in the wavelength direction , offering significant advantages over simply masking the affected pixels .
more specifically , we develop an approach for the removal of the dominant oh sky - subtraction residuals in the sdss spectra based on principal component analysis ( pca ) .
pca ( also called karhunen
love transformation ) is a well - established data reduction technique in astronomy , frequently applied to classification and reconstruction problems associated with large numbers of spectra of various types of object ( e.g. galaxies : connolly et al .
1995 ; folkes et al . 1996
; madgwick et al . 2002 , 2003 ; yip et al .
quasars : francis et al .
1992 ; yip et al .
2004b . stars : whitney 1983 ) .
the idea of employing pca for sky - subtraction has been suggested by @xcite , who present a complex method to remove the sky signal without the need for concurrent sky observations .
however , the scheme appears to have generated little interest , perhaps because of the ambitious goal of the technique and the targeting of observations in which sky spectra are unavailable .
by contrast , we develop a much more specific application of the pca technique to the sdss dr2 spectra that results in an improvement by over a factor of 2 in the s / n of those pixels longward of @xmath1 affected by oh sky lines , dramatically increasing the potential use of the red half of the sdss spectra for a variety of scientific investigations .
the technique is equally applicable to the spectra in the recent dr3 release .
we adopt the same convention as employed in the sdss and use vacuum wavelengths throughout the paper . in section [ sec_method ]
we present our method which is applied to sdss galaxy and quasar spectra in section [ sec_recon ] .
section [ sec_tests ] uses subsets of the sdss dr2 spectroscopic catalogue to demonstrate application of the method to various astronomical studies .
the code and documentation with which to carry out the method on public release sdss spectra is available on the web at http://www.ast.cam.ac.uk/research/downloads + /code
our overall approach is determined by the extent of the spectroscopic information provided in a sdss public data release .
the raw ccd exposures , typically 4 or 5 per spectroscopic plate , that make up the total exposure in each spectroscopic field are not available , precluding any attempt to improve the sky - subtraction on an exposure by exposure basis .
however , the sdss data releases do provide the final reduced spectra for both target objects and for the fibres allocated to blank sky regions employed to define the underlying spectrum of the night sky .
these contain significant diagnostic information about the quality of sky subtraction for each spectroscopic plate . for each observation of a sdss spectroscopic plate , approximately 32 fibres , 16 for each of the 2 spectrographs ,
are assigned to blank sky regions , selected from areas containing no detected objects in the sdss imaging survey .
the sky fibres are identified as `` sky '' in the catalogue s `` objtype '' and `` specclass '' field .
the sky spectra are combined to create a `` master - sky '' spectrum for each spectroscopic plate , which is then scaled and subtracted from each of the 640 spectra , producing 608 sky - subtracted object spectra and 32 sky - subtracted sky spectra , all of which are part of the standard sdss data releases . throughout the remainder of the paper we will refer to the sky - subtracted sky spectra as `` sky spectra '' , and the sky - subtracted object spectra as `` object spectra '' .
the availability of large numbers of sky spectra , over 18,500 in dr2 , thus provides a direct empirical measure of the noise resulting from counting statistics as well as any systematic residuals from the sky - subtraction procedure .
the scale of the sdss releases is such that the application of empirical self - calibration techniques , based on the statistical properties of the data set itself , can be a powerful tool
. visual inspection of the spectra of faint objects , or sky spectra ( fig .
[ fig_skyegs ] ) , immediately reveals the presence of `` patterns '' due to small imperfections in sky subtraction .
the presence of systematic deviations that are correlated over extended wavelength ranges suggests that a technique capable of quantifying the form and amplitude of the correlated deviations could allow the removal of a substantial fraction of the sky - subtraction noise .
pca is a well - established technique with a number of desirable properties for such an application .
the principles behind our sky - subtraction method are simple to understand : a pca of the sky spectra produces a set of orthogonal components that provides a compact representation of the systematic residuals resulting from the sky - subtraction process .
the components are then added in linear combinations to remove the systematic sky residuals from the spectra of target objects , such as galaxies .
our analysis focuses on the red part of the sdss spectra ( @xmath11 ) , as the strong emission features blueward of @xmath1 are small in number and easily masked .
the upper limit of @xmath12 is set by the requirement that the entire wavelength region be included in all spectra when creating the pca components from the sky spectra .
although techniques exist to overcome this @xcite , it was believed they would not greatly improve the applicability of the procedure for scientific objectives due to the very small fraction of plates affected .
the goal of the sky - subtraction scheme is to remove any systematic residuals from the bulk of the sdss spectra .
pathological spectra with very unusual deviations can have a disproportionate effect on the generation of the pca components .
therefore , we identify a restricted subset of the full 18,764 sky spectra for use in deriving the pca components . for the wavelength range used in the analysis ( @xmath11 ) the spectra must satisfy the following criteria , where the numerical values are in the units of the sdss spectra ( @xmath13 ) : 1 .
@xmath14 mean flux @xmath15 , ensuring the spectra have a mean close to zero
variance of the flux @xmath16 , ensuring that the amplitude of pixel - to - pixel fluctuations is not unusually high 3 .
`` spectral colours '' have values @xmath17 , @xmath18 and @xmath19 ensuring that the spectra do not exhibit significant large scale gradients .
@xmath20 , @xmath21 and @xmath22 are calculated by averaging the flux in three wavelength regions largely free of oh sky emission ( 7000:7200 [ a ] , 8100:8250 [ b ] and 9100:9180 [ c ] ) .
we further only accept spectra with a minimum of 3800 good pixels ( `` ngood '' parameter ) over the entire wavelength range of the spectrum , to eliminate spectra with substantial numbers of missing pixels .
following a trial run of the pca , 300 spectra are identified as outliers in the distribution of principal components , therefore dominating particular components , and discarded ( see section [ sec_pca ] ) .
application of these selection criteria leaves sky spectra .
finally , the spectra were normalised to have a median flux of zero over the wavelength range @xmath11 .
none of the results in the paper depend on the details of the criteria used to define the subset of sky spectra to be used in the pca - analysis .
the true error on each pixel in each spectrum is made up of components due to poisson noise and systematic sky - subtraction errors .
our method for removing sky residuals relies on determining the best - fit pca - components via a minimum least - squares criterion over all pixels within the @xmath11 wavelength range .
a pixel which varies greatly between spectra will be weighted highly during the generation of the pca - components and in the subsequent reconstruction of the sky - residuals present in individual spectra .
for example , there will be a greater variance among spectra at wavelengths in the vicinity of strong oh emission lines , purely due to poisson noise and independent of whether there is a contribution due to sky - subtraction errors .
it follows that to achieve effective sky - subtraction it is important to normalise each spectrum by the poisson noise expected at each pixel .
failure to do so results in over - subtraction for certain pixels , with `` corrected '' pixels apparently exhibiting fluctuations below the poisson limit .
each sdss spectrum includes a companion error array based on the original poisson errors derived from photon counts , ccd read - noise and so forth .
unfortunately for our application , the error arrays have been systematically increased in regions of poor sky subtraction .
the procedure is carried out separately for each plate and details can be found in the data reduction source code at http://spectro.princeton.edu/idlspec2d_doc.html#skysubtract .
[ fig_hedge1 ] shows the flux standard deviation ( rms ) of the sky spectra with and without normalisation by the associated sdss noise arrays .
if the sdss noise arrays accounted for all the variance present in the spectra , the lower plot in fig .
[ fig_hedge1 ] would be flat .
in fact the sdss noise arrays account well both for the increased `` continuum '' noise at red wavelengths , the presence of the atmospheric a - band at @xmath23 and the contribution from the increased sky emission associated with the presence of the broad o@xmath24 airglow emission centred on @xmath25 .
however , significant additional residual variance resulting from the incomplete subtraction of the barely resolved oh emission lines remains .
for many applications the resulting scaled error arrays , which better reflect the true error in each spectrum , are an improvement .
however , using the modified sdss noise arrays for our sky - subtraction procedure does not allow it to reach its full potential , due to the excessive down - weighting of strong oh features .
it is not possible to recover directly the original poisson errors that we require , so instead we develop an empirical approximation from investigation of several spectroscopic plates prior to sky subtraction , kindly made available to us by the sdss project ( e. switzer , p. macdonald and d. schlegel ) .
these data confirmed the expectation that scaling of the error arrays is applied by the sdss reduction pipeline predominantly at positions of oh emission features .
the same noise scaling is applied to all spectra on a plate , allowing us to derive a single rescaling for each plate . for each one of the 574 plates we calculate the median noise at each wavelength increment ( pixel ) from the sdss error arrays of the sky spectra .
the continuum noise is calculated over the full wavelength range of interest interpolating across pixels containing oh sky emission lines .
the amplitude of this continuum noise is dominated by counting statistics , and rises towards the red ( fig .
[ fig_hedge1 ] ; upper panel ) due to an increase in the sky level and instrumental effects . for this plate , which occurs at the wavelength of @xmath26 .
] we use a simple function to approximate the rescaling of the noise , @xmath27 , for each pixel @xmath28 on each plate : [ eq_scale ] s_i = 1+^ where @xmath29 is the median noise of the plate , @xmath22 the continuum noise , and max@xmath30 the maximum noise , above the continuum level , over all pixels . fig .
[ fig_egplate ] shows an example plate median noise array , with estimated continuum and max@xmath30 marked .
@xmath2 and @xmath7 are determined empirically as described below and represent , respectively , the relative amount of rescaling between pixels with differing amounts of noise , and the total rescaling of the pixel with the highest noise above continuum level .
as the exact effects of the sdss noise scaling are not understood , the form of this function is chosen to provide flexibility in both the total rescaling and dependence with wavelength .
( @xmath31 ) .
for clarity , the upper spectra have been displaced vertically by 2.5 units ( upper panel ) and 1.5 units ( lower panel ) .
the effectiveness of the sky - residual subtraction procedure is seen by the reduction by more than half of the systematic spikes in the top figure , and complete removal in the bottom figure . ]
anticipating the results of section [ sec_recon ] , the effectiveness of the pca sky - residual subtraction procedure can be seen in fig .
[ fig_hedge2 ] .
the top panel shows the effect of the sky - residual subtraction on the raw rms of the sky spectra : the lower spectrum shows the uncorrected rms as a function of wavelength ( reproducing the top spectrum from fig .
[ fig_hedge1 ] ) ; the upper spectrum shows the raw rms after application of the sky - residual subtraction scheme . note how the features due to the atmospheric a - band ( @xmath32 ) and o@xmath24-emission ( @xmath33 ) are unaffected , while the height of the spikes associated with the oh emission lines is greatly reduced . the variation with wavelength of the rms , post sky - residual subtraction ,
is explained almost entirely by counting statistics .
the bottom panel of fig .
[ fig_hedge2 ] shows the rms of the sky spectra weighted by our modified noise arrays ( with @xmath31 ) : the lower and upper spectra show the rms before and after subtraction of the sky residuals respectively . .
from bottom to top @xmath34 and @xmath35 . ] fig .
[ fig_beta ] shows the effect of varying the noise rescaling ( @xmath7 in eq .
( [ eq_scale ] ) ) applied to the sky spectra prior to creating the eigenspectra . from bottom to top @xmath36 . for low values of @xmath7 ( small amounts of rescaling of the noise arrays )
the noise weighted rms is approximately constant with wavelength , i.e. the sky - residual subtraction procedure is not removing signal below the poisson limit for any pixels , although the procedure may also not be removing the full systematic contribution to the errors . as beta increases , initially the rms remains constant until a point is reached where significant dips appear in the red half of the spectra .
the presence of these dips show that an unphysical situation has resulted , where , following the rescaling , the noise falls below the poisson limit for those pixels .
the same effect is seen if we undertake the sky residual subtraction without weighting the flux by the noise arrays , i.e. the noise is constant , independent of wavelength .
the fact that we can rescale the noise arrays by some amount before these unphysical dips appear , indicates that the sdss noise arrays have indeed been scaled to take account of the presence of a systematic contribution to the noise at wavelengths where the oh - emission is significant . under - subtraction of the sky - residual signal due to over estimation of the poisson errors
prevents our sky - residual subtraction method from reaching its full potential . over - subtraction of the sky - residual signal causes downward spikes to appear at the location of the oh emission lines ( fig .
[ fig_beta ] ) , as we begin to erroneously subtract poisson noise .
this suggests a simple method to estimate the values of @xmath2 and @xmath7 in eq .
[ eq_scale ] : we require the weighted flux rms , post sky - residual subtraction , to be as constant with wavelength as possible whilst maximising the noise rescaling .
we take a grid of values for @xmath2 and @xmath7 and for each combination first run the pca on the sky spectra with noise normalisation scaled according to eq .
( [ eq_scale ] ) , then perform with the sky - residual subtraction with the same noise weighting . by calculating the standard deviation of the rms array , post sky - residual subtraction , for each value of ( @xmath2 , @xmath7 )
we can derive the best empirical rescaling of the sdss noise arrays : we require the greatest value of @xmath7 that does not introduce upward or downward spikes , whilst @xmath2 allows purely for potential wavelength variations .
we find that @xmath35 and @xmath37 .
our final results are insensitive to small changes in @xmath2 or @xmath7 , and the measured value of @xmath35 suggests that the empirical rescaling of the noise arrays within the sdss spectroscopic pipeline is predominantly a function of the height of the oh spikes above the continuum noise . by systematically decreasing the amplitude of the noise arrays by 30% at positions of oh lines whilst retaining a constant noise weighted rms with wavelength , we substantially increase the effectiveness of the method , without over - subtracting residual sky features beyond the limit set by counting statistics .
we emphasise that it is not possible to derive the true noise array , based on counting statistics , for each sdss spectrum from the information contained in the publically available sdss data releases
. however , the close to constant rms as a function of wavelength achieved following our empirical rescaling of the noise arrays ( fig .
[ fig_hedge2 ] ; bottom panel ) is strong evidence that , while approximate , the scheme does achieve a highly effective correction , reducing the rms close to the level set by counting statistics at all wavelengths .
if a version of our sky - residual subtraction scheme could be applied to the sdss spectra using the true noise arrays , based on counting statistics , then further improvements should result .
the strong sky emission lines , which lead to the presence of systematic sky - residuals , occupy only a fraction of the @xmath11 wavelength range and it is necessary to identify those pixels which suffer significantly from sky subtraction errors .
the affected pixels can be found via the calculation of the rms of the sky spectra , normalised by the corresponding sdss noise arrays ( as shown in the lower panel of fig .
[ fig_hedge1 ] ) .
all pixels above a threshold value are defined as `` sky pixels '' .
the `` non - sky pixels '' are not included in the pca sky - residual removal procedure but provide an empirical estimate of the level of poisson noise ( i.e. excluding systematic sky - residuals ) present in each spectrum . adopting a threshold level of 0.85 for the sky / non - sky boundary , the scheme results in 670 sky pixels and 697 non - sky pixels . the results shown in subsequent sections are not sensitive to reasonable variations in the threshold level .
finally , to minimise their effect during the calculation of the pca components , sky pixels in each spectrum where the associated sdss noise array is zero ( i.e. no data ) are set to the mean value of the pixel in all the sky spectra that contain data .
pca is a simple statistical method for data reduction which looks for components in a dataset that vary the most .
the technique can be visualised by imagining an @xmath38-dimensional array , where @xmath38 is the number of pixels in the spectra . each spectrum
is then represented by a point in the @xmath38-dimensional array ; pca searches for lines of greatest variance through the array , each one being orthogonal to all previous lines .
the line of greatest variance defines the first pca component , and so on until the line of least variance defines the @xmath39 pca component .
the projection of an individual spectrum back onto a pca component gives the amplitude of the component contained in that particular spectrum .
a spectrum is reconstructed by summing the principal components multiplied by the relative amplitudes for that spectrum .
appendix [ ap_pca ] presents the mathematics of pca .
having created a sample of sky spectra containing only those pixels with sky signal and weighted by our empirically derived noise array for the relevant plate , we run the pca .
the output consists of @xmath38 principal components which are @xmath38 pixels long .
each component has an associated variance , which gives the percentage of the total variance of the dataset contained in that component .
as pca can be sensitive to spectra with particularly unusual features in which we are not interested , we remove spectra which dominate the signal in individual components after one trial run .
the pruning , of 300 spectra in total , is achieved by removing those spectra with principal component amplitudes more than @xmath40 from the mean .
the pca is then rerun on the resulting sample of sky spectra .
[ fig_temp ] shows the first 5 principal components resulting from the analysis .
note the distinctive , correlated features present in each component .
once the components are created , they can be used as templates to reconstruct the input sky spectra and later the sky residuals in the object spectra ( section [ sec_reconobj ] ) .
this is done by projection of each spectrum onto the components and summation of the components weighted by the projection coefficients ( eq . [ eq_recon ] and [ eq_recon ] ) .
the reconstruction may then be subtracted from the sky spectra leaving residual - free spectra , termed sky - residual subtracted spectra .
the number of components to use in the reconstruction is not well defined . the use of too many results in the artificial suppression of noise below the poisson limit , with the pca acting as an ( undesirable ) high - frequency filter .
the use of too few components means that the removal of sky residuals is sub - optimal and , in some cases , the overall quality of the spectra can decrease .
[ fig_nrecon ] shows the mean ratio of the rms of the noise - weighted flux in the sky pixels to that in the non - sky pixels , for the sky - residual subtracted sky spectra , as a function of the number of components used in the reconstructions .
the rms of the non - sky pixels in each spectrum remains constant and the ratio decreases monotonically as more components are used in the reconstructions , with an increasing fraction of the noise in the sky pixels removed .
the reduction of the noise in the sky pixels below the noise in the non - sky pixels is clearly unphysical , and we therefore estimate the number of components to employ in the reconstruction of each spectrum by adopting the non - sky pixel rms of that spectrum as a reference .
the reconstruction of a spectrum proceeds one component at a time , with the rms ratio calculated for the sky - residual subtracted spectrum .
reconstruction is stopped when the rms ratio reaches unity , i.e. when the noise weighted flux rms is the same for the sky and non - sky pixels .
the scheme is largely self - calibrating .
for example , in spectra with high s / n the systematic sky residuals typically contribute only marginally to the sky - pixel noise , the rms ratio thus starts close to unity and only a small number of components are necessary to achieve equality in the rms ratio .
the large number of pixels that contribute , combined with the robust estimator of the non - sky rms , mean the amplitude of the non - sky rms is well determined , providing an excellent indicator of when to halt the reconstructions .
[ fig_hist ] shows the distribution of the number of components used for dr2 sky , galaxy and quasar spectra . for bright objects
the contribution to the noise from imperfections in the sky subtraction is small and few components are required to bring the sky and non - sky rms to the same level .
the presence of many spectra with high s / n produces the `` spike '' at low component numbers in both the galaxy and quasar sample histograms . the fainter limiting magnitude of the quasar sample make sky - residuals a problem in a larger fraction of the spectra and , on average , more components are required to remove the systematic errors present following the default sdss - pipeline sky - subtraction . to limit the effect of very occasional poor estimation of the non - sky rms we set the maximum number of components to 150 and 200 for the galaxy and quasar samples respectively . in the preceding sections
we have developed a procedure that achieves a substantial reduction in the amplitude of the systematic sky residuals in spectra that possess no large scale signal , i.e. the successfully sky spectra are known , by definition , to possess zero signal at all wavelengths .
we now turn to the more interesting application of removing the systematic sky - residuals from the sdss science spectra of targets such as galaxies and quasars .
the problem is made tractable by the success of the sky - subtraction procedure performed as part of the standard sdss spectroscopic pipeline .
detailed examination of the properties of the sky spectra shows that both the mean level and the large scale shape of the sky spectrum to be subtracted from each spectrum have been determined to very high accuracy . as a result , there is effectively no sky `` continuum '' to remove , rather the problem is confined to removing the high - frequency structure due to the presence of the strong oh sky emission lines .
the key goal is to ensure that object continuum and intrinsic absorption and emission line features are not removed as a result of the procedure used to reduce the amplitude of the systematic sky - residuals .
as it is not necessary to identify any sky continuum present , we can remove the continuum of the object spectrum using a median filter before projection of the spectrum onto the principal components derived from the sky spectra .
the smallest filter size can be found from median filtering the sky spectra as it is important that the filter is unaffected by the oh residuals ( about 40 pixels ) . as the filter size
is increased above this minimum , some features intrinsic to the object fail to be removed .
this generally causes an increase in the rms of the non - sky pixels used as a reference point to halt the reconstruction ( see previous section ) and therfore a decrease in the number of components used and slight decrease in the effectiveness of the sky - residual subtraction procedure .
however , the final effect is small and insignificant for a reasonable range of filter sizes ( up to about 80 pixels ) . we have used a filter size of 55 pixels throughout this paper . in certain circumstances
, it may be desirable to reduce the filter scale in order to follow the broad emission line profiles in quasars closer to the line centroids .
narrow line features present the greatest challenge , as these can be mistaken for an oh sky residual by the pca .
the advantage of pca in this task is that it looks for patterns in the spectrum , linking lines together which are correlated in the input sky spectra , and weights all bins equally .
however , if an emission or absorption feature occurs exactly at the location of an oh line , some combination of principal components can sometimes be found to reconstruct the feature , without increasing the rms in the rest of the spectrum significantly .
such behaviour is particularly likely if the line feature lies in a noisy part of the spectrum . for most applications in which the sky - residual subtraction scheme might be employed , it is possible to mask known emission and absorption features , thereby circumventing the problem . in section [ sec_cat ]
we show how such masked features are unaffected by the sky - residual subtraction procedure .
the disadvantage is that real sky features which happen to fall in the masked regions do not contribute to the projection onto the principal components and the sky - residual subtraction may not be fully effective in these regions . in some potential applications , it is not known in advance where absorption and emission features may occur and we present such an example in section [ sec_dla ] .
features are masked by removing the relevant pixels during projection onto the principal components .
alternatively , replacing the pixels with the local mean results in almost identical reconstruction .
we similarly mask bad pixels ( where the sdss noise array is set to zero ) . once the object spectrum has been continuum subtracted , we divide by the sdss noise spectrum of the relevant plate and project onto the sky principal components ( eq . [ eq_pcs ] ) , leaving out those pixels identified as possibly containing emission and absorption features .
the resulting principal component amplitudes are used to reconstruct the residual sky - subtraction signal ( eq . [ eq_recon ] ) and the reconstruction is subtracted from the object spectrum , including masked pixels .
re - multiplication by the noise spectrum , followed by the addition of the object continuum , returns the object spectrum cleaned of oh sky emission line residuals .
[ fig_gal1 ] illustrates the process of sky - residual subtraction on a typical sdss galaxy spectrum .
each class of spectroscopic science targets has different spectral characteristics .
the effective application of the sky residual subtraction scheme requires the removal of the large scale `` continuum '' signal from the target object and the identification of wavelength intervals where narrow emission or absorption features may be present .
the extremely high identification success rate achieved in the sdss means that the wavelengths of strong absorption and emission features in the spectra of stars and galaxies are known ; there are only 3,002 spectra without secure identifications among the 329,382 object spectra included in dr2 .
potential systematic biases in the sky residual subtraction can be prevented by masking the wavelength intervals that include such features .
the exact nature of the pre - processing applied to spectra prior to the implementation of the sky residual subtraction depends on the scientific goal . however , in subsequent sections we describe schemes that are likely to have wide application in the analysis of the 3 main classes of sdss science targets : galaxies , quasars and stars .
the dr2 catalogue contains 249,678 unique objects spectroscopically classified as galaxies .
the strong systematic trends in the absorption- and emission - line properties of the galaxies as one moves from early through to late type objects necessitates a sub - classification of the population prior to the application of the sky - residual subtraction .
therefore , we classify each galaxy as either an absorption line , emission line or extreme emission line object .
the spectral type parameter ( `` eclass '' ) in the sdss catalogue provides the basis for the absorption line ( eclass@xmath41 ) and emission line ( eclass @xmath42 ) classification .
the extreme emission line galaxies are identified through the presence of emission lines with very large equivalent widths ( ew ) ( @xmath43 , @xmath44 ; or @xmath45 } } > 200$] , @xmath46 ) .
an appropriate feature mask is then applied depending on the galaxy type ( e.g. see fig . [ fig_gal1 ] ) .
a total of 331/286/469are masked in absorption / emission / extreme emission line objects respectively , over the entire rest - frame wavelength range of @xmath47 , i.e. in the observed - frame wavelength range of @xmath11 only a small fraction of pixels are affected by the line masks .
[ fig_gal2 ] shows examples of galaxy spectra before and after sky residual removal .
the dr2 catalogue contains 34,674 unique objects spectroscopically classified as quasars or high - redshift quasars .
a feature mask includes narrow emission lines ( e.g. [ oiii ] @xmath48 ) and @xmath49intervals ( observed frame ) centred on the broad emission lines ( e.g. civ @xmath50 ) .
the wings of the broad emission features can , in practice , be regarded as `` continuum '' in the context of the sky residual subtraction as they are removed by the median filtering .
[ fig_qso ] shows examples of quasar spectra before and after sky subtraction .
there are 42,027 unique objects classified as stars or late - type stars in the dr2 . due to their single redshift , application of the sky - subtraction procedure
is even simpler .
depending on the final science required and the type of stars involved , a suitable feature mask and continuum estimation can be straightforwardly derived but our adopted median filter scale of 71 pixels , employed for the galaxies and quasars , produces very satisfactory results .
in this section we present the quantitative results of applying the sky - residual subtraction to a variety of object spectra . [ cols="^,^,^,^,^,^,^ " , ] the detection of intervening absorption features in quasar spectra is an example of an investigation in which it is not possible to simply mask the wavelength regions containing features before the application of the sky residual subtraction .
that is not to say that the sky - residual subtraction scheme is not of potential interest .
the improvement in the quality of the spectra is such that the s / n of specific features can increase , albeit at the potential cost of modification of the feature properties . in practice , a hybrid scheme , involving one or more iterations , with detected features masked in a second application of the sky - residual subtraction ,
is straightforward to implement . in any particular application ,
a full understanding of the advantages of employing the sky - residual subtraction scheme would require the use of ( straightforward ) monte carlo simulations to quantify the probability of feature detection as a function of wavelength .
such an investigation is beyond the scope of the present paper but to illustrate the impact of the sky - residual subtraction on unmasked absorption features we create a composite spectrum of damped ly@xmath2 ( dla ) systems identified by @xcite .
each quasar spectrum is shifted to the rest frame of the dla absorber and normalised to possess the same signal in the absorber rest - frame wavelength interval @xmath51 .
the composite spectrum is then constructed using an arithmetic mean of all the spectra with signal at each wavelength , taking care to account for the flux / term in the sdss spectra .
[ fig_dla1 ] shows the resulting composite spectrum , calculated with and without the application of the sky - residual subtraction to the constituent quasar spectra . as in section [ sec_qso ] the centres of prominent broad emission lines in the quasars
are masked during the sky - residual subtraction .
[ fig_dla2 ] shows 3 versions of the composite spectrum , focusing on the rest - frame wavelength region that derives from observed - frame wavelengths @xmath52 .
the 3 versions of the composite were calculated using quasar spectra without sky - residual subtraction ( bottom ) , with sky - residual subtraction ( middle ) and with sky - residual subtraction _ and _ the absorption features masked .
the absorption - line mask consisted of 14@xmath53 intervals , in the absorber rest - frame , centered on each absorption line .
absorption - line equivalent widths ( ews ) are measured in the absorber rest - frame , using the method of section [ sec_cat ] .
table [ tab_dla ] includes the ews of the absorption features in the 3 composite spectra , together with the wavelength intervals used in the line and continuum measurement .
for each composite the same continuum is used for measuring ews , in this case calculated from the spectrum with lines masked , although in practice any consistent continuum would suffice .
the sky - residual subtracted composite , with masking of absorption lines , provides the best reference .
as expected , the ews of several of the absorption lines in the unmasked composite are systematically reduced .
however , the illustration is a `` worst case '' example of the sky - residual subtraction scheme involving only a single iteration .
a second application of the sky - residual subtraction using a mask based on the identification of features following the first application , offers the prospect of significant improvement in the identification and measurement of features at initially unknown wavelengths .
poor sky subtraction affects most those spectra with low s / n , and the fainter quasars , including those at high redshift ( @xmath54 ) , suffer from significant contamination from sky subtraction residuals .
the extended wavelength range of the sdss spectra combined with a redshift range for the quasars of @xmath55 , allows the construction of composite spectra over unprecedented ranges in rest - frame wavelength and luminosity @xcite .
however , the number of quasars that occupy the extremes of the wavelength and luminosity ranges is relatively small and the sky subtraction residuals at @xmath56 limit both the determination of the continuum properties and the detection of weak emission features .
we create composite spectra of all quasars in the sdss with @xmath57 , in redshift slices of @xmath58 .
each input quasar spectrum is normalised to have a mean flux of 1 between @xmath59 , prior to combination using an arithmetic mean .
[ fig_qsocomp ] illustrates how the sky - residual subtraction technique improves the quality of these composites .
subtracting the underlying spectrum using a median filter and removing the centers of emission lines from the calculation , we calculate the error weighted absolute deviation from zero on the composites with and without sky - residual subtraction .
the average improvement in s / n over the entire wavelength region due to the sky subtraction is about 20% for all redshift bins .
the sdss spectra represent a vast improvement in data quantity , quality and wavelength coverage over previous surveys .
the sky - subtraction carried out by the sdss reduction pipeline has in general achieved excellent results , however the common problem of the subtraction of undersampled oh emission lines results in substantial systematic residuals over almost half the spectral range .
a technique is presented to remove these residual oh features , based on a principal component analysis of the observed - frame sky spectra which have been sky subtracted along with all other spectra in the sdss data releases .
the scheme takes advantage of the high degree of correlation between the residuals in order to produce corrected spectra whose noise properties are very close to the limit set by counting statistics .
the sky - residual subtraction scheme is applied to the sdss dr2 catalogue and is immediately applicable to the recent sdss dr3 release .
the precise form of the implementation depends on the spectral energy distributions of the target objects and the scientific goal of any analysis . as a consequence ,
the main classes of target objects , galaxies , quasars and stars , are treated slightly differently .
constructing composites of high - redshift quasars before and after sky - residual subtraction illustrates the significant increase in the overall s / n of the spectra that is achieved .
the application of the procedure when narrow absorption or emission features are present requires the use of a mask to prevent the sky - subtraction scheme suffering from bias .
both the calcium - triplet in low redshift galaxies and metal absorption lines in damped ly@xmath2 systems at high redshifts fall in the wavelength range affected by the oh - forest ( @xmath60 ) . by employing these two rather different examples it is shown that masking absorption and emission features prior to the reconstruction of the sky - residual signal in each spectrum ensures that the properties of the features themselves are not systematically biased in the resulting sky - residual subtracted spectra . the significant improvement in s / n achieved over some @xmath5 of a large fraction of sdss spectra , particularly for the fainter objects such as the high - redshift quasars , should benefit a wide range of scientific investigations .
we have made available datafiles , idl code and a comprehensive guide on using the code , with which the procedure can be applied to all sdss spectra ( http://www.ast.cam.ac.uk/research/downloads + /code
further improvements should be possible if the sky - residual subtraction procedure is adapted to run on the sdss spectra using the original noise arrays based on counting statistics alone .
the nature of the sky - residual subtraction scheme is such that application to other large samples of spectra should be relatively straightforward .
many thanks to eric switzer , david schlegel and patrick mcdonald of the sdss team who made available information and answered questions about the sdss spectroscopic reduction pipeline .
we would also like to thank the referee , simon morris , for his careful reading of the manuscript and helpful comments and suggestions .
vw acknowledges the award of a pparc research studentship .
funding for the sloan digital sky survey ( sdss ) has been provided by the alfred p. sloan foundation , the participating institutions , the national aeronautics and space administration , the national science foundation , the u.s .
department of energy , the japanese monbukagakusho , and the max planck society .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium ( arc ) for the participating institutions .
the participating institutions are the university of chicago , fermilab , the institute for advanced study , the japan participation group , the johns hopkins university , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , university of pittsburgh , princeton university , the united states naval observatory , and the university of washington .
k. , adelman - mccarthy j. k. , ag " ueros m. a. , et al .
( the sdss collaboration ) , 2004a , , 128 , 502 k. , adelman - mccarthy j. k. , ag " ueros m. a. , et al .
( the sdss collaboration ) , 2004b , preprint ( astro - ph/0410239 ) a. s. , burles s. , schlegel d. j. , eisenstein d. j. , brinkmann j. , 2004 , , 127 , 1860 a. j. , szalay a. s. , 1999 , , 117 , 2052 a. j. , szalay a. s. , bershady m. a. , kinney a. l. , calzetti d. , 1995 , , 110 , 1071 a. i. , terlevich e. , terlevich r. , 1989 , , 239 , 325 a. , 1984 , , 286 , 97 g. , fall s. m. , 1984 , , 206 , 453 s. r. , lahav o. , maddox s. j. , 1996 , , 283 , 651 p. j. , hewett p. c. , foltz c. b. , chaffee f. h. , 1992 , , 398 , 476 hewett p. c. , irwin m. j. , bunclark p. , bridgeland m. t. , kibblewhite e. j. , he x. t. ,
smith m. g. , 1985 , , 213 , 971 d. d. , 2003 , , 115 , 688 m. g. , 1975 , multivariate analysis .
griffen , london m. j. , mink d. j. , 2000 , , 533 , l183
d. s. , coil a. l. , conselice c. j. , cooper m. c. , davis m. , ellis r. s. , faber s. m. , finkbeiner d. p. , gerke b. , guhathakurta p. , kaiser n. , koo d. c. , newman j. a. , phillips a. c. , steidel c. c. , weiner b. j. , willmer c. n. a. , yan r. , 2003 , , 599 , 997 d. s. , lahav o. , baldry i. k. , et al .
( the 2dfgrs team ) 2002 , , 333 , 133 f. , heck a. , 1987 , multivariate data analysis .
astrophysics and space science library , dordrecht : reidel , 1987 r. n. , long d. c. , snedden s. a. , et al .
( for the sdss collaboration ) , 2004 , preprint ( astro - ph/0408167 ) prochaska j. x. , herbert - fort s. , 2004 , pasp , 116 , 622 g. a. , hesser j. e. , stetson p. b. , mateo m. , simard l. , bolte m. , friel e. d. , copin y. , 1997 , , 109 , 883 c. , lupton r. h. , bernardi m. , et al .
( the sdss collaboration ) , 2002 , , 123 , 485 e. , diaz a. i. , terlevich r. , 1990 , , 242 , 271 d. , yip c. , connolly a. , jester s. , stoughton c. , 2004 , in asp conference series , volume 311 p. 21
c. a. , 1983 , , 51 , 443 c. w. , connolly a. j. , szalay a. s. , budav ' ari t. , subbarao m. , frieman j. a. , nichol r. c. , hopkins a. m. , york d. g. , okamura s. , brinkmann j. , csabai i. , thakar a. r. , fukugita m. , ivezi ' c .
, 2004a , , 128 , 585 c. w. , connolly a. j. , vanden berk d. e. , ma z. , frieman j. a. , subbarao m. , szalay a. s. , richards g. t. , hall p. b. , schneider d. p. , hopkins a. m. , trump j. , brinkmann j. , 2004b , aj , 128 , 2603 d. g. , adelman j. , anderson j. e. , et al .
( the sdss collaboration ) , 2000 , , 120 , 1579
for completeness we include the basic mathematics of pca .
further details can be found in @xcite and , for astronomical applications , in @xcite and @xcite . for an @xmath61 spectra by @xmath38 pixel data array @xmath62 , where @xmath63 , @xmath64 and @xmath65 , the covariance matrix is c_jk = _
i=1^n x_ijx_ik .
the eigenvectors ( @xmath66 , principal components in the language of this paper ) and eigenvalues ( @xmath67 ) of the covariance matrix are _ j
= _ j _ j. it can be shown that @xmath68 is the axis along which the variance is maximal , @xmath69 is the axis with second greatest variance , and so on until @xmath70 has the least variance .
the principal component amplitudes for each noise weighted input spectrum @xmath71 are given by [ eq_pcs ] a_j = _
j. the @xmath38 eigenvectors can be used as a basis set on which to project any spectrum of the same dimensions .
the principal component amplitudes of this projection and the eigenvectors can then reconstruct the input spectrum from the templates : [ eq_recon ] f_k = _ j=1^m
the reconstructed spectrum only contains information present in the templates , which may or may not be a fair representation of the input spectrum . in general as the variance of the eigenvectors decrease , so does the useful information contained in the spectra , allowing pca to be a useful form of data compression .
the exact number of eigenvectors required to reconstruct the input spectrum satisfactorily depends on the dataset and purpose of the analysis . | the sloan digital sky survey ( sdss ) currently provides by far the largest homogeneous sample of intermediate signal - to - noise ratio ( s / n ) optical spectra of galaxies and quasars .
the fully automated sdss spectroscopic reduction pipeline has provided spectra of unprecedented quality that cover the wavelength range @xmath0 . however , in common with spectra from virtually all multi - object surveys employing fibres , there remain significant systematic residuals in many of the spectra due to the incomplete subtraction of the strong oh sky emission lines longward of @xmath1 .
these sky lines affect almost half the wavelength range of the sdss spectra , and the s / n over substantial wavelength regions in many spectra is reduced by more than a factor of 2 over that expected from counting statistics .
we present a method to automatically remove the sky residual signal , using a principal component analysis ( pca ) which takes advantage of the correlation in the form of the sky subtraction residuals present in each spectrum .
application of the method results in spectra with essentially no evidence for degradation due to the incomplete subtraction of oh emission features . a dramatic improvement in the quality of a substantial number of spectra , particularly those of faint objects such as the bulk of the high - redshift quasars ,
is achieved .
we make available idl code and documentation to implement the sky residual subtraction scheme on sdss spectra included in the public data releases . to ensure that absorption and emission features intrinsic to the object spectra do not affect the subtraction procedure line masks
must be created that depend on the scientific application of interest .
we illustrate the power of the sky - residual subtraction scheme using samples of sdss galaxy and quasar spectra , presenting tests involving the near - infrared caii triplet absorption , metal absorption line features in damped ly@xmath2 systems and composite spectra of high - redshift quasars .
methods : statistical surveys techniques : spectroscopic |
the problem of the formation of the structures in the universe is one of the most important problem of modern extragalactic astronomy and cosmology .
one of interesting aspects of this problem is the analysis of the orientation of galaxies inside the galaxy structures .
the very important question is if there exists dependence on the alignment to the mass of the analyzed structures .
it is because the investigation of the orientation of galaxies planes is regarded as a standard test of galaxies formation scenarios . @xcite .
there was a lot of investigation of the orientations of galaxies inside clusters ( see @xcite for the latest review ) .
godowski et al @xcite suggested that alignment of galaxies in cluster should increase with the number of objects in a particular cluster .
that suggestion was confirmed by aryal et al .
@xcite , based on the series of papers @xcite .
however both @xcite and @xcite analysis were qualitative only .
for this manner godowski at al .
@xcite analyzed sample of 247 rich abell clusters using statistical tests and found that alignment increases with the richness of the clusters .
the analysis of the orientation of galaxies in poor galaxy structures - tully groups of galaxies was performed by godowski et al .
@xcite . in this paper
it was found that the group do not exhibit clear evidence for existence of alignment in the investigated structures .
however they concluded , that observational effect generated by the process of deprojection of galaxies @xcite , later confirmed by @xcite , masks to the high degree any possible alignment during analysis of the spatial orientation of galaxies in clusters .
for these reasons we analyzed the orientation of galaxy in tully galaxy group in more details .
.test for isotropy of the orientations of galaxy plane . the distribution of the angle @xmath2 of galaxies , inclination taken directly from ngc catalog .
[ cols="^,^,>,>,>,>,>,^,>",options="header " , ] at first , following godowski et al .
@xcite but adding the k - s test , we analyzed orientation of galaxies in tully groups using , for obtaining @xmath0 and @xmath1 angles , inclinations angles taken directly from nbg catalog @xcite ( sample a ) .
the results were presented in the tables 1 - 3 .
analysis of the supergalactic position angles showed that only one group ( @xmath3 ) exhibits alignment of galaxies .
analysis of the distribution of the angles giving spatial orientation of galaxies ( @xmath0 and @xmath1 ) seems to show weak alignment . for @xmath0 angle three tests
showed that the distribution is non random in the case of the clusters @xmath4 , @xmath5 and @xmath6 .
two tests showed nonrandomness in the case of clusters @xmath7 and @xmath8 . for @xmath1 angle three tests showed alignment in the case of clusters @xmath4 , @xmath9 , @xmath7 , @xmath8 and @xmath10 .
two tests showed it in the case of clusters @xmath5 and @xmath6 . for more detailed analysis we used the method described in @xcite .
the question which arose , is if we could say that we found an alignment in the analyzed sample of @xmath11 tully groups of galaxies .
so we computed the mean value and variance of analyzed statistics : @xmath12 , @xmath13 , @xmath14 ( i.e. the same statistics as was analyzed in @xcite ) for our sample of @xmath11 groups and compared it with results of numerical simulations .
we performed 1000 simulations of 18 fictious clusters , each with number of randomly oriented , members galaxies , the same as in real clusters . in the table 4 we present , obtained from numerical simulations average values of the analyzed statistics , their standard deviations , standard deviations in the sample as well as their standard deviations for distribution of @xmath15 angles .
one should note that there are some differences in results of numerical simulations for @xmath15 , @xmath0 and @xmath1 angles but it does not change our further conclusions . the mean values and variance of analyzed statistics for sample of real clusters are presented in the table 5 . on can show that ( for sample a ) analysis of the position angles does not show significant deviation from the values expected in the case of random distributions , while it seems that analysis of angles @xmath0 and @xmath1 shows existence of alignment at the @xmath16 level ( with exception of @xmath14 statistics for @xmath0 angle ) .
however , below we will argue against such interpretation .
the tully groups were analyzed also by godowski and ostrowski @xcite .
for every cluster the parameter @xmath17 describing the galactic axes alignment with respect to a chosen cluster pole , divided by its formal error @xmath18 ( @xmath19 ) were mapped .
the cluster pole coordinates change along the entire celestial sphere .
the resulting maps were analyzed for correlations of their maxima with important points on the maps ( see fig.1 for details ) .
it was found that maxima correlated well with the line of sight direction .
godowski and ostrowski @xcite concluded that the strong systematic effect , generated by the process of galactic axis de - projection from its optical image , is present in the catalogue data .
the example of that effect is presented in the fig.1 . to avoid possible influence of global alignment inside ls ,
we choose for presentation the cluster @xmath9 ( ursa major cloud ) because it is far from the center of the local supercluster ( i.e. virgo cluster ) . to remove above effect we should avoid the assumption that the `` true '' axial ratio is @xmath20 , which is a rather poor approximation , especially for non - spiral galaxies .
fortunately nbg catalog contains morphological types of galaxies .
these allowed us to use different values of @xmath21 depending on morphological type @xcite .
now , with help of fouque & paturel @xcite formulae , which convert @xmath22 to standard photometrical axial ratios , we compute new inclination angle @xmath23 for all galaxies in ngb catalog .
we repeated our investigation with that `` new '' sample of galaxies ( sample b ) .
as one can see from right panel of the fig.1 , the `` line of sight '' effect disappeared which shows that our procedure for computation of the inclination angles are much better than the previous one . from the tab.6 and tab.7
we can show that during analysis of the spatial orientation of galaxies , the alignment is observed only in the case of @xmath1 angle for clusters @xmath4 and @xmath7 ( virgo cluster and virgo - libra cloud ) .
of course our procedure does not change position angles .
in the tab.8 we presented analyzes of the distribution of @xmath15 where only galaxies with the certainly measure positions angles were taken into account . on can show ( tab.5 , sample @xmath24 ) that now the mean value of analyzed statistics does not exhibit any significant deviation from the values expected in the case of random distributions .
the above results allowed us to conclude that we did not observe any significant alignment for tully groups of galaxies .
we investigated the orientation of galaxies inside @xmath11 tully groups of galaxies belonging to the local supercluster , do not founding any significant alignment .
so we conclude that orientations of galaxies in the tully groups are random .
we also analyzed observational effect generated by the process of deprojection of galaxies found by godowski and ostrowski @xcite , which masks to the high degree any possible alignment during analysis of the spatial orientation of galaxies in clusters .
we showed that using `` true shape '' of galaxies @xmath21 depending on morphological type according to heidmann et al .
@xcite with help of fouque & paturel @xcite corrections of @xmath22 to standard photometrical axial ratios , allowed us to avoid this problem .
this gives much more powerful investigation of the spatial orientation of galaxies . in our prvious papers we found that in the sample of @xmath25 rich abell galaxy clusters we observed an alignment , which increased with the cluster richness @xcite
now , we found that the orientation of galaxies inside the poor galaxy structure is random .
it confirms our suggestion that alignment of galaxies increases with the mass of the structures @xcite . usually such dependence between the angular momentum and the mass of the structure
is presented as empirical relation @xmath26 @xcite .
the observed relation between mass of the galaxy structure and the alignment is compatible with the prediction of the li model @xcite in which galaxies are forming in the rotating universe .
however , in our opinion , it is due to tidal torque , as suggested by @xcite .
also the results of the analysis of the linear tidal torque theory which noticed the connection of the alignment with the considered scale of the structure is pointing in the same direction @xcite .
aryal , b. , saurer , w. 2004 , a&a , 425 , 871 aryal , b. , saurer , w. 2005 , a&a , 432 , 431 aryal , b. , saurer , w. 2006 , mnras , 336 , 438 aryal , b. , paudel , s. , saurer , w. 2007 , mnras , 379 , 1011 baier , f. w. , godowski , w. , macgillivray , h. t. 2003 , a&a , 403,847 bower , r. g. , benson , a.j . , malbon , r. , helly , j. , frenk , c. s. , baugh , c. m. , cole , s. , lacey , c. g. 2006 , mnras , 370 , 645 brosche , p. 1986 , comm .
astroph . , 11 , 213 carrasco , l. , roth , m. , serrano , a. 1982 , a&a , 106 , 89 catelan , p. , theuns , t. 1996 mnras , 282 , 436 dekel , a. 1985 , apj , 298 , 461 doroshkevich , a. g. 1973 , apl , 14 , 11 flin , p. , godowski , w. 1986 , mnras , 222 , 525 flin , p. , godowski , w. , 1989 sov . astron
15 , 374 ( pisma w astronomiczeskij zurnal 15 , 867 ) flin , p. , godowski , w. , 1990 sov .
65 , 209 ( pisma w astronomiczeskij zurnal 16 , 490 ) fouque , p. , paturel , g. , 1985 a&a , 150 , 192 godowski , w. 1993 , mnras , 265 , 874 godowski , w. 1994 , mnras , 271 , 19 godowski , w. 2011 , ijmpd 20 , 1643 godowski , w. , 2011 , arxiv 1110.2245 godowski , w. , baier , f.w .
macgillivray , h.t . , 1998 , a&a 339 , 709 godowski , w. , flin , p. , 2010 , apj 708 , 902 godowski , w. , ostrowski , m. , 1999 , mnras 303 , 50 godowski , w. , szydowski , m. , flin , p. 2005
rel . grav .
37 , ( 3 ) 615 godowski , w. , piwowarska , p. , panko , e. , flin , p. , 2010 ,
apj , 723 , 985 hawley , d. i. , peebles , p. j. e. 1975 , aj , 80 , 477 heavens , a. , peacock , j. 1988 , mnras 232 , 339 heidmann , j. , heidmann , n. , de vaucouleurs 1972 mem.r.ast.soc .
75 , 85 holmberg , e. 1946 , medd . lund .
117 jaaniste , j. , saar , e. 1978 , in : the large scale structures of the universe . , eds . m. s. longair and j. einasto , d. reidel , dordrecht ( iau symp .
79 ) , p.488 labuerts , a. 1982 , eso / uppsalasurvey of the eso b atlas , eso : garching lauberts , a. , valentijn , e. 1989 , the surface photometry catalogue of the eso - uppsala galaxies , eso : garching li , li - xin . , 1998 ,
grav , , 30 , 497 nilson , p. 1973
, uppsala general catalogue of galaxies , astr .
v , vol.1 : uppsala nilson , p. 1974 catalogue of selected non - ugc galaxies , uppsala astr .
. rep . 5 : uppsala noh , y. , lee , j. 2006 , astro - ph/0602575 noh , y. , lee , j. 2006 , apj , 652 , l71 pik , e.j .
1970 , irish aj , 9 , 211 peebles , p.j.e .
1969 , apj , 155 , 393 shandarin , s.f . 1974 , sov .
18 , 392 silk , j. , efstathiou , g. a. 1983 , the formation of galaxies , fundamentals of cosm .
phys . 9 , 1 sunyaev , a. r. , zeldovich , ya . b. , 1972 a&a , 20 , 189 tammann , g. a. , sandage , a. 1976 apj , 207 , l1 tully , r. b. 1987 , apj , 321 , 280 tully , r. b. 1988 , nearby galaxy catalog , cambridge wesson , p. s. 1979 , a&a , 80 , 269 wesson , p. s. 1982 , vistas astron . , 26 , 225 wesson , p. s. 1983 , a&a , 119 , 313 zeldovich , b. ya .
1970 , a&a , 5 , 84 | we investigated the sample of galaxies belonging to the tully groups of galaxies .
we analyzed the orientation of galaxies inside the group .
we did not found significant deviation from isotropy both in orientation of position angles and angles @xmath0 and @xmath1 giving the spatial orientation of galaxy planes .
moreover we analyzed consequences of different approximation of `` true shape '' of galaxies and showed possible influence of this problem for investigation of spatial orientation of galaxies .
implications of the obtained results for theory of galaxy formation was discussed as well . |
much effort has been devoted to understand epidemics on networks , mainly because of the increasing threats from cybercrime and the expected outbreak of new fatal viruses in populations .
epidemics have been studied on undirected networks for a long time and many authors ( see @xcite ) addressed the existence of an epidemic threshold @xmath1 in the susceptible - infected - susceptible * * * * ( sis ) epidemic process * * * * @xcite .
we consider the sis epidemic process in an undirected network @xmath9 , characterized by a symmetric adjacency matrix @xmath10 consisting of elements @xmath11 that are either one or zero depending on whether node @xmath12 is connected to @xmath13 or not .
each node @xmath12 has two possible states at time @xmath14 : either @xmath15 , meaning that the node is healthy and susceptible , or @xmath16 , for an infected node .
initially , a certain percentage of nodes are randomly selected as infected .
the infection from an infected node to each of its healthy neighbors and the curing of an infected node are assumed to be independent poisson processes with rates @xmath17 and @xmath18 , respectively .
every node @xmath12 at time @xmath14 is either infected , with probability @xmath19 @xmath20 $ ] or healthy ( but susceptible ) with probability @xmath21 .
this is the general continuous - time description of the simplest type of a sis epidemic process on a network .
the epidemic threshold @xmath1 separates two different phases of the epidemic process on a network : if the effective infection rate @xmath22 is above the threshold , the infection spreads and becomes persistent ; if @xmath23 , the infection dies out exponentially fast . the epidemic threshold @xmath24 of the n - intertwined mean - field approximation ( nimfa ) @xcite of the sis model lower bounds the real threshold , where @xmath3 is the largest eigenvalue of the adjacency matrix @xmath10 , also called the spectral radius .
topologies of undirected networks have been mostly modeled by erds and rnyi nodes by randomly assigning a link with probability @xmath25 to each pair of nodes . ]
@xcite as binomial networks , by brabasi and albert nodes . at every time step ,
we add a new node with @xmath26 links that connect the new node to @xmath26 different nodes already present in the graph .
the probability that a new node will be connected to node @xmath12 in step @xmath14 is proportional to the degree @xmath27 of that node .
this is referred to as preferential attachment . ]
@xcite as power - law networks , or by watts and strogatz nodes and @xmath28 edges per node , by rewiring each link at random with probability @xmath25 . ]
@xcite as * * * * small - world networks . more complicated static and dynamic models , such as the configuration model @xcite ,
are also proposed to approximate real - world networks .
however , many real - world networks are _ directed _ networks , some examples are shown in table [ realworld_table1 ] .
the dataset of the real - world networks is obtained from @xcite and the description of these networks is attached in the appendix a. @xmath29{cccc}\hline\hline \text{real - world networks } & n & l_{arcs } & \xi\\\hline \text{enron } & 69\,,244 & 276\,,143 & 84.29\%\\ \text{ljournal-2008 } & 5\,,363,\,260 & 79\,,023\,,142 & 25.32\%\\ \text{twitter-2010 } & 41,\,652,\,230 & 1\,,468\,,365,\,182 & 64.29\%\\ \text{wordassociation-2011 } & 10\,,617 & 72,\,172 & 76.77\%\\ \text{cnr-2000 } & 325,557 & 3,216,152 & 70.33\%\\ \text{in-2004 } & 1,382,908 & 16,917,053 & 60.68\%\\ \text{eu-2005 } & 862,664 & 1,935,140 & 67.80\%\\ \text{uk-2007 - 05@100000 } & 100,000 & 3,050,615 & 82.23\%\\ \text{uk-2007 - 05@1000000 } & 1,000,000 & 41,247,159 & 79.71\%\\\hline\hline \end{array } $ ] two kinds of links , namely bidirectional links and unidirectional links , exist in directed networks . if node @xmath12 is connected to node @xmath13 ( denoted by @xmath30 ) then @xmath13 is also linked to @xmath12 ( denoted by @xmath31 ) , one bidirectional link exists between nodes @xmath12 and @xmath13 ; and if either @xmath30 or @xmath32 exists , but not both in between the node pair @xmath12 and @xmath13 , a unidirectional link exists . here , we define the directionality as @xmath33 , where the number of arcs ( the number of @xmath34-elemets in the adjacency matrix ) @xmath35 , ( @xmath36 is the all - one vector ) , can also be calculated by @xmath37 .
a directed network with directionality @xmath0 is denoted by @xmath38 .
the network @xmath39 is a bidirectional network or an undirected network , whose adjacency matrix is symmetric , when @xmath40 .
the network @xmath41 is a directed network without any bidirectional link , when @xmath42 .
* * * * a high directionality is observed in twitter , as shown in table * * * * [ realworld_table1 ] .
a link runs from user a to user b if user a follows user b in twitter , where user a is called the `` follower '' of user b. the fact that user a `` follows ''
user b , does not necessarily mean that the reverse is also true .
for example , a famous person could have millions of followers but he / she may not follow many others .
this explains the high directionality @xmath0 of twitter .
the virtual - community social networks , such as livejournal , have a low directionality ( see table [ realworld_table1 ] ) , mainly because they aim to construct virtual connections in between real - life friends , and friendship relations are usually mutual .
there has been an increasing interest in the study of directed networks .
topological properties of directed networks , such as the short loops , closure connectivity , degree , domination and communities on realistic directed networks have already been studied in @xcite .
garlaschelli and loffredo @xcite investigated the reciprocity @xcite in directed networks , where the reciprocity is equal to @xmath43 .
processes taking place on networks , such as synchronization , percolation and epidemic spread , have also been researched @xcite in real directed networks .
percolation theory for directed networks with @xmath44 was firstly developed by newman et al .
then , bogu and serrano @xcite pointed out that even a small fraction of bidirectional links suffices to percolate the network .
moreover , meyers et al .
@xcite used a generating function method to predict the epidemic threshold in directed networks with @xmath45 and the size of the infected cluster .
recently , van mieghem and van de bovenkamp have proven that the nimfa epidemic threshold @xmath24 of the sis epidemic process also holds for directed networks @xcite .
stimulated by the directed social networks with different directionalities , here , we focus on the influence of the directionality @xmath0 on the epidemic threshold @xmath24 and other spectral properties .
this paper is organized as follows . in section [ algorithms ] , we propose two algorithms that could be applied to a bidirectional network to generate a directed network with an arbitrary given directionality @xmath0 , by rewiring or resetting links .
the in- and out- degree distribution of the generated directed network is the same as the degree distribution of the original bidirectional network .
chen and olvera - cravioto @xcite proposed an algorithm to generate a directed network with a given in- and out - degree distribution , which is similar to the configuration model .
however , their algorithm in @xcite can not generate a directed network with a given directionality @xmath0 . in section [ sec_spectral ]
, we investigate the effect of the directionality @xmath0 on the spectral radius @xmath3 , the principal eigenvector @xmath4 ( the eigenvector corresponding to @xmath3 ) , the spectral gap @xmath46 and the algebraic connectivity @xmath6 in both directed binomial .
a general binomial network could possibly have an assortativity @xmath8 within a large range . ] and power - law networks , whose in - degree and out - degree both follow a binomial ( or power - law ) distribution .
interestingly , we find that the spectral radius @xmath3 of networks @xmath39 is larger than that of directed networks @xmath41 when the degree distribution and the assortativity of these networks are the same .
this means that the epidemic threshold @xmath1 in undirected networks is smaller than that in directed networks with the same degree distribution and assortativity .
furthermore , we explore the influence of the pearson degree correlation coefficient @xmath8 ( also called the assortativity ) on the epidemic threshold @xmath47 in both directed binomial and directed power - law networks with different @xmath0 , in sec .
[ section_infect_assortativity ] .
the @xmath8 is the pearson correlation coefficient of degrees @xcite at either ends of a link and lies in the range @xmath48 $ ]
. actually , there are four degree correlations , namely the in - degree and in - degree correlation , the in - degree and out - degree correlation , the out - degree and in - degree correlation and the out - degree and out - degree correlation , in directed networks .
we consider directed networks where the in - degree and out - degree of each node are the same . in this case
, the four degree correlations are equal to each other and can be all referred as the degree correlation ( or the assortativity ) .
the decrease of the spectral radius @xmath3 with @xmath0 is large when the assortativity @xmath8 is large in directed binomial networks , whereas the opposite is observed in directed power - law networks .
the networks mentioned in this paper are simple , without self - loops and multiple links from any node to any other node . here , we propose two algorithms , in - degree and out - degree preserving rewiring algorithm ( iopra ) and link resetting algorithm ( lra ) , which both can be applied to any network to generate a directed network with a given directionality @xmath0 . in this study
, we only apply these two algorithms to generate directed networks with the same in- and out- degree distribution .
the difference is that iopra preserves the in- and out- degree of each node , while , lra may change the in- and out - degree of any node .
iopra is inspired by the degree preserving rewiring , which has been presented in @xcite .
we firstly introduce the degree - preserving rewiring , which monotonously increases or decreases the assortativity @xmath8 , while maintaining the node degrees unchanged , in undirected networks .
afterwards , we describe iopra and lra in detail . the degree - preserving rewiring @xcite can either increase or decrease the assortativity of a bidirectional network : ( a ) the degree - preserving assortative rewiring : randomly select two links associated with four nodes and then rewire the two links such that the two nodes with the highest degree and the two lowest - degree nodes are connected * * , * * respectively .
if any of the newly rewired links exists before rewiring , discard this step and a new pair of links is randomly selected ; ( b ) the degree - preserving disassortative rewiring : randomly select two links associated with four nodes and then rewire the two links such that the highest - degree node and the lowest - degree node are connected , and the remaining two nodes are also connected , as long as the newly rewired links do not exist before rewiring . either rewiring step ( a ) or ( b ) can be repeated to monotonically increase or decrease the assortativity in a bidirectional network .
iopra can be applied to change the directionality of networks .
we define our in - degree and out - degree preserving rewiring algorithm ( iopra ) as follows : randomly choose two unidirectional links with four end nodes , and rewire the two unidirectional links . in iopra ,
the head of one unidirectional link only can rewire with the head of the other unidirectional link , in order to maintain both the in - degree and out - degree of the four nodes unchanged ( see figure [ picture1 ] ) .
we do nt rewire if such rewiring can introduce duplicated links from any node to any other .
we discard the rewiring step if this rewiring step does nt change the directionality @xmath0 towards the given directionality . in both cases ,
we randomly reelect a pair of unidirectional links associated four nodes .
we illustrate the process of iopra changing the directionality in algorithm [ iopraalgo ] ( see appendix c ) .
iopra actually changes the directionality @xmath0 of a given network @xmath49 without changing the in- and out - degree of each node .
if the original network @xmath49 is an undirected network , the in - degree sequence is exactly the same as the out - degree sequence in the directed network @xmath38 generated by iopra .
picture1.eps iopra changes the directionality , as well as randomizing the connections of the original network , without changing the degree of any node . hence ,
if the initial network is a random network , e.g. an erds - rnyi ( er ) network or a scale - free network generated by the configuration model , where the connections are originally laid at random , iopra changes only the directionality @xmath0 .
however , if we apply iopra to a nonrandom network , e.g. a lattice , the resulting network has not only a different directionality but also a more randomized structure .
we start with a bidirectional network @xmath49 , and use the link resetting algorithm ( lra ) to change the directionality ( see algorithm [ lraalgo ] in appendix c ) .
we randomly choose a fraction @xmath0 of the bidirectional link pairs from @xmath49 .
then , we randomly choose only one unidirectional link from each bidirectional link , and randomly relocate the selected unidirectional links to a place without any link .
in this work , we only apply lra to er networks . in this case , the in- and out - degree of the generated directed network follow the same binomial degree distribution as the original network .
however , the in - degree and out - degree of any node in the generated network may differ from those in the original network @xmath49 . when lra is applied to other types of networks , such as the power - law networks , the original in- and out - degree distributions
are destroyed , and tend to be binomial . in summary ,
two types of directed binomial networks can be generated : one is generated by iopra ( called the iopra directed binomial networks ) , whose nodes have the same in - degree and out - degree ; the other , created by lra ( called the lra directed binomial networks ) , has the same in- and out- degree distribution , while allowing the in- and out- degree of any node to be different .
the adjacency matrix of a directed network is an asymmetric matrix , whose spectral radius @xmath3 is still real by the perron - frobenius theorem ( see @xcite ) .
we generate directed networks , with the directionality @xmath0 ranging from @xmath50 to @xmath34 , by applying iopra to the * * * * er ( @xmath51 , @xmath52ln@xmath53 ) and the ba ( @xmath51 , @xmath54 ) networks gradually .
an er network is connected , if @xmath55 ln@xmath53 for large @xmath56 , where @xmath57 is the disconnectivity threshold . in this work ,
we choose @xmath58 to be sure that the original networks are connected .
the influence of the directionality @xmath0 on the spectral radius @xmath3 and the assortativity @xmath8 is studied in both directed power - law networks and directed binomial networks ( see figure [ spectral_figure1 ] ) .
@xmath29{cc}\raisebox{-0.8752in}{\parbox[b]{3.0511in}{\begin{center } \includegraphics [ height=1.6457 in , width=3.0511 in ] { er1.eps}\\ ( a ) \end{center } } } & \raisebox{-0.8648in}{\parbox[b]{3.2024in}{\begin{center } \includegraphics [ height=1.6362 in , width=3.2024 in ] { ba1.eps}\\ ( b ) \end{center}}}\\{\parbox[b]{3.0511in}{\begin{center } \includegraphics [ height=1.5843 in , width=3.0511 in ] { er1ass.eps}\\ ( c ) \end{center } } } & \raisebox{-0.0311in}{\parbox[b]{3.2197in}{\begin{center } \includegraphics [ height=1.5843 in , width=3.2197 in ] { ba1ass.eps}\\ ( d ) \end{center}}}\end{array } $ ] apart from some wobbles , the spectral radius @xmath3 decreases almost linearly with the directionality @xmath0 .
the same phenomenon can also be observed in large , sparse directed networks ( see appendix d ) .
moreover , the assortativity @xmath8 of the network fluctuates slightly around @xmath50 .
we also have observed a similar phenomenon in large sparse networks .
we observe that the tiny leaps of spectral radius @xmath3 happen when the assortativity @xmath8 has a rise , which is understandable , because it has been shown in @xcite that the spectral radius @xmath3 increases with the increase of the assortativity @xmath8 .
figure [ picture2 ] exemplifies that the spectral radius @xmath3 may increase instead of decreasing when the directionality increases due to the assortativity @xmath8 .
we will study the effect of the assortativity @xmath8 on the decrease of the spectral radius @xmath3 with the directionality @xmath0 in * * * * section [ section_infect_assortativity ] .
picture3.eps with lra , we generate directed binomial networks with directionality @xmath0 from @xmath50 to @xmath34 with step @xmath59 .
the assortativity @xmath8 of all the directed binomial networks generated by lra is around @xmath50 .
hence , the effect of the assortativity @xmath8 * * * * can be ignored here .
the spectral radius @xmath3 is calculated in directed networks with different directionality @xmath0 .
we performed all the simulations for @xmath60 network realizations .
the spectral radius @xmath3 is plotted as a function of the directionality @xmath0 for directed binomial networks with @xmath61 and @xmath62 in figure [ lra ] .
methodlra.eps from the observation , the spectral radius @xmath3 is inversely proportional to the directionality @xmath0 with the factor @xmath63 @xmath64 , which is independent from the link density @xmath25 of the networks .
this observation can be explained by the following proposition .
let @xmath65 be a connected erds - rnyi ( er ) random graph with a finite @xmath56 , and let @xmath38 be a directed binomial network generated by lra whose in- and out - degree follow the same binomial distribution as * * * * @xmath66 .
the average spectral radius satisfies[proposition ] @xmath67\simeq e[\lambda_{1}(g_{p}(n))]-\xi \label{spectral_decrease}\ ] ] * arguments : * a directed binomial network @xmath38 generated by lra with link density @xmath25 , can be equivalently constructed by randomly adding @xmath68 unidirectional links to a bidirectional er network @xmath69 with size @xmath56 and link density @xmath70 . the average spectral radius ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
137 ) of @xmath69 is @xmath71=(n-2)p(1-\xi)+1+o(\frac{1}{\sqrt{n}})\text{.}\ ] ] the principal eigenvector of an adjacency matrix @xmath10 is denoted by @xmath4 obeying the normalization @xmath72 .
let @xmath73 denote the adjacency matrix of the resulting network after adding one unidirectional link to network @xmath49 .
the largest eigenvalue is increased due to the addition of the link ( @xmath30 ) ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* , lemma 7 ) as @xmath74 where the increase is strict if the adjacency matrix @xmath10 is irreducible .
hence , the average increase of the spectral radius by adding @xmath26 unidirectional links in random networks is obtained as@xmath75\simeq me[(x_{1})_{i}(x_{1})_{j}]\text{.}\ ] ] the sum of the product of components in the principal eigenvector of erds - rnyi networks is approximated by a function of link density @xmath25 ( see figure [ sumproduct ] ) .
the fitting function can be expressed as,@xmath76 = n-\frac{1}{p}+o(1)\text{,}\ ] ] when the network is connected .
since @xmath72 and since the expectation @xmath77 $ ] is a linear operator , we obtain @xmath78 = \frac{n-\frac{1}{p}-1}{n(n-1)}+o(\frac{1}{n^{2}})\text { , } \label{product}\ ] ] when @xmath79
. directed binomial networks generated by lra from er with @xmath56 and @xmath25 , have the same @xmath80 $ ] .
hence , the average spectral radius of the directed network obtained by adding @xmath81 unidirectional links to the network @xmath69 can be approximated by @xmath82\simeq e[\lambda_{1}(g_{p(1-\xi)}(n))]+2\left ( n(n-1)/2\right ) p\xi e\left [ ( x_{1})_{i}(x_{1})_{j}\right ] \text{.}\ ] ] using ( [ product ] ) , @xmath82\simeq(n-2)p+1-\xi+o(\frac{1}{\sqrt{n}})\text{,}\ ] ] which leads to ( 1 ) .
@xmath83 juhsz @xcite also pointed out that the largest eigenvalue @xmath84 of a directed random network with link density @xmath25 and size @xmath56 is almost surely @xmath85 , when @xmath56 is large . in er random networks * * , * * the spectral radius @xmath86\rightarrow np+1 $ ] , when @xmath56 is large ( see ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* , art . 137 ) ) .
both earlier results are consistent with proposition [ proposition ] , and support that the proportionality factor between the spectral radius @xmath3 and the directionality @xmath0 is around @xmath64 .
sumproduct.eps proposition [ proposition ] also reveals the effect of the size @xmath56 on the relative largest decrease of the spectral radius @xmath87 .
we predict that @xmath88 if @xmath89 for directed binomial networks , because the decrease of the spectral radius ( @xmath90 ) is almost a constant value , whereas the spectral radius @xmath91 of dense directed binomial networks increases with the size of the networks .
this implies that the effect of the directionality @xmath0 on the spectral radius is small in large dense binomial networks .
the principal eigenvector @xmath4 was first proposed as a centrality metric by bonacich @xcite in @xmath92 , to indicate the influence of each node .
for example , the decrease of the spectral radius @xcite by removing nodes , can be characterized by the corresponding principal eigenvector components . in this section
, we explore the principal eigenvector in directed networks .
the principal eigenvector of the directed networks , with the directionality @xmath0 from @xmath50 to @xmath34 with step @xmath59 , are calculated .
then , the components of the principal eigenvector are sorted in an ascending order . for each @xmath0
, we simulate @xmath60 network realizations and compute the average sorted principal eigenvector components .
figure [ eigenvector ] illustrates that the components of the principal eigenvector are more uniform in directed binomial networks : the principal eigenvector @xmath93 as @xmath94 ; moreover , the variance of components of the principal eigenvector linearly decreases with the directionality @xmath0 in both directed binomial networks and the directed power - law networks ( see figure [ spectral_figure3 ] ) .
the observation implies that when the directionality is larger , the influence of each node on the spectral radius is more similar .
this experimental evidence suggests that increasing the directionality enables all nodes to contribute more similarly to the robustness against epidemic in directed networks .
the decrease of the variance @xmath95 $ ] in directed binomial networks by lra is larger than that in directed binomial networks by iopra ( see figure [ spectral_figure3](a ) ) .
the connections in lra directed binomial networks are more random than that in iopra directed binomial networks , in the sense that lra allows each node to have a different in- and out- degree , although the in- and out- degree distribution are the same in both lra and iopra binomial networks . as a consequence ,
the principal eigenvector @xmath96 is more uniform with a smaller @xmath95 $ ] in lra directed binomial networks than that in iopra directed binomial networks when the directionality is the same .
thus , nodes in lra directed binomial networks have more equal contributions to the spectral radius than nodes in iopra directed binomial networks with the same directionality .
_ @xcite have shown that both a large variance of the degree and a large assortativity @xmath8 contribute to a large variance @xmath95 $ ] of the components of the principal eigenvector @xmath4 . here , we point out further that a large directionality @xmath0 leads to a small variance @xmath95 $ ] of the components of @xmath4 .
principaleigenvector.eps @xmath29{cc}\raisebox{0.0104in}{\parbox[b]{3.1618in}{\begin{center } \includegraphics [ height=1.7045 in , width=3.1618 in ] { varer.eps}\\ ( a ) \end{center } } } & \raisebox{-0.0104in}{\parbox[b]{3.3114in}{\begin{center } \includegraphics [ height=1.7193 in , width=3.3114 in ] { varba10times.eps}\\ ( b ) \end{center } } } \end{array } $ ] the difference @xmath97 between the largest eigenvalue @xmath3 and the second largest eigenvalue @xmath98 is called the spectral gap .
all eigenvalues of the symmetric adjacency matrix of an undirected network are real . here
we focus on the directed networks , whose adjacency matrix is asymmetric .
the eigenvalues of directed networks can be complex numbers ( as exemplified in figure [ eigenvalues ] in appendix b ) . in directed networks ,
the spectral gap is defined as @xmath99 , where @xmath100 is the modulus of @xmath98 .
the spectral gap @xmath101 increases with the directionality @xmath0 in both the directed binomial networks and the directed power - law networks ( see figure [ spectral_figure4 ] ) . as introduced in sec .
[ sec_spectral_radius ] , the spectral radius decreases with the directionality . * * * * our observation implies that the second largest eigenvalue @xmath102 decreases with the directionality faster than the spectral radius .
the larger the spectral gap is , the faster a random walk converges to its steady - state @xcite .
thus , the dynamic process in a directed network reaches the steady - state faster than that in an undirected network with the same degree distribution .
figure [ spectral_figure4 ] ( a ) implies that a dynamic process is slightly faster to reach the steady - state in iopra directed binomial networks than in lra directed binomial networks . the existence of large spectral gap together with a uniform degree distribution results in higher structural sturdiness and robustness against node and link failures @xcite . hence , directed networks with high directionality @xmath7 and a uniform degree distribution are more robust than undirected networks with large variance of degree .
@xmath29{cc}{\parbox[b]{3.1868in}{\begin{center } \includegraphics [ height=1.8031 in , width=3.1868 in ] { spectralgap1000.eps}\\ ( a ) \end{center } } } & { \parbox[b]{3.2379in}{\begin{center } \includegraphics [ height=1.8031 in , width=3.2379 in ] { spectralgapba1000.eps}\\ ( b ) \end{center } } } \end{array } $ ] the laplacian matrix @xcite is defined as @xmath103 , where the incidence matrix @xmath104 is an @xmath105 matrix with elements @xcite@xmath106{c}1\text { \ if link } e_{l}=i\rightarrow j\\ -1\text { if link } e_{l}=j\rightarrow i\\ 0\text { \ \ \ \ \ \ \ \ \ otherwise.}\end{array } \right.\ ] ] the laplacian matrix can be equivalently expressed as @xmath107 , where @xmath108 , @xmath109 and @xmath110 are diagonal matrices which contain the in - degree and out - degree of each node respectively , and @xmath111 . if the network is an undirected network , @xmath112 is the adjacency matrix @xmath10 and @xmath113 @xmath114 is the degree matrix .
the second smallest eigenvalue @xmath6 of the laplacian @xmath115 was named algebraic connectivity by fiedler @xcite .
the laplacian @xmath115 is always symmetric as defined .
hence , the algebraic connectivity of a directed and connected network is a positive real number .
the algebraic connectivity , together with the spectral gap , quantifies the robustness and the network s well - connectedness .
the larger the algebraic connectivity is , the more difficult it is to cut the network into disconnected parts . here , we study the influence of the directionality @xmath0 on the algebraic connectivity @xmath6 of directed networks . as illustrated in figure
[ spectral_figure5 ] , the algebraic connectivity increases with the directionality @xmath0 in both the directed binomial networks and the directed power - law networks .
this suggests that the directed networks with high directionality are more difficult to break into parts and synchronize faster . as the directionality increases , the number of none - zero elements of @xmath112 increases and the variance of the elements of @xmath112 decreases
this could be one possible reason why the network is better connected .
moreover , the algebraic connectivity @xmath6 is greater in the lra directed binomial networks than in the iopra binomial networks ( see figure [ spectral_figure5](a ) ) .
the algebraic connectivity @xmath6 approaches the spectral gap @xmath116 , as the network tends to be regular bidirectional networks @xcite , which suggests the spectral gap is related to the algebraic connectivity .
figures [ spectral_figure4 ] and [ spectral_figure5 ] show that both , the spectral gap and the algebraic connectivity , increase with the directionality @xmath0 in directed networks , which is consistent with the relation between the algebraic connectivity and the spectral gap .
@xmath29{cc}\raisebox{0.0104in}{\parbox[b]{3.1557in}{\begin{center } \includegraphics [ height=1.8239 in , width=3.1557 in ] { alger.eps}\\ ( a ) \end{center } } } & { \parbox[b]{3.3425in}{\begin{center } \includegraphics [ height=1.8222 in , width=3.3425 in ] { algba.eps}\\ ( b ) \end{center } } } \end{array } $ ]
in the directed networks generated by applying iopra to er or ba networks , the in- and in- degree correlation , the in- and out- degree correlation , the out- and in- degree correlation and the out- and out- degree correlation are the same .
thus , the four correlations are all referred as the degree correlation ( or the assortativity ) . in section [ sec_spectral ] , we have discussed how the spectral properties change with the directionality in directed networks , where the assortativity is always close to zero . here ,
we study how the spectral radius @xmath3 changes with the directionality @xmath0 when the assortativity @xmath8 is the same , and how the change of the spectral radius @xmath3 with @xmath117is influenced by the assortativity in directed networks .
two approaches are applied to investigate this problem .
@xmath29{cc}{\parbox[b]{3.1384in}{\begin{center } \includegraphics [ height=1.7608 in , width=3.1384 in ] { effectofdirection.eps}\\ ( a ) one network realization \end{center } } } & \raisebox{0.0104in}{\parbox[b]{3.3425in}{\begin{center } \includegraphics [ height=1.74 in , width=3.3425 in ] { exactass100times.eps}\\ ( b ) $ 10 ^ 2 $ network realizaitons \end{center } } } \end{array } $ ] @xmath118 * approach * @xmath34 * * : * * first , we perform degree - preserving rewiring on er networks ( or ba networks ) to obtain a set of bidirectional networks with assortativity @xmath8 from @xmath119 to @xmath120 ( or @xmath121 to @xmath122 ) with step @xmath59 .
second , we alter the directionality @xmath0 of all bidirectional networks with each assortativity using iopra .
the directionality @xmath0 is changed from @xmath50 to @xmath34 with step @xmath59 .
iopra randomizes network connections , and thus pushes the assortativity of the resulting directed network towards zero , if the original network has a non - zero assortativity .
figure [ binomial_1 ] plots the simulation results of one binomial network realization and @xmath123 binomial network realizations .
the simulation of one realization is almost the same as the result of a large number of network realizations , which points to _ almost sure behavior _
the results in the directed power - law networks are shown in figure [ spectraldecvsassba ] .
effectofdirectionba.eps @xmath118 * approach * @xmath124 * * : * * first , we generate er networks ( or ba networks ) @xmath39 whose directionality @xmath40 .
second , we apply iopra to er ( or ba ) networks @xmath39 to generate directed binomial networks ( or directed power - law networks ) @xmath41 with directionality @xmath42 .
then , we change the assortativity of er networks ( or ba networks ) @xmath39 and directed binomial networks ( or directed power - law networks ) @xmath41 by degree preserving rewiring and in- and out- degree preserving rewiring , respectively , without changing the directionality .
note that the in- and out- degree preserving rewiring can be applied not only to change the directionality , but also to change the assortativity .
figure [ spectradecvsass ] plots the spectral radius of the networks @xmath39 and @xmath41 as a function of the assortativity . @xmath29{cc}{\parbox[b]{3.1695in}{\begin{center } \includegraphics [ height=1.7513 in , width=3.1695 in ] { binomial1000biggestdec1000.eps}\\ ( a ) \end{center } } } & { \parbox[b]{3.3217in}{\begin{center } \includegraphics [ height=1.7513 in , width=3.3217 in ] { powerlaw1000biggestdec1000.eps}\\ ( b ) \end{center } } } \end{array } $ ] the two approaches both change the assortativity and the directionality of the networks , while , the order of change is different : approach @xmath34 changes the assortativity firstly and then the directionality ; approach 2 is the opposite
. figures [ binomial_1 ] , [ spectraldecvsassba ] and [ spectradecvsass ] , show that the spectral radius @xmath3 always decreases with the directionality @xmath0 when the networks have the same degree distribution and the same assortativity @xmath8 .
moreover , the degree distribution of the network also influences the change range of the spectral radius @xmath3 with @xmath0 .
the decrement of the spectral radius @xmath3 with @xmath0 increases with the assortativity in directed binomial networks ( see figures [ binomial_1 ] and [ spectradecvsass ] ( a ) ) . on the contrary ,
the decrement of the spectral radius @xmath3 with @xmath0 goes down with the assortativity in directed power - law networks ( see figures [ spectraldecvsassba ] and [ spectradecvsass ] ( b ) ) . furthermore , the decrease of the spectral radius in directed power - law networks is larger than that in directed binomial networks , when the assortativity is zero .
many real - world networks are directed power - law networks , where @xmath3 could possibly be tuned within a large range by controlling the directionality in real - world networks .
summarizing , the spectral radius @xmath3 decreases with the directionality @xmath0 when the assortativity remains constant . in order to protect the network from virus spreading via increasing the epidemic threshold
, while maintaining the degree distribution and the assortativity , increasing the directionality of networks is recommended .
meanwhile , the spectral gap and the algebraic connectivity are also increased , which means that the topological robustness is also enhanced in return .
in this work , two algorithms to generate directed networks with a given directionality @xmath0 are proposed .
this allows us to study the influence of the directionality @xmath0 on the spectral properties of networks .
the spectral radius @xmath3 , which is the inverse of the sis nimfa epidemic threshold @xmath47 , is studied in directed networks .
a universal observation is that , the spectral radius decreases with the directionality when the degree distribution and the assortativity of the network is preserved .
we may , thus , increase the epidemic threshold to suppress the virus spread via increasing the directionality of the network .
the possible range to increase the epidemic threshold is relatively large in directed binomial networks with a high assortativity and directed power - law networks with a low assortativity .
the variance of the components of the principal eigenvector decreases with the directionality , which indicates that the influence of each node on the spectral radius is similar in networks with a high directionality .
moreover , the spectral gap and the algebraic connectivity increase with the directionality , implying that an increase of the directionality enhances the connectivity of the network .
furthermore , we observe that the spectral gap increases faster with the directionality in iopra than in lra directed binomial networks , on the contrary , the algebraic connectivity increases with the directionality faster in lra than in iopra directed binomial networks
. this observation may be due to the fact that the in- and out- degree of each node could be different in lra directed binomial networks , while , are exactly the same in iopra directed binomial networks .
the influence of the difference between in- and out- degree of nodes on spectral properties for directed power - law networks remains an open question .
j. o. kephart and s. r. white , _ directed - graph epidemiological models of computer viruses _ , proceedings of the 1991 ieee computer society symposium on research in security and privacy , ieee computer society , los alamitos , ca , pages 343 - 359 , 1991 .
y. wang , d. chakrabarti , c. wang and c. faloutsos , epidemic spreading in real networks : an eigenvalue viewpoint , in reliable distributed systems , proceedings , 22nd international symposium on ( pp .
25 - 34 ) , ieee computer society press , los alamitos , ca , 2003 . p. boldi , m. rosa , m. santini and s. vigna , _ layered label propagation : a multiresolution coordinate - free ordering for compressing social networks _ , proceedings of the 20th international conference on world wide web , acm press , 2011
. p. van mieghem , h. wang , x. ge , s. tang and f.a .
kuipers , _ influence of assortativity and degree - preserving rewiring on the spectra of networks _ , the european physical journal b * 76*(4 ) : 643 - 652 , 2010 .
p. van mieghem , g. hooghiemstra and r. w. van der hofstad , _ stochastic model for the number of traversed routers in internet _ , proceedings of passive and active measurement ( pam ) , ripe ncc , amsterdam , the netherlands , 2001 .
this data set was made public by the federal energy regulatory commission during its investigations : it is a partially anonymised corpus of e - mail messages exchanged by some enron employees ( mostly part of the senior management ) . this data set is a directed graph , whose nodes represent people and with an arc from x to y whenever y was the recipient of ( at least ) a message sent by x. livejournal is a virtual - community social site started in 1999 : nodes are users and there is an arc from x to y if x registered y among his friends .
it is not necessary to ask y permission , so the graph is directed .
this graph is the snapshot used by chierichetti , flavio , et al . in `` on compressing social networks .
'' proceedings of the 15th acm sigkdd international conference on knowledge discovery and data mining .
acm , 2009 , and was kindly provided by the authors .
twitter is a website , owned and operated by twitter inc .
, which offers a social networking and microblogging service , enabling its users to send and read messages called tweets .
tweets are text - based posts of up to 140 characters displayed on the user s profile page .
this is a crawl presented by kwak , haewoon , et al . in `` what is twitter , a social network or a news media ? '' , proceedings of the 19th international conference on world wide web .
acm , 2010 .
nodes are users and there is an arc from x to y if y is a follower of x. in other words , arcs follow the direction of tweet transmission . the free word association norms network is a directed graph describing the results of an experiment of free word association performed by more than 6000 participants in the united states : its nodes correspond to words and arcs represent a cue - target pair ( the arc x-@xmath125y means that the word y was output by some of the participants based on the stimulus x ) . the networks , `` cnr-2000 '' , `` in-2004 '' , `` eu-2005 '' , `` uk-2007 - 05@100000 '' and `` uk-2007 - 05@1000000 '' are small www networks that were crawled from the internet .
the `` cnr-2000 '' is crawled from the italian cnr domain .
a small crawl of the .in domain performed for the nagaoka university of technology is in data `` in-2004 '' .
the `` eu-2005 '' is a small crawl of the .eu domain .
this network `` uk-2007 - 05@100000 '' and `` uk-2007 - 05@1000000 '' have been artificially generated by combining twelve monthly snapshot of the .uk domain and collected for the delis project .
the spectral radius @xmath3 and the spectral gap @xmath97 are considered as important metrics for the percolation processes on networks . here
we also present all eigenvalues in directed networks in a image - real figure .
the eigenvalues are calculated on @xmath60 simulation realizations .
the changes of the eigenvalues @xmath126 with the directionality @xmath0 from @xmath50 to @xmath34 with step @xmath59 in directed binomial networks ( @xmath127 , @xmath128 ) are shown on figure [ eigenvalues ] .
surprisingly , the real part of all eigenvalues tends to @xmath50 , when the directionality @xmath0 increases . | epidemics have so far been mostly studied in undirected networks . however , many real - world networks , such as the online social network twitter and the world - wide web , on which information , emotion or malware spreads , are directed networks , composed of both unidirectional links and bidirectional links . we define the _ directionality _ @xmath0 as the percentage of unidirectional links .
the epidemic threshold @xmath1 for the susceptible - infected - susceptible ( sis ) epidemic is lower bounded by @xmath2 in directed networks , where @xmath3 , also called the spectral radius , is the largest eigenvalue of the adjacency matrix . in this work
, we propose two algorithms to generate directed networks with a given * * directionality @xmath0 .
the effect of @xmath0 on the spectral radius @xmath3 , principal eigenvector @xmath4 , spectral gap @xmath5 and algebraic connectivity @xmath6 is studied .
important findings are that the spectral radius @xmath3 decreases with the directionality @xmath0 , whereas the spectral gap and the algebraic connectivity increase with the directionality @xmath7 .
the extent of the decrease of the spectral radius depends on both the degree distribution and the degree - degree correlation @xmath8 .
hence , * * * * in directed networks , the epidemic threshold is larger and a random walk converges to its steady - state faster than that in undirected networks with the same degree distribution . |
a lot of effort has justifiably been spent on the central detectors of the lhc experiments so that hard interactions can be triggered on , and observed , in order to expose new physics .
however the vast majority of interactions are soft with particles predominantly going forward , and the physics of this domain is one of the subjects of the workshop .
the relevant processes , forward elastic scattering and soft diffraction , are driven by pomeron ( or rather by multi - pomeron ) exchange .
they are interesting in their own right , although in the period after the advent of qcd they did not attract so much attention .
however in the last few years the situation has changed .
a stimulus came from the observation of diffractive processes at hera and the tevatron , characterised by the presence of rapidity gaps .
moreover , diffractive processes have been proposed as possible ways to identify new physics .
for example , the central production of a higgs boson with a rapidity gap on either side is advocated as a possible discovery channel at the lhc . the chance that these gaps survive the soft rescattering of the colliding hadrons was one of the topics of discussion .
the second subject of the workshop was the way to measure the luminosity of the lhc .
three methods were discussed : * the classic method based on the optical theorem , [ eq : a1 ] .
|_t = 0 = ( 1 + ^2 ) , where @xmath5 is the ratio of the real to the imaginary part of the forward elastic amplitude ( coulomb effects have been neglected in ( [ eq : a1 ] ) ) , * to measure pure qed @xmath6 or @xmath7 production via photon - photon fusion [ eq : a2 ] pp p + ^+ ^- + p , * to measure inclusive @xmath2 or @xmath3 production . at first sight
it might appear that an accurate measurement of the luminosity is not essential .
however , for example , precision measurements in the higgs sector of accuracy of about 7% require the uncertainty in the measurement of the luminosity to be 5% or less @xcite . at this workshop ,
tapprogge @xcite summarised the physics reasons why a precise measurement of the luminosity of the lhc is important .
bozzo @xcite presented the totem experimental programme , which will take place in the very first runs at the lhc .
the plan is to measure @xmath9 by a luminosity - independent method [ eq : a3 ] _ tot = , with an absolute error of about 1 mb .
the totem inelastic detector will be installed inside the cms experiment , with the elastic scattering roman pot detectors located at distances in the interval 100 to 200 m from the crossing point .
the totem measurements need special runs at the low initial luminosity , with high @xmath10 optics for an accurate measurement of the small scattering angles
. the detector should be efficient to within 2 mm of the beam and can reach down to about @xmath11 .
they will collect about 100 events / sec for @xmath12 at a luminosity @xmath13 .
the differential cross section will also be measured in the interval @xmath14 in the high luminosity runs .
totem also plans an inclusive trigger for the measurement of single diffraction .
we also heard at the workshop about the novel microstation concept for forward measurements @xcite .
the microstation is a light , compact device which could be integrated with the beam pipe .
it could be used in situations where there are severe space and mass limitations for the inelastic detector , such as in the atlas experiment .
the elastic proton measurement would be like that for totem , so again it should be possible to reach down to @xmath15 .
piotrzkowski @xcite considered the possibility of using the lhc as a @xmath16 collider , and noted how the present totem layout may be modified to allow the tagging of protons with small energy losses . in principle , this would allow the study of @xmath17 etc . in a broad region of @xmath16
centre - of - mass energy about 200 gev , with an effective @xmath16 luminosity that is reduced by about @xmath18@xmath19 of that of the lhc .
guryn @xcite presented the pp2pp experimental physics programme at rhic .
their main goal is to make a detailed study of the spin dependence of the proton - proton interaction in the kinematic range [ eq : a4 ] 50 < < 500 gev 4 10 ^ -4 < |t| < 1.5 gev^2 , in order to probe features of the pomeron . at the beginning
they will measure @xmath20 down to @xmath21 , and study the @xmath22 dependence of the slope .
later they will explore the coulomb - nuclear interference region , and determine @xmath5 . by measuring the spin asymmetries
@xmath23 they will obtain information on the five independent @xmath24 helicity amplitudes .
block @xcite discussed the determination of @xmath25 at @xmath26 tev from cosmic ray data .
he emphasized the importance of the relation between the slope @xmath27 of the elastic scattering distribution and @xmath9 for energies well above the accelerator regime , and the reliance on a model of proton - air interactions .
vorobyov @xcite reminded us how at isr energies , and below , it was possible to make precision measurements of the elastic differential cross section at small momentum transfer down into the coulomb - nuclear interference region . from these experiments one could make a determination of the local slope [ eq : a5 ] b ( t ) = as a function of @xmath22 , measure the ratio @xmath5 of the real to imaginary part of the forward amplitude and reliably determine the total cross section from the optical point . however
the application of this method becomes more and more difficult as we go up to tevatron and then lhc energies , due to the decrease of the typical scattering angle and the inability to experimentally reach the coulomb - nuclear interference region .
figs . 1 , 2 and 3 show data at isr , cern @xmath30 and tevatron energies respectively .
in the first two plots we see the coulomb interference spike at very small @xmath22 , whereas at the tevatron the spike is no longer accessible and the data of the two experiments extrapolate to different values at @xmath29 .
moreover at the isr we see , from fig . 1 , a change of the local slope with @xmath22 , [ eq : a6 ] b ( 0 ) - b ( | t | = 0.2 gev^2 ) 2 gev^-2 , whereas at the tevatron the data suggest less variation for @xmath31 , see fig . 3 .
a global description of these and other forward data gives the @xmath22 dependence of the local slopes shown in fig .
the dashed curves should be ignored as they come from a description which does not include high - mass diffraction .
the remaining two curves ( continuous and dotted ) are obtained using two extremum models for high - mass diffraction , and give a measure of the level of the theoretical ambiguity in @xmath32 .
the predictions at the lhc energy mean that we should be able to extrapolate the totem measurements to @xmath29 with an error of less than 0.5% coming from the variation of the local slope .
even if the cross section were measured only in the interval @xmath33 , then the uncertainty at @xmath29 would be less than 3% due to the variation of @xmath32 .
high energy @xmath8 scattering at small momentum transfer @xmath22 is driven by pomeron @xmath22-channel exchange .
in fact we were reminded at the workshop @xcite how in 1960 , gribov startled the community by showing that the behaviour of the @xmath8 amplitude [ eq : a7 ] a ( s , t ) = i s f ( t ) , which gives the asymptotic behaviour crossing invariance is an important ingredient in the second equality in ( [ eq : a8 ] ) , the pomeranchuk theorem . ]
[ eq : a8 ] _ tot = , ( pp ) = ( p|p ) , contradicts @xmath22-channel unitarity .
gribov noted that a possible solution was to introduce the regge behaviour [ eq : a9 ] a ( s , t ) = i s^_(t ) f ( t ) , with a pomeron trajectory @xmath34 , with intercept @xmath35 and slope @xmath36 .
this is in agreement with @xmath22-channel unitarity , while still giving relations ( [ eq : a8 ] ) .
in addition to the pomeron pole , there are also regge cuts coming from multiple pomeron exchange , which led to the development of gribov s reggeon calculus , involving renormalisation of the bare pole , mandelstam crossed diagrams @xcite , agk cutting rules @xcite , etc .
this soft pomeron was the subject of the talks by landshoff @xcite and kaidalov @xcite .
landshoff @xcite showed that a good description of the available data for high energy @xmath9 and @xmath37 at small momentum transfer is given by a simple effective pomeron pole trajectory [ eq : a10 ] _ eff ( t ) = 1.08 + ^t with @xmath38 .
this is a remarkable simplification , but of course it is necessarily incomplete for the reasons that are mentioned below .
kaidalov @xcite explained that to describe phenomena in the forward region , it is important to extend single - pomeron exchange so as to explicitly include the multi - pomeron effects . at very high energy such effects are needed to restore @xmath39-channel unitarity , but , as we will see in a moment , they are also necessary at current energies .
5 shows examples of typical corrections to the bare pomeron pole of diagram ( a ) .
first , iterations of the pole amplitude via elastic unitarity gives contributions of the type shown in diagram ( b ) .
if we take into account the possibility of proton excitations @xmath40 in intermediate states , then we must include contributions such as that in diagram ( c ) .
these iterations are implemented in terms of a two - channel eikonal formalism @xcite .
note that diagrams ( b ) and ( c ) are very symbolic by implication they incorporate the mandelstam crossed diagrams @xcite and satisfy the agk cutting rules @xcite .
the second @xmath41 channel effectively allows for _ low mass _ diffractive dissociation .
the excitation into _ high mass _ @xmath42 states is described by the triple - pomeron graph ( d ) for single - diffractive dissociation ( with cross section @xmath43 ) , and by ( e ) for double - diffractive dissociation @xmath44 . to be self - consistent , the pomeron lines in ( d ) and ( e ) actually represent the final pomeron amplitude with all the screening effects included .
in addition to ( d ) , there is an equal contribution @xmath43 from dissociation of the lower proton only .
the contribution of graphs of the type ( b)(e ) are not negligible . indeed , using the agk rules
, it may be estimated that the correction to ( a ) is [ eq : a11 ] _
d/_tot ( _ el + 2 _ sd + _
dd)/_tot ~ 0.4 at the lhc .
a most convincing way to see the necessity of multi - pomeron rescattering effects at current energies is to look at the energy behaviour of @xmath43 .
if only the bare pomeron is used , as in diagram ( d ) , then the cross section grows as [ eq : a12 ] _ sd ~ s^2 _ ( t ) - 2 , whereas when rescattering effects are included [ eq : a13 ] _ sd ~ _ tot / s .
the difference is dramatic , as fig .
the physical origin of this result is that the high energy proton - proton interaction reaches the black - disc limit for central values of the impact parameter , and forces inelastic diffraction to come from the peripheral region with @xmath45 .
a related observation was made by schlein @xcite .
he showed that the single diffractive data in the isr - tevatron energy range can be described by the triple - regge formula , provided that the effective intercept @xmath46 decreased with increasing @xmath39 .
such an @xmath39 dependent behaviour of the effective intercept may be interpreted as a manifestation of the multi - pomeron exchange effects.the bare pomeron , together with the multi - pomeron corrections , should give a global description of all forward phenomena , including @xmath47 and @xmath48 .
several analyses have been undertaken at various levels of sophistication to all or parts of the data .
recent examples are given in @xcite .
perhaps the most complete study to date is that of ref .
@xcite , which is in the spirit of the much earlier pioneering work of kaidalov et al .
it describes the forward phenomena in high energy @xmath8 ( and @xmath49 ) collisions using a multi - pomeron approach which embodies : * _ pion - loop _ insertions in the bare pomeron pole , which represent the nearest singularity generated by @xmath22-channel unitarity , * a _ two - channel eikonal _ which incorporates the pomeron cuts generated by elastic and quasi - elastic ( with @xmath50 intermediate states ) @xmath39-channel unitarity .
* high mass single and double _ diffractive dissociation_. the resulting description of @xmath37 data is shown in figs .
13 and the corresponding @xmath22 dependence of the local slope @xmath32 is given in fig .
surprisingly , the bare pomeron pole parameters turn out to be [ eq:14 ] ( 0 ) - 1 0.10 , ^ = 0 , which is to be contrasted to those of the effective pole ( [ eq : a10 ] ) .
thus the shrinkage of the diffraction cone comes not from the bare pole , but rather has components from the three ingredients , ( i)(iii ) , of the model .
that is , in the isr - tevatron energy range @xcite[eq : a15 ] _
( 0.034 + 0.15 + 0.066 ) gev^-2 from the @xmath51-loop , @xmath39-channel eikonalization and diffractive dissociation respectively .
we saw @xcite that at lower energies the fixed target data require @xmath52 , which is consistent with ( [ eq : a15 ] ) since as the energy decreases the effect of the eikonal and higher mass diffractive dissociation reduces .
moreover eikonal rescattering suppresses the growth of the cross section with @xmath53 , so to describe the same data we require @xmath54 .
another observable difference between the naive effective pole and the multi - pomeron approach is in the energy behaviour of the slope @xmath55 .
this is illustrated in fig .
7 , together with the behaviour of @xmath9 .
it is seen that , while the two approaches give similar values of @xmath9 up to lhc energies , the predictions for the slope @xmath55 at the lhc already differ significantly . in table 1
we show some predictions for the forward observables at the lhc .
note that the bh predictions @xcite are based on a single - channel eikonal , and do not therefore involve inelastic diffraction .
the kmr prediction is an update of @xcite .
.predictions for forward observables at the lhc . [ cols="^,^,^,^ " , ] an experimental manifestation of the survival probability @xmath56 comes from comparing hard diffraction at the tevatron with that observed at hera .
the relevant plot was presented by snow @xcite , see also @xcite .
it shows the recent cdf measurements of diffractive dijet production as a function of @xmath57 , together with the expectations based on convoluting the parton densities of the pomeron ( and those of the proton ) with the partonic - level cross sections of the hard subprocess .
@xmath57 is the momentum fraction of the pomeron entering the hard subprocess .
the parton densities of the pomeron are determined from diffractive deep - inelastic - scattering data at hera @xcite .
the partonic structure of the tevatron dijet production and hera diffractive dis are sketched in fig .
however , as the cdf plot @xcite shows , when the hera pomeron densities are used to estimate tevatron dijet production , the factorized prediction turns out to be about an order of magnitude larger than the data .
a key assumption of the factorization estimate is that the survival probability of the rapidity gap , associated with the pomeron exchange , is the same in figs .
8(a ) and 8(b ) .
comparison of the diagrams shows that the breakdown of factorization is an inevitable consequence of qcd , and occurs naturally , due to the small probability @xmath56 for the rapidity gap to survive the soft rescattering which occurs between the incoming hadrons at the tevatron , but which is absent between the electron and the proton at hera .
in fact , a detailed study of this comparison , including the @xmath57 dependence , tells us about the dependence of the survival probability on the kinematic variables @xcite , and opens the door to the application to many other hard processes with rapidity gaps , including those discussed by khoze @xcite .
lipatov gave a comprehensive survey of the pomeron before and after qcd @xcite .
since he was either close to , or pioneered , all these developments , it was a masterly and informative summary .
the pomeron before qcd was the subject of section 4 .
that pomeron was a reggeon , even if it has a rather unusual regge trajectory , which in the time - like ( @xmath58 ) region is mixed with the @xmath59 and @xmath60 meson trajectories @xcite . in qcd
the gluon and quark @xmath22-channel exchanges are themselves reggeized .
the qcd pomeron ( or bfkl pomeron as it is frequently called ) is a compound state of two reggeized gluons .
in general it may have multi - gluon components .
the bfkl framework allows the behaviour of the scattering of hadronic objects with transverse scale @xmath61 at centre - of - mass energy @xmath53 to be predicted in the domain @xmath62 . in the leading @xmath63 ( ll )
approximation the cross section [ eq : a18 ] = _ n = 0^ c_n ( _ s s)^n .
since @xmath64 a resummation of the @xmath63 terms is necessary .
bfkl carried out this ll summation about 25 years ago with the result [ eq : a19 ] ~ s^ = |_s 4 2 , where @xmath65 , see @xcite .
hence we speak of the bfkl or qcd pomeron . if @xmath66 , then @xmath67 .
this result appears to be in contradiction with the observed rise of the relevant cross sections at large @xmath39 .
the data at small @xmath68 , or of forward jets in deep - inelastic scattering at hera , or of @xmath69 scattering at lep , or jets separated by a large rapidity gap at the tevatron . ] indicate a power growth more like 0.3 than 0.5 .
it was expected that the resummation of the nll corrections , that is of the @xmath70 terms , would remove the discrepancy .
recently the computation of these terms has been completed @xcite and the corrections found to be large , giving [ eq : a20 ] = |_s 4 2 ( 1 - 6.3 |_s ) , which puts the usefulness of the whole perturbative approximation into question . fortunately it was observed that a major part of remaining higher order corrections may be resummed to all orders @xcite .
these all - order resummations bring the bfkl programme back under control .
at small bjorken @xmath0 the gluon distribution , @xmath59 , unintegrated over its transverse momentum @xmath71 , should exhibit bfkl behaviour .
that is a characteristic @xmath72 growth as @xmath73 , accompanied by diffusion in @xmath71 .
the transverse momenta are not ordered in the small @xmath0 bfkl evolution leading to a gaussian - type form of @xmath59 in @xmath74 with a width which grows as @xmath75 as @xmath73 .
ultimately the diffusion will be a problem since it leads to increasingly important contributions from the infrared domain of @xmath76 where the bfkl equation is not expected to be valid .
stirling @xcite discussed ways of studying the bfkl pomeron at the lhc .
he concentrated on the original idea of mueller and navelet , that is to study the correlations between two jets widely separated in rapidity .
it was argued that the best bfkl indicator is the rate of the weakening of the azimuthal ( back - to - back type ) correlation between the jets , as the rapidity interval increases a manifestation of the diffusion in @xmath71 .
the tevatron data show much less decorrelation than predicted by naive bfkl , but a much more realistic monte carlo has been developed @xcite which allows the proper constraints on phase space to be imposed .
predictions for the lhc were presented . de roeck @xcite discussed the opportunities at the lhc to observe the behaviour of parton densities at very small @xmath0 .
in particular he emphasized that it may be possible to probe the gluon distribution in the @xmath77 and @xmath78 domain by observing either prompt photon production @xmath79 or drell - yan production at very large rapidities .
the latter process involves a convolution to allow for the @xmath80 transition , which is required for a gluon - initiated reaction ; consequently somewhat large values of the gluon @xmath0 are probed .
he pointed out that this domain may allow the shadowing corrections to @xmath81 to be studied .
prompted by this talk , a quantitative study was performed using a unified evolution equation which embodies both bfkl and dglap behaviour and which incorporates the leading @xmath82 triple - pomeron vertex @xcite .
the shadowing corrections were found to be small in the hera domain , but lead to about a factor of two suppression of the gluon in the @xmath83 region , which should be accessible in the experiments at the lhc .
one possibility to measure the luminosity at the lhc is to observe exclusive lepton - pair production via photon - photon fusion [ eq : a21 ] pp p + ^+ ^- + p , where @xmath84 or @xmath85 , see @xcite . the born amplitude ( fig .
9(a ) ) may be calculated within qed @xcite , and there are no strong interactions involving the leptons in the final state . the main phenomenological questions concern , first , the size of the absorptive corrections arising from inelastic proton - proton rescattering ( sketched symbolically in fig . 9(b ) ) and , second , how to suppress the proton dissociation contributions of fig .
these questions are addressed in ref .
the dissociation contributions vanish as @xmath86 , due to gauge invariance , where the @xmath87 are defined in fig .
since it is difficult to measure a leading proton with @xmath88 mev , it is proposed @xcite to select events with very small transverse momentum of the lepton pair [ eq : a22 ] p_t ( ^+ ^- ) | _ ^+ t + _ ^- t | , with typically @xmath89 mev .
moreover the rescattering correction of fig .
9(b ) is suppressed because the main part of the born amplitude ( fig .
9(a ) ) comes from large impact parameters by a @xmath90 selection rule . ]
@xmath91 , whereas the rescattering occurs at smaller @xmath91 .
the detection of the @xmath7 and @xmath6 processes differ .
hence their use as a luminosity monitor is different .
first we discuss @xmath7 production @xcite . to identify muons ( and separate them from @xmath92 mesons ) they have to have rather large transverse energy , @xmath93 gev .
it is still possible to satisfy the @xmath94 mev cut , but the cross section is significantly reduced .
the main contribution to the rescattering correction comes from the @xmath95 domain , where @xmath96 is the loop momentum in fig .
9(b ) . as a consequence
the correction is given by @xmath97 , where @xmath98 is a known , small numerical coefficient @xcite . if @xmath99 mev then the correction is only 0.13% .
in addition to the small @xmath100 cut , it is proposed @xcite to fit the observed distribution in the muon acoplanarity angle @xmath101 in order to distinguish the elastic mechanism , fig .
9(a ) , via its prominent peak at @xmath102 , from the background processes which are flat in @xmath101 .
although the @xmath103 and @xmath100 cuts reduce the cross section , the muons have the advantage that we may trace the tracks back to the interaction vertex , and hence isolate the interaction in pile - up events .
so , in principle , @xmath7 production can act as a luminometer in high luminosity lhc runs .
an accuracy of @xmath104 is claimed provided the muon trigger is good enough @xcite .
for @xmath6 production we do not need to select events with large @xmath105 , and so we may consider the small @xmath105 domain where the @xmath6 production cross section is much larger , and where the rescattering correction becomes totally negligible .
this will require a dedicated detector in the forward regions for electrons of energy about 5 gev .
if the threshold energy were reduced to 1 gev then the signal is increased by 15 and the signal - to - background ratio is substantially improved @xcite . it is claimed that an absolute luminosity measurment down to @xmath106 is possible for low luminosities @xmath107 , but for high luminosity the @xmath6 method may be limited by pile - up effects .
@xmath2 and @xmath3 production in high energy @xmath8 collisions have clean signatures through their leptonic decay modes , @xmath108 and @xmath109 , and so may be considered as potential luminosity monitors @xcite .
a vital ingredient is the accuracy with which the cross sections for @xmath2 and @xmath3 production can be theoretically calculated .
the cross sections depend on parton distributions , especially quark densities , in a kinematic region where they are believed to be reliably known .
recent determinations of @xmath110 at the lhc are shown in fig .
the solid squares and triangles are from the nlo parton analyses of @xcite and the final two predictions are from the nlo and nnlo analyses of @xcite .
the two major uncertainties appear to be due to the value of @xmath111 and to using different parton densities labelled by @xmath112 and @xmath113 .
the @xmath114 and @xmath115 values correspond to changing @xmath116 by @xmath117 , which is probably too conservative , so a @xmath118 uncertainty in @xmath110 is more realistic from this source .
the normalisation of the input data used in the global parton analyses is another source of uncertainty in @xmath110 .
the hera experiments provide almost all of the data used in the global analyses in the relevant small @xmath0 domain .
the @xmath112 and @xmath113 parton sets correspond to separate global fits in which the hera data have been renormalized by @xmath119 respectively . allowing for these uncertainties ,
we conclude that the cross sections of @xmath2 and @xmath3 production are known to be about @xmath120 at the lhc energy . for a precise measurement allowance
should be made for @xmath121 pair production and for @xmath2 bosons produced via @xmath22-quark decays , which produce about @xmath122 of the total signal .
caron @xcite discussed the experimental requirements of using @xmath2 and @xmath3 production to determine the luminosity at the lhc .
the luminosity determinations based on the measurement of the forward elastic cross section and on two - photon @xmath6 production can only be made in low luminosity runs , and require dedicated forward detectors and triggers
. on the other hand , the measurement of @xmath2 or @xmath3 and two - photon @xmath7 production may be performed at high luminosity with the central detector .
the latter process requires a good di - muon trigger with thresholds of @xmath123 gev for each muon . in principle
, we may monitor the luminosity using any process , with a significant cross section , which is straightforward to detect cleanly .
for example , it could be single - pion production in a given rapidity and @xmath124 domain or inclusive @xmath7 production in a well defined kinematic region . in this way we may determine the relative luminosity and calibrate the monitor by comparing the number of events detected for the monitor reaction with the number of events for a process whose cross section is known .
this has the advantage that the calibration may be carried out in a low luminosity run .
the desired goal of measuring the luminosity to better than @xmath125 definitely seems attainable .
we note that for applications where it is sufficient to know the parton - parton luminosity , better accuracy can be achieved .
one view of forward or soft physics , which dominates the interactions at the lhc , is that it is an unfortunate unavoidable complication to the exciting rare events which we hope to see
. it may be useful as a luminosity monitor , but the pomeron is boring . another view expressed at the meeting is that to avoid the pomeron is to deny our birthright .
almost all qcd is contained in the pomeron .
the pomeron indirectly spawned string theory and sent a galaxy of physicists spinning off into higher dimensions .
most of the rich hera physics is driven by the pomeron .
the soft pomeron in the non - perturbative domain has fascinating regge properties , leading to glueballs and mixing with @xmath126 states . with the advent of qcd we were reminded that the quark and gluon are not elementary , but reggeons
and that the hard or qcd pomeron is the compound state of two ( or more ) reggeized gluons .
the correct interpolation between the hard and the soft regimes is an exciting fundamental problem yet to be solved .
it is a special pleasure to thank , on behalf of all of the participants , dan - olof riska , risto orava and the other members of the organizing committee , together with laura salmi , for arranging such an excellent workshop , and for their hospitality in helsinki .
i gratefully acknowledge the help that i have received from aliosha kaidalov , valery khoze , misha ryskin and stefan tapprogge on the subject of this workshop .
bjorken , int . j. mod
a7 * ( 1992 ) 4189 . f. gianotti and m. pepe - altarelli , nucl .
* 89 * ( 2000 ) 177 .
s. tapprogge , these proceedings .
m. bozzo , these proceedings .
v. nomokonov , these proceedings .
k. piotrzkowski , these proceedings ; phys . rev .
* d63 * ( 2001 ) 071502(r ) .
w. guryn , these proceedings .
block , these proceedings ( talk 2 ) .
vorobyov , these proceedings .
khoze , a.d .
martin and m.g .
ryskin , eur .
j. * c18 * ( 2000 ) 167 .
lipatov , these proceedings .
a. martin , these proceedings . s. mandelstam , nuovo cimento * 30 * ( 1963 ) 1113 , 1127 .
abramovskii , v.n .
gribov and o. kancheli , sov . j. nucl
. phys . * 18 * ( 1974 ) 308 .
landshoff , these proceedings .
kaidalov , these proceedings .
ter - martirosyan , itep reports 70,71 ( 1975 ) ; 7,11,133135,158 ( 1976 ) .
kaidalov , phys . rep .
* 50 * ( 1979 ) 157 .
p. schlein , these proceedings .
block , these proceedings ( talk 1 ) and references therein ; + m.m . block and f.
halzen , hep - ph/0101022 .
e. gotsman , e. levin and u. maor , phys . lett .
* b452 * ( 1999 ) 387 .
kaidalov , l.a .
ponomarev and k.a .
ter - martirosyan , sov .
* 44 * ( 1986 ) 468 .
khoze , a.d .
martin and m.g .
ryskin , nucl .
* 99 * ( 2001 ) 213 .
dokshitzer , v.a .
khoze and s.i .
troyan , sov .
* 46 * ( 1987 ) 712 .
bjorken , phys . rev .
* d47 * ( 1993 ) 101 .
e. gotsman , e. levin and u. maor , phys . rev .
* d60 * ( 1999 ) 094011 and references therein .
khoze et al . ,
these proceedings and references therein .
albrow and a. rostovtsev , hep - ph/0009336 .
m.l . good and w.d .
walker , phys .
* 126 * ( 1960 ) 1857 .
gribov , sov .
jetp * 19 * ( 1969 ) 483 .
j. pumplin , physica scripta * 25 * ( 1982 ) 191 .
g. snow , these proceedings .
cdf collaboration : t. affolder et al .
rev . lett .
* 84 * ( 2000 ) 5043 .
p. marage , these proceedings .
kaidalov et al . , in preparation .
salam , jhep * 9807 * ( 1998 ) 19 ; + m. ciafaloni , d. colferai and g.p .
salam , phys .
* d60 * ( 1999 ) 114036 ; + r.s .
thorne , phys .
* b474 * ( 2000 ) 372 ; hep - ph/0103210 ; + see also j. kwiecinski , a.d .
martin and p.j .
sutton , zeit .
* c71 * ( 1996 ) 585 .
stirling , these proceedings .
j. andersen et al . , jhep 0102:007 ( 2001 ) .
a. de roeck , these proceedings .
kimber , j. kwiecinski and a.d .
martin , hep - ph/0101099 , phys . lett .
* b * ( in press ) .
a.g . shamov and v.i .
telnov , these proceedings .
b. caron , these proceedings .
see , for example , v.m .
budnev , i.f .
ginzburg , g.v .
meledin and v.g .
serbo , phys . lett . *
b39 * ( 1972 ) 526 .
frisch , cdf / physics / public/2484 ( 1994 ) ; + m. dittmar , f. pauss and d. zrcher , phys .
* d56 * ( 1997 ) 7284 .
khoze , a.d .
martin , r. orava and m.g .
ryskin , hep - ph/0010163 , eur .
j. * c * ( in press ) .
martin , r.g .
roberts , w.j . stirling and r.s .
thorne , eur .
j. * c14 * ( 2000 ) 133 .
martin , r.g .
roberts , w.j . stirling and r.s .
thorne , eur .
j. * c18 * ( 2000 ) 117 .
felix proposal , cern / lhcc 97 - 45 , lhcc / i10 ( 1997 ) . | an attempt is made to summarize the discussion at the workshop , except for the panel discussion on the ability of the lhc detectors to accommodate forward reactions .
the workshop focused on two main topics .
the first topic was forward physics at the lhc .
predictions were made for forward reactions , including elastic scattering and soft diffractive processes , in terms of ( multi ) pomeron exchange , using knowledge gained at lower energies .
the survival probability of rapidity gaps accompanying hard subprocesses was studied .
the nature of the pomeron , before and after qcd , was exposed , and some aspects of small @xmath0 physics at the lhc were considered .
the second topic of the workshop concerned the accuracy of the luminosity measuring processes at the lhc .
attention concentrated on three methods .
the classic approach based on the optical theorem , secondly , the observation of the pure qed process of lepton - pair ( @xmath1 ) production by photon - photon fusion and , finally , the measurement of inclusive @xmath2 and @xmath3 production . plus 2 mm minus 2 mm 23.0 cm 17.0 cm -1.0 in -42pt ippp/01/13 + dcpt/01/26 + 27 march 2001 + * summary talk : * + + * luminosity determination at the lhc * + a.d .
martin@xmath4 + @xmath4 department of physics and institute for particle physics phenomenology , university of durham , durham , dh1 3le |
the ligo interferometric gravitational wave ( gw ) detector held its first science run ( s1 ) in 2002 , from 23 august to 9 september @xcite .
ligo consists of an interferometer ( ifo ) with 4 km arms in livingston , la , usa ( the ligo livingston observatory , or llo ) , called l1 for short , and two ifos , with arms of 4 km and 2 km , in hanford , wa , usa ( the ligo hanford observatory , or lho ) , called h1 and h2 , respectively .
the data were analysed to search for gw bursts @xcite , signals from inspiralling neutron star binaries @xcite , periodic signals from a rotating neutron star @xcite , and stochastic backgrounds @xcite .
this paper summarizes the analysis method and results of the search for a stochastic background of gravitational waves ( sbgw ) , which are explained in more detail in @xcite .
a sbgw is assumed for simplicity to be isotropic , unpolarized , gaussian , and stationary .
subject to these assumptions , the sbgw is completely described by its power spectrum .
it is conventional to express this spectrum in terms of the gw contribution to the cosmological parameter @xmath1 : @xmath2 note that @xmath3 has been constructed to be dimensionless , and represents the contribution to the overall @xmath4 per _ logarithmic _ frequency interval . in particular , it is _ not _ equivalent to @xmath5 .
note also that since the critical density @xmath6 , which is used in the normalization of @xmath3 , is proportional to the square of the hubble constant @xmath7 @xcite , it is convenient to work with @xmath8 , which is independent of the observationally determined value of @xmath9 .
is now much more accurately known than it once was , we still work in terms of @xmath8 to facilitate comparisons with prior results . ]
the standard method to search for such a background is to cross - correlate the outputs of two gw detectors @xcite .
if the noise in the two detectors is uncorrelated , the only non - zero contribution to the average cross - correlation ( cc ) will come from the sbgw . in the optimally - filtered cc method ( described in more detail in @xcite ) , one calculates a cc statistic @xmath10 where @xmath11 are the data streams from the two detectors , @xmath12 are their fourier transforms , and @xmath13 ( with fourier transform @xmath14 ) is a suitably - chosen optimal filter . the choice which optimizes signal - to - noise ratio for a constant-@xmath3 background is @xcite @xmath15 the normalization of the optimal filter is conventionally chosen so that in the presence of a sbgw of strength @xmath16 , the expected mean value of the cc statistic is @xmath17 where @xmath18 is the duration of the analysed data sets .
the expected variance of the cc statistic is @xmath19 ^ 2}{p_1(f)p_2(f ) } \right)^{-1 } \ .\ ] ] the method is sensitive to backgrounds on the order of @xmath20 ^ 2}{f^6 { p_1(f ) p_2(f ) } } \right)^{-1/2 } \ .\ ] ] the sensitivity of this method improves with time and is limited by the power spectral densities @xmath21 of the noise in the two detectors .
the factor @xmath22 in the numerator of the integral is the _ overlap reduction function _
@xcite , which describes the observing geometry . here
@xmath23 is a projector onto symmetric traceless tensors transverse to a direction @xmath24 and @xmath25 are the _ detector response tensors _ for the two detectors .
these are the tensors with which the metric perturbation @xmath26 at the detector should be contracted to obtain the _ gw strain _ @xmath27 .
if @xmath28 and @xmath29 are unit vectors pointing in the directions of an ifo s two arms , its response tensor is @xmath30 the overlap reduction function is equal to unity for the case of a pair of ifos at the same location with their arms aligned , and is suppressed as the detectors are rotated out of alignment or separated from one another .
it also oscillates with frequency as correlations are suppressed for detectors whose separation is comparable to or greater than the corresponding gw wavelength .
figure [ fig : alloverlap ] shows the overlap reduction functions for combinations of detectors which were operational during s1 . for combinations of the ligo livingston observatory ( llo ) and ligo hanford observatory ( lho ) with each other and with the geo600 site .
( geo600 was also operational during s1 , but has not been included in this analysis because it was considerably less sensitive than the ligo detectors . )
note that the overlap reduction function for correlations between the two detectors at lho ( h1 and h2 ) is identically equal to unity .
the solid lines ( at @xmath31 and @xmath32 ) show the range of frequencies used in our analysis .
, height=336 ]
the previous best upper limit on a sbgw from direct observation with gw detectors was @xmath33 @xcite , set by correlating the resonant bar detectors explorer ( in geneva , switzerland ) and nautilus ( near rome , italy ) .
a broad - band limit of @xmath34 was set using a pair of ` prototype ' ifos @xcite .
more stringent upper limits can be set on astrophysical grounds .
they are detailed elsewhere @xcite , but we mention the bound from big - bang nucleosynthesis @xcite , which states that a cosmological sbgw is limited by @xmath35 this broad - band limit implies that any cosmologically interesting sbgw very likely lies several orders of magnitude below the existing limits .
each of the three combinations of detectors ( h1-l1 , h2-l1 , and h1-h2 ) was analysed separately for ccs .
since the power spectra @xmath21 varied over the course of the s1 run , the concident data for each pair of ifos were divided into 15 min blocks , and an optimal filter constructed for each such block .. ] to maximize overall signal - to - noise ratio @xcite , we combined the cc statistics from the different blocks using a weighting factor of @xmath36 , where @xmath37 is the theoretical standard deviation defined in ( [ eq : ccvar ] ) . note that this can be calculated from the individual power spectra , without cross - correlating the data . to avoid problems from noisy and presumably less gaussian data , we discarded the 15 min blocks with the highest @xmath37 values , corresponding to a 5% total contribution to @xmath38 . for each block ,
an optimal filter was constructed with a frequency resolution of @xmath39 according to the discrete frequency - domain analogue of ( [ eq : qdef ] ) .
the range of frequencies included in the calculation of the cc statistic was chosen to be 40314hz for h1-l1 and h2-l1 , and 40300hz for h1-h2 .
given the power spectra of the instruments and the expected spectrum of correlations associated with a constant @xmath3 , frequencies outside that range were not expected to improve the sensitivity appreciably .
additionally , individual frequency bins associated with cross - correlated instrumental noise were omitted from the sum over frequencies ( which means the optimal filter was effectively set to zero there ) .
these were integer multiples of 16hz and 60hz , as well as a few frequencies which had a coherence over the entire run above 0.2 , namely 250hz for l1-h2 , and 168.25hz and 168.5hz for h1-h2 .
each 15 min block was divided into ten 90 s segments using a tukey window which consisted of half s hann transitions on either side of an 89 s flat top .
the cc statistic was calculated for each segment , using the discrete analogue of the frequency - domain form of ( [ eq : ccstat ] ) , and these were combined to give a cc statistic for the entire block . in this way
, we were able to obtain , through the measured standard deviation of the ten cc statistics within a block , a statistical measure of the error associated with the cc statistic for the block .
we also estimated the systematic error associated with the change in sensitivity and calibration over the course of each block .
we then combined all three , appropriately weighted over the whole run , to obtain a total estimated error @xmath40 associated with the point estimate @xmath41 calculated from the weighted average of all the cc statistics using ( [ eq : ccmean ] ) . in the absence of cross - correlated noise , the 90% confidence level upper limit on @xmath42 , the constant value of @xmath3 , is @xmath43
the results for the three ifo pairs are summarized in table [ tab : results ] .
.summary of the point estimate @xmath41 and total estimated error @xmath44 for the three ifo pairs considered .
note that no upper limit is set from h1-h2 , the two ifos at the hanford site , since there was evidence of cross - correlated noise . associated with each of the values quoted is an additional 20% uncertainty arising from the calibration of the instruments . [
cols="<,^,^,^,^",options="header " , ] there is a statistically significant anti - correlation observed between h1 and h2 , two ifos which share the same vacuum envelope at the lho site in hanford , wa .
time - shift and @xmath45 analyses show that this anti - correlation is inconsistent with constant-@xmath46 sbgw , so we conclude it is due to instrumental cross - correlations between the two colocated detectors . for the inter - site measurements ( h1-l1 and h2-l1 ) , the lack of statistically significant cross - correlations makes these checks trivial , and we proceed to setting an upper limit from each pair .
the stronger upper limit is set by h2 and l1 , and it is @xmath47 this represents a factor of 23 improvement over the previous direct upper limits described in sec .
[ s : prior ] , and an improvement by a factor of over 1000 over the previous measurements with interferometric detectors . for more details on the analysis ,
the reader is directed to @xcite .
the authors gratefully acknowledge the support of the united states national science foundation for the construction and operation of the ligo laboratory and the particle physics and astronomy research council of the united kingdom , the max - planck - society and the state of niedersachsen / germany for support of the construction and operation of the geo600 detector .
the authors also gratefully acknowledge the support of the research by these agencies and by the australian research council , the natural sciences and engineering research council of canada , the council of scientific and industrial research of india , the department of science and technology of india , the spanish ministerio de ciencia y tecnologia , the john simon guggenheim foundation , the david and lucile packard foundation , the research corporation , and the alfred p. sloan foundation .
jtw also wishes to thank his colleagues in the ligo scientific collaboration , especially the stochastic sources upper limits group , chaired by j romano and p fritschel , as well as the albert einstein institute in golm , germany .
jtw was supported by the national science foundation under grant phy-0300609 and by the max - planck - society .
19 abbot b et al ( ligo scientific collaboration ) 2003 detector description and performance for the first coincidence observations between ligo and geo _ preprint _ gr - qc/0308043 ( _ nucl .
instrum .
methods _ at press ) abbot b et al ( ligo scientific collaboration ) 2003 first upper limits on gravitational wave bursts from ligo _ preprint _ gr - qc/0312056 abbot b et al ( ligo scientific collaboration ) 2003 analysis of ligo data for gravitational waves from binary neutron stars _ preprint _
gr - qc/0308069 ( d submitted ) abbot b et al ( ligo scientific collaboration ) 2004 d * 64 * at press ( _ preprint _ gr - qc/0308050 ) abbot b et al ( ligo scientific collaboration ) 2003 analysis of first ligo science data for stochastic gravitational waves _
gr - qc/0312088 ( d to be submitted ) kolb e w and turner m s 1990 _ the early universe _
( addison - wesley , 1990 ) christensen n 1992 d * 46 * 5250 allen b 1997 in _ proceedings of the les houches school on astrophysical sources of gravitational waves _ , eds marck j a and lasota j p , cambridge , 373 ; allen b 1996 _ preprint _ gr - qc/9604033 allen b and romano j d 1999 d * 59 * 102001 ; allen b and romano j d 1997 _ preprint _ gr - qc/9710117 whelan j t et al 2001 1521 ; whelan j t et al 2001 _ preprint _ gr - qc/0110019 flanagan 1993 d * 48 * 2389 ; flanagan 1993 _ preprint _
astro - ph/9305029 astone p et al 1999 _ astron . astrophys._*351 * 811 ; astone p et al 1998 in _ proceedings of the second edoardo amaldi conference on gravitational waves _
, eds coccia e , pizzella g and veneziano g , singapore , 192 compton c , nicholson d and schutz b f 1994 in _ proceedings of the seventh marcel grossman meeting on general relativity _ , eds jantzen r t and keiser g m , singapore , 1078 maggiore m 2000 _ phys .
rep._*331 * 28 ; maggiore m 1999 _ preprint _ gr - qc/9909001 adhikari r , gonzlez g , landry m and oreilly b ( for the ligo scientific collaboration ) 2003 s903 | i describe analysis of correlations in the outputs of the three ligo interferometers from ligo s first science run , held over 17 days in august and september of 2002 , and the resulting upper limit set on a stochastic background of gravitational waves . by searching for cross - correlations between the ligo detectors in livingston , la and hanford , wa
, we are able to set a 90% confidence level upper limit of @xmath0 . |
the study of charged relativistic objects in general relativity is achieved by solving the einstein - maxwell system of equations and imposing conditions for physical acceptability .
this is not easy to achieve because of the nonlinearity of the field equations .
the exact solutions found have many applications in relativistic astrophysics .
the models generated have been used in the description of neutron stars and black hole formation by ray et al .
@xcite and de felice et al .
particular models have also helped in the establishment of the absolute stability limit for charged spheres by giuliani et al .
@xcite and bohmer and harko @xcite .
several models of charged relativistic matter have been used to study strange stars by mak and harko @xcite , bombaci @xcite , komathiraj and maharaj @xcite and thirukkanesh and maharaj @xcite .
charged models have been also used in the description of strange quark matter by discus et al .
@xcite , hybrid protoneutron stars by nicotra et al .
@xcite , and bare quark stars by usov et al . @xcite .
a geometric approach is to assume the existence of a group of conformal motions on spacetime ; exact solutions have been found by mak and harko @xcite for strange quark matter and esculpi and aloma @xcite for anisotropic relativistic charged matter by assuming the existence of a conformal killing vector in static spherically symmetric spacetimes .
models with an equation of state are desirable in the description of realistic astrophysical matter .
however most explicit solutions of the einstein - maxwell system that have been found do not satisfy this property .
there have been some attempts made recently to find exact analytic solutions of the einstein- maxwell system with a linear equation of state .
these include the treatments of ivanov @xcite , sharma and maharaj @xcite , and thirukkanesh and maharaj @xcite .
particular solutions with a quadratic equation of state , relating the radial pressure to the energy , where found by feroze and siddiqui @xcite .
this is an important advance since the complexity of the model is greatly increased because of the nonlinearity of the radial pressure in terms of the energy density .
however the investigations mentioned above all suffer from the undesirable property of possessing a singularity in the property charge density at the centre of sphere .
an essential requirement for a well behaved electromagnetic field is regularity of the proper charge density throughout the matter distribution , particularly at the stellar centre .
the importance of this feature has been highlighted in the analysis of varela et al .
@xcite whose treatment offers a general approach of dealing with anisotropic charged matter with linear or nonlinear equations of state .
it is desirable to eliminate the singularity in the charge density for a detailed and complete analysis of physical properties of charged compact objects .
our results may be helpful in the study of compact stars and gravitational collapse relating to neutron stars and black holes . in this regard
we refer to particular papers some of which have static spherical geometry and others are dynamical .
novikov @xcite showed in the case of spherical geometry that collapse of electrically charged matter may replaced by expansion and infinite densities are avoided . a general treatment of collapsing charged matter was completed by bekenstein @xcite who showed that nonzero pressure plays a significant role .
the analysis of raychaudhuri @xcite for charged dust distributions showed that conditions for collapse and oscillation depend on the ratio of matter density to charge density .
if this ratio is large , corresponding to weakly charged dust spheres , then shell crossings can not be avoided in gravitational collapse as proved by ori @xcite .
krasinski and bolejko @xcite showed that there exist initial conditions for a charged dust sphere with finite radius so that a full cycle of pulsation can be completed by the outer layer with no internal singularity . a full and comprehensive analysis of charged
, dissipative collapse is provided by di prisco _
@xcite for the free - streaming and diffusion approximations .
a related and detailed analysis in the gravitational collapse of a charged medium was performed by kouretsis and tsagas @xcite where the role of raychaudhuri equation is highlighted .
exact solutions with an equation of state , such the quadratic case considered in this paper , are helpful in such studies .
the objective of this paper is to find new exact solutions of the einstein - maxwell field equations with a charged anisotropic matter distribution and a quadratic equation of state .
we indicate that particular models found in the past with an equation of state are part of our general analytical framework .
previous solutions with a linear or quadratic equation of state are regained in our treatment .
we ensure that the charge density is regular at the centre of the compact body and the physical criteria are satisfied . in sect .
[ basic ] , we give the einstein - maxwell field equations for a static spherically symmetric line element as an equivalent system of differential equations utilizing a transformation due to durgapal and bannerji @xcite . in sect .
[ c1 ] , we present new exact solutions to the einstein - maxwell system with a quadratic equation of state .
the solution is regular at the centre of the compact object .
this analysis extends the treatment of thirukkanesh and maharaj @xcite , and feroze and siddiqui @xcite .
known solutions with an equation of state are presented in sect .
[ d ] , as particular cases of our new results . in sect .
[ p ] , a physical analysis of the new solutions is performed ; the matter variables and the electromagnetic quantities are plotted .
we summarise the results obtained in this paper .
in standard coordinates the line element for a static spherically symmetric spacetime , modelling the interior of the relativistic object , has the form @xmath0 we take the energy momentum tensor to be of the form @xmath1 where the quantity @xmath2 is the tangential pressure , @xmath3 is the radial pressure , @xmath4 is the density , and @xmath5 is the electric field intensity . then the einstein - maxwell equations can be written in the form @xmath6^{\prime } & = & \rho+\dfrac{1}{2}e^{2},\\ \label{f3 } -\dfrac{1}{r^{2}}(1-e^{-2\lambda})+\dfrac{2\nu^{\prime}}{r}e^{-2\lambda } & = & p_{r}-\dfrac{1}{2}e^{2},\\ \label{f4 } e^{-2\lambda}\bigg(\nu^{\prime\prime}+\nu^{\prime2}+\dfrac{\nu^{\prime}}{r } -\nu^{\prime}\lambda^{\prime}-\dfrac{\lambda^{\prime}}{r}\bigg ) & = & p_{t}+\dfrac{1}{2}e^{2},\\ \label{f5 } \sigma & = & \dfrac{1}{r^{2}}e^{-\lambda}(r^{2}e)^{\prime},\end{aligned}\ ] ] where primes represent differentiation with respect to _
r_. the quantity @xmath7 represents the proper charge density .
we are utilising units where the coupling constant @xmath8 and the speed of light @xmath9 .
the mass within a radius _ r _ of the sphere is @xmath10 we now introduce a new independent variable _ x _ and define new functions _
y _ and _ z _ so that @xmath11 where _ a _ and _ c _ are constants .
we assume an equation of state of the general form @xmath12 for the matter distribution .
we take the quadratic form @xmath13 relating the radial pressure @xmath3 to the energy density @xmath4 . in the above @xmath14 , @xmath15 , and @xmath16
. then the einstein - maxwell equations governing the gravitational behaviour of a charged anisotropic sphere , with a quadratic equation of state , are represented by @xmath17\frac{\dot{y}}{y}- \alpha \left[\frac{(1-z)}{x}-2\dot{z}-\frac{e^{2}}{2c } \right]\nonumber\\ & & -c\gamma\left[\frac{(1-z)}{x}-2\dot{z}-\frac{e^{2}}{2c } \right]^{2 } + \dot{z}-\frac{e^{2}}{2c}+\frac{\beta}{c},\\ \label{f12 } \dfrac{\dot{y}}{y } & = & \frac{(1-z)(1+\alpha)}{4z}-\frac{(1+\alpha)e^{2}}{8cz}- \frac{\alpha\dot{z}}{4z}-\frac{\beta}{4cz}\nonumber\\ & & + \frac{c\gamma}{4z}\left[\frac{(1-z)}{x}-2\dot{z}-\frac{e^{2}}{2c } \right]^{2},\\ \label{f13 } \dfrac{\sigma^{2}}{c } & = & \dfrac{4z}{x}\left(x\dot{e}+e\right)^{2},\end{aligned}\ ] ] where dots denote differentiation with respect to the variable @xmath18 .
equations ( [ f8])-([f13 ] ) are similar to the field equations of thirukkanesh and maharaj @xcite ; however in this case the equation of state is quadratic .
the quantity @xmath19 is called the measure of anisotropy and vanishes for isotropic pressures .
the nonlinear system as given in ( [ f8])-([f13 ] ) consists of six independent equations in six variables involving the matter and electromagnetic quantities @xmath20 and the two gravitational potentials _ y _ and _ z_. the nonlinearity of the einstein - maxwell system ( [ f8])-([f13 ] ) has been increased , when compared with many earlier treatments , become of the appearance of the quadratic term in ( [ f9 ] ) ; when @xmath21 then there is a linear equation of state .
in addition equation ( [ f12 ] ) now contains terms with @xmath22 increasing the complexity of system since @xmath23 in general .
to integrate the einstein - maxwell system we make the particular choices @xmath24 the gravitational potential _ z _ is well behaved and finite at the origin .
the electric field intensity _ e _ is continuous , regular at the origin and approaches a constant value for increasing values of _ x_. the constants _ a _ , _ b _ , _ k _ and _ s _ are real .
the general analytic functional forms for _ z _ and _ e _ regain particular cases studied in the past with an equation of state .
on substituting ( [ s1 ] ) and ( [ s2 ] ) into ( [ f12 ] ) we obtain the first order equation @xmath25}+\dfrac{{\alpha}(a - b)}{2(1+ax)[1+bx]}- \frac{\beta(1+ax)}{4c[1+bx]}\nonumber\\ & & -\frac{(1+\alpha)[k(3+ax)+sa^{2}x^{2}]}{8(1+ax)[1+bx]}+ \frac{c\gamma[(3+ax)(2a - ab - k)-sa^{2}x^{2}]^{2}}{16(1+ax)(1+bx ) } , \label{s4}\end{aligned}\ ] ] for the metric function _
y_. in spite of complexity of equation ( [ s4 ] ) it can be solved in general .
on integrating @xmath26 we get @xmath27^{n}\exp\left[f(x)\right],\label{s5}\ ] ] where _ d _ is the constant of integration . the function _
f(x ) _ is given explicitly by @xmath28^{2}\left[\frac{2(2b - a)(1+ax)+(b - a)}{8(b - a)^{2}(1+ax)^{2}}\right]\nonumber\\ & & -cs\gamma\left [ \frac{(a - b)^{2}(ax+2)-a(2a+s)(1+ax)}{4(a - b)(1+ax)^{2}}\right]\nonumber\\ & & -cs\gamma\left[\frac{(a - b)(4k+s)+(2k(b-3a)+3bs)(1+ax)}{32(a - b)^{2}(1+ax)^{2 } } \right]\nonumber\\ & & + \frac{ax}{16bc}[c^{2}s^{2}\gamma -2cs(1+\alpha)-4\beta].\label{sf}\end{aligned}\ ] ] the constants _ m _ and _ n _ have the form @xmath29^{2 } \left[\frac{b^{2}}{(b - a)^{3}}+\frac{b}{(b - a)^{2}}+ \frac{1}{4}\right]\nonumber\\ & & + \frac{cs\gamma}{8(a - b)^{3}}\left[(a - b)[2s(a - b)+a+b]+3ab(k-2b)-b^{2}k\right.\nonumber\\ & & \left .
+ 2b^{3}(2a-1 ) \right]\nonumber\\ n & = & \frac{(1+\alpha)}{8b}[2(a - b)-k]+\frac{(1+\alpha)k-2\alpha(a - b)}{4(b - a ) } + \frac{\beta ( a - b)}{4cb^{2}}\nonumber\\ & & + \gamma [ 2(a - b)-k]^{2}\left[\frac{b^{2}}{(b - a)^{3}}+\frac{b}{(b - a)^{2}}+ \frac{1}{4}\right]+ \frac{csa^{2}(1+\alpha)}{8b^{2}(b - a)}\nonumber\\ & & + \frac{cs\gamma}{16b^{2}(b - a)^{3}}\left[a^{4}(s+4b)+(k+2b)(6a^{2}b^{2}-2a^{3}b)\right].\nonumber\end{aligned}\ ] ] then we can generate an exact model for the einstein - maxwell system ( [ f8])-([f13 ] ) in the form @xmath30^{2n}\exp[2f(x)],\\ \label{s8 } \dfrac{\rho}{c } & = & \dfrac{(2a-2b - k)(3+ax)-sa^{2}x^{2}}{2(1+ax)^{2}},\\ \label{s9 } p_{r } & = & \gamma \rho^{2}+\alpha \rho-\beta , \\
\label{s10 } p_{t } & = & p_{r}+\delta,\\ \label{s11 } \frac{\delta}{c } & = & \frac{4x(1+bx)}{1+ax}\left[\frac{m(m-1)a^{2}}{(1+ax)^{2 } } + \frac{2mnab}{(1+ax)(1+bx ) } + \frac{2ma\dot{f}(x)}{1+ax}+\right.\nonumber\\ & & \left .
\frac{b^{2}n(n-1)}{(1+bx)^{2 } } + \frac{2nb\dot{f}(x)}{1+bx}+ \ddot{f}(x)+ \dot{f}(x)^{2}\right ] + \left[-\frac{2(a - b)x}{(1+ax)^{2}}+\dfrac{4(1+bx)}{(1+ax)}\right]\nonumber\\ & & \left[\frac{am}{1+ax}+ \frac{bn}{1+bx } + \dot{f}(x ) \right]-c\gamma\left[\frac{c(2(a - b ) -k)(3+ax)-csa^{2}x^{2}}{2(1+ax)^{2}}\right]^{2}\nonumber\\ & & -\frac{1}{2(1+ax)^{2}}[2c(a - b)+k(3+ax)+sa^{2}x^{2}-\frac{2\beta}{c}(1+ax)^{2}]\nonumber\\ & & -\alpha\left[\frac{(2(a - b)-k)(3+ax)-sa^{2}x^{2}}{2(1+ax)^{2}}\right],\\ \label{s12 } \frac{e^{2}}{c}&= & \frac{k(3+ax)+sa^{2}x^{2}}{(1+ax)^{2}},\\ \label{s13 } \frac{\sigma^{2}}{c } & = & \frac{c[1+bx]\left(\sqrt{k}(a^{2}x^{2}+3ax+6 ) + 2\sqrt{s}ax\sqrt{3+ax}(2+ax)\right)^{2}}{x(3+ax)(1+ax)^{5}},\end{aligned}\ ] ] where _
f(x ) _ is given by ( [ sf ] ) .
we observe that the exact solution ( [ s6])-([s13 ] ) of the einstein - maxwell system has been written solely in terms of elementary functions .
for this solution the mass function is given by @xmath31x^{1/2}}{3a(1+ax)}\right.\nonumber\\ & & \left.-\frac{5s\arctan(\sqrt{ax})}{a^{3/2}}\right].\end{aligned}\ ] ] the gravitational potentials , matter variables and electromagnetic variables are well behaved and regular in the stellar interior .
however in general there is a singularity in the charge density at the centre which is evident in ( [ s13 ] ) .
this singularity is avoidable when @xmath32 , so that we have @xmath33(2+ax)^{2}}{(1+ax)^{5}}.\ ] ] at the centre of the star @xmath34 and the charge density vanishes .
we have found a general class of exact solutions to the einstein - maxwell system with a quadratic equation of state .
it is interesting to observe that for particular parameter values we can regain uncharged anisotropic and isotropic models @xmath35 , @xmath36 from our general solution ( [ s6])-([s13 ] ) .
we regain the following particular cases of physical interest .
this is a special case of our general solution with the quadratic equation of state @xmath3 = @xmath37 .
if we set @xmath38 , @xmath39 and @xmath40 , then we regain the line element @xmath41dt^{2}+\frac{1+ar^{2}}{1+br^{2}}dr^{2}\nonumber\\ & & + r^{2}(d\theta^{2}+\sin^{2}\theta d\phi^{2})\label{s19},\end{aligned}\ ] ] where @xmath42^{2}\left[\frac{2(2b - a)(1+ar^{2})+(b - a)}{8(b - a)^{2}(1+ar^{2})^{2}}\right],\nonumber\\ m & = & \frac{\alpha}{2}-\frac{(1+\alpha)k}{4(b - a)}+\gamma [ 2(a - b)-k]^{2}\left[\frac{b^{2}}{(b - a)^{3}}+\frac{b}{(b - a)^{2}}+ \frac{1}{4}\right],\nonumber\\ n & = & \frac{(1+\alpha)}{8b}[2(a - b)-k]+\frac{(1+\alpha)k-2\alpha(a - b)}{4(b - a)}+ \frac{\beta ( a - b)}{4b^{2}}\nonumber\\ & & + \gamma [ 2(a - b)-k]^{2}\left[\frac{b^{2}}{(b - a)^{3}}+\frac{b}{(b - a)^{2}}+ \frac{1}{4}\right].\nonumber\end{aligned}\ ] ] the line element ( [ s19 ] ) was found by feroze and siddiqui @xcite which was the first model with quadratic equation of state . some minor misprints in @xcite
have been corrected in our result .
this solution may be used to model a compact body .
if we set @xmath43 then we have the linear equation of state @xmath3 = @xmath44 . also setting @xmath39 , @xmath38 and @xmath45
, we get the line element @xmath46^{2n}\exp\left [ \frac{-a\beta r^{2}}{2(a-\tilde{b})}\right]dt^{2}\nonumber\\ & & + \frac{1+ar^{2}}{1+(a-\tilde{b})r^{2}}dr^{2}+r^{2}(d\theta^{2}+\sin^{2}\theta d\phi^{2})\label{s18},\end{aligned}\ ] ] where @xmath47.\nonumber\end{aligned}\ ] ] the metric ( [ s18 ] ) was found by thirukkanesh and maharaj @xcite .
this solution may be used to model realistic charged compact spheres and strange stars with quark matter in the presence of electromagnetic field .
if we set @xmath43 , @xmath48 , then we regain the linear equation of state @xmath49 where @xmath50 is the density at the surface . by setting @xmath32 @xmath38 , @xmath45 , @xmath39 and @xmath40
we find the following form of the line element @xmath51^{2n}\exp\left ( \frac{-a\beta r^{2}}{2(a-\tilde{b})}\right)dt^{2}\nonumber\\ & & + \frac{1+ar^{2}}{1+(a-\tilde{b})r^{2}}dr^{2}+r^{2}(d\theta^{2}+\sin^{2}\theta d\phi^{2})\label{s17},\end{aligned}\ ] ] where @xmath52 the line element ( [ s17 ] ) represents an uncharged anisotropic sphere and was found by sharma and maharaj @xcite .
it may be used to describe strange stars with a linear equation of state with quark matter .
if we set @xmath43 , @xmath53 , then we obtain the linear equation of state @xmath54 .
on setting @xmath32 , @xmath38 , @xmath45 , @xmath55 , @xmath39 and @xmath40 we regain the line element @xmath56 where @xmath57
the metric ( [ s16 ] ) was first found by lobo @xcite which represents uncharged anisotropic matter .
this solution serves as a stellar interior with @xmath58 and may be matched to the schwarzchild exterior for dark energy stars .
we observe that @xmath59 in general and the model remains anisotropic .
however , we can show for particular parameter values that @xmath60 in the general solution ( [ s6])-([s13 ] ) .
if we set @xmath61 , @xmath32 , @xmath38 , @xmath62 , then we obtain @xmath63\nonumber\\ \delta & = & \frac{x}{4c(1-x)}[\beta-3(1+\alpha)c][\beta-(1 + 3\alpha)c].\label{s20}\end{aligned}\ ] ] two different cases arise from ( [ s20 ] ) by setting @xmath60 .
firstly , we observe that when @xmath53 and @xmath64 then @xmath60 .
the equation of state becomes @xmath65 . with the line element @xmath66 where we have set @xmath67 and @xmath68 .
we mention that the metric ( [ s21 ] ) is the isotropic uncharged de sitter model .
secondly , we observe that when @xmath53 and @xmath69 then @xmath60 .
the equation of state becomes @xmath70 with the line element @xmath71 where @xmath72 and @xmath73 .
the metric ( [ s22 ] ) is the isotropic uncharged einstein model .
.2 cm .2 cm .2 cm .2 cm .2 cm .2 cm .2 cm .2 cm
we show that the exact solutions of the einstein - maxwell system found in section 3 are well behaved by generating graphical plots of matter and electromagnetic variables .
we used the software package mathematica @xcite and we make the particular choices @xmath39 , @xmath74 , @xmath75 , @xmath76 , @xmath77 , @xmath78 and @xmath79 .
we generated the plots for the energy density ( fig .
1 ) , radial pressure ( fig .
2 ) , electric field intensity ( fig .
3 ) , charge density ( fig .
4 ) , mass ( fig .
5 ) , speed of sound ( fig .
6 ) , tangential pressure ( fig . 7 ) and the measure of anisotropy ( fig . 8) . the energy density @xmath4 is a finite and monotonically decreasing function .
the radial pressure @xmath3 is similarly well behaved and continuous .
the electric field intensity @xmath5 is initially small and approaches a maximum value as the boundary is approached .
the proper charge density @xmath7 is nonsingular at the origin , increases and then decreases after reaching a maximum value .
the mass function is a strictly increasing function which is continuous and finite .
the speed of sound is less than the speed of light and causality is maintained throughout the stellar interior .
the radial pressure is decreasing and does reach a finite value of the radial coordinate .
the tangential pressure is also a decreasing function .
the measure of anisotropy is a decreasing function as the boundary is approached and remains finite in the interior .
thus all the matter variables , electromagnetic variables and the gravitational potentials are nonsingular and regular in the region containing the stellar centre .
in particular the proper charge density @xmath7 is finite at the centre unlike earlier treatments .
it is desirable to study comprehensively the stability of our new models ; this is a objective for the future reseach .
the solutions generated may be matched to the exterior reissner - nordstrom spacetime @xmath80 across the boundary @xmath81 .
this generates the following conditions @xmath82 and @xmath83 relating the constants _ a , b , a , c , _ @xmath14 , @xmath15 and @xmath16 . there are sufficients number of free parameters to ensure the continuity of the metric coefficients across the boundary of the star .
it is possible to study the astrophysical significance of the exact solutions to the einstein - maxwell equations found in this paper .
this is the object of future research .
we point out that for suitable parameter values we regain the mass @xmath84 of dey @xmath85 @xcite-@xcite corresponding to a strange star model when there is no electromagnetic field .
therefore the solutions found in this paper may be used to generalise earlier results and to model charged relativistic strange and quark stars .
our aim in this paper was to find new regular exact solutions to the einstein - maxwell system for spherically symmetric gravitational field with an eq of state .
in particular we selected a quadratic equation of state relating the energy density to the radial pressure .
the new models presented in this paper may be used to model relativistic compact objects in astrophysics .
sdm and pmt thank the national research foundation and the university of kwazulu - natal for financial support .
sdm acknowledges that this work is based upon research supported by the south african research chair initiative of the department of science and technology and the national research foundation .
we are grateful to the referees for valuable comments . | we provide new exact solutions to the einstein - maxwell system of equations which are physically reasonable .
the spacetime is static and spherically symmetric with a charged matter distribution .
we utilise an equation of state which is quadratic relating the radial pressure to the energy density .
earlier models , with linear and quadratic equations of state , are shown to be contained in our general class of solutions .
the new solutions to the einstein - maxwell are found in terms of elementary functions .
a physical analysis of the matter and electromagnetic variables indicates that the model is well behaved and regular .
in particular there is no singularity in the proper charge density at the stellar centre unlike earlier anisotropic models in the presence of the electromagnetic field . |
recent low temperature experiments on clean materials near ferromagnetic quantum critical points ( fqcp ) have revealed a remarkable range of unusual properties , including non - fermi liquid scalings over a large phase space , unusual transport , and novel quantum ground states , particularly coexisting ferromagnetism and superconductivity in some materials .
although criticality usually implies a certain universality , present experiments show considerable material dependent aspects that are not well understood , laughlin _
e.g. _ the differences between uge@xmath1 and urhge saxena , aoki and zrzn@xmath1 , @xcite which both show coexisting ferromagnetism and superconductivity but very different phase diagrams , in contrast to mnsi , where very clean samples show no hint of superconductivity around the qcp , possibly because of the lack of the inversion symmetry .
@xcite moreover , by far not every magnetic material can be driven to a qcp by pressure or by other means of supressing ferromagnetism .
typically , the transition becomes first order as the curie temperature , @xmath2 is depressed . if this happens too far away from the fluctuation dominated regime , nothing interesting is seen .
also , more pedestrian effects are often important .
for example , impurities or other defects can lead to scattering that smears out the quantum critical region .
one of the fingerprints of a fqcp , maybe the most universal one , is a substantial overestimation of the tendency to magnetism in conventional density functional theory ( dft ) calculations , such as within the local density approximation ( lda ) .
generally , approaches based on density functional theory ( dft ) are successful in accounting for material dependence in cases where sufficiently accurate approximations exist .
density functional theory is in principle an exact ground state theory .
it should , therefore , correctly describe the spin density of magnetic systems .
this is usually the case in actual state of the art density functional calculations .
however , common approximations to the exact density functional theory , such as the lda , may miss important physics and indeed fail to describe some materials .
a well know example is in strongly hubbard correlated systems , where the lda treats the correlations in an orbitally averaged mean field way and often underestimates the tendency towards magnetism .
overestimates of magnetic tendencies , especially in the lda , are considerably less common , the exceptions being materials near magnetic quantum critical points ( qcp ) ; here the error comes from neglect of low energy quantum spin fluctuations . in particular , the lda is parameterized based on the uniform electron gas at densities typical for atoms and solids . however , the uniform electron gas at these densities is stiff against magnetic degrees of freedom and far from magnetic qcp s .
thus , although the lda is exact for the uniform electron gas , and therefore does include all fluctuation effects there , its description of magnetic ground states in solids and molecules is mean field like .
this leads to problems such as the incorrect description of singlet states in molecules with magnetic ions as well as errors in solids when spin fluctuation effects beyond the mean field are important . in solids
near a qcp , the result is an overestimate of the magnetic moments and tendency toward magnetism ( _ i.e. _ misplacement of the position of the critical point ) due to neglect of the quantum critical fluctuations .
@xcite examples include three types of materials : paramagnets that are ferromagnetic in the lda , ferromagnets where the equilibrium magnetic moment is substantially overestimated in the lda , and paramagnets where the paramagnetic susceptibility is substantially overestimated . lllllll material & zrzn@xmath1 & ni@xmath3al & sc@xmath3 in & feal & ni@xmath3ga + type & 1 & 1 & 1 & 2 & 2 + calc / exp & 0.72/0.17 & 0.71/0.23 & 1.05/0.20 & 0.80 & 0.79 + ref . &
@xcite & @xcite & @xcite & @xcite & @xcite + + material & sr@xmath3ru@xmath1o@xmath4 & srrho@xmath3 & na@xmath5coo@xmath1 & pd + type & 2 & 2 & 2 & 3 + calc / exp & 0.80 & 0.9 & 0.5 & 11.6/6.8 + ref .
& @xcite & @xcite & @xcite & @xcite + [ materials ] we list examples of materials in all three categories in table [ materials ] .
at least two of these are cases where a large deviation between the lda and experimental magnetic properties were noted , followed by transport measurements that suggest a nearby ferromagnetic quantum critical point . in particular , in sr@xmath0ru@xmath6o@xmath7 , lda calculations with the experimental crystal structure found a sizeable moment , @xcite while experimentally the material was known to be a paramagnetic metal .
grigera and co - workers then showed that sr@xmath0ru@xmath6o@xmath7 has a metamagnetic quantum critical point at moderate field .
@xcite pd metal provides another example : the calculated lda magnetic susceptibility is nearly twice larger that the experimental one .
correspondingly , nicklas @xmath8 @xmath9@xcite found a fqcp in the pd@xmath10ni@xmath11 system at @xmath12 , where the transport properties become non - fermi liquid .
we emphasize that substantial overestimates of the tendency of metals towards ferromagnetism within the lda is a rare occurance , and propose that it be used as an indicator of critical fluctuations in a material .
however , for this to be an effective screen , competing states , like antiferromagnetism need to be ruled out in each material .
an interesting case study is liv@xmath6o@xmath13 , which is a paramagnetic metal and occurs in the cubic spinel structure .
remarkably , it was discovered by kondo and co - workers that this material behaves at low temperature like a heavy fermion metal .
@xcite lda calculations showed that the material is unstable against ferromagnetism with a sizeable moment . @xcite but calculations also show that the interactions are antiferromagnetic , and as a result it is more unstable against antiferromagnetism , which however is frustrated on the spinel lattice .
while liv@xmath6o@xmath13 may be near an antiferromagnetic qcp , it is not a material near an fqcp .
a popular way to add quantum or termal fluctuation to a mean - field type theory is @xmath14 fluctuation corrections to ginzburg - landau expansion of the free energy . for a detailed discussion
we refer the reader to the book of moriya @xcite and the review article of shimizu @xcite . in short
, one writes the free energy ( or the magnetic field ) as a function of the ferromagnetic magnetization , @xmath15 @xmath16(obviously , @xmath17 gives the inverse spin susceptibility without fluctuations ) , and then assume gaussian zero - point fluctuations of an r.m.s .
magnitude @xmath18 for each of the @xmath19 components of the magnetic moment ( for a 3d isotropic material like pd , @xmath20 after averaging over the spin fluctuations , one obtains a fluctuation - corrected functional .
the general expression can be written in the following compact form : @xmath21for instance,@xmath22 the unrenormalized coefficients can be taken from fixed spin momen lda calculations , in which case @xmath18 becomes the amplitude of those fluctuations only , which are not taken into account in lda ( as mentioned , lda includes some quantum fluctuation , specifically short - range fluctuations present in the interacting uniform electron gas ) . in principle
, one can estimate @xmath18 from the fluctuation - dissipation theorem , which states that ( see , e.g. , refs .
@xcite ) @xmath23where @xmath24 is the magnetic susceptibility and @xmath25 is the brillouin zone volume .
it is customary to approximate @xmath24 by its small @xmath26 small @xmath27 expansion @xcite : @xmath28 with the expansion ( [ ab ] ) the integrations can be performed analytically , and the final result reads : @xmath29.\ ] ] where @xmath30 and @xmath31 is the cutoff parameter for momentum integration in eqn .
[ mumu ] ( the frequency integration at a given @xmath32 is usually assumed to be cut off at @xmath33 ) . to proceed along these lines
one needs to find a way to calculate the crucial parameters of the expansion ( [ ab ] ) .
it was suggested by moriya @xcite that these can be expressed as certain integrals over the fermi surface , by expanding the rpa expression for @xmath34 .
below , we offer a derivation equivalent to that of moriya , but rendering the results in more computable form .
we start with the rpa expressions for the real and imaginary parts of @xmath35 @xmath36 ( e_{\mathbf{k+q}}-e_{\mathbf{k}% } ) ^{-1 } \label{re } \\
\mathop{\rm im}\chi _ { 0}(\mathbf{q,}\omega ) & = & \sum_{\mathbf{k}}[f(e_{% \mathbf{k}})-f(e_{\mathbf{k+q}})]\delta ( e_{\mathbf{k+q}}-e_{\mathbf{k}% } -\omega ) , \label{im}\end{aligned}\]]where @xmath37 is the fermi function , @xmath38 .
expanding eqn . [ re ] in @xmath39
we get to second order in @xmath32@xmath40 .\end{aligned}\]]the odd powers of @xmath41 cancel out and we get ( @xmath42)@xmath43where @xmath44 @xmath45 the last equality assumes cubic symmetry ; generalization to a lower symmetry is trivial .
using the following relation,@xmath46one can prove that@xmath47therefore @xmath48 similarly , for eqn .
[ i m ] one has @xmath49\ ] ] after averaging over the directions of @xmath50 this becomes , for small @xmath51@xmath52although in real materials the fermi velocity is obviously different along different directions , it is still a reasonable approximation to introduce an average @xmath53 .
then the above formulae reduce all parameters needed for estimating the _ r.m.s .
_ amplitude of the spin fluctuations to four integrals over the fermi surface , specifically , the density of states , @xmath54 , @xmath55 , @xmath56 and @xmath57 .
the physical meaning of these parameters is as follows .
@xmath58 defines the rate at which the static susceptibility @xmath59 falls away from the zone center , _
i.e. _ the extent to which the tendency to ferromagnetism is stronger than that to antiferromagnetism .
this translates into the phase space in the brillouin zone where the spin fluctuations are important .
@xmath60 controls the dynamic effects in spin susceptibility .
note that the cutoff parameter @xmath61 remains the only undefined quantity in this formalism .
one obvious choice is @xmath62 @xmath63 because for larger @xmath32 the approximation ( [ ab ] ) gives unphysical negative values for the static susceptibility . on the other hand
, one may argue that @xmath61 should reflect mainly the geometry of the fermi surface and thus not depend on @xmath58 at all .
we will come back to this issue later in this paper and will propose an approach that avoids using @xmath61 whatsoever .
here we use the closely related compounds ni@xmath0al and ni@xmath0ga to illustrate some of the above ideas . further details may be found in ref .
these have the ideal cubic cu@xmath0au @xmath64 structure , with very similar lattice constants , @xmath65 and @xmath66 , respectively , and have been extensively studied by various experimental techniques .
ni@xmath0al is a weak itinerant ferromagnet , @xmath67 = 41.5 k and magnetization , @xmath68=0.23 @xmath69/cell ( 0.077 @xmath69/ni atom ) @xcite with a qcp under pressure at @xmath70=8.1 gpa , @xcite while ni@xmath0ga is a strongly renormalized paramagnet . @xcite
further , it was recently reported that ni@xmath0al shows non - fermi liquid transport over a large range of @xmath71 and @xmath72 range down to very low @xmath72 .
@xcite al ( solid lines ) and ni@xmath3ga ( dotted lines ) .
@xmath73 is at 0 ev .
, title="fig:",height=151 ] 1.2 cm al ( solid lines ) and ni@xmath3ga ( dotted lines ) .
@xmath73 is at 0 ev .
, title="fig:",height=151 ] previous lda calculations showed that the magnetic tendency of both materials is overestimated within the lda , and that ni@xmath0ga is incorrectly predicted to be a ferromagnet .
@xcite moreover , in the lda the tendency to magnetism is stronger in ni@xmath0ga than ni@xmath0al , _ opposite to the experimental trend_. this poses an additional challenge to any theory striving to describe the material dependent aspects of quantum criticality .
the two materials are expected to be very similar electronically ( the small difference between the two is due to relativistic effects associated with ga in ni@xmath0ga ) .
thus these two very similar metals offer a very useful and sensitive benchmark for theoretical approaches .
we use this to test an approach based on the fluctuation dissipation theorem applied to the lda band structures with an ansatz for the cut - off @xmath61 .
we find that this approach corrects the ordering of the magnetic tendencies of the materials , and gives the right ground states at ambient pressure as well as a reasonable value of @xmath70 for ni@xmath0al .
the lda calculations were done using the general potential linearized augmented planewave ( lapw ) method with local orbital extensions @xcite as decribed in ref .
@xcite , with the exchange - correlation functional of hedin and lundqvist with the von barth - hedin spin scaling @xcite .
the lda electronic structure is given in fig .
[ bands ] and table 2 , while results of fixed spin moment calculations of the magnetic properties at the experimental lattice parameters and under hydrostatic compression are given in figs .
[ fsm ] and [ p ] .
the two compounds are very similar in both electronic and magnetic properties , the main apparent difference being the higher equilibrium moment of ni@xmath0ga ( 0.79 @xmath74/f.u .
_ vs. _ 0.71 @xmath74/f.u . )
, in agreement with other full potential calculations .
@xcite the propensity towards magnetism may be described in terms of the stoner criterion , @xmath75 , where @xmath76 is the so - called stoner parameter , which derives from hund s rule coupling on the atoms . for finite magnetizations ,
the so - called extended stoner model @xcite , states that , to the second order in the spin density , the magnetic stabilization energy is given by @xmath77 , \label{estoner}\ ] ] where @xmath78 is the density of states averaged over the exchange splitting corresponding to the magnetization @xmath79 fitting the fixed spin moment results to this expression , we find @xmath80 ev and @xmath81 ev .
these gives @xmath821.21 and @xmath83 = 1.25 for ni@xmath3al and ni@xmath3ga , respectively .
both numbers are larger than unity , corresponding to a ferromagnetic instability , and the value for ni@xmath3ga is larger than that for ni@xmath3al .
importantly , the difference comes from the density of states , since @xmath84 . in both compounds ,
magnetism is suppressed by compression , with an lda critical point at a value @xmath85 -0.05 -0.06 . in ni@xmath0al , the critical point at @xmath86
= -0.058 corresponds to the pressure of @xmath8750 gpa , @xcite which is much higher than the experimental value .
it is interesting that , as in zrzn@xmath6 @xcite , the exchange splitting is very strongly * k*-dependent ; for instance , in ni@xmath0al at some points it is as small as 40 mev/@xmath74 near the fermi level , while at the others ( of pure ni d character ) it is close to 220 mev/@xmath74 . notwithstanding
the general similarity of the two compounds , there is one important difference near the fermi level , specifically , the light band crossing the fermi level in the middle of the @xmath88-m or @xmath88-x directions is steeper in ni@xmath3al ( fig .
[ bands ] ) .
this , in turn , leads to smaller density of states .
this comes from a different position of the top of this band at the @xmath89 point , 0.56 ev in ni@xmath3ga and 0.85 ev in ni@xmath3al .
the corresponding electronic state is a mixture of ni @xmath90 and al ( ga ) @xmath90 states , and is the only state near the fermi level with substantial al ( ga ) content . due to relativistic effects , the ga @xmath90 level is lower than the al @xmath91 level and this leads to the difference in the position of the corresponding hybridized state .
note that this is a purely scalar relativistic effect .
including spin orbit does not produce any further discernible difference .
al and ni@xmath0ga at the experimental lattice parameters .
the energy zero is set to the non - spin - polarized value.,width=240 ] returning to magnetism , the fixed spin moment calculations provide the energy @xmath92 as a function of the magnetization @xmath68 ( fig . [ fsm ] ) .
one can write a landau expansion for @xmath93 as in eqn .
[ eexp ] , which may then be treated as a mean field expression adding the effects of spin fluctuations .
@xcite treating this as a mean field expression and adding the effects of spin fluctuations @xcite leads to renormalization of the expansion coefficients . the renormalized coefficients @xmath94 are written as power series in the averaged square of the magnetic moment fluctuations beyond the lda , @xmath95 as in eqn .
[ renorm ] .
@xmath18 may then be estimated by requiring that the corrected landau functional reproduces the experimental magnetic moment ( for ni@xmath3al ) or experimental magnetic susceptibility ( for ni@xmath3ga ) .
experimental " @xmath18 s obtained in this manner are are 0.47 and 0.55 , respectively , which implies that spin fluctuation effects must be stronger in ni@xmath3ga than in ni@xmath3al .
al and ni@xmath0ga , respectively.,width=384 ] .magnetic energy ( see text ) , magnetic moment in @xmath96/cell and @xmath54 in ev@xmath97 for ni@xmath0al and ni@xmath0ga on a per spin per formula unit basis . [ cols="^,^,^,^,^",options="header " , ] [ tab : table1 ] a link can now be made between this fact and the electronic structures , using the formalism outlined in the previous section .
as discussed , the cutoff parameter @xmath31 is the least well defined quantity in this formalism .
furthermore , the fermiology of these compounds is very complicated : in the paramagnetic state , there are four fermi surfaces , two small and two large ( one open and one closed ) . in this situation , it is hardly possible to justify any simple prescription for @xmath61 .
therefore , we chose a different route : we assume that @xmath61 is the _ same _ for both materials , and choose a number which yields a good description of both the equilibrium moment in ni@xmath0al and the paramagnetic susceptibility in ni@xmath0ga , @xmath98 @xmath99
. note that this is larger that the diameters of the small fermi surfaces but smaller than the radius of the brillouin zone , @xmath100 @xmath99 . to calculate the above quantities , especially @xmath101 we need accurate values of the velocities on a fine mesh .
numerical differentiation of energies within the tetrahedron method proved to be too noisy .
therefore we use the velocities obtained analytically as matrix elements of the momentum operator , computed within the _ optic _ program of the wien package . a bootstrap method , @xcite as described in ref .
@xcite , was used to obtain stable values for @xmath102 .
we found for ni@xmath3al , using as the energy unit ry , the length unit bohr , and the velocity unit ry@xmath103bohr , @xmath104 , @xmath105 , @xmath106 , and @xmath107 @xmath74 . for ni@xmath3ga @xmath108 , @xmath109 , @xmath110 , and @xmath111 @xmath74 . using the resulting values of @xmath18 each compound
we obtain a magnetic moment of @xmath112 @xmath113/cell for ni@xmath114al and a paramagnetic state with the renormalized susceptibility @xmath115 emu / g for ni@xmath3ga , thus correcting the incorrect ordering of the magnetic tendencies of these two compounds and reproducing extremely well the experimental numbers of @xmath116 @xmath74 , @xmath117 emu / g , respectively .
this qualitative behavior is due to the different coefficient @xmath101 _ i.e. _ , different @xmath32 dependencies of @xmath118 at small @xmath32 , which relates to the phase space available for soft fluctuations .
now we turn to the pressure dependence .
the above results imply that beyond - lda fluctuations are already larger than the moments themselves at @xmath119 . in this regime
, we may assume that the size of the beyond - lda fluctuations is only weakly pressure dependent .
then we can apply eqn .
[ renorm ] to the data shown in fig .
[ p ] using @xmath120 as needed to match the @xmath119 value of @xmath68 .
this yields a value @xmath70=10 gpa in quite good agreement with the experimental value , @xmath70=8.1 gpa .
the results for ni@xmath0al and ni@xmath0ga , discussed above , and in ref .
@xcite , show that an approach based on correction of the lda using the fluctuation dissipation theorem has promise .
however , the results hinge on an unknown cut - off , which serves the purpose of including fluctuations that are associated with the fqcp and are not included in the lda , from those that are included in the lda . while it is apparently possible to obtain useful results using reasonable ansatz for this cut - off
, it would be much better to have a truly first principles theory , with no parameters . in order to construct such a theory
, one should find a way of solving the double counting problem , _
i.e _ including in the correction only those fluctuations that are not already taken into acount at the lda level .
this amounts to subtracting from eqn .
[ mumu ] the fluctuations already included in the lda .
since the lda is known to work well for materials far from an fqcp , this means that the correction should be zero or close to it for the most materials .
we suggest that a consistent way to accomplish this is by introducing a reference " susceptibility @xmath121 and subtracting it from @xmath122 @xmath123 , \label{mumuref}\ ] ] we shall use the same expansion[ab ] for both @xmath24 and @xmath124 , to derive equivalent expansions @xmath125where @xmath126 ( density of states per spin ) is the bare ( noninteracting ) static uniform susceptibility , and the stoner parameter @xmath76 is only weakly dependent on * q * and @xmath27 .
note that @xmath127 @xmath128 where @xmath58 and @xmath60 are the coefficients introduced in eq.[ab ] .
we also introduce a notation , @xmath129 as long the same functional form ( [ expchi ] ) is used for @xmath24 and @xmath130 the condition for the convergence of the integral ( [ mumuref ] ) is that the coefficients @xmath131 and @xmath132 controlling the short - range and high frequency fluctuations are the same . of course , the parameter @xmath133 , defining the proximity to the qcp , is different in the reference system , which like the uniform electron gas upon which the lda is based , should be far from any qcp ( let us call @xmath133 for the reference system @xmath134 ) . to calculate the integral ( ( [ mumuref ] ) , we write it in the following form:@xmath135.\]]for instance , @xmath136 represents the susceptibility right at the fqcp .
this diverges for @xmath137 , @xmath138 .
the derivation then proceeds as follows : @xmath139=\frac{q}{2b}% \ln [ \frac{(\delta + aq^{2})^{2}+b^{2}\omega _ { c}^{2}/q^{2}}{(\delta + aq^{2})^{2}}].\ ] ] where we introduce the landau cutoff frequency , @xmath140 ( here @xmath141 is an average fermi velocity ) and the notation @xmath142 .
we will also need the following function:@xmath143 \nonumber\\ & = & \frac{(\delta + x^{2})^{2}+\beta ^{2}}{4}\{\ln [ ( \delta + x^{2})^{2}+\beta ^{2}]-1\ } \nonumber \\ & -&\frac{\delta ( \delta + x^{2})}{2}\{\ln [ ( \delta + x^{2})^{2}+\beta ^{2}]-2\}+\beta \delta \tan ^{-1}\frac{\beta } { \delta + x^{2}\ } \nonumber\end{aligned}\]]now @xmath144.\end{aligned}\]]this is particularly easy to evaluate at @xmath145 . the result is@xmath146 \label{xi } \\ & & \xi ^{2}(s_{0 } ) = \frac{n^{2}bv_{f}^{2}}{2\omega a^{2}}[4s_{0}\tan ^{-1}(s_{0}^{-1})+\ln ( 1+s_{0}^{2})-s_{0}^{2}\ln ( 1+s_{0}^{-2 } ) ] , \label{xi2}\end{aligned}\]]where @xmath147
. obviously , for arbitrary @xmath148 the answer is simply@xmath149given that usually the reason for a quantum criticality is a large density of states , it makes sense to take the stoner parameter for the reference system the same as for the system in question .
the point is that the density of states is a highly non - local parameter ( note that it involves a delta function integral in energy ) , which can hardly be discerned from local information about the charge density , while the stoner parameter is a very local quantity associated with the exchange - correlation potential .
the difference between @xmath148 and @xmath134 then comes from the difference between @xmath150 and the density of states , @xmath151 of the reference system .
one may think about several different ways for choosing @xmath152 one may be to take average @xmath153 over the width of the valence band , @xmath154 where @xmath155 is the total number of states in the band and @xmath156 is its width .
one can also think about the density of states of the uniform electron gas with the same stoner parameter .
there may be other , more sophisticated prescriptions .
probably , the most practical approach will be found after several trial and error tests with real materials .
the failure of the usual approximations to density functional theory , for example , the lda , to describe the magnetic properties of materials near ferromagnetic quantum critical points is associated with renormalization due to critical fluctuations .
it is pointed out that since such fluctuations are invariably antagonistic to ferromagnetic ordering , deviations between experiment and lda calculations in which the lda is overly ferromagnetic can be a useful screen for materials near fqcps .
these errors in the lda can be corrected using a phenomenalogical landau function approach with the fluctuation amplitude as a parameter . however , there is hope that this parameter can be obtained from the electronic structure via the fluctuation dissipation theorem and a suitable reference system .
the key remaining challenges in our view are to define the reference system to be used , and to use calculations to determine the usefulness of this approach for real materials near a critical point .
saxena , p. agarwal , k. ahilan , f.m .
grosche , r.k.w .
haselwimmer , m.j .
steiner , e. pugh , i.r .
walker , s.r .
julian , p. monthoux , g.g .
lonzarich , a. huxley , i. sheikin , d. braithwaite , and j. flouquet , nature * 406 * , 587 ( 2000 ) .
s. kondo , d.c .
johnston , c.a .
swenson , f. borsa , a.v .
mahajan , l.l .
miller , t. gu , a.i .
goldman , m.b .
maple , d.a .
gajewski , e.j .
freeman , n.r .
dilley , r.p .
dickey , j. merrin , k. kojima , g.m .
luke , y.j .
uemura , o. chmaissem and j.d .
jorgensen , phys .
lett . * 78 * , 3729 ( 1997 ) .
niklowitz , f. beckers , n. bernhoeft , d. braithwaite , g. knebel , b. salce , j. thomasson , j. floquet and g.g .
lonzarich ( unpublished ) ; presented at conference on quantum complexities in condensed matter , 2003 .
p. blaha , k. schwarz g.k.h .
madsen , d. kvasnicka , and j. luitz , _
wien2k , an augmented plane wave + local orbitals program for for calculating crystal properties _ ( k. schwarz , techn .
universitat wien , austria , 2001 ) , isbn 3 - 9501031 - 1 - 2 . to compute pressure , we used @xmath157 $ ] , where @xmath158 is the volume compression , @xmath159 and @xmath160 are the bulk modulus and its derivative .
we used the experimental bulk modulus of ni@xmath0al , @xmath159=174 gpa .
@xcite for @xmath160 we used the calculated value @xmath160=5.2 . | we discuss the application of the density functional theory in the local density approximation ( lda ) near a ferromagnetic quantum critical point .
the lda fails to describe the critical fluctuations in this regime .
this provides a fingerprint of a materials near ferromagnetic quantum critical points : overestimation of the tendency to magnetism in the local density approximation .
this is in contrast to the typical , but not universal , tendency of the lda to underestimate the tendency to magnetism in strongly hubbard correlated materials .
we propose a method for correcting the local density calculations by including critical spin fluctuations .
this is based on ( 1 ) landau expansion for the free energy , evaluated within the lda , ( 2 ) lowest order expansion of the rpa susceptibility in lda and ( 3 ) extraction of the amplitude of the relevant ( critical ) fluctuations by applying the fluctuation - dissipation theorem to the difference between a quantum - critical system and a reference system removed from the quantum critical point .
we illustrate some of the aspects of this by the cases of ni@xmath0al and ni@xmath0ga , which are very similar metals on opposite sides of a ferromagnetic quantum critical point .
lda calculations predict that ni@xmath0ga is the more magnetic system , but we find that due to differences in the band structure , fluctuation effects are larger in ni@xmath0ga , explaining the fact that experimentally it is the less magnetic of the two materials .
quantum criticality , magnetism , density functional theory , first - principles calculation . |
let @xmath6 be a subdivision of a region in @xmath7 by convex polytopes .
@xmath8 denotes the set of piecewise polynomial functions ( splines ) on @xmath6 that are continuously differentiable of order @xmath9 .
splines are a fundamental tool in approximation theory and numerical analysis @xcite ; more recently they have also appeared in a geometric context , describing the equivariant cohomology ring of toric varieties @xcite .
practical applications include surface modelling , computer - aided design , and computer graphics @xcite .
one of the fundamental questions in spline theory is to determine the dimension of the space @xmath10 of splines of degree at most @xmath2 . in the bivariate , simplicial case ,
these questions are studied by alfeld - schumaker in @xcite and @xcite using bernstein - bezier methods .
a signature result in @xcite is a formula for @xmath11 when @xmath12 and @xmath13 is a generic simplicial complex . for @xmath13 simplicial and nongeneric , hong @xcite and ibrahim - schumaker @xcite derive a formula for @xmath11 when @xmath14 as a byproduct of constructing local bases for these spaces .
an algebraic approach to the dimension question was pioneered by billera in @xcite using homological and commutative algebra . in @xcite ,
billera - rose show that @xmath15 , the @xmath2th graded piece of the algebra @xmath16 of splines on the cone @xmath17 over @xmath6 .
the function @xmath18 is known as the _ hilbert function _ of @xmath16 in commutative algebra , and a standard result is that the values of the hilbert function eventually agree with the _ hilbert polynomial _
@xmath19 of @xmath16 .
an important invariant of @xmath16 is the _ postulation number _ @xmath20 , which is the largest integer @xmath2 so that @xmath21 . in this terminology
the alfeld - schumaker result above could be viewed as a computation of @xmath22 plus the bound @xmath23 .
the goal of this paper is to provide upper bounds on the postulation number @xmath24 for _ central _ polytopal complexes @xmath25 , where @xmath0 is the algebra of _ mixed splines _ over @xmath6 .
central _ polytopal complex is one in which the intersection of all interior faces is nonempty ; if @xmath6 is central then splines on @xmath6 are a graded algebra . _ mixed splines _ are splines in which different smoothness conditions are imposed across codimension one faces .
the main reason for bounding @xmath24 is that the hilbert polynomial of @xmath0 has been computed in situations where there are no known bounds on @xmath24 , rendering these formulas impractical .
currently , bounds which do not make heavy restrictions on the complex @xmath6 are known only in the simplicial case .
these bounds are recorded in table [ tbl : postbounds ] .
for particular types of complexes @xmath6 , better and sometimes exact bounds are known for @xmath26 . bound &
context & computed by + @xmath23 & _ generic _ simplicial @xmath13 & alfeld - schumaker @xcite + @xmath27 & _ all _ simplicial @xmath13 & hong @xcite + & & ibrahim - schumaker @xcite + @xmath28 & _ all _ simplicial @xmath13 & alfeld - piper - schumaker @xcite + @xmath29 & _ generic _ simplicial @xmath30 & alfeld - schumaker - whiteley @xcite + bound & context & computed by + @xmath31 & _ all _ simplicial @xmath13 & mourrain - villamizar @xcite + @xmath32 & _ generic _ simplicial @xmath13 & billera @xcite + in contrast , the hilbert polynomial @xmath3 has been computed for _ all _ central polytopal complexes @xmath1 . this is done in the simplicial case with mixed smoothness by schenck - geramita @xcite , in the polytopal case with uniform smoothness by schenck - mcdonald @xcite , and in the polytopal case with mixed smoothness and boundary conditions , by the author @xcite . in this paper
we provide the first bound on @xmath24 for all central polytopal complexes @xmath1 . specifically , given _
smoothness parameters _
@xmath33 associated to each codimension one face @xmath34 , our first result is the following .
* theorem [ thm : main2 ] * let @xmath1 be a central , pure , hereditary three - dimensional polytopal complex .
set @xmath35 where @xmath36 denotes the star of @xmath37 and @xmath38 denotes the @xmath39-faces of @xmath36 .
then @xmath40 in particular , @xmath41 for @xmath42 . from an algebraic perspective ,
another reason for bounding @xmath24 is that almost all existing bounds , including most in table [ tbl : postbounds ] , have been computed using analytic techniques .
there are a few instances where algebraic techniques are applied to bound @xmath24 . in @xcite , billera proves @xmath32 for _ generic _ simplicial complexes ( this result relies on a computation of whiteley @xcite ) .
the most general bound produced by homological techniques to date is by mourrain - villamizar @xcite ; building on work of schenck - stillman @xcite they prove that @xmath31 for @xmath43 a planar simplicial complex , recovering an earlier result of alfeld - schumaker @xcite .
our second result is the following .
* theorem [ thm : main3 ] * let @xmath30 be a central , pure , hereditary three - dimensional simplicial complex .
for a @xmath39-face @xmath44 , set @xmath45 then @xmath46 in particular , @xmath47 for @xmath48 .
setting @xmath49 for all @xmath44 , we recover that @xmath50 for @xmath14 .
this was originally proved via constructing local bases by hong @xcite and ibrahim - schumaker @xcite , and is the best bound valid for all planar simplicial complexes recorded in table [ tbl : postbounds ] .
a key tool we use to prove these results is the _ castelnuovo - mumford regularity _ of @xmath0 , denoted @xmath51 .
the relationship between @xmath51 and @xmath24 is discussed in detail in [ sec : regularity ] .
this invariant is also used by schenck - stiller in @xcite .
our particular way of using regularity is inspired by an observation used in the gruson - lazarsfeld - peskine theorem bounding the regularity of curves in projective space . in the context of splines
this observation is roughly that , if we are lucky , we can bound @xmath51 by the regularity of a ` bad ' approximation .
this statement is made precise in proposition [ prop : regdepth ] and theorem [ thm : main0 ] .
we take as our approximation certain locally - supported subalgebras of splines introduced in @xcite .
this could be viewed as an algebraic analogue of locally - supported bases used in @xcite .
the paper is organized as follows . in [ sec : prelim ] we give some background on the spline algebra @xmath0 , in particular the algebraic approach pioneered by billera @xcite and billera and rose @xcite .
we recall the construction of lattice - supported splines @xmath52 introduced in @xcite
. these will provide approximations to @xmath0 . in [ sec : int ] and [ sec : splineint ] we fit lattice - supported splines into a cech - like complex . in [ sec : regularity ] we recall the definition of the regularity of a graded module and prove proposition [ prop : regdepth ] , which is our main tool for bounding regularity . in [ sec : lowpd ] we prove our main results for bounding regularity of spline modules of low projective dimension and prove theorem [ thm : main2 ] bounding the regularity of @xmath0 where @xmath1 is a central polytopal complex . in
[ sec : sstar ] we build on work of tohaneanu - minac @xcite and prove the more precise regularity estimate for central simplicial complexes @xmath30 in theorem [ thm : main3 ] .
we close in [
sec : conjectures ] with conjectured regularity bounds generalizing those derived in this paper .
the two following examples illustrate our results .
let @xmath53 be a subdivision of a topological @xmath39-disk by convex polytopes .
@xmath8 is the algebra of @xmath9-splines on @xmath6 , where @xmath49 for every interior edge and @xmath54 for every boundary edge . by corollary 3.14 of @xcite ,
the hilbert polynomial of @xmath16 is @xmath55 where @xmath56 are the number of @xmath57-faces and interior @xmath57-faces of @xmath6 , @xmath9 is the smoothness parameter , and the constants @xmath58 record the dimension of certain vector spaces coming from ideals of powers of linear forms .
[ ex : hpolytopal ] the complex @xmath59 in figure [ fig : hpolytopal ] has @xmath60 .
it is shown in 4 of @xcite that there are @xmath61 constants @xmath58 in the formula , and they are all equal to the constant @xmath62 hence by equation , @xmath63 ( -1,-1 )
( 1,-1 ) (0,1 ) node [ right ] @xmath64 (0,3 ) node [ above ] @xmath65(-3,-2 ) node [ below ] @xmath66
( 3,-2 ) node [ below ] @xmath67 ; at ( -1.6,-.8 ) @xmath68 ; at ( 1.5,-.8 ) @xmath69 ; ( 1,-1)(3,-2 ) ( 0,3 ) ; ( -3,-2)(-1,-1)(0,1 ) ; ( -1,-1 ) circle [ radius=.05 ] ; ( -3,-2 ) circle [ radius=.05 ] ; ( 1,-1 ) circle [ radius=.05 ] ; ( 3,-2 ) circle [ radius=.05 ] ; ( 0,1 ) circle [ radius=.05 ] ; ( 0,3 ) circle [ radius=.05 ] ; by theorem [ thm : main2 ] , @xmath70 , where @xmath71 the star of each interior edge of @xmath59 has @xmath72 edges which are interior .
so @xmath73 and the hilbert function @xmath74 agrees with the hilbert polynomial @xmath75 above for @xmath76 .
computations in macaulay2 @xcite suggest that @xmath77 ( in fact the behavior is the same as example [ ex : schlegelcube ] in [ sec : examples ] ) , indicating that there is room for improvement in theorem [ thm : main2 ] . in ( * ?
* theorem 4.3 ) , geramita and schenck give a formula for @xmath78 , where @xmath43 is a planar simplicial complex and @xmath79 is the vector of smoothness parameters associated to codimension one faces .
[ ex : hsimplicial ] triangulate the polytopal complex @xmath59 in example [ ex : hpolytopal ] to obtain the simplicial complex @xmath43 below , with @xmath80 and @xmath81 .
take smoothness parameters @xmath82 on the edges of the center triangle and @xmath83 on the six edges which connect interior vertices to boundary vertices .
( -1,-1 ) ( 1,-1 ) (0,1 ) (0,3)(-3,-2 ) ( 3,-2 ) ; ( 1,-1)(3,-2 ) ( 0,3 ) ; ( -3,-2)(-1,-1)(0,1 ) ; ( -1,-1)(3,-2 ) ; ( -3,-2)(0,1 ) ; ( 0,3)(1,-1 ) ; ( -1,-1 ) circle [ radius=.05 ] ; ( -3,-2 ) circle [ radius=.05 ] ; ( 1,-1 ) circle [ radius=.05 ] ; ( 3,-2 ) circle [ radius=.05 ] ; ( 0,1 ) circle [ radius=.05 ] ; ( 0,3 ) circle [ radius=.05 ] ; in example 4.5 of @xcite , schenck and geramita compute @xmath84 by theorem [ thm : main3 ] , @xmath85 , where @xmath86 .
this yields @xmath87 , so the polynomial above gives the correct dimension of @xmath88 for @xmath89 .
macaulay2 gives @xmath90 , so the formula is actually correct for @xmath91 .
we begin with some preliminary notions . a _
polytopal complex _
@xmath92 is a finite set of convex polytopes ( called _ faces _ of @xmath6 ) in @xmath7 such that * if @xmath93 , then all faces of @xmath94 are in @xmath6 . * if @xmath95 then @xmath96 is a face of both @xmath94 and @xmath37 ( possibly empty ) .
the _ dimension _ of @xmath6 is the greatest dimension of a face of @xmath6 .
the faces of @xmath6 are ordered via inclusion ; a maximal face of @xmath6 is called a _ facet _ of @xmath6 , and @xmath6
is said to be _ pure _ if all facets are equidimensional .
@xmath97 denotes the underlying space of @xmath6 .
@xmath98 and @xmath99 denote the set of @xmath57-faces and the set of interior @xmath57-faces , respectively , while @xmath100 and @xmath101 . in the case
that all facets of @xmath6 are simplices , @xmath6 is a simplicial complex and will be denoted by @xmath43 .
the boundary of @xmath6 , denoted @xmath102 , is a polytopal complex of dimension @xmath103 , and is pure if @xmath6 is pure . given a complex @xmath6 and a face @xmath93 , the _ star _ of @xmath94 in @xmath6 , denoted @xmath104 , is defined by @xmath105 this is the smallest subcomplex of @xmath6 which contains all faces which contain @xmath94 .
if the complex @xmath6 is understood we will write @xmath106 . for @xmath92 , we define @xmath107 to be the graph with a vertex for every facet ( element of @xmath108 ) ; two vertices are joined by an edge iff the corresponding facets @xmath109 and @xmath110 satisfy @xmath111 .
@xmath6 is said to be _ hereditary _ if @xmath112 is connected for every nonempty @xmath93 . throughout this paper , @xmath92 is assumed to be a pure , @xmath113-dimensional , hereditary polytopal complex .
let @xmath114 $ ] be the polynomial ring in @xmath113 variables , and @xmath115 @xmath116 $ ] the polynomial ring in @xmath117 variables .
we will typically use @xmath118 in inhomogeneous and @xmath119 in homogeneous situations .
we now recall the definition of the ring of splines @xmath8 . for @xmath120 ,
let @xmath121 denote the set of functions @xmath122 continuously differentiable of order @xmath9 . for @xmath123 a function and @xmath124 , @xmath125 denotes the restriction of @xmath126 to @xmath109 .
the module @xmath8 of piecewise polynomials continuously differentiable of order @xmath9 on @xmath6 is defined by @xmath127 the polynomial ring @xmath118 includes in @xmath8 as globally polynomial functions ( these are the _ trivial _ splines ) ; this makes @xmath8 an @xmath118-algebra via pointwise multiplication .
if @xmath6 is a hereditary complex , the global @xmath128 condition can be expressed as a differentiability condition across internal faces of codimension one . for a codimension one face @xmath129 , let @xmath130 denote a choice of affine form ( unique up to scaling ) which vanishes on @xmath37 .
then a function @xmath123 which restricts to a polynomial on each facet is in @xmath8 iff @xmath131 for every pair of facets @xmath132 which intersect in a codimension one face @xmath37 @xcite . in @xcite , schenck and geramita
study the dimension of _ mixed spline _ spaces , in which the order of differentiability across codimension one faces varies .
specifically , let @xmath133 be a list of _ smoothness parameters _
@xmath33 associated to each codimension one face .
we require that @xmath134 for @xmath129 and @xmath135 for @xmath136 .
then the algebra @xmath0 of mixed splines on @xmath6 is defined as the set of splines @xmath137 satisfying * @xmath138 for @xmath129 with @xmath139 * @xmath140 for @xmath136 with @xmath141 for hereditary complexes , the usual ring of splines @xmath8 is recovered by setting @xmath49 for every @xmath129 and @xmath54 for every @xmath142 . the following variant of ( * ? ? ?
* proposition 4.3 ) encodes the mixed spline conditions as a matrix .
[ lem : seq1 ] if @xmath6 is hereditary and @xmath133 , @xmath0 fits into the graded exact sequence @xmath143 @xmath144 , @xmath145 and the matrix @xmath146 is the top dimensional cellular boundary map of @xmath6 .
since our results apply in the context of mixed splines , we will use these throughout the paper .
let @xmath147 and @xmath148 be the set of polynomials @xmath149 of degree @xmath150 and degree @xmath2 , respectively . for @xmath92
we have a filtration of @xmath0 by @xmath151-vector spaces @xmath152 a polytopal complex @xmath6 is called a _ central _
complex if all interior codimension one faces share a common face
. we will always assume the origin @xmath153 is contained in this common face .
for central complexes we make the assumption that @xmath54 for all codimension one faces @xmath142 so that @xmath154
. then the diagonal portion of the matrix @xmath155 in lemma [ lem : seq1 ] consists of forms of degree @xmath156 and @xmath0 is graded .
the graded pieces are the vector spaces @xmath157 given @xmath92 , the cone @xmath158 over @xmath6 is formed by taking the join of @xmath159 with @xmath160 , where @xmath161 is defined by @xmath162 .
this is clearly a central complex .
if @xmath6 comes with smoothness parameters @xmath79 , extend these to smoothness parameters @xmath163 on @xmath17 by assigning * @xmath164 for @xmath165 which is a cone over @xmath142 and * @xmath166 for @xmath165 so that @xmath167 since this extension is natural we will abuse notation and drop the hat on @xmath79 , denoting @xmath168 by @xmath169 . in practice one computes the algebra @xmath168 by replacing the polynomial ring @xmath118 by @xmath119 in lemma [ lem : seq1 ] and homogenizing the entries of the matrix @xmath155 used to compute @xmath0 .
the following lemma is proved the same way as theorem 2.6 of @xcite .
[ lem : splinehom ] let @xmath92 be a polytopal complex with smoothness parameters @xmath79 .
then @xmath170 as @xmath151-vector spaces . in @xcite
certain subalgebras @xmath171 are constructed as approximations to @xmath8 .
this construction carries over directly to mixed splines ; we will denote the corresponding submodules by @xmath52 .
we briefly summarize the construction . for a pure @xmath113-dimensional subcomplex @xmath172 , not necessarily
hereditary , define @xmath173 let @xmath174 denote the set of faces of @xmath6 which are contained in a codimension one face @xmath37 so that @xmath54 ; this is a subcomplex of @xmath102 .
let @xmath92 be a polytopal complex and @xmath79 a list of smoothness parameters . 1 . for @xmath34 a face
, @xmath175 denotes the affine span of @xmath37 .
2 . @xmath176 denotes the hyperplane arrangement @xmath177 .
3 . @xmath178 denotes the intersection semi - lattice @xmath179 of @xmath176 .
the elements @xmath180 are called _ flats_. these consist of the whole space @xmath7 , the hyperplanes @xmath181 , and all nonempty intersections of these , ordered with respect to reverse inclusion
. the _ rank _ of a flat @xmath182 , denoted @xmath183 , is its codimension as a vector space .
to each flat @xmath180 we associate a _ lattice complex _ @xmath184 as follows .
form a graph @xmath185 whose vertices correspond to facets which have a codimension one face @xmath37 so that @xmath186 .
connect two vertices corresponding to facets @xmath132 if @xmath187 is a codimension one face of both and @xmath188 .
each connected component @xmath189 of @xmath185 is the dual graph of a unique subcomplex @xmath190 .
the lattice complex @xmath184 is defined as the disjoint union of these @xmath190 , which we call components of @xmath184 .
then define @xmath191 the submodule generated by splines which vanish outside a component of @xmath184 .
then @xmath52 is defined by @xmath192 it is equivalent to let the sum in the definition of @xmath52 run across maximal components ( with respect to inclusion ) occuring among lattice complexes @xmath184 with the rank of @xmath182 at most @xmath193 . to make this more precise ,
let @xmath194 be the poset of components of lattice complexes @xmath184 with @xmath195 , ordered with respect to inclusion .
let @xmath196 be the set of maximal subcomplexes appearing in @xmath194 .
then we have ( * ? ? ?
* proposition 4.9)[prop : top ] @xmath197 since we will use this construction primarily in the cases @xmath198 and @xmath199 , we describe @xmath200 and @xmath201 precisely . if @xmath94 is a face of @xmath6 of some dimension , we use @xmath202 and @xmath203 interchangeably to denote the subring of splines which vanish outside of the star of @xmath94 , as long as no confusion results .
so @xmath204 for @xmath124 denotes the subring of splines supported on a single facet , @xmath205 for @xmath129 denotes the ring of splines supported on the two facets of @xmath36 , etc .
[ cor : lowkdescription ] let @xmath92 be a polytopal complex .
then @xmath206 for @xmath198 , the only flat @xmath180 of rank zero is the whole space @xmath7 .
the corresponding lattice complex @xmath207 is the disjoint union of the facets of @xmath6 .
hence @xmath208 for @xmath199 , the flats @xmath180 of rank one are precisely the hyperplanes @xmath175 with @xmath134 , where @xmath142 .
the components of the lattice complex @xmath209 are the complexes @xmath106 for all @xmath94 satisfying @xmath210 . if @xmath211 , then @xmath106 consists of two facets and all their faces ; otherwise @xmath212 and @xmath106 consists of a single facet of @xmath6 and all its faces .
however , as long as @xmath6 is hereditary and has more than one facet , every facet @xmath124 has a codimension one face @xmath94 which is interior .
hence @xmath213 .
it follows that @xmath214 consists of stars of interior codimension one faces of @xmath6 . by proposition
[ prop : top ] we have @xmath215 theorem 4.3 of @xcite makes precise the sense in which @xmath216 is an approximation to @xmath8 . this result and its proof extend directly to mixed splines ,
so we state the result in this context .
[ thm : latticegens ] let @xmath92 be a polytopal complex .
then @xmath52 fits into a short exact sequence @xmath217 where @xmath218 has codimension @xmath219 and the primes in the support of @xmath218 with codimension @xmath220 are contained in the set @xmath221 . to use the submodules @xmath52 effectively
, it will be useful to fit @xmath52 into a chain complex whose pieces are easier to understand . in the next section
we describe such a complex .
in this section we introduce a cech - type complex for finite sums of submodules of a given @xmath119-module @xmath222 and give a criterion for its exactness .
we apply this to the submodules @xmath223 in [ sec : splineint ] . for an integer @xmath224 ,
let @xmath225 be the set of all subsets of size @xmath226 formed from the index set @xmath227 .
thinking of @xmath228 as a @xmath193-simpex of the @xmath224-simplex @xmath43 , we have the complex @xmath229 with @xmath230 below whose homology is the simplicial homology of @xmath43 with coefficients in @xmath119 .
@xmath231 if @xmath232 and @xmath233 is the idempotent corresponding to @xmath234 , then @xmath235 it will be convenient to augment this complex with a final map @xmath236 defined by @xmath237 for every @xmath57 .
we denote this augmented complex as @xmath238 .
the homology of @xmath238 computes the _ reduced _ homology of @xmath43 with coefficients in @xmath119 .
we extend these complexes to an @xmath119-module @xmath222 by tensoring ; let @xmath239 denote @xmath240 and @xmath241 denote @xmath242 .
now suppose @xmath243 , where each @xmath244 is a submodule of @xmath222 . for @xmath245 let @xmath246 denote the intersection @xmath247 .
define submodules @xmath248 .
since @xmath249 for every @xmath250 , the differential @xmath251 of @xmath239 restricts to a map @xmath252 , so @xmath253 is a subcomplex of @xmath239 .
for example , if @xmath254 , @xmath253 is the complex @xmath255 where @xmath256 . given any submodule @xmath257 containing all the @xmath244 , we may augment @xmath253 with the map @xmath258 where @xmath259 .
we denote this augmented complex by @xmath260 .
now consider the condition @xmath261 on @xmath262 given by @xmath263 we only need to check this condition on subsets @xmath264 with @xmath265 , since if there is some @xmath266 then @xmath267 and both sides are equal to @xmath246 .
[ prop : augexact ] if @xmath268 satisfies @xmath261 then @xmath269 for @xmath270 and @xmath271 .
the assertion @xmath271 is always true , so we prove @xmath269 for @xmath270 .
we proceed by induction on the cardinality @xmath224 of @xmath262 . if @xmath254 then @xmath272 is the complex @xmath273 which satisfies the conclusion of proposition [ prop : augexact ] .
now suppose @xmath274 .
let @xmath275 and @xmath276 , where @xmath277 .
we have a short exact sequence of complexes @xmath278 , shown below . here
@xmath279 denotes the complex @xmath218 with shifted grading @xmath280 .
this short exact sequence follows from the fact that @xmath272 can be constructed as the mapping cone of the ( appropriately signed ) inclusion @xmath281 .
it is also not difficult to check exactness of this sequence directly .
& & 0 & 0 & 0 + c^a_(,m ) & 0 & m_1, ,n-1 & _ i=1^n-1 m_i & m + c^a_(,m ) & m_1, ,n & _ ii(n-1 ) m_i & _ i=1^n m_i & m + c^a_(,m_n)(-1 ) & m_1, ,n & _ m_i & m_n & 0 + & 0 & 0 & 0 & clearly @xmath282 satisfies @xmath261 , inheriting the necessary conditions from the fact that @xmath283 satisfies @xmath261 . since @xmath284 , @xmath285 for @xmath270 by induction .
we claim @xmath286 also satisfies @xmath261 .
interpreted for the set @xmath286 , the condition @xmath261 is @xmath287 first note that for any subset @xmath288 , @xmath289 since @xmath283 satisfies @xmath261 .
so the left hand side of @xmath290 is equivalent to @xmath291 .
again , since @xmath283 satisfies @xmath261 , @xmath292 .
so @xmath293 satisfies @xmath261 . since @xmath294 , @xmath295 for @xmath296 by induction
it follows from the long exact sequence in homology that @xmath297 for @xmath296 . for @xmath298
we have the exact sequence 0 & h_1(c^a_(,m ) ) & h_1(c^a_(,m_n)(-1 ) ) & + & h_0(c^a_(,m ) ) & h_0(c^a_(,m ) ) & 0 we have @xmath299 = & \dfrac{m_n}{m_n\cap(\sum_{i=1}^{n-1 } m_i)}\\[15 pt ] = & \dfrac{m_n + \sum_{i=1}^{n-1}m_i}{\sum_{i=1}^{n-1 } m_i } , \end{array}\ ] ] where the second equality comes from the fact that @xmath283 satisfies @xmath261 .
but this is precisely the kernel of the natural surjection @xmath300 it follows that @xmath301 and we are done .
[ cor : exactintc ] let @xmath243 be a set of submodules of @xmath222 .
then if @xmath283 satisfies @xmath261 , @xmath302 is exact .
in this section we apply the complex constructed in [ sec : int ] to the case of lattice - supported splines .
recall from
[ sec : lat ] that @xmath194 is the poset of components appearing among lattice complexes of the form @xmath184 with the rank of @xmath182 at most @xmath193 , ordered with respect to inclusion .
@xmath196 is the set of maximal complexes appearing in @xmath303 . with this notation , proposition [ prop : top ]
states @xmath304 where @xmath305 is the subalgebra of splines vanishing outside of @xmath306 .
now set @xmath307 .
@xmath52 fits into the complex @xmath308 if @xmath309 are pure @xmath113-dimensional subcomplexes of @xmath92 , then @xmath310 this extends to any finite intersection , hence we may write @xmath311 where @xmath59 runs across all intersections of @xmath57 subcomplexes from @xmath312 . as will be evident below , the same subcomplex can appear multiple times as an intersection .
we prove exactness of @xmath313 .
[ prop : fakeres ] let @xmath314 .
then the augmented complex @xmath315 is exact .
we show that @xmath316 satisfies the condition @xmath261 from the previous section ; then by corollary [ cor : exactintc ] the proposition will be proved .
first suppose given @xmath317 codimension one faces @xmath318 . if these are all faces of a common facet @xmath109 , then @xmath319 . otherwise , this intersection has dimension less than @xmath113 and no splines are defined on it .
hence to show @xmath261 for @xmath316 amounts to showing that , given a set @xmath320 of codimension one faces of @xmath6 , the following equalities hold .
keep in mind that for two subcomplexes @xmath321 , @xmath322 . 1 .
for any facet @xmath124 , @xmath323 2 . for any codimension one face @xmath129 ,
@xmath324 \(1 ) if @xmath325 for some @xmath326 , then both sides are equal to @xmath327 . otherwise both sides are trivial .
( 2 ) if @xmath328 , then both sides are equal to @xmath329 .
otherwise , set @xmath330 and let @xmath331 . since @xmath332 , @xmath126 must vanish along @xmath37 to order @xmath33 . letting @xmath132
be the two facets of @xmath36 , we see @xmath333 for @xmath334 .
it follows that @xmath335 now by ( 1 ) the intersections @xmath336 distribute . it would be interesting to know if proposition [ prop : fakeres ] holds for any @xmath337 , where @xmath338 . [
prop : m1res ] let @xmath314 , where @xmath6 is a pure @xmath113-dimensional hereditary polytopal complex . for a facet @xmath124 ,
let @xmath339 denote the set of codimension one faces of @xmath109 that are interior faces of @xmath6 .
set @xmath340 .
the complex @xmath313 satisfies
@xmath341 by definition @xmath342 is the direct sum of all the submodules of @xmath316 .
in general we have @xmath343 where the direct sum runs over all subcomplexes @xmath172 which are intersections of @xmath193 distinct subcomplexes chosen from the set @xmath344 . if @xmath345 then @xmath59 is the intersection of two or more stars of codimension one faces , say @xmath346 .
hence @xmath59 contains at most one facet , and that facet must have @xmath347 as faces .
so if @xmath345 , @xmath348 where @xmath349 is the number of edges of @xmath109 which are interior to @xmath6 . from this
we also see that @xmath350 for @xmath351 .
in this section we briefly summarize some commutative algebra .
the first chapter of @xcite is an excellent introduction to the graded approach we take here .
most of the material of this section comes from this source .
let @xmath222 be a graded module over the polynomial ring @xmath352 $ ] , where @xmath193 is a field of characteristic @xmath353 .
let @xmath354 denote the hilbert function of @xmath222 in degree @xmath2 .
a standard result states that for @xmath355 , @xmath356 agrees with a polynomial function @xmath357 , the hilbert polynomial of @xmath222 . as we noted in the introduction , the largest integer @xmath2 for which @xmath358 is called the _ postulation number _ of @xmath222 , denoted by @xmath359 .
the degree of @xmath357 is one less than the krull dimension of @xmath222 , denoted @xmath360 .
the _ codimension _ of @xmath222 is defined by @xmath361 .
@xmath222 has a minimal graded free resolution @xmath362 with @xmath363 .
the index @xmath364 of the final free module appearing in this resolution is called the projective dimension of @xmath222 , denoted @xmath365 . for an integer @xmath366 , let @xmath367 denote the polynomial ring with grading shifted by @xmath366 , so @xmath368 .
an important invariant of the module @xmath222 which ( among other things ) governs when @xmath356 becomes polynomial is the _ castelnuovo - mumford regularity _ of @xmath222 .
[ defn : regdef ] let @xmath222 be a graded @xmath119 module and @xmath369 the minimal free resolution of @xmath222 , with @xmath370 .
the castelnuovo - mumford regularity of @xmath222 , denoted @xmath371 , is defined by @xmath372 [ rm : degrees ] note that , according to this definition , @xmath371 bounds the minimal degree of generators of @xmath222 as an @xmath119-module . from definition [ defn : regdef ] one
derives the following theorem .
recall an @xmath119-module @xmath222 is _ cohen - macaulay _ if @xmath373 .
* theorem 4.2)[thm : hfhp ] let @xmath222 be a finitely generated graded module over @xmath119
. then 1 .
@xmath374 for @xmath375 .
equivalently , @xmath376 .
2 . if @xmath222 is a cohen - macaulay module , the bound in ( 1 ) is sharp .
another characterization of regularity is obtained via local cohomology , so we introduce this notion .
see ( * ? ? ?
* appendix 1 ) for more details .
let @xmath377 be an ideal of @xmath119 .
the local cohomology modules @xmath378 of @xmath222 with respect to @xmath377 are the right derived functors of the the @xmath377-torsion functor @xmath379 , where @xmath380 we will only be concerned with the case @xmath381 , where @xmath382 is the graded maximal ideal of @xmath119 .
[ thm : cohoreg ] let @xmath383 be the maximal ideal of @xmath119 and @xmath222 a graded @xmath119-module .
then @xmath384 the benefit of this description of regularity is that it interacts well with short exact sequences .
for instance , the following result is a straightforward application of theorem [ thm : cohoreg ] .
* corollary 20.19)[prop : regseq ] let @xmath385 be a graded exact sequence of finitely generated @xmath119 modules
. then 1 .
@xmath386 2 .
@xmath387 3 .
@xmath388 proposition [ prop : regseq ] can be extended to bound the regularity of a module appearing in an exact sequence of any length by breaking the exact sequence into short exact pieces .
we will use the following corollary to proposition [ prop : regseq ] .
[ cor : lesreg ] let @xmath389 and @xmath390 an exact sequence of @xmath119-modules .
then @xmath391 one more concept that is relevant to our situation is that of depth .
the _ depth _ of a graded @xmath119-module @xmath222 with respect to the homogeneous maximal ideal @xmath392 , denoted @xmath393 , is the length of a maximal sequence @xmath394 satisfying that @xmath395 is a non - zerodivisor on @xmath222 and @xmath396 is a non - zerodivisor on @xmath397 for @xmath398 .
such a sequence is called an @xmath222-sequence .
we will use the following result of auslander and buchsbaum to move back and forth between the notions of depth and projective dimension .
[ thm : ab ] let @xmath222 be an @xmath352$]-module . then @xmath399 observe that , according to this formula , @xmath400 .
this inequality is known as the hilbert syzygy theorem .
the following proposition is one of the ingredients used in the proof of the gruson - lazarsfeld - peskine theorem on bounding the regularity of curves in projective space ( * ? ? ?
* proposition 5.5 ) .
it is the main tool we will use for bounding regularity of spline modules .
[ prop : regdepth ] let @xmath222 be an @xmath119-module and @xmath401 a submodule of @xmath222 with @xmath402 , or equivalently @xmath403
. then @xmath404 .
we prove @xmath404 if @xmath402 .
the equivalence of the statements @xmath402 and @xmath403 follows directly from theorem [ thm : ab ] .
set @xmath405 . by (
* proposition a1.16 ) , @xmath406 for @xmath407 and @xmath408 for @xmath409 .
the long exact sequence in local cohomology resulting from the short exact sequence @xmath410 yields a surjection @xmath411 and isomorphisms @xmath412 for @xmath413 . since @xmath406 for @xmath407 , theorem [ thm : cohoreg ] yields @xmath414 .
we give a construction motivating the regularity bounds we derive in corollary [ cor : freegens ] , theorem [ thm : main2 ] , and theorem [ thm : main3 ] .
these results suggest that in general regularity bounds for @xmath0 might be obtained by taking the maximal sum of smoothness parameters @xmath156 appearing in certain subcomplexes of @xmath6 . in the following example , starting with a polytope @xmath415 , we construct a polytopal complex @xmath6 so that @xmath124 and @xmath416 ( @xmath0 if @xmath6 is central ) has a minimal generator supported the facet @xmath417 ( @xmath109 if @xmath6 is central ) .
such generators have degree @xmath418 . since @xmath51 in particular bounds
the degrees of generators of @xmath0 ( see remark [ rm : degrees ] ) , this construction indicates that a bound on @xmath51 will need to be at least as large as the maximal sum of smoothness parameters over codimension one faces occurring in any facet of @xmath6 ( or at least boundary facets - see conjecture [ conj : c1 ] ) .
this example generalizes the construction in ( * ? ?
* theorem 5.7 ) .
for simplicity we restrict the construction to the case of uniform smoothness without imposing boundary vanishing .
the generalization to arbitrary smoothness parameters should be clear .
[ ex : highgen ] suppose that @xmath419 is a polytope with a codimension one face @xmath420 so that @xmath421 is the graph of a piecewise linear function over @xmath37 .
remark [ rm : polygraph ] below shows that this can be accomplished for any polytope by a projective change of coordinates . for instance
this is true if @xmath422 is the join of @xmath37 with the origin @xmath153 .
let @xmath130 be a choice of affine form vanishing on @xmath37 and let @xmath423 be coordinates on @xmath7 .
we further assume that 1 .
@xmath37 is parallel to the coordinate hyperplane @xmath424 2 .
@xmath422 lies between the hyperplanes @xmath424 and @xmath425 .
3 . for any two
codimension one faces @xmath426 ,
@xmath427 and @xmath428 intersect the coordinate hyperplane @xmath424 in distinct linear subspaces of codimension @xmath39 .
\(1 ) can be obtained by rotating the original polytope , ( 2 ) and ( 3 ) can be obtained by translation .
if @xmath422 is the join of @xmath37 with the origin , ( 3 ) may be obtained by slight perturbations of the non - zero vertices of @xmath422 ( within the plane @xmath425 ) .
let @xmath429 be the reflection of @xmath422 across the hyperplane @xmath424 .
for a face @xmath430 , let @xmath431 denote the corresponding face of @xmath429 obtained by reflection .
for @xmath432 , let @xmath433 denote the polytope formed by taking the convex hull of @xmath94 and @xmath431 .
now define @xmath434 as the polytopal complex with facets @xmath435 and @xmath436 .
see figure [ fig : highgens ] for examples of this construction in @xmath437 and @xmath438 .
.45 .45 take the cone @xmath439 over @xmath434 and consider the graded @xmath440$]-module @xmath441 .
let @xmath442 be the @xmath119-linear map obtained by restricting splines @xmath443 to the facet @xmath444 .
this is a splitting of the inclusion @xmath445 as global polynomials on @xmath446 .
let @xmath447 be the kernel of @xmath448
. then @xmath449 let @xmath450 $ ] and , for @xmath451 , set @xmath452 .
define an @xmath119-linear map @xmath453 by @xmath454 where @xmath451 , @xmath455 , and @xmath456 is the restriction of @xmath126 to the facet @xmath457 .
set @xmath458 , where @xmath459 is a choice of homogeneous form vanishing on @xmath460 .
we claim that the image of @xmath155 is the principal ideal @xmath461 @xmath155 is surjective since the spline @xmath462 , defined by @xmath463 goes to the generator of @xmath464 under @xmath155 . to see that @xmath465 , let @xmath455 .
then , since @xmath466 , @xmath467 for every @xmath468
. we also have @xmath469 for every @xmath470 .
hence @xmath471 . but @xmath472 and @xmath473 differ at most by a scalar multiple and a sign on the variable @xmath474 , so @xmath475 and @xmath476 as claimed .
property ( 3 ) above is used in the first equality - this guarantees all the forms @xmath477 are distinct .
it follows that the spline @xmath462 , which is supported only on the facet @xmath457 and generates splines supported on @xmath457 , is a minimal generator of @xmath441 .
if @xmath422 is the join of @xmath37 with @xmath478 , then @xmath434 is central and @xmath479 is graded over the polynomial ring @xmath114 $ ] . in this case
it is unnecessary to take the cone over @xmath434 above .
[ rm : polygraph ] given a convex polytope @xmath480 and a choice @xmath37 of codimension one face , there is a projective change of coordinates which makes @xmath481 into the graph of a piecewise linear function over @xmath37 .
if @xmath422 is the join of @xmath37 with the origin @xmath153 , then this is easily done by a linear transformation .
otherwise , this can be accomplished by choosing a hyperplane @xmath482 which is parallel to @xmath37 and very close to @xmath483 without intersecting @xmath483 .
then make a projective change of coordinates which sends @xmath484 to the hyperplane at infinity ( this argument is due to sergei ivanov ) .
as long as @xmath484 is chosen close enough to @xmath37 , this has the effect of making the face @xmath37 huge and the rest of the polytope the graph of a piecewise linear function over @xmath37 ( once we restrict to affine coordinates again ) .
hence , given any polytope @xmath419 and a choice of codimension one face @xmath420 , the construction in example [ ex : highgen ] allows us to build a polytopal complex @xmath434 so that @xmath485 and the generator of @xmath486 is a minimal generator of @xmath441 .
the construction in example [ ex : highgen ] is inherently nonsimplicial .
some other construction needs to be used to obtain high degree generators in the simplicial case . in the planar simplicial case
, there is an example in @xcite of a planar simplicial complex @xmath43 with minimal generator in degree @xmath487 .
in this section we combine the observations so far to bound the regularity of the spline algebra @xmath0 , where @xmath25 is a central , pure , hereditary , @xmath488-dimensional polytopal complex .
recall a central complex is one in which the intersection of all interior codimension one faces is nonempty .
we assume this intersection contains the origin and that @xmath54 for every codimension one face @xmath489 so that @xmath154 ; this makes the ring @xmath0 a graded @xmath440$]-algebra with respect to the standard grading on @xmath119 .
the following corollary is critical to our analysis .
* proposition 3.4)[cor : splinepd ] if @xmath6 is a central , pure , hereditary , @xmath488-dimensional polytopal complex , then 1 .
@xmath490 2 .
@xmath491 .
\(1 ) follows from lemma [ lem : seq1 ] .
@xmath0 is the kernel of a map between free @xmath119-modules , so it is a second syzygy module . by the hilbert syzygy theorem , any @xmath119-module has projective dimension at most @xmath117 . since @xmath0 is a second syzygy module , @xmath490 .
( 2 ) follows from ( 1 ) and theorem [ thm : hfhp ] .
[ thm : main0 ] let @xmath25 be a pure @xmath488-dimensional hereditary polytopal complex which is central
. then @xmath492 more generally , if @xmath493 , then @xmath494 the first statement follows from the second by corollary [ cor : splinepd ] . to prove the second statement , note that by theorem [ thm : latticegens ] , the cokernel of the inclusion @xmath52 has codimension at least @xmath220 . by proposition [ prop : regdepth ] , @xmath495 . to simplify the statements of later results , we introduce some additional notation .
given a pure @xmath488-dimensional subcomplex @xmath172 , let @xmath496 denote the set of @xmath113 dimensional boundary faces of @xmath59 .
define @xmath497 and set @xmath498 as a first application of theorem [ thm : main0 ] , we give a bound on the degree of generators of @xmath0 when @xmath0 is free .
[ cor : freegens ] suppose @xmath0 is free and set @xmath499
. then @xmath0 is generated in degrees at most @xmath500 .
for a free module , regularity is the maximum degree of generators ( this follows from definition [ defn : regdef ] ) , so we need to show @xmath501 .
@xmath0 is free iff @xmath502 . by theorem [ thm : main0 ] , @xmath503 by corollary [ cor : lowkdescription ] ,
since the support of each summand is disjoint , this is a direct sum , so @xmath505 . also , @xmath204 consists of splines @xmath126 supported on the single facet @xmath109 .
such splines are characterized by @xmath506 being a polynomial multiple of @xmath507 .
it follows that @xmath508 .
hence @xmath509 we now apply theorem [ thm : main0 ] to the case where @xmath0 has projective dimension at most one .
in particular , this includes central complexes in @xmath438 by corollary [ cor : splinepd ] .
[ thm : main1 ] suppose @xmath510 .
let @xmath499 and @xmath511 .
then @xmath512 .
by corollary [ thm : main0 ] , @xmath513 by proposition [ prop : fakeres ] , @xmath514 fits into the exact sequence @xmath515 from proposition [ prop : m1res ] , @xmath516 where @xmath517 . as we saw in the proof of corollary [ cor :
freegens ] , @xmath508 , hence @xmath518 for every @xmath193 with @xmath519 .
now the conclusion follows from corollary [ cor : lesreg ] . at this point
we see that to obtain more precise results for projective dimension one it is necessary to understand the ring @xmath329 of splines vanishing outside the star of a codimension one face .
[ prop : idsyz ] let @xmath520 be an interior codimension one face of @xmath6 , and @xmath132 the two facets of @xmath36 , the star of @xmath37 .
set @xmath521 , @xmath522 .
define the ideal @xmath523 by @xmath524 we have a graded isomorphism @xmath525 where @xmath526 is the module of syzygies on the ideal @xmath523 .
let @xmath527 and set @xmath528 .
then there are polynomials @xmath529 satisfying the following relations .
@xmath530 taking the alternating sum of the above equations yields @xmath531 hence @xmath532 gives rise to a syzygy on the columns of the matrix @xmath533 now suppose given a syzygy @xmath534 on the columns of @xmath222 .
we obtain a spline @xmath527 by setting @xmath535 , hence @xmath329 is isomorphic to the syzygies on the columns @xmath222 . if @xmath523 is minimally generated by @xmath536 and @xmath537 , we obtain @xmath538 .
otherwise we obtain the cases listed above .
for instance , if @xmath539 , then there exist polynomials @xmath540 so that @xmath541 and @xmath542 is generated by @xmath543 of degrees @xmath544 and @xmath545 , respectively .
the other cases follow similarly .
[ prop : edgereg ] let @xmath1 be a central complex , and @xmath546 a codimension one face of @xmath6 . define @xmath547 then @xmath548 unless @xmath549 for all @xmath550 , when @xmath551 .
let @xmath552 be as defined in proposition [ prop : idsyz ] .
then @xmath553 if the ideal @xmath554 is not minimally generated by @xmath555 , and @xmath537 , then @xmath329 is free , generated in degrees indicated by proposition [ prop : idsyz ] . by that description @xmath556
unless @xmath549 for all @xmath550 , when @xmath551 .
so assume @xmath523 is minimally generated by @xmath552 and @xmath557 .
we define a submodule @xmath558 of @xmath559 as follows .
let @xmath132 be the two facets of @xmath36 and @xmath560 the free @xmath119-module on generators @xmath561 corresponding to @xmath132 .
define @xmath558 to be the submodule of @xmath559 generated by @xmath562 , and @xmath563 .
there is a single nontrivial syzygy among @xmath564 given by @xmath565 .
so @xmath558 has minimal free resolution @xmath566 from definition [ defn : regdef ] and the free resolution above we see that @xmath567 .
now we show @xmath568 . it suffices to show that @xmath569 for every prime of codimension one . since @xmath119 is a ufd , primes of codimension one
are principle , generated by a single irreducible polynomial . if @xmath570 for any @xmath571 then @xmath572 if @xmath573 for some @xmath574 , then @xmath575 if @xmath576 meets only one face @xmath577 or @xmath578 if @xmath576 meets both @xmath579 and @xmath580 in a codimension one face .
if @xmath581 , then @xmath582 @xmath583 follows by corollary [ cor : splinepd ] , because we assumed @xmath1 .
since @xmath568 , @xmath584 follows from proposition [ prop : regdepth ] .
[ thm : main2 ] let @xmath1 be a pure @xmath585-dimensional polytopal complex which is central and set @xmath586
. then 1 .
@xmath587 2 .
@xmath588 in particular , @xmath589 for @xmath42 .
\(1 ) follows by applying theorem [ thm : main1 ] to proposition [ prop : edgereg ] .
( 2 ) follows from ( 1 ) by corollary [ cor : splinepd ] .
example [ ex : hpolytopal ] indicates that the bound given in theorem [ thm : main2 ] can be far from optimal . in the next section we bound @xmath590 more precisely for @xmath30 a central simplicial complex .
in this section we analyze the regularity of the ring of splines @xmath591 vanishing outside the star of @xmath39-face , for @xmath30 a pure three - dimensional hereditary simplicial complex which is central .
again we assume @xmath54 for @xmath592 with @xmath154 , so that @xmath593 is a graded module over the polynomial ring @xmath594 $ ] .
this means that @xmath36 has at most five @xmath39-faces @xmath94 for which @xmath595 ( @xmath134 is required ) .
we prove the following theorem .
[ thm : simpedgereg ] let @xmath44 be a @xmath39-face . define @xmath596 then @xmath597 . before proving theorem [ thm : simpedgereg ]
we derive a couple of corollaries .
[ thm : main3 ] let @xmath30 be a pure @xmath585-dimensional hereditary simplicial complex which is central . for @xmath44 , let @xmath598 be defined as in theorem [ thm : simpedgereg ]
. then 1 .
@xmath599 2 .
@xmath600 in particular , @xmath47 for @xmath601 .
\(1 ) follows by applying theorem [ thm : main1 ] to theorem [ thm : simpedgereg ] , ( 2 ) follows by applying theorem [ thm : hfhp ] to ( 1 ) . setting @xmath49 for all @xmath44 , we obtain [ cor : simpreg ] let @xmath30 be a pure @xmath585-dimensional hereditary simplicial complex which is central
. then 1 .
@xmath602 2 .
@xmath603 in particular , @xmath604 for @xmath14 .
this result was obtained in the case of @xmath605 , for simplicial @xmath13 , by hong @xcite and ibrahim and schumaker @xcite ( see table [ tbl : postbounds ] in the introduction ) . before proving theorem
[ thm : main3 ] we set up some notation .
figure [ fig : estar ] depicts our situation .
we will abuse notation and write @xmath606 both for the corresponding edge of @xmath36 and for the vector we obtain by taking positive real multiples of this edge .
= [ fill = gray!50!white , line width=1 pt , join = round ] = [ fill = gray , thick , join = round ] ( v1 ) at ( 4,.4,1 ) ; ( v2 ) at ( .4,4,1 ) ; ( v4 ) at ( 1,2,-3 ) ; ( v3 ) at ( 2,1,5 ) ; ( 0,0,0)node[at end , right]@xmath607(4,0,0 ) ; ( 0,0,0)node[at end , below]@xmath608(0,4,0 ) ; ( 0,0,0)node[at end , right]@xmath609(0,0,4 ) ; ( 0,0,0)node[at end , below]@xmath610 ( 4.4,.44,1.1 ) ; ( 0,0,0)node[at end , below right]@xmath611 ( .44,4.4,1.1 ) ; ( 0,0,0)node[at end , left]@xmath612 ( 2.2,1.1,5.5 ) ; ( 0,0,0)node[at end , left]@xmath613 ( 1.1,2.2,-3.3 ) ; ( v2)(0,0,0)(v4)cycle ; ( 7/15 , 2 , -2/3 ) node @xmath614 ; ( v1)(0,0,0)(v4)cycle ; ( 5/3 , 4/5 , -2/3 ) node @xmath615 ; ( v1)(0,0,0)(v2)cycle ; ( 22/15 , 22/15 , 2/3 ) node @xmath37 ; ( v1)(0,0,0)(v3)cycle ; ( 2 , 7/15 , 2 ) node @xmath616 ; ( v2)(0,0,0)(v3)cycle ; ( 4/5 , 5/3 , 2 ) node @xmath617 ; let @xmath618 be the forms corresponding to the @xmath39-faces @xmath619 , let @xmath620 be the forms corresponding to the @xmath39-faces @xmath621 , and @xmath130 be the form corresponding to @xmath37 ( for now do this without coordinates ) .
let @xmath622 be the exponents to appear on @xmath623 corresponding to the smoothness parameters specified by @xmath79 .
the following lemma is a special case of proposition [ prop : idsyz ] .
[ lem : splineideal ] let @xmath624 .
then we have a graded isomorphism @xmath625 where @xmath526 is the module of syzygies on the ideal @xmath523 .
if @xmath626 or @xmath627 then @xmath597 is clear from lemma [ lem : splineideal ] . if @xmath628 then @xmath629 , @xmath630 , and @xmath597 from lemma [ lem : splineideal ] .
so we may assume @xmath523 is minimally generated by the three given forms and @xmath631 .
in this case @xmath632 by two applications of proposition [ prop : regseq ] ( equality holds but we will not need this ) .
so it suffices to show that @xmath633 .
four special cases are given by 1 .
@xmath634 2 .
@xmath635 3 .
@xmath636 4 .
@xmath637 since @xmath523 is minimally generated by the three given forms , ( * ? ? ?
* theorem 2.7 ) applies in cases ( 1 ) and ( 2 ) .
for example , in case ( 1 ) we have @xmath638 a similar argument holds for case ( 2 ) . in cases ( 3 ) and ( 4 ) , @xmath523 is a complete intersection of its generators and @xmath639 follows from the koszul resolution . if at most one of @xmath640 vanishes we show @xmath633 by fitting @xmath641 into exact sequences and using proposition [ prop : regseq ] .
let @xmath642 .
we have the short exact sequence @xmath643 @xmath377 is a complete intersection with @xmath39 generators in degrees @xmath644 and @xmath645 , so @xmath646 the ideal @xmath377 decomposes as @xmath647 .
then @xmath648 where @xmath649 since @xmath650 is @xmath651-primary and @xmath652 , @xmath653 similarly , @xmath654 from proposition [ prop : describei ] below , if @xmath655 and @xmath656 then @xmath657 are complete intersections and @xmath658 we consider four final special cases before moving on to the general case . a : : @xmath659 b : : @xmath660 note that @xmath661 falls under * a * and @xmath662 falls under * b*. by the exact sequence and proposition [ prop : regseq ] we have the corresponding bounds a : : @xmath663 b : : @xmath664 if we use multiplication by @xmath665 in the exact sequence then we have the corresponding ideals @xmath666,@xmath667 and @xmath668 .
we then have the analogous cases c : : @xmath669 d : : @xmath670 note that @xmath671 falls under * c * and @xmath672 falls under * d*. the corresponding bounds are c : : @xmath673 d : : @xmath674 .
we have reduced to the case where * @xmath675 ( equivalently @xmath676 ) for @xmath334 * @xmath677 ( equivalently @xmath678 ) for @xmath334 * @xmath679 for @xmath334 and @xmath680 .
in particular , @xmath681 implies that the vectors @xmath682 are linearly independent and @xmath683 implies @xmath684 are linearly independent in figure [ fig : estar ] .
it follows that we may make a change of coordinates so that @xmath610 points along the @xmath608-axis , @xmath611 points along the @xmath607-axis , and @xmath612 points along the @xmath609-axis . applying appropriate scaling in the @xmath607 , @xmath608 , and positive @xmath609 directions
, we can assume that the vector defined by @xmath613 points in the direction of @xmath685 . under this change of coordinates , @xmath36 has four possible configurations , shown in figure [ fig : starconfigs ] .
the ideal @xmath523 is the same for all of these .
we have @xmath686 and @xmath687 by corollary [ cor : initial3 ] in the next section , @xmath688 . by the exact sequence and
lemma [ prop : regseq ] , the proof is complete .
= [ fill = gray!50!white , line width=1 pt , join = round ] = [ fill = gray , thick , join = round ] ( -1.2,0,0)node[at end , below]@xmath607(1.2,0,0 ) ; ( 0,-1.2,0)node[at end , below]@xmath608(0,1.2,0 ) ; ( 0,0,-1)node[at end , right]@xmath609(0,0,1.2 ) ; ( 0,1,0)(0,0,0)(0,0,1)cycle ; ( 1,0,0)(0,0,0)(0,0,1)cycle ; ( 1,0,0)(0,0,0)(1,1,-1)cycle ; ( 0,0,0)(0,1,0)(1,1,-1)cycle ; ( 1,0,0)(0,0,0)(0,1,0)cycle ; ( .33,.33,0 ) node@xmath37 ; = [ fill = gray!50!white , line width=1 pt , join = round ] = [ fill = gray , thick , join = round ] ( 0,0,0)(0,-1,0)(1,1,-1)cycle ; ( 1,0,0)(0,0,0)(0,-1,0)cycle ; ( .33,-.38,0 ) node@xmath37 ; ( -1.2,0,0)node[at end , below]@xmath607(1.2,0,0 ) ; ( 0,-1.2,0)node[at end , below]@xmath608(0,1.2,0 ) ; ( 0,0,-1)node[at end , right]@xmath609(0,0,1.2 ) ; ( 0,-1,0)(0,0,0)(0,0,1)cycle ; ( 1,0,0)(0,0,0)(0,0,1)cycle ; ( 1,0,0)(0,0,0)(1,1,-1)cycle ; = [ fill = gray!50!white , line width=1 pt , join = round ] = [ fill = gray , thick , join = round ] ( -1,0,0)(0,0,0)(0,-1,0)cycle ; ( -.33,-.33,0 ) node@xmath37 ; ( 0,0,-1)node[at end , right]@xmath609(0,0,1.2 ) ; ( 0,-1,0)(0,0,0)(0,0,1)cycle ; ( -1,0,0)(0,0,0)(0,0,1)cycle ; ( 0,0,0)(0,-1,0)(1,1,-1)cycle ; ( -1,0,0)(0,0,0)(1,1,-1)cycle ; ( -1.2,0,0)node[at end , below]@xmath607(1.2,0,0 ) ; ( 0,-1.2,0)node[at end , below]@xmath608(0,1.2,0 ) ; = [ fill = gray!50!white , line width=1 pt , join = round ] = [ fill = gray , thick , join = round ] ( -1,0,0)(0,0,0)(1,1,-1)cycle ; ( -1,0,0)(0,0,0)(0,1,0)cycle ; ( -.65,.23,0 ) node@xmath37 ; ( -1.2,0,0)node[at end , below]@xmath607(1.2,0,0 ) ; ( 0,-1.2,0)node[at end , below]@xmath608(0,1.2,0 ) ; ( 0,0,-1)node[at end , right]@xmath609(0,0,1.2 ) ; ( -1,0,0)(0,0,0)(0,0,1)cycle ; ( 0,0,0)(0,1,0)(1,1,-1)cycle ; ( 0,1,0)(0,0,0)(0,0,1)cycle ; let @xmath689 in @xcite , tohaneanu and minac compute the hilbert function of the ideal ( up to change of coordinates ) @xmath690 it is not so obvious how to apply their methods directly to the ideal @xmath691 .
building on their work , however , we show how to construct enough of the initial ideal ( with respect to the lexicographic order ) of @xmath692 to give a fairly tight bound on the socle degree of @xmath693 .
the methods synthesize descriptions of such ideals in terms of linear and commutative algebra .
we first compute the initial ideal of @xmath694 in the ring @xmath695 $ ] with standard lexicographic order .
we assume @xmath696 , so @xmath697 .
this is equivalent to requiring @xmath698 .
[ prop : describei ] let @xmath699 be as above , with @xmath698 . then @xmath464 is a complete intersection generated by two polynomials of 1
. degrees @xmath700 if @xmath701 .
degrees @xmath702 if @xmath703 .
3 . degrees @xmath704
if @xmath705 and @xmath706 . * @xmath707 * in this case @xmath708 .
let @xmath709 for some polynomials @xmath710 , where @xmath711 has no term divisible by @xmath712 .
it is immediate that @xmath713 and @xmath714 the polynomial @xmath711 is not divisible by @xmath715 since it has a term which is a constant multiple of @xmath716 .
it follows that @xmath712 ang @xmath711 are relatively prime and @xmath464 is a complete intersection .
since @xmath711 has degree @xmath717 , ( 1 ) is proved .
* @xmath718 * the argument is identical to the previous case .
* @xmath705 and @xmath706 : * let @xmath719 since we assume @xmath698 as well , @xmath720 is minimally generated by the three given generators .
we describe @xmath464 in terms of the minimal free resolution of the ideal @xmath720 .
set @xmath721 and @xmath722 .
the assumption @xmath698 guarantees that @xmath723 .
@xmath720 is a codimension two cohen - macaulay ideal with hilbert - burch resolution of the form below ( * ? ? ?
* theorem 2.7 ) @xmath724 where @xmath725 is a matrix of forms satisfying @xmath726 .
it follows that the module of syzygies on @xmath720 has two generators , corresponding to the relations @xmath727 and @xmath728 in terms of the entries of the matrix @xmath155 we may write @xmath729 where @xmath730 , and @xmath731 . since @xmath732 and @xmath733 , any common factor of @xmath218 and @xmath126 would give a common factor of @xmath734 and @xmath735 , so @xmath218 and @xmath126 are relatively prime .
so @xmath464 is a complete intersection of the required degrees . as an immediate corollary we have the following lemma .
[ cor : hilbert ] with @xmath736 as above , minimally generated by two forms of degree @xmath737 , we have @xmath738 from proposition [ prop : describei ] , @xmath464 is a complete intersection of polynomials @xmath739 with @xmath730 .
so @xmath464 has minimal resolution of the form @xmath740 the result follows from the additivity of hilbert functions across exact sequences . given a hilbert function @xmath741 , let @xmath742 be the vector space spanned by the @xmath741 greatest monomials of degree @xmath2 with respect to lex order . then the direct sum @xmath743 is an ideal , known as the _ lex - segment _
ideal for the hilbert function @xmath744 ( * ? ? ?
* proposition 2.21 ) . since two _ generic _ forms in @xmath745 $ ] of degrees @xmath737 form a complete intersection , the ideal they generate has the same hilbert function as @xmath464 .
denote by @xmath746 the corresponding lex - segment ideal .
we will show that @xmath746 is the initial ideal of @xmath747 .
to prove this we will use a matrix condition on the coefficients of a form @xmath748 of degree @xmath2 which distinguishes when @xmath749 . from corollary [ cor : hilbert ] ,
@xmath750 for @xmath751 . since @xmath731
, this matrix condition we derive will be nontrivial for @xmath752 .
suppose @xmath753 satisfies @xmath754 .
then by definition we have @xmath755 since the ideal on the right is a monomial ideal , @xmath756 every monomial of @xmath757 is divisible by either @xmath712 or @xmath758 .
expanding @xmath759 and multiplying by @xmath748 gives @xmath760 setting @xmath761 and @xmath762 gives @xmath763 @xmath749 iff the only nonzero coefficients in this expression occur when @xmath764 or @xmath765 . since @xmath766 , @xmath765 is equivalent to @xmath767 .
so @xmath749 iff for @xmath768 we have the condition
@xmath769 here we follow the convention that @xmath770 when @xmath771 or @xmath772 .
these fit together into the following matrix condition on the coefficients of @xmath748 : @xmath773 denote the @xmath774 matrix on the left by @xmath775 .
@xmath775 has entries @xmath776 where @xmath777 and @xmath778 . with this choice of indexing
, column @xmath58 of @xmath775 corresponds to the coefficient @xmath779 .
the following lemma is fundamental for understanding @xmath775 .
[ lem : schur ] let @xmath780 be a partition with @xmath220 parts so that @xmath781 .
let @xmath782 be the square matrix with entries @xmath783 for @xmath784 , @xmath785 .
then @xmath782 has nonzero determinant .
this observation is made in @xcite , where it is noted that determinants of such matrices play a role in the representation theory of the special linear group @xmath786 , where @xmath787 is an @xmath9-dimensional vector space .
in particular , if @xmath788 , the conjugate partition to @xmath789 , @xmath790 is the dimension of the weyl module @xmath791 , which is a nontrivial irreducible representation of @xmath786 .
more explicitly , @xmath792 , where @xmath793 is the schur polynomial in @xmath9 variables of the partition @xmath788 .
in particular , @xmath782 has nonzero determinant .
see @xcite and ( * ? ? ?
* appendix a.1 ) for more details .
[ cor : schurinv ] let @xmath794 be the @xmath774 matrix defined as above , @xmath795 a _ nonzero _ entry of @xmath222 , and @xmath193 a nonnegative integer so that @xmath796 and @xmath797
. then 1 .
the @xmath798 submatrix of @xmath222 formed by the entries @xmath799 is invertible .
the rank of @xmath222 is the minimum of the number of nonzero rows of @xmath222 and the number of nonzero columns of @xmath222 .
\(1 ) the submatrix of @xmath794 above has entries @xmath800 for @xmath784,@xmath785 .
since we assume @xmath801 , @xmath802 and the first statement follows from lemma [ lem : schur ] by taking @xmath803 .
\(2 ) observe that either @xmath804 or , if @xmath805 , then the first entry @xmath806 ( @xmath807 ) which is nonzero is equal to @xmath808 .
the last entry in the first column is @xmath809 ( we assumed @xmath810 ) , so there is at least one nonzero entry in the first column of @xmath222 . if the row of @xmath222 with index @xmath57 is nonzero , every row with index @xmath811 is also nonzero . if the column of @xmath222 with index @xmath812 is zero , every column with index @xmath813 is also zero .
now the second statement follows by taking the largest square submatrix of @xmath222 whose upper left corner is the first nonzero entry of the first column of @xmath222 .
this is a @xmath814 submatrix of @xmath222 where @xmath193 is the minimum of the number of nonzero rows of @xmath222 and the number of nonzero columns of @xmath222 . by the first statement
, this submatrix is invertible , and from the earlier observations @xmath193 must be the rank of @xmath222 .
[ lem : initial ] the initial ideal of @xmath736 is the lex - segment ideal @xmath746 , where @xmath737 are the degrees of the generators of @xmath464 . for fixed degree @xmath2 ,
let @xmath815 be the rank of @xmath775 . by definition , @xmath816 hence @xmath817
. from the submatrix constructed to prove part ( 2 ) of corollary [ cor : schurinv ] , the first @xmath815 columns of @xmath222 are linearly independent .
it follows that for any column @xmath818 of @xmath775 with @xmath819 , there is a unique ( up to scaling ) relation @xmath820 where @xmath821 .
this gives rise to the polynomial @xmath822 with leading monomial @xmath823 .
these monomials are the largest @xmath824 monomials of degree @xmath2 with respect to lex ordering , so the result follows .
[ cor : initial1 ] let @xmath825 , generated in degrees @xmath366 and @xmath826 , with @xmath737 .
the initial ideal of @xmath464 with respect to the standard lexicographic order is @xmath827 by lemma [ lem : initial ] it suffices to show that the lex - segment ideal @xmath746 has the form above .
the hilbert function of a complete intersection @xmath464 generated in degrees @xmath366 and @xmath826 is @xmath738 more explicitly , we have @xmath828 recall @xmath829 is the vector space spanned by the @xmath830 greatest monomials of degree @xmath2 with respect to lex order . if @xmath831 , then the @xmath832 greatest monomials are @xmath833 .
these are all divisible by @xmath834 . if @xmath835 , the @xmath836 greatest monomials are @xmath837 . if @xmath838 , @xmath839 is divisible by @xmath834 .
if @xmath840 , then @xmath841 , where @xmath842 .
this is divisible by @xmath843 , which proves the corollary .
conca and valla @xcite parametrize of all ideals in two variables with a given initial ideal . using this
, one can show that the lex - segment ideal @xmath746 is the initial ideal of any ideal generated by two _ generic _ forms of degree @xmath366 and @xmath826 . here _
generic _ means there are certain polynomials in the coefficients of the forms that must not vanish ( the condition is not equivalent to the two forms being relatively prime ) .
lemma [ lem : initial ] can be viewed as a proof that the ideal @xmath464 , which is generated by two forms , is generic in this sense .
[ prop : initial2 ] set @xmath844 $ ] , @xmath845 $ ] both with standard lexicographic orders . for positive integers @xmath737 , @xmath846 , let @xmath847 $ ] be ideals satisfying @xmath848 and @xmath849 , respectively .
let @xmath845 $ ] and define ring maps @xmath850 by @xmath851 and @xmath852 .
set @xmath853 , @xmath854 , and @xmath855 .
then @xmath856 it suffices to show that @xmath857 .
we have @xmath858 let @xmath859 be a monomial of @xmath119 with degree @xmath224 .
we claim @xmath860 .
if @xmath861 or @xmath862 then @xmath863 or @xmath864 and we are done .
so set @xmath865 , where @xmath866 .
if @xmath867 then @xmath868 . so @xmath869 and @xmath870 .
if @xmath871 then @xmath872 . so @xmath873 and @xmath874 .
[ cor : initial3 ] let @xmath689 where @xmath875 and @xmath876 for @xmath334 .
also assume @xmath679 for @xmath334 and @xmath680 .
let @xmath877 as in the statement of theorem [ thm : simpedgereg ] .
then @xmath878 we use the short exact sequence @xmath879 and proposition [ prop : regseq ] . from proposition
[ prop : describei ] , @xmath880 and @xmath881 .
we show @xmath882 ; then we are done by proposition [ prop : regseq ] .
equivalently , we show @xmath883 for @xmath884 .
let @xmath657 be generated in degrees @xmath885 respectively .
by lemma [ lem : initial ] and proposition [ prop : initial2],@xmath883 for @xmath886 .
so we need to show that @xmath887 .
we consider @xmath61 cases . 1 .
@xmath888 for @xmath334 .
2 . @xmath889 and @xmath890 .
3 . @xmath890 and @xmath891 .
4 . @xmath892 for @xmath334 . * case 1 : * by proposition [ prop : describei ] , @xmath893 and @xmath894 .
suppose @xmath895 .
then @xmath896 , so @xmath897 and @xmath898 , where @xmath899 .
similarly if @xmath900 then @xmath901 and @xmath902 , where @xmath903 .
if @xmath904 then @xmath905 and @xmath906 .
hence @xmath907 .
* case 2 : * by proposition [ prop : describei ] , @xmath908 and @xmath909 . since @xmath910 by assumption , @xmath911 and @xmath912 . by proposition [ prop : describei ] , @xmath913 d=&\left\lceil \dfrac{\alpha_2+\alpha_\tau-\beta_2}{2 }
\end{array}\ ] ] hence @xmath914 @xmath915 , so @xmath916 it follows that @xmath917 * case 3 : * by arguing exactly as in case @xmath39 we obtain @xmath918 * case 4 : * by proposition [ prop : describei ] , @xmath919 is generated in degrees @xmath920 and @xmath921 is generated in degrees @xmath922 .
we have @xmath910 and @xmath923 by assumption .
it follows that @xmath924 and @xmath925 .
so @xmath926 the final inequality follows since we assumed @xmath927 are all at least @xmath808 .
we give several examples to illustrate both how the bounds in theorems [ thm : main2 ] and [ thm : main3 ] may be used and how well they approximate the actual regularity of the spline algebra .
these examples also elucidate a difference between _ complete _ central complexes @xmath6 ( in which the intersection of all facets of @xmath6 is an _ interior _ face of @xmath6 ) and central complexes which are not complete .
this difference is key to conjecture [ conj : c1 ] in the following section .
[ ex : schlegelcube ] in this example we apply theorem [ thm : main2 ] to bound the regularity of @xmath0 where boundary vanishing is imposed . consider the two dimensional polytopal complex @xmath59 in figure [ fig : sevl ] with five faces , eight interior edges , and four interior vertices . ]
impose vanishing of order @xmath9 along interior codimension one edges and vanishing of order @xmath715 along boundary codimension one faces .
the following hilbert polynomials are computed in ( * ? ? ?
* example 8.5 ) .
if @xmath928 , then @xmath929 by theorem [ thm : main2 ] , @xmath930 and @xmath931 for @xmath932 .
we compare the regularity bound @xmath933 with @xmath934 as computed in macaulay2 in table [ tbl : reg1 ] .
@xmath934 appears to have alternating differences of @xmath808 and @xmath585 and grows roughly as @xmath935 .
in fact @xmath934 appears to agree with the regularity of @xmath9-splines on the complex from example [ ex : hpolytopal ] . .
[ cols="^,^,^,^,^,^,^,^,^,^,^",options="header " , ] now suppose that vanishing of degree @xmath936 is imposed along @xmath102 .
then @xmath937 this formula is correct when @xmath938 are not too small ; for instance if @xmath939 and @xmath940 , the above formula has constant term @xmath941 while the actual constant , according to macaulay2 , is @xmath942 . by theorem [ thm : main2 ] , @xmath943 and @xmath944 for @xmath945 a comparison of the bound on @xmath946 and its actual value computed in macau - lay2 appears in table [ tbl : reg2 ] for @xmath947 .
c|ccccc + & @xmath940 & @xmath948 & @xmath949 & @xmath950 & @xmath951 + @xmath952 & 7 & 8 & 10 & 12 & 14 + @xmath953 & 13 & 14 & 15 & 17 & 19 + @xmath954 & 19 & 20 & 21 & 22 & 24 + @xmath939 & 25 & 26 & 27 & 28 & 29 + @xmath955 & 31 & 32 & 33 & 34 & 35 c|ccccc + & @xmath940 & @xmath948 & @xmath949 & @xmath950 & @xmath951 + @xmath952 & 4 & 4 & 5 & 6 & 7 + @xmath953 & 4 & 5 & 6 & 8 & 9 + @xmath954 & 7 & 8 & 8 & 9 & 10 + @xmath939 & 8 & 8 & 9 & 10 & 12 + @xmath955 & 11 & 11 & 12 & 12 & 13 + [ ex : largeface ] we now give an example which has very different behavior from example [ ex : highgen ] .
consider the two - dimensional polytopal complex @xmath59 formed by placing a regular ( or almost regular ) @xmath113-gon inside of a scaled copy of itself and connecting corresponding vertices by edges .
@xmath59 has one facet with @xmath113 edges and @xmath113 quadrilateral facets .
an example for @xmath956 is shown in figure [ fig : tenspoke ] .
we may or may not perturb the vertices so that the affine spans of the edges between the inner and outer @xmath113-gons do not all meet at the origin .
this does not appear to have much effect on regularity , although it does change the constant term of @xmath75 . according to theorem [ thm : main2 ] , @xmath957 as long as @xmath958 .
however , according to computations for @xmath959 and @xmath960 in macaulay2 , @xmath961 regardless of what value @xmath113 takes .
it appears that having a facet @xmath109 with many codimesion one facets may only significantly effect the regularity of @xmath0 if @xmath962 , as in example [ ex : highgen ] .
[ ex : oct ] consider a regular octahedron @xmath30 triangulated by placing a centrally symmetric vertex , shown in figure [ fig : octfig ] . in (
* example 5.2 ) , schenck shows that @xmath963 is free , generated in degrees @xmath964 and @xmath965 .
thus the regularity bound for @xmath963 given by corollary [ cor : freegens ] is tight .
computations in macaulay2 suggest that the regularity of @xmath963 stays at @xmath965 for generic perturbations of the noncentral vertices .
let @xmath92 be a pure hereditary @xmath113-dimensional polytopal complex .
we end with a conjecture in the case of uniform smoothness , where @xmath49 for all @xmath129 and @xmath54 for @xmath136 .
this conjecture is a slight refinement of ( * ? ? ?
* conjecture 5.6 ) . for @xmath124 , recall
@xmath349 is the number of codimension one faces of @xmath109 which are interior to the complex .
we will call a central complex @xmath6 _ complete _ if the intersection of all facets of @xmath6 is an _ interior _
face of @xmath6 .
[ conj : c1 ] let @xmath25 be a pure , central , hereditary @xmath113-dimensional polytopal complex .
call a facet @xmath966 a boundary facet if it has a codimension one face @xmath967 so that @xmath37 contains the cone vertex .
set @xmath968 then , 1 .
if @xmath6 is central and complete , then @xmath969 and this bound is tight .
2 . if @xmath6 is central but not complete , then @xmath970 .
furthermore , the bound is attained by free modules @xmath8 in both cases .
example [ ex : highgen ] shows that generators can be obtained in degree @xmath971 for the complete central case and degree @xmath972 in the non - complete central case , so these are the lowest possible regularity bounds that we can conjecture .
if @xmath8 is free , then @xmath973 by corollary [ cor : freegens ]
. example [ ex : oct ] , coupled with theorem [ thm : main3 ] , shows that conjecture [ conj : c1 ] is true in the complete , central , three dimensional , simplicial case .
if @xmath1 is complete , central , and non - simplicial , then conjecture [ conj : c1 ] should be provable using the methods of
[ sec : sstar ] .
the difficulty is in analyzing the ideal @xmath523 from lemma [ lem : splineideal ] .
conjecture [ conj : c1 ] part ( 2 ) is a natural generalization of a conjecture of schenck @xcite , that @xmath974 for @xmath13 .
this is a highly nontrivial conjecture in the simplicial case ; it implies , for instance , that @xmath975 . to date
, it is unknown whether @xmath976 .
the difficulty of this problem is in large part due to the fact that non - local geometry plays an increasingly important role in low degree @xcite .
since our methods hinge on using the algebras @xmath52 , which are locally supported approximations to @xmath0 , our approach will not be effective in proving conjecture [ conj : c1 ] part ( 2 ) .
in the non - simplicial case , conjecture [ conj : c1 ] part ( 2 ) appears to run contrary to the spirit of the regularity bounds we have proved in this paper , since no account is taken of interior facets of @xmath6 , which may have many codimension one faces .
it is nevertheless consistent with example [ ex : highgen ] , where the minimal generator of high degree is supported on a boundary facet , and example [ ex : largeface ] , where an interior facet with many codimension one faces appears to have no contribution to @xmath977 .
an example of a polytopal complex @xmath6 with a minimal generator of high degree ( relative to the number of codimension one faces of boundary facets ) , supported on interior facets , would be very interesting .
i thank hal schenck for his guidance and patient listening .
i also thank stefan tohaneanu and jimmy shan for helpful conversations .
macaulay2 @xcite was indispensable for performing the computations in this paper .
several of the images were generated using mathematica . | we derive bounds on the regularity of the algebra @xmath0 of mixed splines over a central polytopal complex @xmath1 . as a consequence we bound the largest integer @xmath2 ( the postulation number ) for which the hilbert polynomial @xmath3 disagrees with the hilbert function @xmath4 .
the polynomial @xmath3 has been computed in @xcite , building on @xcite .
hence the regularity bounds obtained indicate when a known polynomial gives the correct dimension of the spline space @xmath5 . in the simplicial case with all smoothness parameters
equal , we recover a bound originally due to hong @xcite and ibrahim and schumaker @xcite . |
neural networks have been successfully implemented in a plethora of prediction tasks ranging from speech interpretation to facial recognition . because of ground - breaking work in optimization techniques ( such as batch normalization [ 5 ] ) and model architecture ( convolutional , deep belief , and lstm networks ) , it is now tractable to use dnn methods to effectively learn a better feature representation compared to hand - crafted methods .
however , one area where such methods have not been utilized is the space of adversarial multiagent systems - specifically when the multiagent behavior comes in the form of trajectories .
there are two reasons for this : i ) procuring large volumes of data where deep methods are effective is difficult to obtain , and ii ) forming an initial representation of the raw trajectories so that deep neural networks are effective is challenging . in this paper
, we explore the effectiveness of deep neural networks on a large volume of basketball tracking data , which contains the @xmath0 locations of multiple agents in an adversarial domain . to throughly explore this problem , we focus on the following task : given the trajectories of the players and ball in the previous five seconds
, can we accurately predict the likelihood that a player with role @xmath1 will make the shot ? "
we express this as a ten - class prediction problem by considering the underlying role of the player . for this paper
, player role refers to the position of the player such as a point guard . since we are utilizing an image - based technique , the cnn will not know which role shoots the ball , especially if there are multiple offensive players nearby .
thus , solving as a ten - class problem is ideal for our deep - learning approach .
our work contains three main contributions .
first , we create trajectories for the offense , ball , and defense as an eleven channel image .
each channel corresponds to the five offensive and defensive players , as well as the ball . in order to encode the direction of the trajectories , we fade the paths of the ball and players . in our case ,
an instance is a possession that results in an shot attempt .
second , we apply a combined convolutional neural network ( cnn ) and feed forward network ( ffn ) model on an adversarial multiagent trajectory based prediction problem .
third , we gain insight into the nature of shot positions and the importance of certain features in predicting whether a shot will result in a basket .
our results show that it is possible to solve this problem with relative significance .
the best performing model , the cnn+fnn model , obtains an error rate of 26% .
in addition , it is able to accurately create heat maps by shot location for each player role . during training we found that one particular feature , whether the player received a pass just before the shot , is highly predictive .
other features , which are un - surprisingly important , are the number of defenders around the shooter and location of the ball at the time of the shot .
with the rise of deep neural networks , sports prediction experts have new tools for analyzing players , match - ups , and team strategy in these adversarial multiagent systems .. trajectory data was not available at the time , so much previous work on basketball data using neural networks have used statistical features such as : the number of games won and the number of points scored .
for example , bauer et .
al . [ 7 ] , use statistics from 620 nba games and a neural network in order to predict the winner of a game .
another interested in predicting game outcomes is mccabe [ 11 ] . on the other hand , nalisnick [ 12 ] in his blog
discusses predicting basketball shots based upon the type of shot ( layups versus free throws and three - point shots ) and where the ball was shot from . in other sports related papers , chang et.al . [ 4 ] use a neural network to predict winners of soccer games in the 2006 world cup . also , wickramaratna et.al .
[ 16 ] predict goal events in video footage of soccer games . although aforementioned basketball work did not have access to raw trajectory data , lucey et.al .
[ 9 ] use the same dataset provided by stats llc . for some of their work involving basketball .
they explore how to get an open shot in basketball using trajectory data to find that the number of times defensive players swapped roles / positions was predictive of scoring
. however , they explore open versus pressured shots ( rather than shot making prediction ) , do not represent the data as an image , and do not implement neural networks for their findings .
other trajectory work includes using conditional random fields to predict ball ownership from only player positions [ 18 ] , as well as predicting the next action of the ball owner via pass , shot , or dribble [ 20 ] .
goldsberry et .
[ 1 ] use non - negative matrix factorization to identify different types of shooters using trajectory data . because of a lack of defensive statistics in the sport , goldsberry et .
create counterpoints ( defensive points ) to better quantify defensive plays [ 2][3 ] .
al . [ 13 ] make use of trajectory data by segmenting a game of basketball into phases ( offense , defense , and time - outs ) to then analyze team behavior during these phases .
wang and zemel [ 17 ] use trajectory data representations and recurrent neural networks ( rather than cnn s ) to predict plays . because of the nature of our problem ,
predicting shot making at the time of the shot , there is not an obvious choice of labeling to use for a recurrent network .
they also fade the trajectories as the players move through time .
similar to us , they create images of the trajectory data of the players on the court .
our images differ in that we train our network on the image of a five second play and entire possession , while their training set is based on individual frames represented as individual positions rather than full trajectories .
they use the standard rgb channels , which we found is not as effective as mapping eleven channels to player roles and the ball for our proposed classification problem . also , the images they create solely concentrate on offensive players and do not include defensive positions . the final model that we implement , the combined network , utilizes both image and other statistical features .
there is work that utilizes both image and text data with a combined model .
recently , bengio et .
al . [ 19 ] , fei fei et . al . [ 6 ] , ng et . al . [ 14 ] , erham et . al . [ 15 ] , and mao et
10 ] all explore the idea of captioning images , which requires the use of generative models for text and a model for recognizing images .
however , to the best of our knowledge , we have not seen visual data that incorporates a fading an entire trajectory for use in a cnn .
the dataset was collected via sportsvu by stats llc .
sportsvu is a tracking system that uses 6 cameras in order to track all player locations ( including referees and the ball ) .
the particular data used in this study was from the 2012 - 2013 nba season and includes thirteen teams , which have approximately forty games each .
each game consists of at least four quarters , with a few containing overtime periods .
the sportvu system records the positions of the players , ball , and referees 25 times per second . at each recorded frame
, the data contains the game time , the absolute time , player and team anonymized identification numbers , the location of all players given as ( x , y ) coordinates , the role of the player , and some event data ( i.e. passes , shots made / missed , etc . ) .
it also contains referee positions , which are unimportant for this study , and the three - dimensional ball location .
this dataset is very unique in that before sportvu , there was very little data available of player movements on the court and none known that provides frame - by - frame player locations . since it is likely that most events in basketball can be determined by the movements of the players and the ball , having the trajectory data along with the event data should provide a powerful mixture of data types for prediction tasks .
there are a few ways to extract typical shot plays from the raw data .
one is to choose a flat amount of time for a shooting possession . in our case
we choose to include five seconds of a typical possession . to obtain clean plays , those that lasted less than 5 seconds due to possession changes and
those in which multiple shots were taken were thrown out . after throwing out these cases ,
we were left with 70,000 five second plays .
the other way of obtaining play data would be to take the entire possession .
thus , rather than having plays be limited to five seconds , possessions can be much longer or shorter . since the raw data does not contain labels for possession , we had to do this ourselves . to identify possession ,
we calculate the distance between the ball and each of the players .
the player closest to the ball would be deemed the ball possessor .
since this approach may break during passes and other events during a game , we end the play of the possession when the ball is closer to the defensive team for 12 frames ( roughly 0.5 seconds ) .
the procedure yields 72,200 possession examples .
since the classification problem is also dependent upon a player s role , a player s role must be chosen for each play .
one way to do this would be to identify the role of a player at the beginning of the play and hold it constant .
however , a player s role at the beginning of the play is usually not the same as a player s role at the time of the shot .
therefore , we decide to label a player s role via their position at end of the play in which the ball is shot .
in terms of applying deep neural networks to multiagent trajectories , we first have to form an initial representation .
a natural choice for representing these trajectories is in the form of an image . given that the basketball court is @xmath2x@xmath3 feet , we can form a @xmath2x@xmath3 pixel image . in terms of the type of image we use , there are multiple choices : i ) grayscale ( where we identify that a player was a specific location by making that pixel location 1 ) , ii ) rgb ( we can represent the home team trajectories in the red channel , the away team in the blue channel and the ball in the green channel , and the occurrence of a player / ball at that pixel location can be representing by a 1 ) , and iii ) 11-channel image ( where each agent has their own separate channel ) .
examples of the grayscale and rgb approach are shown in figure 1 .
the 11-channel approach requires some type of alignment . in this paper
, we apply the ` role - representation ' which was first deployed by lucey et al . [
the intuition behind this approach is that for each trajectory , the role of that player is known ( i.e. , point - guard , shooting guard , center , power - forward , small - forward ) .
this is found by aligning to a pre - defined template which is learnt in the training phase .
figure 1 shows examples of the methods we use to represent our data for our cnn . the grayscale image , which appears on the left , can accurately depict the various trajectories in our system
however , because the image is grayscale , the cnn will treat each trajectory the same .
since we have an adversarial multiagent system in which defensive and offensive behavior in trajectory data can lead to different conclusions , grayscale is not the best option for representation .
therefore , to increase the distinction between our agents , we represent the trajectories with an rgb scale .
we choose red to be offense , blue to be defense , and green to be the ball .
this approach takes advantage of multiple channels to allow the model to better distinguish our adversarial agents .
although the ball may be part of the offensive agent structure , we decide to place the ball in a channel by itself since the ball is the most important agent .
this approach , although better than the gray images , lacks in distinguishing player roles . since we classify our made and missed shots along with the role of the player that shoots the ball , a cnn will have trouble distinguishing the different roles on the court .
therefore , for our final representation , we decide to separate all agents into their own channel so that each role is properly distinguished by their own channel .
the above ideas nearly creates ideal images ; however , it does not include time during a play .
since each trajectory is equal brightness from beginning to end , it may be difficult to identify where the ball was shot from and player locations at the end of the play .
therefore , we implement a fading variable at each time frame . we subtract a parameterized amount from each channel of the image to create a faded image as seen in figures 1 and 2 .
thus , it becomes trivial to distinguish the end of the possession from the beginning and leads to better model performance .
we create two options with our final faded image data : five seconds before the ball was shot and the total possession .
five seconds likely performs better since full possession data contains more temporal information than cnn s can handle . in longer possessions , trajectories tend to cross more , which confuses the cnn model by adversely affecting the fading since new trajectories will be higher in value .
in order to fully utilize the power of this dataset , we implement a variety of networks for our prediction task . for our base model
, we use logistic regression with 197 hand - crafted features detailed later . to improve upon this basic model
, we use a multi - layer ffn with the same features and utilize batch normalization during training .
because of the nature of these two models , we could only include the positions of the players at the time of the shot .
therefore , we craft images to include the position of the players throughout the possession
. we then apply a cnn to these new image features .
finally , we create a combined model that adopts both images and the original ffn features for training .
for the first models , features with the knowledge of the game in mind were crafted .
the list of features includes : * player and ball positions at the time of the shot * game time and quarter time left * player speeds over either five seconds or an entire possession * distances and angles ( with respect to the hoop ) between players * whether the shooting player received a pass two seconds or less before the shot * number of defenders in front of the shooter ( @xmath4 of the shooter ) and within six feed based upon the angles calculated between players * individual time of ball possession for each offensive player logistic regression and fnn both use the same calculated features .
in addition , only the cnn does not incorporate the above features . for our model
, we use a cnn that consists of three full convolutional layers , each with 32 3x3 filters and a max - pooling layer with a pool - size of 2x2 following each convolutional layer .
after the last pooling layer there is a fully connected layer with 400 neurons , and finally an output layer consisting of a softmax function and ten neurons .
in addition , we use the relu function for our nonlinearity at each convolutional layer and the fully connected layer .
we also implement alexnet and vgg-16 , but we did not garner significant improvement from either model . the final network implemented is a combination of both the feed forward and convolutional networks . for this model , we use both the feed forward features and the fading trajectory images from the cnn . the idea behind the combined network is to have the model identify trajectory patterns in the images along with statistics that are known to be important for a typical basketball game .
the cnn and ffn parts of the combined network have the exact same architecture as the stand - alone versions of each model .
the final layers just before the softmax layer of each stand - alone network are then fully - connected to a feed - forward layer that consists of 1,000 neurons .
this layer is then fed into the final softmax layer to give predictions . after performing experiments and measuring log loss , we found that adding layers to this final network or adding additional neurons to this layer did not improve our final results .
all of the models ( fnn , cnn , and fnn + cnn ) use the typical log loss function as the cost function with a softmax function at the output layer .
the weights are initialized with a general rule of @xmath5 where @xmath6 is the number of incoming units into the layer . for training
, we implement the batch stochastic gradient method utilizing batch normalization .
batch normalization is used on the convolutional and feed forward layers of the model . in addition , we train the models on a nvidia titan x using theano . for the cnn and cnn+fnn networks
we utilize our eleven channel images with fading . to justify our representation
, we predict our classification problem with each crafted image set with the previously mentioned cnn architecture .
figure 3 displays the log loss and error rate for each of our representations .
evaluating our representations by log loss and error rate show a dramatic difference between our images .
the eleven channel method is understandably the best choice since having eleven channels gives the cnn much more role information .
figure 4 gives a summary of the results of the three different models .
it is intriguing how poorly the convolutional neural network captures the adversarial multiagent properties without the addition of the hand - crafted features used in the ffn .
theoretically , the image data contains all pertinent information related to the play .
therefore , in our problem , hand - crafted features clearly remain essential for successful prediction .
since the combined model surpasses the performance of either model separated , we presume that the networks find different features during training .
while the performance of the eleven channel images is an improvement ( a 10% gain ) , there is still much potential progress .
other methods include making each trajectory a different color in rgb space , varying the strength of the fading effect , and including extra channels of heat maps depicting the final ball position .
therefore , successfully representing trajectory data in sport that outperforms traditional metrics is a nontrivial problem and is not further explored here .
the remaining analyses are based on the combined model .
in addition to assessing the accuracy of our model , we explore a basic heat map of basketball shots based upon the raw data . at the very least
, we expect that our complete model should be able to recreate a heat map created via raw data .
we make the heat map by taking a count of shots made against missed within a square foot of the basketball court . since our classification model gives probabilities of making a shot ( rather than a binary variable ) , we take the maximum probability to create a heat map equivalent to the raw data map . in the raw data heat map ,
figure 6 , we note that the best probability of making a shot lies on top of the basket .
as we get farther back , the probability decreases with two dead zones ( lines of dark blue signifying very low probability ) : one right outside the paint and another just inside the three point line .
the model results , figure 5 , expectedly prefers shots near the basket .
the model also predicts a larger high value area surrounding the basket , which extends further into the paint of the court .
the model dead - zones are less arc - like as well .
there is a single dark - blue arc around the paint / basket , but the model predicts more pocket like behavior for missed shots than the raw data heat map . to further explore our results ,
we create heat maps solely based on the role of the player ( to break down scoring chances by agent ) .
as before , each role represents an offensive player . in figure 7
we present a few player roles and their representative heat maps .
role 3 is obviously the center position from their respective shot selection and role 5 is the left guard .
note that the model predicts a much smaller area of midrange scoring probability than from the raw data for role 3 .
the model heat map for role 3 strictly covers the paint , while the raw data has significantly higher shot probabilities outside of the paint .
role 5 , which depicts a guard has very similar heat maps for both the raw and model predictions .
the only notable differences are the larger hot area at the basket and the pocket of low probability shots just outside the paint detailed in the model heat map .
the probabilities that the combined model predicts for each shot may also provide useful insight into the game and our model s interpretation of high versus low value shots .
we create these histograms by finding which examples the model gives the highest probability as a shot made or missed by player with role @xmath7 .
we then group these examples together , and the probability of making a shot is reported in the histogram . in the histogram figure 8
we see that the majority of shots have a low probability .
this agrees with common basketball knowledge because many of a guard s shots are beyond the paint .
on the other hand , a center lives primarily under the basket and in the paint .
therefore , many of their shots are much more likely . watching a game live , a center getting a clean pass right under the basket
often results in a successful goal .
in addition , most roles tend to follow the probability pattern of role 5 with the exception of role 3 ( the center ) , which has a wider distribution and higher average probability of making a shot .
a brief glance at nba statistics agrees with this interpretation as the players with the highest shooting percentage ( barring free throws ) tend to be centers . the large number of low probability shots is not limited to the combined model .
the ffn has many low probability shots while the cnn does not .
we probe this phenomena by leaving single features out of the ffn to find the ones responsible for causing low probability shots .
we found that one feature in particular , pass - received , is the root cause .
this feature detects whether the shooter received a pass two seconds before the shot was taken .
we found that when a pass was received by the shooter during the two seconds before the shot the likelihood of missing the shot is much higher .
in fact , when a player received a pass , 94% of the shots are missed .
although counter - intuitive at first , two seconds is a small window to shoot . since two seconds
is the maximum time for this particular feature , many of the shooters may have shot immediately after the pass .
therefore , many of these shots are hurried explaining why a majority of them are missed .
this one feature led to a nearly 10% decrease in the error rate .
we also exhibit figure 9 to provide additional visual context for our model probabilities .
these figures depict the final positions of all players on the court at the time of the shot .
thus we can see how `` open '' the shot maker is at the time of the shot and the relative position of both the offensive and defensive players .
the offensive players are blue , defensive players red , and the ball is green .
each offensive and defensive player has the letter `` o '' and `` d '' respectively followed by a number signifying the role of that player at that time .
there are some times where the model makes some pretty questionable predictions .
for example , the three - point shot that is exhibited in the top left of figure 9 with a 0.892 probability is much too high .
even unguarded the best three - point shooters in the game can not hit 89.2% .
thankfully , these examples are very rare in our model . for the most part ,
three - point shots are rated extremely low by the model garnering probabilities of less than 20% .
in addition to three - point shots to having a generally low probability , shots that were well - covered by defenders had a much lower probability of success .
this is an unsurprising well - known result , but it does add validity to our model .
in addition , shots that are open and close to the basket are heavily favored in our model .
for example , in the right pictures , both players have open shots right under the basket with the defense well out of position . since our combined model predicts shots with more accuracy , we are curious to find the features that deep learning with a cnn observes with raw trajectory data .
our images contain both more temporal and spatial data since we fade images and do not include all multiagent interactions ( since it is difficult to label all inter - player events ) .
to investigate our raw trajectory approach , we use a simple gradient ascent method to create an image that most strongly reacts to the filters in our network .
for example , in a problem that contains cats and dogs , some filters may be looking for long whiskers to help identify a cat .
in addition , these filters help determine whether the network has been well - trained .
networks having filters with little to no interpretation or appear random are usually poorly trained .
figure 10 shows a few representative images that we craft with the gradient ascent method . in the first image of figure 10 from the left
, we see that the model is searching for defensive positioning ( light blue ) and the ball ( green ) .
there are clusters of these positions , which shows that the model is searching for the role that shoots the ball via ball positioning and the number of defensive players near the ball . on the other hand ,
the filter on the far right only depicts ball locations and does not include defensive positioning .
the second figure from the left is actually very similar to the far right filter ; however , it includes much more than simple ball information .
this filter is specifically looking for ball location as well as offensive / defensive player locations near the ball .
thus , the cnn model is likely more effective in capturing defensive structure and paths for shot prediction , which is difficult to construct by hand .
for further research , it would be very interesting to identify time dependency in basketball plays . in our image data ,
we subtract a flat amount at each equally spaced frame to cause the fading effect .
however , this assumes that the data in time is linearly related .
since this is not necessarily true , designing a recurrent model to find this temporal dependency could be a very interesting problem . instead of having a fading effect in the image data
, we can design an lstm that takes a moving window of player and ball trajectories .
one last aspect that was not taken into account during this study was the identities of teams and particular players .
the focus of this research was to gather more insight on the average shooting plays of teams in the nba .
however , teams in the nba have drastically different strategies .
for example , the golden state warriors tend to rely on a three - point strategy while bigger teams , such the thunder , build their offensive strategy around being inside the paint .
thus , new knowledge on basketball could be gathered if models were applied to different teams and possibly identify some overall team strategies .
such a more fine - grained analysis will require much more data .
[ 1 ] a. miller , l. bornn , r. adams , and k. goldsberry . ( 2014 ) `` factorized point process intensities : a spatial analysis of professional basketball . ''
_ international conference on machine learning_. [ 4 ] k.y .
huang and w.l .
( 2010 ) `` a neural network method for prediction of 2006 world cup football game . '' _ the 2010 international joint conference on .
institute of electrical and electronics engineers . _ 74 [ 5 ] s. ioffe and c. szegedy .
( 2015 ) `` batch normalization : accelerating deep network training by reducing internal covariate shift . '' _ international conference on machine learning_. [ 8 ] p. lucey , a. bialkowski , p. carr , s. morgan , i. matthews and y. sheikh . `` representing and discovering adversarial team behaviors using player roles . '' in _ ieee international conference on computer vision and pattern recognition . _ [ 9 ] p. lucey , a. bialkowski , p. carr , y. yue and i. matthews .
( 2014 ) `` how to get an open shot : analyzing team movement in basketball using tracking data . '' in _ mit
sloan sports analytics conference_. [ 14 ] r. socher , a. karpathy , q. le , c. manning , a. ng .
( 2014 ) `` grounded compositional semantics for finding and describing images with sentences . '' _ transactions of the association for computational linguistics 2 : 207 - 218 . _ [ 18 ] x. wei , l. sha , p. lucey , p. carr , s. sridharan and i. matthews .
( 2015 ) `` predicting ball ownership in basketball from a monocular view using only player trajectories . ''
_ knowledge discover and data mining workshop on large - scale sports analytics _ ,
sidney , australia .
[ 19 ] k. xu , j. lei ba , r. kiros , k. cho , a. courville , r. salakhutdinov , r. zemel , and y. bengio .
`` show , attend and tell : neural image caption generation with visual attention . ''
_ international conference on machine learning ( icml-15 ) . | in this paper , we predict the likelihood of a player making a shot in basketball from multiagent trajectories .
previous approaches to similar problems center on hand - crafting features to capture domain specific knowledge .
although intuitive , recent work in deep learning has shown this approach is prone to missing important predictive features . to circumvent this issue
, we present a convolutional neural network ( cnn ) approach where we initially represent the multiagent behavior as an image . to encode the adversarial nature of basketball
, we use a multi - channel image which we then feed into a cnn . additionally , to capture the temporal aspect of the trajectories we use `` fading . '' by using gradient ascent , we were able to discover what the cnn filters look for during training .
last , we find that a combined fnn+cnn is the best performing network with an error rate of 26% . |
consider an isometric immersion of @xmath2 into @xmath3 , where @xmath2 denotes a connected , simply connected @xmath0-dimensional space form of constant sectional curvature @xmath4 .
such an immersion must be * a cylinder over a complete plane curve if @xmath5 , which was proved by hartman - nirenberg @xcite .
( massey @xcite gave an alternative proof for @xmath6 . ) * totally geodesic if @xmath7 , which was proved by oneill - stiel @xcite .
roughly speaking , if @xmath4 is non - negative , such an isometric immersion is trivial .
however , if @xmath4 is negative , there exist many nontrivial examples of such isometric immersions as shown in @xcite .
a complete classification was given by abe - mori - takahashi @xcite ( see also @xcite ) .
on the other hand , even if @xmath4 is non - negative , there are many nontrivial examples of such isometric immersions admitting some _ singularities _ ( cf . figure [ fig : gallery ] ) . in the seminal paper of murata - umehara @xcite , they classified complete _ flat fronts _ , which are regarded as a generalization of isometric immersions of @xmath8 into @xmath9 ( cf .
section [ sec : transformation ] ) .
moreover , complete flat fronts were proved to have various global properties ( see fact [ fact : mu ] for the precise statement ) .
hence , the following problem naturally arises : _ what occurs in the non - flat or higher dimensional cases ? _
c@c@c & & + ( a ) totally geodesic . &
( b ) non - totally - geodesic . &
( b ) the same , transparent . in this paper
, we investigate ` generalized ' isometric immersions of @xmath10 into @xmath11 , where @xmath11 is a sphere of positive sectional curvature @xmath7 .
to clarify the settings , we first give a definition of _ coherent tangent bundles of space form type _ , which is a generalization of space forms ( definition [ def : ctb_spaceform ] ) .
recall that , in @xcite , saji - umehara - yamada introduced _
coherent tangent bundles _ which can be regarded as a generalized notion of riemannian manifolds with singularities ( for the precise definitions , see section [ sec : prelim ] ) .
we remark that there exist coherent tangent bundle structures of space form type on @xmath12 whose sectional curvature is non - positive ( cf .
example [ ex : simp_conn_ctb ] ) , while the sectional curvature of a simply connected compact space form must be positive .
a _ wave front _
( or a _ front _ , for short ) is a generalized notion of immersions admitting singularities .
coherent tangent bundles can be considered as an intrinsic formulation of fronts , because fronts ( more generally _ frontals _ ) induce coherent tangent bundle structures on the source manifold ( cf .
@xcite , see also example [ ex : induced_ctb ] ) .
the main theorem of this paper is as follows .
[ thm : main ] let @xmath13 be a connected smooth @xmath0-manifold , and @xmath14 a front .
if the induced coherent tangent bundle on @xmath13 is of space form type with positive sectional curvature @xmath4 , then @xmath15 is either totally geodesic or an umbilic - free tube of radius @xmath16 whose center curve is a closed regular curve in @xmath17 . in particular , @xmath13 is diffeomorphic to either @xmath18 or @xmath19 , hence orientable . to prove theorem [ thm : main ] , we classify _ weakly complete _ constant curvature @xmath20 fronts in @xmath21 ( theorem [ thm : w - complete ] ) , where completeness implies weak completeness ( cf
fact [ fact : mu_complete ] ) . using these
, we then investigate some local and global properties of the caustics of constant curvature @xmath20 fronts .
caustics are defined as the singular loci of parallel fronts .
it is known that caustics of flat fronts in @xmath22 are also flat fronts ( cf.@xcite for the case of @xmath23 ; @xcite for the case of @xmath24 ; @xcite for the case of @xmath25 ) .
we can prove a similar result for constant curvature @xmath20 fronts in @xmath21 ( proposition [ prop : caus - tube ] ) . even if the original front is connected or co - orientable , its caustic will , in general , not be connected or co - orientable .
however , the following holds .
[ thm : caustic ] let @xmath26 be a connected non - totally - geodesic weakly complete constant curvature @xmath20 front . then
, the caustic @xmath27 of @xmath15 is also a connected non - totally - geodesic weakly complete constant curvature @xmath20 front .
moreover , if @xmath15 is complete @xmath28resp .
non - co - orientable@xmath29 then @xmath27 is also complete @xmath28resp .
non - co - orientable@xmath29 .
finally in section [ sec : transformation ] , we investigate constant curvature @xmath20 fronts in the case of @xmath6 . on the regular point set of a constant curvature @xmath20 front
, its extrinsic curvature vanishes identically .
thus , we also call such fronts _ extrinsically flat_. although the local nature of extrinsically flat surfaces in @xmath25 is the same as that of flat surfaces in @xmath24 , via central projection of a hemisphere to a tangent space ( cf .
example [ ex : dev - moebius ] ) , they may display different global properties ( cf .
example [ ex : non - co - ori ] ) . on the other hand , in the particular case of @xmath6
, we can define the _ duals _ of tubes of radius @xmath30 in @xmath25 .
such duals may not be fronts in general ( cf .
remark [ rem : ins ] ) .
we classify tubes of radius @xmath30 in @xmath25 whose duals are congruent to the original ones ( i.e. , _ self - dual _ developable tubes ) in theorem [ prop : self - dual ] .
this paper is organized as follows . in section
[ sec : prelim ] , we review the definition and fundamental properties of coherent tangent bundles and fronts . in section [ sec : ctb ] , we give a definition of coherent tangent bundles of space form type and give some examples . in section
[ sec : proof_a ] , we shall investigate constant curvature @xmath20 fronts in @xmath21 , and prove theorem [ thm : main ] and theorem [ thm : caustic ] . in section [ sec : transformation ] , we restrict to the case of @xmath6 and investigate their duals .
in this section , we review the definitions and basic properties of space forms , coherent tangent bundles and ( wave ) fronts .
let @xmath21 be the unit sphere @xmath31 where @xmath32 is the canonical inner product of the euclidean space @xmath33 .
we denote by @xmath34 the hyperbolic space @xmath35 where @xmath36 is the lorentz - minkowski space equipped with the canonical lorentzian inner product @xmath37 .
that is , @xmath38 , @xmath39 and @xmath40 , where @xmath2 is the connected , simply connected @xmath0-dimensional space form of constant sectional curvature @xmath4 .
then , the unit tangent bundles of @xmath41 , @xmath21 and @xmath34 are identified with @xmath42 respectively , where @xmath43 is the de sitter space . for @xmath44 , set a vector @xmath45 as @xmath46 where @xmath47 is the canonical basis of @xmath33 and @xmath48 for each @xmath49 . in @xcite ,
saji - umehara - yamada gave a definition of _ coherent tangent bundles _
, which is a generalized notion of riemannian manifolds .
[ def : ctb ] let @xmath50 be a vector bundle of rank @xmath0 over a smooth @xmath0-manifold @xmath13 .
we equip a fiber metric @xmath32 on @xmath50 and a metric connection @xmath51 on @xmath52 .
if @xmath53 is a bundle homomorphism which satisfies @xmath54)=0\ ] ] for smooth vector fields @xmath55 , @xmath56 on @xmath13 , then @xmath57 is called a _ coherent tangent bundle _ over @xmath13 .
it is known that , in general , coherent tangent bundles can be constructed only from positive semi - definite metrics ( so - called _ kossowski metrics _ ) .
this was proved in @xcite for the case of @xmath6 , and in @xcite for the higher dimensional case .
a point @xmath58 is called _
@xmath59-singular _ , if @xmath60 is not a bijection , where @xmath61 is the fiber of @xmath50 at @xmath62 .
let @xmath63 be the set of @xmath59-singular points .
on @xmath64 , the levi - civita connection of the pull - back metric @xmath65 coincides with the pull - back of @xmath51 .
thus , we may recognize the concept of coherent tangent bundles as a generalization of riemannian manifolds . a coherent tangent bundle @xmath57 is called _ orientable _ , if @xmath13 is orientable .
moreover , @xmath50 is _ co - orientable _ if the vector bundle @xmath50 is orientable ( i.e. , there exists a smooth section @xmath66 of @xmath67 defined on @xmath13 such that @xmath68 holds for any orthonormal frame @xmath69 on @xmath50 ) .
if the @xmath59-singular set @xmath63 is empty , the orientability is equivalent to the co - orientability .
moreover , even if the coherent tangent bundle is non - co - orientable ( resp .
non - orientable ) , there exists a double covering @xmath70 such that the pull - back of @xmath50 by @xmath71 ( resp .
@xmath72 ) is co - orientable ( resp .
orientable ) coherent tangent bundle .
on the other hand , let @xmath73 be a positive semi - definite metric on a smooth @xmath0-manifold @xmath13 .
a point @xmath58 is called _
@xmath73-singular _ , if @xmath73 is not positive definite at @xmath62 .
let @xmath74 be the set of @xmath73-singular points .
[ def : completectb ] if there exists a symmetric covariant tensor @xmath75 on @xmath13 with compact support such that @xmath76 gives a complete metric on @xmath13 , @xmath73 is called _
complete_. we call a coherent tangent bundle _ complete _ , if the pull - back metric is complete .
such a completeness was used in the study of maximal surfaces in @xmath77 @xcite , flat fronts in @xmath23 @xcite and in @xmath24 @xcite . since @xmath74 is closed , we have the following . [
lem : sing_compact ] the @xmath73-singular set @xmath74 of a complete positive semi - definite metric @xmath73 is compact , if @xmath74 is non - empty .
if @xmath13 is compact , every positive semi - definite metric on @xmath13 is complete .
let @xmath78 be a riemannian manifold and @xmath79 the unit tangent bundle . a smooth map @xmath80 is called a _ frontal _ , if for each point @xmath58 , there exist a neighborhood @xmath81 of @xmath62 and a smooth vector field @xmath82 along @xmath15 such that @xmath83 holds for each @xmath84 and @xmath85 . as in (
* example 2.4 ) , frontals induce coherent tangent bundles .
[ ex : induced_ctb ] for a frontal @xmath86 , set * @xmath87 is the subbundle of the pull - back bundle @xmath88 perpendicular to @xmath89 , * @xmath90 defined as @xmath91 , * @xmath51 is the tangential part of the levi - civita connection on @xmath92 , * @xmath32 is the metric on @xmath87 induced from the metric on @xmath92 .
then , @xmath93 is a coherent tangent bundle .
we call @xmath87 the _ induced coherent tangent bundle_. a @xmath59-singular point of the induced coherent tangent bundle @xmath87 is a _ singular _ point of @xmath15 ( i.e. , @xmath94 holds ) .
a non - singular point is called _
regular_. a frontal @xmath15 is called complete ( resp .
orientable , co - orientable ) if the induced coherent tangent bundle @xmath87 is complete ( resp .
orientable , co - orientable ) .
we remark that a frontal @xmath15 is co - orientable if and only if @xmath89 is well - defined on @xmath13 .
we now consider the case that @xmath92 is the simply connected space form @xmath3 of constant sectional curvature @xmath4 . if @xmath5 ( resp .
@xmath20 , @xmath95 ) , the unit normal vector field @xmath89 can be considered as a smooth map into @xmath18 ( resp .
@xmath21 , @xmath96 ) .
the map @xmath97 is called the _
legendrian lift _ of @xmath15 ( cf . ) .
let @xmath3 be the simply connected space form of constant sectional curvature @xmath4 , @xmath98 a frontal with a @xmath28locally defined@xmath29 unit normal vector field @xmath89 , and @xmath99 the induced coherent tangent bundle .
then , the identity @xmath100 holds , where @xmath55 and @xmath56 are smooth vector fields on @xmath13 , @xmath101 and @xmath102 are smooth sections of @xmath87 , and @xmath103 is the curvature tensor of the connection @xmath51 given by @xmath104}\xi.\ ] ] the equation is called the _
gauss equation_. a frontal @xmath105 is called a ( _ wave _ ) _ front _ , if its legendrian lift @xmath106 is an immersion .
then , @xmath107 defines a riemannian metric on @xmath13 which we call the _
lift metric_. if @xmath108 is complete , @xmath15 is called _
weakly complete_. then the following holds .
[ fact : mu_complete ] a complete front is weakly complete .
since the proof of ( * ? ? ?
* lemma 4.1 ) can be applied in our setting , we omit the proof .
a point @xmath58 is called _ umbilic _ , if there are real numbers @xmath109 , @xmath110 such that @xmath111 and @xmath112 hold . for a front @xmath105 and @xmath113 , set @xmath114
then we can check that @xmath115 is a front ( called the
_ parallel front _ of @xmath15 ) with a unit normal @xmath116 given by @xmath117 umbilic points are common in its parallel family .
[ lem : singular - umb ] a singular point @xmath62 of a front @xmath15 is umbilic if and only if @xmath118 . if @xmath118 , then @xmath62 is umbilic
thus , we shall prove the converse .
assume that @xmath62 is umbilic .
there exist real numbers @xmath109 , @xmath110 @xmath119 such that @xmath112 .
if we suppose @xmath120 , then we have @xmath121 , and hence @xmath122 , where we put @xmath123 .
since @xmath62 is a singular point of @xmath15 , then there exists a non - zero tangent vector @xmath124 such that @xmath125 .
therefore , we have @xmath126 which contradicts the fact that @xmath15 is a front .
a coherent tangent bundle @xmath127 over a smooth @xmath0-manifold @xmath13 is called _ of constant sectional curvature _ @xmath128 , if @xmath129 holds for smooth vector fields @xmath55 , @xmath56 on @xmath13 and a smooth section @xmath101 of @xmath50 . in particular ,
if @xmath130 , such a coherent tangent bundle is called _ flat_. we say a frontal @xmath15 has _ constant sectional curvature _ or _ constant curvature _ ( resp . is _ flat _ ) , if the induced coherent tangent bundle @xmath87 has constant sectional curvature ( resp .
is flat ) .
[ def : ctb_spaceform ] a complete coherent tangent bundle of constant sectional curvature is called _ space form type_. if the @xmath59-singular set @xmath63 is empty , a coherent tangent bundle of space form type is regarded as a space form ( i.e. , a complete riemannian manifold of constant sectional curvature ) .
a simply connected space form is isometric to either a sphere , euclidean space or hyperbolic space . in particular
, any simply connected compact space form has positive sectional curvature .
however , if the @xmath59-singular set @xmath63 is non - empty , there are simply connected compact coherent tangent bundles of space form type whose sectional curvature is non - positive .
[ ex : simp_conn_ctb ] set @xmath131 where we regard @xmath18 as a subset of @xmath41 ( cf .
figure [ fig : flat_ctb ] ) .
the singular point set of @xmath132 is the equator @xmath133 . since an unit normal vector field @xmath89 along @xmath132 is given by @xmath134 , @xmath132 is a co - orientable frontal ( but not a front ) . as in example
[ ex : induced_ctb ] , @xmath132 induces a coherent tangent bundle @xmath135 . since @xmath18 is compact , the pull back metric @xmath136 is complete .
moreover , the gauss equation implies that @xmath137 ( cf .
lemma [ lem : degenerate - nu ] ) .
hence @xmath138 is a flat coherent tangent bundle of space form type .
similarly , let @xmath139 be the co - orientable frontal defined by @xmath140 where @xmath141 is the diffeomorphism @xmath142 as in the case of @xmath138 , the coherent tangent bundle @xmath143 induced from @xmath144 is of space form type with sectional curvature @xmath95 .
c@c@c & & the unit normal vector field @xmath89 of either the frontals @xmath132 or @xmath145 in example [ ex : simp_conn_ctb ] satisfies @xmath146 .
in general , the following holds .
[ lem : degenerate - nu ] let @xmath147 be a frontal with a unit normal vector field @xmath89
. then @xmath15 has constant sectional curvature @xmath4 if and only if @xmath148 holds on @xmath13 .
assume @xmath15 is constant sectional curvature @xmath4 , that is , the induced coherent tangent bundle @xmath93 is of constant sectional curvature @xmath4 . by and
, @xmath149 holds for any vector fields @xmath55 , @xmath56 on @xmath13 and sections @xmath101 , @xmath102 of @xmath87 , which is equivalent to @xmath148 .
conversely , assume @xmath148 , that is holds . by
, @xmath150 holds for any vector fields @xmath55 , @xmath56 on @xmath13 and sections @xmath101 , @xmath102 of @xmath87 , which implies with @xmath151 .
in this section , we exhibit examples of constant curvature @xmath20 fronts in the unit sphere @xmath21 .
then , we give the proof of theorem [ thm : main ] . finally , we investigate the caustics of constant curvature @xmath20 fronts and prove theorem [ thm : caustic ] . in the following ,
we respectively denote by @xmath152 , @xmath153 the derivatives @xmath154 , @xmath155 , etc .
the totally geodesic hypersphere @xmath156 in @xmath21 is a constant curvature @xmath20 front ( immersion , in fact ) satisfying @xmath146 , which can be considered as a trivial example .
we shall give an example of @xmath157 as follows .
let @xmath158 be a complete regular curve parametrized by arclength .
if there exists @xmath159 such that @xmath160 holds for any @xmath161 , we call @xmath162 _ closed _ or _
periodic_. the minimum of @xmath163 is called the _ period _ of @xmath164 . moreover ,
if @xmath165 for each @xmath161 , we call @xmath166 _ antiperiodic_. then , @xmath167 is called the _
antiperiod_. for a complete regular curve @xmath162 , set @xmath168 as @xmath169 where @xmath69 is an orthonormal frame of the normal bundle @xmath170 .
we call @xmath15 the _ tube of radius @xmath30 _ or the _ developable tube_. if @xmath164 is closed with period @xmath106 ( resp
. antiperiodic with antiperiod @xmath167 ) , we can regard @xmath15 as a map @xmath171 ( resp .
@xmath172 ) , where @xmath173 .
we remark that @xmath15 gives a parametrization of the set @xmath174 which consists of points whose geodesic distance to @xmath162 is @xmath30 .
since @xmath175 , we call @xmath164 and @xmath176 the center curves of @xmath174 . [ lem : dev - tube ] let @xmath168 be a developable tube given by .
then , for each @xmath161 , there exists a point @xmath177 such that @xmath178 is a singular point of @xmath15 .
set @xmath179 .
let @xmath180 be a _ bishop frame _
@xcite of @xmath164 .
that is , there exist smooth functions @xmath181 @xmath182 such that @xmath183 for @xmath184 .
remark that the curvature function @xmath185 of @xmath166 is given by @xmath186 differentiating , we have @xmath187 hence , the first fundamental form @xmath188 is calculated as @xmath189 where @xmath190 is the standard metric of @xmath191 . the singular point set @xmath192
is given by @xmath193 at a singular point , @xmath194 holds . by
, we have @xmath195 is a subset of @xmath192 . hence
, if @xmath196 holds at @xmath161 , @xmath197 holds for any point @xmath177 . for a singular point @xmath198 , set @xmath199 , where @xmath200 for @xmath184
then we have @xmath201 , where
@xmath202 thus , if @xmath203 holds at @xmath161 , there exists @xmath177 such that @xmath204 , and hence @xmath205 .
this completes the proof . by the above proof
, it also holds that every singular point @xmath206 is of corank one @xmath28i.e .
, @xmath207 .
[ prop : dev - tube ] a developable tube @xmath168 is an umbilic - free weakly complete constant curvature @xmath20 front . if @xmath15 is complete , then the center curve @xmath164 is closed .
moreover , @xmath15 is non - co - orientable if and only if @xmath164 is antiperiodic . by , we have that @xmath208 gives the unit normal vector field along @xmath15 .
thus , @xmath157 holds and @xmath15 is a constant curvature @xmath20 frontal by lemma [ lem : degenerate - nu ] .
then , the lift metric @xmath209 satisfies @xmath210 which implies that @xmath15 is a front and weakly complete .
moreover , since @xmath157 , @xmath15 has no umbilic points .
assume that @xmath15 is complete .
if @xmath166 is not closed , by lemma [ lem : dev - tube ] , @xmath192 is not closed . by lemma [ lem : sing_compact ] ,
the singular point set @xmath192 of @xmath15 is compact .
therefore , @xmath166 must be closed .
the third assertion follows because @xmath164 gives the unit normal of @xmath15 .
now , we give a proof of theorem [ thm : main ] . in the following lemmas ( lemma [ lem : sp_coord ] and lemma [ lem : massey ] ) , we fix a non - totally - geodesic constant curvature @xmath20 front @xmath26 .
first , assume @xmath15 is co - orientable .
let @xmath211 be a ( globally defined ) unit normal vector field .
by lemma [ lem : degenerate - nu ] , @xmath212 .
thus , a point @xmath58 is umbilic if and only if @xmath213 .
we denote by @xmath214 the umbilic point set of @xmath15 .
[ lem : sp_coord ] a singular point @xmath58 of @xmath15 is of corank one .
moreover , around any non - umbilic point @xmath215 , there exist a local coordinate system @xmath216 and a smooth function @xmath217 on @xmath81 such that @xmath218 hold for each @xmath219 and @xmath220 is a frame on @xmath81 . setting @xmath221 , the restriction @xmath222 is a totally geodesic embedding for each @xmath223 .
let @xmath224 be a local coordinate system such that @xmath225 for each @xmath219 . since @xmath15 is a front , @xmath226 for each @xmath219 .
hence @xmath227 is a singular point if and only if @xmath228 , and then @xmath229 is linearly independent .
in this case , for each @xmath230 , the parallel front @xmath115 ( cf . ) is an immersion around @xmath62 .
then , we have @xmath231 for each @xmath219 .
thus , the principal curvatures @xmath232 of @xmath115 are given by @xmath233 , @xmath234 @xmath235 for some function @xmath236 on @xmath237 .
let @xmath216 be a curvature line coordinate system of @xmath115 around @xmath215 .
that is , @xmath238 and @xmath239 hold for @xmath219 .
since @xmath240 is smooth on @xmath81 and , we have with @xmath241 . with respect to the third assertion
, @xmath242 gives a unit normal vector field of @xmath243 .
then , @xmath244 and @xmath245 yield @xmath246 on @xmath247 for each @xmath248 .
together with @xmath249 @xmath235 on @xmath247 , we have the conclusion . on the regular point set ,
the principal curvatures @xmath250 of @xmath15 are given by @xmath251 , @xmath252 @xmath235 .
we call @xmath253 and @xmath254 : u \rightarrow p^1({\boldsymbol{r}})$ ] the _ curvature radius function _ and the _ principal curvature map _ , respectively as in @xcite , where @xmath255 is the real projective line
. we may regard @xmath256 as the proportional ratio @xmath257=[-\nu_{u_1}:f_{u_1}]$ ] .
then , in our setting , a point @xmath58 is singular ( resp .
umbilic ) if and only if @xmath258 $ ] ( resp .
@xmath259 $ ] ) . by the proof above
, the principal curvature map @xmath256 can be continuously extended to the whole of @xmath13 ( cf .
* proposition 1.6 ) ) .
let @xmath260 be an open interval .
a curve @xmath261 is called _ asymptotic _ , if @xmath262 holds for each @xmath263 , where @xmath264 .
if @xmath15 is a front , @xmath265 does not vanish .
thus , we can parametrize @xmath266 by an arclength parameter @xmath267 , i.e. , @xmath268 for each @xmath263 .
[ lem : massey ] let @xmath269 be an asymptotic curve parametrized by arclength passing through a non - umbilic point @xmath270 . then
the restriction of the curvature radius function @xmath253 to @xmath266 satisfies @xmath271 for each @xmath263
. moreover , the closure of @xmath272 and @xmath214 do not intersect .
let @xmath216 be a coordinate system centered at @xmath273 as in lemma [ lem : sp_coord ] .
moreover , by lemma [ lem : sp_coord ] , we may assume that each @xmath274-curve is a geodesic of @xmath275 .
lemma [ lem : sp_coord ] yields that @xmath276 on @xmath81 .
thus , by lemma [ lem : sp_coord ] , @xmath277 holds . on the other hand , we have @xmath278 since @xmath279 and @xmath280
, we have @xmath281 , which proves the first assertion .
the general solution is @xmath282 with constants @xmath283 , @xmath284 .
therefore , we have @xmath285.\ ] ] suppose that @xmath272 accumulates at an umbilic point @xmath286 .
since @xmath287 passes through a non - umbilic point @xmath288 , there exists a sequence @xmath289 such that @xmath290 .
let @xmath291 be a sequence such that @xmath292 for each @xmath293 .
then , there exist constants @xmath294 such that @xmath295 .
in particular , @xmath253 is bounded . on the other hand , @xmath257=[1/\rho:1]$ ]
implies that @xmath296 must diverge as @xmath297 , which is a contradiction .
hence , we may conclude that the closure of @xmath272 and @xmath214 have no intersection .
[ thm : w - complete ] let @xmath298 be a weakly complete co - orientable constant curvature @xmath20 front .
if @xmath15 is not totally geodesic , there exists a complete regular curve in @xmath166 in @xmath21 such that @xmath15 is a developable tube of @xmath166 given by .
let @xmath216 be a coordinate system centered at @xmath273 as in lemma [ lem : sp_coord ] .
on @xmath81 , @xmath89 is a regular curve in @xmath21 and we may represent as @xmath299 . by lemma [ lem : sp_coord ] , for each @xmath223 , there exist a @xmath0-subspace @xmath300 of @xmath33 such that the image of @xmath301 is included in @xmath302 . since @xmath300 is perpendicular to both of @xmath166 and @xmath303 , @xmath300 is the normal space @xmath304 at @xmath166 .
thus , we have the image of @xmath301 is given by the embedding @xmath305 defined as @xmath306 , where @xmath307 is an orthonormal basis of @xmath304 at @xmath166 and @xmath308 , which implies @xmath15 is a developable tube given by of @xmath309 on @xmath81 .
as in lemma [ lem : dev - tube ] , the lift metric @xmath108 of @xmath15 is given by @xmath310 on @xmath81 , where @xmath190 is the standard metric of @xmath191 . since @xmath15 is weakly complete , we have @xmath311 .
suppose that the umbilic point set @xmath214 of @xmath15 is not empty , and take an umbilic point @xmath312 .
then there exists a sequence @xmath313 such that @xmath314 . for each @xmath315
, we set @xmath316 as an asymptotic curve passing through @xmath315 . since @xmath317 is a great circle , there exists a subsequence @xmath318 such that @xmath319 is also a great circle .
since the inverse image @xmath320 of @xmath321 through @xmath15 passes through @xmath288 and by lemma [ lem : massey ] , every point on @xmath320 is umbilic . on the other hand , for each @xmath322 , there exist @xmath323 such that @xmath324 $ ] , by . if we take the limit as @xmath325 , we have @xmath326 .
therefore , by the continuity of the principal curvature map @xmath256 , there exists a point on @xmath320 such that @xmath258 \,(\neq[0:1])$ ] , which implies there exists a singular point on @xmath214 .
however , by lemma [ lem : singular - umb ] and lemma [ lem : sp_coord ] , we have a contradiction and hence @xmath214 is empty . by a homothety
, it suffices to consider the case of @xmath327 and we may assume @xmath298 is a complete constant curvature @xmath20 front .
assume @xmath15 is not totally geodesic .
first , consider the case that @xmath15 is co - orientable .
since the completeness implies the weak completeness by fact [ fact : mu_complete ] , theorem [ thm : w - complete ] implies that @xmath15 is a tube of radius @xmath30 as in whose center curve @xmath158 is regular . by proposition [ prop : dev - tube ] ,
the center curve @xmath164 is closed .
let @xmath159 be the period of @xmath164 .
we shall prove the orientability of @xmath15 .
take a positively oriented orthonormal frame @xmath328 of @xmath164 such that @xmath329 ( cf . ) .
if @xmath15 is not orientable , there exists a real number @xmath330 such that the orientation of @xmath331 does not coincide with that of @xmath332 .
then , @xmath333 holds .
however , by the closedness of @xmath164 , we have @xmath334 , and hence @xmath335 holds .
then @xmath336 yields @xmath337 , which contradicts the regularity of @xmath166 . therefore , @xmath15 must be orientable . in the case that @xmath15 is not co - orientable , there exists a double covering @xmath338 such that @xmath339 is co - orientable .
then , @xmath340 is a co - orientable developable tube whose unit normal @xmath341 is closed .
moreover , since @xmath15 is not co - orientable , @xmath341 is antiperiodic and regular by proposition [ prop : dev - tube ] .
assume @xmath13 is non - orientable .
then , by the above argument , we have @xmath342 for some @xmath330 , which is a contradiction .
and hence @xmath15 must be orientable . for a non - totally - geodesic constant curvature @xmath20 front @xmath26
, the map @xmath343 is called the _ caustic _ ( or the _ focal hypersurface _ ) of @xmath15 , where @xmath253 is the curvature radius function ( cf .
lemma [ lem : sp_coord ] ) .
we remark that , even if @xmath15 is non - co - orientable , this definition is also well - defined .
also note that caustics may not be connected .
[ prop : caus - tube ] for a non - totally - geodesic constant curvature @xmath20 front in @xmath21 , its caustic is also a constant curvature @xmath20 front without umbilic points .
moreover , if @xmath15 is a developable tube as in of a regular curve @xmath162 , the unit tangent vector @xmath344 gives the center curve of @xmath27 .
using the coordinate system as in lemma [ lem : sp_coord ] , the first assertion can be proved by a straight - forward calculation .
if @xmath15 is a developable tube as in , we have @xmath345 . substituting @xmath15 , @xmath341 and @xmath253 into
, we have @xmath346 for @xmath347 , which proves the assertion . in ( * ? ? ?
* theorem 6.6 ) , kokubu - rossman - umehara - yamada proved that any flat front in @xmath23 is locally the caustic of some flat front .
a similar result is proved as follows .
[ cor : inverse_caustic ] any umbilic - free constant curvature @xmath20 front in @xmath21 is locally the caustic of some constant curvature @xmath20 front .
let @xmath348 be an umbilic - free constant curvature @xmath20 front . by the proof of theorem [ thm : w - complete ]
, @xmath15 is a part of developable tube .
if @xmath349 gives a center curve of @xmath15 , let @xmath350 be an integral curve of @xmath351 .
then , by proposition [ prop : caus - tube ] , the caustic of the developable tube of @xmath166 coincides with @xmath15 . by theorem [ thm : w - complete ] and theorem [ thm : main ] ,
@xmath15 is a developable tube as in .
let @xmath352 be a center curve of @xmath15 .
without loss of generality , we may assume @xmath166 is parametrized by arclength .
by proposition [ prop : caus - tube ] , @xmath353 gives a center curve of the caustic @xmath27 .
let @xmath354 be the bishop frame of @xmath355 such that @xmath356 for @xmath347 .
then the caustic is given by @xmath357 with the unit normal @xmath358 , where @xmath359 .
if @xmath360 is the curvature function of @xmath166 , @xmath361 holds .
the lift metric @xmath362 of @xmath27 is calculated as @xmath363 which proves the first assertion . if @xmath15 is complete ( resp .
non - co - orientable ) , by proposition [ prop : dev - tube ] , @xmath162 is closed ( resp .
antiperiodic ) . then , @xmath364 is also closed ( resp .
antiperiodic ) , and hence the caustic @xmath27 is also complete ( resp .
non - co - orientable ) by proposition [ prop : dev - tube ] .
in this section , we stick to the case of @xmath6 and investigate their duals . in @xcite , a front in @xmath24
is called _ flat _ , if its ( locally defined ) unit normal vector field @xmath89 degenerates everywhere . by lemma [ lem : degenerate - nu ] , we have that this flatness coincides with our definition of flatness ( cf .
section [ sec : ctb ] ) .
murata - umehara @xcite proved the following .
let @xmath368 be a constant curvature @xmath20 front .
if @xmath15 is an immersion , the gaussian curvature @xmath369 of the induced metric is identically @xmath20 . by the gauss equation @xmath370 ,
the extrinsic curvature @xmath371 vanishes on @xmath365 .
therefore , we also call a constant curvature @xmath20 front _ extrinsically flat _ ( or _ e - flat _ ) . by lemma [ lem : degenerate - nu ] ,
a front is e - flat if and only if its ( locally defined ) unit normal vector field @xmath89 degenerates everywhere .
it is known that any flat immersed surface in @xmath24 is regarded as an e - flat one in @xmath25 via a central projection ( cf .
example [ ex : dev - moebius ] ) and vise versa .
that is , their local properties are the same .
however , they may have different global properties .
for example , there are non - totally - geodesic complete flat immersions in @xmath24 ( cylinders over plane curves @xcite ) , while a complete e - flat immersion in @xmath25 must be totally geodesic @xcite .
similarly , in the case of fronts , the global properties of e - flat fronts in @xmath25 are not necessarily equal to those of flat fronts in @xmath24 .
for example , although complete flat fronts in @xmath24 with non - empty singular set must be co - orientable by fact [ fact : mu ] , we can not deduce the co - orientability of complete e - flat fronts in @xmath25 .
( theorem [ thm : main ] just implies their orientability . ) in fact , there exist compact non - co - orientable e - flat fronts in @xmath25 as follows .
[ ex : non - co - ori ] let @xmath372 be a developable tube as in . since @xmath373 , we may put @xmath374 , @xmath375 . and
hence , we have the following parametrization : @xmath376 we call a regular curve in @xmath25 whose curvature and torsion are constant _ helix_. * figure [ fig : gallery ] ( b ) shows a graphic of a developable tube @xmath377 whose center curve is a helix @xmath166 with @xmath378 . in this case , @xmath15 is co - orientable and compact .
* figure [ fig : non - co - ori ] shows a graphic of a developable tube @xmath377 whose center curve is a helix @xmath166 with @xmath379 . in this case
, @xmath166 is antiperiodic , and hence @xmath15 is non - co - orientable by proposition [ prop : dev - tube ] . in each case
, the restriction of @xmath15 to the regular point set is an embedding . by this example
, we can not expect any evaluation of the number of singular points other than cuspidal edges as in the case of complete flat fronts in @xmath24 ( cf .
fact @xcite ) .
[ ex : dev - moebius ] the central projection @xmath385 is a diffeomorphism defined by @xmath386 where @xmath387 is the hemisphere @xmath388 .
then the _ klein model _ of @xmath25 is @xmath24 equipped with the metric induced by @xmath389 .
it is known that if @xmath390 is a flat immersion , @xmath391 gives an e - flat immersion . since there exist
non - orientable flat surfaces in @xmath24 ( cf .
@xcite ) , there are non - orientable e - flat ones in @xmath25 through the klein model .
e - flat mbius strips are studied in @xcite .
let @xmath392 be a great circle in @xmath25 .
there exists a linear @xmath366-subspace @xmath393 of @xmath394 such that @xmath395 coincides with the image of @xmath396
. then we call a great circle @xmath397 the _ dual _ of @xmath396 , if the image of @xmath397 coincides with @xmath398 , where @xmath399 is the orthogonal complement of @xmath393 in @xmath394 . for a great circle @xmath400 @xmath401 , take an orthonormal basis @xmath402 of @xmath394
. then , the dual great circle @xmath397 of @xmath396 is parametrized as @xmath403 for an open interval @xmath404 , let @xmath405 be a ruled surface @xmath406 given by a curve @xmath407 . as in the case of great circles
, we may define the _ dual _ @xmath408 as @xmath409 where @xmath410 is an orthonormal frame .
it suffices to show that @xmath411 gives the unit normal of @xmath412 .
since @xmath413 is an orthonormal frame , the dual @xmath412 is given by @xmath414 thus @xmath415 hold , which proves the assertion .
[ rem : ins ] in ( * ? ? ?
* section 7 ) , izumiya - nagai - saji studied the dual of a ruled e - flat ( i.e. , developable ) surface .
we can check that our definition of duals coincides with the definition given by izumiya - nagai - saji @xcite .
we also remark that , since @xmath416 at a point @xmath417 ( cf . ) , _ duals of developable tube fronts are not necessarily fronts _ ( see figure [ fig : mu ] ) , where @xmath418 is the torsion function of @xmath166 .
now , we consider the fixed point set of the dual operation . for a regular curve @xmath166 in @xmath25 ,
the following is called the _ frenet - serret _ formula : @xmath419 where @xmath420 and @xmath364 are the curvature function and the velocity vector field @xmath421 of @xmath164 , respectively . if @xmath166 is congruent to @xmath411
, @xmath424 holds . on the other hand , by the frenet - serret formula
, we have @xmath425 .
thus , we have @xmath426 for each @xmath427
. we shall prove the converse .
without loss of generality , we may assume that @xmath428 for all @xmath427 .
by the frenet - serret formula , we have @xmath429 and @xmath430 thus , the curvature and torsion of @xmath411 are given by @xmath431 , @xmath20 respectively .
therefore , @xmath166 is congruent to @xmath411 .
the author would like to thank professors masaaki umehara , kotaro yamada , miyuki koiso , masatoshi kokubu , jun - ichi inoguchi , yu kawakami and kosuke naokawa for their valuable comments and constant encouragements .
he also expresses gratitude to professor udo hertrich - jeromin and gudrun szewieczek for careful reading of the first draft .
this work is partially supported by grant - in - aid for challenging exploratory research no .
26610016 of the japan society for the promotion of science . | in this paper , we give a definition of _ coherent tangent bundles of space form type _ , which is a generalized notion of space forms .
then , we classify their realizations in the sphere as a wave front , which is a generalization of a theorem of oneill and stiel : any isometric immersion of the @xmath0-sphere into the @xmath1-sphere of the same sectional curvature is totally geodesic . |
n49 is a bright supernova remnant ( snr ) in the large magellanic cloud ( lmc ) @xcite .
the low metal abundances of n49 @xcite indicated a substantial mixture of the supernova ( sn ) ejecta material with the lmc interstellar medium ( ism ) , which is consistent with the relatively old dynamical age as derived to be @xmath25000 @xmath0 10000 yr @xcite .
the most remarkable feature of n49 is strong enhancements of the optical / x - ray surface brightness in the south - east .
this asymmetric intensity distribution has been attributed to the interaction of the blast wave with the nearby molecular clouds located in the east of the snr @xcite .
the radio images show similar intensity distribution to the x - ray morphology , but with a complete shell in the west as well as in the east @xcite .
the radio spectral index of @xmath3 = @xmath00.58 @xcite is typical for a shell - type snr .
although the mean fractional polarization over the whole remnant is low ( @xmath45% ) , the western side of the snr is highly polarized . the highly organized magnetic field in the west of the snr is supportive of the less - disturbed blast wave expanding into the low density ism in the west @xcite .
vancura et al .
( 1992 ) have performed extensive optical / uv spectroscopy and have found that n49 is interacting with dense clouds ( @xmath5 @xmath2 150 @xmath6 on average ) while the preshock intercloud density is @xmath20.9 @xmath6 .
this dense ism is most likely associated with the co clouds detected in the east of the snr @xcite .
the bright optical / uv emission arises from `` sheets '' of emission generated by slow radiative shocks of @xmath7 @xmath8 140 km s@xmath9 as the blast wave interacts with the dense ambient material , while the x - ray emission appears from the fast shock ( @xmath7 @xmath2 730 km s@xmath9 ) propagating in the intercloud medium .
the x - ray spectrum of n49 as obtained with _
asca _ was fitted with a non - equilibrium ionization ( nei ) model with @xmath10 @xmath2 0.6 kev and @xmath11 @xmath2 10@xmath12 @xmath6 s @xcite .
the fitted electron temperature and the ionization age are consistent with the sedov evolution without significant alteration of the preshock ism by the stellar winds from the progenitor ( the model suggested by shull et al .
[ 1985 ] ) , which supports the overall picture proposed by vancura et al .
the overall x - ray spectrum was fitted with low metal abundances , which is also consistent with the optical / uv spectral analysis .
the position of the outburst of the soft gamma - ray repeater ( sgr ) sgr 0526@xmath066 in 1979 march 5 has been known to be within the boundary of n49 @xcite and is consistent with the unresolved x - ray source rx j05260.3@xmath0660433 @xcite , suggesting the identification of this source as the x - ray counterpart of sgr 0526@xmath066 .
this unresolved x - ray source appears spatially associated with snr n49 based on its positional coincidence within the boundary of the snr although an incidental alignment along the line of sight between the two can not be ruled out @xcite .
although an 8 s pulsation from sgr 0526@xmath066 has been detected during the @xmath13-ray burst in 1979 @xcite , the x - ray timing / spectral analysis of this unresolved x - ray source has been limited with previous detectors ( e.g. , marsden et al .
no counterparts in the infrared and radio bands have been found for sgr 0526@xmath066 @xcite .
these results from sgr 0526@xmath066 share typical characteristics of the anomalous x - ray pulsars ( axp ) , provided that the detection of the x - ray pulsar is confirmed .
high resolution x - ray imaging with the moderate spectral capability of _ chandra _ observations would be critical to confirm the detection of the x - ray counterpart of sgr 0526@xmath066 .
we here report the results from the imaging and spatially resolved spectral analysis of snr n49 from the _ chandra _ observations .
while this work is primarily devoted to the analysis of the diffuse x - ray emission from snr n49 , we also report some preliminary results from the detected point source , the x - ray counterpart of sgr 0526@xmath066 .
we describe the observation and the data reduction in [ sec : obs ] .
the image and spectral analyses are presented in
[ sec : analysis ] , and a discussion in
[ sec : disc ] . a summary will be presented in
[ sec : summary ] .
snr n49 was observed with the advanced ccd imaging spectrometer ( acis ) on board
the _ chandra x - ray observatory _ on 2001 september 15 as part of the _ chandra _ guaranteed time observation ( gto ) program ( obsid 1041 ) .
the acis - s3 chip was chosen to utilize the best sensitivity and energy resolution of the detector in the soft x - ray bands .
the pointing ( @xmath3@xmath14 = 05@xmath15 25@xmath16 [email protected] , @xmath18@xmath14 = @xmath065@xmath19
59@xmath20 06@xmath219 ) was roughly toward the x - ray centroid of the nearby lmc snr n49b with the aim point shifted by 245 detector pixels ( @xmath22@xmath20 ) from the nominal s3 chip aim point in order to simultaneously observe both of n49 and n49b on the same ccd chip .
this pointing direction results in an angular offset of @xmath26@xmath225 from the aim point at the position of n49 .
the results from the data analysis of the pointed target snr n49b are presented elsewhere @xcite .
we have utilized the data reduction techniques developed at the pennsylvania state university for correcting the spatial and spectral degradation of the acis data caused by radiation damage , known as charge transfer inefficiency ( cti ) @xcite .
the expected effects of the cti correction include an increase in the number of detected events and improved event energies and energy resolution @xcite .
we screened the data with the flight timeline filter and then applied the cti correction before further data screenings by status and grade .
we removed `` flaring '' pixels and selected _ asca _ grades ( 02346 ) .
the overall lightcurve was examined for possible contamination from time - variable background .
the overall lightcurve from the whole s3 chip shows relatively high background ( by a factor of @xmath232 ) for the last @xmath210% of the observation time .
this contamination from the variable background is however negligible for the n49 source region ( @xmath21@xmath224 diameter ) and we ignore it in the data analysis . after these data reduction process ,
the effective exposure was 34 ks that contains @xmath2241 k photons in the 0.3 @xmath0 8 kev band for the whole remnant within a 1@xmath224 diameter circular region . during _ chandra _ cycle 1 ( 2000 january 4 )
, n49 was observed with the acis - s3 as a guest observer ( go ) program , which is now publicly available in the archive ( obsid 747 ) .
the pointing was toward the position of sgr 0526@xmath066 ( @xmath3@xmath14 = 05@xmath15 26@xmath16 [email protected] , @xmath18@xmath14 = @xmath066@xmath19 04@xmath20 35@xmath210 ) . in this observation , instead of the whole ccd and the standard 3.2 s frame time , a 128-row subarray with a 0.4 s frame time was used for the temporal study of the target object sgr 0526@xmath066 . due to the selection of the subarray of the ccd chip ,
the coverage of the snr was incomplete and only the northern parts of the snr were observed .
the total exposure was @xmath240 ks . although the angular coverage is incomplete , this on - axis observation provides better angular resolutions than our gto data , thus we utilize these archival data as supplements in the analysis .
we have reprocessed the raw archival data in exactly the same way as the gto data were processed , including the cti correction .
the `` true - color '' x - ray images of snr n49 from the _
chandra_/acis observations are presented in figure [ fig : fig1 ] .
figure [ fig : fig1]a is the image from the gto observation and figure [ fig : fig1]b is the one from the archival go data .
the differences in the point spread function ( psf ) between two images due to the angular offset of @xmath26@xmath225 from the aim point in the gto data are apparent .
the high resolution on - axis archival data reveal the detailed diffuse filamentary structure , particularly in the east of the snr , far better than the off - axis gto data .
the detection of the point source as the counterpart of sgr 0526@xmath066 is evident in the middle of the northern part of the snr as shown by a bright white spot ( figure [ fig : fig1]b ) .
this point source is also detected in the off - axis image ( figure [ fig : fig1]a ) but appears extended because of the psf degradation . despite the superior angular resolution of the on - axis data , the selection of the subarray of the ccd allows only a partial coverage of the snr : i.e. , about a third of the snr in the south - west was not covered ( figure [ fig : fig1]b ) . the imaging and spectral analysis over
the whole remnant is important to understand the global nature of the snr and we primarily utilize our gto data for the data analysis . in order
to take advantage of the high angular resolution , we use the archival data for supplementary purposes . in figure
[ fig : fig1 ] , the soft x - ray emission ( red - yellow ) reveals various filamentary structures across the snr .
the circumferential shell - like soft x - ray features around the snr are spatially coincident with the radio shells ( figure [ fig : fig2]a ) , representing the blast wave shock front . by comparison ,
the hard x - ray emission features show little correlation with the radio intensities ( figure [ fig : fig2]b ) . based on astrophysical considerations
, we believe that it is reasonable to suppose that shocks are reflecting off the denser molecular gas in the east and may be propagating back into the interior of n49
. this could be a partial explanation for the higher temperatures we find in the spectral analysis of the eastern region ( [ subsubsec : east ] ) , although a definitive conclusion would require a detailed investigation of the hydrodynamics of n49 that is beyond the scope of this work .
other evidence , albeit more speculative , in support of reflected shocks in the eastern region of n49 include ( 1 ) the morphology of the hard x - rays which peak closer to the interior on the eastern side than the soft x - rays do ( figure [ fig : fig2 ] ) , and ( 2 ) the clear arc - like feature in the si ew map ( [ subsec : ew ] and figure [ fig : fig5]f ) which traces hot gas , that perhaps delineates the edge of the inward moving reflected shock .
the western half is , for the most parts , dominated by hard x - ray emission ( green - blue ) , which may indicate that the blast wave is propagating through a low density ism , as suggested by radio data @xcite .
the brightest diffuse emission is in the southeast , as recognized by previous x - ray observations , where the snr is interacting with dense molecular clouds .
the overall bright x - ray emission in the eastern half of the snr is significant in all colors and likely has complex spectral structure .
while filamentary soft , red - yellow emission prevails in the eastern half of the snr , there are apparent blue regions in the east ( primarily to the immediate north of the southeast brightest region ; figure [ fig : fig1 ] ) .
this diffuse x - ray emission with spectral variations appears localized to produce spatially distinctive layers of x - ray colors in the eastern half of the snr .
the interactions of the blast wave shock with the clumpy dense molecular clouds in these regions appear responsible for the observed spatio - spectral substructures of the diffuse x - ray emission ( see [ subsec : spec ] ) .
for example , the bright filametary emission features from the optical oiii images are largely anti - correlated with the blue , hard x - ray intensity and more correlated with the soft , red x - ray emission ( figure [ fig : fig3]a and figure [ fig : fig3]b ) .
the optical emission from n49 is believed to originate primarily from slow radiative shocks entering dense molecular clouds , while the x - ray emission is in general from fast shocks propagating through the intercloud medium @xcite . the high resolution x - ray images from the acis data reveal that the hard x - rays are indeed from intercloud regions where little optical emission originates , and that the soft x - ray emission is spatially more coincident with the bright optical emitting regions .
these features support the interactions of the blast wave with the clouds as the origin of the x - ray spectral variations observed in the color images .
we detect an interesting , previously unknown feature of n49 with our _ chandra _ gto data : i.e. , the blue , hard x - ray knot in the southwest , beyond the main boundary of the snr ( figure [ fig : fig1]a ) , from which faint radio emission is also associated ( figure [ fig : fig2]b ) .
the extended shape of this feature appears caused by the off - axis psf and the true morphological nature of this emission is uncertain with the current off - axis data .
we can not utilize the high resolution archival data , because this feature lies outside the field of view ( figure [ fig : fig1]b ) .
we find no proper counterparts for this feature in other wavelengths , which suggests that it is unlikely a projection of a background / foreground point source ( [ subsubsec : bullet ] ) .
the x - ray spectrum of this feature indicates metal - rich thermal emission with strong atomic emission lines rather than a power law dominated spectrum ( [ subsubsec : bullet ] ) .
these spectral features support the proposition that this hard x - ray knot is a part of the snr , probably fragment(s ) of sn ejecta .
the overall spectrum of n49 shows some broad atomic emission lines ( figure [ fig : fig4 ] ) and we generate _ equivalent width _ ( ew ) images of k - shell line emission from the atomic elements o , ne , mg , si , s , and the l - shell line emission of fe by selecting photons around the broad line complexes ( table [ tbl : tab1 ] ) , following the methods described by park et al .
( 2002a ) .
images in the bandpasses of interest for both the lines and continuum were extracted with 2@xmath24 pixels and smoothed by a gaussian with @xmath25 = 7@xmath24 .
the underlying continuum was calculated by logarithmically interpolating between images made from the higher and lower energy `` shoulders '' of each broad line .
the estimated continuum flux was integrated over the selected line width and subtracted from the line emission .
the continuum - subtracted line intensity was then divided by the estimated continuum on a pixel - by - pixel basis to generate the ew images for each element . in order to avoid noise in the ew maps caused by poor photon statistics near the edge of the snr
, we have set the ew values to zero where the estimated continuum flux is low .
we also set the ew to zero where the integrated continuum flux is greater than the line flux .
we present the ew images for o , ne , mg , si , and fe ( figure [ fig : fig5 ] ) .
the s ew image is unreliable due to the low photon statistics .
the broad line complexes at @xmath20.8 @xmath0 1.1 kev are from various ionization states of ne k and fe l - shell line emission , and we present tentative identifications in figure [ fig : fig5 ] and table [ tbl : tab1 ] .
although mg and si lines show he - like and h - like substructures in the overall spectrum ( figure [ fig : fig4 ] ) , we present the ew images from the combined broad lines in figure [ fig : fig5 ] since the h - like lines of these elements are too weak to separately produce useful ew images over the whole snr .
the o ew distribution is generally consistent with the optical oi , iii intensity distributions and shows higher ew in the east and south of the snr ( figure [ fig : fig5]a ; vancura et al .
1992 ) . the fe ew ( figure [ fig : fig5]d ) shows similar distribution to the o map .
ne / fe and mg ew ( figure [ fig : fig5]b , figure [ fig : fig5]c , and figure [ fig : fig5]e ) distributions appear to be opposite to o , showing higher intensities in the west of the snr .
the si ew is relatively strong in both east and west ( figure [ fig : fig5]f ) .
the o and fe ew enhancements are spatially correlated with the location of the dense molecular clouds where the x - ray and optical emission is generally bright ( figure [ fig : fig5]a and figure [ fig : fig5]d ) . since the x - ray emission in these regions
is presumably dominated by the shocked dense ism , these high o and fe ews are unlikely due to abundance enhancements , but appear to trace shock temperature variations .
optical spectroscopic studies have indicated that metal abundances of n49 are close to the lmc values @xcite and support temperature variations , rather than abundances , as the origin of these ew distributions .
based on the plane - parallel shock model with a moderate ionization timescale ( @xmath11 @xmath2 10@xmath26 @xmath6 s ) and the lmc abundances , the fe l ew ( figure [ fig : fig5]d ) is in fact representative of the broad line complex at @xmath27 @xmath2 0.8 kev , which is characteristic of a `` cool '' plasma ( @xmath10 @xmath28 0.5 kev ) as opposed to the `` hot '' component ( @xmath10 @xmath2 1 kev ) that would produce a prominent fe l line complex at @xmath27 @xmath21 kev . in direct contrast , the h - like ne ( and/or hot fe l ) ew ( figure [ fig : fig5]c ) , which is sensitive to high temperatures of @xmath10 @xmath2 1 kev , is enhanced in the west where the blast wave is expanding into a low density ism .
these ew distributions may thus indicate the presence of an _ undisturbed _
( by dense ism ) blast wave shock in the west , which may have a higher temperature than that in the eastern side of the snr . if the dominant contributor to the enhanced ew at @xmath27 @xmath2 1.1 kev is ne , comparisons between he - like and h - like ne ew images ( figure [ fig : fig5]b and figure [ fig : fig5]c ) may suggest progressive ionizations of ne by the reverse shock as detected in other snrs @xcite .
the si ew enhancements in both east and west are intriguing and may suggest the presence of hot phases in addition to cool phase shocks in the south - east parts of the snr ( see more discussion in [ subsec : spec ] ) .
the si ew enhancements in the southeastern quadrant are largely shell - like , as indicated by a dotted line in figure [ fig : fig5]f .
this may also be interpreted as hot , highly ionized plasma in the southeastern regions of the snr .
the hard x - ray knot in the southwest beyond the snr boundary , as revealed with the true - color image , shows strong enhancements in the si , mg , and hot ne / fe ew images .
the actual spectral analysis indicates that these high ews are caused by metal abundances as well as advanced ionizations and high temperature ( see [ subsubsec : bullet ] ) .
the enhanced ews for the elemental species caused by high metal abundances support that this hard x - ray knot may not be a background extragalactic source , and that it is likely emission from metal - rich sn ejecta associated with n49 .
the ew around the detected point source is , on the other hand , very low for all elemental species .
these low ews are strong indications of the continuum - dominated x - ray spectrum for the point source .
this is in good agreements with the presumed identification of the point source as the neutron star , or the x - ray counterpart of sgr 0526@xmath066 in the quiescent state .
the high resolution acis data , for the first time , allow us to perform spatially - resolved spectral analysis in x - rays for the study of snr n49 . despite the psf degradations due to the large angular offset from the aim point , the gto data provide moderate angular resolution of a few arcsec , which are still sufficient to resolve major diffuse structure of n49 .
moreover , the complete spatial coverage of the snr with the gto data is advantageous over the archival data for the efficient and consistent study of the global structures of the snr .
combining two data sets may not be desirable because of the significant difference in the psf .
we thus make a simple and secure approach for the spatially - resolved spectral analysis by utilizing the gto data as the primary data and the archival data as the supplementary data : i.e. , we perform all spectral analysis with the gto data and then verify the results with the archival data as necessary .
we extract the spectrum from several small regions over the snr .
the regional selections were based on the characteristics revealed by the broad band color images and the ew images , and are presented in figure [ fig : fig6 ] . in all spectral fittings , we made a low energy cut at the photon energy of @xmath27 = 0.5 kev because of unreliable calibrations of the acis spectrum in the soft bands .
in most cases , we bin the data to contain a minimum of 20 counts for the spectral fitting . for the spectral analysis of our
cti corrected data , we have utilized the response matrices appropriate for the spectral redistribution of the ccd , as generated by townsley et al .
( 2002b ) .
the low energy ( @xmath27 @xmath4 1 kev ) quantum efficiency ( qe ) of the acis has been degraded because of the molecular contamination on the optical blocking filter .
we have corrected this time - dependent qe degradation by modifying the ancillary response function ( arf ) for each extracted spectrum by utilizing the idl version of acisabs software developed by chartas et al .
( 2002 ) .
we select three regions from the western half of the snr ( regions 1 , 2 , and 3 ; figure [ fig : fig6 ] ) .
these regions are where the broad band surface brightness is relatively low , while the ew images reveal some noticeable features ( figure [ fig : fig5 ] and figure [ fig : fig6 ] ) .
region 1 is where o and ne ews are enhanced and contains @xmath23500 photons .
region 2 is where ne / fe , mg and si are enhanced containing @xmath28100 counts .
region 3 shows low ew for most of the elements and contains @xmath23200 photons .
we fit these spectra with the plane - parallel shock model @xcite as presented in figure [ fig : fig7]a , figure [ fig : fig7]b , and figure [ fig : fig7]c .
the elemental abundances were fixed for he (= 0.89 ) , c (= 0.30 ) , n (= 0.12 ) , ca (= 0.34 ) and ni (= 0.62 ) at values for the lmc @xcite ( hereafter , all abundances are with respect to the solar ; anders & grevesse 1989 ) since the contribution from these species in the spectral fitting is expected to be insignificant in the selected energy range .
other elements were allowed to vary .
the results of the spectral fittings for the west regions of the snr are summarized in table [ tbl : tab2 ] and table [ tbl : tab3 ] .
these spectra in the western half of the snr can be described by a thermal plasma with the best - fit @xmath10 of 0.52 @xmath0 0.65 kev .
the implied absorptions are @xmath20.6 @xmath02.7 @xmath29 10@xmath30 @xmath31 .
these results are consistent with the results from the _ asca _ data @xcite .
the fitted abundances are generally consistent with the ew distributions .
the results from the region 2 spectrum are perhaps the most interesting .
this is where ne / fe , mg , and si ews are enhanced and the best - fit abundances for those elemental species ( except for fe ) are also relatively high , although the magnitudes of the enhancements are moderate only at around the solar level .
the fitted abundances for this region suggest that the high ew responsible for the emission line at @xmath27 @xmath2 1.1 kev ( figure [ fig : fig5]c ) is due to highly ionized h - like ne rather than hot fe .
the best - fit shock temperature and the ionization timescale for this region are higher than region 1 in which he - like ne ew is enhanced ( figure [ fig : fig5]b ) .
this supports the advanced ionization in region 2 as suggested by the ew images .
the region 3 spectrum is on the other hand fitted with low lmc - like abundances for all elements .
the broad band x - ray surface brightness is also low for this region , indicating x - ray emission from the shocked low density lmc ism .
we select two regions from the eastern half of the snr ( regions 4 and 5 ; figure [ fig : fig6 ] ) .
region 4 is within the bright eastern rim which is evidently dominated by the hard x - ray emission as displayed in figure [ fig : fig1 ] .
a total of @xmath210000 photons are extracted from region 4 .
region 5 is the brightest knot in the southeast , where we extract @xmath217000 counts .
we fit these spectra in the same way as we did for the western regions : i.e. , we use the plane - parallel shock model with o , ne , mg , si , s , and fe abundances variable in the fittings .
both of the spectra in the eastern half of the snr show significant si ly@xmath3 line ( @xmath27 @xmath2 2.0 kev ) as well as strong soft x - ray emission ( @xmath32 1 kev ) ( figure [ fig : fig7]d and figure [ fig : fig7]e ) . in order to adequately describe these spectral features , a hot component shock with @xmath10 @xmath23 1 kev
is required in addition to the soft component with @xmath10 @xmath2 0.4 @xmath0 0.5 kev .
the hot components have larger ionization timescales ( @xmath11 @xmath23 10@xmath26 @xmath6 s ) than the cool components ( @xmath11 @xmath2 10@xmath12 @xmath6 s ) .
the fitted ionization timescales for the hot component are not constrained in the upper bounds and a thermal plasma in the complete collisional ionization equilibrium ( cie ) may also fit the data with the equivalent statistical significance .
the implied absorbing columns are @xmath33 @xmath2 3 @xmath0 4 @xmath29 10@xmath30 @xmath31 for these regions .
the soft component model for the region 5 spectrum indicates low metal abundances of @xmath28 0.5 solar for fitted elemental species .
this soft component dominates the x - ray emission at @xmath27 @xmath28 1 kev , which consistently supports that the enhanced o and fe ews ( figure [ fig : fig5]a and figure [ fig : fig5]d ) are not caused by high abundances .
the best - fit model for the hard component implies overabundances ( typically @xmath2 1 @xmath0 2 solar ) compared with the lmc abundances for some elemental species .
the statistical uncertainties for the fitted abundances are however large enough that the reality of these high abundances are inconclusive with the current data .
for example , this region 5 spectrum may also be fitted with subsolar abundances , for both of the soft and hard components of the shock , with equivalent statistical significance based on the f - test .
the ews distributions in the east of the snr are therefore more likely dominated by the shock temperature and the ionization states rather than by metal abundances .
the fitted model parameters for the region 4 spectrum are also generally consistent with the results from region 5 , which also supports the temperature and ionization states for the origins of the observed ew distributions .
we note that region 4 ews appear relatively weak for the atomic lines at low photon energies ( o and fe ; figure [ fig : fig5]a and figure [ fig : fig5]d ) and stronger in higher energies ( si ; figure [ fig : fig5]f ) compared with region 5 , even though there is no significant difference in the fitted shock parameters between these two regions .
these features may imply a difference in the local interstellar environments due to the clumpy structure of the interacting dense molecular clouds : i.e. , a density gradient of the clouds between these two regions might have caused deficient soft x - ray emission in region 4 as displayed by the lack of red , soft emission there ( figure [ fig : fig1 ] ) .
the anticorrelations between optical and hard x - ray emission are supportive of this case .
the fitted volume emission measure ( @xmath34 ) of the soft component plasma for the region 5 spectrum is indeed significantly higher than that for region 4 , which also supports this interpretation .
the results of the spectral fittings for the east regions of the snr are summarized in table [ tbl : tab4 ] .
the hard x - ray knot in the southwest beyond the main boundary of the snr ( region 6 in figure [ fig : fig6 ] ) contains @xmath2800 counts and we binned the spectrum to contain a minimum of 12 counts per bin .
the spectrum can be fitted with a single - temperature plane - parallel shock ( figure [ fig : fig7]f ) .
the spectrum features strong atomic emission lines and is most likely emission from metal - rich ejecta .
the strong si ly@xmath3 line ( @xmath27 @xmath2 2.0 kev ) constrains the plasma temperature to @xmath10 @xmath23 1 kev .
although statistical uncertainties are large due to the poor photon statistics of this faint feature , the best - fit model indicates strong enhancements in metal abundances at several times larger than solar abundances .
the broad band line features between @xmath20.8 kev and @xmath21.2 kev represent fe l line complex as well as he - like and/or h - like ne k lines . with the photon statistics of the current data ,
it is difficult to identify these individual atomic lines .
we have thus attempted fitting the spectrum by assuming dominant contributions from ne or fe as well as from both elements .
the overall fits are statistically indistinguishable among these cases .
we however note that the model fits fe as dominant contributor ( fe abundance @xmath2 0.8 , ne @xmath2 0 ) when both ne and fe are simultaneously allowed to vary . even with fe abundance fixed at lmc value , the fitted ne abundance is still low ( @xmath20.4 ) .
we therefore obtain the best - fit model by allowing fe abundance to vary with ne fixed at the lmc abundance .
the fitted parameters are summarized in table [ tbl : tab2 ] and table [ tbl : tab3 ] .
although the metal - rich spectrum indicates that this feature is likely emission from shocked sn ejecta , the isolated morphology of this hard x - ray knot outside of the main boundary of the snr may suggest the possibility of a projected background object .
the source position for the knot is @xmath3@xmath14 = 05@xmath15 25@xmath16 [email protected] , @xmath18@xmath14 = @xmath066@xmath19 05@xmath20 13@xmath215 , and the acis astrometric uncertainties at @xmath26@xmath20 off - axis is generally @xmath282@xmath24 ( e.g. , feigelson et al .
the _ chandra_/acis position of sgr 0526@xmath066 based on the on - axis data has a conservative astrometric uncertainty of @xmath282@xmath213 ( kaplan et al .
2001 and references therein ) .
the position of the sgr of our off - axis data as determined by _ wavdetect _ algorithm is @xmath21@xmath218 offset from the on - axis sgr position .
the source positions with our off - axis data therefore appear to have @xmath44@xmath24 accuracy as a conservative limit .
we have searched for counterparts around the source position in simbad and ned catalogs as well as @xmath2100 multi - wavelengh on - line catalogs available through heasarc .
we find no candidate for an extragalactic counterpart within 1@xmath20 radius of the source position .
the derived x - ray flux is @xmath35 @xmath2 6.0 @xmath29 10@xmath36 ergs s@xmath9 @xmath31 and @xmath37 @xmath2 1.5 @xmath29 10@xmath36 ergs s@xmath9 @xmath31 .
based on the logn @xmath0 logs relations of the _ chandra _ deep field @xcite , these x - ray fluxes imply the probability of only @xmath2810@xmath38 within a generous area of 5@xmath24 radius around the source position for a coincidental detection of an extragalactic object .
we find an lmc star , macs j0525@xmath0660#037 @xcite which has an angular offset from the position of the hard x - ray knot of @xmath26@xmath212 .
considering the positional accuracy of the macs catalog ( @xmath20@xmath215 ; tucholke et al .
1996 ) , the apparent angular offset between the hard x - ray knot and optical position of the macs star is significant .
we find another nearby lmc star within the ob association lh 53 as cataloged ( i d # 1630 ) by hill et al .
the relatively small angular offset ( @xmath22@xmath24 ) of this star from the x - ray hard knot position suggests that this star could be the optical counterpart of the x - ray knot .
the derived x - ray luminosity of @xmath39 @xmath2 3.4 @xmath29 10@xmath40 ergs s@xmath9 ( at the distance of 50 kpc ) for the x - ray knot is however higher than that of typical late - type young stellar objects ( @xmath39 @xmath2 10@xmath41 @xmath0 10@xmath42 ergs s@xmath9 ; e.g. , imanish et al .
2001 ) by a few orders of magnitude
. a massive early - type star should be brighter ( @xmath39 @xmath2 10@xmath43 ergs s@xmath9 ; e.g. , seward & chlebowski 1982 ) , but may also not match the derived high @xmath39 of the hard x - ray knot .
moreover , although there are @xmath280 lh 53 stars within 1@xmath20 radius of this hard x - ray knot position @xcite , none of them is detected with our acis observation at the flux level of this hard x - ray knot .
it is unlikely for only one star to be detected with the acis at this particular position at this flux level .
if this object is a foreground young late - type galactic star with a typical x - ray luminosity of @xmath39 @xmath2 10@xmath41@xmath010@xmath42 ergs s@xmath9 , this object _ must _ be nearby with a distance of @xmath230 @xmath0 300 pc , and a v magnitude of @xmath211 @xmath0 14 is expected from the measured x - ray flux .
we have , however , found no such bright stars in galactic star catalogs .
we find a faint star in the guide star catalog with an @xmath22@xmath24 offset from the source position , but the cataloged b magnitude is only 19.44 .
the best - fit absorbing column from the x - ray spectral modeling may also be higher ( @xmath33 = 1.1 @xmath29 10@xmath30 @xmath31 ) than the galactic absorption toward the position of this hard x - ray knot ( @xmath33 @xmath2 6 @xmath29 10@xmath44 @xmath31 ; dickey & lockman 1990 ) .
this hard x - ray knot is therefore most likely a feature spatially associated with snr n49 .
this feature then appears to be fragments from the sn explosion ejected outside of the snr shock boundary or `` shrapnel '' as discovered in other galactic snrs such as vela and tycho @xcite .
the true angular size and shape of this hard x - ray knot is however uncertain because of the poor off - axis psf .
we note that an on - axis _ chandra_/acis observation with a full angular coverage of n49 , targeted on sgr 0526@xmath066 has recently been performed ( obsid 2515 ) .
the detailed results from the diffuse x - ray emission of the snr have yet to be reported .
the spectrum from the point - like source is extracted from region 7 ( figure [ fig : fig6 ] ) . due to the angular offset from the aim point ,
the psf at the position of the point source is significantly degraded ( @xmath232@xmath24 ) , and we select a relatively large area ( a circular region with a radius of 6@xmath24 ) , which contains @xmath214000 photons .
although the total count rate ( @xmath20.4 counts s@xmath9 ) is relatively high , pileup is insignificant due to the large off - axis angle .
since the source location is within fairly bright diffuse emission from the snr , the degraded psf also makes the background estimation difficult .
we thus include a thermal component for the background spectrum in the spectral fitting instead of subtracting the background from the source spectrum .
in a direct contrast to the diffuse emission spectra , the point source spectrum is dominated by continuum , as perhaps expected from a neutron star ( figure [ fig : fig8]a ) .
the source spectrum can be described with a single power law model of photon index @xmath1 = 2.98@xmath45 .
the embedded background thermal component is a @xmath10 @xmath2 0.33 kev plasma with lmc - like metal abundances , which is primarily responsible for the soft emission at @xmath27 @xmath4 1 kev .
a black body model may also fit the data ( @xmath10 @xmath2 0.52 kev ) with acceptable statistics , but appears , qualitatively , inappropriate to describe the hard tail of the specrum .
since the archival data are pointed in the direction of the position of the point source or sgr 0526@xmath066 , the on - axis high angular resolution allows us to manipulate the background subtraction with less confusion .
we extract the source spectrum from a circular region with a radius of 2@xmath24 , which contains @xmath210000 photons . because of the utilized short time - frame ( 0.4 s )
, pileup is negligible with this on - axis observation .
sgr 0526@xmath066 is on top of complex filamentary structure of the diffuse emission from the snr and the background emission for the sgr is non - uniform ( figure [ fig : fig1]b ) .
we thus extract the background emission from the immediate annular region centered on the point source position ( inner radius of 2@xmath215 and outer radius of 4@xmath24 ) , in order to obtain an `` average '' spectrum over the variable emission features around the sgr . the sgr spectrum from the on - axis data
can be fitted with a single power law model ( figure [ fig : fig8]b ) .
the best - fit photon index is @xmath1 = 2.72@xmath46 which is in good agreement with the result from the off - axis data .
the results from the fit with a black body model ( @xmath10 @xmath2 0.50 kev ) are also identical to the off - axis data and appear inadequate by the reason of the same hard - tail discrepancy .
the implied x - ray luminosity is @xmath39 = 2.1 @xmath29 10@xmath47 ergs s@xmath9 in the 2 @xmath0 10 kev band with the off - axis data ( @xmath39 = 2.5 @xmath29 10@xmath47 ergs s@xmath9 based on the on - axis data ) , which is consistent with the asca upper limit @xcite and that of the galactic sgr 1806@xmath020 @xcite .
the results from the spectral model fits of the point source are summarized in table [ tbl : tab5 ] .
the x - ray color images and the ew images indicate complex spectral structure of the diffuse x - ray emission across the snr n49 .
the filamentary and clumpy emission features in the eastern half of the snr show spatially distinctive bright structures in different colors .
the spatial and spectral structure of the x - ray emission in the western half is relatively uniform , primarily represented by intermediate - to - hard emission ( green - blue ) with low surface brightness .
the soft ( red - yellow ) x - ray shell - like emission features represent the blast wave shock front as observed in the radio band
. the overall features of the ew images indicate enhancements of soft line emission ( @xmath32 1 kev ) in the south - east and hard emission lines ( @xmath48 1 kev ) in the west of the snr .
these features are consistent with the overall density distributions of the ambient ism ; i.e. , the blast wave is propagating into a low density ism in the west and interacting with dense clouds in the east .
the comparisons between the x - ray and optical images also suggest that the hard x - ray emission ( @xmath48 1.6 kev ) appears to be produced in less - dense intercloud regions while the soft x - rays ( @xmath32 0.75 kev ) originate near the dense clouds .
the results from the spatially - resolved spectral analysis confirm the speculations inferred from the image analysis by revealing that shock temperatures and ionization states in the east of the snr are significantly different from those in the western half : i.e. , the western portion of the snr is described by a simple one - temperature plane - parallel shock with relatively large ionization timescales while two - temperature model with multiple ionization timescales is required in the eastern half .
the global spectral structure of the snr may then be represented by a multi - phase shock model with three characteristic temperatures , i.e. , the `` cool '' ( @xmath10 @xmath2 0.44@xmath49 kev ) and `` hot '' ( @xmath10 @xmath2 1.25@xmath50 kev ) components in the east , and the `` intermediate '' ( @xmath10 @xmath2 0.59@xmath51 kev ) component in the west . the presence of multiple phases of the shock due to the interactions of the blast wave with the dense molecular clouds has been put forth by hester & cox ( 1986 ) . recently , based on the high resolution _
chandra_/acis observation of the galactic snr cygnus loop , levenson et al .
( 2002 ) reported evidence for the three distinct regions characterised as : ( i ) the forward shock decelerated by the clouds , ( ii ) the postshock snr interior , and ( iii ) the reflected shock that re - heated the postshock regions . in this context , we propose that the cool component of x - ray emission from n49 corresponds to the forward shock entering the dense clouds , the intermediate component to emission from the postshock region , and the hot component to emission from the re - heated postshock region , respectively . comparisons of the x - ray emission features with optical images indicate that the soft x - rays are spatially coincident with the bright optical knots while the hard x - ray emission is anti - correlated with the optical intensity distribution .
since the bright filamentary optical emission features are believed to originate from the slow , cool radiative shock entering dense molecular clouds , the observed correlation / anti - correlation between x - ray and optical images provide good support for the suggested multi - phase shock interpretation .
hester et al . (
1994 ) have developed a simple 1-dimensional model in order to describe such a multi - phase shock structure of the nearby galactic snr cygnus loop , probably a `` magnified '' version of the distant snr n49 .
our best - fit electron temperatures indicate that the ratios of temperatures between the re - heated postshock and the postshock without re - heating are @xmath52 = @xmath21.8 @xmath0 2.3 , which implies the preshock cloud density @xmath53 @xmath2 30 @xmath0 2500 @xmath54 based on the model by hester et al .
( 1994 ) , where @xmath54 is the preshock intercloud density . assuming the preshock intercloud density @xmath54 @xmath2 0.9 @xmath6 as estimated from the optical spectroscopy @xcite , we derive @xmath53 @xmath2 27 @xmath0 2300 @xmath6 for the interacting dense molecular clouds .
this is in plausible agreements with the previous estimations based on optical / uv spectroscopy ( @xmath53 = 20 @xmath0 940 @xmath6 ; vancura et al .
the best - fit @xmath34s in the eastern regions imply ( @xmath53/@xmath54)@xmath55(@xmath56/@xmath57 ) = 2 @xmath0 13 , where @xmath56 is the x - ray emission volume of the cool forward shock component and @xmath57 is the x - ray emission volume of the hot re - heated component ( we assume an x - ray emitting volume filling factor of @xmath21 , hereafter ) .
provided that @xmath53/@xmath54 is in the order of @xmath210@xmath58 @xmath0 10@xmath59 , this implies @xmath56/@xmath57 @xmath28 0.01 .
the soft x - ray emission from the cool forward shock is then from the blast wave entering small - volume , clumpy clouds while the hot component originates from the large - volume , re - heated intercloud postshock regions .
for an adiabatic shock , the relations between the shock temperature and the shock velocity is @xmath60 = @xmath61 and can be written as @xmath62 where @xmath60 is the shock temperature , @xmath63 is the shock velocity , and mean atomic mass @xmath64 @xmath65 0.7@xmath66 ( @xmath66 = proton mass ) .
the cool component shock velocity is then @xmath67 = 582@xmath68 km s@xmath9 , the hot component velocity is @xmath69 = 981@xmath70 km s@xmath9 , and the intermediate velocity is @xmath71 = 674@xmath72 km s@xmath9 . for comparisons ,
vancura et al . (
1992 ) have derived a shock velocity of @xmath7 = 730 km s@xmath9 for the x - ray emission of n49 ( note : for simplicity , we assumed the electron and ion temperatures in equalibration in these calculations ) .
the ionization timescale ( @xmath11 ) ratios between the cool and the hot components are ( @xmath11)@xmath73/(@xmath11)@xmath74 @xmath28 0.1 , which implies @xmath75 @xmath28 0.01@xmath76 , where @xmath75 is the ionization age of the shock entering the clouds and @xmath76 is the ionization age of the re - heated regions . assuming an sn explosion energy of @xmath77 = 10@xmath78 ergs , the snr radius of @xmath79 = 10 pc based on the x - ray image of n49 , and the preshock density @xmath54 = 0.9 @xmath6 , the implied sedov dynamical age @xcite , @xmath80 is @xmath2 6600 yr . assuming @xmath76 is on the order of @xmath80 or @xmath210@xmath81 yr , the measured @xmath11 ratios imply @xmath75 @xmath28 100 yr .
this short ionization age suggests that the soft x - ray emission ( @xmath10 @xmath2 0.44 kev ) has been produced by a rapidly decelerated shock entering dense molecular clouds as was observed from the cygnus loop snr @xcite . the hard x - ray knot in the southwest beyond the snr blast wave shock front ( region 6 ; figure [ fig : fig6 ] )
is most likely associated with the snr rather than a background / foreground object .
the x - ray spectrum is dominated by emission lines and can be described with a @xmath10 @xmath2 1 kev plasma with metal - rich abundances , which supports the shocked ejecta origin .
the fitted plane - parallel shock model indicates a highly advanced ionization state ( @xmath11 @xmath25 @xmath29 10@xmath82 @xmath6 s@xmath9 ) and the spectrum may also be equally described with a thermal plasma in the cie .
the high temperature and the large ionization timescale of this feature are also supportive of its origin as the high - speed ejecta fragments created at the time of the sn explosion . because of the psf degradation
, the apparent angular extension of this feature ( @xmath25@xmath24 in radius ) is likely an artifact and the true morphology is uncertain with the current data . in order to estimate the electron density @xmath83 ,
we thus assume two extreme cases of a circular shape with a radius of 5@xmath24 ( for the apparent angular size and so 1.2 pc radius in physical size at the distance of @xmath84 = 50 kpc ) and 0@xmath215 ( for the case of an unresolved point - like source , and so the physical size of @xmath28 0.125 pc radius at @xmath84 = 50 kpc ) .
the assumed range of the physical size of the knot ( @xmath20.1 pc @xmath0 1.2 pc in radius ) is reasonably consistent with the sn ejecta fragments detected in other galactc snrs of vela and tycho @xcite .
assuming the distance of 50 kpc to the snr and a simple spherical geometry for the x - ray emitting volume , the best - fit @xmath34 implies @xmath83 of 2.3 @xmath6 to @xmath23 68 @xmath6 depending on the assumed emission volume . based on the best - fit abundances
, we may derive the ejecta mass of @xmath20.001 @xmath85 @xmath0 0.03 @xmath85 .
although this range of the ejecta mass is in reasonable agreements with other galactic sn ejecta fragments @xcite , we caution that these estimations are crude due to the large uncertainties in the fitted abundances and assumed emission volumes .
the angular offset of the hard x - ray knot from the geometrical center of n49 is @xmath248@xmath24 .
the spatial separation of this knot from the center of the snr is then @xmath86 = 11.6 @xmath87 ( @xmath88)@xmath9 pc , where @xmath87 is the distance to n49 in the units of 50 kpc , and @xmath89 is the angle between the line of sight and the moving direction of the knot . assuming that the sn explosion site , so the birth place of the observed ejecta fragments , is close to the geometrical center of the current x - ray remnant of n49 , the mean transverse velocity is @xmath90 = 1717 @xmath91 @xmath92 km s@xmath9 , where @xmath91 is the spatial separation of the knot from the geometrical center of the snr in the units of 11.6 pc , and @xmath93 is the age of the snr in the units of 6600 yr . the analysis on this hard x - ray knot is however substantially limited by the poor psf .
follow - up on - axis _
chandra _ observations would be essential to unveil the detailed nature of this feature .
chandra_/acis observations conclusively detect an x - ray point source at the position of sgr 0526@xmath066 within the boundary of snr n49 .
the x - ray spectrum of this point source is dominated by continuum , which can be described with a power law model ( see [ subsubsec : sgr ] ) .
this point source is thus confidently identified with the x - ray counterpart of quiescent phase sgr 0526@xmath066 .
the single power law model indicates a soft spectrum with @xmath1 @xmath2 3 for sgr 0526@xmath066 , which resembles the typical spectrum of the axps ( mereghetti et al .
2002 and references therein ) , while the x - ray spectra from galactic sgrs are described with harder single power laws of @xmath1 @xmath2 2 @xcite . as x - ray spectra from the axps and sgrs can be described with the power law + black body models , we have made an attempt to fit the sgr 0526@xmath066 spectrum with such a multi - component model . with the off - axis gto data , the fitted black body temperature ( @xmath10 @xmath2 0.55 kev ) is consistent with typical axp / sgr spectra , and
the improvement in the overall fit is also statistically significant based on the f - test .
the power law component , with the addition of the black body component , however implies substantially harder spectrum ( @xmath1 @xmath2 1.7 ) than those of the axps ( @xmath1 @xmath2 3@xmath04 ; mereghetti et al .
2002 and references therein ) .
nonetheless , the two component model is still in good agreements with galactic sgrs as they are fitted with @xmath1 @xmath2 1@xmath02 and black body temperature @xmath10 @xmath2 0.5 kev @xcite .
we however note that this two component model is not required to describe the on - axis data .
since the embedded confusion level caused by the background in the spectral model fitting is more severe for the off - axis data than the on - axis data , we consider that the presence of the black body component is inconclusive , although the improvement in the fit for the off - axis data is statistically significant with the additional black body component .
recently , kulkarni et al . ( 2002 ) have reported the results from the spectral analysis of sgr 0526@xmath066 utilizing two _ chandra_/acis go observations : one ( obsid 747 ) of those two data sets was identical to that we uitilze in the current work .
they found that the observed spectrum of the sgr was best described by a black body + power law model , which appears to be inconsistent with our results .
although the origin of the inconsistency is unclear , we may attribute the descrepancy to the background estimations utilized in the spectral analysis . as kulkarni et al .
( 2002 ) pointed out , their sgr source spectrum apparently include residual emission from the diffuse snr , whereas such a contamination in our on - axis source spectrum appears to be negligible .
the soft x - rays from the diffuse emission of the snr might have caused the slight differences in the results of the spectral analysis as presented by kulkarni et al .
( 2002 ) . we have also noticed that kulkarni et al .
( 2002 ) have not incorperated the cti and qe corrections in their data reduction . although the substantial contributions by the cti and the qe degradation effects on a point - like source in the observation taken in a relatively early phase of the _ chandra _ mission are unlikely , we speculate that potential contaminations from these artifacts in the spectral analysis of any acis data can not be completely ruled out .
in addition , we would like to comment on the spectral analysis of some parts of the diffuse emission from n49 by kulkarni et al .
( 2002 ) . they have used a mekal model for such a spectral analysis , which should be inappropriate for the complicated diffuse snr emission as we presented in the current work .
more complex , physically realistic models should be utilized in order for an appropriate spectral analysis of such data .
the results from our simple spectral analysis may support the speculation that sgr 0526@xmath066 is a transition object bridging the axp and the sgr phases of neutron stars @xcite .
extensive analyses with the on - axis observations , which are beyond the scope of the current work , would however be necessary to understand the detailed astrophysical characteristics of this object .
for example , the measurements of the period ( @xmath94 ) and the period derivative ( _ p _ ) would be critical to determine the magnetic field strengh @xmath95 and the spin - down age @xmath96 of sgr 0526@xmath066 .
this information can provide important clues on the nature of sgr 0526@xmath066 such as the identification with a high - field ( @xmath95 @xmath23 10@xmath97 g ) neutron star or the `` magnetar '' @xcite .
the measurement of the spin - down age would also be useful to test the physical association of sgr 0526@xmath066 with snr n49 , which should provide important implications on the possible astrophysical connections between the sgrs and the axps ( e.g. , gaensler et al . 2001 ) .
kulkarni et al .
( 2002 ) have recently reported a detection of the 8.04 s pulsation from sgr 0526@xmath066 by utilizing @xmath285 ks _
chandra_/acis observations .
the derived magnetic dipole field strength is high ( @xmath95 @xmath2 7 @xmath29 10@xmath98 g ) and supports the magnetar interpretation .
the estimated spin - down age ( @xmath94/2_p _ @xmath2 2000 yr ) is consistent with the dynamical age of snr n49 , which support the physical association between n49 and sgr 0526@xmath066 . the location of sgr 0526@xmath066 in the lmc and the likely association with n49 is also supported by the fitted columns : i.e. , the absorbing columns for n49 and sgr 0526@xmath066 are consistent .
this sgr location may confirm the enormous energy release in the 1979 march @xmath13-ray burst .
if sgr 0526@xmath066 is indeed the compact remnant of the sn explosion which has also created snr n49 and it was born near the center of n49 , we may estimate its transverse velocity .
the angular offset of sgr 0526@xmath066 from the geometrical center of n49 is @xmath222@xmath24 , which implies that the spatial separation of the sgr from the center of the snr is @xmath86 = 5.3 ( @xmath88)@xmath9 pc at the distance of 50 kpc . assuming that sgr 0526@xmath066 is the neutron star born at the time of the sn explosion that produced n49 , and that the age of the sgr is @xmath96 @xmath2 6600 yr
, the mean transverse velocity is @xmath90 @xmath23 787 km s@xmath9 .
this would suggest that sgr 0526@xmath066 is a young , high velocity neutron star that is in a different class than typical axps instead of being the same class object as other axps , simply in a more evolved phase .
snr n49 in the lmc has been observed with _ chandra_/acis .
the high resolution `` true - color '' acis images reveal complex x - ray filamentary structures which represent the blast wave shock sweeping through the ambient ism and the shock interacting with the nearby dense molecular clouds .
the overall features from the ew images of the detected elemental species indicate that soft line emission is enhanced in the south - east while hard line emission is enhanced in the west of the snr .
these overall ew features are most likely dominated by the variations in the shock temperature and the ionization states of the plasma across the snr .
the spatially - resolved spectral analysis of n49 suggests that the x - ray emission from the snr can be described with multiple - phase shocks representing the singly - postshock region , the re - heated postshock region by the reflected shocks , and the cool forward shock entering the dense molecular clouds .
we detect a hard x - ray knot in the southwest of n49 beyond the snr shock boundary .
the x - ray spectrum of this knot is emission line dominated and can be described with a metal - rich hot thermal plasma .
this result implies that this hard knot appears to be ejecta fragments created at the time of the sn explosion that produced snr n49 .
the psf degradation of the off - axis observation however restricts the data analysis of this feature and preclude resolving the angular structure of this knot .
follow - up on - axis observations with a decent exposure might be necessary to address the detailed nature of this feature . the acis observations of n49 also firmly detect the x - ray counterpart of the quiescent phase of sgr 0526@xmath066 .
the x - ray spectrum of sgr 0526@xmath066 is continuum dominated and can be fitted with a power law model .
the implied photon index is @xmath1 @xmath2 3 .
this power law spectrum is generally consistent with those from the axps although the observed @xmath13-ray burst of sgr 0526@xmath066 is a rare characteristic of typical axps .
the authors thank j. dickel for providing his atca data .
sp also thank a. bykov for a valuable discussion on the sn ejecta fragments .
this work has been in parts supported by nasa under contract nas8 - 01128 for the _ chandra x - ray observatory_. jph acknowledges support from _ chandra _ grants go1 - 2052x , go2 - 3068x , and go2 - 3080b to rutgers .
brandt , w. n. , alexander , d. m. , hornschemeier , a. e. , garmire , g. p. , schneider , d. p. , barger , a. j. , bauer , f. e. , broos , p. s. , cowie , l. l. , townsley , l. k. , burrows , d. n. , chartas , g. , feigelson , e. d. , griffiths , r. e. , nousek , j. a. , & , sargent , w. l. w. 2001 , aj , 122 , 2810 cline , t. l. , desai , u. d. , teegarden , b. j. , evans , w. d. , klebesadel , r. w. , laros , j. g. , barat , c. , hurley , k. , niel , m. , vedrenne , g. , estulin , i. v. , kurt , v. g. , mersov , g. a. , zenchenko , v. m. , weisskopf , m. c. , & grindlay , j. 1982 , apj , 255 , l45 mereghetti , s. , chiarlone , l. , israel , g. l. , & stella , l. 2002 , `` neutron stars , pulsars and supernova remnants '' , proceedings of the 270 .
we - heraeus - seminar , bad honnef , january 2002 , eds . w. becker , h. lesch & j. trmper , mpe - report in press ( astro - ph/0205122 ) cccc o & 590 @xmath0 740 & 300 @xmath0 550 & 1200 @xmath0 1280 + fe l & 750 @xmath0 870 & 300 @xmath0 550 & 1200 @xmath0 1280 + ne ( he@xmath3 ) & 880 @xmath0 950 & 300 @xmath0 550 & 1200 @xmath0 1280 + ne ( ly@xmath3)/fe l & 980 @xmath0 1100 & 300 @xmath0 550 & 1200 @xmath0 1280 + mg & 1290 @xmath0 1550 & 1200 @xmath0 1280 & 1630 @xmath0 1710 + si & 1750 @xmath0 2100 & 1630 @xmath0 1710 & 2140 @xmath0 2280 + ccccccc 1 & 2.7@xmath99 & 0.52@xmath100 & 8.6@xmath101 & 2.01@xmath102 & 43.3 & 60 + 2 & 1.2@xmath103 & 0.65@xmath104 & 70.0@xmath105 & 2.79@xmath106 & 77.0 & 93 + 3 & @xmath4 2.5 ( 0.6 ) & 0.59@xmath107 & 31.5@xmath108 & 1.71@xmath109 & 59.9 & 56 + 6 & @xmath4 3.6 ( 1.1 ) & 1.04@xmath110 & @xmath111 41.0 ( 500 ) & @xmath4 0.27 ( 0.12 ) & 19.5 & 36 + ccccccc 1 & 0.90@xmath112 & 1.01@xmath113 & 0.53@xmath114 & 0.39@xmath115 & @xmath4 5.77 ( 1.91 ) & 0.17@xmath116 + 2 & 1.23@xmath117 & 1.80@xmath118 & 0.78@xmath119 & 0.86@xmath120 & 1.75@xmath121 & 0.25@xmath122 + 3 & 0.46@xmath123 & 0.30@xmath124 & 0.16@xmath125 & 0.19@xmath126 & @xmath4 4.02 ( 1.50 ) & 0.18@xmath127 + 6 & 5.71@xmath128 & 0.33 ( fixed ) & @xmath4 502.7 ( 1.11 ) & 5.92@xmath129 & @xmath4 1334.2 ( 3.98 ) & 0.94@xmath130 + cccccccc 4 & 3.9@xmath131 & 67.7 & 94 & soft & 0.47@xmath132 & 0.8@xmath133 & 3.81@xmath134 + & & & & hard & 1.08@xmath135 & @xmath111 14.6 ( 40.9 ) & 2.19@xmath136 + 5 & 3.6@xmath137 & 86.7 & 96 & soft & 0.41@xmath138 & 1.1@xmath139 & 19.83@xmath140 + & & & & hard & 1.41@xmath141 & @xmath111 9.3 ( 22.1 ) & 1.53@xmath142 + | the supernova remnant ( snr ) n49 in the large magellanic cloud has been observed with the advanced ccd imaging spectrometer ( acis ) on board the _ chandra x - ray observatory_. the superb angular resolution of the _
chandra_/acis images resolves a point source , the likely x - ray counterpart of soft gamma - ray repeater sgr 0526@xmath066 , and the diffuse filaments and knots across the snr .
these filamentary features represent the blast wave sweeping through the ambient interstellar medium and nearby dense molecular clouds .
we detect metal - rich ejecta beyond the main blast wave shock boundary in the southwest of the snr , which appear to be explosion fragments or `` bullets '' ejected from the progenitor star .
the detection of strong h - like si line emission in the eastern side of the snr requires multi - phase shocks in order to describe the observed x - ray spectrum , whereas such a multi - phase plasma is not evident in the western side .
this complex spectral structure of n49 suggests that the postshock regions toward the east of the snr might have been reheated by the reverse shock off the dense molecular clouds while the blast wave shock front has decelerated as it propagates into the dense clouds .
the x - ray spectrum of the detected point - like source is continuum dominated and can be described with a power law of @xmath1 @xmath2 3 .
this provides a confirmation that this point - like x - ray source is the counterpart of sgr 0526@xmath066 in the quiescent state . |
although the basic weak interaction is relatively well known , there exist large uncertainties of the weak couplings of mesons and baryons at the hadron level relevant at low and intermediate energies up to about 1 gev , _
i.e. _ in nuclear physics .
for example the efforts to determine the parity nonconserving ( pnc ) couplings @xcite report uncertainties of the order of 100 % or more of the recommended `` best '' values .
( for reviews see _ e.g. _ refs .
@xcite . ) in this situation even the signs of couplings may be suspect . considering the weakness of the couplings their experimental determination is a big challenge and attempts often utilize nuclear structure to enhance pnc effects . however , the analysis of the basic couplings is then complicated by the environment . among few - nucleon systems
also polarized photoreactions have been used with the deuteron . in principle , the most direct way without external disturbances would be @xmath2 scattering . presently pnc @xmath6 scattering experiments are unlikely and even @xmath5 scattering experiments are scarce @xcite . among these of particular interest is the measurement of the parity violating spin observable @xmath7 in the triumf experiment e497 at 221.3 mev @xcite .
the energy was chosen so that incidentally the strong interaction phases conspire to cause the @xmath8 contribution to cross zero @xcite .
while in this amplitude both @xmath9 and @xmath10 mesons are equally important , the next @xmath11 parity mixing amplitude is strongly dominated by @xmath9 exchange enabling a drive towards pnc @xmath12 coupling .
the above logic is based on the idea that the dominant pnc effect in @xmath5 scattering should be due to vector mesons , since long - ranged single pion exchange is forbidden by barton s theorem @xcite .
this theorem forbids in general neutral @xmath13 mesons to couple with nucleons in the pnc interaction because of the simultaneous violation of @xmath14 and @xmath15 symmetries .
so in the pnc @xmath5 interaction at low energies only @xmath9- , and @xmath10-meson exchanges are expected to be significant and the two - pion exchange ( tpe ) is assumed negligible @xcite . with these assumptions the triumf experiment e497 @xcite would , in principle , result in the determination of the weak @xmath12 coupling constant @xmath16 from @xmath17 scattering .
the lower energy experiments @xcite have already determined the independent combination of @xmath18+@xmath19 ( where @xmath20 ) and so both the @xmath18 and @xmath19 are supposedly determined separately . in view of such
dedicated and very time consuming experiments it is important to carefully study the validity of the assumptions and uncertainties in their interpretation .
in fact , it was shown in ref .
@xcite that @xmath21-isobar excitation by weak @xmath9 and strong ( mainly ) pion exchange has a significant effect on @xmath7 at all energies and should also be considered . in spite of pnc single pion exchange being excluded in @xmath5 scattering ,
crossed charged two - pion exchange is allowed .
charged pions are also possible in two - pion exchanges involving the excitation of the @xmath22 resonance .
the former was studied already in the early works @xcite , while calculations for the latter were performed above the pion production threshold in refs .
however , the @xmath21 contribution extends also below this to low energies ( as shown with vector meson exchanges in ref .
@xcite ) . to our knowledge
the contribution of all these two pion effects to pnc observables has not been investigated systematically on the same footing together .
closest comes the recent ref .
@xcite in deriving pnc the two - pion potential in chiral perturbation theory .
our aim in the present paper is to calculate this potential including realistic form factors and extending the calculation to the observable @xmath23 .
further interest in tpe lies in the fact that it should be the longest ranged pnc effect in @xmath5 scattering , which might show up in energy dependence .
the organization of the paper is as follows .
first we evaluate numerically pnc tpe as a potential in the momentum representation assuming the baryons to be static .
this potential can be well fitted by lorentz functions of the momentum transfer , which in turn can be expressed in terms of yukawa functions in the coordinate representation .
it may be noted that the result will be local ( excluding relativistic corrections ) and a comparison with the local parts of vector meson exchanges is then straightforward .
next we discuss non - static effects in three different kinematics and dynamics allowing kinetic energies for baryons .
these are found to be significant but not necessarily dominant .
the hamiltonians for the parity conserving ( pc ) @xmath24 , @xmath25 , and pnc @xmath24 interactions , which describe the needed nonrelativistic vertices are @xmath26 in this calculation only the @xmath2 and @xmath27 intermediate states are considered and @xmath28 intermediate state ignored , because the pnc @xmath29 vertex is concluded to be very small in ref .
we take it to be zero .
time orderings of parity nonconserving ( black spot ) two - pion exchange ( dashed line ) via nucleon ( solid line ) delta ( bar ) intermediate state .
both the crossed and box graphs include the @xmath27 intermediate state while the @xmath2 intermediate state can appear only in the crossed graphs .
the direction of time is upwards . ] unfortunately , in literature choices of the signs in the above definitions vary , both overall and even in the @xmath24 _ vs. _ @xmath30 vertices . in purely strong interactions
the signs cancel , but in the presence of a weak vertex the sign matters .
our choice follows the standard one ( ddh @xcite ) so that our pnc ope potential @xmath31 would be as given in _
e.g. _ ref .
@xcite . here
@xmath32 represents the yukawa function @xmath33 ( possibly modified by a form factor ) .
the change of the @xmath24 coupling to the @xmath30 would be a replacement of the nucleon spin - isospin operators @xmath34 and @xmath35 by the transition operators @xmath36 and @xmath37 with the normalization @xmath38 @xcite .
time orderings of the pnc tpe via @xmath27 intermediate state can be grouped under eight different groups , which all contain six time orderings that have the same spin and isospin structure .
two of the groups correspond to the graphs in fig .
[ grf ] , where the upper row is for the box ( b ) and the lower for the crossed ( c ) graphs , in which the parity is broken in vertex 1 ( the vertices are numbered in fig . [ kin ] ) .
the rest of the groups can also be formed by using the graphs in fig .
[ grf ] : two of the groups include the same graphs except the parity is broken in vertex 2 .
the remaining four groups can be obtained in the same manner from fig .
[ grf ] with the exchange @xmath39 and breaking the parity respectively in vertices 3 and 4 .
time orderings of the pnc tpe via @xmath2 intermediate state correspond only to the crossed graphs in fig .
this contribution has four different groups , @xmath40 the parity is broken once in each vertex .
overall there are totally 72 time - ordered graphs to add up .
a symmetric and practical choice of the meson mediated momentum transfers is @xmath41 shown in fig .
[ kin ] , where @xmath42 is an integration parameter and @xmath43 the overall momentum transfer .
the symmetry of the kinematics for the box and crossed graphs , where @xmath43 is the momentum transfer , @xmath42 the free loop momentum , and @xmath44 the relative momentum between the nucleons .
the pions carry the momentum @xmath41 .
numbers 1 - 4 denote respectively the number of the vertex . ] in our basic model the initial and final nucleons and all the intermediate baryons are considered static .
the energy ( mass ) difference between the isobar and nucleon is denoted by @xmath45 and the energies of the exchanged pions are @xmath46 .
the energy denominators of the different graph types given in fig .
[ grf ] act as propagators of the old - fashioned time ordered second - order perturbation theory each of these propagators corresponds also to a different spin - isospin structure acting in the numerator .
however , a simplification arises from the even parity of the propagators , which does not affect pnc .
then their angular dependence is only due to @xmath48 , which in a good approximation can be taken the angular average @xmath49 ( _ i.e. _ one can replace @xmath50 ) .
thus pnc actually arises from the vertex structure of eqs .
( [ h1])([h3 ] ) .
in fact , only the terms with even powers of @xmath51 survive , when the angular integral is carried out .
.[tab : params ] parameter values for meson - nucleon couplings .
the weak couplings @xmath52 ( @xmath53 for pions ) , are the `` best '' estimates of refs .
@xcite ( ddh ) and @xcite ( fcdh ) . the strong couplings and cut - offs of the vector mesons are from the coordinate space obe of ref
[ cols="^,^,^,^,^,^ " , ] the total and the partial contributions of the pnc tpe given by eq .
( [ part]).,width=415 ] following ref .
@xcite we replace eq .
( [ upot ] ) with the fit in ( [ pot ] ) and get the configuration space potential as the fourier transform @xmath54 with @xmath55-\frac{e^{-br}}{r } ( b+\frac{1}{r})\bigr\}\ ; .\end{aligned}\ ] ] the magnitudes of eq .
@xmath56 for the partial contributions are illustrated in fig .
[ figpotr ] , where the @xmath2 crossed box and @xmath57 direct box are by far dominant . as expected , the crossed @xmath2 contribution is of the longest range , but the @xmath57 excitation becomes as important or even larger inside @xmath58 fm . partly due to the weight factors present in eq .
( [ upot ] ) the crossed @xmath57 contribution is nearly negligible . a comparison with the potential obtained from chiral perturbation theory @xcite may be in order .
that contains also the ( reducible ) box diagram and a triangle diagram with @xmath59-wave rescattering .
the former is identically zero in @xmath5 scattering ( barton s theorem ) , while the latter is chirally suppressed in the @xmath5 case .
so our results should be similar .
in fact , for distances larger than about 1.2 fm they are indistinguishable , whereas inside the radius of 0.8 fm our result is significantly softer due to the form factors involved . both have a range corresponding roughly to vector meson exchanges , so actually probably one may expect similar results .
the analyzing power @xmath7 arising from the different partial contributions of the pnc 2@xmath60 exchanges using the ddh `` best '' value for the weak coupling @xmath61.,width=415 ] we now use the above pnc 2@xmath60 exchange potential to calculate the asymmetry of the parity - violating spin observable @xmath62 where @xmath63 are the scattering cross sections of the two helicity states of the longitudinally polarized beam .
the reid soft - core potential @xcite is used to obtain the strong interaction distortions for weak interactions , while to minimize theoretical uncertainties otherwise the empirical phase shifts and strong interaction amplitudes are used and taken from ref .
it has been seen in the past that the dependence on the strong potential is relatively weak and we do not go in this in detail @xcite . therefore , apart from the easily scalable @xmath64 weak coupling and the form factor , the results in fig .
[ al_partial ] may be considered fairly model independent .
contributions of the five lowest parity mixed partial wave amplitudes ( @xmath65 ) , ( @xmath66 ) , ( @xmath67 ) , ( @xmath68 ) , and ( @xmath69 ) are included in this calculation .
structurally the @xmath8 and @xmath11 contributions are similar to earlier results and the expansion is already converged for those ( @xmath70 is negligible at our energies ) . analyzing power @xmath71 for the pnc tpe and the local parts of the @xmath10 and @xmath9 exchanges using the ddh `` best '' values for the weak couplings .
the experimental data points are bonn at 13.6 mev @xcite , psi at 45 mev @xcite , triumf at 221.3 mev @xcite , and los alamos at 800 mev @xcite.,width=415 ] as the tpe effect in @xmath7 is clearly large , it is of utmost interest to compare the present result against vector meson effects considered earlier .
these potentials are given in ref .
@xcite in which we also incorporate the monopole form factors of the type ( [ form ] ) in each vertex and use two sets of weak couplings @xcite .
thus the pnc vector meson potentials for @xmath9 and @xmath10 read @xmath72\bigr),\ ] ] @xmath73 with the relevant parameters given in table [ tab : params ] .
in the case of the vector mesons we use only the dominant local ( latter ) term of eq .
( [ omegapot ] ) and neglect its nonlocal ( former ) term , which only causes a minor contribution on the interaction @xcite and also does not have a direct correspondence with the present local interaction .
the pnc tpe effect is comparable in size to those given by vector meson exchanges using both sets ( figs .
[ al1 ] and [ al2 ] ) . the older set ( ddh )
would give the sum as an overestimate , while the new analysis ( fcdh ) gives a satisfactory agreement .
however , one should remember that the nonlocal pnc would increase the result somewhat . the same as fig .
[ al1 ] , except the fcdh best values are employed .
the 2@xmath60 and @xmath9 exchanges are now nearly indistinguishable.,width=415 ] so far the calculation has totally neglected any kinetic energies of baryons .
this assumption gives simplicity and clarity while still being probably reasonably realistic due to the large baryon masses .
we now make various attempts to overcome this approximation and to model non - static effects .
first we still take the initial kinetic energy ( and baryon momenta ) to be zero , while the intermediate and final baryons have nonzero energies . in this case
the final relative momentum is @xmath74 and energy is not conserved . in one boson exchange potentials without internal excitations this is not a problem and presently we are just making an energy independent potential .
the relatively trivial generalization of eq .
( [ props ] ) with @xmath75 in fig .
[ kin ] yields a qualitatively similar but weaker potential , since the excitation is larger , and a smaller @xmath7 ( `` non - static asymmetric '' dashed curve in fig .
[ last ] and parametrization in table [ tab : fit ] ) .
another way of considering non - static effects is to allow kinetic energy to be present in the initial state with the consequences on momenta . in this case
the assumption of the conservation of energy would give a simplification making @xmath76 and @xmath74 orthogonal and kinematics could be symmetric .
however , in this case numerically one meets a pole in the integration over @xmath77 at higher energies , _
i.e. _ for large incident momenta . taking for definiteness @xmath78 to be zero it is , nevertheless , possible to get a result without the pole disturbing too much numerics ( keeping @xmath79 ) see fig .
[ last ] ( dotted curve ) and table [ tab : fit ] .
model dependence of pnc tpe using the ddh `` best '' value for the @xmath53 .
the solid line is the sum of coupled channels , the residual part , and the @xmath2 crossed contributions .
the dash - dotted line is the static , dotted non - static symmetric , and the dashed non - static asymmetric pnc tpe as discussed in the text.,width=415 ] probably the best way to include the kinetic energy of the baryons is to incorporate them dynamically in the equation of motion .
this can be done at the two - baryon level by the coupled - channel schrdinger equation .
although this does not cover crossed mesons , following ref .
@xcite it is possible to rearrange the direct and crossed @xmath57 diagrams as @xmath80 where the numerators involving the spin - isospin structure are denoted by @xmath81 .
the first term is an iteration of static @xmath60 exchange potentials , since the sum of all propagators turns out to be just @xmath82 and @xmath83 has the corresponding vertex structure . so this term can be generated by coupled channels iterating an @xmath84 transition potential , while the second is presumably a smaller residual correction to be dealt with perturbatively . in terms of the previous diagrammatic integrals of eq .
( [ upot ] ) this rearrangement means @xmath85 here the part ( [ cc ] ) is to be treated in the coupled - channels approach ; the same term would arise from the iteration of the strong eq .
( [ nd ] ) and weak eq .
( [ wtrpot ] ) ope transition potentials .
one advantage of coupled channels is that different centrifugal barriers related to the orbital angular momenta are automatically taken into account .
therefore , coupled channels results are state dependent .
also opening of different channels causes energy dependence , which can not be simulated by energy independent potentials alone .
the first two basic hamiltonians , eqs .
( [ h1 ] ) and ( [ h2 ] ) , lead to the strong ope transition ( @xmath86@xmath27 ) potential @xmath87y_\pi(\bm{r})\ , , \end{aligned}\ ] ] whereas from eqs .
( [ h2 ] ) and ( [ h3 ] ) we get the weak ope transition ( @xmath86@xmath27 ) potential @xmath88\frac{\partial}{\partial_r}y_\pi(\bm{r}).\end{aligned}\ ] ] iterating these two should give the effect of the potential ( [ cc ] ) , a claim borne out in a comparison of the dashed and solid ( coupled channels + residual + crossed @xmath2 ) curves in fig .
[ last ] . at low energies
the result is very similar to the non - static asymmetric potential result but deviates drastically in energy dependence in particular in the proximity of the @xmath57 threshold .
partial wave contributions to @xmath7 from pnc tpe using the coupled - channels approach and the ddh `` best '' value for the @xmath53.,width=415 ] it is of interest to study in more detail the dip at 600 mev in fig .
[ last ] . this kind of cusp structures arise in dynamical coupled channels calculations ( as isospin one `` dibaryon '' resonances in the case of @xmath57 thresholds giving rise to maxima e.g. in pion production ) , but the present one is somewhat sharper than the wide maximum seen in ref .
however , there it was produced by pnc @xmath9 exchange interfering with the strong transition involving both pion and a destructive @xmath9 . in the present case , to be consistent with the tpe potential , we have exclusively the long ranged pion exchange in both transition potentials without damping other than the form factor .
the structure is due to the favoured transition @xmath89 through the intermediate state without a centrifugal barrier .
the other intermediate @xmath57 states coupling with tensor - coupled states and with centrifugal barrier are not particularly favoured @xcite .
these phenomenological arguments are confirmed in fig .
[ iter ] , which shows splitting the above total result into partial wave amplitudes : the structure is not seen in the otherwise dominant @xmath8 amplitude , which has the same structure ( dictated by strong interaction @xcite ) as for other potentials .
we evaluated the pnc tpep for the elastic @xmath5 scattering and calculated the longitudinal asymmetry @xmath7 . compared to the local contributions of the pnc @xmath10 and @xmath9 exchanges our results are of the same order indicating that also this mechanism should be seriously considered in interpreting pnc data . with the old ddh couplings @xcite
this additional contribution leads to an overestimate , but the newer ones @xcite can give a tolerable agreement ( though we do not include nonlocal pnc here ) especially , if non - static effects are included .
however the ddh and fcdh `` best '' values for the @xmath61 might be too large in the light of the experimental restrictions given by the @xmath90f parity violating measurements @xcite , which bound the upper limit in the range @xmath91 @xmath92 .
some theoretical predictions are also within this limit @xcite .
further , one should note that the relative sign of the pion strong and weak coupling may be unknown in the vertex definitions ( [ h1])([h3 ] ) even if the weak magnitude were given .
the knowledge of this coupling is essential , since the pions are more than five times lighter than the vector mesons and thus the pnc tpe presumably represents the longest range part of the weak interaction in @xmath5 scattering . in particular relating to the triumf experiment at 221.3 mev we get at that energy the tpe contribution 0.48 to @xmath7 in the static model and in the nonstatic `` symmetric '' model 0.26 ( 0.28 in the `` asymmetric '' one ) .
these may set rather realistic limits though the even larger result 0.70 using the coupled channels model should be noted .
however , these are obtained using the older ddh coupling @xmath93 and can be easily scaled for the value from the fcdh analysis @xmath94 or any other .
the long ranged ope transition potential produces also a manifest cusp peak at 600 mev in a coupled channels calculation . whether or not this will be diminished in the presence of vector mesons will be discussed in a subsequent work @xcite .
there are two competing effects : destructive interference from the @xmath9 exchange in the strong transition potential and the contribution from the @xmath9 in the pnc transition as discussed earlier @xcite . in any case , as suggested in ref .
@xcite an experimental point between the triumf and los alamos energies would be of interest . | parity nonconserving two - pion exchange in elastic @xmath0 scattering is investigated in the presence of phenomenological strong distortions in various models .
parity violation is included in @xmath1 vertex considering @xmath2 and @xmath3 intermediate states in box and crossed box diagrams . using the derived parity nonconserving two - pion exchange potential we calculate the longitudinal analyzing power @xmath4 in elastic @xmath5 scattering .
the predicted effect is of the same order as vector meson exchanges . |
66 blazar - type agns have been detected by the egret instrument on board the _ compton gamma - ray observatory _ as sources of @xmath0-rays above 100 mev ( hartman et al .
these objects are identified with flat - spectrum radio sources classified as bl lac objects or flat - spectrum radio quasars ( fsrqs ) .
many of these objects exhibit variability at all wavelengths , generally with the most rapid variability , on time scales of hours to days , observed at the highest @xmath0-ray energies ( e.g. , bloom et al .
1997 ; wagner et al . 1995 ; mukherjee et al .
the broadband ( radio@xmath0-ray ) emission from blazars is most probably emitted via nonthermal synchrotron radiation and comptonization of soft photons by energetic particles in relativistic outflows powered by accreting supermassive black holes .
potential sources of soft photons which are compton - scattered to produce the @xmath0-ray emission include internal synchrotron photons ( e.g. , marscher & gear 1985 , maraschi et al .
1992 , bloom & marscher 1996 ) , jet synchrotron radiation rescattered by circumnuclear material ( ghisellini & madau 1996 , bednarek 1998 , bttcher & dermer 1998 ) , and accretion - disk radiation which enters the jet directly ( dermer & schlickeiser 1993 ) and/or after being scattered by surrounding blr clouds and circumnuclear debris ( e.g. , sikora , begelman & rees 1994 ; blandford & levinson 1995 ; dermer , sturner , & schlickeiser 1997 ; protheroe & biermann 1997 ) . in recent years
, extensive simultaneous broadband observations of blazars have enabled detailed modeling of their broadband spectra
. a general result of these modeling efforts appears to be that the spectra of high - frequency peaked bl lac objects ( hbls ) are well reproduced using a synchrotron - self compton ( ssc ) model , where the radio through x - ray radiation is produced by nonthermal synchrotron radiation of ultrarelativistic electrons and the @xmath0-ray emission , extending in some cases up to tev energies , results from compton scattering of the synchrotron radiation by the same population of electrons ( e. g. , mastichiadis & kirk 1997 for mrk 421 , pian et al .
1998 for mrk 501 ) . however , a critical question in this case is whether the primary acceleration mechanism is able to accelerate electrons up to the required ultrarelativistic energies , which are @xmath1 tev if the recent hegra detection of @xmath2 tev photons from mrk 501 ( aharonian et al .
1999 ) is real . on the other hand ,
fsrqs are more successfully modeled assuming that the soft seed photons for comptonization are external to the jet ( ec for external comptonization ; e. g. , dermer et al .
1997 for 3c 273 , sambruna et al .
1997 and mukherjee et al .
1999 for pks 0528 + 134 , bttcher et al . 1997 for 3c 279 ) .
there appears to be a more or less continuous sequence of spectral properties of different objects , hbl @xmath3 lbl @xmath3 fsrq , characterized by increasing @xmath0-ray luminosity , increasing dominance of the energy output in the @xmath0-ray component over the synchrotron component , and a shift of the peaks in the @xmath4 spectra of both components towards lower energies .
this sequence can be understood in terms of increasing dominance of the ec over the ssc mechanism ( fossati et al .
1997 , ghisellini et al . 1998 ) .
recently , bttcher & collmar ( 1998 ) and mukherjee et al . ( 1999 ) have suggested that in the case of fsrqs a similar sequence might also occur between different intensity states of the same object , and have applied this idea to the various states of pks 0528 + 134 .
there , it was assumed that the high states of pks 0528 + 134 are characterized by a high bulk lorentz factor of the ultrarelativistic material in the jet , implying that the external radiation field is more strongly boosted into the comoving rest frame than during the quiescent state .
this effect , leading to a stronger dependence of the ec radiation on the doppler factor than of the ssc radiation , was first pointed out by dermer ( 1995 ) . in fig .
1 , the fit results to two extreme states of pks 0528 + 134 from mukherjee et al . ( 1999 ) are compared .
the figure reveals an important prediction of the model adopted in that paper : during the @xmath0-ray high state we expect that the synchrotron spectrum peaks at lower frequencies than in the low state .
unfortunately , the peaks of the @xmath4 synchrotron spectra of most fsrqs are in the infrared and thus very hard to observe . for this reason ,
the results of detailed modeling of the synchrotron component of an fsrq have to be regarded with caution since in most cases the shape of the synchrotron spectrum is not well enough constrained to allow an exact determination of jet parameters .
note , for example , that in ghisellini et al .
( 1998 ) in many cases the 9-parameter model adopted there is used to effectively fit less than 10 data points , and that the model ir optical spectra of pks 0528 + 134 calculated in mukherjee et al .
( 1999 ) for several observing periods are extremely poorly constrained . in the very low @xmath0-ray states of pks 0528 + 134 ( vps 39 , 337 , and 616 )
, even the @xmath0-ray spectrum is very poorly constrained .
therefore it is important to demonstrate that the predicted synchrotron peak shift does not depend on the details of the adopted jet model and is not a consequence of fine - tuning of parameters . in section 2
, i will present an analytical estimate of the synchrotron peak shift on the basis of very simple , general arguments . in section 3
, several fsrqs are suggested as promising candidates to test the prediction made in this letter .
i summarize in section 4 .
the shift of the @xmath4 peak of the synchrotron spectrum during a @xmath0-ray flare of an fsrq will be determined on the basis of the following general assumptions : ( a ) the @xmath0-ray flare is predominantly caused by an enhancement of the energy density @xmath5 of external photons in the rest frame comoving with a component ( blob ) of ultrarelativistic material moving along the jet .
this enhancement can be caused by an increasing bulk lorentz factor @xmath6 of the jet material .
we have @xmath7 .
( b ) the electrons in the blob have a non - thermal distribution with a peak at the energy @xmath8 which is determined by the balance of an energy - independent acceleration rate @xmath9 to the radiative energy loss rate ( cf .
ghisellini et al .
( c ) during a flare , the radiative energy loss rate of relativistic electrons in the blob is dominated by compton scattering of external photons in the thomson regime .
the assumptions ( b ) and ( c ) imply that @xmath10 .
the jets of blazars are directed at a small angle @xmath11 with respect to our line of sight .
thus , to a good approximation , the doppler factor @xmath12 \bigr)^{-1 } \approx \gamma$ ] .
then , the observed peak of the synchrotron component depends on the bulk lorentz factor and the external photon density as @xmath13 where @xmath14 is the magnetic field strength in the comoving frame .
the apparent bolometric luminosity in the synchrotron component varies acording to @xmath15 a plausible way to estimate the magnetic field strength might be the assumption of equipartition of magnetic field energy density to the energy density of ultrarelativistic electrons in the jet . however , our conclusions do not depend on the particular choice of @xmath14 .
if equipartition applies , then @xmath16 , and eqs.(1 ) and ( 2 ) become ( the superscript @xmath17 denotes the equipartition case ) @xmath18 and @xmath19 assuming thomson scattering at the @xmath4 peak of the compton component ( this is a reasonable assumption for the peak at several mev 10 gev , while beyond that energy klein - nishina effects may become important ) , the @xmath0-ray spectrum peaks at @xmath20 where @xmath21 is the mean photon energy of the external photon field in the stationary frame of the agn .
the apparent bolometric luminosity in the compton component depends on @xmath6 and @xmath22 through @xmath23 independent of the external photon density .
remarkably , this implies that an enhancement of the external photon density @xmath22 , e. g. , due to structural changes of the circumnuclear material does not produce a @xmath0-ray flare , although it leads to spectral variability , shifting both spectral components to lower frequencies .
now let us consider the effect of a variation of the bulk lorentz factor as a possible cause of a @xmath0-ray flare .
( 1 ) , ( 3 ) , and ( 5 ) predict a shift of the synchrotron peak to lower frequencies , while the @xmath0-ray peak should remain at basically the same photon energy .
the ratio of the luminosities in both components varies according to @xmath24 or @xmath25 , respectively .
if @xmath22 and @xmath9 remain constant and equipartition applies , this yields a variation @xmath26 , thus predicting a much stronger relative variation of the two spectral components than predicted by the ssc model which was ruled out for 3c 279 for this reason by the observation of a variation of the @xmath0-ray component by amplitudes greater than the square of the amplitudes of variation of the synchrotron component ( wehrle et al . 1998 ) .
the synchrotron peak is still shifted to lower energies even if the increase of the bulk lorentz factor is physically related to a more efficient particle acceleration in the comoving frame , as long as during the flare the product @xmath27 is lower than during the quiescent state . in this case ,
( 1 ) , ( 3 ) , and ( 5 ) predict that the @xmath0-ray spectrum becomes spectrally harder , while at the same time the synchrotron peak shifts to lower energies . in fact , spectral hardening of the @xmath0-ray spectrum during flares is a common feature in egret - detected fsrqs ( e. g. , collmar et al .
1997 , hartman et al .
1996 , wehrle et al .
1998 ) .
this flaring behavior is in qualitative contrast to the flares observed in hbls where a shift of both the synchrotron and the compton peaks to higher frequencies is observed ( e. g. , catanese et al . 1997 ,
pian et al .
1998 , kataoka et al . 1999 ) .
these objects are believed to be qualitatively different from fsrqs because ( a ) the broad - line regions surrounding the central engine are weak or absent and the isotropic luminosity of the central accretion disk is generally weaker than in quasars , implying that the external soft photon field is much weaker than in the fsrq case and hence negligible compared to the intrinsic synchrotron radiation field , and ( b ) compton scattering events near the @xmath4 peak of the @xmath0-ray spectrum most probably occur in the extreme klein - nishina regime , thus rendering compton cooling rather inefficient .
the flaring behavior of these sources seems to be dominated by an enhanced efficiency of electron acceleration in the jet which shifts both spectral components towards higher photon energies ( mastichiadis & kirk 1997 , pian et al .
1998 ) .
therefore , i propose that the qualitatively different behavior of the synchrotron peak during a @xmath0-ray flare may serve as a diagnostic to determine the dominant electron cooling and radiation mechanism at @xmath0-ray energies .
as mentioned earlier , most of the bright and well - observed fsrqs ( e. g. , pks 0528 + 134 : mukherjee et al .
1996 , or 3c 279 : wehrle et al .
1998 ) have their synchrotron peak in the infrared where it is very hard to observe .
3 of werhle et al .
( 1998 ) seems to indicate a shift of the synchrotron peak of 3c 279 to lower frequencies during the 1996 flare state .
however , the peak frequency range is not covered in any of the observing periods presented there .
thus , the determination of the peak frequency from the existing data on 3c 279 would necessarily be model dependent .
a much more promising candidate to test the predictions presented in this letter is its `` sister '' source 3c 273 ( von montigny et al .
this source appears to be ideal for such a study for several reasons .
( 1 ) it is a very bright radio and ir source , persistently detectable from radio through ir and optical frequencies .
( 2 ) during flares it is a strong egret source , allowing an easy detection of a @xmath0-ray flare even with the degraded sensitivity of egret . ( 3 ) the strong big blue bump yields a very exact determination of the luminosity ( @xmath28 erg s@xmath29 ) and spectrum of the underlying accretion disc of the agn which facilitates the normalization of model calculations . in fig .
2 , the simultaneous radio spectra of three epochs before , during , and after the prominent @xmath0-ray flare of 3c 273 in 1993
november ( von montigny et al . 1997 ) are presented .
a comparison of the flare - state radio spectrum ( open circles connected by solid lines ) to the pre - flare and post - flare spectra seems to indicate a softening of the synchrotron spectrum during the flare , in perfect agreement with the expectation if the @xmath0-rays are produced by the ec mechanism . at frequencies above 100 ghz ,
the radio spectra shown in fig .
2 are simultaneous to within 1 day .
however , the radio spectrum shortly ( i. e. a few days ) prior to the 1993 @xmath0-ray flare was not monitored simultaneously with reasonable spectral coverage at @xmath30 ghz . the pre - flare period closest to the 1993 @xmath0-ray flare for which such a simultaneous radio spectrum was available , was about 100 days before the flare . at that time , 3c273 was not in the field of view of egret so that we can not be sure that the source was in its quiescent state .
furthermore , the broadband spectrum of 3c273 shown in von montigny et al .
( 1997 ) and in kubo et al .
( 1998 ) suggests that the actual synchrotron peak is located at higher frequencies , @xmath31 hz , which have not been monitored in regular , short time intervals in the von montigny et al .
( 1997 ) campaign .
therefore , although this result seems very promising , it needs more solid confirmation by future broadband campaigns with good spectral and temporal coverage of the @xmath32 ghz
@xmath33 hz frequency range in order to allow a reliable , model - independent determination of the synchrotron peak of precisely simultaneous snapshot spectra in different @xmath0-ray states of the source .
in order to test the predictions presented here , it is particularly important that the synchrotron spectrum _
prior to _ a @xmath0-ray flare is known .
an apparent shift of the synchrotron peak towards higher frequencies following a @xmath0-ray flare is also predicted by alternative models , for example by the marscher & gear ( 1985 ) ssc model where a radio flare is delayed with respect to an outburst at @xmath0-rays due to the finite synchrotron cooling time of electrons in the jet .
a few other fsrqs also emit a high flux at multi - ghz frequencies and might therefore allow the precise determination of the synchrotron peak and its shift during @xmath0-ray flares in future campaigns . in particular , pks 0420 - 014 ( radecke et al .
1995 , von montigny et al .
1995 ) , pks 0521 - 365 ( pian et al . 1996 ) , and 3c454.3 (= pks 2251 + 158 ; von montigny et al .
1995 ) have successfully been observed with near - complete frequency coverage of their synchrotron peak , located at @xmath34
@xmath35 hz .
these sources may serve as test cases for the predictions presented in this letter .
on the basis of very general arguments i have shown that the synchrotron @xmath4 peak in the broadband spectra of egret - detected fsrqs is expected to shift towards lower frequencies during @xmath0-ray flares , if comptonization of external radiation is the dominant electron cooling and radiation mechanism at @xmath0-ray energies .
this behavior is qualitatively different from most probably ssc - dominated hbls where both the synchrotron and the @xmath0-ray component shift towards higher frequencies during the flare state .
i propose this qualitative difference as a new diagnostic tool to distinguish between the two competing radiation mechanisms potentially responsible for the production of @xmath0-rays in the jets of blazars .
results of a broadband campaign on 3c 273 have been used to support the prediction of the external - comptonization model , but future campaigns with more continuous frequency and temporal coverage in the radio ir spectral range are needed to draw solid conclusions .
other promising candidates for such campaigns include the fsrqs pks 0420 - 014 , pks 0521 - 365 , and 3c454.3 .
i thank r. mukherjee and p. sreekumar for valuable discussions , drawing my attention to this problem , and for useful comments on the manuscript .
i also thank the referee , s. d. bloom , for very useful suggestions .
this work was supported by nasa grant nag 5 - 4055 . | a prediction for the energy shift of the synchrotron spectrum of flat - spectrum radio quasars ( fsrqs ) during high - energy flares is presented . if the @xmath0-ray emission of fsrqs is produced by comptonization of external radiation , then the peak of the synchrotron spectrum
is predicted to move to lower energies in the flare state .
this is opposite to the well - known broadband spectral behavior of high - frequency peaked bl lac objects where the external radiation field is believed to be weak and synchrotron - self compton scattering might be the dominant @xmath0-ray radiation mechanism .
the synchrotron peak shift , if observed in fsrqs , can thus be used as a diagnostic to determine the dominant radiation mechanism in these objects .
i suggest a few fsrqs as promising candidates to test the prediction of the external - comptonization model . _ to appear in astrophysical journal letters _ |
the determination of the weighted mean is the fundamental problem for chi - square ( @xmath7 ) minimization methods .
the goodness - of - fit between an observation of @xmath14 data values , @xmath15 , with errors , @xmath16 , and a model , @xmath17 , can be determined by using the standard chi - square statistic : @xmath18 ^ 2\ , .
\label{eq : chisq_std}\ ] ] the theory of least - squares states that the optimum value of all the parameters of the model are obtained when the chi - square statistic is minimized with respect to each parameter simultaneously .
for example , the standard formula of the weighted mean can be derived by assuming that the model is a constant and then solving the equation , @xmath19 ^ 2 = 0 \ , \ ] ] for that constant : @xmath20 the standard weighted - mean formula thus weights every data value , @xmath15 , inversely by its own variance ( i.e. @xmath21 ) .
let us assume that all the data values come from a pure counting experiment where each data value , @xmath1 , is a random integer deviate drawn from a poisson ( @xcite ) distribution , @xmath22 with a mean value of @xmath23 .
let us also make the common assumption that the error of each data value is the square root of the mean of the parent poisson distribution . using these transformations , @xmath24 and @xmath25 , we see that equation ( [ eq : weighted_mean ] ) becomes @xmath26 which reduces to become the definition of the sample mean : @xmath27 in the limit of a large number of observations of the poisson distribution @xmath28 , we find that equation ( [ eq : standard_mean ] ) will , on average , determine the mean of the parent poisson distribution for all true mean values @xmath23 : @xmath29 & \equiv & \lim_{n \rightarrow \infty }
\left [ \frac{1}{n } \sum_{i=1}^n n_i \right ] \nonumber \\ & \approx & \lim_{n \rightarrow \infty } \left [ \frac{1}{n } \sum_{k=0}^{\infty } k \left\ { \rule[-0.1em]{0em}{1.2em }
n p(k;\mu ) \right\ } \right ] \nonumber \\ & = & \sum_{k=0}^{\infty } k \left\ { \rule[-0.1em]{0em}{1.2em } p(k;\mu ) \right\ } \nonumber \\ & = & \sum_{k=0}^{\infty } k \frac{\mu^k}{k!}e^{-\mu } \nonumber \\ & = & 0 \frac{\mu^0}{0!}e^{-\mu } + e^{-\mu } \sum_{k=1}^{\infty } \frac{\mu^k}{(k-1 ) ! } \nonumber \\ & = & e^{-\mu } \mu \sum_{k=1}^{\infty } \frac{\mu^{k-1}}{(k-1 ) ! } \nonumber \\ & = & e^{-\mu } \mu \sum_{j=0}^{\infty } \frac{\mu^j}{j ! } \nonumber \\ & = & e^{-\mu } \mu e^{\mu } \nonumber \\ & = & \mu \label{eq : weighted_mean_pearson_limit } \ .\end{aligned}\ ] ] _ applying the standard weighted mean formula , @xmath0 / \left[\sum_i { \sigma^{-2}_i}\right ] $ ] , to determine the weighted mean of data , @xmath1 , drawn from a poisson distribution , will , on average , determine the mean of the parent poisson distribution for all true mean values if a constant weight is assigned to all data values _
( i.e. @xmath30 constant ) .
it is a common practice to assume that the error of a poisson deviate @xmath31 is @xmath32 .
unfortunately , this practice causes the standard weighted - mean formula to be undefined for data values of zero .
a simple solution to this computational problem is to arbitrarily assign a non - zero constant error to all poisson deviates with a value of zero .
let us make the common assumption that the error of each data value , @xmath1 , is equal to @xmath33 or 1 whichever is greater . using the following transformations , @xmath34 and @xmath35
, we see that equation ( [ eq : weighted_mean ] ) becomes @xmath36 in the limit of a large number of observations of the poisson distribution @xmath28 , we find that @xmath37 & \equiv & \lim_{n \rightarrow \infty } \left [ \frac{\displaystyle \sum_{i=1}^n \frac{n_i}{\max(n_i,1 ) } } { \displaystyle \sum_{i=1}^n \frac{1}{\max(n_i,1 ) } } \right ] \nonumber \\ & \approx & \lim_{n \rightarrow \infty } \left [ \frac{\displaystyle \sum_{k=0}^\infty \frac{k}{\max(k,1 ) } \left\ { \rule[-0.1em]{0em}{1.2em } n p(k;\mu ) \right\ } } { \displaystyle \sum_{k=0}^\infty \frac{1}{\max(k,1 ) } \left\ { \rule[-0.1em]{0em}{1.2em } n p(k;\mu ) \right\ } } \right ] \nonumber \\ & = & \frac{\displaystyle \sum_{k=0}^\infty \frac{k}{\max(k,1 ) } \left\ { \rule[-0.1em]{0em}{1.2em } p(k;\mu ) \right\ } } { \displaystyle \sum_{k=0}^\infty \frac{1}{\max(k,1 ) } \left\ { \rule[-0.1em]{0em}{1.2em } p(k;\mu ) \right\ } } \nonumber \\ & = & \frac{\displaystyle \frac{0}{1 } p(0;\mu ) + \sum_{k=1}^\infty \frac{k}{k } p(k;\mu ) } { \displaystyle \frac{1}{1 } p(0;\mu ) + \sum_{k=1}^\infty \frac{1}{k } p(k;\mu ) } \nonumber \\ & = & \frac{\displaystyle 1-e^{-\mu } } { \displaystyle e^{-\mu } + e^{-\mu } \sum_{k=1}^\infty \frac{\mu^k}{k\,k ! } } \nonumber \\ & = & \frac{\displaystyle e^\mu - 1 } { \displaystyle 1 + \left [ \sum_{k=1}^\infty \frac{\mu^k}{k\,k ! } \right ] } \label{eq : wmean_limit_neyman_with_summation } \\ & = & \frac{\displaystyle e^\mu - 1 } { \displaystyle 1 + \left [ \rule[-0.1em]{0em}{1.2em } { \rm{ei}}(\mu ) - \gamma - \ln(\mu ) \right]~ , } \label{eq : wmean_limit_neyman}\end{aligned}\ ] ] where ei@xmath38 is the exponential integral of @xmath39 , @xmath40 ( for @xmath41 ) , and @xmath42 is the euler - mascheroni constant : @xmath43 = 0.5772156649\cdots $ ] ( see , e.g. , abramowitz & stegun @xcite ) .
let us now investigate the limit of equation ( [ eq : wmean_limit_neyman ] ) with large poisson mean values .
the transformation of equation ( [ eq : wmean_limit_neyman_with_summation ] ) to equation ( [ eq : wmean_limit_neyman ] ) used the power series of ei@xmath38 , @xmath44 which has the following asymptotic expansion : @xmath45 from the following limit , @xmath46 = -1~,\ ] ] we see that ei@xmath38 asymptotically approaches the function @xmath47 for large values of @xmath39 .
for @xmath48 this approximation has an error of @xmath491% ; for @xmath50 the error is @xmath510.1% . in the limit of large mean poisson values
, we see that the numerator of equation ( [ eq : wmean_limit_neyman ] ) is dominated by the @xmath52 term while the denominator is dominated by the ei@xmath53 term which asymptotically approaches the value of @xmath54 .
_ we then have come to the surprising conclusion that for poisson distributions with large mean values , @xmath55 $ ] approaches the value of @xmath56 instead of the expected value of @xmath23 .
_ equation ( [ eq : wmean_limit_neyman ] ) can also be investigated graphically .
figure [ fig - weighted_mean_modified_neyman]a plots the difference between the weighted mean computed using equation ( [ eq : weighted_mean_modified_neyman ] ) and the true mean for poisson - distributed data with true mean values between 0.001 and 1000 .
each open square represents the weighted mean of @xmath57@xmath58@xmath59 poisson deviates at each given true mean value .
the solid curve through the data [ open squares in fig .
[ fig - weighted_mean_modified_neyman]a ] is the difference between equation ( [ eq : wmean_limit_neyman ] ) and the true mean .
note that equation ( [ eq : wmean_limit_neyman ] ) underestimates the true mean by @xmath2@xmath3 for large true mean values ( as predicted above ) .
_ applying the standard weighted mean formula , @xmath0 / \left[\sum_i { \sigma^{-2}_i}\right ] $ ] , to determine the weighted mean of data , @xmath1 , drawn from a poisson distribution , will , on average , underestimate the true mean by @xmath2@xmath3 for all true mean values larger than @xmath2@xmath4 when the common assumption is made that the error of the @xmath5th observation is @xmath6 . _
we will now develop a weighted - mean formula for poisson - distributed data that will , on average , determine the true mean of the parent distribution for all true mean values .
let us assume that the error of each data value , @xmath1 , is equal to @xmath60 instead of @xmath61 . using the following transformations , @xmath24 and @xmath62
, we see that equation ( [ eq : weighted_mean ] ) becomes @xmath63 in the limit of a large number of observations of the poisson distribution @xmath28 , we find that @xmath64 & \equiv & \lim_{n \rightarrow \infty } \left [ \frac{\displaystyle \sum_{i=1}^n \frac{n_i}{n_i+1 } } { \displaystyle \sum_{i=1}^n \frac{1}{n_i+1 } } \right ] \nonumber \\ & \approx & \lim_{n \rightarrow \infty } \left [ \frac{\displaystyle \sum_{k=0}^\infty \frac{k}{k+1 } \left\ { \rule[-0.1em]{0em}{1.2em } n p(k;\mu ) \right\ } } { \displaystyle \sum_{k=0}^\infty \frac{1}{k+1 } \left\ { \rule[-0.1em]{0em}{1.2em } n p(k;\mu ) \right\ } } \right ] \nonumber \\ & = & \frac{\displaystyle \sum_{k=0}^\infty \frac{k}{k+1 } \left\ { \rule[-0.1em]{0em}{1.2em } p(k;\mu ) \right\ } } { \displaystyle \sum_{k=0}^\infty \frac{1}{k+1 } \left\ { \rule[-0.1em]{0em}{1.2em } p(k;\mu ) \right\ } } \nonumber \\ & = & \frac{\displaystyle \frac{1}{\mu}\left ( \mu - 1 + e^{-\mu } \right ) } { \displaystyle \frac{1}{\mu}\left ( 1 - e^{-\mu } \right ) } \nonumber \\ & = & \frac{\mu}{1-e^{-\mu } } - 1 \ .
\label{eq : weighted_mean_limit_alpha}\end{aligned}\ ] ] figure [ fig - weighted_mean_modified_neyman]b graphically confirms this finding . increasing the error estimates from @xmath65 to @xmath60 has only yielded a minor improvement .
notice that the dip in the solid curve in fig.[fig - weighted_mean_modified_neyman]a at @xmath66 is not present in the solid curve in fig.[fig - weighted_mean_modified_neyman]b .
a more radical change appears to be required in order for us to develop a weighted - mean formula for poisson - distributed data .
let us now add one to all data values and assume that the error of each data value is the square root of the new data value . using these transformations , @xmath67 and @xmath62 , we see that equation ( [ eq : weighted_mean ] ) becomes @xmath68 in the limit of a large number of observations of the poisson distribution @xmath28 , we find that @xmath69 & \equiv & \lim_{n \rightarrow \infty } \left [ \frac{\displaystyle \sum_{i=1}^n \frac{n_i+1}{n_i+1 } } { \displaystyle \sum_{i=1}^n \frac{1}{n_i+1 } } \right ] \nonumber \\ & \approx & \lim_{n \rightarrow \infty } \left [ \frac{\displaystyle \sum_{k=0}^\infty \frac{k+1}{k+1 } \left\ { \rule[-0.1em]{0em}{1.2em } n p(k;\mu ) \right\ } } { \displaystyle \sum_{k=0}^\infty \frac{1}{k+1 } \left\ { \rule[-0.1em]{0em}{1.2em } n p(k;\mu ) \right\ } } \right ] \nonumber \\ & = & \frac{\displaystyle \sum_{k=0}^\infty \left\ { \rule[-0.1em]{0em}{1.2em } p(k;\mu ) \right\ } } { \displaystyle \sum_{k=0}^\infty \frac{1}{k+1 } \left\ { \rule[-0.1em]{0em}{1.2em } p(k;\mu ) \right\ } } \nonumber \\ & = & \frac{\displaystyle 1 } { \displaystyle \frac{1}{\mu}\left ( 1 - e^{-\mu } \right ) } \nonumber \\ & = & \frac{\mu}{1-e^{-\mu } } \ .
\label{eq : weighted_mean_limit_beta}\end{aligned}\ ] ] figure [ fig - weighted_mean_modified_neyman]c graphically confirms this finding .
we have now made significant progress towards our goal of developing a weighted - mean formula for poisson - distributed data .
applying equation ( [ eq : weighted_mean_beta ] ) to determine the weighted mean of poisson - distributed data , will , on average , estimate the true mean with @xmath13@xmath3% errors for true poisson mean values @xmath70 .
the deviation of the solid curve in figure [ fig - weighted_mean_modified_neyman]c from zero can be eliminated by making just a minor change to our transformations . using the same errors as above , @xmath62 , but now adding one to only those data values that are initially greater than zero , @xmath71 , we see that equation ( [ eq : weighted_mean ] ) becomes @xmath72 in the limit of a large number of observations of the poisson distribution @xmath28 , we find that @xmath73 & \equiv & \lim_{n \rightarrow \infty } \left [ \frac{\displaystyle \sum_{i=1}^n \frac{n_i+\min(n_i,1)}{n_i+1 } } { \displaystyle \sum_{i=1}^n \frac{1}{n_i+1 } } \right ] \nonumber \\ & \approx & \lim_{n \rightarrow \infty } \left [ \frac{\displaystyle \sum_{k=0}^\infty \frac{k+\min(k,1)}{k+1 } \left\ { \rule[-0.1em]{0em}{1.2em } n p(k;\mu ) \right\ } } { \displaystyle \sum_{k=0}^\infty \frac{1}{k+1 } \left\ { \rule[-0.1em]{0em}{1.2em } n p(k;\mu ) \right\ } } \right ] \nonumber \\ & = & \frac{\displaystyle \sum_{k=0}^\infty \frac{k+\min(k,1)}{k+1 } \left\ { \rule[-0.1em]{0em}{1.2em } p(k;\mu ) \right\ } } { \displaystyle \sum_{k=0}^\infty \frac{1}{k+1 } \left\ { \rule[-0.1em]{0em}{1.2em } p(k;\mu ) \right\ } } \nonumber \\ & = & \frac{\displaystyle \frac{0}{1}p(0;\mu ) + \sum_{k=1}^\infty \frac{k+1}{k+1 } p(k;\mu ) } { \displaystyle \sum_{k=0}^\infty \frac{1}{k+1 } p(k;\mu ) } \nonumber \\ & = & \frac{\displaystyle 1 - e^{-\mu } } { \displaystyle \frac{1}{\mu}\left ( 1 - e^{-\mu } \right ) } \nonumber \\ & = & \mu \ .
\label{eq : weighted_mean_limit_gamma}\end{aligned}\ ] ] figure [ fig - weighted_mean_modified_neyman]d graphically confirms this finding .
we have now achieved our goal of developing a weighted - mean formula for poisson - distributed data . _
applying equation ( [ eq : weighted_mean_gamma ] ) to determine the weighted mean of poisson - distributed data , will , on average , estimate the true mean for all true poisson mean values _ ( @xmath74 ) .
based on my finding that the weighted mean of data drawn from a poisson distribution can be determined using the formula @xmath10\left(n_i+1\right)^{-1 } \right ] / \left [ \sum_i \left(n_i+1\right)^{-1 } \right ] $ ] , i propose that , given @xmath14 observations ( @xmath1 ) and a model ( @xmath17 ) , a new @xmath7 statistic , @xmath75 ^ 2 } { n_i+1 } \ , \ ] ] should always be used to analyze poisson - distributed data in preference to the modified neyman s @xmath7 statistic , @xmath76 because the weighted - mean formula for the modified neyman s @xmath7 statistic [ @xmath77 : equation ( [ eq : weighted_mean_modified_neyman ] ) ] systematically underestimates the true mean value of poisson - distributed data with true mean values @xmath78 ( see fig . [
fig - weighted_mean_modified_neyman]a ) .
for poisson - distributed data , it has long been observed that , in many cases , chi - square fits using the modified neyman s @xmath7 statistic and the pearson s @xmath7 statistic , @xmath79 will underestimate and overestimate the total area , respectively , while the usage of the maximum likelihood ratio statistic for poisson distributions , @xmath80 \label{eq : x2l } \ , \ ] ] preserves the total area ( e.g. , baker & cousins @xcite and references therein ) .
it has been known for decades that chi - square minimization techniques using the modified neyman s @xmath7 statistic to analyze poisson - distributed data will typically predict a total number of counts ( total area ) that underestimates the true total counts by about @xmath3 count per bin ( e.g. , bevington @xcite , wheaton @xcite ) .
the reason why this underestimation occurs is now obvious : the application of the modified neyman s @xmath7 statistic to poisson - distributed data causes the fitted model value at each bin , @xmath17 , to be , on average , underestimated by @xmath2@xmath3 count for all true poisson model mean values @xmath81@xmath4 .
the underestimation of the true mean by one count gives a very large @xmath82% error when the true mean of the data is 5 but only a @xmath3% error when the true mean of the data is 100 .
it would clearly be difficult to detect such a small systematic error with _
small samples _ of poisson - distributed data with _ large true mean values_. figure [ fig - weighted_mean_modified_neyman]a shows that this underestimation is real and is easily measurable with _ large samples _ of poisson - distributed data .
the number of degrees of freedom , commonly represented with the symbol @xmath83 , of a chi - square minimization problem is the difference between the number of observations ( sample size ) and the number of free parameters ( @xmath84 ) of the model : @xmath85 .
the reduced chi - square of the pearson s @xmath7 statistic is , by definition , the value of pearson s @xmath7 statistic divided by the number of degrees of freedom : @xmath86 on average , the expected reduced chi - square value of a proper @xmath7 statistic with a perfect model is one given a large number of observations .
now let us assume that our data comes from a poisson distribution with a mean value of @xmath23 . in this case ,
the model @xmath17 will be a constant , @xmath87 [ equation ( [ eq : weighted_mean_pearson ] ) ] , which will , on average , have a value , @xmath88 , given by equation ( [ eq : weighted_mean_pearson_limit ] ) in the limit of a large number of observations ( n.b .
@xmath89 ) . the model is a constant and therefore there is only one degree - of - freedom : @xmath90 . given these assumptions , we find that , in the limit of a large number of observations , the reduced chi - square of the pearson s @xmath7 statistic with the model @xmath91 is @xmath92 & \equiv & \lim_{n\rightarrow\infty } \left [ \frac{1}{n - m } { \sum_{i=1}^{n } } \frac { ( n_i - m_i)^2 } { m_i } \right ] \nonumber \\ & = & \lim_{n\rightarrow\infty } \left [ \frac{1}{n-1 } { \sum_{i=1}^{n } } \frac { ( n_i - \mu_{\rm{p}})^2 } { \mu_{\rm{p } } } \right ] \nonumber \\ & \approx & \lim_{n\rightarrow\infty } \left [ \frac{1}{n-1 } { \sum_{k=0}^{\infty } } \frac { ( k - \mu_{\rm{p}^\prime})^2 } { \mu_{\rm{p}^\prime } } \left\ { \rule[-0.1em]{0em}{1.2em } n p(k;\mu ) \right\ } \right ] \nonumber \\ & = & { \sum_{k=0}^{\infty } } \frac { ( k - \mu_{\rm{p}^\prime})^2 } { \mu_{\rm{p}^\prime } } \left\ { \rule[-0.1em]{0em}{1.2em } p(k;\mu ) \right\ } \nonumber \\ & = & { \sum_{k=0}^{\infty } } \frac { ( k - \mu)^2 } { \mu }
\left\ { \rule[-0.1em]{0em}{1.2em } p(k;\mu ) \right\ } \nonumber \\ & = & \frac{1}{\mu } \left ( \left [ { \sum_{k=0}^{\infty } } k^2 p(k;\mu ) \right ] - 2\mu \left [ { \sum_{k=0}^{\infty } } k p(k;\mu ) \right ] + \mu^2 \left [ { \sum_{k=0}^{\infty } } p(k;\mu ) \right ] \,\right ) \nonumber \\ & = & \frac{1}{\mu } \left ( \left [ \rule[-0.1em]{0em}{1.2em } \mu^2 + \mu \right ] - 2\mu \left [ \rule[-0.1em]{0em}{1.2em } \mu \right ] + \mu^2 \left [ \rule[-0.1em]{0em}{1.2em } 1 \right ] \,\right ) \nonumber \\ & = & 1~.\end{aligned}\ ] ] the reduced chi - square of the modified neyman s @xmath7 statistic is , by definition , the value of the modified neyman s @xmath7 statistic divided by the number of degrees of freedom : @xmath93 now let us assume that our data comes from a poisson distribution with a mean value of @xmath23 . in this case , the model @xmath17 will be a constant , @xmath77 [ equation ( [ eq : weighted_mean_modified_neyman ] ) ] , which will , on average , have a value , @xmath94 , given by equation ( [ eq : wmean_limit_neyman ] ) in the limit of a large number of observations .
given these assumptions , we find that , in the limit of a large number of observations , the reduced chi - square of the modified neyman s @xmath7 statistic with the model @xmath95 is @xmath96 & \equiv & \lim_{n \rightarrow \infty } \left
[ \frac{1}{n - m } { \sum_{i=1}^{n } } \frac { ( n_i - m_i)^2 } { \max(n_i,1 ) } \right ] \nonumber \\ & = & \lim_{n \rightarrow \infty } \left [ \frac{1}{n-1 } { \sum_{i=1}^{n } } \frac { ( n_i - \mu_{\rm{n}})^2 } { \max(n_i,1 ) } \right ] \nonumber \\ & \approx & \lim_{n \rightarrow \infty } \left [ \frac{1}{n-1 } { \sum_{k=0}^{\infty } } \frac { ( k - \mu_{\rm{n}^\prime})^2 } { \max(k,1 ) } \left\ { \rule[-0.1em]{0em}{1.2em } n p(k;\mu ) \right\ } \right ] \nonumber \\ & = & { \sum_{k=0}^{\infty } } \frac { ( k - \mu_{\rm{n}^\prime})^2 } { \max(k,1 ) } \left\ { \rule[-0.1em]{0em}{1.2em } p(k;\mu ) \right\ } \nonumber \\ & = & \mu_{\rm{n}^\prime}^2 e^{-\mu } + { \sum_{k=1}^{\infty } } \frac { ( k - \mu_{\rm{n}^\prime})^2 } { k } p(k;\mu ) \nonumber \\ & = & \mu_{\rm{n}^\prime}^2 e^{-\mu } + \left [ \rule[-0.1em]{0em}{1.2em } { \sum_{k=1}^{\infty } } k p(k;\mu ) \right ] -2\mu_{\rm{n}^\prime } \left [ \rule[-0.1em]{0em}{1.2em } { \sum_{k=1}^{\infty } } p(k;\mu ) \right ] + \mu_{\rm{n}^\prime}^2 \left\ { \rule[-0.1em]{0em}{1.2em } { \sum_{k=1}^{\infty } } \frac{1}{k } p(k;\mu ) \right\ } \nonumber \\ & = & \mu_{\rm{n}^\prime}^2 e^{-\mu } + \left [ \rule[-0.1em]{0em}{1.2em } \mu \right ] -2\mu_{\rm{n}^\prime } \left [ \rule[-0.1em]{0em}{1.2em } 1-e^{-\mu } \right ] + \mu_{\rm{n}^\prime}^2 \left\ { \rule[-0.1em]{0em}{1.2em } e^{-\mu } \bigl [ \mbox{ei}(\mu ) - \gamma - \ln(\mu ) \bigl ] \right\ } \nonumber \\ & = & \mu_{\rm{n}^\prime}^2 e^{-\mu } \bigl [ 1 + \mbox{ei}(\mu ) - \gamma - \ln(\mu ) \bigl ] -2\mu_{\rm{n}^\prime } \left [ \rule[-0.1em]{0em}{1.2em } 1-e^{-\mu } \right ] + \left [ \rule[-0.1em]{0em}{1.2em } \mu \right ] \nonumber \\ & = & \mu_{\rm{n}^\prime}^2 e^{-\mu } \left [ \frac{e^{\mu}-1}{\mu_{\rm{n}^\prime } } \right ] -2\mu_{\rm{n}^\prime } \left [ \rule[-0.1em]{0em}{1.2em } 1-e^{-\mu } \right ] + \left [ \rule[-0.1em]{0em}{1.2em } \mu \right ] \nonumber \\ & = & \left\ { \mu_{\rm{n}^\prime } \right\ } \left [ e^{-\mu } - 1 \right ] + \mu \nonumber \\ & = & \left\ { \frac{e^\mu-1}{1 + \mbox{ei}(\mu ) - \gamma - \ln(\mu ) } \right\ } \left [ e^{-\mu } - 1 \right ] + \mu \nonumber \\ & = & \frac{\displaystyle 2 - e^{\mu } - e^{-\mu } } { \displaystyle { \rm{ei}}(\mu ) - \gamma - \ln(\mu ) + 1 } + \mu \label{eq : reduced_x2n } { } ~.\end{aligned}\ ] ] in the limit of large mean poisson values , we see that the numerator of the first term of equation ( [ eq : reduced_x2n ] ) is dominated by the @xmath97 term while the denominator of the first term is dominated by the ei@xmath53 term which asymptotically approaches the value of @xmath54 .
we then conclude that the reduced chi - square of the @xmath98 statistic applied to a poisson distribution [ equation ( [ eq : reduced_x2n ] ) ] approaches the value of one for large true poisson mean values .
figure [ fig - reduced_chi_square ] graphically confirms this finding ; we see that equation ( [ eq : reduced_x2n ] ) reaches a value of @xmath2@xmath3 only for very large true poisson mean values ( @xmath99 ) .
the reduced chi - square of the new @xmath12 statistic is , by definition , the value of the @xmath12 statistic divided by the number of degrees of freedom : @xmath100 ^ 2 } { n_i + 1 } \ .\ ] ] now let us assume that our data comes from a poisson distribution with a mean value of @xmath23 . in this case , the model @xmath17 will be a constant , @xmath101 [ equation ( [ eq : weighted_mean_gamma ] ) ] , which will , on average , have a value , @xmath102 , given by equation ( [ eq : weighted_mean_limit_gamma ] ) in the limit of a large number of observations ( n.b .
@xmath103 ) . given these assumptions , we find that , in the limit of a large number of observations , the reduced chi - square of the new @xmath12 statistic with the model @xmath101 is @xmath104 & \equiv & \lim_{n \rightarrow \infty } \left [ \frac{1}{n - m } { \sum_{i=1}^{n } } \frac { [ n_i + \min(n_i,1 ) - m_i]^2 } { n_i + 1 } \right ] \nonumber \\ & = & \lim_{n \rightarrow \infty } \left [ \frac{1}{n-1 } { \sum_{i=1}^{n } } \frac { [ n_i + \min(n_i,1 ) - \mu_\gamma]^2 } { n_i + 1 } \right ] \nonumber \\ & \approx & \lim_{n \rightarrow \infty }
\left [ \frac{1}{n-1 } { \sum_{k=0}^{\infty } } \frac { [ k + \min(k,1 ) - \mu_{\gamma^\prime}]^2 } { k + 1 } \left\ { \rule[-0.1em]{0em}{1.2em } n p(k;\mu ) \right\ } \right ] \nonumber \\ & = & { \sum_{k=0}^{\infty } } \frac { [ k + \min(k,1 ) - \mu_{\gamma^\prime}]^2 } { k + 1 } \left\ { \rule[-0.1em]{0em}{1.2em } p(k;\mu ) \right\ } \nonumber \\ & = & { \sum_{k=0}^{\infty } } \frac { [ k + \min(k,1 ) - \mu]^2 } { k + 1 } \left\ { \rule[-0.1em]{0em}{1.2em } p(k;\mu ) \right\ } \nonumber \\ & = & \mu^2 e^{-\mu } + { \sum_{k=1}^{\infty } } \frac { [ k + 1 - \mu]^2 } { k + 1 } \left\ { \rule[-0.1em]{0em}{1.2em } p(k;\mu ) \right\ } \nonumber \\ & = & \mu^2 e^{-\mu } + \left [ \rule[-0.1em]{0em}{1.2em } { \sum_{k=1}^{\infty } } k p(k;\mu ) \right ] + \left ( 1 - 2\mu \right ) \left [ \rule[-0.1em]{0em}{1.2em } { \sum_{k=1}^{\infty } } p(k;\mu ) \right ] + \mu^2 \left [ \rule[-0.1em]{0em}{1.2em } { \sum_{k=1}^{\infty } } \frac{1}{k+1 } p(k;\mu ) \right ] \nonumber \\ & = & \mu^2 e^{-\mu } + \left [ \rule[-0.1em]{0em}{1.2em } \mu \right ] + \left ( 1 - 2\mu \right ) \left [ \rule[-0.1em]{0em}{1.2em } 1-e^{-\mu } \right ] + \mu^2 \left [ \rule[-0.1em]{0em}{1.2em } \frac{1}{\mu}\left(1-e^{-\mu}\right ) - e^{-\mu } \right ] \nonumber \\ & = & 1 + e^{-\mu}\left ( \mu - 1 \right ) \label{eq : reduced_x2 g } { } ~.\end{aligned}\ ] ] figure [ fig - reduced_chi_square ] shows that the reduced chi - square of the @xmath105 statistic applied to a poisson distribution [ equation ( [ eq : reduced_x2 g ] ) ] approaches the value of one for small true poisson mean values ( i.e. @xmath106 ) .
figure [ fig - variance_reduced_chi_square ] shows the variance of the reduced chi - square of the @xmath107 , @xmath98 , @xmath12 , and @xmath108 statistics as a function of the true poisson mean .
this figure was derived by analyzing the data used in figure [ fig - weighted_mean_modified_neyman ] .
i now demonstrate the new @xmath12 statistic by using it to study a dataset of simulated x - ray power - law spectra . this dataset is based on my duplication of the simple numerical experiment of nousek & shue ( @xcite ) .
the number of x - ray photons per energy interval ( bin ) of a x - ray power - law spectrum is @xmath109 over an energy range , @xmath110 kev , the expectation value for the total number of counts can be determined as follows @xmath111 which implies that @xmath112 following nousek & shue , i chose the slope value of @xmath113 and used the energy range of 0.0950.845 kev which was split into 15 equal bins of 0.050 kev per bin .
i simulated @xmath114 x - ray spectra for each of the theoretical @xmath14 values used by nousek & shue : 25 , 50 , 75 , 100 , 150 , 250 , 500 , 750 , 1000 , 2500 , 5000 , and @xmath114 photons per spectrum .
figure [ fig - spectra ] shows four of the simulated x - ray power - law spectra .
i determined the best - fit model parameters @xmath115 and @xmath116 for each simulated spectrum with powell s function minimization method using the modified neyman s @xmath7 statistic ( @xmath98 ) , pearson s @xmath7 statistic ( @xmath107 ) , and the new @xmath12 statistic .
i used the following crude initial guesses : @xmath117 and @xmath118 , where @xmath1 is the observed number of photons in the @xmath5th channel ( bin ) .
i computed the robust mean ( average ) and robust standard deviation@xmath119 where @xmath119 is the standard deviation of all values within two average deviations of the standard mean values .
] of the ratios @xmath120 and @xmath121 for the @xmath114 simulated spectra of each dataset .
the results of powell s method with two free parameters ( @xmath122 ) using the @xmath98 , @xmath107 , @xmath12 statistics are presented in table [ tbl - powell_2fp ] and figure [ fig - powell_2fp ] .
the first column , @xmath14 , of table [ tbl - powell_2fp ] corresponds to the total theoretical number of counts in the spectrum .
the columns `` @xmath120 '' and `` @xmath121 '' are the robust mean values of the ratios of the best - fit parameters divided by the original value that was used to create the datasets .
the parenthetical numbers are the robust standard deviations which can be used to determine the significance of the deviation from the perfect ratio value of one .
for example , the first value of the 2nd column of table [ tbl - powell_2fp ] is 1.002(11 ) which represents the value of @xmath123@xmath124@xmath125 .
the deviation of this value from one ( i.e. @xmath126 ) is statistically significant since the error of the mean is only @xmath2@xmath127 or @xmath2@xmath128 .
figure [ fig - powell_2fp ] indicates that the new @xmath12 statistic gives the best results . using a 5% criteria for both fitted parameters @xmath129
, we see that the @xmath12 statistic gives good results for spectra with @xmath81@xmath130 photons . by comparison
, pearson s @xmath7 statistic requires @xmath81@xmath131 photons and the modified neyman s @xmath7 statistic requires @xmath81@xmath132 photons in order to get the same quality of results .
baker & cousins ( @xcite ) noted that , in many cases , @xmath7 fits using the the modified neyman s @xmath7 statistic will underestimate the total number of counts while @xmath7 fits using pearson s @xmath7 statistic will overestimate the total number of counts ; both systematic errors are clearly seen in the bottom panel of figure [ fig - powell_2fp ] .
i stated in the previous section that the usage of the modified neyman s @xmath7 statistic with poisson - distributed data will typically underestimate the total counts by one count per bin .
my results for the @xmath98 statistic clearly exhibit this systematic error : the results of the ratio @xmath121 for spectra with @xmath133 photons ( squares in the bottom panel of fig .
[ fig - powell_2fp ] ) are well modeled by the function @xmath134 where 15 is the number of bins ( channels ) in our spectra [ see the dashed curve in the bottom panel of fig . [ fig - powell_2fp ] ] .
a comparison of my analysis of @xmath120 using the modified neyman s @xmath7 statistic ( 2nd column of table [ tbl - powell_2fp ] ) with the analysis of nousek & shue for powell s method ( 3rd column of their table 3 ) shows nearly identical results . in my version of this numerical experiment
, i used the two parameters @xmath14 and @xmath42 while nousek & shue used @xmath135 and @xmath42 .
a comparison of my analysis of @xmath121 ( 3rd column of table [ tbl - powell_2fp ] ) with their powell s method analysis of @xmath136 ( 2nd column of their table 3 ) shows that my analysis with @xmath121 has produced better estimates . this should not be surprising because the parameter @xmath135 is not an independent parameter
@xmath135 depends on both the slope of the spectrum and the theoretical number of photons in the spectrum . as a general rule ,
one gets better results by solving for independent parameters instead of dependent parameters .
i determined the best - fit model parameters @xmath115 and @xmath116 for each simulated spectrum with levenberg - marquardt method using the modified neyman s @xmath7 statistic ( @xmath98 ) , pearson s @xmath7 statistic ( @xmath107 ) , and the new @xmath12 statistic .
i used the previous crude initial guesses : @xmath117 and @xmath118 .
i computed the robust mean and robust standard deviation of the ratios @xmath120 and @xmath121 for the @xmath114 simulated spectra of each dataset .
the results of levenberg - marquardt method with two free parameters ( @xmath122 ) using the @xmath98 , @xmath107 , @xmath12 statistics are presented in table [ tbl - marquardt_2fp ] and figure [ fig - marquardt_2fp ] .
figure [ fig - marquardt_2fp ] indicates that the new @xmath12 statistic gives the best results . using a 5% criteria for both fitted parameters @xmath129
, we see that the @xmath12 statistic gives good results for all the spectra ( @xmath137 photons ) . by comparison
, pearson s @xmath7 statistic requires @xmath81@xmath138 photons and the modified neyman s @xmath7 statistic requires @xmath81@xmath139 photons in order to get the same quality of results .
the results for the @xmath12 and @xmath98 statistics are nearly identical with either powell s method ( table [ tbl - powell_2fp ] ) or the levenberg - marquardt method ( table [ tbl - marquardt_2fp ] ) .
this finding refutes the determination by nousek & shue ( @xcite ) that powell s method gives more accurate results than the levenberg - marquardt method . the results for pearson s @xmath7 improved significantly by using the levenberg - marquardt method instead of powell s method .
an inspection of the individual fits showed that the levenberg - marquardt method with the @xmath107 statistic produced a best - fit value for @xmath14 that was within a one - tenth of one percent of the total number of photons in the spectrum .
needless to say , with such an improvement in the determination of @xmath14 , a much better estimate for the slope @xmath42 could be determined .
this peculiar result tells us something important about this particular minimization problem : an excellent estimate of the total number of photons in the best - fit spectrum is the total number of photons in the actual spectrum .
thus by setting @xmath14 to be a constant , @xmath140 , we can eliminate one parameter and solve for @xmath42 alone .
i determined the best - fit model parameter @xmath115 for each simulated spectrum with powell s function minimization method using the modified neyman s @xmath7 statistic ( @xmath98 ) , pearson s @xmath7 statistic ( @xmath107 ) , and the new @xmath12 statistic .
i set @xmath141 and used the crude initial guess of @xmath117 .
i computed the robust mean and robust standard deviation of the ratios @xmath120 for the @xmath114 simulated spectra of each dataset .
the results of powell s method with two free parameters ( @xmath122 ) using the @xmath98 , @xmath107 , @xmath12 statistics are presented in table [ tbl - powell_1fp ] and figure [ fig - powell_1fp ] .
figure [ fig - powell_1fp ] indicates that the new @xmath12 statistic gives the best results . using a 5% criteria
, we see that the @xmath12 statistic gives good results for all the spectra ( @xmath137 photons ) . by comparison
, pearson s @xmath7 statistic requires @xmath81@xmath131 photons and the modified neyman s @xmath7 statistic requires @xmath81@xmath132 photons in order to get the same quality of results .
fitting only for the slope @xmath42 has improved the results for the new @xmath12 statistic and the modified neyman s @xmath7 statistic .
the results for pearson s @xmath7 show no improvement over the two free parameter result .
i determined the best - fit model parameter @xmath115 for each simulated spectrum with the levenberg - marquardt minimization method using the modified neyman s @xmath7 statistic ( @xmath98 ) , pearson s @xmath7 statistic ( @xmath107 ) , and the new @xmath12 statistic .
i set @xmath141 and used the crude initial guess of @xmath117 .
i computed the robust mean and robust standard deviation of the ratios @xmath120 for the @xmath114 simulated spectra of each dataset .
the results of the levenberg - marquardt method with one free parameter ( @xmath42 ) using the @xmath98 , @xmath107 , @xmath12 statistics are presented in table [ tbl - marquardt_1fp ] and figure [ fig - marquardt_1fp ] .
figure [ fig - marquardt_1fp ] indicates that pearson s @xmath7 statistic gives the best results . using a 5% criteria
, we see that both the new @xmath12 statistic and the @xmath107 statistic give good results for all the spectra ( @xmath137 photons ) . by comparison , the modified neyman s
@xmath7 statistic still requires @xmath81@xmath132 photons in order to get the same quality of results .
once again , we note that the results for the @xmath12 and @xmath98 statistics are nearly identical with either powell s method ( table [ tbl - powell_1fp ] ) or the levenberg - marquardt method ( table [ tbl - marquardt_1fp ] ) .
one expects the quality of the slope determination to degrade as the total number of photons in the x - ray spectra decline .
figure [ fig - marquardt_powell_1fp ] shows the distribution of the best - fit values for the slope @xmath42 for the faintest spectra with a theoretical total of 100 , 50 , and 25 photons .
as expected , the range of best - fit slope values measured for spectra with only @xmath142 photons is considerably larger than the range of values for spectra with @xmath143 photons .
the levenberg - marquardt method not only provides best - fit values for parameters but it also provides an error estimate ( approximately @xmath144 errors ) of those fitted parameters . how believable are these error estimates ?
figure [ fig - error_estimates ] shows an analysis of the errors estimated by the levenberg - marquardt method when the new @xmath12 statistic was used to analyze spectra with theoretical totals of 100 , 50 , and 25 photons .
the top panel of figure [ fig - error_estimates ] shows the error analysis of spectra with @xmath143 photons .
the median slope value is 1.989 and the median error estimate is 0.194 .
a total of 15.87% of the spectra have estimates of @xmath145 and 15.87% of the spectra have estimates of @xmath146 . for a normal distribution ,
one expects 68.26% of the deviates to be found within one standard deviation of the mean . assuming that the distribution of best - fit @xmath42 values approximates a normal distribution , then half of the difference between the 84.13 and 15.87 percentile values of @xmath42 can be used as an estimate for the slope error : @xmath147 .
this value is 8.8% larger than the median levenberg - marquardt error estimate ; a fractional error of 10.6% instead of the predicted 9.8% .
the middle panel of figure [ fig - error_estimates ] shows the error analysis of spectra with @xmath148 photons .
the median slope value is 2.009 and the median error estimate is 0.301 .
a total of 15.87% of the spectra have estimates of @xmath149 and 15.87% of the spectra have estimates of @xmath150 .
this gives an estimated slope error of @xmath151 .
this value is exactly equal to the median levenberg - marquardt error estimate .
the bottom panel of figure [ fig - error_estimates ] shows the error analysis of spectra with @xmath142 photons .
the median slope value is 2.071 and the median error estimate is 0.484 .
a total of 15.87% of the spectra have estimates of @xmath152 and 15.87% of the spectra have estimates of @xmath153 .
this gives an estimated slope error of @xmath154 .
this value is 9.3% less than the median levenberg - marquardt error estimate ; a fractional error of 21.2% instead of the predicted 23.4% . _ the errors estimated by the levenberg - marquardt method are seen to be reasonable .
_ figure [ fig - spectra_fits ] shows the simulated x - ray spectra of fig .
[ fig - spectra ] now plotted with @xmath12 fits produced by the levenberg - marquardt method with one free parmater .
the levenberg - marquardt method has done a good job even with the two faintest spectra which have actual totals of only 28 and 101 photons . for the sake of completeness
, i determined the best - fit model parameter @xmath115 for each simulated spectrum with powell s function minimization method using the maximum likelihood ratio statistic for poisson distributions , @xmath108 [ equation ( [ eq : x2l ] ) ] , and cash s @xmath155 statistic , @xmath156\ ] ] [ equation ( 6 ) of cash @xcite ] .
i set @xmath141 and used the crude initial guess of @xmath117 .
i computed the robust mean and robust standard deviation of the ratios @xmath120 for the @xmath114 simulated spectra of each dataset .
the results of powell s method with one free parameter ( @xmath42 ) using the @xmath108 statistic and cash s @xmath155 statistic are presented in table [ tbl - powell_1fp_x2l ] and figure [ fig - x2l ] .
table [ tbl - powell_1fp_x2l ] and the right panel of figure [ fig - x2l ] shows that cash s @xmath155 statistic and the maximum likelihood ratio statistic for poisson distributions , @xmath108 , give identical results .
this is not surprising because cash s @xmath155 statistic is a variant of the more well - known @xmath108 statistic which has been discussed in the literature for over 70 years ( e.g. , neyman & pearson @xcite ) .
i also determined the best - fit model parameter @xmath115 for each simulated spectrum with the levenberg - marquardt minimization method using the maximum likelihood ratio statistic for poisson distributions , @xmath108 .
i set @xmath141 and used the crude initial guess of @xmath117 .
i computed the robust average and robust standard deviation of the ratios @xmath120 for the @xmath114 simulated spectra of each dataset .
the results of the levenberg - marquardt method with one free parameter ( @xmath42 ) using the @xmath108 statistic is presented in table [ tbl - marquardt_1fp_x2l ] and figure [ fig - x2l ] .
the maximum likelihood ratio statistic for poisson distributions , @xmath108 , produces nearly identical results with either powell s method or the levenberg - marquardt minimization method .
of the two statistics , @xmath108 and the new @xmath12 , which is better ?
although tables [ tbl - marquardt_1fp_x2l ] and [ tbl - marquardt_1fp ] indicate that the @xmath108 is slightly better , we see that the actual differences between the distributions presented in figure [ fig - x2l ] are really quite negligible when compared with the overall uncertainty caused by simple sampling errors ( counting statistics ) of the simulated x - ray spectra .
i have demonstrated that the application of the standard weighted mean formula , @xmath0 / \left[\sum_i { \sigma^{-2}_i}\right ] $ ] , to determine the weighted mean of data , @xmath1 , drawn from a poisson distribution , will , on average , underestimate the true mean by @xmath2@xmath3 for all true mean values larger than @xmath2@xmath4 when the common assumption is made that the error of the @xmath5th observation is @xmath6 .
this small , but statistically significant offset , explains the long - known observation that chi - square minimization techniques which use the modified neyman s @xmath7 statistic , @xmath8 , to compare poisson - distributed data with model values , @xmath9 , will typically predict a total number of counts that underestimates the true total by about @xmath3 count per bin . based on my finding that the weighted mean of data drawn from a poisson distribution can be determined using the formula @xmath10\left(n_i+1\right)^{-1 } \right ] / \left [ \sum_i \left(n_i+1\right)^{-1 } \right ] $ ]
, i proposed that a new @xmath7 statistic , @xmath11 ^ 2 / \left [ n_i + 1 \right]$ ] , should always be used to analyze poisson - distributed data in preference to the modified neyman s @xmath7 statistic .
i demonstrated the power and usefulness of @xmath12 minimization by using two statistical fitting techniques ( powell s method and the levenberg - marquardt method ) and five @xmath7 statistics ( @xmath98 , @xmath107 , @xmath12 , @xmath108 , and cash s @xmath155 ) to analyze simulated x - ray power - law 15-channel spectra with large and small counts per bin .
i showed that @xmath12 minimization with the levenberg - marquardt or powell s method can produce excellent results ( mean slope errors @xmath13@xmath4% ) with spectra having as few as 25 total counts .
this analysis shows that there is nothing inherently wrong with either the levenberg - marquardt method or powell s method in the low - count regime provided that one uses an appropriate @xmath7 statistic for the type of data being analyzed .
given poisson - distributed data , one should always use the new @xmath12 statistic in preference to the modified neyman s @xmath7 statistic because that statistic produces small , but statistically significant , systematic errors with poisson - distributed data . while the new @xmath12 statistic is not perfect , neither is the more well - known @xmath108 statistic ( e.g. , see figures [ fig - reduced_chi_square ] and [ fig - variance_reduced_chi_square ] ) .
both statistics have problems in the very - low - count regime .
the new @xmath12 statistic complements but does not replace the older @xmath108 statistic .
which statistic is `` best '' will generally depend on the particular problem being analyzed .
an important difference between these two statistics is that the @xmath108 statistic assumes that all data is perfect . with data from perfect counting experiments ,
the @xmath108 statistic may give slightly better results than the new @xmath12 statistic .
however , data is typically obtained under less - than - perfect circumstances with multiple imperfect detectors .
the @xmath12 statistic , by definition , is a _ weighted _
@xmath7 statistic which makes it easy to assign a _
lower _ weight to data from poor detectors .
thus in the analysis of real data obtained with noisy and imperfect detectors , the @xmath12 statistic may well outperform the classic @xmath108 statistic because low - quality data can be given a lower weight instead of being completely rejected .
finally , i note in passing that two simple transformations may make it possible to retrofit many existing computer implementations ( i.e. executable binaries ) of @xmath98 minimization algorithms to do @xmath12 minimization through the simple expedient of _ changing the input data _ from @xmath157 $ ] to @xmath158 $ ] , and error estimates , @xmath16 , from @xmath159 $ ] to @xmath160 $ ] .
i would like to thank my former colleagues at mount stromlo and siding spring observatories and current colleages at the national optical astronomy observatories with whom i have had many useful discussions about various topics related to this research .
keith arnaud , martin elvis , and andy ptak provided valuable practical information about the implementation details of the @xmath108 statistic within the x - ray spectral - fitting program .
i would also like to thank the anonymous referee whose thoughtful comments and suggestions improved the final article .
i am grateful to the anonymous referee of the previous version of this paper ( based on the @xmath161 statistic ) for providing a copy of the baker & cousins ( @xcite ) article .
special thanks are due to christopher burke who checked all the proofs in an early draft of the manuscript and to ian dellantonio who carefully read the second - to - last revision of the manuscript .
mary guerrieri has my heartfelt thanks for cheerfully helping me acquire copies of the many articles in statistical journals which are not available in the noao library collection .
kjm was supported by a grant from the national aeronautics and space administration ( nasa ) , order no .
s-67046-f , which was awarded by the long - term space astrophysics program ( nra 95-oss-16 ) .
publication costs were partially paid by nasa through grant no .
stsci go-5386 from the space telescope science institute , which is operated by the by the association of universities for research in astronomy , inc .
, under nasa contract no .
nas5 - 26555 .
this research has made use of nasa s astrophysics data system abstract service which is operated by the jet propulsion laboratory at the california institute of technology , under contract with nasa .
abramowitz , m. , & stegun , i. 1964 , `` handbook of mathematical functions with formulas , graphs , and mathematical tables '' , national bureau of standards , applied mathematics series # 55 , edited by m. abramowitz & i. stegun ( washington d.c . : u.s . government printing office ) 01cap [ fig - weighted_mean_modified_neyman ] analysis of four weighted - mean formulae applied to poisson - distributed data .
each open square represents the weighted mean of @xmath57@xmath58@xmath59 poisson deviates at each given true mean value : @xmath162 .
the difference between the weighted mean computed using equation ( [ eq : weighted_mean_modified_neyman ] ) , @xmath77 , and the true mean , @xmath163 .
the solid curve is the difference between equation ( [ eq : wmean_limit_neyman ] ) and the true mean : @xmath164[1+{\rm{ei}}(\mu)-\gamma-\ln(\mu)]^{-1}\right\}-\mu.$ ] the difference between the weighted mean computed using equation ( [ eq : weighted_mean_alpha ] ) , @xmath165 , and the true mean , @xmath163 .
the solid curve is the difference between equation ( [ eq : weighted_mean_limit_alpha ] ) and the true mean : @xmath166^{-1}-1\right\}-\mu$ ] .
the difference between the weighted mean computed using equation ( [ eq : weighted_mean_beta ] ) , @xmath167 , and the true mean , @xmath163 .
the solid curve is the difference between equation ( [ eq : weighted_mean_limit_beta ] ) and the true mean : @xmath166^{-1}\right\}-\mu$ ] .
the difference between the weighted mean computed using equation ( [ eq : weighted_mean_gamma ] ) , @xmath101 , and the true mean , @xmath163 .
the solid curve is the difference between equation ( [ eq : weighted_mean_limit_gamma ] ) and the true mean .
the difference is zero because @xmath101 is the weighted - mean formula for poisson - distributed data .
02cap [ fig - reduced_chi_square ] reduced chi - square @xmath168 as a function of true poisson mean , @xmath23 , for 4 @xmath7 statistics : pearson s @xmath7 @xmath169 $ ] , the modified neyman s
@xmath7 @xmath170 $ ] , the new @xmath12 statistic @xmath171 $ ] , and the maximum likelihood ratio statistic for poisson distributions @xmath172 $ ] .
the poisson distributions of figure [ fig - weighted_mean_modified_neyman ] were analyzed to produce this plot .
the formula for the curve connecting the values for modified neyman s @xmath7 statistic ( @xmath98 ) is given in equation ( [ eq : reduced_x2n ] ) .
the formula for the curve connecting the values for new @xmath12 statistic is given in equation ( [ eq : reduced_x2 g ] ) .
the dotted line shows the ideal value of one . 03cap
[ fig - variance_reduced_chi_square ] the variance of the reduced chi - square @xmath173 as a function of true poisson mean , @xmath23 , for 5 @xmath7 statistics : the standard @xmath7 , pearson s @xmath7 @xmath174 , the modified neyman
s @xmath7 @xmath175 , the new @xmath12 statistic , and the maximum likelihood ratio statistic for poisson distributions @xmath176 .
the poisson distributions of figure [ fig - weighted_mean_modified_neyman ] were analyzed to produce this plot .
the formula for the variance of the reduced pearson s @xmath7 statistic is @xmath177 .
the dotted line shows the ideal value of two . 04cap
[ fig - spectra ] the dashed lines show 4 ideal x - ray power - law spectra with a total of 25 , 100 , 1000 , and 10000 photons .
four simulated x - ray spectra with totals of 28 , 101 , 1015 , and 9938 photons are shown with 1@xmath119 error bars estimated with equations ( 9 ) and ( 14 ) of gehrels ( @xcite ) .
( n.b . some errorbars overlap and the bottom two spectra have identical data values at the 0.47 and 0.67 kev bins . ) 05cap [ fig - powell_2fp ] results of powell s method with two free parameters ( @xmath122 ) for three statistics : @xmath12 ( circles ) , @xmath98 ( squares ) , and @xmath107 ( triangles ) .
this figure uses the data given in table [ tbl - powell_2fp ] .
the dotted lines show the ideal ratio value of one .
the dashed curve in the bottom panel shows the function @xmath134 which is a good model for the @xmath121 results of the @xmath98 statistic for all spectra with @xmath178 photons .
06cap [ fig - marquardt_2fp ] results of the levenberg - marquardt method with two free parameters ( @xmath122 ) for three statistics : @xmath12 ( circles ) , @xmath98 ( squares ) , and @xmath107 ( triangles ) .
this figure uses the data given in table [ tbl - marquardt_2fp ] .
the dotted lines show the ideal ratio value of one .
the dashed curve in the bottom panel shows the function @xmath134 which is a good model for the @xmath121 results of the @xmath98 statistic for all spectra with @xmath178 photons .
07cap [ fig - powell_1fp ] results of powell s method with one free parameter ( @xmath42 ) for three statistics : @xmath12 ( circles ) , @xmath98 ( squares ) , and @xmath107 ( triangles ) .
this figure uses the data given in table [ tbl - powell_1fp ] .
the dotted lines show the ideal ratio value of one . 08cap [ fig - marquardt_1fp ] results of the levenberg - marquardt method with one free parameter ( @xmath42 ) for three statistics : @xmath12 ( circles ) , @xmath98 ( squares ) , and @xmath107 ( triangles ) . this figure uses the data given in table [ tbl - marquardt_1fp ] .
the dotted lines show the ideal ratio value of one .
09cap [ fig - marquardt_powell_1fp ] a comparison of the results of the analysis of the simulated x - ray spectra with theoretical totals of 100 , 50 , and 25 photons using the levenberg - marquardt method with 1 free parameter ( left ) and powell s method with 1 free parameter ( right ) . note that the histograms for the @xmath12 and @xmath98 are nearly identical for both methods .
the statistical analysis of this data is presented in tables [ tbl - marquardt_1fp ] and [ tbl - powell_1fp ] .
10cap [ fig - error_estimates ] error analysis of the levenberg - marquardt method results using the @xmath12 statistic with one free parameter .
the thick curve in each panel shows the cumulative distribution of the best - fit estimates of the slope @xmath42 .
the right ( left ) thin curve in each panel shows the cumulative distribution of @xmath42 plus ( minus ) @xmath179 which is the error estimate of the best - fit slope value .
the statistical analysis of this data is presented in table [ tbl - marquardt_1fp ] .
11cap [ fig - spectra_fits ] the simulated x - ray spectra of fig .
[ fig - spectra ] now plotted with @xmath12 fits .
the best fits are shown with solid curves .
the upper and lower @xmath3@xmath180 slope estimates are shown with long dashed curves .
12cap [ fig - x2l ] a comparison of the results of the levenberg - marquardt method with 1 free parameter ( left ) and powell s method with 1 free parameter ( right ) for three statistics : @xmath12 , @xmath108 , cash s @xmath155 .
the statistical analysis of this data is presented in tables [ tbl - powell_1fp ] , [ tbl - powell_1fp_x2l ] , and [ tbl - marquardt_1fp_x2l ] . | applying the standard weighted mean formula , @xmath0 / \left[\sum_i { \sigma^{-2}_i}\right ] $ ] , to determine the weighted mean of data , @xmath1 , drawn from a poisson distribution , will , on average , underestimate the true mean by @xmath2@xmath3 for all true mean values larger than @xmath2@xmath4 when the common assumption is made that the error of the @xmath5th observation is @xmath6 .
this small , but statistically significant offset , explains the long - known observation that chi - square minimization techniques which use the modified neyman s @xmath7 statistic , @xmath8 , to compare poisson - distributed data with model values , @xmath9 , will typically predict a total number of counts that underestimates the true total by about @xmath3 count per bin .
based on my finding that the weighted mean of data drawn from a poisson distribution can be determined using the formula @xmath10\left(n_i+1\right)^{-1 } \right ] / \left [ \sum_i \left(n_i+1\right)^{-1 } \right ] $ ] , i propose that a new @xmath7 statistic , @xmath11 ^ 2 / \left [ n_i + 1 \right]$ ] , should always be used to analyze poisson - distributed data in preference to the modified neyman s @xmath7 statistic .
i demonstrate the power and usefulness of @xmath12 minimization by using two statistical fitting techniques and five @xmath7 statistics to analyze simulated x - ray power - law 15-channel spectra with large and small counts per bin .
i show that @xmath12 minimization with the levenberg - marquardt or powell s method can produce excellent results ( mean slope errors @xmath13@xmath4% ) with spectra having as few as 25 total counts .
# 1#1 # 1 # 1 |
we show that compact non - singular axisymmetric equilibrium magnetic field configurations are attractors ( more precisely intermediate asymptotics ) in a weakly dissipative plasma .
this means that some initial magnetic field configurations evolve into solitary magnetic bubbles .
because of the attractor property , such free - floating solitary magnetic bubbles should exist in nature ( see braithwaite 2010 and references therein ) .
first consider magnetic equilibria of ideally conducting incompressible plasmas .
the magnetic field @xmath0 will be in equilibrium if the magnetic @xmath1 force is balanced by the pressure gradient : @xmath2 it turns out , that due to topological obstacles ( arnold 1986 ) , generic magnetic equilibria , that is generic solutions of eq.([equi ] ) , contain singular current layers ( gruzinov 2009 ) .
if the plasma electric conductivity is finite , but still very high , the magnetic diffusivity will disperse the singular current layers by the reconnection process , without dispersing the entire magnetic field configuration .
this is the intermediate - asymptotic stage we are interested in . during the intermediate - asymptotic stage
, the magnetic equilibrium configuration should be non - singular .
ideally , we want to find all compact non - singular magnetic equilibria .
we think we have a full solution of this problem in two dimensions ( 3 ) . in 2d ,
the only compact non - singular equilibrium is a circular - symmetric magnetic bubble . in 3d
, there is a class of non - singular equilibrium configurations compact axisymmetric magnetic bubbles ( 4 ) . the non - singular axisymmetric equilibrium bubbles are attractors ( 5 ) .
but unlike the 2d case , we do not know if these bubbles are the only non - singular compact attractors .
we study magnetic equilibria using the relaxation method ( e.g. gruzinov 2009 ) .
we put an arbitrary compact magnetic field into high - viscosity small - magnetic - diffusivity plasma , and let the plasma move under the resulting magnetic stress : @xmath3 the first equation ( with @xmath4 ) says that the magnetic field @xmath0 is frozen into the fluid moving with velocity @xmath5 .
non - zero @xmath6 describes magnetic diffusivity , @xmath7 .
the second equation says that the fluid incompressibly yields to the magnetic force .
the equations are solved numerically , using the fast fourier transformation .
needless to say , the creeping flow evolution is not what happens in nature .
but the final non - singular equilibria that we find should be real .
the idea is that the route to the equilibrium is irrelevant , we just find generic non - singular equilibria .
consider 2d magnetic fields , that is magnetic fields of the form @xmath8 take a generic initial magnetic stream function @xmath9 vanishing at spatial infinity and let it evolve according to eq.([relax ] ) with zero magnetic diffusivity , @xmath4 .
then each saddle of @xmath9 develops into a singular current layer ( gruzinov 2009 ) .
if we want to have a non - singular equilibrium , we must start with the initial @xmath9 without saddles .
then the isolines of @xmath9 are nested closed curves .
the ideal relaxation of magnetic energy simply turns all the nested curves into concentric circles , giving the final state @xmath10 where @xmath11 is the distance from some center .
now consider slightly non - ideal relaxation , with magnetic diffusivity @xmath12 but still very small .
numerical experiments ( some 300 different runs ) show that the generic initial state with compact support evolves as follows : * sharp current layers form at saddle points * reconnection occurs at current layers * the field splits into a number of circular magnetic bubbles .
* the bubbles have monotonic @xmath13 , with a maximum or minimum at @xmath14 and vanishing @xmath9 at large @xmath11 .
* bubbles of the same polarity ( same sign of @xmath9 ) merge , bubbles of the opposite polarity repel each other .
* ultimately , two bubbles form , one positive and one negative , and the distance between the two final bubbles is gradually increasing .
this is illustrated in fig.1 . at high numerical resolution ,
when the small magnetic diffusivity can be achieved , the different stages of the relaxation process are easily distinguished . at first
, ideal relaxation quickly moves the field into the minimum - energy - for - a - given - topology state .
then reconnection occurs , on the time scale which is much longer than the ideal relaxation time .
after that , an almost steady non - singular equilibrium state is established .
the 2d summary is : ( i ) generic equilibrium is singular , ( ii ) the only non - singular compact equilibrium configurations are circular bubbles , ( iii ) for arbitrary compact initial field , circular solitary magnetic bubbles are the only attractors ( more precisely , intermediate asymptotics of the weakly dissipative mhd ) .
this behavior formation of two infinitely separated circular bubbles seems too simple to be true .
but in fact its origin is quite clear . in 2d
, there are infinitely many robust invariants , which are constant even when the magnetic energy can decrease due to magnetic diffusivity . from eq.([relax ] ) , @xmath15 this is very small , because the rate of dissipative magnetic energy damping is given by the integral of @xmath16 .
the final state is simply the state with minimal energy for the given area per each interval @xmath17 .
this minimum is given by the state with two monotonic circular bubbles , one positive and one negative .
now we want to find the weakly - dissipative 3d attractors . unlike the 2d case , we can not claim a full understanding of this problem .
this is because the runs take longer , and the magnetic fields are harder to visualize in 3d .
also , in 2d , we know that all saddles become singular current layers under the ideal relaxation . in 3d ,
generic ideal equilibria are also singular , with infinitely many singular current layers ( gruzinov 2009 ) , but it remains unclear where the current layers form .
however one thing seems to emerge from our numerical experiments with the weakly dissipative magnetic relaxation in 3d ( about 100 different runs ) .
there exist non - singular axisymmetric compact magnetic configurations , which are attractors . in this section
we describe the non - singular axisymmetric equilibrium bubbles , and in 5 we demonstrate their attractor and stability properties . in cylindrical coordinates @xmath18 , the axisymmetric magnetic field can be written as follows @xmath19 where @xmath20 , @xmath21 .
then the current is @xmath22 where @xmath23 .
the equilibrium equation ( [ equi ] ) then reads @xmath24 @xmath25 @xmath26 from eq.([aequi3 ] ) , @xmath27 . then from eqs.([aequi1 ] , [ aequi2 ] ) , @xmath28 , and we get the grad - shafranov equation @xmath29 where @xmath30
. the invariant functionals @xmath34 and @xmath35 , being linear , can be replaced , without loss of generality , by some invariant functions of a single variable , say @xmath36 $ ] , @xmath37 $ ] the volume and the toroidal flux distributions functions .
thus , we see that the axisymmetric magnetic equilibria are parametrized by two functions of a single variable . in the 2d case , we only have one invariant function @xmath34 .
the final state is also a function of a single variable , @xmath13 .
thus , it is not surprising that the 2d non - singular equilibria are all circular symmetric . in the 3d axisymmetric case
, we have two constraints , and therefore the shape of the non - singular solution can not be universal .
if one keeps @xmath34 fixed , and changes @xmath35 , the tori of constant @xmath9 will incompressibly deform .
the grad - shafranov equation ( [ gs ] ) can be easily solved numerically ( and even analyzed theoretically for some @xmath38 and @xmath39 ) .
perhaps the simplest non - trivial case ( coming from @xmath40 , @xmath41 ) is @xmath42 here @xmath43 is the potential of the shrodinger equation , and @xmath44 is the energy level .
the potential is confining in the @xmath11 direction .
confinement along @xmath45 is self - generated by positive @xmath9 .
it is clear that non - singular solutions exist , at least for @xmath44 slightly smaller than the ground state energy of the radial part of the hamiltonian ( the operator -@xmath46 ) . to confirm the non - singularity , and for the stability studies of 5 ,
we have solved eq.([gsex ] ) numerically , using the relaxation method @xmath47 where the time step @xmath48 should be small enough , and the normalization factor @xmath49 at each time step is chosen so as to keep the maximum value of @xmath9 at a pre - fixed value .
the value of @xmath44 is given by the asymptotic value of @xmath49 , @xmath50 .
the results are shown in fig.(2 ) . in summary , a family of non - singular axisymmetric equilibria parametrized by two functions of a single variable exists .
are the axisymmetric non - singular bubbles stable in 3d ?
are these bubbles attractors of the weakly dissipative mhd ?
stability of plasma in magnetic fields is the most elaborate part of plasma physics ( kadomtsev 1966 ) .
the development of the mhd stability theory was motivated by the practical problem of confining the nuclear burning plasma by external magnetic fields . in astrophysics ,
the mhd stability theory has found another application . instead of asking which external magnetic configurations
will stably confine the plasma , one asks what magnetic configurations can be stably supported by the plasma .
what are the possible relic magnetic fields in non - convective stars , or what kinds of magnetic relics can exist in the ism ?
this problem ( formulated by chandrasekhar and fermi 1953 ) is still unsolved .
one can show that certain magnetic field configurations are unstable ( tayler 1982 and references therein ) .
but the proof of stability is more difficult .
one needs to show that the energy functional is non - decreasing under an _ arbitrary _ virtual displacement , and this turned out to be hard ( although gruzinov 2008 proves stability for some artificial magnetic configuration ) .
a possible way to check the stability of a given magnetic field numerically is by the ideal ( meaning @xmath51 ) relaxation method ( [ relax ] ) .
if the equilibrium is the energy saddle or maximum , rather than minimum , the relaxation will destroy it .
we checked that the relaxation scheme does catch the classical kink and screw instabilities ( for definitions , see kadomtsev 1966 ) when we take the equilibrium bubble ( [ gsex ] ) with @xmath52 ( shown in fig.2 ) , add a random weak magnetic field ( to seed the possible instabilities ) , and subject it to the 3d relaxation ( [ relax ] ) , we see just the relaxation of the random component ( magnetic energy decreases a bit ) .
we also see that the noise slightly tilts the bubble ( fig.3 ) .
of course we can not conclude that the bubble is stable , especially given the danger of resonant surfaces where the field lines are closed .
but the instabilities , if any , should be mild and globally non - destructive . given that the non - singular axisymmetric bubbles are stable , it is tempting to suggest that they behave just like the circular 2d bubbles .
this is true , except for one major difference .
recall that the opposite polarity 2d bubbles repel , and the same polarity bubbles merge . in 3d ,
the role of polarity is played by the magnetic helicity @xmath53 , @xmath54 .
but unlike the 2d case , the 3d bubbles always seem to ultimately merge , regardless of the signs of helicity and axis orientations .
the difference between the 2d and the 3d cases must be due to the fact that in 2d we have infinitely many robust invariants @xmath34 .
in 3d , @xmath34 and @xmath35 are invariant ( and even defined ) only under axisymmetry .
the only generic robust invariant in 3d is , apparently , the helicity .
then , if the total helicity is zero , the magnetic field has no reason to stay , and indeed it does seem to decay to zero . more work is needed here .
we can only show what we have seen : * fig.4 . just like the 2d non - circular bubbles
relax into circular bubbles , in 3d , the non - axisymmetric bubbles relax into axisymmetric .
* fig.5 . just like
the 2d bubbles of the same polarity merge into a single circular bubble , in 3d , bubbles of the same helicity merge into a single axisymmetric bubble , conserving helicity but decreasing energy .
. shows the merger of the bubbles with the same helicity and with parallel axes . *
the merger of the bubbles with the same helicity and with anti - parallel axes . * fig.7 .
the 2d bubbles of the opposite polarity repel . in 3d
, the opposite helicity bubbles still merge .
if the total helicity is zero , apparently a total destruction of the magnetic field occurs .
fig.7 shows annihilation of the opposite helicity bubbles with anti - parallel axes .
we have shown ( tentatively , numerically ) that non - singular axisymmetric equilibria are weakly dissipative attractors .
the unsolved problems are : | stability and attractor property of free - floating axisymmetric magnetic bubbles in high - conductivity plasmas is ( tentatively , numerically ) demonstrated .
the existence of compact non - singular axisymmetric magnetic equilibria is proved .
being attractors , the solitary magnetic bubbles should exist in nature . |
multiple analyses of observed gamma rays @xcite provide growing evidence for the existence of inter - galactic magnetic fields ( for reviews see @xcite ) .
the existence of such magnetic fields poses new questions for cosmology and probably also for particle physics @xcite .
in addition , a primordial magnetic field can play an important role during structure formation in the universe and could help us understand the ubiquity of magnetic fields in astrophysical bodies .
a critical challenge at this stage is to sharpen observational techniques so that we can better observe and measure inter - galactic magnetic fields . of the various probes of inter - galactic magnetic fields ,
blazar - induced gamma ray cascades hold certain key advantages .
the gamma ray cascades originate in the voids in the large - scale structure and are mostly immune to complications of a noisy environment .
the cascade develops in a relatively small spatial volume and hence is a local probe of the magnetic field in the voids .
this is distinct from other methods , such as the faraday rotation of the cosmic microwave background polarization , that probe an integrated measure of the magnetic field .
gamma ray cascades are also highly sensitive probes and can trace very weak cosmological magnetic fields .
in this paper we study the effect of stochastic inter - galactic magnetic fields on blazar induced gamma ray halos and some results overlap with those of refs .
the cascade process is complicated and all analyses use some simplifying assumptions .
for example , the analysis in ref .
@xcite only considered non - stochastic magnetic field configurations .
other simplification schemes , such as the `` large spherical observer '' method employed in ref .
@xcite , transport arrival directions of gamma rays for distant observers to a single earth - bound observer .
this technique is certainly useful to study spectral properties of the cascade , but there is a danger that it loses or shuffles the spatial information of gamma ray arrival directions that is crucial for morphological studies .
our focus is on the effect of stochastic magnetic fields that are statistically isotropic and with or without helicity .
so we carefully analyse the spatial information of the gamma rays that is useful for deducing properties of the magnetic field but , for the present , we only include an approximate description of the cascade development .
an important helpful concept that we develop in this paper is that of the `` pp surface '' ( see sec .
[ sec : morph ] ) .
this spatial surface holds the key to halo morphology and many of the features that we see in our simulations can be understood in terms of the shape of the pp surface and its intersection with the blazar jet .
we have applied a refined version of the q - statistics first proposed in ref .
@xcite to study the morphology of halos .
our results show that this statistic can successfully extract the helicity of the magnetic field .
our simulations also reveal that the plot of @xmath0 , where @xmath1 is a variable that will be explained below , has an additional bump .
we are able to show that this bump is a genuine feature of the q - statistic and explain it in terms of properties of the pp surface in sec .
[ sec : bump ] .
thus this extra feature of @xmath0 may become an observational tool in future .
we give some background information in sec .
[ background ] , discuss our simulation techniques in sec .
[ sec : morph ] , discuss features of the halo in sec .
[ halofeatures ] , present results with non - stochastic fields in sec .
[ qanalytical ] and with stochastic fields in sec .
[ qstochastic ] .
as mentioned above , we discuss the bump feature in @xmath0 in sec .
[ sec : bump ] .
we summarize our conclusions in sec .
[ conclusions ] .
our stochastic magnetic field generation scheme is described in appendix [ bgeneration ] .
tev photons from blazars induced electromagnetic cascades through pair production with the extragalactic background light , @xmath2 in the presence of a magnetic field , the charged leptons follow spiral paths as they propagate and lose energy due to inverse compton scattering with the cosmic microwave background ( cmb ) photons .
the up - scattered cmb photons have gamma ray energies and produce extended halos around the direction of the blazar . in this section
we briefly discuss the formation of the halo under simplifying assumptions .
consider a blazar located at the origin of our coordinate system described by the unit basis vectors @xmath3 .
we choose @xmath4 so that earth is located at @xmath5 where @xmath6 is the comoving distance to the source , d_s=_0^z_sdz . to perform the integral ,
we have used @xmath7 , @xmath8 , @xmath9 as found in ref .
@xcite and we have also assumed that @xmath10 and used natural units so that @xmath11 .
for all the simulations in this paper we will choose @xmath12 .
the blazar will typically emit photons in a collimated jet which we approximate to be a conical region with half - opening angle @xmath13 .
the energy @xmath14 of these photons must lie above some threshold of about a tev if they are to produce an electron - positron pair from interaction with the extragalactic background light ( ebl ) .
due to the opacity of the ebl , the tev photons will travel a mean free path ( mfp ) determined by the pair production cross section @xmath15 and the number density of the ebl photons @xmath16 , d_0= _
^-1(80 ) ( ) , we have assumed that @xmath17 to approximate the mfp in the final equality @xcite .
following @xcite we will set @xmath18 as this dimensionless constant is estimated to lie in the range of @xmath19 .
the comoving distance from the source to the pair production event is given by @xmath20 .
the redshift of the produced lepton pairs will depend on the relative position of the leptons to the source .
since @xmath21 , we make the approximation @xmath22 and we can write , [ tevcomfp ] d^c_0(80 ) ( )
. the energy of each of the produced leptons will be @xmath23 .
these leptons are expected to travel a distance @xmath24 before losing most of their energy through inverse compton ( ic ) cooling which occurs by upscattering cmb photons .
the cooling distance is d_e= ( 31 ) ( ) ( ) ^4 [ eq : de ] where @xmath25 is the thomson scattering cross section and @xmath26 is the cmb energy density .
note that we can assume the whole cascade development happens around redshift @xmath27 as @xmath28 . at that redshift ,
the average energy of a cmb photon is e_cmb ( 610 ^ -4)(1+z _ ) , which implies that , from energy conservation , the upscattered photons will have energy [ igamogamenergy ] e_=e_cmb(77 ) ( ) ^2 . as the lepton propagates , it upscatters @xmath29 photons , and produces a gamma ray cascade in the 1 - 100 gev range if the initial gamma ray had an energy of a few tev
. clearly not every photon upscattered by the leptons will reach earth .
those that do must come from a set of events that satisfy a set of three constraints given in ref .
@xcite that we now describe . after pair production
, the lepton s initial velocity will be almost parallel to the momentum of the parent photon with a negligible deviation of order the inverse lorentz boost factor @xmath30 .
their subsequent trajectory will be determined by the magnetic field @xmath31 through the lorentz force .
if the magnetic field is incoherent on length scales smaller than the cooling distance @xmath32 , the lepton trajectories will be diffusive and this situation is much harder to analyze .
so we focus on magnetic fields that are coherent on scales that are much larger than @xmath24 .
then the lepton trajectories are bent in an effectively constant magnetic field and follow a helical trajectory with gyroradius r_l = r_l0 |_| , r_l0= , which depends on the lepton s perpendicular velocity to @xmath33 , _
i.e. _ @xmath34 .
the quantity @xmath35 is useful as it denotes the distance the lepton must travel in order to perform a full revolution .
the value of @xmath36 is a function of the redshift as it depends on @xmath37 . for magnetic fields frozen in the plasma , the field strength redshifts as @xmath38 where @xmath39 is the magnetic field magnitude today . with r_l0 3.5 ( ) ( ) ^-1()^-2 , we can evaluate the ratio 0.0106 ( ) ^-1()()^-2 [ bending ] which determines the angular deflection of the leptons . before scattering off an ebl photon and pair producing leptons .
the lepton trajectories are bent due to the local magnetic field over a very short distance compared to the distance to the source , @xmath6 , and are shown in the insets .
( the insets show huge bending whereas we have only considered magnetic fields that give small bending . )
inverse - compton scattering of a lepton and cmb photons results in a cascade of gev energy gamma rays arriving at earth from the direction of the pair production .
[ sketch taken from @xcite.],scaledwidth=50.0% ] as depicted in fig .
[ halocreation ] , let us introduce the following angles : @xmath40 is the arrival angle of the gev photon with respect to the source location , @xmath41 represents the angle between the upscattered photon and the tev photon , @xmath42 is the angle subtended by the tev photon s momentum , @xmath43 the vector from the source to earth , and finally @xmath44 is the azimuthal angle in which the whole ( planar ) scattering process take place .
we also introduce the polar vectors @xmath45 and @xmath46 , in the @xmath47 plane .
it is important to emphasize that the whole process occurs in a plane to a good approximation because @xmath48 and the length of the lepton trajectory can be ignored .
then there are only 3 points that are relevant ( the source , the pair production point , and the observer ) and they always lie in a plane .
applying the sine formula to the triangle in fig .
[ halocreation ] we get our first constraint [ const1 ] d_s ( ) = d_0 ( ) . where @xmath49 is the distance traveled by the tev photon and is a random variable drawn from a distribution that depends on the mfp @xmath50 in eq .
( [ tevcomfp ] ) .
this is discussed in detail in sec .
[ sec : morph ] . the bending angle @xmath41 is related to the distance traveled by the lepton through the local magnetic field which we write as @xmath51 .
we also decompose the lepton s initial velocity at time @xmath52 , @xmath53 , where @xmath54 . at some later time @xmath55
the velocity is [ leptonvel ] ( t)=v_||_||+v_(t)_v_(t)(__|| ) , here we introduced the angular frequency of the orbital motion @xmath56 and the @xmath57 sign refers to the positron ( electron ) trajectory .
a cmb photon upscattered at time @xmath58 will be directed along the lepton s trajectory and so the deflection angle of figure [ halocreation ] can be expressed as @xmath59 .
( [ leptonvel ] ) we can derive the second constraint , [ const2 ] 1-()=(1-((0))^2)(1-(t_ic / r_l0 ) ) .
the time of inverse compton scattering @xmath58 is a stochastic variable .
given its value and the magnetic field direction , the constraints determine the bending angle , @xmath41 .
a single propagating lepton will be able to upscatter cmb photons towards earth only at certain times when the lepton s momentum is directed towards earth .
photons upscattered at other times will not reach earth and we can safely ignore them .
the number of photons upscattered by a lepton is very large ( @xmath60 ) , with mean deviation angles between the photons @xmath61 ( see eq .
( [ bending ] ) ) .
this angle is large enough that we only expect @xmath62 of the cascade photons from any lepton to reach earth .
this allows us to adopt the strategy that we first select a value of @xmath58 from an exponential probability distribution as described in sec .
[ sec : morph ] and then solve the constraint equations to find all tev gamma rays from the blazar that upscatter cmb photons that reach earth . for different values of @xmath58 , different tev gamma rays from the blazar will lead to observed photons . in this way
, we will be able to track the photons that arrive on earth and not waste computational effort on those that go elsewhere .
the third and final constraint is that the cascade gamma ray lies in the plane specified by @xmath63 .
this requires that the lorentz force in the azimuthal @xmath63 direction vanishes between the time of pair production and ic scattering .
namely the @xmath44 component of the impulse must vanish , j_== ( e_0^t_ic dt ( t ) ) = 0 . the impulse can be simplified by pulling out the assumed constant magnetic field of the integral and defining _ = _ 0^t_ic dt ( t ) .
the geometrical setup of fig .
[ halocreation ] forces @xmath64 to bisect the angle @xmath41 and therefore its unit vector can be written as _ =
( /2- ) -(/2- ) . decomposing @xmath33 as = b_+b_+b_z allows us to write [ const3 ] _
= -b_(/2-)-b_z(/2-)=0 . to summarize this section , eqs .
( [ const1 ] ) , ( [ const2 ] ) and ( [ const3 ] ) are the constraints that need to be satisfied by the variables @xmath65 given a magnetic field realization and initial velocity of the tev gamma ray ( both of which depend on @xmath65 ) , the source - observer distance ( @xmath6 ) , the distance to pair production ( @xmath49 ) , and the photon upscattering time ( @xmath58 ) .
for events that satisfy the constraints in eqs .
( [ const1 ] ) , ( [ const2 ] ) and ( [ const3 ] ) , an observer on earth will receive flux at a polar angle of @xmath40 from the line of sight ( los ) to the blazar and at an azimuthal angle @xmath44 .
it is possible to solve for these angles analytically in certain regimes for specific magnetic fields @xmath66 while using the mfps of the particles : @xmath67 as given by eq .
( [ eq : de ] ) and @xmath50 as given in eq .
( [ tevcomfp ] ) @xcite
. we will refer to these values as `` canonical '' propagating distances ( pds ) , in contrast to the `` stochastic '' pds we will use later .
numerical methods are required when considering more general @xmath68 and when including the stochasticity in the pds of the initial gamma ray and pair produced leptons .
note that @xmath24 , @xmath50 depend on the energy @xmath69 of the final gamma ray reaching earth .
therefore , in the analysis with canonical pds , once we specify @xmath68 we can expect to find a 1 dimensional curve @xmath70 , @xmath71 by solving the constraints .
the analysis with the canonical pds is helpful because it gives an idea of the overall shape that the extended emission will ressemble even when we include stochasticity .
for the case of stochastic pd the constraint equations are solved with @xmath72 drawn from the probability density function p[d_e ] = e^-d_e / d_e [ cticdist ] and similarly the distance @xmath49 traversed by a tev gamma ray is drawn from p[d_0 ] = e^-d_0/d^c_0 .
[ xdist ] we will solve the constraint equations in a variety of magnetic field backgrounds , starting with simple analytic configurations for illustration purposes , and then move on to the more realistic case of stochastic , isotropic magnetic fields . our procedure to generate stochastic , isotropic magnetic fields is described in appendix [ bgeneration ] . turning to the energy distribution of gamma rays in our simulations , we assume a power law for the blazar spectrum @xcite and follow ref .
@xcite by choosing a spectral index of @xmath73 that is characteristic of the tev sources .
this yields a spectrum given by , ~()^-2.5 .
we can use the relation between @xmath14 and @xmath69 to turn the blazar spectrum into one for the observed photons ~()^-1.75 .
[ dnde ] our simulation algorithm proceeds in the following steps .
first we draw an observed energy , @xmath69 , from the distribution in eq .
( [ dnde ] ) .
the drawn value of @xmath69 determines the energy of the high energy photon , @xmath14 , from eq .
( [ igamogamenergy ] ) . with the values of @xmath69 and @xmath14
we can find @xmath50 and @xmath24 from eqs .
( [ tevcomfp ] ) and ( [ eq : de ] ) , then use eqs .
( [ cticdist ] ) and ( [ xdist ] ) to draw values of @xmath74 and @xmath49 .
now we try to solve the constraint equations ( [ const1 ] ) , ( [ const2 ] ) and ( [ const3 ] ) .
if a solution does not exist we go back to the first step and draw an observed energy @xmath69 .
if a solution does exist , it gives the coordinates @xmath65 of an event on earth at energy @xmath69 that satisfies all the constraints .
we repeat the entire process until 1000 solutions are found . for a jet source ,
we only retain the events whose initial tev photon lie within the jet .
these small number of events give us the observed halo that will be shown in our plots .
+ to start we consider a source that radiates tev photons isotropically in two different magnetic field backgrounds .
the first background , [ constantb ] * b * = b_0(()- ( ) ) , with @xmath75 , is a uniform magnetic field pointing at an angle @xmath76 from the line of sight .
the second background is a maximally helical field [ helb ] * b * = b_0((2x/ ) + ( 2x/ ) ) .
here @xmath77 is the coherence length of the helical field .
we will take @xmath78 , @xmath79 and @xmath80 as the prototypical values and eventually vary them one at a time to see their effect on the halo morphology .
next we solve the constraint equations with the canonical pds and determine the arrival directions @xmath40 , @xmath44 for several different energies , @xmath69 , for an isotropically emitting source . with canonical pds , the events lie on a one dimensional curve shown in fig .
[ haloexample ] in black .
the corresponding simulations with stochastic pds results in the distribution of points in fig .
[ haloexample ] .
the points at lower energies are located further away from the source direction as is to be expected since the tev gamma ray travels further at low energies before pair producing leptons ( eq . ( [ tevcomfp ] ) ) and
leptons at low energies bend more before up - scattering cmb photons ( eq . ( [ bending ] ) ) .
looking closely at fig .
[ haloexample ] , the drawn points are triangular ; upright triangles are gamma rays that originate from electrons and inverted triangles are those that originate from positrons .
the distinction is made clearer in fig .
[ haloexample2 ] where red ( black ) points originate from positron ( electron ) processes .
if the source was taken to be a jet , gamma rays predominantly from one of the two leptons will be observed . ) and ( [ helb ] ) @xcite .
the colors denote the energy of the observed gamma ray.,title="fig:",scaledwidth=45.0% ] ) and ( [ helb ] ) @xcite .
the colors denote the energy of the observed gamma ray.,title="fig:",scaledwidth=45.0% ] but now red ( black ) points originate from inverse compton scattering due to positrons ( electrons ) .
if the source was taken to emit along a jet , most of the observed gamma rays would originate from either positron or electron processes but not both.,title="fig:",scaledwidth=45.0% ] but now red ( black ) points originate from inverse compton scattering due to positrons ( electrons ) .
if the source was taken to emit along a jet , most of the observed gamma rays would originate from either positron or electron processes but not both.,title="fig:",scaledwidth=45.0% ] the constraint equations are quite complicated to solve but there is a helpful visualization .
first consider the third constraint equation , eq .
( [ const3 ] ) , and note that @xmath81 and @xmath82 are also functions of @xmath40 , @xmath41 and @xmath44 .
so eq . ( [ const3 ] ) provides one functional relation between these variables that only depends on the magnetic field background .
hence the magnetic field defines a two - dimensional surface in space .
we will call this the `` pair production surface '' or the `` pp surface '' since only lepton pair production at this surface can send gev gamma rays to the observer . in fig .
[ fluxsources ] we show the pp surface for the magnetic fields of eqs .
( [ constantb ] ) and ( [ helb ] ) . on these plots
we also show the pair production locations , `` pp locations '' , that resulted in the halos of fig .
[ haloexample ] .
note that a gamma ray from the source will propagate a certain distance , @xmath49 and then pair produce .
so the pair production points also lie on a sphere of radius @xmath49 .
this is partly enforced by the law of sines in eq .
( [ const1 ] ) , which gives a relation between @xmath41 and @xmath40 .
the intersection of this sphere and the pp surface define a one - dimensional curve in space ; cmb photons that are inverse compton scattered along the one - dimensional curve can propagate to earth .
however , not all points on this one - dimensional curve will satisfy the final constraint .
namely , eq . ( [ const2 ] ) , picks out a limited set of points on the one - dimensional curve and these give the trajectories of the gamma rays that are observed .
the pp surface can be found analytically for simple cases .
for instance , the constraint in eq .
( [ const3 ] ) with the helical magnetic field from eq .
( [ helb ] ) , which has @xmath83 , reduces to , b_(/2-)=0 , with b_==(2z/+ ) .
as @xmath84 has only one solution at @xmath85 in the physical range @xmath86 $ ] , @xmath87 $ ] , the surface is mainly determined by @xmath88 which translates to , = - .
[ analyticpp ] this equation describes a spiral structure as seen in figure [ fluxsources ] .
until now , we have been assuming that the source emits photons isotropically .
below , we will also consider the case when the source emits photon in a collimated jet . in that case , there is a fourth constraint restricting the relevant part of the pp surface to where it intersects the jet , and it is quite possible that there is no solution .
we ignore such cases as they are observationally irrelevant .
in following figures we will show pp locations , even if they do not lie within the jet . only those pp locations that lie within the jet
will lead to observed gamma rays . )
( top ) and the maximally helical magnetic field of eq .
( [ helb ] ) ( bottom ) .
the source is located at the orange star ; the observer is at @xmath89 .
the blue points are the events that give rise to the halos shown in figs .
[ haloexample ] . , title="fig:",scaledwidth=45.0% ] )
( top ) and the maximally helical magnetic field of eq .
( [ helb ] ) ( bottom ) .
the source is located at the orange star ; the observer is at @xmath89 .
the blue points are the events that give rise to the halos shown in figs .
[ haloexample ] . ,
title="fig:",scaledwidth=45.0% ] fig .
[ fluxsourcesjet ] presents an example in which the source has a jet with half - opening angle @xmath90 and the magnetic field is given by eq .
( [ helb ] ) .
the jet direction is chosen so that the earth lies within the cone of the jet and the blazar can be seen directly .
the top plots in both figures shows that the jet picks out a small region of the pp surface , and the bottom plots show the resulting halo . ) with @xmath91 g , @xmath80.,title="fig:",scaledwidth=45.0% ] ) with @xmath91 g , @xmath80.,title="fig:",scaledwidth=45.0% ] ) and with the same setup as in fig .
[ fluxsourcesjet ] but with @xmath39 reduced to @xmath92 . compared to fig .
[ fluxsourcesjet ] , we see that the high energy gamma rays ( blue points ) are more clustered and so the halo size is smaller at fixed energy .
however , the reduced field strength also causes lower energy gamma rays to come into the field of view , and the overall halo size does not change due to a reduced @xmath39 .
, title="fig:",scaledwidth=45.0% ] ) and with the same setup as in fig .
[ fluxsourcesjet ] but with @xmath39 reduced to @xmath92 .
compared to fig .
[ fluxsourcesjet ] , we see that the high energy gamma rays ( blue points ) are more clustered and so the halo size is smaller at fixed energy .
however , the reduced field strength also causes lower energy gamma rays to come into the field of view , and the overall halo size does not change due to a reduced @xmath39 . ,
in this section we discuss the structure of the halo as the parameters @xmath39 , @xmath77 and the sign of the helicity of the maximally helical magnetic field in eq .
( [ helb ] ) are varied .
the concept of the pp surface will be a useful tool for this discussion as it allows us to clearly see how the magnetic field dictates the halo s shape .
the magnetic field strength directly affects the amount of bending of the lepton trajectories since the gyroradius @xmath93 .
therefore a weak magnetic field will require that the initial tev gamma ray is already propagating nearly towards earth . thus reducing @xmath39
will shrink the size of the halo at any given gamma ray energy , although lower energy gamma rays may now enter the field of view .
this can be seen in fig .
[ halo_b_reduced ] which was created using @xmath94 and @xmath95 .
the plot looks almost identical to fig .
[ fluxsourcesjet ] , which was created using @xmath79 , except that the extent of the halo in the @xmath96 and @xmath97 directions , for photons of the same energy , has shrunk by a factor of @xmath98 .
if one does not track the photon s energy , the effect of a change in @xmath39 is not easily seen through the morphology of the halos as their shapes and sizes are determined by the intersection of the jet and the pp surface .
we show a few example of this interplay in the sketch of fig .
[ haloshapecartoon ] .
it is then possible to learn valuable information about the inter - galactic magnetic field in the region of the pp locations by observing the halo s shape . in a real situation
, we can not observe the full halo shape as it will be contaminated by background photons coming from other sources .
however , we can still extract certain useful halo information since the background is expected to be stochastically isotropic and certainly not parity odd . as can be seen from eq .
( [ tevcomfp ] ) , the mfp @xmath50 of the tev photons are of the order of @xmath99 mpc . any magnetic field with coherence length much larger than @xmath50 will appear constant in space . on the other hand , for @xmath100 , the halo will be produced from a rapidly varying part of the pp surface and will be more scattered .
the helicity of the magnetic field in eq .
( [ helb ] ) can be flipped by changing @xmath101 .
a flip in the helicity simply leads to a parity inversion of the pp surface and the halo spiral also changes handedness .
however , to get more statistics , we will investigate both helicities using independent simulations in sec .
[ qanalytical ] .
one of the main goals of this work is to determine if the helicity of the inter - galactic magnetic field can be deduced from the shape of the blazar halos .
as we have seen , under certain conditions , a helical inter - galactic magnetic field can produce a clear spiral - like structure in a gamma ray halo .
hence it is important to develop a statistical technique that is sensitive to this structure .
a statistic , called @xmath102 , was developed in ref .
@xcite , and was applied to the _ diffuse _ gamma ray background observed by the fermi telescope in refs .
a non - zero value of @xmath102 was observed with high confidence in comparison to monte carlo simulations that assume no inter - galactic magnetic field .
one can see from the halo plots above , _
e.g. _ fig .
[ haloexample ] , that the arrival direction of high energy photons tend to be closer to the blazar line - of - sight than those for lower energy photons .
this is mainly due to the smaller magnetic bending at higher energies .
hence different locations of the pp surface are sampled by photons of different energies and the observed gamma rays can carry an imprint of any curvature or twist of the pp surface .
more precisely , the work of ref .
@xcite showed that a left ( right ) handed helical magnetic field will create left ( right ) handed spiral patterns in the observed photons .
) points out of the plane of the page as denoted by the arrow tip .
the left sketch shows a situation where @xmath103 since @xmath104 ; similarly the sketch of the blazar on the right shows a @xmath105 situation .
, scaledwidth=35.0% ] below we briefly review the idea behind a slightly modified version of the q statistic proposed in ref . @xcite .
we will apply the statistics on regions surrounding an observed blazar whose angular position will be denoted by the unit vector @xmath106 .
we consider a disk of radius @xmath1 centered on the location of the source and consider the set of photons within this disk .
these photons are binned according to their energies into non - overlapping bins @xmath107 .
we use @xmath108 to denote the number of photons in bin @xmath109 within the disk of radius @xmath1 .
we then perform the sum , @xmath110 where @xmath111 is the unit vector denoting the arrival direction of photon @xmath112 in bin @xmath113 ; @xmath114 is the unit vector obtained by projecting @xmath115 on to the xy - plane : @xmath116 .
the original q statistic in ref .
@xcite was defined without the heaviside function ( @xmath117 ) in eq .
( [ qdefn ] ) .
we find that the @xmath117 factor improves the resolution of q. without the @xmath117 factor , the q - statistic is determined by the cross product of the _ average _ arrival direction of the photons ( in two energy bins ) within a radius @xmath1 of the source . because of the electron - positron symmetry , photons tend to arrive on either side of the source ( see fig .
[ haloexample2 ] ) , and the average arrival direction tends to be near the origin . introducing the @xmath117 factor ensures that for every high energy photon selected , we only average the low energy photons that arrive on the same side with respect to the source .
then there is a larger contribution to the value of @xmath102 .
essentially the @xmath117 term limits the sum to gamma rays within the electron ( or the positron ) branch of the halo ( see fig .
[ haloexample2 ] ) .
we illustrate the @xmath118statistic in fig .
[ qidea ] . .
this simulation did not include any stochasticity in the mean free path of the tev @xmath119-ray or the pair produced leptons , as such the energy of the photons are in one - one correspondence with their propagation distance .
the surrounding magnetic field is given by eq .
( [ helb ] ) with @xmath120 and @xmath95.,scaledwidth=45.0% ] .
the curves start at @xmath121 because there are insufficient photons within that distance.,scaledwidth=48.0% ] in the presence of background gamma rays in addition to the blazar gamma rays , we expect @xmath0 to start near @xmath122 at @xmath123 , grow to a peak value near @xmath124 , where @xmath125 is the angular radius of the halo , and finally come back down towards @xmath122 at large @xmath1 where the signal becomes background dominated .
however in mock maps with no background , the value of @xmath102 should asymptotically flatten out to its maximal value attained at @xmath125 , and its value will be negative ( positive ) for right ( left ) handed magnetic fields .
we can see this behavior in fig .
[ nostoqexample ] where we simulate the halo without any stochasticity and with the magnetic field of eq .
( [ helb ] ) . to showcase the q - statistic in this paper we will separate the gamma rays in three energy bins : @xmath126 all numbers in gev .
the clean result of fig . [ nostoqexample ] will get noisy once we include stochastic magnetic fields and stochastic pds . indeed , a randomly generated magnetic field can sometimes create halos whose q - statistics suggest the wrong helicity .
it is therefore important to average over many realizations of the magnetic field and the jet orientation .
each realization will simulate a blazar with a jet of half - opening angle @xmath127 and having earth in its los .
the jet is also constrained to generate a halo with at least 3 events in order for the statistics to be applied ; this condition is generally automatically satisfied if earth is in the jet s los and the root mean squared magnetic field strength satisfies @xmath128 g .
larger values of @xmath39 create too much bending and very few , if any , upscattered photons reach earth from seen blazars .
jets pointing further away from the los might still yield observable photons but we would not be able to identify these blazars .
this bending is also the reason for the empty regions around the axis of the pp surface seen in fig .
[ fluxsourcesjet ] . the empty region gets larger for larger @xmath39
. we will consider magnetic fields of the form , [ rmfonemode_notappen ] * b*(*x*)=_*k*k * b*(*k*,f_,b_)e^i * k * with the set @xmath129 consisting of @xmath130 vectors which have magnitude @xmath131 and whose directions are approximatively uniformly spread over the unit sphere .
half of the fourier coefficients @xmath132 are drawn from their respective distribution as outlined in appendix [ bgeneration ] , while the other half are set by the requirement @xmath133 , necessary for obtaining a real value for the magnetic field . the value of @xmath134 controls the handedness of the field , namely @xmath135 corresponds to a maximally right - handed ( left - handed ) helical field .
finally @xmath136 determines the root mean square of @xmath137 . in fig .
[ nostoq ] we compute the q - statistics for 100 realizations of halos created with canonical pds and with a random magnetic field created using the parameters [ param1_sim ] b_rms&=&210 ^ -15 , k_mag=0.01/ , + f_&=&+1 , 2n^2 + 2=27 , where @xmath130 is the number of directions of the @xmath138 vector in eq .
( [ rmfonemode_notappen ] ) . in fig .
[ nostoq2 ] we plot the average of @xmath102 over all the realizations , denoted @xmath139 , for the same runs as in fig . [ nostoq ] and for the three gamma ray energy bin combinations . by doing so
, we have in mind of averaging the q - statistics obtained from small regions around multiple observed blazars .
the plot also shows the standard error in @xmath139 which is given by the standard deviation of @xmath0 divided by the square root of the number of realizations in the monte carlo simulations . the standard error follows from the central limit theorem and is the error in using the sample mean to estimate the population mean .
however , it assumes that the samples , monte carlo simulations in our case , from which values of @xmath0 are drawn are independent and identically distributed .
this is certainly true in our setup but may not be true for actual observations in which the same photons might contribute to the @xmath0 calculated for blazars that are close to each other .
in addition , there will be variation in the distance to observed blazars and other source characteristics .
we plan to take some of these factors into account in a follow - up analysis . versus @xmath1 for 100
monte carlo runs when the stochastic magnetic fields are generated using the parameters shown in eq .
( [ param1_sim ] ) with @xmath140 and canonical pds .
the mean @xmath139 is shown by the orange curve .
, scaledwidth=50.0% ] of fig .
[ nostoq ] for stochastic magnetic fields and canonical pds .
the width of the error band is given by the standard error _ i.e. _ standard deviation of the 100 monte carlo @xmath0 values divided by the square root of the sample size ( 100 ) .
, scaledwidth=50.0% ] versus @xmath1 for @xmath141 ( red , solid ) , @xmath142 ( black , dashed ) and @xmath143 ( blue , dotted ) for the three energy bin combinations : @xmath144 ( top row ) , @xmath145 ( middle row ) and @xmath146 ( bottom row ) where the bins are defined in eq .
( [ energybins ] ) .
the other parameters for the stochastic magnetic fields are given in eq .
( [ param1_sim ] ) ; the pds are also stochastic as described in sec .
[ sec : morph ] .
the bands denote standard error of @xmath147 .
, scaledwidth=50.0% ] versus @xmath1 for @xmath148 simulations with @xmath149 ( green , dotted ) , @xmath150 ( blue , solid ) , and @xmath151 ( orange , dashed ) and @xmath152 , @xmath153 .
, scaledwidth=50.0% ] it is very clear from fig .
[ nostoq2 ] that @xmath154 is a good measure of the magnetic field helicity when we assume canonical pds .
the next step is to see if @xmath155 remains a useful discriminant of magnetic helicity , even after we include stochasticity in both the magnetic field and the pds .
we have plotted @xmath139 versus @xmath1 for 1000 monte carlo simulations for @xmath156 ( fig .
[ unib_q_100_hel1 ] ) .
the plots show a clear correlation between @xmath157 and the helicity of the magnetic field .
it is also reassuring that the @xmath140 plot is the mirror image of the @xmath158 curve , just as we would expect due to parity reflection .
next we examine the dependence of @xmath139 on magnetic field parameters . in fig .
[ q_hel1_bvar_5em15_em15 ] we plot @xmath139 for several different magnetic field strengths and for fixed helicity @xmath152 .
we see that increasing the magnetic field strength leads to an increasing amplitude of @xmath147 for _ all _ energy combinations .
the increase is due to the magnitude of @xmath159 which becomes larger as the bending allows @xmath160 , @xmath161 to point further apart .
the effect of changing the magnetic field coherence scale is shown in fig .
[ q_hel1_kvar ] where the magnetic field strength and other parameters are fixed and only @xmath131 is varied . the magnetic field with larger coherence length gives a larger signal , but there is a turning point as extremely large coherence scale fields will behave like uniform fields .
the signal for smaller correlation length is washed out but the suppression depends on the energy combination .
this is to be expected from the analysis of ref .
@xcite since @xmath147 with a certain energy combination is sensitive to the magnetic helicity power spectrum at a definite coherence scale that is determined by the combination of energies . to probe magnetic fields on small length scales ,
it is necessary to consider gamma rays whose energies are close together @xcite .
thus the energy bins also have to be smaller and this means that the statistics is poorer .
the above results are from 1000 realizations in the monte carlo simulations .
each run gives a certain asymptotic value @xmath162 .
the 1000 values of @xmath162 for different magnetic field helicities and for different energy combinations are plotted in histograms in fig .
[ qhisto_3hist ] .
non - vanishing helicity causes the distribution of @xmath162 to skew in a direction that is in correspondence with the sign of the helicity . versus @xmath1 for @xmath148 simulations with @xmath153 ( solid , blue ) , @xmath163 ( orange , dashed ) , @xmath164 ( green , dotted ) with @xmath165 and @xmath140 . ,
the q - statistic is essentially a measure of the differential rotation found in the arrival direction of photons of different energies .
q will be negative ( positive ) if the rotation is right ( left)-handed as is depicted in fig .
[ qidea ] .
the results of our monte carlo simulations , for example in fig .
[ unib_q_100_hel1 ] , show that @xmath147 has a bump at small @xmath1 . here
we provide an explanation of this small @xmath1 bump .
magnetic field and canonical pds .
the lower the energy of the observed photon , the further away is the pp location from the source .
, scaledwidth=45.0% ] magnetic field and stochastic pds .
( the corresponding halo is shown in fig .
[ bumpskymap ] . )
the color coding is the same as in fig .
[ 3dnosto ] .
now the low energy observed photons have pp locations that are spread out in the direction of the los and some of these photons ( red dots ) originate closer to the source than the high energy pp locations ( blue dots ) . , scaledwidth=45.0% ] first note that the small @xmath1 bump does not appear in fig .
[ nostoq2 ] that was produced by assuming canonical , not stochastic , pds .
therefore the appearance of the bumps must be tied to the stochasticity of the pds .
when the pds take their canonical values , low energy gamma rays arise from interactions on a region of the pp surface that is far from the source , whereas high energy gamma rays arise closer to the source .
hence the pp locations are distributed in a natural progression , from close to the source to further away from the source , as the energy of the photons is decreased .
this can be seen in fig .
[ 3dnosto ] . on the other hand , when the pds are stochastic
, the low energy photons can arise from close to the source and also from far away as seen in fig .
[ bump3d ] .
the ones that come from close to the source are also closer to the los due to the geometry of the jet and are observed at smaller values of @xmath1 .
these low energy photons bend more than the high energy photons and can have pp locations that are higher up along the los than the high energy photons .
hence they measure the twist of the pp surface as one traverses the pp surface from low positions along the los to high positions ( instead of the usual direction of traversal from high to low ) .
if @xmath1 is large , then the @xmath102 statistics takes into account all the low energy photons and these are predominantly far away from the source .
so at large @xmath1 , the sign of the @xmath102 statistics is the same as that in the case of canonical pds .
thus @xmath0 changes sign once @xmath1 is large enough so that the mean position of the pp locations of the low energy photons is below that of the high energy photons .
this effect results in a halo of the form seen in fig .
[ bumpskymap ] and translates into a bump in @xmath0 at small values of @xmath1 and with a sign that is opposite to that of the asymptotic value of @xmath102 . ) but with @xmath158 .
the q - statistics is sensitive to half the halo , as depicted by the red shading . considering the remaining photons in the dark - shaded region , the red triangle and the purple square show the mean arrival directions of the low energy and high energy gamma rays within a disk of small @xmath1 in the upper diagram and of large @xmath1 in the lower diagram . comparing to fig .
[ qidea ] , we find @xmath103 for small @xmath1 and @xmath105 at large @xmath1 , implying a bump in the @xmath0 plot at small @xmath1 .
the color coding is the same as in fig .
[ 3dnosto ] .
, scaledwidth=45.0% ]
we have studied the effect of stochastic inter - galactic magnetic fields on the morphology of gamma ray halos .
the dependence of the morphology on the magnetic field strength , the coherence length , and the helicity were investigated .
most importantly , we have provided an understanding of the structure of the halo in geometrical terms , as arising due to the `` pp surface '' as determined by the magnetic field . in simple cases , the pp surface can be found analytically ( for example , eq . ( [ analyticpp ] ) ) . to analyze the halo morphology ,
we have proposed a sharper version of the q - statistic in eq .
( [ qdefn ] ) and applied it to simulated halos .
our key finding is that @xmath102 is a powerful diagnostic of the magnetic helicity ( fig .
[ unib_q_100_hel1 ] ) , field strength ( fig . [ q_hel1_bvar_5em15_em15 ] ) and coherence scale ( fig .
[ q_hel1_kvar ] ) . based on the analytical work of ref .
@xcite , we expect the sensitivity of @xmath102 to the coherence scale to depend crucially on the energies of the gamma rays that are used .
it would be interesting to quantitatively examine how the sensitivity of @xmath102 to the coherence scale can be improved with a choice of energy bins .
in addition , our monte carlo simulations have revealed a bump in @xmath139 at small values of @xmath1 ( see fig .
[ unib_q_100_hel1 ] ) . we have understood and explained this feature in terms of the pp surface in sec .
[ sec : bump ] .
this new feature may become useful in the analysis of real data in the future .
our present study is limited in a few ways that we plan to overcome in future work .
first , we have not included any background gamma rays .
these will introduce noise in the evaluation of @xmath147 and the error bars will increase .
we have also limited ourselves to stochastic isotropic magnetic fields but with only one @xmath166mode .
this is useful at this stage as it allows us to diagnose the effects of changing the coherence scale . in future
, we plan to include a spectrum of the magnetic field as motivated by current observations @xcite . in future
we also plan to incorporate the full development of the electromagnetic cascade into our numerical code , perhaps along the lines of ref .
@xcite or @xcite .
once we have understood individual blazar halos , we will apply our techniques to the diffuse gamma ray background which is expected to contain halos due to unseen blazars as well as those due to identified blazars .
we thank andrew long for comments .
tv is supported by the u.s .
department of energy , office of high energy physics , under award no .
de - sc0013605 at asu .
to generate helical magnetic fields , we first decompose the magnetic field @xmath137 in circularly polarized modes with basis vectors @xmath167 that are divergence - free eigenfunctions of the laplace operator ^(*k*)=*e*^(*k*)^i * k * ^i * k * . the triad of unit vectors , @xmath168 , is constructed as _ 1 , * e*_2 , * e*_3= where @xmath169 is any chosen unit vector such that @xmath170 .
hence any magnetic field can be decomposed as , @xmath172,\end{aligned}\ ] ] with the condition that b^(*k*)=-b^(-*k * ) [ realb ] to ensure that @xmath137 is real .
the divergence - free condition , @xmath173 , is automatically satisfied in this procedure .
we are interested in generating random magnetic fields with given energy ( @xmath174 ) and helical ( @xmath175 ) power spectra .
the relations between the modes @xmath176 and the power spectra are given by [ averagebsquare ] |*b*(*x*)|^2 = e_b(k ) d ( k ) and [ adotb ] ( * x*)(*x * ) = h_b(k ) d ( k ) .
the ratio of @xmath177 and @xmath178 will be written in terms of a function @xmath179 as @xcite , h_b(k)=f_h(k)e_b(k ) , [ defnfh ] and the `` realizability condition '' leads to the restriction -1f_h(k ) 1 the field is non - helical if @xmath180 , maximally right - handed if @xmath140 , and maximally left - handed if @xmath158 . in this paper
we focus on stochastic magnetic fields that are _ isotropic _ but have power on a single length scale @xmath185 and that have either @xmath186 or @xmath187 .
this corresponds to a delta function distribution for @xmath174 and vanishing or maximal helicity of either sign . to ensure that the magnetic fields are stochastically isotropic
, we choose @xmath188 vectors @xmath189 ( @xmath190 ) that discretize half of the two - sphere of directions in @xmath191space , @xmath192 with @xmath193 for @xmath194 , and @xmath195 once we have @xmath138 and @xmath176 as described above , we compute ( * k*)= b^+*k*^++b^- * k*^- for every @xmath196 .
we also find @xmath197 using the reality condition ( -*k * ) = * b*^*(*k * ) .
finally we obtain the random magnetic field , [ rmfonemode ] * b*(*x)=_*k*k * b*(*k*)e^i * k * where @xmath129 is the set of vectors @xmath198 for @xmath199 . * a. neronov and i. vovk , science * 328 * , 73 ( 2010 ) doi:10.1126/science.1184192 [ arxiv:1006.3504 [ astro-ph.he ] ] . s. ando and a. kusenko , astrophys . j. * 722 * , l39 ( 2010 ) doi:10.1088/2041 - 8205/722/1/l39 [ arxiv:1005.1924 [ astro-ph.he ] ] .
w. essey , s. ando and a. kusenko , astropart .
* 35 * , 135 ( 2011 )
doi:10.1016/j.astropartphys.2011.06.010 [ arxiv:1012.5313 [ astro-ph.he ] ] . h. tashiro , w. chen , f. ferrer and t. vachaspati , mon . not .
* 445 * , no .
1 , l41 ( 2014 ) doi:10.1093/mnrasl / slu134 [ arxiv:1310.4826 [ astro-ph.co ] ] . w. chen , b. d. chowdhury , f. ferrer , h. tashiro and t. vachaspati , mon . not .
450 * , no . 4 , 3371 ( 2015 ) doi:10.1093/mnras / stv308
[ arxiv:1412.3171 [ astro-ph.co ] ] . w. chen , j. h. buckley and f. ferrer , phys .
lett . * 115 * , 211103 ( 2015 ) doi:10.1103/physrevlett.115.211103 [ arxiv:1410.7717 [ astro-ph.he ] ] . j. d. finke , l. c. reyes , m. georganopoulos , k. reynolds , m. ajello , s. j. fegan and k. mccann , astrophys .
j. * 814 * , no . 1 , 20 ( 2015 ) doi:10.1088/0004 - 637x/814/1/20 [ arxiv:1510.02485 [ astro-ph.he ] ] .
a. brandenburg and k. subramanian , phys .
rept . * 417 * , 1 ( 2005 ) doi:10.1016/j.physrep.2005.06.005 [ astro - ph/0405052 ] .
a. elyiv , a. neronov and d. v. semikoz , phys .
d * 80 * , 023010 ( 2009 ) doi:10.1103/physrevd.80.023010 [ arxiv:0903.3649 [ astro-ph.co ] ] .
a. j. long and t. vachaspati , jcap * 1509 * , no . 09 , 065 ( 2015 ) doi:10.1088/1475 - 7516/2015/09/065 [ arxiv:1505.07846 [ astro-ph.co ] ] .
t. fitoussi , r. belmont , j. malzac , a. marcowith , j. cohen - tanugi and p. jean , doi:10.1093/mnras / stw3365 arxiv:1701.00654 [ astro-ph.he ] .
h. tashiro and t. vachaspati , phys .
d * 87 * , no .
12 , 123527 ( 2013 ) doi:10.1103/physrevd.87.123527 [ arxiv:1305.0181 [ astro-ph.co ] ] . | inter - galactic magnetic fields can imprint their structure on the morphology of blazar - induced gamma ray halos .
we show that the halo morphology arises through the interplay of the source s jet and a two - dimensional surface dictated by the magnetic field . through extensive numerical simulations ,
we generate mock halos created by stochastic magnetic fields with and without helicity , and study the dependence of the halo features on the properties of the magnetic field .
we propose a sharper version of the q - statistics and demonstrate its sensitivity to the magnetic field strength , the coherence scale , and the handedness of the helicity .
we also identify and explain a new feature of the q - statistics that can further enhance its power . |
it is generally accepted that most astrophysical fluids are turbulent with magnetic fields playing important role . in this situation
the properties of the fluids depend of the properties of the magnetized turbulence . in most cases astrophysical scales
are much larger than the scales determined by the plasma effects and therefore the magneto - hydrodynamic ( mhd ) description of astrophysical fluids is adequate . scattering and acceleration of cosmic rays ( see @xcite and references therein ) , acceleration of charged interstellar dust particles @xcite , conduction of heat ( see @xcite ) depend intimately on the properties of mhd turbulence
. rates at which the topology of otherwise `` frozen in '' astrophysical magnetic field changes , i.e. magnetic reconnection rates , according to @xcite are controlled by mhd turbulence .
it happened that for years mhd turbulence was a rather speculative area of research , where many models could coexist without much chance of being tested .
an important progress in the field was stimulated by _ in situ _ measurements of the statistics of interplanetary turbulence ( see @xcite ) .
however , these measurements were not sufficiently detailed to get a good insight into the structure of the magnetized eddies .
a study of _ incompressible _ mhd turbulence by goldreich & sridhar @xcite became a milestone for the subject .
there a universal scaling was proposed for magnetic eddies in the system of reference related to the _ local _ mean magnetic field of the eddy .
considerations applied by goldreich & sridhar to pseudo - alfvn waves have been generalized and extended to slow modes of _ compressible _ mhd ( @xcite ) .
recent theoretical and numerical work ( see @xcite , @xcite @xcite ) shed light on what is happening with mhd turbulence in partially ionized gas , on turbulence intermittency , on coupling of compressible and alfvnic modes , on rates of turbulence decay etc . in its turn
this work stimulated revisions of some earlier paradigms for basic astrophysical processes ( see @xcite ) .
it is important to stress that for the first time ever the theoretical predictions were tested with direct 3d numerical simulations .
does this mean that we are sure that we know the final truth about the actual astrophysical turbulence ?
surely , not .
it is well known that flows are similar when their reynolds numbers , which are the ratio of the advection to dissipation terms in the navie - stocks equations are the same . for mhd ,
magnetic diffusivity is important and therefore magnetic reynolds numbers should be also the same . as the corresponding numbers for the numerical simulations and astrophysics are different by a factor of @xmath0 and larger , from this point of view , the flows simulated in computers and those in astrophysical fluids are very different .
in other words , while theoretical expectations from elegant , but simplified models exist , their observational testing is essential .
moreover , such observational studies can provide an insight into astrophysically important questions of driving turbulence and the scales of energy injection .
the statistics of density , which is readily available in many cases , is related in a non - trivial way with the magnetic and velocity fluctuations ( @xcite ) .
obtaining of the statistics of velocity fluctuations from observations is a problem that has been subjected to scrutiny recently ( see @xcite and references therein ) . in what follows we shall discuss the problem of determining the statistics of magnetic fields by observing fluctuation of polarization arising from aligned dust and atoms .
starlight polarization measurements by hall and hiltner @xcite revealed that interstellar dust grains are aligned .
it did not take long to realize that grains tend in most cases tend to be aligned with their long axes perpendicular to magnetic field .
however , progress in theoretical understanding of the alignment has been surprisingly slow in spite of the fact that great minds like l. spitzer and e. purcell worked on grain alignment ( see @xcite ) .
the problem happened to be both very and tough as a lot of relevant physics had to be uncovered .
currently , grain alignment theory is an exciting branch of research where a lot of insight has been obtained lately ( see reviews @xcite ) .
we shall refer the reader to these reviews which contain dozens of references to the original important work , while here we present a few facts about dust alignment that are needed to understand how well the alignment represents magnetic field direction .
originally it was widely believed that interstellar grains can be well aligned by a paramagnetic or davis - greenstein mechanism @xcite .
this mechanism based on the direct interaction of rotating grains with the interstellar magnetic field required magnetic fields that are stronger than those obtained by other techniques .
later , a pioneering work by purcell @xcite showed a way how to make grain alignment more efficient .
purcell introduced several processes that are bound to make grains very fast rotators and noticed that grains rotating at high rates are not so susceptible to the randomization induced by gaseous collisions . for decades
this became a standard explanation for grain alignment , even though it could not explain several observational facts , e.g. why small grains are less aligned than the large ones ( @xcite ) .
new physics of grain internal motion uncovered recently explains inefficiency of alignment of small grains by purcell s mechanism .
lazarian & draine @xcite found that small grains flip frequently due to the coupling of rotational and vibrational degrees of freedom of a grain .
as the result regular torques , e.g. torques due to ejection of h@xmath1 molecules , get averaged out and grains rotate at thermal velocities .
the paramagnetic alignment of thermally rotating grains as we mentioned earlier is inefficient ( see @xcite ) .
interestingly enough , lazarian & draine @xcite found that coupling of rotational and vibrational degrees of freedom happens most efficiently through a process the origin of which is related to _
nuclear _ spins !
this relaxation makes grains of size @xmath2 cm rotate thermally , which makes the purcell mechanism inefficient for most of dust in _ diffuse _ interstellar medium .
a group of alternative mechanisms of alignment related to the relative gas - grain motion have their particular niches .
the first mechanical alignment mechanism was pioneered by gold @xcite .
later work included driving grains by ambipolar diffusion @xcite and alfvn waves @xcite .
although new efficient processes of mechanical alignment @xcite , this did not make the mechanical alignment universally applicable .
quite recently , radiative torque ( henceforth rt ) have been accepted as the most promising mechanism for explaining grain alignment over vast expanses of interstellar space .
introduced first by dolginov @xcite and dolginov & mytrophanov @xcite the rt were mostly forgotten till a more recent study by draine & weingartner @xcite , where their efficiency was demonstrated using numerical simulations ( see also @xcite , @xcite ) .
the rt make use of interaction of radiation with a grain to spin the grain up .
unpolarized light can be presented as a superposition of photons with left and right circular polarization . in general , the cross - sections of interaction of such photons with an irregular grain are different . as the result of preferential extinction of photons with a particular polarization
the grain experiences regular torques and gets spun up .
the predictions of rt mechanism are roughly consistent with the molecular cloud extinction and emission polarimetry @xcite and the polarization spectrum measured @xcite .
rt have been demonstrated to be efficient in a laboratory setup @xcite .
evidence in favor of rt alignment was found for the data obtained at the interface of the dense and diffuse gas ( see @xcite ) .
while it was originally believed that rt can not align grains at optical depths larger than @xmath3 , cho & lazarian @xcite demonstrated that the efficiency of rt increases sharply with the grain size and therefore bigger grains that exist within molecular clouds can be aligned for @xmath4 more of 30 .
in addition , as large grains do not flip frequently the purcell alignment gets efficient as well . ]
large grains may constitute an appreciable part of the total mass of dust within a cloud , while still be marginal in terms of light extinction .
therefore a non - detectable polarization in optical and near infrared does not preclude substantial polarization to be present in submillimeter .
this makes submillimeter polarimetry the preferred tool for studies magnetic fields and magnetic turbulence in molecular clouds . to align grains with rt mechanism requires grains to be comparable to the wavelength they interact with .
therefore grains smaller than @xmath5 cm are not aligned by rt in typical interstellar conditions .
these grains flip frequently and therefore they can not be aligned by @xcite process according to @xcite .
however , these grains can be aligned by the original davis - greenstein process ( see detailed calculations in @xcite ) .
the degree of the expected alignment for typical interstellar conditions does not exceed several percent , but this alignment is sensitive to the _ magnitude _ of magnetic field . the data in @xcite is consistent with the existence of a population of small aligned grains .
such aligned grains may be revealed with uv polarimetry .
most of the processes produce grain alignment in respect to magnetic field , even if the alignment mechanism is of non - magnetic nature .
this is true due to the rapid precession of grains about magnetic field .
this precession stems from the substantial magnetic moment that grains get due to their rotation @xcite .
the corresponding period of grain precession @xmath6 is @xmath7 s , where the the external magnetic field is normalized over its typical interstellar value of @xmath8 g and grain size is chosen to be @xmath9 cm .
this means that for turbulent motions on time scales longer than @xmath6 grain alignment in respect to magnetic field lines does not change as the consequence of the adiabatic invariant conservation ( see more discussion in @xcite ) .
if the alignment happens on the time scales shorter than @xmath6 the dust orientation may not reflect the magnetic field . for the rt such a fast alignment will happen with longer grain axes perpendicular _ to the direction of radiation _ , while the fast mechanical alignment will happen with longer axes parallel to the flow .
the mechanical and rt alignment takes place on the time scale of approximately gaseous damping time , which is for interstellar medium is @xmath10 s , where typical temperatures and densities of cold interstellar medium , which are respectively 100k and @xmath11 @xmath12 , were used for the normalization .
note , that magnetic alignment takes place over even longer time scales , namely , @xmath13 s. therefore in most cases the magnetic field indeed acts as the alignment axis .
it is worth noting that the turbulent fluctuations with periods longer than @xmath6 _ do not _ suppress alignment .
the rapidly precessing grains preserve their orientation to the local direction of magnetic field and undergo the alignment even when this local direction is changing in space . in this respect grain alignment
is a local process that can reflect local direction of magnetic field for magnetic ripples larger than @xmath14 , where @xmath15 is alfvn speed . whatever is the process of alignment ,
if it aligns grains over timescale larger than @xmath6 the alignment is _ perpendicular or parallel _ to magnetic field .
the issues of the mechanism and the efficiency of alignment is essential when averaging along the line of sight is involved .
for instance , if we study turbulence in a molecular cloud we would like to know whether the polarization signal that we measure is coming from a thin skin layer of the cloud or from the whole cloud volume .
we also would like to know whether the direction of alignment changes along the line of sight .
it follows from the discussion above that in most cases the assumption that grains are aligned with long axes perpendicular to magnetic fields is a valid one .
the fact , that rt have been identified as a dominant alignment mechanism plus that this process seems to be efficient even at substantial optical depths simplifies the interpretation of the polarization fluctuations in terms of magnetic field fluctuations . if we limit ourselves to discussing grains aligned by rt , the grains in diffuse interstellar medium can be considered perfectly aligned with longer axes perpendicular to magnetic field .
optical and near infrared emission from stars will be polarized via the differential extinction by the aligned grains so that the direction of linear polarization will be parallel to the local interstellar magnetic field .
the variations of magnetic field directions along the line of sight from the star to the observer will influence the resulting degree of polarization @xmath16 as well as the positional angle of the polarization vector @xmath17 . these observational parameters can be used to obtain the polarization statistics .
the geometry of lines of sight is also important for the interpretation of interstellar polarization statistics .
it can be easily shown ( see @xcite , @xcite ) that when a signal is collected along the converging lines of sight ( see fig . 1 )
, the turbulence provides a universal @xmath18 spectrum of fluctuations when the correlated lines of sight are separated by an angle larger than the ratio of the injection scale of turbulence and the distance to the stars , i.e. @xmath19 . the underlying spectrum of magnetic turbulence can be revealed . in the opposite
limit the universal assymptotics @xmath18 is present .
_ right _ : spectra in terms of spherical harmonics for the degree of polarization for stars at a fixed at at a random distance from the observer . from @xcite.,title="fig:",scaledwidth=40.0% ] the underlying spectrum of magnetic turbulence can be revealed . in the opposite limit the universal assymptotics
@xmath18 is present .
_ right _ : spectra in terms of spherical harmonics for the degree of polarization for stars at a fixed at at a random distance from the observer . from @xcite.,title="fig:",scaledwidth=40.0% ] the variations of @xmath16 were studied in @xcite for a sample of polarization measurement available in heiles catalog .
they provided a spectrum @xmath20 that was interpreted by @xcite as the result of the superposition of the contributions from distant and nearby stars in the presence of mhd turbulence in the interstellar dusty gas ( see fig . 1 ) . a more systematic study of polarization using only close and only distant stars could test the predictions in @xcite .
it shall also be able to establish the correlation scale of galactic turbulence @xmath21 .
interestingly enough , confirming that the polarization fluctuations in milky way arise from mhd turbulence is important from the point of view of cmb studies .
aligned dust provides an important contribution to polarized microwave foreground ( see @xcite and references therein ) .
this foreground must be removed to reveal the polarized emission originated in the early universe .
mhd turbulence has a well defined statistics and therefore this open ways for more reliable ways of weeding out such a foreground .
as we discussed above , polarimetry of aligned dust provides one of the ways of studying mhd turbulence .
polarimetry of some molecular lines using goldreich - kylafis effect @xcite have been shown recently to be a good tool for magnetic field studies in molecular clouds ( @xcite ) . here
we discuss yet another promising technique to study magnetic fields that employs optical and uv polarimetry .
this technique can be used for interstellar and circumstellar studies as well as for studies of magnetic fields in qsos and other exotic objects .
the technique is based on the ability of atoms to be aligned by external radiation in their ground state and be realigned through precession in magnetic field .
it has been known that atoms can be aligned through the interactions with the anisotropic flux of resonance emission ( see review @xcite ) .
alignment is understood here in terms of orientation of vector angular momentum @xmath22 , if we use the language of classical mechanics . in quantum terms
this means a difference in the population of sublevels corresponding to the projections of angular momentum to the quantization axis
. we will call atomic alignment as the alignment of @xmath22 in the ground state .
this state is long lived and therefore is being affected by weak magnetic field .
atomic alignment has been studied in laboratory in relation with early day maser research ( @xcite ) .
alignment of atoms in ground state can change the optical properties of the medium .
this effect was noticed and made use of for the interstellar case by @xcite in case of hyperfine structure of the ground state , for fine structure of the ground state by , e.g. @xcite .
here we shall be concerned with the variations of degree of polarization @xmath16 and the positional angle @xmath17 that is caused by atomic alignment .
a possibility of using atomic alignment to study weak magnetic field in diffuse media was first discussed in @xcite for ideal two - level atoms with fine structures and polarization of emissions from these atoms were considered for a very restricted geometry of observations , namely , when magnetic field is aligned with line of sight and perpendicular to the incident light .
a discussion of alignment of various atoms and ions for arbitrary orientations of magnetic fields towards the line of sight and the direction to the illumination source is given in @xcite .
similar to the case of interstellar dust , the rapid precession of atoms in magnetic field makes the direction of polarization sensitive to the direction of underlying magnetic field .
as the precession of magnetic moments of atoms is much faster than the precession of magnetic moments of grains atoms can reflect much more rapid variations of magnetic field .
more importantly , alignable atoms and ions ( see table 1 ) can reflect magnetic fields in the environments where either properties of dust change or the dust can not survive .
this opens wide avenues for the magnetic field research in circumstellar regions , interstellar medium , interplanetary medium , intracluster medium and quasars etc .
in addition , the degree of alignment @xmath16 can help in determining the angle between the direction towards the illumination source and the direction of magnetic field .
this information is not available by any other technique .
.note : only polarizable and alignable components are listed .
`` m '' stands for marginal polarization . from @xcite [ cols="^,^,^,^,^,^,^",options="header " , ]
the basic idea of the atomic alignment is simple .
the alignment is caused by the anisotropic deposition of angular momentum from photons . in typical astrophysical situations
the radiation flux is anisotropic . as the photon spin is along the direction of its propagation
, we expect that atoms that scattered the radiation can be aligned in terms of their angular momentum .
such an alignment happens in terms of the projection of angular momentum to the direction of the incoming light .
it is clear that to have the alignment of the ground state , the atom should have non - zero angular momentum in its ground state
. therefore fine or hyperfine structure of the ground state is necessary to enable various projection of angular momentum to exist in the ground state . whether polarization arising from aligned atoms is parallel or perpendicular to magnetic field depends on the angle that the anisotropic radiation makes with the magnetic field .
let us discuss a toy model that provides an intuitive insight into the physics of atomic alignment .
first of all , consider a toy model of an atom with lower state corresponding to the total angular momentum @xmath23 and the upper state corresponding to angular momentum @xmath24 . if the projection of the angular momentum to the direction of the incident resonance photon beam is @xmath25 , for the lower state @xmath25 can be @xmath26 , @xmath27 , and @xmath28 , while for the upper state @xmath29 ( see fig .
2_left _ ) .
the unpolarized beam contains an equal number of left and right circularly polarized photons which projection on the beam direction are 1 and -1 .
thus absorption of the photons will induce transitions from @xmath30 and @xmath31 states .
however , the decay of the upper state happens to all three levels . as the result
the atoms get accumulated at the @xmath29 ground state from which no excitations are possible . as a result of that the optical properties of the media ( e.g. absorption ) would change . this above toy model
can also exemplify the role of collisions and magnetic field .
without collisions one may expect that all atoms reside eventually at the state of @xmath29 .
collisions , however , redistribute atoms with different states . however , as disalignment of the ground state requires spin flips , it is less efficient than one can naively imagine .
calculations by @xcite show that to disalign sodium one requires more than 10 collisions with paramagnetic atoms and experimental data by @xcite support this .
this reduced sensitivity of aligned atoms to disorienting collisions makes the effect important for various astrophysical environments .
, as radiation removes atoms from the ground states @xmath32 and @xmath33 ; _ middle _ : resonance scattering of solar light by sodium tail from comet ; _ right : _ polarization caused by sodium aligned in the comet wake .
spatial and temporal fluctuations of polarization carry the information on mhd turbulence.,title="fig:",scaledwidth=32.0% ] , as radiation removes atoms from the ground states @xmath32 and @xmath33 ; _ middle _ : resonance scattering of solar light by sodium tail from comet ; _ right : _ polarization caused by sodium aligned in the comet wake . spatial and temporal fluctuations of polarization carry the information on mhd turbulence.,title="fig : " ] , as radiation removes atoms from the ground states @xmath32 and @xmath33 ; _ middle _ : resonance scattering of solar light by sodium tail from comet ; _ right : _ polarization caused by sodium aligned in the comet wake .
spatial and temporal fluctuations of polarization carry the information on mhd turbulence.,title="fig : " ] magnetic field would also mixes up different @xmath25 states .
however , it is clear that the randomization in this situation will not be complete and the residual alignment would reflect the magnetic field direction in respect to the observer .
magnetic mixing happens if the angular momentum precession rate is higher than the rate of the excitation of atoms from the ground state , which is true for many astrophysical conditions .
note that in order to be aligned , first the atoms should have enough degree of freedom , namely , the quantum number of angular momentum must be @xmath34 .
second , incident flux must be anisotropic
. moreover , the collisional rate should not be too high . while the latter requires special laboratory conditions , it is the case for many astrophysical environments such as the outer layers of stellar atmosphere , interplanetary , interstellar , and intergalactic media , etc .
as long as these conditions are satisfied , atoms can be aligned within either fine or hyperfine structure . for light elements ,
the hyperfine splitting is very small and the line components overlap to a large extent .
however , for resonant lines , the hyperfine interactions causes substantial precession of electron angular momentum @xmath35 about the total angular momentum @xmath36 before spontaneous emission . therefore total angular momentum should be considered and the @xmath37 basis must be adopted ( @xcite ) . for alkali - like atoms ,
hyperfine structure should be invoked to allow enough degrees of freedom to harbor the alignment and to induce the corresponding polarizations . in terms of time scales
, we have a number of those , which makes the problem interesting and allows getting additional information about environments .
the corresponding rates are 1 ) the rate of the precession @xmath38 , 2 ) the rate of the photon arrival @xmath39 , 3 ) the rate of collisional randomization @xmath40 , 4 ) the rate of the transition within fine or hyperfine structure @xmath41 . in many case @xmath42 .
other relations are possible , however . if the larmor precession gets comparable with any of the other rates , it is possible to get information about the _ magnitude _ of magnetic field .
another limitation of our approach is that we consider that @xmath38 is much smaller that the rate of the decay of the excited state , which means that we disregard the hanle effect .
as an illustration , we discuss here a synthetic observation of a comet wake .
though the abundance of sodium in comets is very low , its high efficency of scattering sun light makes it a good tracer ( @xcite ) .
the sodium atoms ejected from the comets are aligned by the solar radiation .
distant from comets , the sun can be considered a point source . as shown in fig .
2_middle _ , the geometry of the scattering is well defined , i.e. , the scattering angle @xmath43 is known .
the polarization of the sodium emission thus provides an exclusive information of the magnetic field in the comet wake .
embedded in solar wind , the magnetic field is turbulent in a comet wake .
we take a data cube directly from mhd simulations of a comet wake . depending on its direction , the embedded magnetic field
alter the degree of alignment and therefore polarization of the light scattered by the aligned atoms .
therefore , fluctuations in the linear polarization are expected from such a turbulent field ( see fig .
2_right _ ) . for interplanetary studies
, one can investigate not only spatial , but also temporal variations of magnetic fields .
this can allow cost effective way of studying interplanetary magnetic turbulence at different scales .
astrophysical environments provide unique opportunities for studying mhd turbulence at reynolds and magnetic reynolds numbers that are not available through either numerical simulations or laboratory experiments . this allows to test our theoretical constructions to see whether numerics is adequate .
it is difficult to overestimate the importance of understanding better astrophysical turbulence including its spectra , injection and dissipation scales .
magnetic field statistics is complementary to the velocity statistics that can be studied using doppler shifted spectral lines ( see @xcite ) .
aligned dust and aligned atoms provide a unique way of studying magnetic turbulence in various astrophysical environments from molecular clouds to quasars .
recent progress in theories of atomic and grain alignment gives us confidence in interpreting polarization in terms of underlying magnetic fields .
aligned atoms provide a possibility of studying temporal variations arising from interplanetary turbulence .
techniques that use aligned atoms and aligned grains are synergetic for studies of magnetic turbulence .
abbas , m. , craven , p. , spann , j. , tankosic , d. , leclair , a .
, gallagher , d. , west , e. , weingartner , j. , witherow , w. , & tielens , a. 2004 , apj , 614 , 781 cho , j. & lazarian , a. 2002 , phy .
lett . , 88 , 245001 cho , j. & lazarian , a. 2002 , apjl , 575 , l63 cho , j. & lazarian , a. , 2003 , mnras , 345 , 325 cho , j. & lazarian , a. 2004 , in proc . of the summer program 2004 , stanford univ . , p. 75
cho , j. & lazarian , a. 2005 , theoret .
fluid dynamics , in press cho , j. & lazarian , a. 2005 , apj , submitted cho , j. , lazarian , a. , vishniac , e. t. 2003 apj , 595 , 812 cho , j. , lazarian , a. , honein , al . , knaepen , b. , kassinos , s. , moin , p. 2003
, apj , 589 , 77l cremonese et al .
1997 , apjl , 490 , l199 davis , j. , & greenstein , j.l .
1951 , apj , 114 , 206 dolginov , a. z. 1972,apss , 16 , 337 dolginov a.z . & mytrophanov , i.g .
1976 , ap&ss , 43 , 291 draine , b.t . , &
weingartner , j.c .
1996 , apj 470 , 551 ( dw96 ) draine , b. & weingartner , j. 1997 , apj , 480 , 633 fosalba , p. , lazarian , a. , prunet , s. , & tauber , j. 2003 , apj , 564 , 762 girart , j. m. ; crutcher , r. m. , & rao , r. 1999 , apj , 525l , 109 gold , t. 1951 , nature , 169 , 322 goldreich , p. & kylafis , n. d. 1982 , apj , 253 , 606 goldreich , p. & sridhar , h. 1995 , apj , * 438 * , 763 hall , j.s .
1949 , science , 109 , 166 higdon , j.c . 1984 , apj , 285 , 109 hiltner , w.a . 1949 ,
apj , 109 , 471 happer , w. 1972 rev .
modern phys . , 44 , 2 hawkins , w. b.1955 , phys .
98 , 478 hildebrand , r. h. , davidson , j. a. , dotson , j. l. , dowell , c. d. , novak , g. , & vailancourt , j. e. 2000 , asp , 112 , 1215 kastler , a. 1956 , j. optical soc .
america , 47 , 460 kim , s .- h . & martin , p. 1995 ,
apj , 444 , 293 landolfi , m. & landi deglinnocenti e. 1986 , a&a , 167 , 200 lazarian , a. 1994 , mnras , 268 , 713 lazarian , a. 1995 , apj , 451 , 660 lazarian , a. 1997 , mnras , 288 , 609 lazarian , a. 1997 , apj , 483 , 296 lazarian , a. 2003 , journal of quant .
spect . and radiative transfer , 79 , 881 lazarian , a. 2004 , jkas , 37 , 563 lazarian , a. & cho , j. 2005 , in astrophysical polarimetry 2004 , hawaii , in press , astro - ph/0408174 lazarian , a. , & draine , b.t .
1999a , apjl , 516 , l37 lazarian , a. , & draine , b.t .
1999b , apjl , 520 , l67 lazarian , a. , efroimsky , m. , & ozik , j. 1996 , apj , 472 , 240 lazarian , a. , & finkbeiner 2003 , new astronomy reviews , 47 , v - viii , p.1107 lazarian , a. , goodman , a.a . , & myers , p.c .
1997 , apj , 490 , 273 lazarian , a. & shutenkov , v. 1990 , pisma v astronom .
zhurnal , 16 , 690 lazarian , a. & vishniac , e. 1999 , apj , * 517 * , 700l lazarian , a. , vishniac , e. t. , & cho , j. 2004 apj , 603 , 180l lazarian a. , & yan , h. 2002 apjl , 566 , l105 lee , h .- w . , blandford , r. d. , & western , l. , 1994 , mnras , 267 , 303 lithwick , y. and goldreich , p. 2001 , apj , * 562
* , 279 montgomery , d. & turner , l. 1981 , phys . fluids , 24 , 825 narayan , r. & medvedev , m. , apj , 562 , 129l purcell , e.m .
1969 , physica , 41 , 100 purcell , e.m . 1975 , in _ the dusty universe _ , eds g.b
. field & a.g.w .
cameron , new york , neal watson , p. 155 .
roberge , w. & hanany , s. 1990 , b.a.a.s . , 22 , 862 purcell , e.m .
1979 , apj , 231 , 404 roberge , w. & hanany , s. 1990 , b.a.a.s . , 22 ,
862 roberge , w.g . , & lazarian , a. 1999 , mnras , 305 , 615 schlickeiser , r. 2002 , cosmic ray astrophysics , springer - verlag , berlin heidelberg spitzer , l. , & mcglynn , t.a .
1979 , apj , 231 , 417 spitzer , l. , jr . ; tukey , j. 1951 apj , 114 , 187 thomas , n. 1992 , suvey geophys .
, 13 , 91 varshalovich , d. a. 1968 , astrofizika , 4 , 519 walkup , r. , migdall , a. l. , & pritchard , d. e. 1982 , phys .
a , 25 , 3114 weingartner , j. & draine , b. 2003 , apj , 589 , 289 yan , h. & lazarian , a. 2002 , phys .
lett . , * 89 * , 281102 yan , h. , & lazarian , a. 2003 , apjl , 592 , l33 yan , h. & lazarian , a. 2004 , apj , 614 , 757 yan , h. & lazarian , a. 2005 , apj , submitted | most astrophysical fluids are turbulent and magnetized . fluctuations of polarization provide a promising way to study astrophysical magnetic turbulence .
we discuss polarization that arises from grains and atoms aligned in respect to magnetic field , describe the processes of alignment , explain when the alignment reflects the direction of magnetic field .
we show that temporal fluctuations of interplanetary magnetic field can be studied using aligned sodium atoms .
address = department of astronomy , university of wisconsin , 475 n. charter st .
, madison , wi 53706 |
active galactic nuclei ( agns ) can be divided into two classes : radio - loud and radio - quiet .
a small subgroup of radio - loud agns have displayed flux variability on time scales ranging from hours to decades ( see a review by fan 2005 ) .
examples of long - term variations can be found in the light curves of 3c 120 ( jurkevich et al .
1971 ) , oj 287 ( sillanp@xmath4 et al . 1988 ) , 3c 345 ( kidger et al . 1992 ) , pks 0735 + 178 ( pollock , 1975 , fan et al .
1997 , qian & tao 2004 ) , bl lacertae ( fan et al .
1998 ) , 3c 66a ( lainela et al . 1999 ) , 3c 279 ( fan 1999 ) , s5 0716 + 714 ( qian et al . 2002 ) , mrk 335 ( tao et al . 2004 ) , pks 1510 - 089 & ma 0829 + 047 ( liller & liller 1975 ) , b2 1101 + 38 ( miller 1975 ) and in various other radio - selected bl lacertae objects ( fan et al . 2002 ) .
3c 390.3 ( z=0.0561 ; osterbrock et al .
1975 ) is a well - known broad - line radio galaxy with broad double peaked emission line profile ( sandage 1966 ; lynds 1968 ) .
it is also an extended double - lobed fr ii radio source .
leahy & perley ( 1995 ) discussed the jets and hotspots of it and they also provided some basic data of this source .
the vlbi observations at 5 ghz show evidence of superluminal motions , with @xmath5 ( alef et al .
3c 390.3 has a well known history , both in the continuum light and in emission lines(e.g .
, dietrich et al . 1998 ; shapovalova et al . 2001 ; sandage 1967 , 1973 ; zheng 1996 ) .
prieto & kotilainen(1997 ) detected optically - resolved emission coincident with the brightest region of the northern radio lobe ( hot spot b ) in b , r , and i bands .
dunn et al.(2006 ) provided a database of ultraviolet continuum light curves for agns , and 3c 390.3 is one of them .
_ rosat pspc _ of 3c 390.3 show resolved x - ray emission spatially located at the position of the northern radio hotspot(prieto 1997 ) .
it is also a highly variable x - ray source ( inda et al .
1994 ; eracleous et al . 1996
; wozniak et al . 1998 ; leighly et al . 1997 ; leighly & obrien 1997 ; gliozzi et al .
gaskell(2006 ) discussed the relationship between the x - ray , uv , and optical variability in 3c 390.3 .
the narrow - line region in 3c 390.3 is probably on the order of 1 lt yr , and a significant part of the nlr is located near the line of sight(zheng et al .
the black hole mass of 3c 390.3 has been estimated through reverberation mapping ( e.g. , mclure & dunlop 2001 ; botte et al .
2004 ; shapovalova et al .
2001 ) , ranging from @xmath6 to @xmath7 .
3c 390.3 has been one of our primary targets for long - term monitoring in the optical band at the shanghai astronomical observatory(shao ) . in this paper , we present v , r , and i data taken in the period from 1995 to 2004 .
we also compile historical data from 1894 to 2004 .
we obtain a total of 786 observational points , to which we adopt the @xmath8 method to search for periodicities .
the observations presented here were obtained with the 1.56-m telescope at shao from 1995 march 6 to 2004 august 14 . at
first a liquid nitrogen cooled ccd camera ( @xmath9 pixels , 1 pixel = 0.019 mm ) was used .
the field of view of this was 4.3 arcminutes ( 1 pixel = 0.25 arcsec ) when used directly at the cassegrain focus , and approximately 13 arcmins ( 1 pixel = 0.76 arcsec ) when used with a focal reducer .
a new liquid nitrogen cooled ccd camera ( @xmath10 pixels , site ccd chip ) has been used since august , 2002 .
the chip of this camera subtends 11 arcmin by 11 arcmin in sky , with a scale is 0.31 arcsec per pixel ( 1 pixel = 0.024 mm ) .
standard johnson - cousins v , r , and i filters were used .
typical integration times were 300 s for i and r , and 600 s for filter v , depending on sky conditions and the brightness of 3c 390.3 .
the flat field images were taken at dusk and dawn .
the bias images were taken at the beginning and the end of the observations , while the dark - field images were taken at the end .
all observing data were processed using the iraf software package .
the seeing at the sheshan station of shao varied from 1.3 to 2.0 arcseconds fwhm and we have discussed this elsewhere ( see fig . 3 of qian et al .
2002 ) . since 3c 390.3 is not a point - like source , seeing fluctuations may have an effect .
thus , we have followed the recommendations of cellone et al .
( 2000 ) and selected a photometric aperture of radius 6.5 arcseconds to minimize potential seeing - dependent effects .
cellone et al.(2000 ) gave the minimum aperture radii in table 2 of their paper , for a v=16 mag .
agn and various host galaxy for magnitudes @xmath11 .
the fainter the @xmath11 , the smaller the photometric radii that can be used .
shapovalova et al.(2001 ) used a 10 arcsec diameter photometric aperture , corresponding a 5 arcsec radius aperture , which is also comparable to the one we have used .
we used comparison stars a , b , and d of penston et al.(1971 ) and hst gs 4951:731 .
v , r , and i magnitudes were taken from dietrich et al.(1998 ) . using differential photometry with respect to each comparison star , @xmath12 ,
within the frame we obtain a magnitude @xmath13 for the object of interest
. then the object s magnitude at that time is @xmath14 where @xmath15 is the number of standard stars , in this case , 4 .
the uncertainty @xmath16 was calculated as @xmath17 the difference between comparison stars is @xmath18 where @xmath19 is the magnitude difference of comparison stars a and b as given by dietrich et al.(1998 ) , and @xmath20 and @xmath21 are the observed magnitudes of comparison stars a and b. the absolute value of @xmath22 was used as an additional indicator of the observational uncertainty .
the observational data are listed in tables 13 , in which the first column gives the julian date , the second column the magnitudes , the third the @xmath16 uncertainties , and the fourth the @xmath23 uncertainties .
figures 1 3 show the light curves of 3c 390.3 and the magnitude differences between comparison stars a and b in the v , r , and i filters , respectively . in our observation period from march 1995 to august 2004 ,
variations of [email protected](15.023 to 16.395 ) in the v band , [email protected](14.352 to 15.592 ) in the r band , and [email protected] ( 13.896 to 15.180 ) in the i band are found .
the observations show that 3c 390.3 was brightening from 1995 and reached its brightest in 1996 september .
then it slowly declined in brightness , but was brightening again in september 2003 .
we have compared our results with those of dietrich et al.(1998 ) .
our data have 10 days of overlap with the data of dietrich et al .
we converted their published flux data of 3c 390.3 to v magnitudes by taking @xmath25 . for each day , the observational data were averaged and they were shown in fig .
this comparison gave the relationship used to convert their fluxes to v magnitudes .
dunn et al.(2006 ) determined the uv light curves of 3c 390.3 using the multimission archive at stsci .
the data were observed with the international ultraviolet explorer , from 1978 to 1996 .
the light curves of the three bands , centered at 1431 , 1816 , and 1912 angstroms , all show a brightening trend from jd 2449750 to jd 2450162 , which is consistent with our v , r , and i light curves for the same period .
we have reconstructed the historical light curve by combining our recent data with data from the literature ( cannon et al .
1971 ; selmes et al . 1975 ; dietrich et al .
1998 ; scott et al . 1976 ; sergeev et al . 2002 ; lloyd 1984 ; shapovalova et al . 2001 ; sandage 1967 , 1973 ; shen et al .
1972 ; yee & oke 1981 ; pica et al .
1980 , 1988 ) .
most of these data set are for the v band or continuum near v band .
however , there are some b band data in the remaining papers . in order to include these data
, we needed to convert them to v band magnitude .
shapovalova et al .
( 2001 ) provide simultaneous b and v data ( see fig . 5 ) . from their data
we find that @xmath26 with a standard deviation of 0.052 mag .
this relationship was then used to convert b - band magnitudes to v - band magnitudes to give a historical light curve in the v band .
this has a total of 786 observational points over a time interval of 111 years ( from 1894 to 2004 ) .
we show the historical light curve in fig .
it shows two outbursts ( 1970 and 1996 ) and a possible third outburst ( near 1939 ) . using the historical light curve we have constructed for 3c 390.3
, we searched for possible periodicities .
as in our previous papers , we use the @xmath8 analysis ( foster 1995 ) to search for periodicities ( fan et al . 2006 , 2007 ) .
the @xmath8 analysis cleans spurious periodicities as follows .
first , the strongest single peak and corresponding aliases are subtracted from the original spectrum , then the residual spectrum is scanned to determine whether the next strongest remaining peak is statistically significant .
if so , then the original data are analyzed to find the pair of frequencies which best models the data , these two peaks and corresponding aliases are subtracted , and the residual spectrum is scanned again .
the process continues , producing successive @xmath8 spectra , until all statistically significant frequencies are included .
the light curve of 3c 390.3 shows that the data are unevenly sampled with most data concentrated in the period from 1965 to 2004 .
observations before 1933 are particularly sparse . after using the @xmath8 algorithm on the 1933 - 2004 data set
we find that there are seven independent frequency components required to `` clean '' the light curve . the @xmath8 spectrum is shown in fig .
7 . for comparison , in fig .
10 we show the results of using only the post 1965 data .
when analyzing unevenly sampled time series , the irregular spacing introduces many complications into the fourier transformations ; it can alter peak frequencies ( slightly ) and amplitudes ( greatly ) , and even introduce extremely large spurious peaks . including all 3c 390.3 data points
would not only give irregular sampling but would also give undue weight to the most recent 40 yr of data . to obtain regular sampling and to give more uniform weighting to different epochs we binned the data from 1933 onwards .
however , binning data inevitably throws away information . to minimize this we adopted a bin size of 0.02 yr ( 7.30 days ) which is short enough compared to the long - term periods we are looking for ( years ) and thus unlikely to distort long - term variations .
this binning gives us 361 points in the binned light curve from 1933 to 2004 . in fig .
8 we show the @xmath8 spectrum for this binned light curve .
we also investigated the effect of removing a long - term linear trend in the binned 19332004 historical light curve using @xmath27 .
in fig.9 we show the @xmath8 spectrum for this light curve . following foster(1996 ) , the variance of a frequency @xmath28 and the variance of the amplitude of the given frequency @xmath29 can be estimated by : @xmath30 where @xmath31 is the variance of the residual data , @xmath32 is the amplitude of the given frequency and @xmath33 is the total time span , and @xmath34 means the number of data values in an observed time series .
the @xmath35 is estimated by @xmath36 where @xmath37 is the variance of residual data , @xmath38 , and @xmath39 is the number of discrete frequencies . @xmath40
represent a vector in a n - dimensional vector space . @xmath41\ ] ] defining the inner product of two vectors , @xmath42 and @xmath43 , as the average value of the product @xmath44 over the sampling @xmath45 @xmath46 defining the constant vector @xmath47 as @xmath48 .\ ] ] the variation of a vector @xmath42 is @xmath49 the results of these four @xmath8 analyses are listed in table 4 . in this table
we also give the false alarm probability ( fap ) of each of the @xmath8 frequency components , which depends on the amplitude .
small fap values support the reality of these periodicities ( see fan et al .
2006 for details ) . the unbinned v - band data , the binned data , the linear - trend removed data , and the post-1965 data ( see table 4 ) all show that there are apparent periodicities .
those matched periods derived from the first three data sets are : @xmath50 days(@xmath51 days ) , @xmath52 days ( @xmath53 days , @xmath54 days ) , @xmath55 days ( @xmath56 days , @xmath57 days ) , @xmath58 days(@xmath59 days ) , and @xmath60 days ( @xmath61 days , @xmath62 days ) .
however , since the period @xmath50 days(@xmath51 days ) is almost the length of the data coverage , the period is not a physically significant one . from fig
6 , we can see that most data was sampled in the period of 1965 to 2004(about 40 years ) .
therefore , we reinvestigated periods based on the post 1965 data .
the matched periods derived from the v band data(1933 - 2004 ) and the post 1965 data(1965 - 2004 ) are : @xmath63 days(@xmath64 days ) , @xmath52 days(@xmath65 days ) , @xmath66 days(@xmath67days ) , @xmath68 days(@xmath68 days ) , and @xmath69 days(@xmath69 days ) . in table 4
, we also present the phase of the corresponding period .
phase and period can be used to identify a signal in multi - data set .
we used the obtained components to fit all the light curves and the results are shown in fig .
according to the data processing theory , the results derived from the linear - trend removed data set is more reliable .
so the possible periods are : @xmath0 , @xmath1 , @xmath2 , and @xmath3 years .
agns are variable throughout the whole electromagnetic spectrum . in our monitoring program
, we have observed the galaxy 3c 390.3 from march 1995 to august 2004 .
the observations clearly show that the source is variable in the optical band with the variation amplitude of [email protected] , [email protected] , and [email protected] in the v , r , and i bands respectively .
the v band historical light curve is compiled , which has a time span of 111 years .
possible periods of @xmath0 , @xmath1 , @xmath2 , and @xmath3 years were found in the light curve by means of @xmath8 method .
it has been suggested that the long - term periodic outbursts of oj 287 may be explained by a binary black hole ( sillanp@xmath4 et al .
1988 ) or such outbursts may be due to thermal and viscous instabilities in a thin accretion disk ( meyer & meyer - hofmeister 1984 ; horiuchi & kato 1990 ) .
the historical light curve of 3c 390.3 shows strong variability .
several possible periods were found by @xmath8 method .
the multiple periods we derived may imply the instabilities in the disk .
the analysis of spectra of 3c 390.3 covering a period of over 20 yr may indicate the binary black hole model(gaskell 1996 ) .
shapovalova et al.(2001 ) have monitored this object between 1995 and 2000 .
their results do not support either the models of outflowing biconical gas streams or those of supermassive binary black holes .
they conclude that they favor the accretion disk model .
this work is partially supported by the joint laboratory for optical astronomy of chinese academy of sciences , the national natural science foundation of china ( 10573005,10633010 ) , the 973 project ( no .
2007cb815405 ) , and science & technology commission of shanghai municipality ( 06dz22101 ) .
we thank martin gaskell for editing the english and for comments .
we are grateful to the referee , dr . paul viita for his valuable comments .
alef , w. , et al .
1996 , a&a , 308 , 376 botte , v. , et al .
2004 , aj , 127 , 3168 cannon , d. r. , et al . 1971 ,
mnras , 152 , 79 cellone , s. a. , romero , g. e. , & combi , j. a. 2000 , aj , 119 , 1534 dietrich , m. , et al .
1998 , apjs , 115 , 185 dunn , j. p. , et al .
2006 , pasp , 118 , 572 eracleous , m. , et al .
1996 , apj , 459 , 89 fan , j.h . , liu , y. , yuan , y.h . ,
2007 , a&a , 462 , 547 fan , j.h . , liu , y. , yuan , y.h .
2006 , pasj , 58 , 797 fan , j. h. , 2005 , chja&a suppl . , 5 , 213 fan , j. h.
, lin r. g. , xie , g. z. , et al .
2002 , a&a , 381 , 1 fan , j.h . , 1999 ,
mnras , 308 , 1032 fan , j. h. , xie , g. z. , lin , r.g . , et al .
1997 , a&as , 125 , 525 foster , g. 1996 , aj , 111 , 541 foster , g. 1995 , aj , 109 , 1889 gaskell , c. m. 1996 , apj , 464 , l107 gaskell , c. m. 2006 , asp conf . ser . , 360 , 111 gliozzi , m. , sambruna , r. , & eracleous , m. 2003 , apj , 584 , 176 horiuchi , t. , & kato , s. 1990 , pasj , 42 , 661 inda , m. , et al .
1994 , apj , 420 , 143 jurkevich , i. , usher , p. d. , & shen , b. s. p. 1971 , ap&ss , 10 , 402 kidger , m. , takalo , l. , & sillanp@xmath4 , a. 1992 , a&a , 264 , 32 lainela , m. , et al . 1999 ,
apj , 521 , 561 leahy , j. p. , & perley , r. a. 1995 , mnras , 277 , 1097 leighly , k. m. , et al .
1997 , apj , 483 , 767 leighly , k. m. , & obrien , p. t. 1997 , apj , 481 , l15 liller , m. h. & liller , w. 1975 , apj , 199 , l133 lloyd , c. 1984 , mnras , 209 , 697 lynds , c. 1968 , aj , 73 , 888 mclure , r. j. , & dunlop , j. s. 2001 , mnras , 327 , 199 meyer , f. , & meyer - hofmeister , e. 1984 , a&a , 132 , 143 miller , h. r. 1975 , apj , 201 , l109 osterbrock , d. e. , koski , a. t. , & phillips , m. m. 1975 , apj , 197 , l41 penston , m. j. , penston , m. v. , & sandage , a. r. 1971 , pasp , 83 , 783 pica , a. j. , et al .
1980 , aj , 85 , 1442 pica , a. j. , et al .
1988 , aj , 96 , 1215 pollock , j. t. 1975 , apj , 198 , l53 prieto , m. a. 1997 , mnras , 284 , 627 prieto , m. a. , & kotilainen , j. k. 1997 , apj , 491 , l77 qian , b. c. , & tao , j. 2004 , pasp , 116 , 161 qian , b. c. , tao , j. , & fan , j. h. 2002 , aj , 123 , 678 sandage , a. 1966 , apj , 145 , 1 sandage , a. 1967 , apj , 150 , l177 sandage , a. 1973 , apj , 180 , 687 scott , r. l. , et al .
1976 , aj , 81 , 7 selmes , r. a. , et al . 1975 ,
mnras , 170 , 15 sergeev , s. g. , et al .
2002 , apj , 576 , 660 shapovalova , a. i. , et al .
2001 , a&a , 376 , 775 shen , b. s. p. , et al .
1972 , apj , 171 , 457 sillanp@xmath4 , a. , haarala , s. , valtonen , m. j. , sundelius , b. , & byrd , g. g. 1988 , apj , 325 , 628 tao , j. , qian , b. c. , & fan , j. h. 2004 , pasp , 116 , 634 wozniak , p. r. , et al .
1998 , mnras , 299 , 499 yee , h. k. c. , & oke , j. b. 1981 , apj , 248 , 472 zheng , w. 1996 , aj , 111 , 1498 zheng , w. , perez , e. , grandi , s. a. , & penston , m. v. 1995 , aj , 109 , 2355 cccc 783.3201 & 15.658 & 0.010 & -0.004 + 783.3259 & 15.503 & 0.014 & -0.016 + 783.3349 & 15.589 & 0.010 & 0.004 + 802.3073 & 15.696 & 0.023 & -0.055 + 802.3107 & 15.450 & 0.014 & 0.003 + 803.3255 & 15.576 & 0.028 & -0.055 + 845.1668 & 15.383 & 0.007 & 0.001 + 845.1734 & 15.594 & 0.004 & 0.000 + 924.0950 & 15.513 & 0.026 & 0.041 + 924.1245 & 15.598 & 0.019 & 0.028 + 935.0480 & 15.585 & 0.038 & -0.051 + 935.0582 & 15.539 & 0.013 & -0.001 + 935.0627 & 15.492 & 0.028 & -0.040 + 943.0259 & 15.470 & 0.015 & -0.025 + 943.0345 & 15.461 & 0.025 & -0.047 + 947.0802 & 15.330 & 0.025 & 0.049 + 947.0870 & 15.436 & 0.025 & 0.049 + 977.9790 & 15.310 & 0.024 & 0.043 + 977.9976 & 15.320 & 0.043 & 0.087 + 978.0020 & 15.322 & 0.053 & 0.072 + 1196.2712 & 15.114 & 0.093 & 0.031 + 1196.2934 & 15.332 & 0.032 & 0.047 + 1197.2786 & 15.303 & 0.048 & -0.053 + 1197.2850 & 15.480 & 0.041 & -0.001 + 1199.2778 & 15.445 & 0.036 & 0.032 + 1199.2833 & 15.146 & 0.008 & 0.012 + 1221.1243 & 15.050 & 0.035 & -0.047 + 1257.0869 & 15.328 & 0.044 & -0.023 + 1257.0953 & 15.157 & 0.063 & 0.053 + 1328.0242 & 15.119 & 0.053 & -0.075 + 1328.0316 & 15.382 & 0.047 & 0.047 + 1328.0402 & 15.047 & 0.073 & -0.020 + 1334.0357 & 15.046 & 0.040 & -0.014 + 1334.0427 & 15.056 & 0.035 & -0.048 + 1371.9932 & 15.681 & 0.043 & 0.024 + 1372.0024 & 15.367 & 0.081 & 0.060 + 1550.3632 & 15.498 & 0.038 & 0.081 + 1577.2818 & 15.386 & 0.018 & -0.040 + 1577.2894 & 15.408 & 0.041 & -0.072 + 1577.2964 & 15.521 & 0.040 & -0.081 + 1577.3030 & 15.579 & 0.041 & -0.068 + 1586.2854 & 15.883 & 0.009 & 0.001 + 1586.3046 & 15.602 & 0.021 & -0.005 + 1594.2563 & 15.361 & 0.059 & -0.073 + 1594.2604 & 15.388 & 0.029 & 0.055 + 1594.2646 & 15.645 & 0.026 & -0.052 + 1599.1615 & 15.570 & 0.017 & 0.019 + 1614.2056 & 15.515 & 0.035 & -0.080 + 1614.2110 & 15.591 & 0.032 & 0.057 + 1614.2210 & 15.427 & 0.098 & -0.034 + 1616.1638 & 15.764 & 0.027 & -0.002 + 1616.1776 & 15.810 & 0.031 & -0.015 + 1660.1593 & 15.590 & 0.074 & -0.077 + 1660.1655 & 15.823 & 0.053 & 0.095 + 1692.0016 & 15.633 & 0.030 & -0.017 + 1692.0042 & 15.741 & 0.024 & -0.031 + 1692.0264 & 15.562 & 0.035 & -0.061 + 1692.0333 & 15.432 & 0.024 & 0.019 + 1701.0160 & 15.731 & 0.024 & -0.032 + 1701.0196 & 15.945 & 0.019 & -0.011 + 2229.3530 & 15.518 & 0.027 & 0.004 + 2318.3192 & 15.853 & 0.045 & -0.003 + 2348.1668 & 16.380 & 0.090 & 0.087 + 2348.1746 & 16.260 & 0.081 & 0.010 + 2437.0529 & 16.395 & 0.078 & 0.110 + 3041.2721 & 15.430 & 0.083 & -0.118 + 3041.2866 & 15.673 & 0.087 & -0.124 + 3094.2237 & 15.272 & 0.065 & 0.092 + 3094.2386 & 15.306 & 0.103 & 0.054 + 3100.1232 & 15.489 & 0.042 & 0.001 + 3112.3400 & 15.581 & 0.093 & 0.132 + 3518.0847 & 15.888 & 0.069 & -0.023 + 3856.0297 & 15.242 & 0.063 & -0.045 + 3856.0446 & 15.170 & 0.059 & -0.083 + 3888.0554 & 15.023 & 0.068 & 0.096 + 3888.0695 & 15.733 & 0.069 & 0.097 + 4231.7777 & 16.060 & 0.060 & -0.063 + cccc 802.2993 & 15.206 & 0.052 & 0.073 + 802.3039 & 15.322 & 0.038 & -0.025 + 803.3159 & 15.270 & 0.018 & -0.003 + 819.0507 & 15.265 & 0.015 & 0.034 + 824.1183 & 15.234 & 0.040 & 0.049 + 845.1945 & 15.111 & 0.014 & 0.019 + 919.0410 & 15.305 & 0.028 & 0.052 + 924.0838 & 15.288 & 0.022 & 0.027 + 935.0480 & 15.149 & 0.015 & 0.027 + 943.0036 & 15.099 & 0.014 & -0.005 + 943.0098 & 15.098 & 0.014 & -0.004 + 947.0513 & 15.191 & 0.037 & 0.032 + 947.0587 & 15.154 & 0.035 & 0.066 + 977.9743 & 15.027 & 0.003 & -0.004 + 977.9790 & 15.033 & 0.007 & -0.010 + 977.9834 & 15.023 & 0.020 & 0.013 + 1196.2656 & 14.660 & 0.021 & 0.021 + 1196.2876 & 14.608 & 0.040 & 0.021 + 1197.2695 & 14.698 & 0.019 & 0.003 + 1197.2740 & 14.661 & 0.024 & -0.029 + 1199.2660 & 14.820 & 0.035 & 0.041 + 1199.2722 & 14.775 & 0.044 & 0.056 + 1221.0993 & 14.769 & 0.024 & 0.035 + 1257.0738 & 14.880 & 0.054 & 0.034 + 1257.0799 & 14.871 & 0.045 & -0.052 + 1328.0826 & 14.352 & 0.019 & 0.028 + 1328.0873 & 14.402 & 0.082 & 0.048 + 1328.0929 & 14.504 & 0.060 & 0.003 + 1332.0358 & 14.944 & 0.067 & -0.095 + 1371.9859 & 14.656 & 0.062 & 0.030 + 1371.9889 & 14.718 & 0.086 & 0.031 + 1550.3524 & 15.144 & 0.012 & -0.027 + 1550.3569 & 15.083 & 0.020 & 0.028 + 1577.3114 & 15.083 & 0.045 & 0.056 + 1577.3167 & 15.029 & 0.039 & 0.061 + 1577.3217 & 15.078 & 0.034 & 0.052 + 1577.3267 & 14.978 & 0.040 & 0.059 + 1577.3330 & 14.990 & 0.039 & 0.062 + 1586.2576 & 14.828 & 0.046 & 0.069 + 1586.2617 & 15.274 & 0.032 & 0.038 + 1586.2704 & 15.169 & 0.020 & 0.007 + 1586.2747 & 15.171 & 0.023 & 0.022 + 1586.2793 & 15.213 & 0.042 & 0.069 + 1594.2340 & 15.122 & 0.042 & 0.065 + 1594.2422 & 14.994 & 0.024 & -0.012 + 1594.2447 & 15.196 & 0.024 & 0.019 + 1599.1528 & 15.192 & 0.047 & 0.060 + 1614.2360 & 15.219 & 0.015 & 0.019 + 1614.2418 & 15.191 & 0.016 & 0.016 + 1616.1541 & 15.138 & 0.020 & 0.029 + 1616.1572 & 15.211 & 0.015 & 0.021 + 1616.1600 & 15.271 & 0.030 & 0.056 + 1660.1492 & 15.339 & 0.034 & 0.008 + 1660.1524 & 15.210 & 0.021 & 0.012 + 1691.9937 & 15.205 & 0.014 & -0.015 + 1691.9959 & 15.198 & 0.012 & 0.020 + 1691.9981 & 15.217 & 0.018 & -0.005 + 1692.0182 & 15.181 & 0.014 & 0.027 + 1692.0209 & 15.170 & 0.027 & 0.032 + 1692.0234 & 15.223 & 0.030 & 0.005 + 1699.1292 & 15.067 & 0.029 & 0.009 + 1699.1339 & 15.181 & 0.034 & 0.045 + 1701.0020 & 15.317 & 0.029 & 0.024 + 1701.0053 & 15.301 & 0.007 & 0.013 + 1701.0087 & 15.373 & 0.006 & 0.003 + 1701.0122 & 15.380 & 0.029 & -0.042 + 2229.3450 & 15.099 & 0.034 & 0.018 + 2318.3073 & 14.846 & 0.034 & 0.049 + 2348.1569 & 15.211 & 0.037 & 0.053 + 2348.1621 & 15.198 & 0.044 & 0.062 + 2437.0420 & 15.273 & 0.049 & 0.070 + 2757.3133 & 15.010 & 0.057 & 0.080 + 3017.2940 & 14.905 & 0.045 & -0.051 + 3017.3345 & 15.014 & 0.049 & -0.070 + 3018.3446 & 14.897 & 0.035 & -0.049 + 3041.2560 & 14.911 & 0.071 & 0.062 + 3041.2636 & 14.955 & 0.038 & 0.078 + 3091.1142 & 14.839 & 0.087 & 0.022 + 3091.1253 & 15.106 & 0.066 & 0.076 + 3093.2495 & 14.950 & 0.057 & 0.026 + 3094.2063 & 14.977 & 0.036 & 0.027 + 3094.2139 & 14.991 & 0.052 & -0.024 + 3100.1131 & 15.085 & 0.082 & -0.023 + 3108.1328 & 15.135 & 0.071 & -0.078 + 3108.1425 & 15.108 & 0.005 & 0.002 + 3112.3128 & 15.127 & 0.063 & -0.040 + 3124.1414 & 15.023 & 0.021 & 0.012 + 3124.1552 & 15.033 & 0.079 & 0.010 + 3518.0390 & 15.211 & 0.119 & 0.168 + 3518.0766 & 15.329 & 0.099 & -0.140 + 3519.0071 & 14.841 & 0.047 & 0.066 + 3519.0510 & 14.828 & 0.068 & 0.096 + 3856.0247 & 15.130 & 0.016 & 0.023 + 3856.0399 & 15.000 & 0.028 & 0.040 + 3873.9964 & 15.075 & 0.052 & 0.074 + 3887.9728 & 14.451 & 0.069 & -0.098 + 3887.9856 & 15.090 & 0.144 & -0.204 + 3887.9983 & 14.409 & 0.053 & 0.045 + 3888.0113 & 14.985 & 0.052 & 0.074 + 4231.7682 & 15.592 & 0.098 & -0.139 + cccc 783.3399 & 14.635 & 0.019 & -0.036 + 783.3429 & 14.627 & 0.028 & -0.008 + 783.3463 & 14.719 & 0.021 & -0.038 + 783.3556 & 14.696 & 0.022 & -0.044 + 802.3145 & 14.470 & 0.014 & 0.002 + 803.3013 & 14.600 & 0.019 & -0.029 + 803.3078 & 14.551 & 0.019 & -0.023 + 845.2020 & 14.532 & 0.023 & 0.004 + 845.2187 & 14.575 & 0.028 & 0.017 + 845.2234 & 14.566 & 0.030 & 0.030 + 919.0351 & 14.645 & 0.041 & 0.023 + 919.0451 & 14.629 & 0.038 & -0.031 + 924.0796 & 14.630 & 0.045 & -0.025 + 924.1049 & 14.690 & 0.056 & 0.020 + 934.9925 & 14.567 & 0.028 & -0.043 + 935.0350 & 14.505 & 0.040 & -0.076 + 935.0395 & 14.491 & 0.041 & -0.036 + 942.9844 & 14.568 & 0.022 & -0.018 + 942.9920 & 14.536 & 0.015 & -0.017 + 947.0385 & 14.547 & 0.065 & -0.001 + 947.0445 & 14.623 & 0.066 & -0.001 + 961.9682 & 14.409 & 0.035 & -0.017 + 961.9741 & 14.452 & 0.054 & -0.032 + 961.9825 & 14.431 & 0.037 & -0.012 + 977.9529 & 14.374 & 0.041 & 0.003 + 977.9578 & 14.432 & 0.039 & -0.032 + 977.9626 & 14.389 & 0.092 & -0.184 + 1196.2602 & 14.105 & 0.023 & -0.013 + 1196.2785 & 14.096 & 0.039 & -0.092 + 1197.2588 & 14.081 & 0.033 & 0.036 + 1197.2646 & 14.223 & 0.032 & 0.044 + 1199.2347 & 14.287 & 0.016 & -0.004 + 1199.2490 & 14.219 & 0.025 & 0.000 + 1199.2556 & 14.220 & 0.022 & -0.026 + 1221.0931 & 14.098 & 0.028 & 0.034 + 1257.0579 & 14.292 & 0.019 & -0.007 + 1257.0645 & 14.238 & 0.031 & 0.004 + 1371.9801 & 14.117 & 0.028 & -0.042 + 1550.3434 & 14.394 & 0.018 & 0.026 + 1550.3469 & 14.346 & 0.037 & 0.053 + 1577.2501 & 14.452 & 0.009 & 0.004 + 1577.2542 & 14.439 & 0.022 & 0.026 + 1577.2606 & 14.537 & 0.024 & 0.004 + 1577.2678 & 14.491 & 0.024 & -0.012 + 1577.2715 & 14.469 & 0.025 & 0.013 + 1578.2352 & 14.378 & 0.040 & 0.035 + 1578.2389 & 14.553 & 0.065 & 0.096 + 1578.2842 & 14.512 & 0.047 & 0.019 + 1578.2877 & 14.503 & 0.069 & 0.075 + 1586.3170 & 14.578 & 0.014 & -0.021 + 1586.3204 & 14.586 & 0.025 & 0.016 + 1586.3255 & 14.575 & 0.027 & 0.035 + 1594.2498 & 14.530 & 0.030 & 0.041 + 1594.2519 & 14.508 & 0.021 & 0.011 + 1594.2536 & 14.519 & 0.036 & 0.040 + 1599.1410 & 14.556 & 0.040 & 0.039 + 1599.1462 & 14.559 & 0.030 & 0.048 + 1614.1925 & 14.625 & 0.040 & -0.085 + 1614.1991 & 14.602 & 0.031 & -0.065 + 1614.2020 & 14.578 & 0.032 & -0.062 + 1616.1458 & 14.620 & 0.013 & 0.011 + 1616.1483 & 14.536 & 0.092 & 0.015 + 1616.1509 & 14.528 & 0.091 & 0.033 + 1660.1375 & 14.905 & 0.041 & -0.027 + 1660.1424 & 14.353 & 0.060 & 0.011 + 1660.1455 & 14.488 & 0.027 & -0.013 + 1691.9837 & 14.575 & 0.028 & -0.058 + 1691.9864 & 14.474 & 0.020 & -0.007 + 1691.9895 & 14.543 & 0.033 & -0.073 + 1692.0115 & 14.525 & 0.042 & 0.034 + 1692.0139 & 14.425 & 0.021 & 0.030 + 1692.0157 & 14.386 & 0.015 & -0.006 + 1699.1103 & 14.432 & 0.009 & 0.009 + 1699.1143 & 14.562 & 0.013 & 0.017 + 1699.1182 & 14.510 & 0.013 & 0.009 + 1699.1234 & 14.268 & 0.035 & -0.070 + 1700.9910 & 14.760 & 0.004 & 0.007 + 1700.9941 & 14.630 & 0.014 & -0.027 + 1700.9972 & 14.784 & 0.035 & -0.055 + 1700.9992 & 14.838 & 0.015 & -0.010 + 2088.9659 & 14.368 & 0.029 & -0.051 + 2229.3354 & 14.606 & 0.011 & 0.000 + 2229.3390 & 14.587 & 0.003 & 0.001 + 2318.2813 & 14.472 & 0.038 & 0.037 + 2318.2947 & 14.514 & 0.049 & 0.028 + 2348.1474 & 14.723 & 0.023 & 0.012 + 2348.1517 & 14.712 & 0.025 & 0.035 + 2437.0209 & 14.834 & 0.033 & 0.024 + 2437.0278 & 14.679 & 0.008 & 0.012 + 2437.0355 & 14.697 & 0.074 & 0.066 + 2633.3729 & 14.730 & 0.057 & -0.088 + 2696.0850 & 14.778 & 0.100 & -0.018 + 2696.1289 & 14.946 & 0.004 & -0.006 + 2701.2598 & 14.527 & 0.013 & -0.018 + 2701.2719 & 14.586 & 0.019 & 0.006 + 2701.3009 & 14.572 & 0.028 & -0.052 + 2702.2897 & 14.803 & 0.017 & -0.015 + 2708.2299 & 14.678 & 0.039 & -0.029 + 2708.2428 & 14.420 & 0.034 & -0.048 + 2708.2686 & 14.429 & 0.082 & 0.035 + 2708.2820 & 14.356 & 0.020 & -0.029 + 2741.2177 & 14.580 & 0.078 & 0.019 + 2742.0505 & 14.651 & 0.048 & -0.068 + 2742.0943 & 14.646 & 0.070 & -0.099 + 2743.0436 & 14.595 & 0.030 & -0.006 + 2743.0924 & 14.711 & 0.081 & -0.019 + 2743.1037 & 14.457 & 0.094 & -0.022 + 2746.0535 & 14.469 & 0.062 & 0.049 + 2746.0703 & 14.712 & 0.069 & -0.092 + 2747.2671 & 14.472 & 0.046 & -0.024 + 2752.2818 & 14.398 & 0.056 & -0.010 + 2752.3235 & 14.309 & 0.055 & 0.066 + 2752.3354 & 14.496 & 0.074 & -0.041 + 2757.2584 & 14.241 & 0.027 & -0.038 + 2757.2833 & 14.270 & 0.006 & -0.009 + 3017.2877 & 14.378 & 0.104 & 0.147 + 3017.3431 & 14.387 & 0.053 & 0.042 + 3018.3268 & 14.355 & 0.090 & -0.036 + 3041.2464 & 14.469 & 0.024 & -0.016 + 3041.2504 & 14.440 & 0.013 & 0.019 + 3091.1076 & 14.415 & 0.036 & 0.053 + 3091.1077 & 14.458 & 0.029 & 0.041 + 3091.1006 & 14.385 & 0.055 & 0.070 + 3091.1077 & 14.537 & 0.045 & 0.011 + 3093.2401 & 14.535 & 0.012 & 0.000 + 3094.1951 & 14.413 & 0.027 & 0.049 + 3094.1994 & 14.440 & 0.035 & 0.024 + 3100.1083 & 14.745 & 0.028 & -0.054 + 3108.1097 & 14.459 & 0.098 & -0.081 + 3108.1206 & 14.575 & 0.064 & -0.013 + 3112.3084 & 14.416 & 0.065 & 0.004 + 3124.1235 & 14.515 & 0.056 & -0.041 + 3124.1323 & 14.510 & 0.039 & -0.042 + 3760.2646 & 14.261 & 0.013 & 0.025 + 3760.2955 & 14.570 & 0.042 & 0.060 + 3795.2234 & 13.896 & 0.047 & -0.067 + 3795.2569 & 13.978 & 0.074 & -0.104 + 3855.1142 & 14.204 & 0.008 & 0.012 + 3855.1228 & 14.499 & 0.068 & 0.097 + 3856.0373 & 14.568 & 0.063 & -0.089 + 3856.0520 & 14.471 & 0.085 & -0.121 + 3857.0662 & 14.425 & 0.000 & 0.000 + 3874.9808 & 14.190 & 0.039 & -0.055 + 3886.0060 & 14.286 & 0.188 & -0.267 + 3886.0105 & 14.106 & 0.016 & -0.022 + 3886.0435 & 14.180 & 0.060 & -0.085 + 3887.9957 & 14.200 & 0.066 & 0.094 + 3888.0086 & 14.187 & 0.127 & -0.180 + 3888.0216 & 13.987 & 0.078 & 0.110 + 4231.7819 & 15.106 & 0.021 & 0.030 + 4231.8011 & 15.180 & 0.042 & -0.059 + llrcperiod & amplitude & phase & fap + ( d ) & & ( @xmath70 ) & + + @xmath71 & @xmath72 & @xmath73 & @xmath74 + @xmath75 & @xmath76 & @xmath77 & @xmath78 + @xmath79 & @xmath80 & @xmath81 & @xmath82 + @xmath83 & @xmath84 & @xmath85 & @xmath82 + @xmath86 & @xmath87 & @xmath88 & @xmath78 + @xmath89 & @xmath90 & @xmath91 & @xmath78 + @xmath92 & @xmath93 & @xmath94 & @xmath95 + + @xmath50 & @xmath96 & @xmath97 & @xmath98 + @xmath99 & @xmath100 & @xmath101 & @xmath98 + @xmath102 & @xmath103 & @xmath104 & @xmath105 + @xmath106 & @xmath107 & @xmath108 & @xmath82 + @xmath109 & @xmath110 & @xmath111 & @xmath82 + @xmath112 & @xmath113 & @xmath114 & @xmath78 + @xmath115 & @xmath116 & @xmath117 & @xmath118 + + @xmath51 & @xmath119 & @xmath120 & @xmath74 + @xmath121 & @xmath122 & @xmath123 & @xmath95 + @xmath124 & @xmath125 & @xmath126 & @xmath95 + @xmath127 & @xmath128 & @xmath129 & @xmath82 + @xmath130 & @xmath131 & @xmath132 & @xmath95 + @xmath133 & @xmath134 & @xmath135 & @xmath82 + @xmath136 & @xmath137 & @xmath138 & @xmath95 + + @xmath139 & @xmath140 & @xmath141 & @xmath74 + @xmath142 & @xmath143 & @xmath144 & @xmath145 + @xmath146 & @xmath147 & @xmath148 & @xmath78 + @xmath149 & @xmath150 & @xmath151 & @xmath145 + @xmath86 & @xmath152 & @xmath153 & @xmath145 + @xmath154 & @xmath155 & @xmath156 & @xmath105 + @xmath157 & @xmath158 & @xmath159 & @xmath105 + | we report v , r , and i band ccd photometry of the radio galaxy 3c 390.3 obtained with the 1.56-m telescope of the shanghai astronomical observatory from march 1995 to august 2004 . combining these data with data from the literature ,
we have constructed a historical light curve from 1894 to 2004 and searched for periodicities using the _ cleanest _ program .
we find possible periods of @xmath0 , @xmath1 , @xmath2 , and @xmath3 years . |
the baryonic acoustic oscillations ( bao ) constitute a characteristic feature within the large scale structure of the universe and can serve as standard ruler for constraining the properties of dark energy ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the bao in the baryon - photon fluid of the pre - recombination era imprint the sound horizon distance at decoupling as a typical scale in the matter correlation function or power spectrum @xcite .
these oscillations were detected in the cosmic microwave background ( e.g. , * ? ? ? * ) and in the spatial distribution of galaxies @xcite and have been confirmed by a number of subsequent studies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . the next generation of large galaxy surveys , like the panoramic survey telescope & rapid response system ( pan - starrs , @xcite ) , the dark energy survey ( des , @xcite ) , the baryonic oscillation spectroscopic survey ( boss , @xcite ) , bigboss @xcite , the hobby eberly telescope dark energy experiment ( hetdex , @xcite ) and the space based euclid mission @xcite , will cover volumes much larger than current datasets , allowing for much more accurate determinations of the bao . in the two - point correlation function
the bao are visible as a unique broad and quasi gaussian peak @xcite .
however , the determination of the shape and location of the peak may be affected by sample variance @xcite , non linear effects and the bias of the tracer galaxy population .
these difficulties lead @xcite to suggest to use the zero - crossing of the two point correlation located at @xmath3 as standard ruler instead of the peak location . yet
, several observations do not show the theoretically predicted zero - crossing at all @xcite . at this stage
it is unclear whether this discrepancy is caused by systematic effects or cosmic variance or whether it represents a challenge for the concordance @xmath4cdm model @xcite .
one basic assumption for interpretation of the bao measurements is that the matter distribution is isotropic at the relevant scales ( @xmath1 ) . in this work we use alignment correlation functions @xcite to show that the amplitudes of the two - point correlation function measured along the orientations of the high density peaks are larger than those derived from spherically averaged ( conventional ) clustering analysis out to scales of @xmath5 and discuss possible effects on measurements of the bao .
we use a set of large scale dark matter simulations which follow the evolution of the cosmic density field including its collapse into high density peaks . based on their orientations
we compute the alignment cross correlation function with the over all mass distribution out to @xmath6 . in this section
we briefly discuss the simulations , the computation of density peak orientations and the definition of the alignment correlation function .
we used gadget-2 @xcite to carry out 50 @xmath7-simulations based on the same concordant @xmath4cdm cosmology but different realizations of the initial density field .
the total volume of the ensemble is almost 70 times larger than the volume used for the detection of the bao signal by @xcite .
the cosmological parameters correspond to those used for the _ millennium simulation _
@xcite : matter density parameter , @xmath8 ; cosmological constant term , @xmath9 ; power spectrum normalization , @xmath10 ; spectral slope , @xmath11 ; and hubble parameter , @xmath12 . for each run
@xmath13 particles are used resulting in a mass resolution of @xmath14 for the dark matter particles .
a plummer - equivalent softening of @xmath15 was employed .
an analysis based on a much lower force resolution showed very similar results .
density peaks are identified by using a friends - of - friends ( fof , * ? ? ?
* ) group - finder with linking parameter @xmath16 times the mean particle separation . in each realization
there are about 30,000 fof - groups with masses greater than @xmath17 .
the orientations are derived by evaluating the eigenvalues of the second moment of the mass tensor . in the following the direction of principal axis associated with the largest eigenvalue
is referred to as the ( 3d ) orientation of the density peak .
observationally 3d orientations are difficult to determine , therefore , we also use projected orientations .
these are determined by collapsing the mass distribution of the density peak along the line of sight ( here equivalent to the z - axis ) and , consecutively , computing the eigenvectors of inertia tensor for the two dimensional mass distribution . despite the poor force resolution
the orientations of the density peaks are expected to trace the large scale tidal field well ( see below ) .
two - point correlation functions , @xmath18 , are a primary tool for quantifying the clustering properties of cosmological structures ( e.g. * ? ? ?
* ) .we employ an extension of the _ two - point cross - correlation function _ which can be used to quantify spatial alignment of objects in one sample with respect to orientations assigned to the objects in the other sample ( see , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
given a sample of objects in question ( sample @xmath19 ) , a sample of reference objects ( sample @xmath20 ) and a random sample ( sample @xmath21 ) we can define the _ alignment two point cross correlation function _
@xmath22 where @xmath23 indicates the distance , @xmath24 , is the angle between the orientation of the objects in sample @xmath19 and the line connecting to objects in either of the two other samples . @xmath25 and @xmath26 are the number of points contained in the reference and random sample , @xmath27 and @xmath28 are the counts of cross pairs between the indicated samples for given @xmath24 and @xmath23 .
the conventional correlation function is the average of the alignment correlation function over the full range of @xmath24 values .
taking symmetries into account the value of @xmath24 ranges from zero ( along the orientation of the main object ) to 90 degrees ( perpendicular to its orientation ) .
thus , higher amplitudes of the alignment correlation functions at small values of @xmath24 indicate that the reference objects are more likely to be aligned along orientation of the main objects . in contrast
, higher amplitudes at larger angles indicate that the reference objects are more likely to be located perpendicular to the orientation of the main objects .
we employ the alignment correlation functions , more exactly , the average of the alignment correlation function within given angular ranges , to quantify the large - scale alignment between the matter distribution in the universe and the orientation of high density peaks ( @xmath29 ) .
the reference samples are generated from 10% random subsets of the overall particle distributions .
if real - space correlations are considered we use the three dimensional orientation and separation vector to determine @xmath24 . alternatively ,
if observational consequences are discussed we use projected orientations but still 3d distances since the effects of reshift - space distortions are small .
in this section we present our findings for the alignment correlation functions in real- and redshift - space and for redshifts z=0 and 0.6 . fig . [
fig : z0 ] shows the conventional and alignment two point cross correlation functions ( upper panels ) and the residuals of the alignment correlation functions about the conventional correlation function ( lower panels ) employing different lower mass cuts ( @xmath30 , @xmath31 and @xmath32 ) for the high density peaks in the main sample .
solid lines show the conventional correlation functions . dotted and dashed lines represent the correlation functions along and perpendicular to the orientations of the density peaks .
the gray regions represent the cosmic variance .
interestingly , the cosmic variance in the lower panels is slightly smaller , this is because the differences between conventional and alignment correlation functions are independent of their absolute values . to check for systematic or numerical errors
we have repeated the above analysis with randomly interchanged orientations within each realization . in this case
we do not detect any significant difference between the conventional and alignment correlation functions . the left panel of fig .
[ fig : z06 ] displays conventional and alignment correlation functions for density peaks with masses larger than @xmath33 at @xmath34 .
the right panel of fig .
[ fig : z06 ] represents the correlation functions based on projected orientations . here , the angles used to select the pairs along and perpendicular to the orientations of the density peaks are computed between the projected orientations and the projected distance vectors .
distances are computed in redshift space using the distant observer approximation . in this case the differences in the amplitudes between the conventional and the alignment correlation functions are somewhat reduced but still significant beyond the bao scale . in conclusion ,
the alignment correlation function between dark matter density peaks with masses @xmath35 and the overall matter distribution reveals anisotropies at scales larger than @xmath36 far beyond the bao scale .
the difference between conventional correlation functions and correlation functions measured along and perpendicular to the orientations of the peaks is most pronounced for the most massive peaks in real - space at @xmath37 .
but it remains visible at earlier time and also if projected orientations are used .
the findings presented here are based on n - body simulations however qualitative similar effects are expected to shape the observed galaxy correlation functions at scales up to @xmath36 .
potential observational implications are : + @xmath38 _ large - scale anisotropy and structure of two point correlation functions : _ our results indicate that matter is distributed anisotropically out to separations far larger then the bao scale . at any given pair separation
the conventional correlation function is the average of the alignment correlation function over the whole range of @xmath24 .
we find that the alignment correlation function measured along high density peaks does not fall below zero ( for separations @xmath39 ) which is counterbalanced by negative amplitudes at much smaller separations for measurements perpendicular to it . at the bao scale galaxy
two point correlation functions can be interpreted as the average of a more highly clustered component along the direction of high density peaks and a less clustered component perpendicular to it .
+ @xmath38 _ shape and location of the bao peak : _ at bao scales the amplitudes of the correlation functions between high density peaks and the overall matter distribution are significantly higher if measured along the orientation of the peaks . in this case
the bao peak is composed of a hump on top of the declining but still positive correlation function . the signal perpendicular to the orientations is dominated by the bao hump itself with negligible underlying clustering signal .
the location of the bao peak is found at somewhat smaller ( larger ) scales if measured parallel ( perpendicular ) to the orientation of high density peaks .
for the measurements along the density peak orientations the bao hump is transformed into a plateau - like feature .
if the survey volume is dominated by large filamentary structures the shape of the conventional correlation function should be close to that found for the parallel signal shown here .
this may relate to a study by @xcite who find no apparent peak in 6% of their mock samples .
+ @xmath38 _ zero crossing and large scale power : _ several publications report the non - detection of zero - crossing out to scales of @xmath40 which indicates ( unexpected ) large scale power ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we find similar results for the correlation functions measured along the orientations of high density peaks .
furthermore , the alignment cross correlation function measured along the orientations of high density peaks at @xmath34 shows zero - crossing in contrast to the behavior at @xmath37 .
the wigglez redshift - space correlation function at @xmath34 @xcite shows a crossover as well .
whether these analogies are coincidental remains to be explored in future work .
+ @xmath38 _ direct measurement of anisotropy : _ with the advent of enormous cluster catalogs ( e.g. , * ? ? ?
* ; * ? ? ?
* ) it should in principle be possible to directly measure the large scale anisotropies with alignment correlation functions if cluster orientations can be determined with sufficient accuracy . an anisotropy signal
may even be extracted simply by using orientations of the cluster central luminous red galaxies because the orientations of central galaxies and host systems are correlated @xcite .
+ @xmath38 _ improvement of bao measurements : _ if the directional effects reported here are observable it would be worthwhile considering to measure the bao peak perpendicular to the orientations of galaxy clusters ( or luminous red galaxies ) since in this direction the bao peak is better confined .
we would like to thank the anonymous referee for very helpful comments .
this work is sponsored by nsfc ( no . 11173045 ) , shanghai pujiang program ( no .
11pj1411600 ) and the cas / safea international partnership program for creative research teams ( kjcx2-yw - t23 ) .
af acknowledges support from the south african ska project .
the simulations used for this paper were performed on the blade centre cluster of the computing center of the max - planck - society in garching , and icc cosmology machine cosma4 , which is part of the dirac facility jointly funded by stfc , the large facilities capital fund of bis , and durham university . , n. , aussel , h. , burke , b. e. , boesgaard , h. , chambers , k. , chun , m. r. , heasley , j. n. , hodapp , k .- w .
, & et al . , .
2002 , in society of photo - optical instrumentation engineers ( spie ) conference series , ed .
j. a. t. .
s. wolff , vol .
4836 , 154164 | tracing the cosmic evolution of the baryonic acoustic oscillation ( bao ) scale with galaxy two point correlation functions is currently the most promising approach to detect dark energy at early times .
a number of ongoing and future experiments will measure the bao peak with unprecedented accuracy .
we show based on a set of n - body simulations that the matter distribution is anisotropic out to @xmath0 , far beyond the bao scale of @xmath1 , and discuss implications for the measurement of the bao . to that purpose
we use alignment correlation functions , i.e. , cross correlation functions between high density peaks and the overall matter distribution measured along the orientation of the peaks and perpendicular to it .
the correlation function measured along ( perpendicular to ) the orientation of high density peaks is enhanced ( reduced ) by a factor of @xmath2 compared to the conventional correlation function and the location of the bao peak shifts towards smaller ( larger ) scales if measured along ( perpendicular to ) the orientation of the high density peaks .
similar effects are expected to shape observed galaxy correlation functions at bao scales . |
naturally - grown high - quality cu@xmath1o is a good candidate for studying excitonic bose - einstein condensation .
first , an inversion symmetry in cu@xmath1o forbids any direct dipole recombination of excitons , which gives the exciton a long lifetime up to microseconds .
second , cu@xmath1o is a direct - gap semiconductor with repulsive exciton - exciton interactions , which means the electron - hole liquid ( ehl ) is not stable @xcite .
moreover , the large binding energy of the excitons ( 150 mev ) is equivalent to a temperature of 1740 kelvin , which allows excitons to exist at room temperature @xcite .
the lowest excitonic state is the 1s-yellow " exciton series , consisting of electrons in the @xmath2 level of the conduction band and holes in the @xmath3 level of the valence band with 2.173 ev gap energy . by electron - hole exchange ,
the yellow " excitons split into a triplet orthoexciton and a singlet paraexciton , which lies 12 mev lower @xcite .
most studies @xcite have concentrated on the higher - lying orthoexciton state since it can be easily created by a single - photon phonon - assisted absorption . in the single - photon phonon - assisted excitation process , however , a hot optical phonon is created which can heat the exciton gas , thereby deterring bec . in the past few years , several groups @xcite have tried resonant two - photon excitation of the orthoexciton state .
a major advantage is that the scattered two - photon laser light will not damage imaging systems because it is far away from the exciton luminescence lines , which allows a measurement during the laser pulse .
it is also expected that two colliding pulses in opposite directions will create excitons directly in the @xmath4 state , where a condensate should appear , instead of in a state with finite momentum .
resonant excitation of orthoexcitons still has the problem that the orthoexcitons will convert down into paraexcitons by a phonon emission @xcite , and the emitted phonon will cause a local heating , which may prevent bec .
our present work aims to solve the above problem with a resonant two - photon excitation of the yellow paraexciton state , because it is the lowest exciton state and primarily only decays by direct single - photon emission .
we create excitons in a harmonic potential trap made with an external stressor @xcite , which confines excitons in a defined volume and changes the local symmetry of cu@xmath1o from o@xmath5 to d@xmath6 .
one problem which becomes more important in the potential trap is the auger recombination process , which happens when two excitons collide and one exciton ionizes , taking the energy of the other exciton that recombined .
therefore the auger process not only gives a severe limit in exciton density but also heats up the exciton gas , which will tend to prevent bec .
the auger recombination rate increases roughly as the square of the stress , with much smaller auger recombination at the stress lower than 1.5 kbar @xcite .
however , because the phonon - assisted orthoexciton and direct paraexciton luminescence lines are separated at high stress but substantially overlapped at low stress @xcite , we have to use high stress to separate the luminescence lines . in this work ,
the inhomogeneous applied stress is kept at 1.9 kbar , a trade - off point which is good enough to spectrally separate the paraexciton line and provides a relatively small auger process . in this paper , we present experimental results for two experimental configurations , with one ir laser beam and with two colliding ir laser pulses .
we demonstrate the resonant creation of paraexcitons under two - photon excitations with several strong evidences .
then we propose a model of the excitation mechanism for the paraexciton and orthoexciton in the two - photon excitations , based on our study on selection rules and polarization dependence of the yellow excitons from group theory .
the cu@xmath1o sample , a @xmath7 mm bulk crystal , is immersed in the liquid helium in a janis optical cryostat , at a temperature controlled by a homemade pid temperature controller .
the lowest temperature in the cryostat is 1.6 kelvin . as shown in fig . 1 ,
in one - beam experiments the sample is oriented as follows : one [ 110 ] surface is illuminated by an ir laser beam and the exciton luminescence is collected from the perpendicular [ @xmath810 ] surface by an imaging lens which focuses the light onto the entrance slit of a @xmath9-meter spectrometer , with a choice of a ccd camera and a photo - multiplier tube ( pmt ) as detectors .
the exciton population dynamics are examined by means of time - correlated single photon detection .
we can also create excitons by splitting the ir laser beam into two counter - propagating parts which meet in the cu@xmath0o sample at the same time from opposite [ 110 ] directions , with a controlled time delay . the repetition rate ,
the duration and the spectral width of the ir laser are 250 khz , 200 fs and 20 nm , respectively . in order to compare
, we also do experiments under single - photon excitation with a cavity - dumped 3.8 mhz and 50 ps red dye - laser tuned to the exciton resonance for the same crystal orientation .
excitons are confined by a harmonic potential trap made with a 6 mm radius curved glass stressor applied along the [ 001 ] axis from the top surface of the cu@xmath0o crystal . under
the [ 001 ] stress , we find strong paraexciton signal , although naka and nagasawa @xcite reported that paraexcitons were too weak to be detected .
the main reason is that our femto - second laser system has much greater instantaneous intensity than their nano - second laser .
moreover , the exciton luminescence is collected from a direction perpendicular to the incident laser beam in this work , which has several advantages over their forward scattering geometry @xcite . for example
, the second harmonic scattering of the incident infrared laser light does not mix with the exciton luminescence .
when we create excitons with an ir laser , the first question which arises is how we can say it is not a three - photon or four - photon excitation , but a two - photon excitation .
one way to answer this question is from the dependence of the exciton luminescence intensity on the ir laser power , that is to say , the excitation is a n - photon process if the intensity of exciton luminescence is proportional to ( laser - power)@xmath10 .
we performed a series of experiments for a broad wavelength range of ir lasers , with an inhomogeneous stress of 1.9 kbar along the [ 001 ] direction and the paraexciton line position at 614.1 nm .
as shown in fig .
2 , our experimental results indicate that the excitation is a two - photon process when the ir laser is at the paraexciton resonant position , and it is a three - photon process under off - resonant ir laser , for example at 1240 nm .
all experiments in the following discussions are under resonant two - photon excitation of the paraexciton state .
we note that even at this wavelength , the three - photon process will be responsible for about 25% of the total exciton creation at the highest laser powers we use , and it will dominate at high excitation power , which may also create another barrier to bec . with
an ir laser tuned to one half the paraexciton ground - state energy , several experimental results convince us that we directly create paraexcitons , and orthoexcitons primarily come from the excitonic auger process . the first strong evidence is shown in fig .
3 , the time - integrated luminescence intensity as a function of ir laser power from 55 mw to 3 mw .
as mentioned above , under the 1.9 kbar stress along the [ 001 ] direction , the direct paraexciton line is at 614.1 nm . although paraexcitons primarily only decay by single - photon emission , there are two emission lines for orthoexcitons , as shown in fig .
3 , the direct orthoexciton luminescence line at 611.7 nm , which corresponds to emission of a single photon , and phonon - assisted orthoexciton luminescence line at 615.7 nm , which corresponds to emission of a single photon and a @xmath11 optical phonon with energy of 13.8 mev .
there is only paraexciton luminescence when the laser power is very low , indicating that the resonant ir laser creates the paraexciton directly .
orthoexcitons start to appear at higher laser power , around 5 mw as seen in fig . 3 , and become stronger and stronger relative to the paraexcitons with increasing the laser power , which implies that orthoexcitons are created not directly by the laser , but by a density - dependent process from paraexcitons . a second evidence is found in time - resolved paraexciton luminescence data
4 shows paraexciton luminescence intensity as a function of time after laser pulses in two different pulse - laser excitations . in one case , two - photon excitation resonant with the paraexciton state
is used , and in the second case , single - photon excitation tuned to the bottom of the orthoexciton phonon - assisted absorption is used .
the average powers of the ir laser and the red laser used in the excitations were adjusted to get the same total integrated exciton luminescence intensity .
these average powers were 70 mw and 0.9 mw , respectively .
there are several similarities between the two cases .
the paraexciton luminescence intensity in both cases reaches a steady - state value long after the laser pulses which indicates lifetimes of paraexcitons longer than the 260 ns period between the laser pulses , and the paraexciton luminescence intensities at time just before the laser pulses are comparable , equal to the steady - state value left from the previous laser pulse .
however , in the case of resonant two - photon excitation , the paraexcitons reach maximum intensity immediately after the laser pulse , consistent with the paraexcitons being created directly by the laser . in the case of single - photon excitation of orthoexcitons
, paraexcitons do not appear right after the laser pulse , but are slowly created in a time interval of more than 10 ns from a process which is known to be phonon - assisted orthoexciton down - conversion @xcite .
the rise time of 10 ns is longer than the reported orthoexciton - paraexciton conversion time of 4 ns , @xcite presumably because the paraexcitons must also cool down before they appear in the single - photon luminescence , which involves only states near the band bottom .
another strong evidence is the power dependence of time - resolved data for both paraexcitons and orthoexcitons under the two - photon excitation resonant with the paraexciton state .
5 shows two important things .
first , paraexcitons appear right after the laser pulse , but orthoexcitons are created slowly after a few nanoseconds .
second , orthoexcitons are created more slowly at lower ir laser power , which again indicates a density - dependent process for the creation of orthoexcitons .
all the above results lead us to a model in which paraexcitons are created directly by ir laser pulses resonant with the paraexciton state , then paraexcitons convert to orthoexcitons through a density - dependent process which may be the well - known excitonic auger process @xcite . in each auger process , two excitons collide and end up with one exciton recombining and the other exciton ionizing . because the orthoexciton is a triplet state and the paraexciton is a singlet state , and spin is randomly selected in ionization , 75 percent of the ionized excitons in the auger process will be returned as orthoexcitons and 25 percent of these excitons will be returned as paraexcitons .
this implies that at high excitation power , the density of orthoexcitons will exceed that of paraexcitons after a few nanoseconds , consistent with the results in fig .
the observation that the creation of orthoexcitons becomes slower at lower exciton density is consistent with the fact that the density - dependent auger process is slower at lower exciton density .
the laser pulses we used for the two - photon excitation have spectral width of around 20 nm .
one would therefore expect that if the orthoexciton and paraexciton cross sections for two - photon absorption are comparable , that we should then see a significant direct creation of orthoexcitons in addition to paraexcitons , even when the laser is tuned to the paraexciton resonance .
the results discussed above indicate that primarily only paraexcitons are being created , however .
is this consistent with the selection rules ?
here we take a close look at selection rules of paraexcitons and orthoexcitons in two - photon excitation . for our one - beam two - photon excitation
, the two input beams have the same energy , @xmath12 , and the same polarization , @xmath13 .
therefore , the absorption coefficient in the dipole transition is proportional to : @xmath14 where @xmath15 are the intermediate states , @xmath16 and @xmath17 are the vacuum state and the 1s-yellow " exciton state respectively .
@xmath18 are possible dipole operators for a certain input polarization @xmath19 , where @xmath20 are operators @xmath21 , @xmath22 and @xmath23 in three dimensions .
because the dipole operator is @xmath24 in the @xmath25 group @xcite for the stressed cu@xmath1o , all allowed internal states should have the negative parity , which could be the @xmath26-states in the yellow " exciton series , the blue " and indigo " exciton series @xcite . by taking into account the energy difference between the yellow " , blue " and indigo " excitons ,
the dominant internal states are the 2@xmath26-yellow " excitons with @xmath27 , @xmath28 , @xmath29 , @xmath30 and @xmath24 symmetries . [
cols="^,^,^",options="header " , ] according to table 2 , when there is no external stress , the two - photon transition is only allowed for orthoexcitons with polarization dependence of @xmath31 in the dipole transition , which is confirmed by experimental reports from kono _ et al .
_ @xcite and kubouchi _ et al . _
@xcite from the same table , for our [ 110 ] laser incident direction , if the stressed crystal is in the d@xmath6 group , paraexcitons should be forbidden in both dipole and quadrupole two - photon transitions , which is consistent with the prediction of inoue - toyozawa @xcite , bader - gold @xcite and ablova - bobrysheva @xcite .
@xmath32 is the angle between the polarization of the input laser beam and the horizontal polarization .
however , our experimental results in figure 6 show that there is strong paraexciton luminescence , and the luminescence from both paraexcitons and orthoexcitons has the same intensity for the vertical and horizontal polarizations .
this can be explained due to the reduction of the symmetry .
that is to say , the crystal is in the d@xmath33 group for most of the harmonic potential well , although the crystal is in the d@xmath6 symmetry group along the axis of the stress . in the d@xmath33 group ,
the polarization dependence of excitons in the direct dipole transition calculated from equation 1 is listed in table 2 . for our [ 110 ] laser incident direction in the d@xmath33 group ,
the paraexciton follows a polarization dependence of @xmath34 , where @xmath35 and @xmath36 are constant factors .
therefore , if @xmath37 , there is no detectable difference between the vertical and horizontal polarization .
because the circular polarization is just a combination of both the vertical and horizontal polarization , we do nt expect to see any difference in the exciton photoluminescence intensity when the input laser is changed to the circular polarization from either vertical or horizontal polarization , which is well confirmed by the experimental results in figure 6 .
the orthoexcitons following the exact polarization dependence of the paraexcitons is another good evidence for our proposed model in the resonant two - photon excitation : paraexcitons are created directly by the ir laser and orthoexcitons are created indirectly through a up - conversion of the paraexcitons .
moreover , we also note that the birefringence changes the polarization of the incident infrared laser beam and the changes are different at different positions in the harmonic potential well . in general
, there is substantial polarization mixing for the whole harmonic trap . at this point ,
one may ask why the paraexciton dipole and quadrupole matrix elements should dominate the creation of excitons in our two - photon excitation , since the orthoexciton transitions are allowed .
one reason is that the ir laser is tuned to the resonant wavelength of the paraexciton state in this work , which gives an on - resonant two - photon excitation for paraexcitons and an off - resonant three - photon excitation for orthoexcitons .
moreover , all our experiments except figure 6 are done with a vertically polarized ir laser , that is to say , @xmath38 in table 2 , which means that the orthoexciton polarization dependence for our [ 110 ] laser incident direction in the d@xmath33 group is @xmath39 .
therefore , the orthoexciton dipole matrix element is forbidden for most of this work . at @xmath40 , the horizontal polarization
, we would expect some orthoexcitons created directly by the ir laser , but the matrix element may be smaller . therefore , we can explicitly express our results as the following : confined by a harmonic potential trap created with an external stress , when an ir laser pulse tuned to one half the paraexciton ground - state energy is shining along the [ 110 ] crystalline direction of the cu@xmath0o sample , paraexcitons are directly created by the ir laser pulse , and orthoexcitons are created primarily from the paraexciton up - conversation process at the early time after the laser pulse .
since we can resonantly create the paraexciton via resonant two - photon excitation , should we see bec of the paraexcitons ?
our estimate of the density of paraexcitons indicates that we should not expect to see this .
7 shows an integrated spatial profile of the paraexciton in the cu@xmath0o crystal at 1.6 kelvin , which gives the paraexciton excitation volume v = @xmath41 . as mentioned in section ( 3.b ) , when detected by pmt , the 250 khz resonant ir laser with average power of 70 mw creates the same exciton luminescence intensity via two - photon excitation as the 3.8 mhz dye - laser at 607.5 nm with 0.9 mw average power does via single - photon excitation .
if we assume that 5% of the 0.9 mw average power is absorbed in the single - photon excitation @xcite , that one absorbed photon creates one exciton in the single - photon excitation , and that comparable luminescence intensity implies comparable exciton density , then we find the density of excitons in the two - photon excitation is @xmath42 where @xmath43 is the speed of light , @xmath44 is a single photon energy in the single - photon excitation with a laser at 607.5 nm , and the ratio @xmath45 is the calibration of the photon - counting efficiencies in the time - correlated single photon detection for the two laser systems with different repetition rates .
this corresponds to @xmath46 absorption of the infrared laser light . to calculate the critical exciton density for bec
, we have to know the exciton temperature , which can be calculated in two ways .
one way is to fit high temperature tail of the paraexciton luminescence with a bose distribution function , @xmath47 , where @xmath48 is paraexciton energy .
the other way is to find the full width at half maximum of the phonon - assisted luminescence line , which is equal to 3.4@xmath49 in a three - dimensional harmonic potential trap @xcite .
our calculations from the above two ways indicate that the exciton has a minimum temperature of 7 kelvin , which means the critical exciton density should be @xmath50 .
@xcite since the created paraexciton density is about thirty times less than the required density for bec , we do not expect to establish bec of paraexcitons even with the maximum laser power from our laser system . with higher laser power or lower temperature ,
however , the paraexciton critical density can be approached with resonant two - photon excitation which creates paraexcitons directly . in order to know whether the density - dependent auger process is the only paraexciton up - conversion process
, we can do an interesting comparison between figure 5 and our former work @xcite .
the former work gives the auger constant of @xmath51 around 1.9 kbar , which implies a decay rate of 2 ns and 4 ns for the input ir laser power of 90 mw and 45 mw , respectively . while figure 5 gives different decay rates , 3 ns for the 90 mw laser and 10 ns for the 45 mw .
there are several reasons for the big inconsistency .
first , there may be other up - conversion processes for the paraexciton besides the density - dependent auger process .
moreover , the paraexcitons must also cool down before they appear in the single - photon luminescence .
besides doing experiments with one - beam , two - photon excitation resonant with the paraexciton state , we can excite excitons with two colliding pulses , that is to say , split the ir laser beam into two parts and shine them into the cu@xmath0o sample , such that both pulses reach the exciton stress trap at the same time . in this case
, the @xmath52 state is directly excited if the two parts travel in opposite directions .
therefore , the created excitons have zero momentum , * @xmath53 * = 0 , which means the ground state is directly excited , where the condensate should appear , if one photon is from each pulse .
the experimental set - up is shown in fig .
1 . in order to make the two beams meet at the same time in the cu@xmath0o sample , we added a two - mirror delay system controlled by a precise translation stage .
8 shows the dependence of the exciton luminescence intensity on the time delay between the two colliding pulses . a comparison between one - beam and two - beam resonant excitations with the same total laser power , 90 mw , is shown in fig .
the exciton luminescence in the one - beam excitation is brighter than that of the two - beam excitation , and creation of orthoexcitons in the one - beam excitation is faster than with the two - beam excitation , which indicates more excitons are created in the one - beam excitation . to explain this phenomenon , one must consider the dependence of the density of states on the momentum * @xmath53*. in the case of two photons travelling in the same direction , the excitons are created in the region of the ( weak ) polariton mixing .
two photons travelling in opposite directions will create an exciton at @xmath54 , where the density of states is much lower . in the case of two - beam excitation , there will also be one - beam excitation with half of the total power of the two - beam excitation . since the exciton density in the two - photon process is proportional to the square of laser power ,
the one - beam excitation with half of the total power will have one fourth the efficiency .
the sum of all processes in the two - beam excitation ends up being less than the total in the one - beam case .
we resonantly create paraexcitons with an ir laser beam tuned to one half the paraexciton ground - state energy .
this is surprising , since it is generally assumed that the cross section for paraexciton creation is always much less than that for orthoexciton .
our experimental results are all consistent with this result , however , including the polarization dependence and time - dependent data . the exciton creation efficiency in resonant two - photon excitation is greater for one - beam excitation than for two colliding pulses , but the colliding pulse method may be useful for direct creation of a condensate in the ground state . at present , the paraexciton density in this work is thirty times less than the required density for bec of paraexcitons , but with higher laser power from stronger ir laser sources or lower temperature , the critical density can be approached with one - beam two - photon excitation resonant with the paraexciton state . *
acknowledgements*. this work has been supported by crdf - mrda award no .
mp2 - 3026 and nsf award no
. dmr-0102457 .
samples of cu@xmath1o were obtained from p. j. dunn of the smithsonian institute .
we would like to thank robert p. devaty for discussions on group theory , albert heberle for suggestions on spectral filtering , m. kuwata - gonokami of tokoyo university on the selection rules and thank vincenzo savona for helpful conversations on the two - beam , two - photon excitation .
99 s. a. moskalenko and d. w. snoke , _ bose - einstein condensation of excitons and biexcitons _ , ( cambridge university press , 2000 ) .
s. denev and d. w. snoke , phys . rev .
b * 65 * , 085211 ( 2001 ) . v. c. negoita , ph.d . thesis , university of pittsburgh , ( 2001 ) .
d. w. snoke and v. negoita , phys .
b * 61 * , 2904 ( 2000 ) .
n. naka and n. nagasawa , phys .
b * 65 * , 075209 ( 2002 )
. s. kono , m. hasuo , and n. nagasawa , j. lumin .
* 66/67 * , 433 ( 1996 ) .
m. kubouchi , k. yoshiot , r. shimano , a. mysyrovwicz and m. kuwata - gonokami , aps/123-qed , in press ( 2004 ) .
a. jolk , m. jorger , and c. klingshirn , phys .
b * 65 * , 245209 ( 2002 ) . y. sun , g. k. l. wong and j. b. ketterson , phys . rev .
b * 63 * , 125323 - 1 ( 2001 ) .
m. y. shen et al .
b * 53 * , 13477 ( 1996 ) .
n. naka and n. nagasawa , solid state comm . * 110 * , 153 ( 1999 ) .
n. naka and n. nagasawa , proceedings of the 26th icps , edingburgh ( 2002 ) .
d. p. trauernicht , j. p. wolfe , and a. mysyrowicz , phys .
b * 34 * , 2561 ( 1986 ) .
yingmei liu and david snoke , `` temperature dependence of exciton auger decay process in cu@xmath0o '' , cond - mat/0408184 ( 2004 ) . g. f. koster , j. o. dimmock , r. g. wheeler , and h. statz , properties of the thirty - two point groups " , ( mit press , cambridge , mass , 1963 ) yingmei liu , ph.d .
thesis , university of pittsburgh , ( 2004 ) .
m. inoue and y. toyozawa , j. phys .
japan * 20 * , 363 ( 1965 ) .
todd r. bader and albert gold , phys .
* 171 * , 997 ( 1968 ) .
l. a. ablova and a. i. bobrysheva , sov .
, vol . 7 , no . 6 , 712 ( 1973 ) .
d. w. snoke and v. negoita , phys .
b * 61 * , 2904 ( 2000 ) . when the ir laser is at 1228 nm , resonant with the paraexciton state , integrated exciton luminescence intensity ( black squares in the figure ) can be best fit to @xmath55 ( black dashed line ) , where @xmath56 is the input laser power .
this is consistent with a two - photon process within the experimental error .
if the laser is off - resonant , at 1240 nm , the integrated exciton luminescence intensity ( black open circles ) can be best fit to @xmath57 ( black solid line in the figure ) , consistent with a three - photon process .
, width=672,height=480 ] time - integrated exciton luminescence intensity in the cu@xmath1o sample for several different laser powers , at 1.6 kelvin and under 1.9 kbar stress along the [ 001 ] direction , with two - photon one - beam excitation resonant with the paraexciton state.,width=672,height=480 ] paraexciton luminescence as a function of the time after the laser pulse , for a cu@xmath1o sample at 1.6 kelvin and under 1.9 kbar stress along the [ 001 ] direction .
black dots : two - photon excitation resonant with the paraexciton state .
black solid line : single - photon excitation tuned to the bottom of the orthoexciton phonon - assisted absorption .
the power of the red and the ir laser were adjusted to give the same total integrated exciton luminescence intensity.,width=672,height=480 ] time - resolved paraexciton and orthoexciton luminescence in the cu@xmath1o sample for two different laser powers , at 1.6 kelvin and under 1.9 kbar stress along the [ 001 ] axis , with two - photon excitation resonant with the paraexciton state .
the black dots and black solid line are the orthoexciton and paraexciton luminescence intensity at laser power of 90 mw , respectively , and the gray dots and gray solid line are the orthoexciton and paraexciton luminescence intensity at laser power of 45 mw , respectively.,width=672,height=480 ] polarization dependence for the total integrated paraexciton and orthoexciton luminescence intensities for a cu@xmath1o sample at 1.6 kelvin under 1.9 kbar stress along the [ 001 ] direction , with two - photon excitation resonant with the paraexciton state.,width=672,height=480 ] spatial profile of the paraexciton luminescence in the stress - induced harmonic potenial trap for the excitons , for the cu@xmath1o sample at 1.6 kelvin under 1.9 kbar stress along the [ 001 ] direction , with two - photon excitations resonant with the paraexciton state.,width=672,height=480 ] total integrated exciton luminescence intensity as a function of the time delay between two laser pulses for two - beam , two - photon excitation resonant with the paraexciton state , for the cu@xmath1o sample at 1.6 kelvin under 1.9 kbar stress along the [ 001 ] direction .
black squares : experimental data .
dashed line : gaussian fit.,width=672,height=480 ] exciton luminescence intensity as a function of the time
after the laser pulses for one - beam and two - beam two - photon excitation resonant with the paraexciton state , with the same total laser power , for the cu@xmath1o sample at 1.6 kelvin under 1.9 kbar stress along the [ 001 ] direction . the black dots and and the black solid line are orthoexciton and paraexciton luminescence intensity , respectively , for one - beam excitation , and the gray dots and gray solid line are the orthoexciton and paraexciton luminescence intensity , respectively , for the two - beam excitation.,width=672,height=480 ] | we have created paraexcitons in cu@xmath0o via resonant two - photon generation , and examined their population dynamics by means of time - correlated single photon detection . confining the excitons to a constant volume in a harmonic - potential trap made with inhomogeneous applied stress along the [ 001 ] axis , we find that paraexcitons are created directly , and orthoexcitons appear primarily through a up - conversion of the paraexcitons .
hot excitons are also created via a three - photon process when the ir laser is non - resonant .
also we generate excitons with two colliding pulses , and the luminescence is weaker than that from one beam excitation with same total laser power .
these results show that resonant one - beam two - photon generation of paraexcitons is a promising way to pursue bose - einstein condensation of paraexcitons . |
transit transmission spectroscopy is an observational technique that can be used to characterize a planet s atmosphere as it transits its host star .
this technique has been used to identify absorption features in some exoplanet atmospheres @xcite , but many planets have flat , featureless spectra @xcite .
flat spectra can be explained by either a high mean molecular weight ( and thus small scale height ) atmosphere or by the presence of high - altitude clouds or hazes , as is common in planets in our own solar system , and has been inferred for the atmospheres of some hot jupiters @xcite . for gj 1214b ,
even high mean molecular weight atmospheres have recently been ruled out , leaving very high altitude clouds or hazes as the only physically plausible explanation for the planet s spectrum @xcite .
the james webb space telescope ( jwst ) and large ground based telescopes such as the european extremely large telescope ( e - elt ) will open up new avenues for characterizing transiting planets .
absorption features for an earth - like or super - earth planet could be detected in the near future with 200 hrs of jwst in - transit observations @xcite , or with @xmath120 hrs of e - elt in - transit observations @xcite .
if there are clouds or hazes present in an atmosphere , they will limit the atmospheric levels that can be probed , making the targets less desirable for characterization .
therefore , it would be beneficial to have a method that could relatively rapidly discriminate between haze - free and hazy planets , which are not easily characterized even in extended transit transmission observations @xcite . here
we examine whether refractive effects on transit transmission spectroscopy could provide a more efficient way of discriminating between hazy , cloudy and clearsky ( free of clouds and hazes ) atmospheres .
@xcite propose that measurements of absorption wing steepness , or a comparison of the depths of multiple absorption bands , could be used to distinguish between cloudy / hazy and clearsky planets .
since both methods require relatively detailed characterization of absorption features , these techniques may not discriminate between a hazy and haze - free planet before considerable amounts of telescope time are used .
in contrast , the refractive signal is independent of absorption features , and could be binned over a wide range of wavelengths , increasing detectability .
while refraction can set a mid - transit maximum transit pressure ( or minimum tangent altitude ) that can be probed by transit transmission spectroscopy @xcite refraction provides the deepest probe of an atmosphere pre- and post - transit @xcite , when it also generates a refractive halo around the exoplanet , increasing the observed flux @xcite .
@xcite derive analytic expressions for the halo brightness for both transparent atmospheres and atmospheres with extinction from rayleigh scattering . @xcite and
@xcite examine the concept further by generating spectra of refracted light for the earth and for venus .
here we expand on previous work by showing that a detection of refracted light in a transit light curve pre - ingress and post - egress would preclude hazy atmospheres , because hazes tend to obscure the layers of the atmosphere that refract light to a distant observer .
we show that this signal could be more readily detectable than spectral absorption features in some cases and could be valuable for selecting targets for more extended follow - up observations .
we used the refraction code that is described in detail in @xcite to calculate refraction angles for a suite of planetary atmospheres .
briefly , refraction is governed by a set of differential equations that we solve at each step along the path through the atmosphere using a runge - kutta integration scheme .
given the planetary radius , surface gravity , atmospheric composition and pressure - temperature profile , the model calculates the angle of deflection due to refraction for a range of tangent altitudes .
we generated refractive light curves to calculate the amount of out - of - transit refracted light .
we first determined whether or not each portion of the atmosphere ( given as an altitude and angle along the annulus of the atmosphere ) is illuminated at each time during the transit event , from half a transit length prior to ingress to half a transit length after egress , and then integrated over the entire atmosphere to generate the light curve .
we quantified the signal of refracted light as the difference in the average value of the transit light curve between two stages of the transit event .
we chose a quarter of a transit length as the time bin to maximize the signal to noise ratio ( s / n ) for the majority of cases we examined .
as can be seen in figure [ fig : lightcurve ] , most of the refracted flux is seen in the quarter of a transit prior to ingress , so dividing the transit into longer stages would reduce the time - averaged signal . stages with shorter durations could increase the time - averaged signal , but would have greater noise levels because of the shorter integration time . refracted light brightness is more strongly peaked just outside of transit for planets with equilibrium temperatures ( t@xmath2 , see @xcite for definition ) @xmath1600 k , but we find that even for these cases adopting a time bin of 5% of the transit length results in poorer s / n for t@xmath3600 k and an increase in s / n by only a factor of @xmath42 for planets with greater temperatures we used a suite of planetary atmospheres to calculate the refracted light signal .
these are shown in table [ tab : params ] .
we have selected a combination of solar system analogs as well possible super - earth and mini - neptune atmospheres to cover a wide range of potential planetary atmospheres .
we assumed the h@xmath5-dominated atmospheres have a solar h / he ratio ( 90% h , 10% he ) for simplicity , but the small change in the refractive index for different h / he ratios should have a negligible effect on our results .
for the super - earth and mini - neptune planets , we ran our models on 4 test cases to span the most likely bulk atmospheric compositions : 100% n@xmath5 , solar composition , 100% h@xmath5o , and 100% co@xmath5 . out - of - transit refracted light
must be deflected by a large enough angle to be scattered into the beam to a distant observer .
the characteristic angle of deflection ( in radians ) is @xmath4r@xmath6/d , ( where r@xmath6 is the stellar radius and @xmath7 is the planet - star distance ) which is also half the angular size of the star , as seen by the planet .
for example , half a transit length prior to ingress , on the trailing side of the planet , light originating at the near and far limb of the star would have to be refracted by r@xmath6/d and 3r@xmath6/d respectively , to reach a distant observer .
more than half a transit length prior to ingress , the required refraction angles would increase , and closer to ingress they would decrease . based on the qualitative description given above , the brightness of the refracted light signal depends on the angles of refraction at each altitude in an atmosphere and the planet - star geometry .
the deflection of light by a planetary atmosphere can be calculated by our model from the atmospheric scale height , the planetary radius ( r@xmath8 ) , and the index of refraction of the atmosphere . for each test case ,
r@xmath8 and the refractive index are given .
the scale height is determined from the surface gravity , mean molecular weight of the atmosphere , and t@xmath2 .
surface gravity is given for each test case , and the mean molecular weight is determined by the composition .
we ran our model simulations over a grid of isothermal atmospheres with t@xmath2 from 100 to 1000 k , covering a wide range of atmospheric scale heights .
we chose to use isothermal atmospheres for simplicity after testing other temperature profiles with realistic tropospheric lapse rates and stratospheric temperature inversions and finding no significant difference in our results . the planet - star geometry is determined by r@xmath8 , r@xmath6 , the impact parameter ( @xmath9 ) , and @xmath7 .
the impact parameter is the sky - projected distance at conjunction , in units of stellar radius @xcite . to cover the full range of planet - star geometries , we ran our simulations over a range of values for @xmath9 , planetary albedo , and stellar types from m9 to f5 , constraining the stellar radius and luminosity .
lcccc earth & 6371 & n@xmath5 & 1.00029 & 9.8 + super - earth & 12742 & n@xmath5 & 1.00029 & 9.8 + mini - neptune & 12742 & h@xmath5 & 1.00012 & 9.8 + h@xmath5o super - earth & 12742 & h@xmath5o & 1.00026 & 9.8 + co@xmath5 super - earth & 12742 & co@xmath5 & 1.00044 & 9.8 + neptune & 24622 & h@xmath5 & 1.00012 & 11.1 + saturn & 58232 & h@xmath5 & 1.00012 & 10.44 + jupiter & 69911 & h@xmath5 & 1.00012 & 24.8 + [ tab : params ] because our model does not explicitly calculate the effect of cloud and aerosol opacity , we simulated the effect of a cloud or haze layer by truncating the depth of the measurable atmosphere at a characteristic pressure layer .
to determine appropriate pressure cutoff layers for the three main aerosol cases under consideration , we used our modeling results and examples of clouds and hazes in our own solar system to select pressure cutoffs at 1 bar ( clearsky case ) , 0.1 bars ( cloudy case ) and 1 mbar ( hazy case ) .
we chose 1 bar as our clearsky pressure cutoff because at pressures @xmath101 bar , our modeling indicates that atmospheres within the range of compositions under consideration are optically thick near 1 @xmath11 m ( the central wavelength for our transit simulations ) when only rayleigh scattering is included .
for the pressure cut - off for cloudy atmospheres we chose 0.1 bars , which is a characteristic lower pressure limit for the tropopause for atmospheres of a range of different compositions @xcite , and the majority of clouds are found within a planet s troposphere .
lastly , we chose 1 mbar as the hazy pressure cutoff because hazes are typically generated via photochemistry in the upper atmosphere at pressures near 1 mbar .
for example , at 1 @xmath11 m venus is optically thick ( @xmath12=1 ) in transit transmission at 90 km ( @xmath40.1 mbar ) @xcite and titan is optically thick at @xmath4240 km ( @xmath00.5 mbar ) @xcite . because hazes form at pressures @xmath01 mbar in both a warm co@xmath5-dominated atmosphere and a cold n@xmath5-dominated atmosphere , we chose 1 mbar as a reasonable cutoff for hazes over the parameter space we explore here .
we used the publicly available exposure time calculators ( etcs ) to estimate the s / n for detecting refracted light in a transit light curve with jwst and e - elt .
we used the etcs to estimate the noise at 1 @xmath11 m with spectral resolving power ( r ) equal to 100 for stellar types from f5v to m9v .
we chose the lowest resolving power available ( r=100 for the jwst etc ) because this technique does not require high resolving power , and could even be performed with broadband filter photometry if necessary .
the input stellar spectra were phoenix nextgen spectra with solar metallicities @xcite with the star placed at a distance of 10 pc . given the parts per million ( ppm ) flux difference for refracted light and the estimated s / ns from the etcs , we calculated the out - of - transit integration time required to detect refracted light at a s / n@xmath13 over all t@xmath2 values for each planetary atmosphere and stellar type .
lccccccccccc earth & 400 & 5780 & clearsky & 0.15 & 0.13 & 4.77 & 1.0 & 0.3 & 999.00 & 781.2 & 231.9 + super - earth & 450 & 5780 & clearsky & 0.15 & 0.29 & 0.91 & 1.0 & 0.2 & 999.00 & 166.9 & 34.8 + mini - neptune & 250 & 5780 & clearsky & 0.15 & 1.98 & 0.02 & 1.0 & 1.2 & 27.65 & 2.0 & 2.4 + h@xmath5o super - earth & 400 & 5780 & clearsky & 0.15 & 0.41 & 0.45 & 1.0 & 0.3 & 640.78 & 73.9 & 21.9 + co@xmath5 super - earth & 600 & 5780 & clearsky & 0.15 & 0.22 & 1.60 & 1.0 & 0.1 & 999.00 & 393.7 & 34.6
+ neptune & 250 & 5780 & clearsky & 0.15 & 3.79 & 0.01 & 1.0 & 1.2 & 7.58 & 1.0 & 1.2 + saturn & 300 & 5780 & clearsky & 0.15 & 10.98 & 0.01 & 1.0 & 0.7 & 0.90 & 1.0 & 0.7 + jupiter & 350 & 5780 & clearsky & 0.15 & 6.39 & 0.01 & 1.0 & 0.4 & 2.66 & 1.0 & 0.4 + [ tab : results ] table [ tab : results ] shows the ppm flux change , and the required integration time , number of transits and total time ( from first transit to last ) for detecting refracted light for each test case over the suite of parameters .
the results shown here are for an albedo of 0.15 , but results for other albedos are available online .
our results indicate that saturn analog planets exhibit the most detectable refracted light of any of the cases because saturn has a radius close to jupiter s radius and a lower surface gravity , which increases the atmospheric scale height at a given temperature .
the amplitude of the refracted light signal ( as defined in section [ sec : modeldesc ] ) is no larger than half a scale height for all cases we have explored here .
the maximum flux amplitude for planets orbiting sun - like stars is 10 ppm for a 300 k saturn analog .
the other h@xmath5 cases have maximum amplitudes of 6 , 4 , and 2 ppm for the jupiter , neptune , and mini - neptune cases , respectively .
the greatest ppm signals are for planets orbiting around m9v stars , for which the signals can increase by nearly two orders of magnitude to 950 ppm for a 200 k saturn analog .
figure [ fig : inttime]a shows the jwst out - of - transit integration time required to detect refracted light for the 4 h@xmath5-dominated atmospheres : jupiter , saturn , neptune analogs and the ` mini - neptune ' , all without clouds or hazes . for many of the saturn and jupiter - analog cases , refracted light could be detected in @xmath010 hours of jwst time .
this integration time can be achieved in 1 transit for jupiter and saturn - analog planets with t@xmath3600 k orbiting f , g and k stars and in @xmath05 transits for t@xmath3400 k orbiting m dwarfs . for cases in which multiple transits are required , the total time from first transit to last is @xmath01 year , and typically @xmath06 months .
figure [ fig : inttime]a shows our results for @xmath9=0.0 , with observing times required increasing by 1% for @xmath9=0.2 , 10% for of @xmath9=0.6 , and 30% for @xmath9=0.9 .
-dominated atmospheres .
the results assume that observations are made at 1 @xmath11 m with r=100 , for a planet at a distance of 10 pc .
the planets with the most detectable refracted light signal are those with t@xmath3600 k for solar - type stars and @xmath0400 k for m dwarfs . for many jupiter and saturn - analog cases , refracted light
could be detected with @xmath010 hour of jwst time . *
b ) * e - elt integration time required to detect refracted light for n@xmath5 and h@xmath5 atmospheres , assuming that 50 wavelength bins can be summed over at r=100 .
refracted light is most detectable for the non - hazy atmospheres , and therefore could be used to distinguish between hazy and non - hazy worlds .
, width=340 ] figure [ fig : inttime]b shows the e - elt integration time required to detect refracted light for super - earth and mini - neptune atmospheres with @xmath9=0.0 .
we calculated the signal levels for n@xmath5 , h@xmath5o , co@xmath5 and h@xmath5 atmospheres , but only a comparison of n@xmath5 and h@xmath5 atmospheres is shown here .
we find that refracted light could be detectable in @xmath010 hrs of e - elt time for many of the clearsky atmospheres , and even some cloudy atmospheres .
in contrast , detecting refracted light for a hazy exoplanet would require @xmath1100 hours for all the planetary atmospheres we considered .
here we have assumed that it is possible to bin over at least 50 spectral resolution elements .
the justification for this is found in figure [ fig : lightcurve - diff ] , which shows the wavelength - dependent refracted light signal for 2 r@xmath13 planet with an earth - like atmosphere .
a larger change in effective radius at a given wavelength means a stronger flux from refraction prior to ingress or after egress . between 0.8 and 1.35 @xmath11 m - shortward of a major h@xmath5o absorption feature and where rayleigh scattering opacities are small - there are @xmath450 spectral resolution elements that could be summed . for earth - analog atmospheres , there is a relatively large flux difference at all these wavelengths .
therefore , we consider binning over multiple spectral resolution elements to decrease the integration time to be a valid approach , at least for n@xmath5-dominated planets like earth . the greatest amplitude of refracted flux for the n@xmath5 super - earth cases around a sun - like star is 0.12 ppm for a 400 k planet . for the cloudy case ,
the maximum amplitude is 0.06 ppm at 200 k. the cloudy h@xmath5 cases have amplitudes between 0.2 and 1.0 ppm , but only for the very cold ( @xmath0200 k ) cases .
the amplitudes for the hazy h@xmath5 cases are all below 0.025 ppm , and below 0.005 ppm for t@xmath14150 k. the number of transits required to detect refracted light with e - elt is 1 for clearsky n@xmath5 super - earths with @xmath15800 k orbiting f , g and k stars and @xmath03 for t@xmath3500 k for those orbiting m dwarfs . as with the jupiter and saturn analogs , 3 transits is , from first transit to last , much less than a year and typically @xmath03 months
. cloudy atmospheres with t@xmath3250 k could exhibit detectable refracted light signals , but hazy atmospheres have largely undetectable refracted light signals except for some very cold ( t@xmath2=100 k ) cases .
super - earth with an earth - like atmosphere , represented as the change in effective radius ( km ) .
the spectra shown are the differences in the spectra between stage 2 and stage 1 of the transit ( see figure [ fig : lightcurve ] ) .
this figure shows that the refracted light signal could be detected over a wide wavelength range , and that it should be possible to bin over multiple spectral resolution elements to reduce the integration time needed to detect refracted light.,width=302 ]
a detection of refracted light implies a haze - free atmosphere because refracted light is much more detectable for a clearsky atmosphere than for a hazy one ( see figure [ fig : inttime]b ) .
however , discriminating between cloudy and hazy worlds could be more challenging . for example
, for a 600 k n@xmath5 super - earth orbiting a sun - like star ( and for the majority of parameter space ) , a null detection of refracted light would be consistent with either a cloudy or hazy atmosphere , with no apparent way to differentiate between the two .
on the other hand , for a 250 k n@xmath5 super - earth orbiting a sun - like star , both the clearsky and cloudy cases are consistent with a detection of refracted light .
to disambiguate these results , one would need to quantify the refracted light , which would require more observing time .
overall , a detection of refracted light is indicative of a non - hazy atmosphere and , for some regions of parameter space , quantifying the refracted light flux could aid in uniquely discriminating between cloudy , hazy and clearsky atmospheres .
the refracted light brightness is strongest for planets with t@xmath2 between 150 - 350 k , and is undetectable for very high temperature planets . for hot , close - in planets , the planet - star distance ( @xmath7 ) is small , meaning that the characteristic deflection angle r@xmath6/d is large , and that large refraction angles are required to produce a strong refracted light signal . for t@xmath14800 k , angles this large would require probing pressures greater than 1 bar , where most atmospheres should be opaque , meaning that atmospheric opacity results in low refracted light signals .
for the coldest ( t@xmath3150 k ) planets , @xmath7 is large and the refraction angles for clearsky atmospheres are often much larger than r@xmath6/d .
this results in more refracted light being observed further away from ingress and egress , increasing the average flux in stage 1 relative to stage 2 and reducing the overall detectability .
( see figure [ fig : lightcurve ] ) . in the near future , e - elt
could be used to identify non - hazy potentially habitable planets , which have 180@xmath0t@xmath3260 k ( @xcite , ravi kopparapu , private communication ) .
as shown in table [ tab : results ] and figure [ fig : inttime ] , refracted light could be detectable with one transit with e - elt , or @xmath05 hrs of out - of - transit e - elt time for potentially habitable n@xmath5 dominated super - earths orbiting f , g and k stars . for planets orbiting m dwarfs ,
the required number of transits is typically less than two , with a total integration time of @xmath05 hrs .
@xcite estimate that it could take up to 10 transits to detect h@xmath5o and co@xmath5 for earth - like planets orbiting f , g and k stars with e - elt using filter photometry , and @xcite find that it would take @xmath120 hours of e - elt time to detect o@xmath5 for earth - like planets orbiting m dwarfs using high resolution ( r@xmath110000 ) spectroscopy .
these estimates are larger than the amount of out - of - transit e - elt time necessary to detect refracted light for potentially habitable planets .
therefore , because refracted light could be more detectable than spectral absorption features , looking for refracted light to distinguish between hazy and non - hazy exoplanets could be a useful tool in selecting exoplanets for extended follow - up observations .
increases in out - of - transit flux due to refraction prior to ingress and subsequent to egress could be detectable with @xmath010 hours of out - of - transit observing time for saturn and jupiter - sized planets with jwst and for super - earths / mini - neptunes with e - elt .
detecting refracted light would be indicative of a haze - free atmosphere , and a quantification of the amount of refracted light could aid in distinguishing between cloudy and clearsky atmospheres for planets with equilibrium temperatures @xmath0300 k. because refracted light can , in some cases , be detectable with less than a few hours of out - of - transit observing time , this method could be an economical way of determining if an exoplanet is haze - free and therefore a good target for extended follow - up observations .
this work was performed by the nasa astrobiology institute s virtual planetary laboratory , supported by the nasa astrobiology institute under cooperative agreement solicitation nnh05zda001c . | we propose a method to distinguish between cloudy , hazy and clearsky ( free of clouds and hazes ) exoplanet atmospheres that could be applicable to upcoming large aperture space and ground - based telescopes such as the james webb space telescope ( jwst ) and the european extremely large telescope ( e - elt ) .
these facilities will be powerful tools for characterizing transiting exoplanets , but only after a considerable amount of telescope time is devoted to a single planet . a technique that could provide a relatively rapid means of identifying haze - free targets ( which may be more valuable targets for characterization )
could potentially increase the science return for these telescopes .
our proposed method utilizes broadband observations of refracted light in the out - of - transit spectrum .
light refracted through an exoplanet atmosphere can lead to an increase of flux prior to ingress and subsequent to egress .
because this light is transmitted at pressures greater than those for typical cloud and haze layers , the detection of refracted light could indicate a cloud- or haze - free atmosphere .
a detection of refracted light could be accomplished in @xmath010 hours for jovian exoplanets with jwst and @xmath05 hours for super - earths / mini - neptunes with e - elt .
we find that this technique is most effective for planets with equilibrium temperatures between 200 and 500 k , which may include potentially habitable planets .
a detection of refracted light for a potentially habitable planet would strongly suggest the planet was free of a global cloud or haze layer , and therefore a promising candidate for follow - up observations . |
the most important experimental goal of the cern large hadron collider ( lhc ) is to unravel the mechanism of electroweak symmetry breaking . in the standard model ( sm ) of electroweak interactions ,
only one higgs doublet is required to generate mass for fermions as well as gauge bosons .
one neutral cp - even higgs boson ( @xmath5 ) appears after spontaneous symmetry breaking .
various extensions of the sm have more complicated higgs sectors and lead to additional physical spin zero fields @xcite .
the minimal supersymmetric extension of the standard model ( mssm ) @xcite has two higgs doublets with vacuum expectation values @xmath6 and @xmath7 .
after spontaneous symmetry breaking , there remain five physical higgs bosons : a pair of singly charged higgs bosons @xmath8 , two neutral cp - even scalars @xmath9 ( heavier ) and @xmath2 ( lighter ) , and a neutral cp - odd pseudoscalar @xmath0 . the higgs sector is strongly constrained by supersymmetry so that , at the tree level , all higgs boson masses and couplings are determined by just two independent parameters , which are commonly chosen to be the mass of the cp - odd pseudoscalar ( @xmath10 ) and @xmath11 . at the one loop level ,
radiative corrections from the @xmath12-quark and the @xmath13 quark yukawa couplings substantially modify the tree level formulae for masses and mixing patterns in the higgs sector .
let us briefly review recent studies on the search for the @xmath0 at the lhc . in most studies ,
the parameters were selected such that supersymmetric particle ( susy particle ) masses were large so that higgs boson decays to susy particles were kinematically forbidden .
the @xmath14 mode might be observable at the lhc for 120 gev @xmath15 and @xmath16 close to one @xcite-@xcite .
the @xmath0 does not have tree level couplings with gauge boson pairs .
the one - loop induced @xmath17 decay mode was found to be a useful discovery channel at the lhc only if @xmath16 is much less than one and @xmath10 is near the @xmath18 threshold @xcite .
if @xmath16 is large , the @xmath19 decay mode might be a promising discovery channel for the @xmath0 in the mssm @xcite .
it has been suggested that the @xmath0 might also be observable via its @xmath20 decays in a large region of the ( @xmath21 ) plane , provided that sufficient @xmath13-tagging capability can be achieved @xcite . for @xmath16 larger than about 10 ,
the @xmath22 discovery mode at the lhc was found to be observable in a large region of the @xmath23 parameter space @xcite , even when higgs decays to susy particles are significant @xcite .
this channel will provide a good opportunity to reconstruct the masses for neutral higgs bosons with high precision .
there are , however , regions of parameter space where rates for higgs boson decays to susy particles , e.g. the charginos and neutralinos , are substantial or even dominant .
while these decays reduce the rates for sm signatures , making conventional detection of higgs bosons even more difficult , they also open up a number of new promising modes for higgs detection @xcite .
a recent study suggests that if the sleptons are light , it might be possible to search for the neutral higgs bosons via their decays into slepton pairs @xcite .
the mssm higgs pseudoscalar ( @xmath0 ) decays into a @xmath1 boson and the lightest higgs scalar ( @xmath2 ) with a large branching ratio for @xmath24 and @xmath25 .
decay is possible for @xmath10 less than about 60 gev , that is slightly above the excluded region of the lep and within the discovery region for the lep ii . ]
the final state where @xmath3 and @xmath26 , was found to be a promising channel to simultaneously detect the @xmath0 and the @xmath2 at the lhc @xcite .
the @xmath4 has a branching fraction about 20 times larger than that of @xmath27 .
it is , therefore , natural to ask whether with a good b - tagging efficiency it is possible to search for the @xmath0 and the @xmath2 in the decay of @xmath28 with the final state of @xmath3 and @xmath4 . in this paper , we assess the prospects for the simultaneous discovery of the @xmath0 and the @xmath2 at the lhc via the process @xmath29 .
the compact muon solenoid detector ( cms ) performance is adopted for a realistic study of observability .
a brief summary of our preliminary analysis , for the case when higgs boson decays to susy particles are kinematically forbidden , has appeared in the cms technical proposal @xcite . in this complete analysis , we have carried out detailed studies for dominant backgrounds , and further , have considered the impact of higgs decays to susy particles .
parton level calculations are presented in section ii .
results from more realistic simulations are discussed in section iii .
we end with a discussion of our results in section iv .
the total cross section for the process @xmath30 is evaluated from the cross section @xmath31 multiplied with the branching fractions of @xmath28 , @xmath3 and @xmath4 .
the parton distribution functions of cteq2l @xcite are chosen to evaluate the cross section of @xmath32 with @xmath33 gev and @xmath34 .
we take @xmath35 gev , @xmath36 , @xmath37 , @xmath38 gev , and @xmath39 gev .
gluon fusion ( @xmath40 ) , via the top quark and the bottom quark triangle loop diagrams , is the major source for the higgs pseudoscalar if @xmath16 is less than about 4 .
the amplitude for the process @xmath40 is a function of the quark mass squared ( @xmath41 ) , @xmath42 , and @xmath16 @xcite : 1 . for @xmath43
, the @xmath12-loop is almost independent of @xmath44 , the @xmath45 coupling can be either obtained from the low energy theorem of the axial anomaly @xcite-@xcite or from the exact calculation for @xmath40 at the limit of @xmath46 .
2 . at the threshold of @xmath47 , the imaginary part of the amplitude turns on . therefore
, the cross section is significantly enhanced when @xmath10 is close to @xmath48 .
when @xmath41 is much less than @xmath42 , the amplitude is proportional to @xmath49 $ ] .
the @xmath12-loop dominates in a large region of @xmath16 .
the cross section is almost proportional to @xmath50 for @xmath51 . only for @xmath16 close to @xmath52
, can the @xmath13-loop dominate and the total cross section be enhanced by large @xmath53 . since the yukawa coupling of @xmath54 is proportional to @xmath16 , the production rate of the @xmath0 from @xmath55
is enhanced for large values of @xmath16 .
if @xmath16 is larger than about 7 , the @xmath0 is dominantly produced from @xmath13-quark fusion ( @xmath55 ) @xcite .
we have evaluated the cross section of the @xmath0 in @xmath56 collisions @xmath31 , with two dominant subprocesses : @xmath40 and @xmath57 .
the cross section of @xmath57 is a good approximation to the ` exact ' cross section of @xmath55 @xcite for @xmath10 less than about 500 gev .
qcd radiative corrections which increase the gluon fusion ( @xmath40 ) production cross section by about 50% to 80% for @xmath58 @xcite are not included in our computation for either the signal or the backgrounds .
we have included complete one loop corrections from both the top and the bottom yukawa interactions to the higgs masses and couplings using the effective potential @xcite-@xcite .
the formulae presented in ref .
@xcite have been employed in our calculations .
the contributions from the d - terms are usually small @xcite , and therefore , are not included . the soft susy breaking parameters @xmath59 and @xmath60
are taken to be zero .
we take @xmath61 and consider three sets of parameters : ( a ) @xmath62 gev , such that the higgs boson decays to susy particles are kinematically forbidden ; ( b ) @xmath63 gev , such that the higgs boson decays to susy particles become significant ; and ( c ) @xmath64 gev , such that the higgs boson decays to susy particles are large and dominant when @xmath16 is less than about 10 . at first , let us take @xmath65 gev , such that the neutralinos ( @xmath66 ) , the charginos ( @xmath67 ) , the squarks ( @xmath68 ) , and the sleptons ( @xmath69 ) , are heavy and the higgs boson decays to susy particles are kinematically forbidden . with qcd
radiative corrections @xcite , the branching fraction of @xmath70 is reduced by about a factor of 2 .
the @xmath20 mode dominates the higgs pseudoscalar decays for @xmath16 larger than about 4 and @xmath71 . for @xmath25 and @xmath24 ,
the branching fraction of @xmath28 is comparable to @xmath72 if the susy decay modes are kinematically forbidden .
the decay of @xmath28 can become dominant if @xmath16 is less than 2 . in this region ,
the lightest higgs scalar mass is modified mainly by the radiative corrections from the top quark yukawa couplings .
the leading radiative corrections to the @xmath73 can be expressed in terms of the third generation doublet and singlet squark masses ( @xmath74 and @xmath75 ) @xcite , @xmath76\ ] ] where @xmath77 ^ 2 + \sin^2 2\beta({m_a}^2+{m_z}^2)^2,\ ] ] and @xmath78.\ ] ] we have , of course , included complete one loop corrections from both the top and the bottom yukawa interactions in our analysis with the formulae in ref .
@xcite .
the total cross section for the process @xmath79 is shown in fig . 1 as a function of @xmath10 for various values of @xmath16 .
we take @xmath61 and consider three sets of parameters similar to those chosen in ref .
@xcite , ( a ) @xmath80 gev , such that the higgs boson decays to susy particles are kinematically forbidden ; ( b ) @xmath81 gev , and ( c ) @xmath82 gev , such that the higgs boson decays to susy particles are large and dominant for @xmath16 less than about 10 .
the signal cross section is largest for case ( a ) where no supersymmetric decays of @xmath0 are kinematically accessible . for the @xmath83 gev cases ( b ) and ( c ) , the branching fraction for susy decays to charginos and neutralinos are larger for the case of positive @xmath84 since @xmath85 and @xmath86 are generally lighter than when @xmath87 .
the reduction in the cross section due to the opening of the susy modes is smaller when @xmath16 is large because the @xmath13-yukawa coupling , and hence the decay rate via the @xmath20 channel , grows with @xmath16 .
we also remark that for @xmath88 gev and @xmath89 gev , and @xmath90 , the cross section is further reduced because the decay @xmath91 become accessible , so that the @xmath92 is reduced by almost 50% . in fig .
2 , we present the total cross section for the process @xmath79 as a function of @xmath16 for various values of @xmath10 .
other parameters are the same as those in fig .
1 . note that the @xmath73 is usually enhanced with a larger @xmath16 . in fig .
2(a ) , the cross section for @xmath93 gev drops sharply for @xmath94 because @xmath95 becomes larger than @xmath10 . in figs .
2(b ) and 2(c ) , the cross section of @xmath93 gev is smaller than that of @xmath96 gev for a large @xmath16 because the phase space is suppressed when @xmath10 is only slightly larger than @xmath95 .
3 shows the contours of total cross section ( @xmath97 ) for @xmath79 in the @xmath23 plane for @xmath98 and 100 fb .
the susy parameters are the same as those in fig
. 1 . also shown as
the dashed line is the mass contour for @xmath99 .
if susy decays of the @xmath0 and the @xmath2 are kinematically forbidden , the cross section can be larger than 100 fb , for @xmath100 , and @xmath16 less than about 2 .
recent studies on @xmath101 @xcite , have shown that more realistic estimations with a detector response could remarkably modify the results of the parton level calculations . to make such a comparison some detector model has to be employed .
the lhc will have two general purpose detectors : the cms and the atlas .
both detectors will provide precise muon and electron momentum reconstruction as well as reasonable jet energy resolution .
we have adopted the cms detector performance parameters @xcite to estimate the signal and backgrounds .
similar results can be expected for the atlas detector @xcite .
the pythia 5.7/jetset 7.4 generator @xcite with the cteq2l @xcite parton distribution functions is used to simulate events at the particle level . to simulate the signal
, we use pythia to generate the subprocess @xmath40 , which is the dominant source of the @xmath0 for low @xmath16 .
but only the kinematics , parton showering and hadronization schemes are used directly from pythia .
the cross section of @xmath102 is very sensitive to the choice of mssm parameters . to be consistent with the parton level calculations ,
the results from pythia are rescaled to the same cross section and the one - loop corrected @xmath73 is incorporated in pythia by hand .
the output from pythia / jetset is processed with the cmsjet program @xcite which has been developed for the fast simulations of the `` realistic '' cms detector response .
the resolution effects are taken into account with the parametrizations obtained from the detailed geant @xcite simulations .
the cmsjet also includes some analysis routines , in particular , a set of jet reconstruction algorithms .
all events containing @xmath103 or @xmath104 pairs and at least two jets are reconstructed and stored for further analysis . at the first stage , the minimal kinematical cuts ( i )
are applied : @xmath105 in principle , the @xmath106 coverage of cms detector for jets is much better , but for @xmath107 the @xmath13-tagging capability rapidly deteriorates .
the lepton pair serves as an event trigger .
the cut on @xmath108 is included to remove the reducible backgrounds from other dilepton sources . the @xmath13-tagging capability of any detector is mainly defined by the tracker system quality .
preliminary cms results indicate that the average @xmath13-tagging efficiency at the level of 40% with a mistagging probability of about 2% can be achieved quite easily with the microvertex detectors @xcite .
the algorithm applied combines common impact parameter and lepton tagging techniques .
these values look quite reasonable in light of the recent cdf results @xcite .
there is some hope that some more sophisticated @xmath13-tagging algorithms will allow one to achieve the @xmath13-tagging efficiency of 60% with a purity of 1% for both the cms and the atlas experiments . because of the existing uncertainties we prefer to use here a rather artificial and simplified but more general way to take into account @xmath13-tagging .
namely , all jets are matched as a @xmath13-jet or a non @xmath13-jet when they are generated . in our analysis , a given jet passing kinematical cuts is matched as a @xmath13-jet if there is a @xmath13-quark in a cone @xmath109 , where @xmath110 of course , there is some ambiguity in this procedure , but we have checked that more than 95% of the @xmath13-jets from the @xmath111 decays are matched correctly .
our procedure then allows us to apply a simplified parameterization of the @xmath13-tagging efficiency to the same event sample simply rejecting or accepting a given jet in the event with a certain probability .
relevant backgrounds can be separated into two groups .
the first one contains irreducible background processes with a @xmath1 and two @xmath13-jets in the final state .
the second one contains all other potential backgrounds without a @xmath1 or two @xmath13-jets in the final state . in our analysis , we have applied double @xmath13-tagging to extract the signal since after a preliminary analysis it quickly became evident that single @xmath13-tagging is not sufficient .
4 shows the invariant mass distribution for various backgrounds @xmath112 with a @xmath13-tagging efficiency of 40% and a purity of 2% after the minimal cuts ( i ) . the dominant background is the irreducible process
@xmath113 , provided that the @xmath13-tagging quality is high enough .
we estimate this process contribution using the @xmath114 subprocess incorporated in pythia @xcite .
the total production cross section is about 600 pb at the lhc and kinematics are in fairly good agreement with the results of our parton level calculations . to estimate the background from @xmath115 , when one or both non @xmath13-jets are mistagged , we use pythia to generate the @xmath116 processes with @xmath117 gev .
the second jet is generated in the parton shower evolution .
to avoid the double counting , events with two tagged @xmath13-jets in the final state were eliminated from the @xmath118 sample .
the background from @xmath119 is at the same level of @xmath120jets after the minimal cuts and can be reduced further by a factor of 3 with the additional cut ( ii ) on the missing @xmath121 , @xmath122 the potentially dangerous background from @xmath123 is below the level of the dominant @xmath124 by a factor of 10 - 20 , so the peak at @xmath125 in the @xmath126 background distribution is nearly invisible
. we also estimated the backgrounds from @xmath127 , @xmath128 and @xmath129 .
they are well below the uncertainty in our estimations of the @xmath130 background . the additional isolation cut ( iii )
@xmath131 effectively reduces the backgrounds with the leptons originating from @xmath132 or @xmath13 hadrons .
the @xmath132-jets are assumed to be mistagged as the @xmath13-jets with a probability of 9@xmath133 .
we have unsuccessfully attempted to improve the significance of the signal using harder kinematical cuts .
it seems that to optimize significance one has to use as soft cuts on @xmath134 and @xmath135 as possible taking into account detector limitations .
this conclusion is valid over almost the whole mass and @xmath16 range except some small region for @xmath136 gev and @xmath137 , where harder cuts provide some small increase in the significance compared to the softest cuts used . the last comment concerning fig .
4 is that all background distributions are peaked around @xmath138 gev .
the large background makes the observation for the @xmath0 very difficult in this mass region in spite of the sizeable production cross section of @xmath32 and a significant branching fraction of @xmath28 .
this discovery channel , @xmath139 , is especially attractive because it allows one , in principle , to reconstruct the masses for both the @xmath0 and the @xmath2 .
the optimal experimental procedure to extract the signal from the background seems to be an iterative one . at the first step
, all @xmath140 events passing the cuts i , ii , iii , and with both @xmath13-tagged jets have to be preselected . then , in the case of known @xmath73 , one can directly use the restriction @xmath141 to improve the observability of the @xmath0 in the @xmath142 distribution .
the window size @xmath143 depends on the detector resolution . for the cms detector
we estimate @xmath144 gev for this particular kinematical region .
it has to be stressed that the reconstructed @xmath126 has the systematic downward shift about 5 - 7 gev when one uses the usual cone jet - finding algorithms . in the case
that @xmath73 is beyond the reach of lep ii , one can use some ` sliding window ' procedure .
namely , one should scan the relevant @xmath145 region by small steps , every time placing the middle of the window into the new position , trying to observe some bump in the @xmath146 distribution .
if there is some peak in the @xmath147 distribution , one should apply the last step to improve observability for the @xmath2 ( if its mass is unknown ) by histogramming the @xmath148 distribution only for events with @xmath147 in some window around the @xmath149 . in fig .
5 , we illustrate the procedure of extracting the signal for @xmath150 gev and @xmath151 .
a signals in the first row are obtained using @xmath148 mass window shifted with respect to the @xmath152 position .
the second row presents the @xmath153 distributions provided the optimal @xmath148 position is chosen . and
the last row demonstrates the @xmath2 signal over the background provided the events are selected within the proper @xmath153 mass range around the bumps found .
we present the @xmath154 significance contours for the cms detector with an integrated luminosity of @xmath155 10 and 30 fb@xmath156 in fig .
all the cuts i , ii , iii , have been applied and the sliding window procedure of fig .
5 has been used to enhance the signal .
we show results for two sets of the mssm parameters : ( 1 ) @xmath157 tev and ( 2 ) @xmath158 gev .
we assume here that @xmath159 fb@xmath156 will be accumulated at the low luminosity regime ( @xmath160 ) and that pile up effects will be small .
a higher luminosity might introduce additional complications due to the pile - up effects .
however , if the reasonable @xmath13-tagging efficiency ( 40% + 2% ) could be achieved at a higher luminosity ( about @xmath161 , the region explorable with @xmath162 fb@xmath156 and a better @xmath13-tagging efficiency 60% + 1% can also be probed with the more modest @xmath13-tagging efficiency ( 40% + 2% ) at @xmath163 fb@xmath156 . for @xmath164 tev
, the signal of @xmath165 can be observable for @xmath166 and 180 gev @xmath15 .
for @xmath158 gev , this discovery channel , @xmath165 , can be observable for @xmath167 and 180 gev @xmath15 .
the cp odd state of the electroweak symmetry breaking sector of the mssm is very difficult to observe at the lhc .
we have found that @xmath30 may provide a promising opportunity to simultaneously search for the @xmath0 and the @xmath2 at the lhc .
if susy particles are too heavy to be produced via decays of @xmath0 , the signal of @xmath165 can be observable for @xmath25 and @xmath166 . 1 . for @xmath168 gev , the signal of @xmath165 can be observable for @xmath167 .
2 . for a positive @xmath84 with @xmath169
, the signal of @xmath165 might be observable only if @xmath170 gev and @xmath16 is close to one .
3 . for @xmath171 gev , susy decay modes reduce the signal to below the observable level . in this case ,
new channels to search for the @xmath0 become available .
the most promising discovery mode is the @xmath172 signal from the decay @xmath173 @xcite .
j. gunion , h. haber , g. kane and s. dawson , _ the higgs hunter s guide _ , ( addison - wesley publishing company , redwood city , ca , 1990 ) .
reviews for the mssm can be found in : h.p .
nilles , phys . rep . * 110 * , ( 1984 ) 1 ; p. nath , r. arnowitt and a. chamseddine , _ applied @xmath174 supergravity _ , ictp series in theoretical physics , vol.i , ( world scientific , singapore , 1984 ) ; h. haber and g. kane , phys .
rep . * 117 * , ( 1985 ) 75 ; x. tata , in _ the standard model and beyond _ , edited by j.e .
kim , ( world scientific , singapore , 1991 ) ; tasi-95 lectures , university of hawaii report , uh-511 - 833 - 95 ( 1995 ) , and references therein .
h. baer , m. bisset , c. kao and x. tata , phys . rev . *
d46 * , 1067 ( 1992 ) .
v. barger , m. berger , a. stange and r.j.n .
phillips , phys . rev .
* d45 * ( 1992 ) 4128 ; v. barger , k. cheung , r.j.n .
phillips , and a.l .
stange , phys .
* d46 * ( 1992 ) 4914 .
gunion and l. orr , phys . rev .
* d46 * ( 1992 ) 2052 ; j.f .
gunion , in _
perspectives on higgs physics _ , ed .
g. kane , ( world scientific , singapore , 1992 ) , and references therein
. z. kunszt and f. zwirner , nucl . phys . * b385 * ( 1992 ) 3 .
j. gunion , h. haber and c. kao , phys . rev .
* d46 * ( 1992 ) 2907 .
cms technical proposal , cern / lhcc 94 - 38 ( 1994 ) . atlas technical proposal , cern / lhcc 94 - 43 ( 1994 ) .
j. dai , j.f .
gunion and r. vega , phys . lett .
* b315 * ( 1993 ) 355 ; phys . lett .
* b345 * ( 1995 ) 29 .
d. froidevaux and e. richter - was , z. phys .
* c67 * ( 1995 ) 213 . c. kao and n. stepanov , phys
* d52 * ( 1995 ) 5025 .
h. baer , m. bisset , d. dicus , c. kao and x. tata , phys . rev .
* d47 * ( 1993 ) 1062 ; h. baer , m. bisset , c. kao and x. tata , phys
. rev . * d50 * ( 1994 ) 316
. m. bisset , s. raychaudhuri and p. roy , tata institute report , tifr / th/96 - 05 ( 1996 ) .
h. baer , c. kao and x. tata , phys .
b303 * ( 1993 ) 284 .
d. froidevaux _ et .
atlas internal note , phys - no-74 ( 1995 ) .
j. botts _ et .
* b304 * , ( 1993 ) 159 . c. kao , phys
. lett . *
b328 * ( 1994 ) 420 .
s. l. adler , phys . rev . * 177 * ( 1969 ) 2426 .
j. bell and r. jackiw , nuovo cim .
* a60 * ( 1969 ) 47 .
s. l. adler and w. a. bardeen , phys .
* 182 * ( 1969 ) 1517 .
d. dicus and s. willenbrock , phys . rev . *
d39 * , ( 1989 ) 751 .
kauffman and w. schaffer , phys .
* d49 * ( 1994 ) 551 .
m. spira , a. djouadi , d. graudenz , and p.m. zerwas , nucl . phys .
* b453 * ( 1995 ) 17 .
y. okada , m. yamaguchi and t. yanagida , phys .
262b * , ( 1991 ) 54 ; prog .
* 85 * , ( 1991 ) 1 .
j. ellis , g. ridolfi and f. zwirner , phys .
257b * , ( 1991 ) 83 .
haber and r. hempfling , phys .
* 66 * , 1815 ( 1991 ) ; h.e .
haber , in _
perspectives on higgs physics _ , ed .
kane , ( world scientific , singapore , 1992 ) , and references therein .
m. bisset , _ ph.d .
_ thesis , university of hawaii report , uh-511 - 813 - 94 ( 1994 ) .
e. braaten , j.p .
leveille , phys .
* d22 * ( 1980 ) 715 .
m. drees and k. hikasa , phys .
b240 * , ( 1990 ) 455 ; ( e)-@xmath175 * b262 * , ( 1991 ) 497 .
t. sjstrand , comput .
* 39 * ( 1986 ) 347 ; cern - th.7112/93 ; t. sjstrand and m. bengtsson , comput .
* 43 * ( 1987 ) 367 ; h. u. bengtsson and t. sjstrand , comput . phys . commun .
* 46 * ( 1987 ) 43 .
s. abdullin , a. khanov , n. stepanov , cms tn/94 - 180 ( 1994 ) .
r. brun _ et al .
_ , geant3 , cern dd / ee/84 - 1 ( 1986 ) . f. abe _ et al .
_ , phys . rev .
* d50 * ( 1994 ) 2966 ; a. caner , talk presented at the la thuile meeting , march 1996 .
b. van eijk and r. kleiss , in _ proceedings of the large hadron collider workshop _ , ed .
g. jarlskog and d. rein , vol .
ii , p.183 , cern 90 - 10 ( aachen , 1990 ) . | the prospects of detecting the cp - odd higgs pseudoscalar ( @xmath0 ) in the minimal supersymmetric model via its decay into a @xmath1 boson and the lighter cp - even higgs scalar ( @xmath2 ) at the cern large hadron collider are investigated .
the final state of @xmath3 and @xmath4 , may provide a promising way to simultaneously detect the @xmath0 and the @xmath2 .
the compact muon solenoid detector performance is adopted for a realistic study of observability . in this discovery channel ,
the masses of the @xmath2 and the @xmath0 can be reconstructed .
the impact of supersymmetric decay modes is considered . |
the need for a coherent theory of physics and mathematics together arises from considerations of the basic relationship between physics and mathematics
. why is mathematics relevant to physics @xcite .
one way to see the problem is based on the widely held platonic view of mathematics .
if mathematical systems have an abstract ideal existence outside of space and time and physics describes the property of systems in space and time , then why should the two be related at all ? yet it is clear that they are very closely related .
the problem of the relationship between the foundations of mathematics and physics is not new .
some recent work on the subject is described in@xcite and in @xcite .
in particular the work of tegmark @xcite is is quite explicit in that it suggests that the physical universe is a mathematical universe .
another approach to this problem is to work towards construction of a theory that treats physics and mathematics together as one coherent whole @xcite .
such a theory would be expected to show why mathematics is so important to physics by describing details of the relation between mathematical and physical systems . in this paper
a possible approach to a coherent theory of physics and mathematics is described .
the approach is based on the field of reference frames that follows from the properties of quantum mechanical representations of real and complex numbers @xcite .
the use , here , of reference frames is similar in many ways to that used by different workers in areas of physics @xcite . in general
, a reference frame provides a background or basis for descriptions of systems .
in special relativity , reference frames for describing physical dynamics are based on choices of inertial coordinate systems . in quantum cryptography ,
polarization directions are used to define reference frames for sending and receiving messages encoded in qubit string states .
the use of reference frames here differs from those noted above in that the frames are not based on a preexisting space and time as a background .
instead they are based on a mathematical parameterization of quantum theory representations of real and complex numbers . in particular , each frame @xmath0 in the field is based on a quantum theory representation , @xmath1 of the real and complex numbers where @xmath2 can be viewed as a set of equivalence classes of cauchy sequences of quantum states of qukit strings .
@xmath3 is a set of pairs of these equivalence classes .
the parameter @xmath4 is the base ( @xmath5 for qubits ) , @xmath6 denotes a basis choice for the states of qukit strings that are values of rational numbers , and @xmath7 denotes an iteration stage .
the existence of iterations follows from the observation that the representations of real and complex numbers are based on qukit string states .
these are elements of a hilbert space that is itself defined as a vector space over a field of complex numbers .
consequently one can use the real and complex numbers constructed in a stage @xmath7 frame as the base of a stage @xmath8 frame .
each reference frame contains a considerable number of mathematical structures .
besides @xmath2 and @xmath9 a frame @xmath0 contains representations of all mathematical systems that can be described as structures based on @xmath2 and @xmath10 however frames do not contain physical theories as mathematical structures based on @xmath11 the reason is that the frames do not contain any representations of space and time .
the goal of this paper is to take a first step in remedying this defect by expansion of the domain of each frame to include discrete space and time lattices .
the lattices , @xmath12 in a frame , @xmath13 are such that the number of points in each dimension is given by @xmath14 the spacing @xmath15 and @xmath16 for each lattice , @xmath17 and @xmath18 are fixed with @xmath17 an arbitrary nonnegative integer .
it follows that each dimension component of the location of each point in a lattice is a rational number expressible as a finite string of base @xmath19 digits .
representations of physical systems of different types are also present in each frame . however , the emphasis here is on strings of qukits , @xmath20 present in each frame .
these strings are considered to be hybrid systems in that they are both physical systems and mathematical systems . as mathematical systems , the quantum states of each string , in some basis ,
represent a set of rational numbers .
as physical systems the motion of strings in a frame is described relative to a space and time lattice in the frame .
this dual role is somewhat similar to the concept that information is physical @xcite .
considerable space in the paper is devoted to how observers in a stage @xmath21 ( parent ) frame view the contents of a stage @xmath7 frame .
for an observer , @xmath22 in a frame @xmath0 the numbers in the real and complex number base of the frame are abstract and have no structure .
the only requirement is that they satisfy the set of axioms for real or complex number systems .
points of lattices @xmath23 in the frame are also regarded as abstract and without structure .
the only requirement is that the lattices satisfy some relevant geometric axioms . the view of the contents of a stage @xmath7 frame as seen by an observer @xmath24 in a stage @xmath21 frame , @xmath25 is quite different .
elements of the stage @xmath7 frame that @xmath26 sees as abstract and with no structure , are seen by @xmath27 to have structure .
numbers in @xmath2 are seen by @xmath27 to be equivalence classes of cauchy sequences of states of stage @xmath21 hybrid systems .
space points of stage @xmath7 lattices with @xmath28 space and one time dimension are seen in a stage j-1 frame to be @xmath28 tuples of hybrid systems with the location of each point given by a state of the @xmath28 tuple .
time points are seen to be hybrid systems whose states correspond to the possible lattice time values .
all this and more is discussed in the rest of the paper .
the next section is a brief summary of quantum theory representations and the resulting frame fields @xcite .
section [ pfctpm ] describes a possible approach to a coherent theory of physics and mathematics as the inclusion of space and time lattices in each frame of the frame field .
properties of the lattices in the frames are described .
qukit strings as hybrid systems are discussed in the next section .
their mathematical properties as rational number systems with states as values of rational numbers are described . also included is a general hamiltonian description of the rational number states as energy eigenvalues and a schrdinger equation description of the dynamics of these systems .
section [ fevpf ] describes frame entities as viewed from a parent frame .
included is a description of real and complex numbers , quantum states and hilbert spaces , and space and time lattices .
section [ lhs ] discusses in more detail a stage @xmath21 views of stage @xmath7 lattice points and locations as tuples of hybrid systems and point locations as states of the tuples .
dynamics of these tuples in stage @xmath21 is briefly described as is the parent frame view of the dynamics of physical systems in general .
the last section is a discussion of several points .
the most important one is that frame field description given here leads to a field of different descriptions of the physical universe , one for each frame , whereas there is just one .
this leads to the need to find some way to merge or collapse the frame field to correspond to the accepted view of the physical universe .
this is discussed in the section as are some other points .
whatever one thinks of the ideas and systems described in this work , it is good to keep the following points in mind .
one point is that the existence of the reference frame field is based on properties of states of qukit string systems representing values of rational numbers .
however the presence of a frame field is more general in that it is not limited to states of qukit strings .
reference frame fields arise for any quantum representation of rational numbers where the values of the rational numbers , as states of some system , are elements of a vector space over the field of complex numbers .
another point is that the three dimensional reference frame field described here exists only for quantum theory representations of the natural numbers , the integers , and the rational numbers . neither the basis degree of freedom @xmath6 , nor the iteration stages , @xmath29 are present in classical representations .
this is the case even for classical representations based on base @xmath19 digit or kit strings .
the reason is that states of digit strings are not elements of a vector space over a complex number field . finally ,
although understandable , it is somewhat of a mystery why so much effort in physics has gone into the description of various aspects of quantum geometry and space time and so little into quantum representations of numbers .
this is especially the case when one considers that natural numbers , integers , rational numbers , and probably real and complex numbers , are even more fundamental to physics than is geometry .
in earlier work @xcite quantum theory representations of natural numbers @xmath30 , integers @xmath31 , and rational numbers @xmath32 , were described by states of finite length qukit strings that include one qubit . to keep the description as simple as possible ,
the strings are considered to be finite sets of qukits and one qubit with the qukits and qubit parameterized by integer labels .
the natural ordering of the integers serves to order the set into a string .
this purely mathematical representation of qukit strings makes no use of physical representation of qukit strings as extended systems in space and/or time .
physical representations are described later on in section [ qsshs ] after the introduction of space and time lattices into each frame of the frame field . the qukit ( @xmath33 ) string states are given by @xmath34 where @xmath35 is a @xmath36 valued function with domain @xmath37 and @xmath38 denotes the sign
. the string location of the sign qubit is given by @xmath18 where @xmath39 @xmath17 is any nonnegative integer .
this expresses the range of possible locations of the sign qubit from one end of the string to the other . by convention
@xmath40 has the sign qubit at the right end of the string , @xmath41 at the left end .
the qubit can occupy the same integer location as a qukit .
the reason for the subscript @xmath6 will be clarified later on .
a compact notation is used where the location @xmath18 of the sign qubit is also the location of the @xmath42 point . as examples ,
the base @xmath43 numbers @xmath44 @xmath45 @xmath46 are represented here by @xmath47 @xmath48 @xmath49 respectively .
strings are characterized by the values of @xmath50 for each @xmath51 @xmath52 and @xmath18 the string states @xmath34 give a unified quantum theory representation of natural numbers and integers in @xmath53 @xmath54 and @xmath55 for numbers in @xmath56 @xmath57 ; for numbers in @xmath58 , @xmath59 and there are no restrictions for @xmath55 here @xmath60 is the set of rational numbers expressible as @xmath61 where @xmath62 is any integer whose absolute value is @xmath63 and @xmath64 @xmath65 and @xmath66 is the set of nonnegative integers in @xmath67 the correspondence between the numbers @xmath61 and the states @xmath34 is given by the observation that each @xmath35 corresponds to an integer @xmath68 also , as noted , @xmath18 is the location of the @xmath42 point measured from the right end of the string . since one is dealing with quantum states of qukit strings , states with leading and trailing
@xmath69 are included . in this case
there are many states that are all arithmetically equal even though they are orthogonal quantum mechanically .
for example @xmath70 even though the two states are orthogonal .
the set of states so defined form a basis set that spans a fock space @xmath71 of states .
a fock space is used because the basis set includes states of @xmath33 strings of different lengths .
linear superposition states in the space have the form@xmath72 the @xmath17 and @xmath18 sums are over all positive integers and from @xmath73 to @xmath17 , and the @xmath35 sum is over all functions with domain @xmath74.$ ] as already noted , the states @xmath34 of @xmath33 strings are values of rational numbers .
quantum mechanically they also represent a basis choice of states in the fock space @xmath75 however the choice of a basis is arbitrary in that there are an infinite number of possible choices . here
the choice of a basis is parameterized by the variable @xmath76 since choice of a basis is equivalent to fixing a gauge , @xmath6 is also a gauge fixing parameter .
one can also describe gauge and base transformation operators on these states .
gauge transformations correspond to a basis change ( @xmath6 to @xmath77 ) and base transformations correspond to a base change ( @xmath19 to @xmath78 ) .
details are given in @xcite .
quantum theory representations of real numbers are defined here as equivalence classes of cauchy sequences of states of finite @xmath33 strings that are values of rational numbers .
let @xmath79 be a function on the natural numbers such that for each @xmath80 @xmath81 is a basis state : @xmath82 for each @xmath80 the interval @xmath83 $ ] is the domain of @xmath84 the sequence @xmath79 is a cauchy sequence if it satisfies the cauchy condition:@xmath85 here @xmath86 is the basis state that is the base @xmath19 arithmetic absolute value of the state resulting from the arithmetic subtraction of @xmath87 from @xmath88 the cauchy condition says that this state is arithmetically less than or equal to the state @xmath89}1_{-\ell}\rangle_{k , g}$ ] for all @xmath90 greater than some @xmath91 . here @xmath92 is a string state that represents the number @xmath93 the subscripts @xmath94 in the definition of the cauchy condition indicate that the operations are arithmetic and are defined for base @xmath19 string states in the gauge @xmath6 .
they are _ not _ the usual quantum theory operations . the definition can be extended to sequences @xmath79 of linear superpositions of basis states . in this case one
defines a probability @xmath95 as a sum over all components of @xmath96 and @xmath87 that satisfy the second line of eq .
[ cauchy ] .
the state @xmath79 is cauchy if the probability @xmath97 where@xmath98 two sequences @xmath79 and @xmath99 are equivalent , @xmath100 if the sequence defined by the termwise arithmetic difference of @xmath79 and @xmath99 converges to @xmath73 .
the specific condition for this is expressed by eq .
[ cauchy ] if one replaces @xmath87 with @xmath101 the relation @xmath102 is used to define equivalence classes of cauchy sequences of @xmath33 string states .
the set of these equivalence classes is a quantum theory representation @xmath103 of the real numbers .
quantum equivalence classes of cauchy sequences of states are larger than classical equivalence classes because each quantum equivalence class contains many sequences of states that have no classical correspondent .
however no new equivalence classes appear .
this is a consequence of the fact that each quantum equivalence class contains a basis valued sequence that corresponds to a classical sequence of finite base @xmath19 digit strings .
one can also define a canonical representation of each equivalence class as a sequence @xmath81 of basis states such that if @xmath104 then @xmath81 is an initial part of @xmath105 this is similar to the usual classical canonical representation of an equivalence class of real numbers as an infinite string of digits with a @xmath42 point .
the quantum representation would be an infinite string of qukits with each qukit in one of @xmath19 basis states .
extension of the above to include quantum representations of complex numbers is straightforward .
one method represents complex rational numbers by pairs of states of finite @xmath33 strings .
this what is actually done in computations involving complex numbers .
the state components of the pair represent values of real and imaginary rational numbers .
cauchy sequences of these state pairs are defined by applying the cauchy condition separately to the component sequences of real rational number states and imaginary rational number states
. two cauchy sequences @xmath79 and @xmath99 of state pairs are equivalent , @xmath100 if the termwise arithmetic differences of the real parts , @xmath106 and of the imaginary parts , @xmath107 each converge to @xmath108 the resulting set of equivalence classes of cauchy sequences is a quantum theory representation of the set of complex numbers .
quantum theory representations of real and complex numbers differ from classical representations in two important ways .
one is the presence of the gauge freedom or basis choice freedom .
this is indicated by the @xmath6 subscript in @xmath109 the other difference is based on the observation that states of qukit strings are elements of a hilbert space or a fock space . from a mathematical point of view
these spaces are vector spaces over the fields of real and complex numbers .
it follows from what has been shown that @xmath33 string states , as elements of a vector space over the field of real and complex numbers , can be used to construct other representations of real and complex numbers .
this suggests the possibility of iteration of the construction described here as the quantum representations of real and complex numbers can in turn be used as the base of hilbert and fock spaces for a repeated construction . here
the iteration stage is another degree of freedom for the frame field .
the third degree of freedom arises from the free choice of the base choice for the humber representation .
this choice , denoted by the @xmath19 subscript , is common to both quantum and classical representations .
these three degrees of freedom can be combined to describe a three dimensional reference frame field .
each reference frame @xmath0 is based on a quantum representation @xmath110 of the real and complex numbers .
the subscripts denote the iteration stage @xmath7 , the base @xmath19 , and the gauge @xmath6 .
each reference frame contains representations of hilbert and fock spaces as mathematical structures over @xmath11 because the iteration degree of freedom is directed , it is useful to use genealogical terms to describe the iteration stages
. frames that are ancestors to a given frame @xmath0 occur at stages @xmath111 where @xmath112 frames that are descendants occur at stages @xmath111 where @xmath113 from a mathematical point of view there are several possibilities for the stages : the frame fields can have a finite number of stages with both a common ancestor frame and a set of terminal frames .
the fields can also be one way infinite with either a common ancestor and no terminal stage , or with a terminal stage and no common ancestor ; or they can be two way infinite
. they can also be cyclic .
these last two cases have no common ancestor or terminal frames .
another way to illustrate the frame field structure is to show , schematically , frames emanating from frames .
figure [ pactpm2 ] illustrates a slice of the frame field for a fixed value of @xmath114 each point @xmath115 in the @xmath116 plane denotes a reference frame @xmath117 at the next iteration stage with @xmath118 as its real and complex number base .
three different viewpoints of the real and complex numbers as frame bases are of use here .
@xmath110 is a view from outside the frame field that denotes the position of the numbers with respect to the field degrees of freedom . to an observer inside @xmath0
the elements of @xmath110 are seen as external , abstract , featureless elements .
they have no structure other than that which follows from the requirement that they satisfy the axioms for real and complex numbers .
this assumption is made because it is the view held , at least implicitly by physicists .
it also corresponds to how numbers are treated in physical theories .
the only properties of numbers relevant to theories are those derived from the appropriate axioms .
an observer in a parent frame @xmath119 sees the elements of @xmath110 as having structure .
they are seen as equivalence classes of cauchy sequences of states of finite @xmath33 strings .
this is in addition to their having properties derived from the relevant axiom sets .
the main consideration of this paper is the proposed use of the reference frame field as a possible approach to a coherent theory of physics and mathematics .
this ensures that quantum theory representations of natural numbers , integers , rational numbers , and real and complex numbers will play a basic role in the theory .
so far the reference frames contain mathematical systems .
these include quantum theory representations of numbers , qukit strings , and representations of other mathematical systems as structures based on the different types of numbers .
physical theories and systems are not present in the frames .
the reason is that there are no representatios of space and time in the frames .
these are needed for theories to describe the kinematics and dynamics of systems moving in space and time .
this must be remedied if the frame field is to be an approach to a coherent theory .
one way to fix this is to expand the domains of the frames to include physical systems and descriptions of their dynamics .
a first step in this direction is to expand the domain of each frame in the field to include discrete lattices of space and time .
the reason for working with discrete instead of continuum space and time will be noted later . to be more specific , each frame
@xmath0 includes a set @xmath120 of space and time lattices .
each lattice @xmath121 in the set is such that the number @xmath122 of points in each dimension is finite and the spacing @xmath123 of points is also finite . in this paper
the number @xmath28 of space dimensions in the lattice will be arbitrary . to keep things simple
, the number @xmath122 of points and spacing @xmath123 in each of the @xmath28 space dimensions and the time dimension will be assumed to be the same .
it should be noted that , to an observer in a frame , the points in each space and time lattice in a frame _ are _ ( emphasis on `` are '' ) points of space and time relative to that frame .
they are not merely mathematical representations or descriptions of some external space and time . in addition
an observer in a frame sees the points of space and time lattices in in the frame as abstract , featureless points with no structure .
the only requirement is that the lattices should satisfy appropriate geometrical axioms .
a restriction on the lattices in frames is that the values of @xmath122 and @xmath123 for each space and time lattice in the set @xmath120 are given by @xmath124 here @xmath17 is an arbitrary nonnegative integer , @xmath125 and @xmath19 is the same integer @xmath126 that is the @xmath19 subscript in the frame label , as in @xmath127 because of this restriction the lattices @xmath121 will be labeled from now on as @xmath128 based on eq .
[ mdelta ] , the location of each space or time point in each dimension @xmath129 is given by a rational number value , @xmath130 where @xmath131 is a nonnegative integer @xmath132 even though these requirements might seem restrictive , they are sufficiently general to allow lattices of arbitrarily small spacing and arbitrarily many points . also they can be used to describe sequences of lattices that become continuous in the limit .
an example of such a sequence is given by setting @xmath133 $ ] and increasing @xmath17 without bound .
it follows from this description that the points of a lattice , @xmath12 with @xmath28 space and one time dimension can be taken to be @xmath134-tuples of rational numbers .
the value of each number in the tuple is expressible as a finite string of base @xmath19 digits .
one may hope that the structure of space and time , as a continuum or as some other structure , will emerge when one finds a way to merge the frames in the frame field . ]
so far each frame contains in its domain , space and time lattices , and strings of qukits that are numbers .
it is reasonable to expect that it also contains various types of physical systems . for the purposes of this paper
the types of included physical systems do not play an important role as the main emphasis here is on qukit string systems . also in this first paper descriptions of system dynamics will be limited to nonrelativistic dynamics .
it follows from this that each frame includes a description of the dynamics of physical systems based on the space and time lattices in the frame .
the kinematics and dynamics of the systems are expressed by theories that are present in the frame as mathematical structures over the real and complex number base of the frame .
this is the case irrespective of whether the physical systems are particles , fields , or strings , or have any other form .
one reason space and time are described as discrete lattices instead of as continua is that it is not clear what the appropriate limit of the discrete description is .
as is well known , there are many different descriptions of space and time that are present in the literature .
the majority of these descriptions arise from the need to combine quantum mechanics with general relativity .
they include use of various quantum geometries @xcite , and space time as a foam @xcite and as a spin network as in loop quantum gravity @xcite .
these are in addition to the often used assumption of a fixed flat space and time continuum that serves as a background arena for the dynamics of all physical systems , from cosmological to microscopic .
space and time may also be emergent in an asymptotic sense @xcite .
the fact that there are many different lattices in a frame each characterized by different values of @xmath17 and @xmath18 and that each can serve as a background space and time is not a problem .
this is no different than the fact that one can use many different space and time lattices with different spacings and numbers of points to describe discrete dynamics of systems .
so far the domain of each frame contains space and time lattices , many types of physical systems ( such as electrons , nuclei , atoms , etc . ) and physical theories as mathematical structures based on the real and complex numbers .
these describe the kinematics and dynamics of these systems on the lattices .
also included are qukit strings .
states of these strings were seen to be values of rational numbers .
these were used to describe real and complex numbers as cauchy sequences of these states .
here it is proposed to consider these strings as systems that can either be numbers , i.e. mathematical systems , or they can be physical systems . because of this dual role , they are referred to as hybrid systems . as such they will be seen to play an important role .
support for this proposal is based on the observation that the description of qukit strings as both numbers and physical systems is not much different than the usual view in physics regarding qubits and strings of qubits . as a unit of quantum information
, the states of a qubit can be @xmath135 and @xmath136 which denote single digit binary numbers .
the states can also be @xmath137 as spin projection states of a physical spin @xmath138 system . in the same way strings of qubits are binary numbers in quantum computation , or they can represent physical systems such as spins or atoms in a linear ion trap @xcite . to be blunt about it , ``
information is physical '' @xcite , and information is mathematical .
also it is reasonable to expect that the domain of a coherent theory of physics and mathematics together would contain systems that are both mathematical systems and physical systems .
the hybrid systems are an example of this in that they are number systems , which are mathematical systems , and they are physical systems .
let @xmath139 denote a hybrid system in @xmath0 that contains @xmath17 qukits and one qubit .
@xmath17 is any nonnegative integer , @xmath18 is any nonnegative integer @xmath140 @xmath7 is the iteration stage of any frame containing these systems , and @xmath78 is the base ( or dimension of the hilbert state space ) of the @xmath141 in the string system .
note that @xmath78 can be different from the base @xmath19 of the frame @xmath0 containing these systems .
the different @xmath141 in the system are distinguished by labels in the integer interval @xmath142.$ ] the qubit has label @xmath18 where @xmath143 the canonical ordering of the integers serves to order the @xmath141 and qubit into a string system . the presence of the sign qubit is needed if the states of the hybrid systems are to be values of rational numbers . since the qubit also corresponds to the @xmath144 point , the value of @xmath18 gives the location of the sign and @xmath144 point in @xmath145 note that there is a change of emphasis from the usual description of numbers . in the usual description
, strings of base @xmath78 digits , such as @xmath146 with @xmath147 are called rational numbers . here
states , such as @xmath148 are called _ values _ of rational numbers .
the hybrid system will also be referred to as a rational number system .
the reason is that the set of all basis states of a hybrid system correspond to a set of @xmath149 values of rational numbers .
the type of number @xmath150 , represented is characterized by the value of @xmath18 and state of the sign qubit in @xmath145 a rational number instead of a rational number system .
this would agree with the usual physical description of systems .
for example a physical system of a certain type is a proton , not a proton system .
however referring to @xmath139 as a rational number instead of a rational number system seems so at odds with the usual use of the term that it is not done here . ]
as was noted , the states of the @xmath139 systems in a frame @xmath0 are elements of a hilbert space @xmath151 in the frame .
the choice of a basis set or gauge @xmath77 fixes the states of @xmath139 that are values of rational numbers .
these states are represented as @xmath152 often the @xmath153 subscripts on the states will be dropped as they will not be needed for the discussion .
the description of the hybrid systems as strings of qukits is one of several possible structures .
for example , as physical systems that move and interact on a space lattice @xmath154 the strings could be open with free ends or closed loops . in this case , aspects of string theory @xcite may be useful in describing the physics of the strings .
whatever structure the hybrid systems have , it would be expected that , as bound systems , they have a spectrum of energy eigenstates described by some hamiltonian @xmath155 if the rational number states of the hybrid system , @xmath156 are energy eigenstates , then one has the eigenvalue equation@xmath157 where @xmath158 is the energy eigenvalue of the state @xmath159 the superscript @xmath160 on @xmath161 allows for the possibility that the hamiltonian depends on the type of hybrid system .
the gauge variable has been removed because the requirement that eq .
[ shenergy ] is satisfied for some choice of @xmath162 fixes the gauge or basis to be the eigenstates of @xmath155 since @xmath161 is not known , neither is the dependence of @xmath158 on @xmath163 the existence of a hamiltonian for the @xmath160 hybrid systems means that there is energy associated with the values of rational numbers represented as states of hybrid systems . from this
it follows that there are potentially many different energies associated with each rational number value .
this is a consequence of the fact that each rational number value has many string state representations that differ by the number of leading and trailing @xmath164 one way to resolve this problem is to let the energy of a hybrid system state with no leading or trailing @xmath69 be the energy value for the rational number represented by the state . in this way one has , for each @xmath51 a unique energy associated with the value of the rational number shown by the state .
one consequence of this association of energy to rational number values is that to each cauchy sequence of rational number states of hybrid systems there corresponds a sequence of energies .
the energy of the @xmath165 state in the sequence is given by @xmath166 it is not known at this point if the sequence of energies associated with a cauchy sequence of hybrid system states converges or not . even if energy sequences converge for cauchy sequences in an equivalence class , the question remains whether or not the energy sequences converge to the same limit for all sequences in the equivalence class .
the above description is valid for one hybrid system . in order to describe more than one of these systems , another parameter ,
@xmath167 is needed whose values distinguish the states of the different @xmath139 systems . to this end the states @xmath168 of a system are expanded by including a parameter @xmath169 as in @xmath170 in this case the state of two @xmath139 is given by @xmath171 where @xmath172 .
this allows for the states of the two systems to have the same @xmath173 and @xmath35 values .
pairs of hybrid systems are of special interest because states of these pairs correspond to values of complex rational numbers .
the state of one of the pairs is the real component and the other is the imaginary component . since these components have different mathematical properties , the corresponding states in the pairs of states of hybrid systems must be distinguished in some way .
one method is to distinguish the hybrid systems in the pairs by an index @xmath174 added to @xmath175 as in @xmath176 in this case states of @xmath177 and @xmath178 are values of the real and imaginary components of rational numbers . in this case
complex numbers are cauchy sequences of states of pairs , @xmath179 of hybrid systems .
as might be expected , the kinematics and dynamics of hybrid systems @xmath139 in a frame @xmath0 are described relative to a space and time lattice in the frame .
for example a schrdinger equation description of two hybrid systems interacting with one another is given by @xmath180 @xmath181 is the discrete forward time derivative where @xmath182 here @xmath183 is the state of the two hybrid systems at time @xmath184 the hamiltonian can be expressed as the sum of a hamiltonian for the separate systems and an interaction part as in@xmath185 for two hybrid systems @xmath186 where@xmath187 the first term of @xmath188 is the kinetic energy operator for the @xmath189 @xmath139 system .
the second term is the hamiltonian for the internal states of the system .
it is given by eq .
[ shenergy ] .
also @xmath190 is the mass of @xmath139 .
@xmath191 and @xmath192 are the discrete forward and backward space derivatives , and @xmath193 is planck s constant .
the dot product indicates the usual sum over the product of the @xmath28 components in the derivatives .
note that a possible dependence of the mass of @xmath139 on @xmath194 has been included .
the question arises regarding how one should view @xmath195-tuple hybrid systems as physical systems .
should they be regarded as @xmath195 independent systems each with its own hamiltonian ( @xmath196 in eq .
[ h0hint ] ) or as systems bound together with energy eigenstates that are quite different from those of the single hybrid systems in isolation .
one way to shed light on this question is to examine physical representations of number tuples in computers .
there @xmath195-tuples of numbers are represented as @xmath195 strings of bits or of qubits ( spin @xmath138 systems ) bound to a background matrix of potential wells where each well contains one qubit .
the locations of the qubits in the background matrix determines their assembly into strings and into tuples of strings .
here it is assumed that @xmath195 tuples of hybrid systems consist of @xmath195 @xmath139 systems bound together in some fashion .
details of the binding , and its effect on the states of the individual @xmath139 in the @xmath195-tuple are not known at this point .
however it will be assumed that the effect is negligible . in this case
the energy of each component state @xmath197 in the @xmath195 tuple will be assumed to be the same as that for an individual @xmath139 system .
then , the energy of the state @xmath198 is the sum of the energies of the individual component states .
also the energy is assumed to be independent of the @xmath169 values .
this picture is supported by the actual states of computers and their computations .
the background potential well matrix that contains the @xmath195-tuples of qubit string states is tied to the computer .
since the computer itself is a physical system , it can be translated , rotated , or given a constant velocity boost . in all these transformations
the states of the qubit strings in the @xmath195-tuples and the space relations of the @xmath195 qubit strings to one another is unchanged .
these parameters would be changed if two computers collided with one another with sufficient energy to disrupt the internal workings .
this picture of each frame containing physical systems and a plethora of different hybrid systems and their tuples may seem objectionable .
however , one should recall that here one is working in a possible domain of a coherent theory of physics and mathematics together . in this case
the domain might be expected to include many types of hybrid systems that have both physical and mathematical properties .
this is in addition to the presence of physical systems and mathematical systems .
so far it has been seen that each frame , @xmath13 in the frame field contains a set @xmath120 of space and time lattices where the number of points in each dimension and the point spacing satisfy eq .
[ mdelta ] for some @xmath17 and @xmath18 .
the frame also contains qukit string systems as hybrid @xmath199 systems and various tuples of these systems . here
@xmath200 need have no relation to @xmath50 each frame also contains physical theories as mathematical structures based on @xmath110 the real and complex number base of the frame .
these theories describe the kinematics and dynamics of physical systems on the space and time lattices in the frame . for quantum systems
these theories include hilbert spaces as vector spaces over @xmath10 since this is true for every frame , it is true for a frame @xmath0 and for a parent frame @xmath201 this raises the question of how entities in a frame are seen by an observer in a frame at an adjacent iteration stage .
as was noted in section [ firf ] , it is assumed that ancestor frames and their contents are not visible to observers in descendant frames ; however , descendant frames and their contents are visible to observers in ancestor frames .
it follows that observers in frame @xmath0 can not see frame @xmath119 or any of its contents .
however , observers in @xmath119 can see @xmath0 and its contents .
one consequence of the relations between frames at different iteration stages is that entities in a frame , that are seen by an observer in the frame as featureless and with no structure , correspond to entities in a parent frame that have structure .
for example , elements of @xmath110 are seen by an observer in @xmath0 as abstract , featureless objects with no properties other than those derived from the relevant axiom sets .
however , to observers in a parent frame @xmath25 numbers in @xmath110 are seen as equivalence classes ( pairs of equivalence classes for @xmath3 ) of cauchy sequences of states of base @xmath19 qukit string systems .
thus entities that are abstract and featureless in a frame have structure as elements of a parent frame .
it is useful to represent these two in - frame views by superscripts @xmath7 and @xmath21 .
thus @xmath202 and @xmath203 denote the stage @xmath21 and stage @xmath7 frame views of the number base of frame @xmath127 they are often referred to in the following as parent frame images of @xmath110 .
the distinction between elements of a frame and their images in a parent frame exists for other frame entities as well .
the state@xmath204 in @xmath205 corresponds to the state @xmath206 in @xmath207 which is the parent frame image of @xmath208 in the above @xmath209 is a featureless abstract complex number in @xmath3 whereas @xmath210 as an element of @xmath211 is an equivalence class of cauchy sequences of hybrid system states .
the use of stage superscripts and subscripts applies to other frame entities , such as hybrid systems , physical systems , and space and time lattices.a hybrid system @xmath199 in @xmath0 has @xmath212 as a parent frame image .
states @xmath213 of @xmath199 are vectors in the hilbert space @xmath214 in @xmath127 states @xmath215 of @xmath212 are vectors in @xmath216 the two states are different in that @xmath213 is an eigenstate of an operator whose corresponding eigenvalue @xmath217 is a rational real number with no structure .
the state @xmath215 of @xmath212 is an eigenstate of a number value operator whose corresponding eigenvalues are rational real numbers in @xmath218 corresponding to @xmath219 these eigenvalues are equivalence classes of cauchy sequences of states of hybrid systems @xmath220 in a @xmath21 stage frame . here
@xmath221 are fixed and @xmath222 and @xmath223 for the @xmath165 term in the sequence .
this accounts for the fact that each state in the sequence is a state of a different hybrid system with @xmath224 held fixed .
the eigenvalue equivalence class is a base @xmath78 real rational number .
since it is an element of @xmath218 , it contains a constant sequence of hybrid system states if and only if all prime factors of @xmath78 are factors of @xmath225 if this is the case then one can equate the equivalence class to a single state of a hybrid system to conclude that the eigenvalue associated with @xmath215 is a state of a hybrid system @xmath220 in stage @xmath21 frame .
is the view of a physical system @xmath226 as seen by an observer in a stage @xmath7 frame .
the image of this view in a stage @xmath21 frame is denoted by @xmath227 the difference between the two is that physical properties of @xmath228 as eigenvalues of operators over @xmath214 are featureless abstract real numbers .
properties of @xmath229 as operator eigenvalues , are equivalence classes of cauchy sequences of hybrid system states .
a frame independent description would be expected to appear only asymptotically when the frames in the field are merged . ] if @xmath78 has prime factors that are not prime factors of @xmath19 ( such as @xmath230 and @xmath231 ) , the eigenvalue for the eigenstate @xmath215 is still a real rational number . however , as an equivalence class of cauchy sequences of base @xmath19 hybrid system states , ] it does not contain a constant sequence of hybrid system states .
instead it contains a sequence that corresponds to an infinite repetition of a base @xmath19 hybrid system state ( just like the decimal expansion of @xmath232 ) .
a similar representation holds for parent frame views of point locations of lattices .
let @xmath23 denote a lattice of @xmath28 space dimensions and one time dimension in a frame @xmath127 the components @xmath233 ( with @xmath234 ) of the @xmath28 space locations @xmath235 of the lattice points , @xmath236 are such that @xmath233 is a rational real number .
the lattice points are abstract and have no structure other than that imparted by the values @xmath237 as rational real numbers in @xmath238 the @xmath233 have no structure other than the requirement that they are both rational and real numbers .
the view or image of @xmath23 from the position of an observer in a stage @xmath21 parent frame is denoted by @xmath239 the image points and locations of points in @xmath240 are denoted by @xmath241 and @xmath242 the space point locations @xmath243 are different from the @xmath235 in that they have more structure .
this follows from the fact that they are @xmath28 tuples of rational real numbers in @xmath244 it follows from this that each component , @xmath245 of a space point location , @xmath246 in @xmath240 is an equivalence class of cauchy sequences of states of hybrid systems .
since each of the @xmath28 equivalence classes is a real number equivalent of a rational number value , the equivalence class includes many ( numerically ) constant sequences of hybrid system states .
the existence of constant sequences follows from the observation that the @xmath19 subscript of the lattice image is the same as that for @xmath218 .
the states in the different sequences differ by the presence of different numbers of leading and trailing @xmath164 a useful way to select a unique constant sequence is to require that all states in the sequence be a unique state of the hybrid system @xmath220 where the @xmath247 subscripts are the same as those for @xmath239 replacement of the sequence by its single component state gives the result that , for each @xmath248 @xmath249 is a state of @xmath250 it follows from this that the locations of space points @xmath251 in @xmath12 as viewed from a stage @xmath21 frame , are seen as states of a @xmath28-tuple @xmath252 of hybrid systems in the stage @xmath21 frame .
these states correspond to @xmath28-tuples of rational number values .
a similar representation of the time points in the lattices is possible for the real rational number time values . in this case
the time values of a lattice are seen in a parent frame as rational number states of a hybrid system .
the above description of a stage @xmath21 frame view of lattices gives point locations of @xmath240 as states of a @xmath28 tuple of hybrid systems @xmath253 for the space part and states of another hybrid system @xmath254 for the time part .
the superscripts @xmath35 and @xmath255 denote the hybrid systems associated with space and time points respectively .
this strongly suggests that each space point image @xmath256 in the space part of a parent frame image , @xmath257 of @xmath23 be identified with a @xmath28 tuple , @xmath253 of hybrid systems and each time point image @xmath258 in the time part be identified with a single hybrid system , @xmath259 for each image point @xmath260 the location is given by the state of the @xmath28 tuple @xmath261 of hybrid systems in a parent frame that is associated with the space point image .
similarly the location of each image time point in @xmath240 is given by the state of the hybrid system , @xmath262 associated with the point .
this shows that set of all parent frame images of the @xmath263 space points of @xmath23 become a set of @xmath263 @xmath28 tuples of parent frame hybrid systems , @xmath261 with the state of each @xmath28 tuple corresponding to the image space point location in @xmath239 the parent frame images of the @xmath264 time points of @xmath23 become a set of @xmath149 hybrid systems @xmath259 each is in a different state corresponding to the different possible lattice time values .
figure [ pactpm3 ] illustrates the situation described above for lattices with one space and one time dimension .
the stage @xmath7 and @xmath265 lattice points are shown by the intersection of the grid lines . for the stage @xmath7 lattice the points correspond to rational number pairs whose locations are given by the values of the pairs of numbers . for the stage @xmath21 image lattice
the points correspond to pairs of hybrid systems , one for the space dimension and one for the time dimension .
point locations are given by the states of the hybrid system pairs .
non relativistic world paths for physical systems @xmath266 and its stage @xmath21 image @xmath267 are also shown .
it is of interest to compare the view here with that in @xcite .
tegmark s explicitly stated view is that real numbers , as labels of space and time points , are distinct from the points themselves .
this is similar to the setup here in that points of parent frame images of lattices are tuples of hybrid systems , and locations or labels of the points of parent frame images of lattice are states of tuples of hybrid systems .
the differing views of hybrid systems as either number systems or physical systems may seem strange when viewed from a perspective outside the frame field and in the usual physical universe .
however it is appropriate for a coherent theory of physics and mathematics together as such a theory might have systems that represent different entities , depending on how they are viewed .
the description of parent frame images of lattice space points and their locations as @xmath28 tuples of hybrid systems and states of the tuples , means that the image of each point has a mass .
the mass is equal to that of the @xmath28 tuple of hybrid systems associated with each point image .
the ( rest ) masses of all space points in an image lattice @xmath240 should be the same as the @xmath28 tuples of hybrid systems associated with each point are the all the same .
however each of the tuples is in a different image state @xmath243 that corresponds to the different locations of each point image .
each component @xmath249 of @xmath243 corresponds to a hybrid system state @xmath268 of a component hybrid system @xmath269 in the @xmath28 tuple ( note the subscript @xmath129 ) .
each of these component hybrid system states is an energy eigenstate of a hybrid system hamiltonian .
the corresponding energy eigenvalue , @xmath270 is defined by eq . [ shenergy ] .
here @xmath173 is positive as the lattice component locations are all @xmath271 if the component hybrid systems in a @xmath28 tuple do not interact with one another , then the energy associated with a parent frame image lattice location is the sum of the component energies . in this case the energy associated with the location image , @xmath272 , of @xmath235 is given by@xmath273here @xmath272 denotes the tensor product of all @xmath28 component states @xmath274 if the component systems in a @xmath28 tuple do interact with one another then the situation becomes more complex as the hamiltonian and energy eigenvalues must take account of the interactions . at this point the specific dependence of the energy on the parent frame image of the lattice point locations
is not known as it depends on the properties of the hybrid system hamiltonian .
nevertheless the existence of energies associated with locations of points of parent frame images of a space lattice is intriguing .
one should note , though , that this association of energy to space points holds only for parent frame images .
it does not extend to lattices @xmath23 when viewed by an observer in @xmath127 this aspect is one reason why one needs to do more work , particularly on the merging of frames in the frame field .
it is quite possible that , in the case of a cyclic frame field , some aspects of the association of energy with space points in the same frame will be preserved .
note that for cyclic frame fields the restriction that an observer can not see ancestor frames or their contents must be relaxed .
the reason is that ancestor frames are also descendant frames .
the description of the motion of hybrid systems and other physical systems in a stage @xmath7 frame , as seen from a parent stage @xmath21 frame , is interesting .
the reason is that the dynamics and kinematics of systems are based on a parent frame image lattice , @xmath257 whose space points and point locations are @xmath28 tuples , @xmath253 , and tensor product states @xmath275 of the @xmath28 tuples of hybrid systems .
the time points and point locations are hybrid systems , @xmath254 and states of these systems .
the @xmath276 and @xmath255 superscripts allow for the possibility that the hybrid systems associated with space point images may be different from those associated with time images .
it follows that the stage @xmath21 frame description of the motion and dynamics of stage @xmath7 systems is described relative to the states of certain hybrid systems in the parent frame . to understand this
consider a simple example where @xmath277 denotes the state of some physical system @xmath266 in a stage @xmath7 frame at position @xmath276 and at time @xmath184 the pair @xmath278 denote a point in a stage @xmath7 frame space and time lattice @xmath23 which , for simplicity , consists of one space and one time dimension .
the stage @xmath7 frame time evolution of the state @xmath277 is given by a discrete schrdinger equation,@xmath279 here @xmath280 is the discrete forward time derivative defined by@xmath281 as a sum of kinetic and potential terms , the hamiltonian , @xmath282 has the form@xmath283 to keep things simple , the description is restricted to just one @xmath266 system interacting with an external potential . in this case @xmath284 also @xmath193 is planck s constant and @xmath285 is the mass of system @xmath266 . in the above ,
the forward and backward discrete derivatives @xmath286 and @xmath287 are defined similar to the forward time derivative .
one has @xmath288
this description of the time development and hamiltonian for a system @xmath266 is a description in a stage @xmath7 frame . viewed from a parent stage @xmath21 frame the image of the schrdinger equation , eq .
[ schrj ] describes the motion of the system @xmath226 in the image lattice @xmath239 since the space and time point locations in the image lattice are states of hybrid systems @xmath289 and @xmath254 , the image schrdinger equation describes the motion of system @xmath226 relative to these states .
the image equation is given by@xmath290 the image state , @xmath291 is the same state in the hilbert space @xmath292 as @xmath293 is in @xmath208 here @xmath294 and @xmath295 are shorthand notations for the hybrid system states , @xmath296 and @xmath297 of @xmath289 and @xmath254 the hamiltonian @xmath298 is given by @xmath299 the potential @xmath300 is a function of the states @xmath296 of @xmath301 the value of @xmath302 is a real number in @xmath218 that is expected to be the same as that of @xmath303 in @xmath304 the forward and backward discrete derivatives are expressed by equations similar to eq .
[ deltafb]:@xmath305 in these equations @xmath306 and @xmath307 denote the hybrid system states @xmath308 and @xmath309 where the subscript @xmath310 denotes arithmetic addition and subtraction .
for example if @xmath311 in binary then @xmath312 this use of states of @xmath289 and @xmath254 as point locations of an image space and time lattice must be reconciled with the observation that these hybrid systems are dynamical systems that move and interact with one another and with other physical systems
. if the @xmath289 and @xmath254 and their states are to serve as points and point locations of a space and time lattice image , they must be dynamically very stable and resistant to change .
this suggests that these systems must be very massive and that their interactions with each other and other systems is such that state changes occur very rarely , possibly on the order of cosmological time intervals .
is described in a stage @xmath21 frame . here
@xmath200 need have no relation to @xmath50 ] the reason for these restrictions on the properties of space and time hybrid systems is that one expects the space and time used to describe motion of systems to be quite stable and to change at most very slowly .
changes , if any , would be expected to be similar to those predicted by the einstein equations of general relativity .
it is to be emphasized that the work presented so far is only a beginning to the development of a complete framework for a coherent theory of physics and mathematics together . not only that but one must also find a way to reconcile the multiplicity of universes , one for each frame , to the view that there is only one physical universe . one way to achieve reconciliation is to drop the single universe view and to relate the multiplicity of frame representations of physics and mathematics to the different many physical universes view of physics .
these include physical universes in existing in different bubbles of space time @xcite , and other descriptions of multiple universes @xcite including the everett wheeler description @xcite . whether any of these are relevant here or not will have to await future work .
if one sticks to the single physical universe view then the frames with their different universes and space and time representations need to be merged or collapsed to arrive in some limit at the existing physical universe with one space time .
this applies in particular to the iteration stage and gauge degrees of freedom as their presence is limited to the quantum representation of numbers .
one expected consequence of the merging is that it will result in the emergence of a single background space time as an asymptotic limit of the merging of the space time lattices in the different frames .
whether the ultimate space time background is a continuum , a foam @xcite , or has some other form , should be determined by details of the merging process .
in addition the merging may affect other entities in the frames .
physical systems , denoted collectively as @xmath266 , may become the observed physical systems in the space time background .
in addition , hybrid systems may split into either physical systems or mathematical systems .
one potentially useful approach to frame merging is the use of gauge theory techniques @xcite to merge frames in different iteration stages .
one hopes that some aspects of the standard model @xcite in physics will be useful here .
this will require inclusion of relativistic treatment of systems and quantum fields in the frames .
this look into a possible future approach emphasizes how much there is to accomplish .
nevertheless one may hope that the work presented here is a beginning to the development of a coherent theory .
the expansion of the frames in the frame field to include , not only mathematical systems , but also space and time lattices , and hybrid systems that are both mathematical systems and physical systems , seems reasonable from the viewpoint of a coherent theory of physics and mathematics together .
one might expect such a theory to contain systems that can be either physical systems or mathematical systems .
the use of massive hybrid systems to be stage @xmath21 frame images of points of space and time lattices in stage @xmath7 frames suggests that there must be different types of hybrid systems .
for example , stage @xmath7 theoretical predictions of the values of some physical quantity @xmath314 are , in general , real numbers in @xmath315 their images in a stage @xmath21 parent frame are equivalence classes of cauchy sequences of states of hybrid systems @xmath220 where @xmath316 are dependent on positions in the cauchy sequences . if the predicted values are rational numbers expressible as a finite string of base @xmath19 digits , then the stage @xmath21 values can be expressed as states of @xmath220 rather than equivalence classes of sequences of states .
if the images of properties of the physical quantities @xmath314 are to be reflected in the properties of the hybrid systems , then different types of hybrid systems , such as @xmath317 must be associated with different physical quantities . whether these descriptions of parent frame images as hybrid systems will remain or will have to be modified , remains to be seen .
however , it should be recalled that these images are based on the dual role played by values of rational numbers both as mathematical systems and as locations of components of points in the lattices .
recall that the notion of a point in a lattice is separated from the location of the point just as the notion of a number , as a mathematical system , is separated from the value of a number .
this use of number and number value is different from the usual use in mathematics in that expressions , as @xmath318 are usually considered as rational numbers and not as values of rational numbers . here a rational number , as a hybrid system , is similar to the usual mathematical concept of a set of rational numbers as a model of the rational number axioms . in conclusion , it is worth reiterating the last paragraphs at the end of the introduction .
whatever one thinks of the ideas presented in the paper , the following points should be kept in mind .
two of the three dimensions of the field of reference frames are present only for quantum theory representations of the real and complex numbers .
these are the gauge or basis degree of freedom and the iteration stage degree of freedom .
they are not present in classical descriptions .
the number base degree of freedom is present for both quantum and classical representations based on rational number representations by digit strings .
the presence of the gauge and iteration degrees of freedom in the quantum representation described here is independent of the description of rational number values as states of qukit string systems .
any quantum representation of the rational numbers , such as states @xmath319 of integer pairs , where the states are elements of a vector space will result in a frame field with gauge and iteration degrees of freedom . finally , the importance of numbers to physics and mathematics should be emphasized .
it is hoped that more work on combining quantum physics and the quantum theory of numbers will be done .
the need for this is based on the observation that natural numbers , integers , rational numbers , and probably real and complex numbers , are even more fundamental to physics than is geometry .
this work was supported by the u.s .
department of energy , office of nuclear physics , under contract no .
de - ac02 - 06ch11357 .
e. wigner , the unreasonable effectiveness of mathematics in the natural sciences , _ commum .
pure and applied math . _ * 13 * 001 ( 1960 ) , reprinted in e. wigner , _ symmetries and reflections _ , ( indiana univ . press , bloomington in 1966 ) , pp222 - 237 .
p. benioff , properties of frame fields based on quantum theory representations of real and complex numbers , contemporary mathematics 482 , _ advances in quantum computation _ , k. mahdavi and d. doslover , eds .
, september 10 - 13 , 2007 , university of texas , tyler , tx , pp .
125 - 163 ; arxiv:0709.2664 .
n. seiberg , emergent spacetime in _ the quantum structure of space and time _ , proceedings of the 23rd solvay conference , brussels , belgium , dec . 1 - 3 , 2005 , d. gross , m. henneaux , a. sevrin , eds . , world scientific press , new jersey . | this work is based on the field of reference frames based on quantum representations of real and complex numbers described in other work . here
frame domains are expanded to include space and time lattices .
strings of qukits are described as hybrid systems as they are both mathematical and physical systems . as mathematical systems
they represent numbers . as physical systems in each frame
the strings have a discrete schrodinger dynamics on the lattices .
the frame field has an iterative structure such that the contents of a stage j frame have images in a stage j-1 ( parent ) frame .
a discussion of parent frame images includes the proposal that points of stage j frame lattices have images as hybrid systems in parent frames .
the resulting association of energy with images of lattice point locations , as hybrid systems states , is discussed .
representations and images of other physical systems in the different frames are also described . |
in the early days after the discovery of extragalactic radio sources , it was a widely held belief that the relativistic electrons responsible for the observed synchrotron emission are secondary electrons from pp - interactions of accelerated protons on ambient gas @xcite .
protons and ions were known to be the by far dominant species in the observed cosmic rays , and this was expected to be mirrored at their acceleration sites .
nevertheless , the picture was soon given up , since it would require enormous amounts of target matter in the jets which is inconsistent with plausible energetics and the non - observation of emission lines and bremsstrahlung .
an electron - positron composition of the jets was also suggested from theory claiming that the plasma feeding radio jets should be due to pair production in the ergosphere of a maximally rotating black hole @xcite .
the dielectric properties of radio jets inferred from observations of weak faraday rotation and circular polarization in the jet of 3c279 seem to lend further support to a light composition @xcite .
however , this picture also bears several fallacies .
target matter does not have to be present to obtain efficient cooling of accelerated protons .
photo - production of secondary particles and synchrotron emission can become important , if the proton energy is high enough , since the cooling rate for these processes decreases with energy .
in fact , in a statistical acceleration process in which there is enough time and space to balance acceleration energy gains against energy losses , it is an inevitable consequence that the protons reach ultrahigh energies @xcite .
even if the original composition of the jet plasma were light , polluting baryons from the ambient medium would quickly take over most of the jet s momentum , so that the acceleration mechanism must eventually tap baryonic kinetic energy .
in fact , for gamma ray bursts , this is believed to be the crucial explanation for the fact that the burst energy is tapped only by shock fronts far away from the site of the original pair fireball @xcite .
therefore , it is the most natural assumption that the observed nonthermal emission is partly due to the accelerated electrons , and partly due to the accelerated protons @xcite .
inferences of the plasma composition based on measurements of polarization and faraday rotation are based on the assumption that the electron distribution observed in the optically thin synchrotron regime traces down to lower energies which is known to lead to theoretical inconsistencies @xcite . since the radiative properties are determined by the particles with the highest energies but the dielectric properties by those with the lowest energies
, there is a general mismatch in the conclusions drawn from observations sensitive to either regime and one must be careful with claims about jet composition .
it may be a good advice not be too narrow - minded when confronted with a new observational result such as multi - tev emission from blazars and to include as many independent facts as possible in its interpretation .
if the gamma rays were indeed primarily of a hadronic origin , there are a number of corollaries which allow to falsify the claim , whereas the arguments for a purely leptonic origin of the gamma rays boil down to some version of ockham s razor ( `` it is more economic to use the observed electrons to model the gamma ray emission '' ) .
i wonder whether this is enough to get around the symptomatic fact that there was no prediction of multi - tev emission from leptonic models prior to the observations .
moreover , there are problems with leptonic models , such as the missing intrinsic curvature in the multi - tev spectrum of mrk 501 and the surprisingly low magnetic field values as was pointed out in ref . @xcite .
although the most recent hegra spectra of mrk 501 @xcite do show some curvature , the lower limits on the infrared background imply that the intrinsic spectrum must be rather flat or even up - turning @xcite .
one might as well argue that it is much more in the sense of ockham s razor ( i.e. more economic ) , if one finds that the same sources which very likely produce most of the extragalactic gamma rays would at the same time produce cosmic rays at ultrahigh energies where it is difficult to find any other astrophysical source supplying enough power .
there are two independent observations and only one model . according to landau
, a third independent fact is needed to consider a theoretical claim seriously .
here i consider the high - energy neutrino emission intimately connected with multi - tev gamma rays from hadronic accelerators as the missing piece of information . to first elucidate the connection between extragalactic high - energy emissions in the framework of global energetics
, sect.2 quantifies the non - thermal energy that may be released by active galactic nuclei and their jets integrated over their cosmic history .
this qualifies extragalactic jets as possible sources of the ultrahigh energy cosmic rays and implies multi - tev emission from the jets due to proton energy losses at their acceleration site .
section 3 discusses the predicted spectra from a simple quasi - stationary unsaturated synchrotron cascade emission model .
bearing on the assumption of a strong evolution of their luminosity density , the neutrino and cosmic ray spectra from extragalactic jets using the assumptions of the hadronic gamma ray emission model are computed in sect.4 .
since the expected peak of the neutrino spectrum lies at energies in excess of 100 tev at which the earth becomes optically thick with respect to neutrino absorption , sect.5 extends the discussion by including the profound effects of neutrino oscillations .
inspection of fig . [ fig1 ] shows an interesting pattern in the present - day energy density of the diffuse isotropic background radiation consisting of a sequence of bumps each with a strength that is decreasing with photon energy .
the microwave bump is recognized as the signature of the big bang at the time of decoupling with its energy density given by the stefan - boltzmann law @xmath2 .
the bump in the far - infrared is due to star formation in early galaxies , since part of the stellar light , which appears as the bump at visible wavelengths , is reprocessed by dust obscuring the star - forming regions .
the energy density of the two bumps can be related to the present - day heavy element abundances .
heavy elements with present - day mass fraction @xmath3 were produced in early bursts of star formation by nucleosynthesis with radiative efficiency @xmath4 yielding the present - day energy density @xmath5 where @xmath6 denotes the mass density of baryonic matter and @xmath7 the formation redshift corresponding to the era of maximum star formation .
this is , of course , only a very rough approximation of the true star formation history , but good enough to set the scale for an argument pertaining to the global energetics . in particular , the ratio between the energy released by stars and by other sources with the same formation history is independent of its details .
let @xmath8 denote the baryon density in terms of the critical density of the universe and @xmath9 km s@xmath10 mpc@xmath10 the dimensionless hubble constant , then the energy density takes the value @xmath11 and should represent the energy density of the sum of the far - infrared and optical bumps .
probably all galaxies ( except dwarfs ) contain supermassive black holes in their centers which are actively accreting over a fraction of @xmath12 of their lifetime implying that the electromagnetic radiation released by the accreting black holes amounts to @xmath13 adopting the accretion efficiency @xmath14 and the black hole mass fraction @xmath15 @xcite .
most of the accretion power emerges in the ultraviolet where the diffuse background is unobservable owing to photoelectric absorption by the neutral component of the interstellar medium .
however , a fraction of @xmath16 taken from the average quasar spectral energy distribution @xcite shows up in hard x - rays producing the diffuse isotropic x - ray background bump with @xmath17 @xcite .
jets with non - thermal @xmath18-ray emission show up only in the radio - loud fraction @xmath19 of all agn and their kinetic power roughly equals the accretion power @xcite .
hence one obtains for the energy density due to extragalactic jets @xmath20 this energy is released in relativistic particles , magnetic fields , and pdv thermodynamic work against the ambient medium into which the jets propagate . adopting a radiative efficiency of @xmath21 for the jets , the gamma ray ( cosmic ray , neutrino ) energy released by the jets can amount to a maximum present - day energy density of @xmath22 which comes remarkably close to the energy density @xmath23 ev @xmath24 of the extragalactic gamma ray background observed between 100 mev and 30 gev gev for extragalactic source populations @xcite ] using the spectrum given in ref .
note that this is consistent with having a particle acceleration efficiency in radio jets which is of the same order of magnitude as the 13% efficiency required for supernova remnants to produce the galactic cosmic rays .
protons can achieve a high radiative efficiency due to photo - production of secondary particles or synchrotron emission only if they reach ultrahigh energies .
this can be seen most easily for synchrotron radiation in which case the proton lorentz factor must be larger than the electron lorentz factor by @xmath25 to produce the same cooling rate .
particles with energies exceeding @xmath26 ev may seem outrageously exotic to many astrophysicists , but they are actually observed in the local cosmic ray spectrum . these particles can not be confined by the milky way , but nevertheless show a near - isotropic sky distribution .
furthermore , the local cosmic ray spectrum flattens above @xmath26 ev indicating a separate extragalactic source population
. it will be argued in sect.4 that only an extragalactic source population as strong as the one supplying the diffuse isotropic gamma rays can be considered for their origin . before highlighting these arguments
, it is shown in the next section that protons at such high energies produce interesting gamma ray spectra owing to the photo - production of secondaries .
a large number of emission models for radio jets based on a shock - in - jet scenario exist and have been shown to explain most of the low - frequency jet phenomenology using a number of parameters such as jet kinetic energy , magnetic field gradient , opening angle of the flow channel , etc . within plausible ranges .
it is straightforward to include relativistic protons in such models assuming they have a spectrum with the same slope but different maximum and minimum energies ( due to different energy losses , gyro - resonant thresholds , and larmor radii ) . among the interesting consequences of the protons
are ( i ) a higher nonthermal pressure and ( ii ) more gamma ray flux ( adding to the compton flux from the accelerated electrons ) .
the equipartition magnetic field strength increases according to @xmath27 where @xmath28 denotes the ratio between the energy density in relativistic protons and electrons , respectively , and this may help to alleviate some problems for pair jets such as the observed pressure support of the radio lobes in ngc 1275 against the surrounding hot intracluster medium of the perseus cluster @xcite . in terms of simplicity and predictive power , the original blandford & knigl conical jet model @xcite is useful and has been investigated for the radiative signatures of the protons @xcite .
the model describes the stationary emission from a conical section of a free relativistic jet and is therefore of limited applicability to non - stationary features in the observed spectra .
the relativistic particle and magnetic energy density decreases as @xmath29 along the jet , i.e. the jet is isothermal .
the emitting conical section of the jet may be thought of as the unresolved superposition of a number of shocks traveling down the jet , and shock acceleration keeps the nonthermal energy constant .
such a jet emits a flat radio spectrum up to the frequency where the spectrum steepens by one power due to the energy losses , typically in the submm - infrared regime .
the flux at this break frequency is dominated by the jet cone near its apex .
the threshold frequency @xmath30 for the production of pions in head - on collisions between photons and protons with lorentz factor @xmath31 is given by @xmath32 showing that the region of maximum surface brightness temperature is most important for the cooling of protons .
owing to electromagnetic cascading the electromagnetic power injected into the jet plasma by the cooling baryons is redistributed smoothly over the x - ray and gamma ray bands .
therefore , the emission component with the largest energy flux dominates the entire high - frequency spectrum and proton - initiated cascade spectra have been computed only for this zone .
an integration along the jet axis ( which is necessary to obtain the flat radio spectrum ) would only lead to marginal corrections in the gamma ray regime .
the geometry is depicted in fig .
[ fig2 ] together with a sketch of the emission components from the various scales . due to the small size of the region dominating the gamma ray energy flux compared with the total volume of the conical jet , traveling shocks would make the gamma ray emission much more susceptible to flux variability than the radio - infrared emission in qualitative agreement with the observations .
as long as the perturbation time scale is larger than the proton acceleration time scale , the emission spectra can still be computed as quasi - stationary spectra , and it is an open theoretical challenge to solve for the spectra in the general time - dependent case .
an interesting time - dependent solution exists for a simplified version of compton - scattering dominated proton - initiated cascades @xcite . using the doppler factor @xmath33 to convert between comoving and observer s frame , the proton acceleration time scale assuming bohm diffusion is given by @xmath34 in comfortable agreement with some of the observed tev flux variation time scales .
the comoving frame gyro - radius of the protons @xmath35 satisfies the constraint @xmath36 if the jet radius is the lightcylinder radius of an mhd jet from a black hole of mass @xmath37 and schwarzschild radius @xmath38 in which case @xmath39 cm . for time scales
longer than the proton acceleration time scale , correlated flux variations from hard x - rays to tev gamma rays are expected as a typical phenomenon ( within the model assumptions ) .
if the hard x - ray emission or the gamma ray emission in the egret band is dominated by emission from the accelerated electrons , the tev variability can be different .
this case would argue for a higher ratio of the photon to magnetic energy density @xmath40 than in the generally assumed case @xmath41 in the proton blazar model .
note also that variability on shorter time scales must be expected under realistic conditions , e.g. due to the presence of inhomogeneities along the jet .
these variations can occur on time scales down to the cooling time scale of the cascade electrons , and require the solution of the time - dependent cascade equations .
the short - term variability behavior of the cascades is very complicated owing to the pair production threshold varying with gamma ray energy .
the tev detections of nearby blazars confirm the prediction of the proton blazar model that the relativistic jets should be optically thin to gamma ray emission below @xmath42 tev .
this means that the conversion of injected gamma rays into pairs and vice versa becomes unimportant below tev ( unsaturated cascades ) .
emission above @xmath42 tev is expected to be optically thick , but nevertheless of approximate power law shape .
synchrotron cooling is assumed to be the dominant process replenishing the gamma rays from the pairs .
for the entire parameter space , synchrotron emission remains non - relativistic so that the characteristic synchrotron photon energy is much less than that of the radiating electrons assuring rapid convergence of the cascade equations ( which can be brought into the form of a volterra integral equation of the second kind ) by banach s fixed point theorem .
the shape of the multi - tev spectrum follows from very simple considerations .
it is assumed that the protons have a differential distribution @xmath43 .
the electron distribution has the same slope in the optically thin range , but in the energy range responsible for producing the target photons for the protons their spectrum is steeper by one power @xmath44 owing to energy losses .
the electrons thus produce a synchrotron flux density spectrum @xmath45 with spectral index @xmath46 . in the original papers @xcite it was assumed that @xmath47 and correspondingly @xmath48 which is the non - relativistic result from 1st order fermi acceleration at strong shocks in the test - particle approximation
. a flatter value @xmath49 corresponding to @xmath50 may be more appropriate for strong shocks in the general case @xcite .
the cascades are initiated by gamma rays from the decay of the neutral pions at ultrahigh energies .
the slope of the differential injection spectrum of the gamma rays is given by @xmath51 , but steepens due to pair creation on the synchrotron target characterized by the optical depth @xmath52 .
the steepening can be described by the energy - dependent escape probability @xmath53/ \tau(e ) } \rightarrow \tau(e)^{-1 } \propto e^{-\alpha}$ ] for @xmath54 , i.e. a steepening of the injection spectrum by @xmath51 .
hence , the stationary injection spectrum has the slope @xmath55 .
the next step in the cascade development involves the creation of pairs which have a stationary distribution with the same slope @xmath56 producing the second generation of gamma rays .
since generally these gamma rays still lie in the optically thick energy range , the stationary gamma ray distribution has the slope @xmath57 .
the same is true for the new generation of pairs produced by these gamma rays , and their synchrotron gamma rays are mostly emitted at optically thin energies where their spectrum has the slope @xmath58 ( in the optically thick range the slope is @xmath59 ) .
thus the predicted multi - tev slope is bracketed by @xmath60 and @xmath59 .
for @xmath48 , the corresponding range is @xmath61 and for @xmath62 it is @xmath63 which is in reasonable agreement with the observations of mrk 421 and mrk 501 @xcite corrected for the expected intergalactic gamma ray attenuation @xcite .
the model spectrum fitted to lower frequency data as published prior to the 1997 hegra multi - tev observations agrees remarkably well with the measurements @xcite .
the shape of the multi - tev spectrum is not sensitive to changes in the maximum energy and can remain constant under large - amplitude changes of the flux associated with changes in the maximum energy .
the photo - production of charged pions leads to the emission of neutrons and neutrinos .
neutrons associated with the production of @xmath64 have no efficient coupling with the magnetized plasma in the jet and therefore escape ballistically .
the neutrons decay to protons after a propagation length @xmath65 mpc , and such extragalactic cosmic rays suffer energy losses traversing the microwave background @xcite . at an observed energy of @xmath66 ev
, the energy - loss distance is @xmath67 owing to pair production .
this distance corresponds to a redshift @xmath68 determined by @xmath69 } $ ] where @xmath70^{1\over 2 } $ ] with @xmath71 .
almost independent on cosmology , the resulting value for @xmath68 is given by @xmath72 where @xmath73 km s@xmath10 mpc@xmath10 .
therefore , when computing the contribution of extragalactic sources to the observed cosmic ray flux above @xmath66 ev , only sources with @xmath74 must be considered .
assuming further that extragalactic sources of cosmic rays and neutrinos are homogeneously distributed with a monochromatic luminosity density @xmath75 where @xmath76 for agn @xcite , their contribution to the energy density of a present - day diffuse isotropic background is given by @xmath77 where @xmath78 denotes the redshift of maximum luminosity density .
the factor @xmath79 accounts for the expansion of space and the redshift of energy . for a simple analytical estimate of the effect of energy losses on the proton energy density at @xmath66 ev
, we collect only protons from sources out to the horizon redshift @xmath80 for @xmath66 ev protons , whereas neutrinos are collected from sources out to the redshift of their maximum luminosity density @xmath81 .
this yields the energy density ratio for neutrinos at an observed energy of @xmath82 ev and protons at @xmath66 ev @xmath83 using @xmath84 from decay and interaction kinematics , and considering an open universe with @xmath85 and a closed one with @xmath86 .
[ fig3 ] shows the exact propagated proton and neutrino spectra for @xmath87 from a full monte - carlo simulation employing the matrix doubling method of protheroe & johnson @xcite .
the assumed neutron spectrum was @xmath88 ( corresponding to @xmath48 in the previous section ) up to @xmath89 ev and @xmath90 above .
the steepening reflects the fact that the maximum energy may vary from source to source .
the muon neutrino spectra follow the same shape , but they are shifted according to a simplified treatment of pion decay and production kinematics . a more accurate treatment yields small corrections @xcite .
the neutron spectrum was normalized to the cosmic ray data yielding a neutrino spectrum consistent with model a from the original work @xcite . the associated gamma ray flux @xmath91 \approx 2\times 10^{-6}$ ] gev @xmath92 s@xmath10 sr@xmath10 corresponds to a sizable fraction of the observed gamma ray background flux .
the neutrino flux is consistent with the bound given in ref .
@xcite , although it is possible to have extragalactic neutrino sources of higher neutrino fluxes without violating the observed cosmic ray data as a bound @xcite .
one could easily construct such models which produce the entire gamma ray background on the same rationale @xcite .
the neutrino flux shown in fig .
[ fig3 ] corresponds to a very low muon event rate even in a detector with an effective area of 1 km@xmath93 ( @xmath42 event per year and per steradian above 100 tev ) .
this event rate is a very conservative estimate , since there must be additional neutrino production due to pp - interactions of escaping nucleons diffusing through the host galaxies and galaxy clusters .
the neutrino flux could also be increased by increasing the number of extragalactic jet sources with proton maximum energies well below @xmath66 ev @xcite . as a matter of fact
, such a model ramification is required if one wants to explain the entire diffuse gamma ray background by hadronic photo - production sources . at this point
the discovery of neutrino mass announced by the super - kamiokande collaboration @xcite changes the situation in a major way .
a deficit of atmospheric muon neutrinos was observed with super - kamiokande at large zenith angles with the most likely explanation being a full - amplitude oscillation of muon flavor eigenstates to tauon flavor eigenstates across the earth at gev energies .
the transition probability @xmath94 is a function of distance and energy @xmath95 , i.e. @xmath96 where the mixing angle @xmath97 specifies the mixing amplitude @xmath98 .
the missing piece of information about the neutrino oscillations is the appearance of tau leptons to which super - kamiokande is not sensitive .
a long - baseline experiment using muon neutrinos from a laboratory beam is extremely difficult if not impossible with existing laboratories , since one must establish a very large distance and a high energy to reach the tau mass shell at @xmath99 gev .
the beam luminosity decreases rapidly with energy and distance which poses an irreducible problem .
however , an astrophysical beam of muon neutrinos , such as the one proposed in this paper and for which the multi - tev observations give us somewhat more confidence that they really exist , is ideally suited for this type of experiment , since both l and e obtain `` astronomically large '' values .
hence it follows that very likely astronomical high - energy neutrino sources such as extragalactic radio jets will help to solve a major puzzle in elementary particle physics .
there is another effect associated with neutrino oscillations which alleviates the problem with the low muon event rate of the predicted flux . to see this
, one must realize that the solid angle for the detection of high - energy muon neutrinos becomes very narrow at energies in excess of @xmath100 tev , since the earth becomes optically thick to muon neutrinos above this energy ( the weak interaction cross section depends on energy ) .
this is called the earth shadowing effect @xcite . below 100
tev the atmospheric background of neutrinos is too strong for the discovery of the rare events due to extragalactic neutrinos , unless the angular resolution of the neutrino telescopes is very good . actually , the atmospheric background is comparable to the predicted astronomical background at around 100 tev .
therefore one is confronted with the problem of a too small solid angle for events above 100 tev and a too low rate below 100 tev .
if the neutrino oscillation hypothesis is correct , the extragalactic muon neutrino beam is fully mixed with tau neutrinos at earth .
tau leptons produced by charged - current interactions inside the earth decay before further interacting .
since the earth is also opaque to tau neutrinos above @xmath101 100 tev , all the tau neutrinos entering on the one side with energies above 100 tev emerge on the other side with energies of around 100 tev @xcite obliterating the earth - shadowing effect . in the proton blazar model with @xmath48 , the number density of neutrinos above 100
tev remains constant up to roughly @xmath102 tev implying an increase in the contained event rate by a factor @xmath103\simeq 3.5\ ] ] where the factor @xmath104 is for the mixing between muon and tauon neutrinos ( @xmath105 would be appropriate for further mixing with electron neutrinos ) .
the neutrinos from extragalactic radio sources can therefore be expected to produce more events than the atmosphere at around 100 tev .
some of these events are not muon events , but direct tauon events . owing to tauon decay ,
the tauon tracks are very short @xmath106 cm and the main signature is the electromagnetic cascade from tau decay .
horizontal events can be of higher energy , producing the famous double - bang events with the first bang indicating the charged - current tauon production event and the second its decay .
thus , the glass is half - full .
the multi - tev spectra from nearby blazars prediced on the basis of the proton blazar model are in accord with the observations if the effect of pair attenuation due to the extragalactic infrared background is taken into account .
variability patterns are similar to those in synchrotron - self - compton models , but more complex , since variations of the target photon flux fold into the cascade development in a non - linear manner .
short time scale variability is likely to reflect the passage of shocks through inhomogeneities and correspond to cooling time scale variations .
if the cosmic ray flux emitted by hadronic accelerators is enough to explain the observed cosmic rays above @xmath66 ev , the associated gamma ray power from these sources is enough to produce at least a sizable fraction of the observed extragalactic gamma ray background . the gamma ray power is larger than that in cosmic rays ,
since the cosmic rays lose energy traversing the low - energy background radiation fields and most sources have high redshifts .
strong evolution of their luminosity density would rule out grbs as possible sources of the highest energy cosmic rays , since their cumulative gamma ray flux is far below the extragalactic gamma ray flux .
although the expected muon event rate in neutrino telescopes is low , neutrino oscillations lead to tau cascades canceling the earth shadowing effect thereby increasing the detection probability .
p. sreekumar , f.w .
stecker , s.c .
kappadath , in : _ proceedings of the fourth compton symposium _ ,
charles d. dermer , mark s. strickman , and james d. kurfess , williamsburg , va , april 1997 : aip conference proceedings 410 ( 1998 ) p. 344 . | three independent predictions follow from postulating the existence of protons co - accelerated with electrons in extragalactic jets ( i ) multi - tev gamma ray emission from nearby blazars , ( ii ) extragalactic cosmic ray protons up to @xmath0 ev , and ( iii ) extragalactic neutrinos up to @xmath1 ev .
recent gamma ray observations of mrk 421 and mrk 501 employing the air - cerenkov technique are consistent with the predicted gamma ray spectrum , if one corrects for pair attenuation on the infrared background . prediction ( ii )
is consistent with cosmic ray data , if one requires that jets are responsible for a at least a sizable fraction of the extragalactic gamma ray background . with kubic kilometer neutrino telescopes
, it will be possible to test ( iii ) , although the muon event rates are rather low .
neutrino oscillations can increase the event rate by inducing tau - cascades removing the so - called earth shadowing effect . |
in relativistic heavy - ion collision experiments ( rhice ) the collision of two nuclei leads to a hot dense region which is expected to rapidly achieve state of thermal equilibrium .
for the relevant range of energies and colliding nuclei at rhic and at lhc there is compelling evidence that a region of quark - gluon plasma is created in these collisions .
simulation results as well as experimental data , such as elliptic flow measurements , all point towards a very rapid thermalization to the qgp phase , within a proper time less than 1 fm .
we thus have a system which starts out in the confining phase , and within proper time of ( probably much less than ) 1 fm makes a transition to the qgp phase ( with maximum temperature estimates ranging from 200 mev to more than 700 mev for the relevant energies in these experiments ) .
lattice results show that for real world qcd with small baryon density ( as appropriate for the central rapidity regions in rhice ) the transition is likely to be a crossover . with that
, the dynamics of the transition depends crucially on the rate of temperature change compared to the time scale of the evolution of the order parameter field . for an equilibrium transition
, we had studied the formation of z(3 ) domains and associated _ qgp strings _ using a first order transition dynamics via bubble nucleation @xcite .
the transition was simulated using the polyakov loop , @xmath4 , as an order parameter for the confinement - deconfinement transition @xcite .
these studies were appropriate for large chemical potential case , as in lower energy collisions , where the transition is expected to be of first order .
though , the results for z(3 ) wall network etc . having certain universal characters , may be applicable in more general context , even for a cross - over , as explained in @xcite .
however , given the very short time scale of initial thermalization to qgp state , an equilibrium dynamics of the transition appears unlikely .
a more appropriate description of the transition should employ quenched dynamics in which the growth of z(3 ) domains will be via spinodal decomposition . in this paper
we carry out such a study using the polyakov loop , @xmath4 , as an an order parameter for the confinement - deconfinement transition , with an effective potential of the kind used in refs.@xcite . for our simulation results
we choose a definite parametrization of the effective potential @xcite as was used in our previous works @xcite .
we model the phase transition in this polyakov loop model , with the initial field configuration ( in the confining phase ) covering a small neighborhood of the confining vacuum @xmath0 as appropriate for the initial @xmath5 = 0 system . in a quench
, the temperature rapidly ( rather suddenly ) increases to its maximum value @xmath6 with the effective potential changed accordingly . the initial field @xmath3 ,
unable to relax to the new equilibrium vacuum state in this short time , becomes unstable and rolls down in different directions from the top of the central hill in the effective potential of @xmath3 .
we study the formation of z(3 ) domain structure during this evolution . when explicit z(3 ) symmetry breaking effects ( arising from dynamical quark effects ) are small , then we find well defined z(3 ) domains which coarsen in time . with a symmetric initial patch of @xmath3 , all the three z(3 )
domains form with random shapes and rapidly increase in size by coarsening .
remarkably , the magnitude plot of @xmath3 shows vacuum bubble like configurations , such as those which arise in a first order transition , arising during the quench in this case ( when the initial field rolls down in different directions ) .
this first order transition like behavior occurs even though there is no metastable vacuum separated by a barrier from the true vacuum for the parameter values used .
these bubble like configurations expand as well , somewhat in similar manner as during a first order transition .
when the initial patch of @xmath3 is only partially symmetric around @xmath7 ( as appropriate for small explicit symmetry breaking from quark effects ) , the dynamics retains these qualitative aspects with expected changes .
thus , true vacuum domains ( with @xmath8 ) are more abundant and they also grow fast at the cost of the other two z(3 ) domains ( which are now metastable due to explicit symmetry breaking ) . still , for small explicit symmetry breaking , all the three types of domains occur during the initial stages of the transition .
there are few , small z(3 ) walls as there are fewer metastable z(3 ) domains embedded in the dominant true vacuum with @xmath8 .
these walls shrink rapidly and eventually only the true vacuum survives .
the dynamics is found to be very different when the explicit symmetry breaking due to quark effects is taken to be strong . in this case
the initial patch of @xmath3 ( around equilibrium point for @xmath9 effective potential ) could be significantly shifted towards the true vacuum for the quenched @xmath10 effective potential .
in such a situation , @xmath3 will roll down roughly along the same direction with angular variations becoming smaller during the roll down . in this case only @xmath8 vacuum survives and no other z(3 ) domains are formed .
also , in this case we do not find bubble - like configurations .
however , in this case we find huge oscillations of @xmath3 with large length scales .
this behavior is known from previous studies of the dynamics of scalar field in a quench @xcite and is expected here when angular variations are small .
the dynamics of field in such a case is dominated by large length scale coherent oscillations .
it leads to novel scenarios of reheating via parametric resonance in the case of inflation in the early universe @xcite . in our case of rhice
also it raises important questions about the possibility of parametric resonance and of novel modes of particle production from these large oscillations of @xmath3 during the early stages of the transition . in the present work
we explore another important effect of these large oscillations , on the evolution of flow anisotropies in rhice .
as most of the flow anisotropies are expected to develop during first few fm of the qgp formation , it becomes an important question if the presence of large oscillations of @xmath3 can affect the development of these flow anisotropies .
as we will see , this indeed happens and these large oscillations lead to large fluctuations in flow anisotropy .
we mention here that we do not discuss which of the cases discussed here ( small , or large explicit symmetry breaking ) may actually be realized in rhice .
this is primarily because of lack of understanding of the magnitude of explicit symmetry breaking term near @xmath11 .
( some discussion of this has been provided in our earlier works @xcite . )
also , the spread of initial field configuration about confining vacuum plays a crucial role here and that in turn depends on details of pre - equilibration stage .
a proper understanding ( or modeling ) of this stage and resulting estimate of the initial spread is essential before one can make more definitive statements about the dynamics of the transition .
we hope to discuss these issues in future .
our main purpose here is to illustrate the possibility of very different types of dynamics of transition depending on the initial configuration .
our results show that in quenched dynamics , z(3 ) domains , and resulting z(3 ) domain walls , can last for any reasonably length of time only when explicit symmetry breaking terms are very small .
otherwise either different domains do not form at all , hence no z(3 ) walls are formed , or fewer metastable z(3 ) domains form embedded in most abundant true vacuum region . in the latter case resulting z(3 ) walls are smaller to begin with , and disappear quickly .
importance of z(3 ) walls in rhice has been discussed by us in previous works , where we have also emphasized non - trivial scattering of quarks from z(3 ) walls .
we have explored its consequences for cosmology as well as for rhice @xcite , including the possibility of cp violating scattering of quarks from z(3 ) walls leading to interesting observational implications @xcite .
recently an interesting possibility has been discussed by asakawa et al . in ref .
@xcite where it is argued that scattering of partons from z(3 ) walls may account for small viscosity as well as large opacity of qgp . our results for the formation of z(3 ) domains during quench ( which are abundant only for small explicit symmetry breaking case ) can be relevant for the studies in ref .
@xcite . in this context
we mention that smallest reasonable size z(3 ) domains ( hence z(3 ) walls ) we find are of order 1 - 2 fm at very early times , and at that stage the magnitude of the polyakov loop order parameter @xmath3 is very small , of order few percent of its vacuum expectation value .
the quark scattering from z(3 ) walls is likely to be small for such a small magnitude of @xmath3 @xcite . by the time the magnitude of @xmath3 becomes significant , domains coarsen to have large sizes , of order several fm .
thus , in the context of our model it appears difficult to form very small z(3 ) domains which still can scatter partons effectively ( as needed in the study of ref.@xcite ) .
it is important to note that our results for domain growth and fluctuations for the quench case are dominated by the spinodal instabilities during the roll down of the field from the top of the hill of the effective potential .
these instabilities arise primarily from the nature of the quench when the initial field configuration becomes unstable to exponential growth of long wavelength modes due to sudden change in the shape of the effective potential ( the quench ) , and will be in general present even if the transition is a cross - over . in this sense
we believe that the qualitative aspects of our results have a wider applicability and are not crucially dependent on the specific form of the effective potential used here .
we briefly mention here that issues related to the physical meaning of z(3 ) domains etc .
have been discussed in the literature ( see , discussion of these in our earlier works @xcite and we will not repeat it here .
we only note that recent work of deka et al .
@xcite has provided a support for the existence of these metastable vacua from lattice .
we also note that a simulation of spinodal decomposition in polyakov loop model has been carried out in ref .
@xcite , where fluctuations in the polyakov loop and growth of long wavelength modes ( representing domain formation ) are investigated . in comparison ,
the main focus of our work is on detailed growth of domains due to coarsening with and without explicit symmetry breaking , new bubble like structures , and existence of large fluctuations affecting flow anisotropies in important ways .
the paper is organized in the following manner . in section
ii , we discuss the essential aspects of the polyakov loop model of confinement - deconfinement phase transition and the effective potential used from ref.@xcite . section iii presents the numerical technique of simulating the phase transition via quench .
we first discuss the case without any explicit symmetry breaking and study the formation and evolution of z(3 ) domains during the quench .
iv discusses the unexpected result of bubble like structures arising during the quench .
we also discuss the case of small explicit symmetry breaking term when true vacuum domains become more abundant , but overall picture of the transition remains roughly similar . in sec.v
we discuss the case when explicit symmetry breaking effects are strong leading to @xmath3 rolling down everywhere roughly along @xmath8 .
this leads to large oscillations of @xmath3 . in sec .
vi we study the effects of these large oscillations on flow anisotropy and show that it leads to large fluctuations in elliptic flow and spatial eccentricity ( as compared to the case of equilibrium transition ) .
section viii presents conclusions .
we briefly recall the salient features of the model we use for the confinement - deconfinement phase transition .
the order parameter is taken to be the the expectation value of polyakov loop @xmath4 , @xmath12 where @xmath13 is the time component of the vector potential @xmath14 , @xmath15 are the generators of @xmath16 in the fundamental representation , @xmath17 denotes path ordering in the euclidean time @xmath18 , g is the gauge coupling , and @xmath19 with t being the temperature .
@xmath20 (= 3 for qcd ) is the number of colors .
the complex scalar field @xmath21 transform under the global @xmath22 ( center ) symmetry transformation as @xmath23 the expectation value of @xmath4 is related to @xmath24 where @xmath25 is the free energy of an infinitely heavy test quark . for temperatures below @xmath26 , in the confined phase , the expectation value of polyakov loop is zero corresponding to the infinite free energy of an isolated test quark .
( hereafter , we will use the same notation @xmath4 to denote the expectation value of the polyakov loop . )
hence the z(n ) symmetry is restored below @xmath26 .
z(n ) symmetry is broken spontaneously above @xmath26 where @xmath4 is non - zero corresponding to the finite free energy of the test quark .
effective theory of the polyakov loop has been proposed by several authors with various parameters fitted to reproduce lattice results for pure qcd @xcite .
we use the polyakov loop effective theory proposed by pisarski @xcite .
the effective lagrangian density can be written as @xmath27 where the effective potential @xmath28 for the polyakov loop , in case of pure gauge theory is given as @xmath29 at low temperature where @xmath30 , the potential has only one minimum .
as temperature becomes higher than @xmath26 the polyakov loop develops a non vanishing vacuum expectation value @xmath31 , and the @xmath32 term , coming from the @xmath33 term above leads to @xmath34 generate vacua .
now in the deconfined phase , for a small range of temperature above @xmath26 , the @xmath30 extremum becomes the local minimum ( false vacuum ) and a potential barrier exist between the local minimum and global minimum ( true vacuum ) of the potential .
( however , the quench temperature used here is much higher than this range and there is no barrier present at such temperatures . )
the effects of dynamical quarks is included in terms of explicit breaking of the @xmath34 symmetry which is represented in the effective potential by inclusion of a linear term in @xmath3 @xcite .
the potential of eq.(4 ) with the linear term becomes , @xmath35 here coefficient @xmath36 measures the strength of explicit symmetry breaking .
the coefficients @xmath36 , @xmath37 , @xmath38 and @xmath39 are dimensionless quantities . with @xmath40 ,
other parameters @xmath37 , @xmath38 and @xmath39 are fitted in ref.@xcite such that that the effective potential reproduces the thermodynamics of pure @xmath41 gauge theory on lattice @xcite .
the values of various coefficients from refs .
@xcite are the same as used in our previous works @xcite ( including discussions about explicit symmetry breaking strength @xmath36 ) and we do not repeat that discussion here .
( with those values of parameters , the transition temperature is taken to be @xmath42 182 mev . )
in this work , we carry out a @xmath43 dimensional field theoretic simulation of the dynamics of confinement - deconfinement transition in a quench .
first we work within the framework of bjorken s boost invariant longitudinal expansion model @xcite ( without any transverse expansion ) for the central rapidity region in rhice . to model the quench , we take the initial field configuration to constitute a small patch around @xmath7 which corresponds to the vacuum configuration for the confining phase .
this is for the case of zero explicit symmetry breaking .
we will discuss the case of explicit symmetry breaking later .
we have taken the initial phase of @xmath3 to vary randomly between 0 and @xmath44 from one lattice site to the other , while the magnitude of @xmath3 is taken to vary uniformly between 0 and @xmath45 .
( we have also taken the initial magnitude to have a fixed value equal to @xmath45 and the results are similar . )
value of @xmath45 is taken to be much smaller than the vacuum expectation value ( vev ) of @xmath3 at @xmath10 and results remain qualitatively the same for similar small values of @xmath45 .
we report results for @xmath46 times the vev of @xmath3 .
we take the quench temperature @xmath47 mev . this initial field configuration , which represents the equilibrium field configuration of a system with @xmath1 ,
is evolved using the effective potential with @xmath48 .
this represents the transition dynamics of a quench .
the field configuration is evolved by time dependent equation of motion in the minkowski space @xcite @xmath49 with @xmath50 at @xmath51 and @xmath52 .
the temperature is taken to decrease starting with the value @xmath6 ( at an initial time @xmath53 which we take to be 1 fm in all the simulations ) as @xmath54 as appropriate for the longitudinal expansion model .
we mention that later in sect.v we will consider an isotropic geometry for the transverse dynamics of qgp as relevant for rhice . there
we will also model non - zero transverse expansion and then the central temperature will decrease faster than @xmath54 . here , we take a @xmath55 square lattice with physical size 20 fm x 20 fm .
we take this lattice as representing the transverse plane of the qgp formed in a central collision and consider mid rapidity region . the evolution of field was numerically implemented by a stabilized leapfrog algorithm of second order accuracy both in space and in time with the second order derivatives of @xmath56 approximated by a diamond shaped grid @xcite .
we evolve the field using the periodic , fixed , and free boundary conditions for the square lattice . here
we present the results with periodic boundary conditions .
we use @xmath57 fm for the present case , later on we use different values of @xmath58 , as shown by the lattice size in the corresponding figures .
we take @xmath59 as well as @xmath60 to satisfy the courant stability criteria .
the stability and accuracy of the simulation is checked using the conservation of energy during simulation .
the total energy fluctuations remains few percent without any net increase or decrease of total energy in the absence of dissipative @xmath61 term in eq.(6 ) .
first , we present results for the formation and evolution of z(3 ) domains .
the initial field configuration in the neighborhood of @xmath7 becomes unstable when evolved with the effective potential ( eq.(5 ) ) with @xmath62 mev ( at @xmath63 1 fm ) . as @xmath3 rolls down in different direction , settling in one of the three z(3 ) vacua , different z(3 ) domains form .
initially , as the phase of @xmath3 is taken to vary randomly from one lattice site to the next , there are no well defined domains . also , the magnitude of @xmath3 is very small initially making any association of z(3 ) structure meaningless at such early stages .
situation remains similar for very early times as seen in fig.1a at an early stage @xmath64 fm ( i.e. 0.2 fm after the quench ) . in fig.1
we have shown the values of the phase of @xmath3 around the three z(3 ) vacua in terms of different shades ( colors ) to focus on the evolution of z(3 ) domain structure .
thus all the values of the phase @xmath65 of @xmath3 are separated in three ranges , between @xmath66 to @xmath67 ( @xmath8 vacuum ) , between @xmath67 to @xmath68 ( @xmath69 vacuum ) , and between @xmath68 to @xmath70 ( @xmath71 vacuum ) .
as the field magnitude grows , the angular variation of @xmath3 also becomes less random over small length scales , leading to a sort of z(3 ) domain structure .
z(3 ) domains become more well defined , and grow in size by coarsening as shown in sequence of figures fig.1b - d .
different shades ( colors ) in fig.1 represent the three z(3 ) domains .
fig.1b - d shows the growth of these domains at @xmath18 = 2.0 , 2.4 , and 2.8 fm , as @xmath3 relaxes to the three z(3 ) vacua and domains grow in size by coarsening .
the magnitude of @xmath3 is about 0.04 , 0.08 , 0.2 , and 0.4 for figs.1a , b , c , and d respectively .
note that domains grow rapidly to size of order 2 fm within a time duration of about 1 fm as shown in fig.1b . within another 1 fm time
, domain size is about 4 fm as seen in fig.1d .
the boundaries of different z(3 ) domains represent z(3 ) walls , and junction of three different z(3 ) domains gives rise to the qgp strings .
these objects have been discussed in detail in our earlier works @xcite .
= 8truecm \(a ) evolution of z(3 ) domain structure during a quench .
note we take the initial temperature to be @xmath6 (= 400 mev ) at proper time @xmath72 fm to set the initial conditions for the simulation .
different shades ( colors ) represent the three z(3 ) domains .
( a ) - ( d ) show the growth of these domains at @xmath18 = 1.2 , 2.0 , 2.4 , and 2.8 fm , ( with corresponding values of temperature @xmath5 = 376 , 317 , 298 , 283 mev ) as @xmath3 relaxes to the three z(3 ) vacua and domains grow in size by coarsening .
the magnitude of @xmath3 is about 0.04 , 0.08 , 0.2 , and 0.4 for ( a ) , ( b ) , ( c ) , and ( d ) respectively .
note that domains grow rapidly to size of order 2 fm within a time duration of about 1 fm as shown in ( b ) . within another 1 fm time ,
domain size is about 4 fm ( as seen in ( d ) ) .
[ fig.1 ] we mention here that smallest reasonable size z(3 ) domains ( hence z(3 ) walls ) , which we find in our simulation , are of order 1 - 2 fm at very early times , such as seen in fig.1b . at this stage ,
the magnitude of the polyakov loop order parameter @xmath3 is still very small , of order few percent of its vacuum expectation value .
this is important when one considers the possibility of nontrivial scattering of partons from z(3 ) walls @xcite .
the quark scattering from z(3 ) walls is likely to be small for such a small magnitude of @xmath3 @xcite . by the time the magnitude of @xmath3 becomes significant , domains coarsen to have large sizes , of order several
fm , as in fig.c , d where the magnitude of @xmath3 is about 20% and 40% , respectively , of its vacuum expectation value .
thus , in the context of our model , with quenched dynamics of transition , it appears difficult to form very small z(3 ) domains which still can scatter partons effectively ( as needed in the study of ref.@xcite ) .
in this section we show a very unexpected result .
fig.2 shows sequence of plots of the magnitude of @xmath3 during the quench described in the previous section .
we note the appearance of bubble like structures .
these structures are also seen to grow in a manner similar to the bubbles for a conventional first order phase transition , as in ref.@xcite .
however , the distribution of the phase of @xmath3 does not show any specific local variation related to these bubble - like configurations . in a roughly uniform region of the phase these localized bubble - like
configurations arise and expand .
we show here the plots of @xmath3 in fig.2 corresponding to the initial central temperature @xmath73 500 mev ( at @xmath63 1 fm ) .
the plots in fig.2a , b , c , d are for @xmath74 2.0 , 2.8 , 3.0 , and 3.3 fm respectively , with the corresponding values of the central temperature being @xmath75 395 , 355 , 345 , 336 mev .
it is important to note that at such high temperatures there is no metastable confining vacuum at @xmath7 in the effective potential @xcite .
( metastable confining vacuum exists from @xmath76 mev upto @xmath77 mev . )
there is no tunneling modeled , ( or thermal hopping over the barrier ) here , nor is expected .
one expects a simple roll down of the field representing the dynamics of spinodal decomposition during the quench .
we mention that similar bubble - like configurations also arise with @xmath47 mev .
however , in that case the temperature range goes below @xmath78 mev . to make a clear case that these bubble like configuration have nothing to do with a first order transition like situation ( even remotely )
, we have shown plots with @xmath79 500 mev .
= 8truecm these plots correspond to initial central temperature @xmath79 500 mev .
( a)-(d ) show the surface plots of @xmath80 during early stages with subsequent formation and growth of bubble like structures just like a first order phase transition
. plots in ( a ) -(d ) are at @xmath81 2 , 2.78 , 3 and 3.3 fm respectively , with corresponding temperature being @xmath75 395 , 355 , 345 , 336 mev .
these bubble - like configurations are surprising as there is no barrier here , and no metastable vacuum in the effective potential for this temperature range .
[ fig.2 ] this result is very unexpected and points to new interesting possibilities for the phase transition dynamics .
for example , such bubble like structures may lead to a dynamics of phase separation in the case of the universe similar to the original witten s scenario @xcite , even when there is no underlying first order transition .
this may also have important implications for rhice .
more importantly this new possibility of transition dynamics needs to be understood and analyzed in detail , see @xcite for a study of these issues .
the studies of this section and the previous section apply to the case without any explicit symmetry breaking , i.e. with @xmath40 in eq.(5 ) .
we have repeated these simulations with small explicit symmetry breaking effects on the initial conditions .
we take @xmath82 as in ref.@xcite here as well as in the next section . by small explicit symmetry breaking effects on the initial conditions we mean that the initial patch of @xmath3 is taken to shift towards @xmath83 vacuum for @xmath10 effective potential , while still overlapping with the initial equilibrium value of @xmath3 .
this forces @xmath3 to roll down to different @xmath65 directions at least at some fraction of lattice points , though major fraction now rolls down towards @xmath8 .
fig.3 shows sequence of plots showing growth of z(3 ) domains in one such case .
we see that one of the vacua ( @xmath8 ) expands dominantly while other domains remain relatively smaller .
the z(3 ) domain walls in this case are smaller and disappear faster compared to the case without explicit symmetry breaking .
the dynamics of bubble like structure retains its qualitative aspects in this case as long as the field rolls down in different directions .
= 8truecm \(a ) field configurations at different times with small explicit symmetry breaking effects .
the shading ( color ) representing dominant region in ( d ) corresponds to the true vacuum with @xmath8 .
( a ) - ( d ) show the growth of domains for @xmath18 = 1.2 , 1.6 , 2.0 , and 2.4 fm ( with corresponding values of temperature @xmath75 376 , 342 , 317 , 298 mev ) .
now we consider the case when explicit symmetry breaking effects make the initial field configuration completely biased towards @xmath8 direction . here the initial patch of @xmath3 rolls down entirely towards @xmath8 direction with angular variations decreasing during the roll down .
the dynamics of transition is entirely different in this case .
clearly there is no possibility of different z(3 ) domains here , hence no z(3 ) interfaces , or qgp strings will form .
we also do not see any bubble like structures here as were seen in fig.2 .
instead we find the @xmath3 settles down to the true vacuum after undergoing huge oscillations , with large length and time scales .
these large oscillations are very similar to coherent oscillations of the inflaton field in the context of inflation in the early universe @xcite . for
the inflation , decay of such coherently oscillating field to particles can be via parametric resonance leading to novel features in the reheating of the universe after inflation .
existence of similar oscillations here raises possibilities of parametric resonance for rhice and similar novel dynamics of particle production during such early stages of qgp evolution . as for the universe
, new possibilities of thermalization and symmetry changes may arise here .
we hope to explore these issues in a future work .
a direct implication of the existence of such huge oscillations of @xmath3 in the context of rhice is its possible effects on the growth of flow anisotropies . in the hydrodynamical studies of elliptic flow in non - central collisions
it is known that much of the flow anisotropy develops during early stages of the plasma evolution @xcite . in such studies
one starts with equilibrium qgp phase where flow develops due to pressure gradients . in view of the possibility of large oscillations during early stages of transition to the qgp phase
, the equilibrium starting point of these hydrodynamics simulations becomes suspicious .
our present study does not allow us to address this issue in the context of hydrodynamical evolution .
however , even with the field theoretical simulation in this work , we can do a comparative study of momentum anisotropy development with and without the presence of large oscillations of the order parameter field @xmath3 . for this
we proceed as follows .
first we need to model the initial qgp system with appropriate spatial anisotropies .
one can not then use the square lattice with uniform temperature .
we use the temperature profile of woods - saxon shape with the size in the x and y directions being different representing elliptical shape for a non - central collision .
this allows us to have a well defined size for the central qgp region , with temperature smoothly decreasing at the boundary of this region .
the transverse size @xmath84 of this system ( i.e. profile of temperature ) is taken to increase with uniform acceleration of 0.015 c per fm , starting from an initial value of @xmath84 equal to the nuclear radius @xcite .
the initial transverse expansion velocity is taken to be zero .
this expanding background of temperature profile is supposed to represent the hydrodynamically expanding quark - gluon plasma in which the evolution of the order parameter field @xmath3 will be studied .
it may appear confusing as @xmath3 is expected to represent the qgp phase .
indeed , the normalization of the effective potential in eq.(5 ) from refs.@xcite is carried out precisely so that it represents energy density and pressure of gluons plus quark degrees of freedom . still , the dynamics of @xmath3 from eq.(5 ) does not carry the information of hydrodynamical degrees of freedom .
various particle modes in @xmath85 need to be excited , which should reach equilibrium , and only then we can expect some sort of hydrodynamical evolution .
clearly the initial field configurations assumed here are far from representing such a hydrodynamical state
. a consistent interpretation of our simulation can be that we are studying long wavelength modes of @xmath3 which are couple to a background of short wavelength modes which are in thermal equilibrium .
this equilibrated background is expanding with velocity as mentioned above , and drives the evolution of large wavelength modes of @xmath3 via eq(6 ) . with this interpretation ,
our task is straightforward .
the central temperature of the woods - saxon profile is taken to decrease by assuming that the total entropy ( integrated in the transverse plane ) decreases linearly as appropriate for bjorken dynamics of longitudinal expansion .
note that , with the transverse expansion being non - zero now , the central temperature will decreases faster than @xmath54 .
we show , in fig.4 , a sequence of surface plots of the magnitude of @xmath3 showing huge oscillations with large length scale during quench .
lattice here is again 2000 @xmath86 2000 but we take a large value of @xmath87 fm so that the physical lattice size is 50 fm @xmath86 50 fm . the woods - saxon temperature profile ( representing qgp region ) is taken to have a diameter of about 16 fm as appropriate for au - au collision for rhice .
large physical size of the lattice allows for the evolution of the qgp region to be free from boundary effects .
= 8truecm surface plots of the magnitude of @xmath3 with circular geometry for the qgp region .
( a ) - ( d ) show large oscillations of @xmath3 during the quench when @xmath3 rolls down everywhere towards the same vacuum with @xmath8 .
plots in ( a)-(d ) correspond to @xmath81 2.6 , 3.6 , 4.4 , 6.2 fm / c , with corresponding values of temperature being @xmath75 281 , 251 , 228 , 191 mev , respectively [ fig.4 ] at any stage we can calculate the energy momentum tensor @xmath88 of @xmath3 .
we then calculate the spatial eccentricity @xmath89 of the @xmath3 field configuration in the standard way , @xmath90 where @xmath91 is the energy density .
we can also calculate the momentum density at any time using @xmath92 and @xmath93 . using these components we know the momentum density vector at every stage . by integrating it in angular bins
we calculate its various fourier coefficients , in particular the elliptic flow coefficient @xmath94 .
to study the effects of large @xmath3 oscillations on flow anisotropy , we consider two separate cases .
note that now we are considering the cases with explicit symmetry breaking , with its effect being strong on the initial field configuration .
first we consider the quench case as described above . here
, we start with the initial field configuration corresponding to a small patch near the equilibrium point of @xmath9 effective potential . the patch is taken , as above , with field magnitude randomly varying in angle between 0 and @xmath44 with random amplitude uniformly varying from 0 to 0.01 times the vev of @xmath3 but now shifted by a constant value of 0.011 times the vev along @xmath8 direction .
this simulates the effect of strong explicit symmetry breaking as the entire patch rolls down towards @xmath8 direction .
this patch is then evolved with the temperature profile of woods - saxon shape as described above with the central temperature having initial value equal to @xmath47 mev . as a sample case
we give results for the case when the initial ( elliptical shaped ) temperature profile has an eccentricity of 0.5 ( with the major and minor axes being along the x and y axes respectively ) , which for uniform energy density will correspond to @xmath95 from eq.(7 ) .
the initial field configuration rolls down in the entire central region towards roughly @xmath8 direction leading to strong @xmath3 oscillations .
@xmath89 , and flow coefficients , e.g. @xmath94 are calculated at each stage during the evolution .
this is shown in fig.5 . in this section
we take the lattice to be 1000 @xmath86 1000 with the physical size of 25 fm @xmath86 25 fm .
this still allows for sufficient separation of the qgp region ( of size 16 fm diamater ) from the boundary .
here we show plots ( both for @xmath89 and for @xmath94 ) for two different realizations of the random initial field configuration ( shown as solid and dashed plots ) . here , and in all the figures below , we will show plots for a time upto @xmath81 10 fm / c , starting with @xmath72 fm / c at the initial stage . the central temperature decreases ( now faster than @xmath54 due to nonzero transverse expansion ) from an initial value @xmath96 400 mev at @xmath72 fm / c to the final value @xmath75 147 mev at @xmath81 10 fm / c .
note from the difference in solid and dashed plots in fig.5 that differences in the initial field configuration , which have very small magnitudes , lead to huge differences in the values of @xmath89 and @xmath94 .
= 5truecm = 5truecm these plots correspond to quench case when initial elliptical shaped temperature profile has eccentricity of 0.5 .
solid and dashed plots correspond to two different realizations of the initial random field configuration .
( a ) and ( b ) show plots for @xmath89 and elliptic flow @xmath94 respectively .
[ fig.5 ] this situation should be compared to the case of equilibrium initial conditions , as appropriate for conventional hydrodynamical simulations .
this equilibrium initial condition is implemented in the following manner here . for the initial temperature profile of woods - saxon shape ( again , with a given eccentricity ) , with initial temperature @xmath97 mev , we first need to determine the appropriate @xmath3 configuration which assumes vacuum expectation value everywhere depending on the local value of the temperature . to achieve this , we first evolve field configuration for 1000 time steps with highly dissipative dynamics while keeping the temperature profile to remain fixed at the initial profile .
a reasonably smooth initial test profile of @xmath3 is used as the initial field configuration . with highly dissipative evolution
, the field everywhere settles to the local minimum of the potential quickly .
the final configuration is found to be reasonably independent of the initial configuration assumed for @xmath3 as long as it is smooth .
this final configuration has correct profile as appropriate for the woods - saxon profile of temperature representing lowest energy configuration everywhere . in order to make a suitable comparison with the quench case
we must incorporate small fluctuations around this _ equilibrium _ configuration everywhere .
for this purpose we add to the value of the field everywhere , a small fluctuating field component with randomly varying angle between 0 and @xmath44 and with random amplitude uniformly varying from 0 to 0.01 times the vev of @xmath3 ( as for the quench case ) .
this new field configuration represents the equilibrium field configuration everywhere , with small fluctuations .
this is taken as the initial field configuration for subsequent evolution where now any extra dissipation is switched off .
this stage is taken as representing the initial time @xmath63 1 fm .
the field now evolves with the field equations , eq.(6 ) .
temperature profile also is now allowed to change in time as mentioned above .
@xmath89 , and @xmath94 etc . are calculated at each stage .
fig.6 shows these plots for this equilibrium case , starting from the time slightly after when extra dissipation ( to achieve equilibrium configuration ) has been switched off .
when random fluctuating component is introduced at every lattice site after the end of dissipative evolution , it introduces large gradients .
thus , for a very short time , there are large fluctuations as the field smoothens to certain level .
thus we show plots slightly after ( by about 0.2 fm / c ) the introduction of random field component .
= 5truecm = 5truecm plots as in fig.5 , now for the equilibrium case .
solid and dashed plots show different realizations of the small random fluctuating part of the equilibrium field configuration .
the temperature profile has same initial eccentricity of 0.5 ( with corresponding value of @xmath98 ) as in fig.5 .
( a ) and ( b ) show plots of @xmath89 and elliptic flow @xmath94 respectively .
note , the initial value of @xmath89 for the equilibrated configuration in ( a ) is very close to -0.143 as shown by the plot in the inset .
[ fig.6 ] comparison of fig.5 and fig.6 shows the dramatic effects of quench induced oscillations on flow anisotropies .
first note that for the equilibrium case the initial value of @xmath89 is close to the value -0.143 ( as shown by the inset in fig.6 ) which exactly corresponds to the initial eccentricity of 0.5 for the temperature profile .
this gives us confidence that our procedure of achieving equilibrated configuration works well .
in contrast , initial value of @xmath89 for the case of quench in fig.5 is very different showing the importance of fluctuations for this case .
we further see huge fluctuations in the values of @xmath89 and @xmath94 in fig.5 compared to the equilibrium case shown in fig.6 . note that though there are oscillations in @xmath94 for the equilibrium case also , they do not change much from one event to another .
in contrast , for the quench case in fig.5 , there are huge variations between the two different realizations , i.e. different events .
these large fluctuations in fig.5 arise due to large oscillations in @xmath3 which itself depends on the nature of randomness in the initial field configuration .
note that plots for @xmath89 in fig.6 show a sharp change near @xmath99 fm after which @xmath89 settles down to a value close to zero .
the central temperature at that stage is about 208 mev .
presumably a large part of the region ( away from the central part where temperature is lower due to a woods - saxon profile ) may be undergoing transition at that stage leading to large changes in field dynamics .
we have checked with the contour plots that the ellipticity of the profile of @xmath3 , as well as that of energy density , does not undergo rapid changes during this stage .
( in fact this is a reflection of a shortcoming in our model where the temperature profile is taken to have definite initial eccentricity , and its further expansion is with definite acceleration as discussed above , starting with zero transverse velocity .
a more appropriate model may be to take the time evolution of the spatial eccentricity of the temperature profile from hydrodynamical models and study the evolution of @xmath3 using that . )
we note strong fluctuations in both these quantities around this stage , which may be responsible for these large changes in @xmath89 at this stage .
we hope to develop better understanding of the dynamics during this stage and discuss in a followup work . at present what is important to note
is that , apart from this peak region , everywhere else @xmath89 shows rather stable values settling down to a value close to zero after @xmath100 5.5 fm , and does not fluctuate much , as compared to the quench case in fig.5a . to further illustrate the importance of fluctuations for the quench case , we show plots for the case with zero eccentricity of the qgp region ( i.e. for the temperature profile ) which also will mean zero value of @xmath89 for uniform density case . thus , any non - zero @xmath89 and @xmath94
arise only from the randomness in the initial field configuration . fig.7a and fig.7b show the plots for the equilibrium case .
we note that @xmath89 and @xmath94 remain very small , apart from a large change near @xmath99 fm for @xmath89 , just as in fig.6 . again , at present what is important to note is that , apart from this peak region , everywhere else @xmath89 and @xmath94 remain very small for equilibrium case , as expected for the zero eccentricity case .
different plots in fig.7a and fig.7b correspond to different realizations of the initial random field configuration .
this situation should be contrasted with the quenched case as shown in fig.7c and fig.7d .
note that @xmath89 and @xmath94 now fluctuate with large amplitudes , even though eccentricity of temperature profile is zero .
further , different realizations of the initial random field configurations lead to widely different plots of these quantities .
this shows the fluctuating nature of development of flow anisotropies if large @xmath3 fluctuations are present .
these results suggest the dynamics of the order parameter field , especially such large length scale oscillations may play an important role in determining flow anisotropies and it needs to be incorporated in the hydrodynamical models .
= 5truecm = 5truecm = 5truecm = 5truecm these plots are for the zero eccentricity for the temperature profile . in all the figures here ,
solid and dashed plots correspond to different realizations of initial random field configuration .
( a ) and ( b ) show plots of @xmath89 and @xmath94 for the equilibrium transition case , while ( c ) and ( d ) correspond to the quench case .
we carry out 2@xmath1011 dimensional simulation to study the dynamics of confinement - deconfinement transition as a quench for the central rapidity region in relativistic heavy - ion collision experiments .
we work in the framework of the polyakov loop model . the initial field ( in the confining phase )
is taken to cover a small neighborhood of the confining vacuum @xmath0 as appropriate for the initial @xmath5 = 0 system .
this initial field @xmath3 , unable to relax to the new equilibrium vacuum state with quenched potential at @xmath62 mev , becomes unstable and rolls down in different directions from the top of the central hill in the effective potential of @xmath3 .
we study the formation of z(3 ) domain structure during this evolution . when explicit z(3 ) symmetry breaking effects ( arising from dynamical quark effects ) are small , then we find well defined z(3 ) domains which coarsen in time . during early stages z(3 ) domains ( and z(3 ) domain walls ) have small sizes of order 1 fm .
however at this stage the magnitude of @xmath3 is very small , of order few present of its vacuum expectation value .
domains coarsen rapidly as the magnitude of @xmath3 grows and by the time @xmath102 fm , domains are of size several fm .
surprisingly , the magnitude plot of @xmath3 shows vacuum bubble like configurations , such as those which arise in a first order transition .
this first order transition like behavior occurs even though there is no metastable vacuum separated by a barrier from the true vacuum for the parameter values used .
these bubble like configurations expand as well , somewhat in a similar manner as during a first order transition .
this result points to new interesting possibilities for the phase transition dynamics .
for example , such bubble like structures may lead to a dynamics of phase separation in the case of the universe similar to the original witten s scenario @xcite , even when there is no underlying first order transition .
this may also have important implications for rhice .
this new possibility of transition dynamics needs to be understood and analyzed in detail , see @xcite for a study of these issues . when the initial patch of @xmath3 is only partially symmetric around @xmath7 ( as appropriate for small explicit symmetry breaking from quark effects ) , the dynamics retains these qualitative aspects , with true vacuum domains ( with @xmath8 ) growing dominantly at the cost of the other two metastable z(3 ) domains . in this case z(3 ) walls are fewer and relatively smaller , and they disappear more quickly . when the initial patch of @xmath3 ( around equilibrium point for @xmath9 effective potential )
is significantly shifted towards the true vacuum for the quenched @xmath10 effective potential ( as will happen when explicit symmetry breaking is strong ) , then @xmath3 rolls down roughly along the same direction with angular variations becoming smaller during the roll down . in this case only @xmath8 vacuum survives and no other z(3 ) domains are formed .
also , in this case we do not find bubble - like configurations .
however , in this case we find huge oscillations of @xmath3 with large length scales .
this is similar to the scenarios of reheating via parametric resonance in the case of inflation in the early universe . in our case of rhice
also it raises important questions about the possibility of novel modes of particle production from these large oscillations of @xmath3 during the early stages of the transition .
we have shown that these large @xmath3 oscillations can strongly affect the evolution of flow anisotropies in rhice . the spatial eccentricity and the flow coefficients ,
e.g. the elliptic flow @xmath94 are found to undergo large fluctuations during the evolution of the system .
these results suggest the dynamics of the order parameter field , especially such large length scale oscillations may play an important role in determining flow anisotropies and it needs to be incorporated in the hydrodynamical models .
we are extremely grateful to umashankar gupta for his help in the simulations , especially for the part involving novel bubble like configurations .
we are also very grateful to rajeev bhalerao , vivek tiwari , sanatan digal , p.s .
saumia , abhishek atreya , and partha bagchi for very useful comments and suggestions .
traschen and r.h .
brandenberger , phys . rev . *
d42 * , 2491 ( 1990 ) ; y. shtanov , j. traschen , and r. brandenberger , phys . rev . *
d51 * , 5438 ( 1995 ) ; l. kofman , a. linde , and a.a .
starobinsky , phys .
* d56 * , 3258 ( 1997 ) . | we model the initial confinement - deconfinement transition in relativistic heavy - ion collisions as a rapid quench in view of expected rapid thermalization to a qgp state .
the transition is studied using the polyakov loop model , with the initial field configuration ( in the confining phase ) covering a small neighborhood of the confining vacuum @xmath0 , as appropriate for @xmath1 .
quench is implemented by evolving this initial configuration with the effective potential at a temperature @xmath2 .
we study the formation of z(3 ) domain structure and its evolution during the transition as @xmath3 rolls down in different directions from the top of the central hill in the effective potential of @xmath3 .
when explicit z(3 ) symmetry breaking effects ( arising from dynamical quark effects ) are small , then we find well defined z(3 ) domains which coarsen in time .
remarkably , the magnitude plot of @xmath3 shows vacuum bubble like configurations arising during the quench .
this first order transition like behavior occurs even though there is no metastable vacuum separated by a barrier from the true vacuum for the parameter values used .
when the initial field configuration everywhere rolls down roughly along the same direction ( as will happen with large explicit symmetry breaking ) then we do not find such bubble - like configurations .
however , in this case we find huge oscillations of @xmath3 with large length scales .
we show that such large oscillations can lead to large fluctuations in the evolution of flow anisotropies compared to the equilibrium transition case .
key words : quark - hadron transition , quenched transition , relativistic heavy - ion collisions , z(3 ) domains , bubble nucleation , elliptic flow |
bell s theorem @xcite and the kochen - specker theorem @xcite are perhaps two of the results which have been most influential in developing the modern understanding of quantum mechanics as an irreducibly non - classical theory @xcite .
moreover , these two no - go theorems are seen as the strongest argument for quantum mechanics being a fundamentally indeterministic theory , rather than one ruled by a deeper determinism below the level of the quantum mechanical description of reality .
bell s theorem , which shows that quantum mechanics predicts statistical correlations between separated particles greater than what would be possible in any local , realistic , classical theory , was the focus of attention for several decades due to its relatively clear ability to be tested experimentally @xcite .
the kochen - specker theorem was proved very shortly afterwards , but was largely ignored due to a perceived lack of testability , and perhaps also its formalisation in terms of partial algebras , until it attracted renewed attention with the more recent advances in quantum information theory and foundations .
in contrast to the bounds on probability distributions given by bell s theorem , the kochen - specker theorem shows that the hilbert - space structure of quantum mechanics makes it impossible to assign ` classical ' definite values to all quantum observables in a consistent manner .
since such a definite value is precisely a ( deterministic ) hidden variable specifying , in advance , the result of a measurement of an observable , this means that the outcomes of all quantum measurements on a system can not be simultaneously predetermined .
more recent developments have significantly reduced the size and difficulty of proofs of the kochen - specker theorem @xcite and converted such proofs into testable inequalities @xcite .
however , in showing the impossibility of a classical deterministic ` two - valued ' measure ( i.e. , value assignment ) the kochen - specker theorem leaves open several possible conclusions .
the kochen - specker theorem , more specifically , finds a contradiction between the following three assumptions , which will be formalised more rigorously a little later : * all observables are assigned a definite value ( i.e. , are ` value definite ' ) ; * this definite value should be noncontextual that is , assigned as a function of the observable alone , and not depending on other compatible observables ; * the definite values for a set of compatible observables must be consistent with the theoretical quantum predictions for the relations between them . condition ( iii ) is largely uncontroversial and hence one must generally conclude that either ( or even both ) ( i ) and ( ii ) must be given up .
several alternative interpretations of quantum mechanics are contextual ( e.g. , @xcite ) , and hence discard ( ii ) . perhaps the more popular interpretation , however , is that the inability to simultaneously assign noncontextual definite values , representing predetermined measurement outcomes , to all observables means that measurement outcomes are not determined in advance at all : that quantum mechanics represents a value indefinite reality .
this interpretation is often referred to simply as ` contextuality ' in the literature ; however we reserve this term strictly for the contextual behaviour of definite values .
if we choose to require ( ii ) to hold , at least for any observables that are assigned definite values , then there remains an oft - overlooked gap between the formal result of the kochen - specker theorem and the general interpretation of it .
indeed , the negation of ( i ) is that _ not all _ observables are assigned definite values : it does not prove that no observable can be assigned a definite value and hence , given that definite values represent predetermined measurement outcomes , does not show that all measurements must result in the _ ex nihilo _ creation of an outcome , nor does it allow one to know _ which _ observables in any set are value indefinite .
we can , of course , postulate that if some observables are value indefinite , then this should , by symmetry or uniformity considerations , be the case for all observables ( or at least those for which the born rule assigns a probability strictly between 0 and 1 to some outcomes ) .
however , it is key to realise that this is not in any sense a formal consequence of the kochen - specker theorem , and constitutes an additional , undesired , assumption . in this paper
we address precisely this issue . as is common in modern treatments of the kochen - specker theorem @xcite we focus on one - dimensional ( rank-1 ) projection observables , and we denote the observable projecting onto the linear subspace spanned by a vector @xmath0 as @xmath1 . by using a modified , weakened set of assumptions , we prove that if one such projection observable @xmath2 is assigned the value 1 , then no other such projection observable @xmath3 can be consistently assigned a definite value unless @xmath2 and @xmath3 commute . in interpreting this result physically , we note that if a system is prepared in the state @xmath0 then the outcome of a measurement of the observable @xmath2 is known to be 1 with certainty , and thus any value assignment representing the outcomes of possible measurements on the system should assign the value 1 to @xmath2 .
if @xmath3 does not commute with @xmath2 it is therefore value indefinite under such a value assignment and hence can not have a consistently predetermined measurement outcome .
this self - contained , analytic proof extends and generalises the results of @xcite . throughout the paper we will assume ( ii ) to hold , as is common in interpretations of the kochen - specker theorem , and
our strengthened results and interpretation of the kochen - specker theorem thus rely on this condition .
we do not attempt to justify this assumption here , as this is an interpretational choice and the subject of much debate ( see ( * ? ? ?
4 ) for an overview ) , which is beyond the scope of this paper . as usual
we denote by @xmath4 the set of complex numbers and use the standard quantum mechanical bra - ket notion ; that is , we denote ( unit ) vectors in the hilbert space @xmath5 by @xmath6 . as mentioned above , we will focus on one - dimensional projection observables and denote by @xmath2 the operator projecting onto the linear subspace spanned by @xmath0 ; that is , @xmath7 . in the following we formalise hidden variables and
the notion of value definiteness in a clear and unambiguous fashion .
this framework is based on that we developed in @xcite , and similar to standard approaches to the kochen - specker theorem @xcite ; we have made several simplifications since we do not wish to explore contextual definite values or hidden variable theories in any detail here .
we fix a positive integer @xmath8 .
let @xmath9 be a nonempty set of one - dimensional projection observables on the hilbert space @xmath10 @xcite .
a set @xmath11 is a _ context of @xmath12 _ if @xmath13 has @xmath14 elements ( i.e. , @xmath15 ) and for all @xmath16 with @xmath17 , @xmath18 . since distinct one - dimensional projection observables
commute if and only if they project onto mutually orthogonal linear subspaces , a context @xmath13 of @xmath12 is thus a maximal set of compatible one - dimensional projection observables on @xmath5 . because there is a direct correspondence ( up to a phase - shift ) between unit vectors and one - dimensional projection observables , a context is uniquely defined by an orthonormal basis of @xmath5 . recall that a partial function is one which may be undefined for some values . if it is defined everywhere , then it is total .
[ def : v ] a _ value assignment function ( on @xmath12 ) _ is a partial two - valued function @xmath19 , assigning values to some ( possibly all ) observables in @xmath12 .
we note that we could , as in @xcite , allow @xmath20 to be a function of both the observable @xmath21 and the context @xmath13 containing @xmath21 , allowing values to be assigned contextually
. it would perhaps be more correct to call @xmath20 , as defined above in definition [ def : v ] , a _
noncontextual _ value assignment function ; however , since we are interested only in the noncontextual case , we avoid this for compactness .
an observable @xmath22 is _ value definite ( under @xmath20 ) _ if @xmath23 is defined ; otherwise it is _ value indefinite ( under @xmath20)_. similarly , we call @xmath12 value definite ( under @xmath20 ) if every observable @xmath22 is value definite . with this terminology , we can state the kochen - specker theorem formally .
we present it in the following form deliberately in order to draw the comparison to our earlier informal description , and to clarify the following discussion , even though the second condition is redundant because we require , by definition , that a value assignment function be noncontextual .
let @xmath24 .
then there exists a ( finite ) set of one - dimensional projection observables @xmath12 on the hilbert space @xmath5 such that there is no value assignment function @xmath20 satisfying the following three conditions : * @xmath12 is value definite under @xmath20 ; that is , @xmath20 is a total function . *
the value @xmath23 of an observable @xmath22 depends only on @xmath21 and not the context containing @xmath21 . * for every context @xmath13 of @xmath12 the following condition holds : is a _
boolean frame function _ with weight 1 @xcite . ]
@xmath25 .
the third condition expresses the fact that only one projection observable in a context can be assigned the value 1 . as we mentioned earlier , this is largely uncontroversial : one can simultaneously measure the observables in a context and quantum mechanics predicts precisely that exactly one of these measurements should give the result ` 1 ' , thus any corresponding definite values assigned to these observables should obey this same condition .
hence , if we assume ( ii ) to be true at least for observables that are value definite for which the statement makes clear sense then the kochen - specker theorem requires us to conclude the negation of ( i ) : that @xmath12 can not be value definite , and hence _ at least one _ observable must be value indefinite . note
that the third condition is not independent of the first : it is not clear how the sum @xmath26 should be evaluated if @xmath23 is undefined .
this is one of the key issues we will clarify in attempting to localise value indefiniteness .
while the kochen - specker theorem certainly succeeds , as was the original intention , in showing that quantum mechanics must obey an entirely non - classical event structure , it does not , as we have pointed out , show that all observables must be value indefinite and their outcomes intrinsically indeterministic . as a consequence of the global nature of the hypothesis of the theorem
that _ all _ observables are value definite one can only draw a global conclusion : that _ not all _ observables are value definite .
that is , the theorem , even under the assumption of noncontextuality , can not ` locate ' value indefiniteness to any particular observable .
this is an important point , not only for the foundational understanding of quantum mechanics , but also in practical applications : quantum random number generators and cryptographic schemes rely on the indeterminism of quantum mechanics providing ` irreducible randomness ' @xcite . to certify such claims , it is important to be able to localise value indefiniteness to ensure it applies to the observables measured in such applications .
we proceed by providing more nuanced and less demanding , localised versions of the kochen - specker assumptions , and use these to localise value indefiniteness ( always under the assumption of noncontextuality for value definite observables ) .
our approach is a conservative one : rather than assuming complete value definiteness of the entire set of observables considered , we require observables to be value definite only when their indefiniteness would allow the possibility of measurements violating the quantum predictions specified in condition ( iii ) of the kochen - specker theorem ( see the more detailed discussion and example below ) . in order for this approach to work , we need , as a premise , at least one observable to be value definite .
we then show that the assumption that _ any other _
observable is value definite leads to a contradiction .
fortunately , there is a justification for this premise : if a system is prepared in an arbitrary state @xmath27 , then measurement of the observable @xmath2 should yield the outcome 1 .
thus , it seems reasonable to require that , if @xmath28 , then @xmath29 . we call this the _ eigenstate assumption _ @xcite , which is similar to , although weaker than , the ` eigenstate - eigenvalue link ' discussed in @xcite . furthermore , since the critical feature of a set @xmath12 of projection observables is the orthogonality relations between these observables rather than the specific form of these observables , we can hence choose our basis at will .
it is thus not unreasonable to consider that some one - dimensional projection observable in @xmath12 has the value 1 , and to fix the basis used to express @xmath12 to that of the state @xmath0 to make this observable coincide with @xmath2 .
let us finally discuss how assumption ( iii ) can be generalised for partial value assignment functions @xmath20 , that is , the case where some observables in @xmath12 may be value indefinite .
[ admis - rules ] let @xmath12 be a set of one - dimensional projection observables on @xmath5 and let @xmath30 be a value assignment function .
then @xmath20 is _ admissible _ if the following two conditions hold for every context @xmath13 of @xmath12 : * if there exists a @xmath31 with @xmath32 , then @xmath33 for all @xmath34 ; * if there exists a @xmath31 with @xmath33 for all @xmath34 , then @xmath32 .
admissibility requires that the quantum predictions of ( iii ) are never violated , while allowing the value indefiniteness of an observable @xmath21 if both outcomes ( 0 and 1 ) of a measurement of @xmath21 would be compatible with the definite values of other observables sharing a context with @xmath21 . for example , if @xmath32 , then a measurement of all the observables in a context @xmath13 containing @xmath21 must yield the outcome 1 for @xmath21 , and hence to avoid contradiction the outcome 0 for the other observables in the context . on the other hand , if @xmath35 , even though measurement of @xmath21 must yield the outcome 0 , any of the other observables in @xmath13 could yield the value 1 or 0 ( as long as only one yields 1 ) , hence we should not conclude the value definiteness of these other observables .
let us illustrate the difference between our weakened assumptions , and in particular admissibility , with the hypotheses of the kochen - specker theorem .
ccc 01 [ scale=0.6 ] = [ line width=1pt ] = [ circle , minimum size=6 ] = [ rectangle , minimum size=9 ] = [ draw = none , circle , minimum size=35 ] = [ draw = none , circle , minimum size=12 ] ( 0:5 ) coordinate(1 ) ( 4.167,1.44 ) coordinate(2 ) ( 3.33,2.88 ) coordinate(3 ) ( 60:5 ) coordinate(4 ) ( 0.833,4.33 ) coordinate(5 ) ( -0.833,4.33 ) coordinate(6 ) ( 120:5 ) coordinate(7 ) ( -3.33,2.88 ) coordinate(8 ) ( -4.167,1.44 ) coordinate(9 ) ( 180:5 ) coordinate(10 ) ( -4.167,-1.44 ) coordinate(11 ) ( -3.33,-2.88 ) coordinate(12 ) ( 240:5 ) coordinate(13 ) ( -0.833,-4.33 ) coordinate(14 ) ( 0.833,-4.33 ) coordinate(15 ) ( 300:5 ) coordinate(16 ) ( 3.33,-2.88 ) coordinate(17 ) ( 4.167,-1.44 ) coordinate(18 ) ; \(1 ) ( 2 ) ( 3 ) ( 4 ) ( 5 )
( 6 ) ( 7 )
( 8) ( 9 ) ( 10 ) ( 11 )
( 13 ) ( 14 ) ( 15 )
( 16 ) ( 17 ) ( 18 )
( 1 ) ; \(2 ) ( 9 ) ; ( 11 ) ( 18 ) ; ( 3 ) ( 14 ) ; ( 5 ) ( 12 ) ; ( 6 ) ( 17 ) ; ( 8) ( 15 ) ; \(2 ) arc ( 450:270:2 and 1.44 ) ; ( 3 ) arc ( 90:270:2 and 1.44 ) ; ( 6 ) arc ( 90:270:2 and 1.44 ) ; ( 9 ) arc ( 90:270:2 and 1.44 ) ; ( 12 ) arc ( 90:270:2 and 1.44 ) ; ( 15 ) arc ( 90:270:2 and 1.44 ) ; \(1 ) coordinate[s1,fill , label=0:@xmath36 ; ( 2 ) coordinate[c1,fill ] ; ( 3 ) coordinate[c1,fill ] ; ( 4 ) coordinate[c1,fill ] ; ( 5 ) coordinate[c1,fill ] ; ( 6 ) coordinate[c1,fill , label=90:@xmath37 ; ( 7 ) coordinate[s1,fill , label=90:@xmath38 ; ( 8) coordinate[c1,fill ] ; ( 9 ) coordinate[c1,fill = white , label=150:@xmath39 ; ( 9 ) node[c1,fill = none ] @xmath40 ; ( 9 ) coordinate[c1,draw ] ; ( 10 ) coordinate[c1,fill ] ; ( 11 ) coordinate[c1,fill ] ; ( 12 ) coordinate[s1,fill , label=180:@xmath41 ; ( 13 ) coordinate[c1,fill ] ; ( 14 ) coordinate[c1,fill ] ; ( 15 ) coordinate[s1,fill , label=270:@xmath42 ; ( 16 ) coordinate[c1,fill ] ; ( 17 ) coordinate[c1,fill ] ; ( 18 ) coordinate[c1,fill ] ; ( c1 ) at ( 10 ) ; ( c2 ) at ( 10 ) ; of each observable ( node ) @xmath21 is represented as follows : @xmath32 black square ; @xmath35 filled circle ; @xmath23 undefined ( value indefinite ) hollow circle . (
a ) the contradiction arising when @xmath43 : @xmath20 can not be admissible , since this would require that @xmath44 and @xmath45 simultaneously , as shown by the cross in the diagram .
( b ) a possible admissible value assignment when @xmath46 and @xmath47 . ] & & 01 [ scale=0.6 ] = [ line width=1pt ] = [ circle , minimum size=6 ] = [ rectangle , minimum size=9 ] = [ draw = none , circle , minimum size=35 ] = [ draw = none , circle , minimum size=12 ] ( 0:5 ) coordinate(1 ) ( 4.167,1.44 ) coordinate(2 ) ( 3.33,2.88 ) coordinate(3 ) ( 60:5 ) coordinate(4 ) ( 0.833,4.33 ) coordinate(5 ) ( -0.833,4.33 ) coordinate(6 ) ( 120:5 ) coordinate(7 ) ( -3.33,2.88 ) coordinate(8 ) ( -4.167,1.44 ) coordinate(9 ) ( 180:5 ) coordinate(10 ) ( -4.167,-1.44 ) coordinate(11 ) ( -3.33,-2.88 ) coordinate(12 ) ( 240:5 ) coordinate(13 ) ( -0.833,-4.33 ) coordinate(14 ) ( 0.833,-4.33 ) coordinate(15 ) ( 300:5 ) coordinate(16 ) ( 3.33,-2.88 ) coordinate(17 ) ( 4.167,-1.44 ) coordinate(18 ) ; \(1 ) ( 2 ) ( 3 ) ( 4 )
( 5 ) ( 6 ) ( 7 ) ( 8) ( 9 ) ( 10 ) ( 11 )
( 12 ) ( 13 )
( 14 ) ( 15 ) ( 16 ) ( 17 ) ( 18 ) ( 1 ) ; \(2 ) ( 9 ) ; ( 11 ) ( 18 ) ; ( 3 ) ( 14 ) ; ( 5 ) ( 12 ) ; ( 6 ) ( 17 ) ; ( 8) ( 15 ) ; \(2 ) arc ( 450:270:2 and 1.44 ) ; ( 3 ) arc ( 90:270:2 and 1.44 ) ; ( 6 ) arc ( 90:270:2 and 1.44 ) ; ( 9 ) arc ( 90:270:2 and 1.44 ) ; ( 12 ) arc ( 90:270:2 and 1.44 ) ; ( 15 ) arc ( 90:270:2 and 1.44 ) ; \(1 ) coordinate[c1,fill , label=0:@xmath36 ; ( 2 ) coordinate[c1,fill = white ] ; ( 2 ) coordinate[c1,draw ] ; ( 3 ) coordinate[c1,fill ] ; ( 4 ) coordinate[c1,fill = white ] ; ( 4 ) coordinate[c1,draw ] ; ( 5 ) coordinate[c1,fill ] ; ( 6 ) coordinate[c1,fill = white ] ; ( 6 ) coordinate[c1,draw ] ; ( 7 ) coordinate[c1,fill = white ] ; ( 7 ) coordinate[c1,draw ] ; ( 8) coordinate[c1,fill = white ] ; ( 8) coordinate[c1,draw ] ; ( 9 ) coordinate[c1,fill = white , label=150:@xmath39 ; ( 9 ) coordinate[c1,draw ] ; ( 10 ) coordinate[c1,fill ] ; ( 11 ) coordinate[c1,fill ] ; ( 12 ) coordinate[s1,fill , label=180:@xmath41 ; ( 13 ) coordinate[c1,fill ] ; ( 14 ) coordinate[c1,fill ] ; ( 15 ) coordinate[c1,fill = white ] ; ( 15 ) coordinate[c1,draw ] ; ( 16 ) coordinate[c1,fill = white ] ; ( 16 ) coordinate[c1,draw ] ; ( 17 ) coordinate[c1,fill = white ] ; ( 17 ) coordinate[c1,draw ] ; ( 18 ) coordinate[c1,fill = white ] ; ( 18 ) coordinate[c1,draw ] ; of each observable ( node ) @xmath21 is represented as follows : @xmath32 black square ; @xmath35 filled circle ; @xmath23 undefined ( value indefinite ) hollow circle . ( a ) the contradiction arising when @xmath43 : @xmath20 can not be admissible , since this would require that @xmath44 and @xmath45 simultaneously , as shown by the cross in the diagram .
( b ) a possible admissible value assignment when @xmath46 and @xmath47.,title="fig : " ] + ( a ) & & ( b ) + consider the greechie orthogonality digram shown in fig . [ fig : cabello ] , in which vertices depict observables and smooth lines or curves represent contexts . this well known diagram represents the ` orthogonality ' relations between the observables used in a well known proof of the kochen - specker theorem due to cabello et al .
@xcite , containing only 18 one - dimensional projection observables on @xmath48 .
the kochen - specker theorem implies that there is no way to assign every observable in this diagram a value such that the admissibility requirements hold : exactly one observable in each context should have the value 1 .
let us suppose for the sake of example that @xmath43 and that @xmath20 is admissible .
then , by working from @xmath49 and @xmath50 and applying the admissibility rule ( a ) one deduces that all observables in a context with @xmath49 or @xmath50 must take the value 0 .
one then notices that there are contexts containing 3 observables with the value 0 , so we can deduce from ( b ) that the fourth must have the value 1 .
if we follow this line of reasoning , we can continue to assign values to observables with the admissibility requirements , as depicted in fig .
[ fig : cabello](a ) , where a black square represents the value 1 , and a black circle the value 0 .
as we can see , by considering the contexts @xmath51 and @xmath52 we can infer that @xmath53 must take both the values 1 and 0 respectively : both possibilities contradict the admissibility of @xmath20 , as does the final possibility that @xmath53 is value indefinite .
note that , in fig .
[ fig : cabello](a ) , the contradiction obtained at @xmath53 marked by the cross is a consequence of a specific succession of applications of the admissibility rules ( a ) and ( b ) in definition [ admis - rules ] . by applying these rules in a different order , one can obtain the contradiction also at @xmath54 , @xmath55 , or @xmath56 .
the most important aspect of this reasoning in this context is that it is deterministic : we proceed only by deducing the value definiteness of observables via ( a ) and ( b ) .
now let us assume that @xmath46 and @xmath47 , as depicted in fig .
[ fig : cabello](b ) .
we again apply ( a ) to observables commuting with @xmath49 ; however , we then see that neither ( a ) nor ( b ) can be used again to deduce the value of another observable .
normally , in proving that this diagram permits no consistent assignment of definite values , one would then proceed by assuming that one of the unfilled observables , such as @xmath53 , must have either @xmath45 or @xmath44 , and trying both possibilities .
one can do this when proving the kochen - specker theorem since one assumes ( i ) : that every observable must have a definite value .
however , in order to localise value indefiniteness we do not make this assumption .
hence , the value assignment in fig . [ fig : cabello](b ) , with the observables represented by unfilled circles being value indefinite ( e.g. , @xmath57 undefined ) represents an admissible value assignment .
thus , under the assumption that @xmath46 , fig .
[ fig : cabello ] does not suffice to prove that @xmath58 must be value indefinite , and hence can not be used to localise value indefiniteness .
it is not difficult to see that we reach the same conclusion irrespective of our choice of observables as @xmath49 and @xmath50 . in this paper , in proving the main theorem , we give a set of observables for which this is the case .
that is , there are observable @xmath49 and @xmath50 such that if @xmath46 then both @xmath47 and @xmath59 lead , via admissibility , to contradictions .
let us now state the strengthened theorem which is the focus of this paper .
as we mentioned , this generalises the results of @xcite and uses a different proof technique allowing for a more symmetrised analytic approach .
the result in @xcite , on the other hand , relies on computational results and the interpretation of graphs .
[ thm : main ] let @xmath24 and @xmath60 be unit vectors such that @xmath61 .
we can effectively find a finite set of one - dimensional projection observables @xmath12 containing @xmath2 and @xmath3 for which there is no admissible value assignment function on @xmath12 such that @xmath29 and @xmath3 is value definite . before we proceed to prove theorem [ thm : main ] ,
let us first discuss some important relevant issues .
this theorem has a slightly different form from the standard kochen - specker theorem because of the requirement that a particular observable in the set @xmath12 be assigned the value 1 .
however , since , as we will see , it is only the orthogonality relations between the observables in @xmath12 which is important , a change of basis can always ensure that the required observable @xmath2 be assigned the value 1 . in order to interpret this result
one has to take into account the eigenstate assumption discussed in the previous section : if a quantum system is prepared in a state @xmath0 in @xmath24 dimensional hilbert space , then every one - dimensional projection observable that does not commute with @xmath2 is value indefinite and hence can not have a predetermined measurement outcome .
the first question to address is whether existing kochen - specker diagrams ( i.e. , greechie diagrams specifying the orthogonality relations of @xmath12 ) could be used to provide a set @xmath12 of observables proving theorem [ thm : main ] ; it is not _ a priori _ obvious that such diagrams are unable to do so .
in section [ sec : cabelloexample ] we showed , as an example , that a particular simple and well - known kochen - specker diagram is not sufficient for this purpose .
a careful search through existing diagrams showed that this is the case in general , and we were unable to find an existing kochen - specker diagram in which there are two observables @xmath49 and @xmath50 with the required property that if @xmath46 , both @xmath47 and @xmath59 lead to a contradiction .
a second conceptual problem with the use of fixed kochen - specker diagrams as in existing proofs is the following .
since , in order to derive a contradiction , we need to assume that an observable @xmath2 in the given observable set has @xmath29 , this limits the observables which can be shown to be value indefinite to , at best , the remaining ones in @xmath62 . however , we wish to prove more : that _ every _ one - dimensional projection observable not commuting with @xmath2 is value indefinite . as a result , we need not only a set of observables with the required properties discussed above , but furthermore an approach to generalise this set of observables to arbitrary other observables .
we overcome this apparent lack of generality via a method of reductions , which we present in the next section and will return to discuss later on .
we prove theorem [ thm : main ] in three main steps : 1 .
we first prove it for the special case that @xmath63 . a similar result ( for @xmath64 )
was shown in ref .
@xcite , but this involved two separate diagrams applying to separate cases . here
we give a single diagram providing a much more compact , clear proof .
2 . we prove a simple reduction for @xmath65 to the first case .
the third and main part of the proof involves finding a reduction in the opposite sense , applying to the final @xmath66 case .
it is this final reduction allowing the complete proof that is the most involved technical aspect of this paper . as is standard in kochen - specker proofs @xcite
, we will work directly in the three - dimensional case of @xmath67 ( in fact , only @xmath68 is needed ) , since the case for @xmath69 can be simply reduced to this situation .
[ lemma : explicitcase ] given any two unit vectors @xmath70 with @xmath71 there exists a finite set of one - dimensional projection observables @xmath12 such that if @xmath46 then @xmath50 is value indefinite under every admissible assignment function @xmath20 on @xmath12 . by choosing an appropriate basis we can assume , without loss of generality , that @xmath72 and @xmath73 .
let us consider the set @xmath74 } of one - dimensional projection observables where the vectors @xmath75 for @xmath76 are defined in table [ table : vectorlist ] ( with the normalisation factors emitted for simplicity ) .
the orthogonality relations between these vectors gives the 26 contexts shown in table [ table : contextlist ] .
note that these observables are ` tightly ' connected : the context - observable ratio is relatively high .
the greechie diagram showing the orthogonality relations is shown in fig .
[ fig : greechie ] .
.the 37 vectors specifying the observables used in the proof of lemma [ lemma : explicitcase ] , with normalisation factors omitted . [ cols= " < ,
< , < , < , < " , ] [ table : prooftable2 ] hence , we must conclude that @xmath50 can not be value definite if @xmath20 is admissible on @xmath12 .
we next show a ` contraction ' lemma that constitutes a simple ` forcing ' of value definiteness : given @xmath49 and @xmath50 with @xmath43 , there is a @xmath77 which is ` closer ' ( i.e. , at a smaller angle of our choosing ; contracted ) to both @xmath78 and @xmath79 , for which @xmath45 also .
this result was proved in @xcite , but we reproduce the short proof here for completeness
. the form of the vectors @xmath80 specified in the lemma will be used several times in the rest of the paper .
[ lemma : reduction1 ] given any two unit vectors @xmath70 with @xmath81 and a @xmath82 such that @xmath83 , we can effectively find a unit vector @xmath77 with @xmath84 , and a finite set of one - dimensional projection observables @xmath12 containing @xmath49 , @xmath50 , @xmath53 such that if @xmath43 , then @xmath45 , for every admissible assignment function @xmath20 on @xmath12 .
furthermore , if we choose our basis such that @xmath85 and @xmath86 , where @xmath87 , then @xmath77 can only be one of the following two vectors : @xmath88 , where @xmath89 , @xmath90 and @xmath91 . without loss of generality , we assume the @xmath92 and choose a basis so that @xmath93 and @xmath94 where @xmath87 and @xmath95 .
note that , since @xmath96 and thus @xmath97 we have @xmath98 if we let @xmath99 we thus have @xmath100 we can then set @xmath101 , making @xmath102 a unit vector such that @xmath103 .
let @xmath104 , @xmath105 and note that @xmath106 also .
thus , if we let @xmath107 and @xmath108 , then @xmath109 , @xmath110 and @xmath111 are all orthonormal bases for @xmath68 and thus @xmath112 , @xmath113 and @xmath114 are all contexts in @xmath115 .
this construction is illustrated in fig . [
fig : reduction1 ] .
if @xmath20 is an admissible assignment function on @xmath12 with @xmath43 then we must have @xmath116 and hence @xmath45 , as required . .
once again , the circles and squares represent observables that have the values @xmath117 and @xmath118 respectively.,title="fig : " ] 01 [ scale=0.6 ] = [ line width=1pt ] = [ circle , minimum size=6 ] = [ rectangle , minimum size=9 ] = [ draw = none , circle , minimum size=4 ] ( 5,0 ) coordinate[s1,fill , label=0:@xmath36 ( b ) ( 10,4 ) coordinate[c1,fill , pos=0.5 ] ( c3 ) coordinate[l1,pos=0.7,label=0:@xmath119 coordinate[c1,fill , label=0:@xmath120 ( beta ) ( 0,4 ) coordinate[s1,fill , pos=0.5,label=90:@xmath39 ( c ) coordinate[l1,pos=0.35,label=25:@xmath52 ] coordinate[c1,fill , label=180:@xmath122 ( alpha ) ( 5,8 ) coordinate[c1,fill , pos=0.5 ] ( c2 ) coordinate[l1,pos=0.3,label=180:@xmath51 ] coordinate[s1,fill , label=0:@xmath41 ( a ) ; we now present a proof for the reduction in the opposite direction : finding ( from @xmath124 ) two vectors @xmath125 specifying observables @xmath126 for which @xmath127 , and which are further apart from each other than @xmath78 is from @xmath79 .
this is made easier by noting that it is not necessary to find a vector @xmath77 ` further ' from @xmath78 than @xmath79 , but rather just two vectors further from each other than @xmath78 is from @xmath79 .
this process is broken into two steps .
we first prove an ` expansion lemma ' which , unlike the contraction lemma , does not find two vectors arbitrarily far apart satisfying the required criteria .
rather , we then show a further lemma , the ` iteration lemma ' , proving that this expansion can be iterated to meet the required conditions .
[ lemma : reduction2 ] given any two unit vectors @xmath70 with @xmath128 , we can effectively find unit vectors @xmath125 with @xmath129 and a finite set of one - dimensional projection observables @xmath12 containing @xmath130 such that if @xmath43 , then @xmath131 , for every admissible assignment function @xmath20 on @xmath12 .
let @xmath132 .
without loss of generality , we will consider only the positive , real case of @xmath133 .
we fix an orthonormal basis such that , written in this basis , @xmath78 and @xmath79 lie in the @xmath134-plane bisected by the @xmath135-axis . in this basis
we thus have @xmath136 where @xmath137 it is readily confirmed that @xmath138 as desired .
note that we thus have @xmath139 figure [ fig : lemma2sphere ] shows the contour representing all the possible vectors specifying observables which can be forced to take the value 1 from the construction in lemma [ lemma : reduction1 ] .
we use two applications of lemma [ lemma : reduction1 ] applied to @xmath124 to give two such vectors @xmath125 lying in the @xmath140-plane .
corresponding to the one - dimensional projection observables that lemma [ lemma : reduction1 ] can force to take the value 1 .
the bold ( red ; colour online ) curve represents the position on the unit sphere of such vectors for given @xmath124 .
note the @xmath77 and @xmath141 are further apart from each other than @xmath78 and @xmath79 . ]
we can also see , at least for the chosen values of @xmath124 that are shown in fig .
[ fig : lemma2sphere ] , that @xmath142 .
indeed it appears that the vectors ` @xmath77 ' , ` @xmath141 ' shown in the @xmath140-plane provide the maximum separation , and the symmetry under exchange of @xmath78 and @xmath79 of lemma [ lemma : reduction1 ] seems to support this .
however , it is not necessary to prove this is the case .
rather , we will show directly that the vectors @xmath125 provide the required expansion .
to do so , we derive a simple explicit form for @xmath125 and thus @xmath143 .
we focus first on finding @xmath77 ; the form of @xmath141 follows immediately . rather than use basis - transformations to attempt to apply lemma [ lemma : reduction1 ] to find the form of @xmath125 in this specific case , we will re - derive the result explicitly making use of our symmetrised basis choice .
the vectors @xmath144 need to follow the orthogonality relations shown in fig .
[ fig : reduction1 ] in order to conclude that @xmath145 .
that is , we need vectors @xmath146 such that @xmath147 is an orthonormal set , and further that @xmath148 . since we choose @xmath77 to be in the @xmath140-plane , we can write it in the parameterised form @xmath149 , where @xmath150 remains to be found .
since @xmath151 should be orthogonal to both @xmath78 and @xmath77 , we have @xmath152 similarly , we have @xmath153 further , the orthogonality of @xmath151 and @xmath154 gives us @xmath155 and hence @xmath156 thus , @xmath157 further , it is readily verified that @xmath158 for @xmath159 and hence for all @xmath160 ( recall eqn .
[ eq : betabound ] ) . similarly , we find @xmath161 using a further two auxiliary vectors @xmath162 forming the orthonormal set @xmath163 where @xmath164 .
thus , if we take @xmath165 , as a result of the orthogonality relationships expressed in fig .
[ fig : reduction1 ] , @xmath127 for any admissible @xmath20 on @xmath12 with @xmath43 .
it remains then just to show that @xmath166 we note that @xmath167 for @xmath168 .
we finish the proof by showing proving eqn .
[ eqn : gleb ] , that is , that @xmath169 , or , equivalently , @xmath170 .
but since we can write @xmath171 we have from eqn . [ eqn : gamma ] @xmath172 concluding the proof .
we note for completeness that we can write @xmath143 directly in terms of @xmath173 from eqns .
[ eqn : beta ] , [ eqn : gamma ] and [ eqn : gleb ] as @xmath174 we now prove that by iterating this procedure we can find a pair of vectors arbitrarily far apart from each other .
[ iteration lemma ] [ lemma : reduction3 ] given any two unit vectors @xmath70 with @xmath175 , we can effectively find unit vectors @xmath125 with @xmath176 and a finite set of one - dimensional projection observables @xmath12 containing @xmath177 such that if @xmath43 , then @xmath131 , for every admissible assignment function @xmath20 on @xmath12 .
we prove by iterating lemma [ lemma : reduction2 ] , and use the notation @xmath178 and @xmath179 , indicating the 0th iteration .
we start with @xmath180 and for each @xmath181 , as long as @xmath182 satisfy @xmath183 , apply the construction used in the proof of lemma [ lemma : reduction2 ] to generate @xmath184 for the next iteration . in particular , @xmath184 satisfy the equality for @xmath185 ( in particular , @xmath186 ) . by lemma [ lemma : reduction2 ] , we know that @xmath187 for each iteration @xmath188 .
we now prove that the process can not produce an infinite sequence @xmath189 , with @xmath190 for all @xmath188 , that is , for some @xmath188 we have @xmath191 .
( the sequence must stop here , since lemma [ lemma : reduction2 ] can not be applied for @xmath192 . ) from eqn .
[ eqn : iprodfn ] we define the function @xmath193 such that @xmath194 giving the inner product of the next pair in the iteration .
we thus have @xmath195 and , more generally , @xmath196 .
we can thus rephrase the problem : _ does there exist a @xmath197 such that @xmath198 ? _ let us , for the sake of contradiction , assume the contrary . then @xmath199 is an infinite strictly decreasing sequence of reals with @xmath200 for all @xmath188 . for any finite @xmath188 we thus have @xmath201 let us define the function @xmath202 such that @xmath203 so that @xmath204 we can show that @xmath205 for @xmath206 : calculating the derivative we have @xmath207 since @xmath208 is thus a strictly decreasing function on @xmath209 and @xmath210 for all @xmath211 , we have @xmath212 for all @xmath211 .
hence we set @xmath213 since @xmath214 is a positive constant , it is not possible that @xmath215 , for all @xmath216 , because in this case we would have @xmath217 , for all @xmath216 , a contradiction .
in fact , if @xmath197 is the smallest positive integer greater than @xmath218 , then @xmath219 , as required .
we note that @xmath220 is not defined . by lemma [ lemma : reduction2 ] , for each @xmath221 there exists a set @xmath222 of one - dimensional projection observables such that @xmath223 under any @xmath20 admissible on @xmath222 satisfying @xmath224 . hence ,
if we take the set @xmath225 we must have @xmath226 under any admissible @xmath20 on @xmath12 satisfying @xmath43 , and @xmath227 , as required .
with these lemmata proved , we are in a position to combine them to prove theorem [ thm : main ] .
if we have @xmath63 then , by lemma [ lemma : explicitcase ] , there exists a finite set @xmath12 of one - dimensional projection observables for which there is no admissible @xmath20 on @xmath12 satisfying the requirements , so we are done . otherwise , we proceed directly to prove that if @xmath12 is a set of one - dimensional projection observables containing @xmath228 then no admissible assignment function @xmath20 on @xmath12 with @xmath29 can have @xmath3 value definite .
we show this in two cases : first that @xmath229 and then that @xmath230 .
let us first show that there is a set @xmath231 for which @xmath229 if @xmath20 is admissible on @xmath231 .
there are two cases : either @xmath232 or @xmath233 .
if @xmath232 , then by lemma [ lemma : reduction1 ] there exists a vector @xmath234 such that @xmath235 and a set @xmath236 of observables containing @xmath237 such that if @xmath20 is admissible on @xmath236 , @xmath238 also .
but , by lemma [ lemma : explicitcase ] , there exists a set @xmath239 of one - dimensional projection observables containing @xmath240 such that if @xmath20 is admissible on @xmath239 and @xmath29 , @xmath241 must be value indefinite . thus , if we take @xmath242 we can not have @xmath243 as required . if @xmath233 , then by lemma [ lemma : reduction3 ] there exist two vectors @xmath244 such that @xmath245 and a set @xmath246 of observables containing @xmath247 such that if @xmath20 is admissible on @xmath246 then @xmath248 also .
but , by lemma [ lemma : reduction1 ] , there exists a vector @xmath249 such that @xmath250 and a set @xmath251 of observables containing @xmath252 such that if @xmath20 is admissible , @xmath253 also .
finally , once more by lemma [ lemma : explicitcase ] , there exists a set @xmath254 for which @xmath20 there is no admissible @xmath20 on @xmath251 satisfying @xmath255 .
hence , there is no admissible @xmath20 on the set @xmath256 such that @xmath243 as required .
this shows that there exists a set @xmath231 of one - dimensional projection observables containing @xmath228 such that we can not have @xmath243 if @xmath29 if @xmath20 is admissible @xmath231 . it remains to show that there exists a set @xmath257 such that we can not have @xmath258 if @xmath20 is admissible on @xmath257 .
let us assume , without loss of generality , that @xmath259 and @xmath260 where @xmath261 .
then let @xmath262 , @xmath263 and @xmath264 . then @xmath265 and @xmath266 are orthonormal bases for @xmath67 and hence @xmath267 and @xmath268 are contexts in @xmath269 .
but if @xmath20 is admissible on @xmath270 and @xmath29 , @xmath258 , admissibility implies that @xmath243 . as we have shown just before , there exists a set @xmath271 of one - dimensional projection observables containing @xmath240 such that there is no admissible assignment function @xmath20 on @xmath271 with @xmath272 , and hence there is no admissible @xmath20 on @xmath273 such that @xmath29 and @xmath258 .
having covered all cases , we are forced to conclude that there is a set @xmath274 of observables containing @xmath2 and @xmath3 such that if @xmath29 , @xmath3 can not be value definite if @xmath20 is admissible on @xmath12 .
the important difference between theorem [ thm : main ] and the kochen - specker theorem lies in what physical conclusions can be drawn from the theorems which , of course , are purely mathematical results .
key to interpreting such theorems is the recognition that a value assignment represents a possible hidden variable assignment for a quantum system , and that the value assigned to an observable thus represents the result that would be obtained upon its measurement . under this interpretation
the kochen - specker theorem shows that , given a system prepared in the quantum state @xmath0 in dimension 3 or higher hilbert space , the results of all possible measurements on the state @xmath0 can not be predetermined ( noncontextually ) as they would in a classical theory .
it says nothing , however , about whether all , or simply a few , outcomes are not predetermined . on the other hand ,
theorem [ thm : main ] implies that no one - dimensional projection observable @xmath21 can have a predetermined measurement outcome for the system unless @xmath0 is an eigenstate of @xmath21 .
this interpretation relies on the eigenstate assumption discussed earlier in the paper , stating that the observable @xmath2 has a predetermined measurement outcome a very weak assumption .
conceptually , this means that theorem [ thm : main ] goes further than the kochen - specker theorem in showing the _ extent _ of non - classicality that the quantum logic event - structure implies .
it is possible to generalise this result that formally applies only to one - dimensional projection observables to the value - indefiniteness of more general classes of observables . since an observable @xmath275 ( formally a hermitian operator in @xmath14-dimensional hilbert space ) with a non - degenerate spectrum , distinct eigenvalues @xmath276 and eigenstates @xmath277 can be expressed as its spectral decomposition @xmath278 ( where @xmath279 , as usual ) , it physically has a predetermined measurement outcome if and
only if all the projectors @xmath280 , @xmath281 , have predetermined measurement outcomes has the predetermined value @xmath118 then one must obtain @xmath282 upon measurement of @xmath275 ; admissibility then requires that all @xmath283 have the definite value 0 for @xmath284 . ] that is , are value definite .
thus , for a system prepared in a state @xmath0 in dimension 3 or higher hilbert space , the outcome of a measurement of an observable @xmath275 with non - degenerate spectra can not be predetermined ( noncontextually ) unless @xmath0 is an eigenstate of @xmath275 . since the first appearance of the kochen - specker theorem @xcite
, much attention has been given to reducing the number of observables and contexts needed to obtain a contradiction and prove the theorem .
the original result used a set of 117 observables , but more recent results have , to quote some notable examples , shown sets containing 31 observables ( in three - dimensional hilbert space ) @xcite and 18 observables ( in four - dimensional hilbert space ) @xcite .
while such results do not affect the interpretation of the theorem , they have merit in showing the depth of the contradiction between the classical and quantum logical structures .
more recently , smaller proofs have been of particular interest since these have been used to derive noncontextuality inequalities that can be experimentally tested @xcite in the same vein as bell - inequalities @xcite ; smaller sets of observables lead to smaller and more readily testable inequalities .
conceptually , however , the key point is probably that the theorem can be proved using a _
finite _ set of observables ; if a contradiction only arose when an infinity of observables were considered , this would potentially raise questions about the constructive and operational character of the theorem and its use of counterfactuals , hence its interpretation would be more questionable @xcite .
the localised nature of theorem [ thm : main ] immediately means that a single finite set @xmath12 of one - dimensional projection observables will never suffice to prove the value indefiniteness of all such projection observables @xmath3 not commuting with @xmath2 for a given state @xmath0 .
there are infinitely many such observables , and one must , by definition , include @xmath3 in @xmath12 to localise value indefiniteness to @xmath3 .
rather , the nature of theorem [ thm : main ] means we must look for constructive methods to obtain a set @xmath285 for a given @xmath3 , which is precisely what we have done in our proof of the result . of course , a given set of orthogonality relations ( i.e. , a greechie diagram ) may be realisable for an infinity of different such sets @xmath12 , as is the case with the diagram depicted in fig .
[ fig : reduction1 ] .
thus , it would be preferable to find a given set of orthogonality relations for which a set @xmath285 of observables realising these relations and containing both @xmath2 and @xmath3 for any @xmath3 . since we were unable to give such a set of relations , we had to iterate lemma [ lemma : reduction2 ] a number of times times depending on @xmath2 , with no upper bound ( but only ever finitely many times ) . furthermore , it seems that it is difficult , if not impossible , to succeed in giving a fixed set of orthogonality relations that works in all cases . in order to show an observable @xmath49
has @xmath46 using the admissibility requirements , one must give a context @xmath286 for which it is already known that @xmath287 .
this implies two observable @xmath54 and @xmath55 such that @xmath288 and @xmath289 .
but this is precisely the case described in lemma [ lemma : reduction1 ] .
however , in lemma [ lemma : reduction2 ] we showed the limitations of this process in ` widening the angle ' between vectors whose corresponding projectors both take the value 1 hence the necessity of iterating lemma [ lemma : reduction2 ] . as a result it seems that , in contrast to the kochen - specker theorem , arbitrarily large ( but always finite ) sets of observables are needed to show that a given observable @xmath3 is value indefinite
nonetheless , the critical point is once again that for _ any given _
@xmath3 , we can show that @xmath3 is value indefinite with a _
finite _ set of observables , and hence that the counterfactual reasoning used is no more problematic than in the kochen - specker theorem .
one of the strengths of the kochen - specker theorem that has been repeatedly emphasised is the fact that the contradiction between its hypotheses is derived independently of the state a quantum system is prepared in ; this is commonly referred to as state - independence .
this is in contrast to violation of bell - type inequalities ( which occur only for particular entangled states ) and shows that the non - classicality results from the structure of quantum mechanics itself , rather than features of particular states , such as entanglement @xcite .
consequently , various experimental inequalities based on the kochen - specker theorem that , although often simpler , are state - dependent have been criticised , and much effort has been expended to find simple , state - independent inequalities to test @xcite .
in contrast to the kochen - specker theorem , the form of theorem [ thm : main ] and , in particular , the interpretation ( relying , of course , on the eigenstate assumption ) that _ for a system prepared in a given state @xmath0 _ , any one - dimensional projection observable @xmath3 not commuting with @xmath2 is value indefinite , may suggest that theorem [ thm : main ] does not share this state - independence . as a result , this issue deserves a little discussion .
the state - independence of the kochen - specker theorem ensures that no quantum state in @xmath24 dimensional hilbert space admits a classical assignment of definite values to all observables within certain finite sets .
this is true also with theorem [ thm : main ] : for any quantum state @xmath0 , all projection observables not contained within the ` star ' of one - dimensional projection observables commuting with @xmath2 ( see fig .
[ fig : star ] ) are value indefinite . with @xmath29 and the ( infinite ) set of compatible observables @xmath3 for which @xmath258 .
this is the maximal extent of value definiteness for a system in state @xmath0 no other one - dimensional projection observables on @xmath67 can be value definite.,title="fig : " ] 01 [ scale=0.8 ] = [ line width=1pt ] = [ draw , line width=1pt , inner sep=0 ] = [ rectangle , minimum size=6 ] = [ circle , draw = none , fill , minimum size=2 ] = [ draw = none , circle , minimum size=45 ] ( 0:0 ) ( 135:3 ) coordinate[c1,minimum size=14,fill , at start ] ( 0 ) coordinate[c1,circle , midway , fill ] ( 1 ) coordinate[c1,circle , at end , fill , label=35:@xmath51 ] ( 2 ) ; ( 0.center ) ( 90:3 ) coordinate[c1,at start ] ( 0 ) coordinate[c1,circle , midway , fill ] ( 3 ) coordinate[c1,circle , at end , fill , label=350:@xmath52 ] ( 4 ) ; ( 0.center ) ( 45:3 ) coordinate[c1,at start ] ( 0 ) coordinate[c1,circle , midway , fill ] ( 5 ) coordinate[c1,circle , at end , fill , label=305:@xmath119 ( 6 ) ; ( 0.center ) ( 315:3 ) coordinate[c1,at start ] ( 0 ) coordinate[c1,circle , midway , fill ] ( 9 ) coordinate[c1,circle , at end , fill , label=215:@xmath290 ( 10 ) ; ( 0.center ) ( 270:3 ) coordinate[c1,at start ] ( 0 ) coordinate[c1,circle , midway , fill ] ( 11 ) coordinate[c1,circle , at end , fill , label=170:@xmath291 ( 12 ) ; ( 0.center )
( 225:3 ) coordinate[c1,at start ] ( 0 ) coordinate[c1,circle , midway , fill ] ( 13 ) coordinate[c1,circle , at end , fill , label=125:@xmath292 ( 14 ) ; ( 0.center )
( 0:3 ) coordinate[c1,fill , at start ] ( 0 ) coordinate[c1,circle , fill , midway , fill ] ( 7 ) coordinate[c1,circle,,fill , at end , label=260:@xmath293 ( 8) ; \(0 ) at ( 0.center ) ; ( . ) at ( 190:2 ) ; ( . ) at ( 180:2 ) ; ( . ) at ( 170:2 ) ; rather , it is not theorem [ thm : main ] that is state - dependent , but the proof we have given : to show that a given observable @xmath3 is value indefinite from the assumption that @xmath29 , we need a set @xmath12 particular to this @xmath294 .
however , as we discussed in the preceding section , this is perfectly reasonable given the form of the theorem .
one can emphasise further the state - independence of theorem [ thm : main ] by restating the theorem in the following form : _ `` only a single one - dimensional projection observable on the hilbert space @xmath5 for @xmath24 can be assigned the value 1 by an admissible , noncontextual value assignment function''_. in this form the state - independence is clear ; the illusion of state - dependence enters because of the connection , via the eigenstate assumption , between the `` one observable assigned the value 1 '' and the particular state @xmath0 ( and corresponding observable @xmath2 with @xmath29 ) which is necessary for the physical interpretation of the theorem .
the importance of the state - independence of the kochen - specker theorem arises , in part , in the use of kochen - specker sets of observables in testable inequalities .
it is important to note that , even though these inequalities are sometimes referred to as `` kochen - specker inequalities '' @xcite , they are better seen simply as noncontextuality inequalities .
these inequalities are derived under the assumption only of noncontextuality , ignoring the admissibility requirements , and bounds on quantities are calculated over all possible noncontextual value assignments .
a key result shows that one can derive such an inequality from any kochen - specker set @xcite .
it is clear that these value assignments can not obey the admissibility requirements , since the kochen - specker theorem shows precisely that no classical value assignment can do so .
the strength of theorem [ thm : main ] , on the other hand , relies precisely on the use of the admissibility requirements to determine when definite values should be assigned . hence , while one can use the methods of @xcite to derive inequalities from the constructions in the proof of theorem [ thm : main ] , these bounds would be calculated over all noncontextual value assignments ( subject to @xmath29 ) , without paying heed to admissibility , and hence would offer no conceptual advantage over existing inequalities .
furthermore , since our construction in lemma [ lemma : explicitcase ] , for example , contains 37 observables , these would pose no experimental benefit to existing , simpler inequalities either @xcite .
nonetheless , the state - independence of the result shows that the value indefiniteness of almost all one - dimensional projection observables in quantum mechanics is indeed a deep feature of the theory of the logical structure of hilbert space rather than a property of particular states .
in summary , we proved a variant of the kochen - specker theorem showing that the non - classicality implied by the kochen - specker theorem is , in a specific sense , maximal .
specifically , under the assumptions that ( 1 ) any value definite observables behave noncontextually , and ( 2 ) contexts obey weak ` admissibility ' rules on any value definite observables they contain , we show that only one one - dimensional projection observable on the hilbert space @xmath5 ( for @xmath24 ) can be assigned consistently the definite value 1 .
if a quantum system is prepared in a state @xmath0 , subject to the assumption that any value assignment function for the system must assign the value 1 to @xmath2 since @xmath295 , the theorem can be interpreted as showing that the measurement of any observable @xmath3 projecting onto the linear subspace spanned by a state @xmath294 that is neither orthogonal nor co - aligned with @xmath0 must be value indefinite that is , indeterministic .
this interpretation , which shows that almost all one - dimensional projection observables are value indefinite for a given system @xcite , is stronger than what can be drawn from the kochen - specker theorem , which , in contrast , shows only that not all observables can be value definite .
this result justifies further the general belief that quantum mechanics is indeterministic that there is no hidden variable or definite value determining the outcome of a measurement in advance .
this eliminates the need to assume that the non - classicality shown by the kochen - specker theorem should apply uniformly , instead deriving this global value indefiniteness . as with the kochen - specker theorem ,
this result relies on the assumption that classical values , should they exist , must behave noncontextually .
finally , these results help theoretically certify quantum random number generators @xcite , since the promises of such devices rely on the indeterministic nature of quantum measurements @xcite . by localising value
indefinite observables , one can be sure that the measurements producing the output bits do not yield any pre - existing element of physical reality .
we emphasise that these results do not hold for two - dimensional systems a class into which many current quantum random number generators unfortunately fall .
22ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) link:\doibase 10.1512/iumj.1968.17.17004 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.65.803 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.49.91 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.101.210401 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physreva.86.062109 [ * * , ( ) ] link:\doibase 10.1103/physreva.89.032109 [ * * , ( ) ] _ _ , @noop ph.d . thesis , ( ) @noop _ _ ( , , ) link:\doibase 10.1512/iumj.1957.6.56050 [ * * , ( ) ] link:\doibase 10.1103/physreva.75.032334 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase physrevlett.109.150401 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) | the kochen - specker theorem proves the inability to assign , simultaneously , noncontextual definite values to all ( of a finite set of ) quantum mechanical observables in a consistent manner .
if one assumes that any definite values behave noncontextually , one can nonetheless only conclude that _ some _ observables ( in this set ) are value indefinite . in this paper
we prove a variant of the kochen - specker theorem showing that , under the same assumption of noncontextuality , if a single one - dimensional projection observable is assigned the definite value 1 , then no one - dimensional projection observable that is incompatible ( i.e. , non - commuting ) with this one can be assigned consistently a definite value . unlike standard proofs of the kochen - specker theorem , in order to localise and show the extent of value indefiniteness this result requires a constructive method of reduction between kochen - specker sets .
if a system is prepared in a pure state @xmath0 , then it is reasonable to assume that any value assignment ( i.e. , hidden variable model ) for this system assigns the value 1 to the observable projecting onto the one - dimensional linear subspace spanned by @xmath0 , and the value 0 to those projecting onto linear subspaces orthogonal to it .
our result can be interpreted , under this assumption , as showing that the outcome of a measurement of _ any _ other incompatible one - dimensional projection observable can not be determined in advance , thus formalising a notion of quantum randomness . |
heat conduction is a hot topic in non - equilibrium physics @xcite .
while there is a large body of works contributing to understand the empirically observed fourier s law of conduction , many issues are still challenging , specially in low dimensions ( see @xcite and references therein ) .
for one - dimensional systems , fourier s law takes the form @xmath6 , where @xmath7 is the flux , @xmath8 the temperature gradient , and @xmath3 the heat conductivity , that depends on the system and can also depend on the temperature but not on system size . while , for systems with momentum non - conservative thermal noise @xcite , an - harmonic pinned systems @xcite , or systems with local reservoirs @xcite , normal transport ( hence , finite @xmath3 ) is observed .
differently , anomalous transport typically occurs in other one - dimensional model systems @xcite . in those cases ,
the conductivity can exhibit a divergent dependence on system size and super - diffusion occurs @xcite , hence the fourier s law is not satisfied .
this scenario has been attributed to momentum conservation @xcite . in apparent contradiction ,
the fourier s law does hold for the momentum conservative model of rotators with nearest - neighbors interactions @xcite .
but , in this case , another quantity , the stretch , is not conserved @xcite . however , when extra mechanical forces act at the extremities @xcite , counter - intuitive effects on the flux and on the temperature profile occur also in the rotators model . besides that , it is known that the range of the interactions can bring new features to a system .
ensemble inequivalence , phase transitions , relaxation times that increase with systems size and formation of quasi - stationary states , amongst others , can emerge when the interactions are sufficiently long - range @xcite . in the present context ,
for instance , the long relaxation times observed in long - range systems @xcite might affect thermal conductivity .
then , a natural question is : which is the influence of the range of the interactions in heat conduction ?
addressing this issue , which can bring new insights to the above scenario , is the aim of the present work . for that purpose
, we consider a paradigmatic model system governed by a hamiltonian dynamics , which generalizes the rotators model .
it is known in the literature as @xmath9-xy @xcite , whose parameter @xmath9 allows to adjust the range of the interactions from the nearest - neighbors to the global ( mean - field ) cases .
the model @xmath9-xy consists of a chain of classical rotators , attached to the sites @xmath10 of a one - dimensional lattice , whose dynamics is governed by the hamiltonian @xcite @xmath11 \\ & \equiv \sum_{i=1}^n \mathcal{h}^\alpha_i \equiv \sum_{i=1}^n \left [ \mathcal{t}_i + \mathcal{u}^\alpha_i \right ] , \end{split } \label{hamiltonian}\ ] ] where @xmath12 is the angular momentum , @xmath13 the rotational inertia , @xmath14 the angular position of the classical rotor , and we set @xmath15 in order to get a proper kac prescription factor @xcite , that guarantees extensive energies in the thermodynamic limit ( tl ) when @xmath16 and is adequate to our ( free ) boundary conditions . in the limits @xmath17 and @xmath18 , we recover the _
infinite range _ or mean - field ( m - f ) and _ first nearest neighbors _ (
n - n ) cases , respectively . at the two ends of the lattice we apply ( short - range ) heat baths with temperatures @xmath19 , with @xmath20 .
to model each reservoir , we use a langevin heat bath .
therefore , the equations of motion are @xmath21 where @xmath22 is the angular velocity , @xmath23 @xmath24 are the damping coefficients of the langevin force and @xmath25 are white noises with correlations @xmath26 the flux is defined through the energy continuity equation for each particle , @xmath27 and , under the condition of local stationarity , we obtain the flow of heat towards the particle @xmath28 due to the particle @xmath29 @xmath30 hence , we define the flux from the left ( right ) particles towards the particle @xmath28 as @xmath31 in the stationary state @xmath32 , for all @xmath28 .
moreover , the `` temperature '' at each particle position is defined as twice the mean kinetic energy @xmath33 , which allows to depict a temperature profile along the system length .
indicated on the figure . for each @xmath9 ,
averages over 50 realizations are computed .
@xmath34 , @xmath35 , and @xmath36 .
, scaledwidth=45.0% ] the equations of motion ( [ eq : alphaxyeqmotion ] ) were integrated by means of a brownian dynamics protocol @xcite that reduces to a velocity - verlet algorithm in the absence of interactions with the heat reservoirs , as in the case of the bulk particles , which are not directly coupled to the reservoirs . the fixed time step @xmath37 for numerical integration was selected so as to keep the energy of the corresponding isolated system constant within an error @xmath38 of order @xmath39 .
initial conditions ( @xmath40 ) were set as follows : angles and momenta were randomly chosen around zero ( and the average momentum subtracted ) , within intervals adequate to reproduce the equilibrium temperature @xmath41 for the isolated system , before switching on the reservoirs .
after the thermal baths are connected and a transient has elapsed , the quantities of interest were averaged over at least 30 different initial conditions and along a time interval @xmath42 . without loss of generality ,
in the numerical simulations reported here , we fixed the following values of the parameters of the hamiltonian ( [ hamiltonian ] ) : @xmath43 , @xmath44 for all @xmath45 , and @xmath46 . in fig .
[ fig : profile ] , we show typical temperature profiles for different values of @xmath9 . these profiles for @xmath47 and @xmath48 do not change substantially for values of @xmath49 larger than the value used in the fig .
[ fig : profile ] .
we observe that , in the bulk region , the profiles are almost linear .
the absolute value of the slope is very close to its maximal value @xmath50 for n - n interactions ( @xmath18 ) , but the curves become less steep as @xmath9 decreases . for sufficiently long - range interactions , when @xmath51 , the bulk profile becomes flat , and also more noisy .
vs @xmath49 , for different values of @xmath9 indicated on the figure .
dashed lines are guides to the eye .
the relative standard error associated to each symbol is about 100% , 10% and 1% for @xmath52 @xmath53 , @xmath54 and @xmath55 , respectively . the solid line , with slope -1 ,
was drawn for comparison . @xmath34 and @xmath35 .
, scaledwidth=50.0% ] the stationary flux @xmath56 through the chain was computed by averaging eq .
( [ flux ] ) over the bulk particles , namely , @xmath57 .
the scaled flux @xmath58 , for fixed @xmath9 , is depicted as a function of the size @xmath49 in fig .
[ fig : flux ] , for the same values of the end temperatures considered in fig .
[ fig : profile ] . for @xmath9 above @xmath59 , the scaled flux grows with @xmath49 attaining a finite value , like in the n - n limit @xcite .
differently , for any @xmath9 below @xmath59 , the scaled flux monotonically decays with @xmath49 .
therefore , a distinct behavior of @xmath58 vs @xmath49 emerges for short and long range interactions . in the limit case @xmath60
, the scaled flux presents a neat decay as @xmath61 , indicating that the flux vanishes in the tl .
apparently , a decay towards zero flux also occurs for any @xmath62 .
it is noteworthy that , although the temperature profiles for small @xmath9 are very similar to those observed for chains of identical masses with n - n harmonic interactions @xcite , differently , in the latter case the flux is significantly non null .
let us analyze the m - f case ( @xmath60 ) .
in the tl , a null current along the chain is expected , because , on the one hand , each rotor interacts with each other rotor with equal intensity , on the other , the contribution of the end rotors ( which are the only ones able to break the m - f symmetry ) becomes negligible compared to the interaction with the bulk .
therefore , the lattice structure where rotors are attached is superfluous and there is not a preferential direction for the flux . from another viewpoint ,
it is as if the two reservoirs were placed anywhere .
even if the current through the chain decays in the tl , there is a flow of energy from the hot to the cold reservoir .
in fact , the end rotors are directly coupled to the heat reservoirs and , as a consequence , there is a current from the hot bath to the first rotor , which is the same current from the last rotor to the cold reservoir .
this current is split in several paths : one is the short - circuit given by the long - range coupling between the two end particles and other paths go through each rotor @xmath45 , that is , passing from the first particle through the bulk particle @xmath45 towards the last one .
but , noteworthily , a net current does not pass `` through '' the chain , then @xmath63 . for a finite system , even if the bulk rotors still interact globally , a small current exists , because in that case the effect over each bulk rotor due to the end rotors is not negligible .
consistently with this view , the current @xmath64 decays with @xmath49 .
this scenario which is clear for the m - f case ( @xmath60 ) , apparently also emerges for any @xmath9 below @xmath59 , while , of course , it breaks down for sufficiently short - range interactions .
once the bulk flux due to given applied end temperatures is computed , the heat conductivity @xmath3 can be estimated through @xmath65 for small enough difference between the temperatures applied at the ends , @xmath50 . therefore , the plots in fig .
[ fig : flux ] for @xmath58 vs @xmath49 ( which were computed for fixed @xmath66 and @xmath67 ) , directly reflect the behavior of the thermal conductivity @xmath3 with @xmath49 and @xmath9 .
we can observe in fig .
[ fig : flux ] that , for fixed size @xmath49 , the scaled flux ( hence @xmath3 ) , continuously increases with @xmath9 . therefore , short range interactions favor heat transport . above @xmath68
, @xmath3 practically attains the level of the n - n dynamics for all @xmath49 . for fixed @xmath9 ,
when the range is short enough ( @xmath69 ) , we notice that @xmath3 tends to a finite value in the large size limit , indicating the validity of fourier s law .
differently , below @xmath59 , @xmath3 decreases with @xmath49 , apparently following a power - law decay @xmath70 , where @xmath71 for @xmath60 , and the exponent @xmath72 decreases with @xmath9 , vanishing above @xmath59 .
this suggests that @xmath3 becomes null in the tl for systems with @xmath62 .
however , much larger sizes , which are computationally infeasible , would be required to determine the precise decay law .
vs temperature @xmath2 , for different system sizes and fixed @xmath9 indicated on each panel . in case @xmath73 , we show a power - law fit to the plots of @xmath74 in the high temperature range . in the inset ,
we show the dependency of @xmath3 with @xmath49 at a low temperature ( @xmath75 ) , for different values of @xmath9 . in all cases , @xmath76 .
dotted lines are guides to the eye .
, scaledwidth=50.0% ] we also investigated the dependency of @xmath3 with the mean temperature @xmath2 , as depicted in fig .
[ fig : kappa ] .
let us comment first on the dependency of @xmath3 with @xmath2 for fixed size @xmath49 .
for any @xmath9 , we observe that , the heat conductivity does not depend on @xmath2 at low temperatures , but there is a crossover to a high temperature regime where @xmath3 decays with @xmath2 . in the limit of sufficiently low temperature ,
particles feel the nearly harmonic bottom of the potential well , reaching the limit of harmonic oscillators . at very high temperatures , a decay with @xmath2
is expected , because , the kinetic energy becomes much larger than the potential one ( which is bounded for the @xmath9-xy hamiltonian ) , then , rotors tend to behave as independent particles , and , concomitantly , any transport process tends to become hindered .
next we discuss the impact of @xmath9 on the dependency of @xmath3 with @xmath49 .
let us look at the three leftside panels of fig .
[ fig : kappa ] .
the scenario for any @xmath77 is the same observed for n - n interactions ( @xmath78 ) . at low temperature ,
@xmath3 increases with @xmath49 ( see also the inset , for @xmath75 ) .
the high temperature decay seems to follow a power - law , with the same exponent observed for n - n interactions ( about -3.2 @xcite ) , as illustrated for @xmath73 in fig .
[ fig : kappa ] .
furthermore , at high @xmath2 , the curves for different sizes tend to coincide , consistently with our previous observation of a limiting value of @xmath3 in the large size limit , when we discussed fig .
[ fig : flux ] which was built for @xmath47 .
when the interactions are short - range ( @xmath79 ) , the level of the flat region of @xmath74 , observed for low temperatures below the crossover , grows linearly with @xmath49 ( see inset in the upper left panel ) , as expected for a chain of harmonic oscillators . however ,
notice that , concomitantly with that increase of the flat level with @xmath49 , the crossover temperature diminishes with @xmath49 , suggesting that the curves @xmath74 for different values of @xmath49 tend to adhere to a same curve as @xmath49 increases ( which occurs progressively at lower temperature and larger conductivity ) . if that were the case , the growth of @xmath3 with @xmath49 would persist only at null temperature , but , at a given finite @xmath2 , the conductivity would increase sublinearly with @xmath49 stabilizing at a finite value in the tl .
then , fourier s law would hold .
this possibility is in accord with previous claims @xcite against the divergence of the conductivity in the tl @xcite , for n - n interactions , but much higher sizes would be required to confirm the result for any @xmath80 . at the particular value @xmath81
, we observe that the flat region of the conductivity profile coincides for different values of @xmath49 ( constant @xmath3 in inset plot ) , while for @xmath82 ( illustrated by the case @xmath83 ) , the conductivity decays with system size .
this points that @xmath81 is a marginal case at low temperatures . at very low temperatures
, a particle feels an harmonic potential , regardless the value of @xmath9 .
however , in the case of long - range interactions , the harmonic approximation does not come solely from n - n interactions
. then the level of the flat region does not increase with @xmath49 , but differently to the harmonic chain behavior , @xmath3 decreases with @xmath49 , becoming presumably vanishingly small in the tl .
now consider the rightside panels of fig .
[ fig : kappa ] ( that is , @xmath84 ) , as well as the inset plot . for temperatures above the crossover ,
a distinctive feature in comparison with the leftside panels appears : the conductivity decays with @xmath49 , suggesting vanishingly small values in the tl ( like in fig .
[ fig : flux ] ) .
therefore , for long - range interactions ( @xmath85 ) , @xmath3 decays with @xmath49 for any @xmath2 , in accord with our discussion for the m - f case where the conductivity vanishes in the limit @xmath86 at any finite temperature . of the different thermal regimes of the conductivity @xmath3
: conductivity attains a finite value ( f ) or not , either decreasing ( d ) or increasing ( i ) with @xmath49 , for the investigated range of @xmath49 . from plots of @xmath87 vs @xmath49 , like in fig .
[ fig : flux ] , we obtained the classification indicated by characters for each pair ( @xmath2,@xmath9 ) , with @xmath76 .
the arrow highlights the outcomes for @xmath47 , in agreement with the results shown in fig .
[ fig : flux ] , although @xmath50 is slightly different .
the black solid line represents the critical value @xmath88 .
, scaledwidth=55.0% ] the observed behavior of @xmath3 with @xmath49 , for different values of @xmath2 and @xmath9 , is summarized in the diagram of fig .
[ fig : diagram ] .
we indicate whether , for the studied range of @xmath49 , @xmath3 vs @xmath49 attains a finite value ( f ) or not , and in this latter case whether the behavior is decreasing ( d ) or increasing ( i ) with @xmath49 .
much larger sizes would be required to determine the tl .
this is limited by the computational cost , taking into account that the integration algorithm is of order @xmath89 , and stabilization times also increase with @xmath49 .
moreover , deviations from the diagram shown in fig .
[ fig : diagram ] may occur for a different ratio @xmath90 , therefore this dependency should be also investigated , but again this is limited by the computational capacity . although one would expect a priori a threshold at @xmath91 ( where @xmath92 is the lattice dimensionality ) , the curve @xmath88 that emerges from simulations goes from @xmath93 at low temperatures to @xmath94 at high temperatures .
also in this case , a distortion due to finite sizes can not be discarded .
but , coincidentally , the change of regime occurs near the critical temperature ( @xmath95 ) at which the isolated system in equilibrium suffers a ferro - paramagnetic transition when @xmath16 ( a transition also exists for @xmath96 , although at smaller @xmath97 ) @xcite .
in summary , we have shown the portrait of heat conduction for a one - dimensional system of interacting particles as a function of the range of the interactions .
the different domains are sketched in the diagram of fig .
[ fig : diagram ] . we conclude that , the longer the range of the interactions , more the thermal conduction is spoiled .
an interesting finding is the occurrence of a kind of insulator behavior for @xmath5 ( white region , denoted by ` d ' , in the diagram of fig . [
fig : diagram ] ) . for @xmath98 ,
the thermal behavior is analogous to that found for nearest neighbor interactions .
that is , for high temperatures , the thermal conductivity stabilizes at a finite value ( ` f ' ) in the large size limit , decaying with @xmath2 following a power law . for low temperatures ,
the conductivity increases with @xmath49 ( ` i ' ) .
although , for the studied sizes , we do not observe stabilization of @xmath3 at a finite value , a picture similar to that observed for n - n interactions @xcite emerges , pointing to the validity of fourier s law .
differently , when interaction are sufficiently long - range ( @xmath99 ) , the scaled flux @xmath64 ( hence the conductivity ) decays with @xmath49 for any @xmath2 ( ` d ' ) , apparently vanishing in the tl , like is expected in the limiting mean - field case @xmath60 .
then , the bulk system presents a flat temperature profile and it behaves like an insulator , in the sense that the flux through the chain is vanishingly small .
this poor thermal conductivity is consistent with the slow relaxation to equilibrium observed in long - range systems , like in the hamiltonian mean - field ( @xmath60 ) @xcite or a modified fermi - pasta ulam model @xcite , where collisional effects act over times that increase with @xmath49 , differently to short - range systems .
the impact on thermal conductivity of other recent results about perturbation propagation , where no finite group velocity limits the spreading of perturbations and supersonic propagation occurs @xcite in long - range systems , also deserves investigation .
finally , note that it is plausible that the nontrivial scenario here reported for one - dimension can be extended to an arbitrary dimension @xmath100 , for @xmath101 , which might also deserve a future extension of this work . | we investigate thermal transport along a one - dimensional lattice of classical inertial rotators , with attractive couplings which decrease with distance as @xmath0 ( @xmath1 ) , subject at its ends to brownian heat reservoirs at different temperatures with average value @xmath2 . by means of numerical integration of the equations of motion
, we show the effects of the range of the interactions in the temperature profile and energy transport , and determine the domain of validity of fourier s law in this context .
we find that fourier s law , as signaled by a finite @xmath3 in the thermodynamic limit , holds only for sufficiently short range interactions , with @xmath4 . for @xmath5 ,
a kind of insulator behavior emerges at any @xmath2 . |
sax j1808.4 - 3658 was the first accreting millisecond x - ray pulsar ( amxp ) to be discovered @xcite .
lying some 3.5 kpc distant @xcite , the neutron star spins with a frequency of @xmath1401 hz and is thought to be orbiting a low - mass brown - dwarf companion @xcite with a 2.01 hr orbital period @xcite .
the established picture is that such amxps are recycled to millisecond periods from more slowly rotating neutron stars , due to the transfer of angular momentum from the accretion flow provided by the companion @xcite . in this sense ,
the amxps bridge the gap between low mass x - ray binaries ( lmxb ) and rotation powered millisecond pulsars .
the accretion flow in amxps differs from that in black hole x - ray binaries ( bhxrb ) due to the stellar magnetic field and the solid neutron star surface .
the magnetic field of sax j1808.4 - 3658 is @xmath2 g @xcite and this is sufficient to channel accreting material approaching within a few neutron star radii along field lines towards the magnetic poles .
the funnelled gas passes through an accretion shock before falling on to the magnetic poles .
x - ray observations of amxps made with the _ rossi x - ray timing explorer _ ( rxte ) and _ xmm - newton _ ( xmm ) have discovered distinct spectral components originating from these sources .
the accretion disc and neutron star surface contribute to a soft component up to @xmath110 kev , and comptonisation in the shock region ( most likely of seed photons from the impact hot - spot ) produces a harder component extending to @xmath1100 kev @xcite .
recent _ xmm - newton _ observations of sax j1808.4 - 3658 have shown two distinct blackbodies at soft energies plus a hard component , interpreted as an accretion disc , a neutron star hot - spot and an accretion shock @xcite . in an analogous manner to reflection in bhxrb @xcite , it is thought that comptonised photons reflect off the thin optically thick accretion disc producing a compton reflection hump and fluorescent iron k@xmath3 line emission at @xmath16.4 kev ( see @xcite for a review of broadened iron lines in neutron star lmxbs ) .
the shape of the line provides information about the distance of the line - forming region , since reflection occurring close to the compact object produces a characteristic asymmetric profile due to doppler - broadening and gravitational redshift .
observations of sax j1808.4 - 3658 have revealed evidence suggestive of such a broadened iron line @xcite .
iron k@xmath3 lines are also of particular importance because they place a constraint on the magnetospheric radius .
it is , as yet , unclear how the magnetic field of a magnetised star truncates the disc , or what effect the star s rotation has on the disc inner edge .
the x - ray emission from amxps is pulsed due to misalignment of the spin and magnetic poles . as the hot - spot and magnetically channelled accretion column rotate around the spin axis , the projected area and the magnitude of the doppler shift changes , causing emission to be pulsed at the spin frequency .
it is therefore not unreasonable in this scenario to imagine that changes in the line shape and/or energy with phase might also occur .
for instance , if the accretion column illuminates significantly different parts of the disc with phase , reflection occurring from approaching and receding parts of the disc should display a change in line energy with phase .
exploring the variation of the line properties with phase , as well as the phase dependent correlation between the line equivalent width ( ew ) ( or normalisation ) and the reflection hump , might be a valuable window on the formation of this spectral feature .
the iron k@xmath3 feature in sax j1808.4 - 3658 has been known about for many years @xcite , and we now know that this feature might be relativistically broadened @xcite . therefore , it is of interest whether this feature is also pulsed at some level , which might be expected if the illuminating continuum is also pulsed as seen from the inner disc ( which might be expected in certain geometries , e.g. where a part of the disc is shielded by the neutron star itself ) . using the correct orbital ephemeris and referencing the photon arrival times to the solar system barycenter ,
it is possible to define an arbitrary phase - zero and ` fold ' all of the incoming photons in to phase - bins representative of a full rotation of the neutron star . by examining the energy spectra of photons from individual phase bins
, one can perform phase - resolved - spectroscopy .
a phase - resolved analysis was performed of the 1998 outburst @xcite , and the self - consistent continuum reflection and iron - line were found to be consistent with having a constant value as a function of phase . in this work ,
we produce phase - resolved spectra of the best - sampled and best - studied 2002 outburst of sax j1808.4 - 3658 as observed by _
rxte _ over the ` slow - decay ' of the outburst @xcite where pulse profiles are most stable and we can investigate the phase - resolved spectrum with high signal - to - noise .
we shall explore changes in the iron - line properties with phase , and investigate the viability of the currently accepted picture of the broadened iron - line emission in amxps .
we have concentrated on rxte observations of sax j1808.4 - 3658 from the slow - decay phase ( mjd 52564 - 52574 ) of the 2002 outburst where the pulse profile is the most stable and the data is best sampled ( in date order obs i d 70080 - 01 - 01 - 02 to 70080 - 01 - 03 - 03 inclusive ) .
the reader is referred to figure 3 of @xcite for a light curve of the outburst illustrating the data segment we used . for simplicity ,
only data taken with the e_125us_64m_0_1s event mode of the _ rxte _ proportional counter array ( pca ) were used in this work .
although during these observations different pca detectors were switched on , we have only analysed data from pcu 2 which was always on , in order to ensure consistency .
gain and offset corrections , which attempt to match the energy scales in different pcu detectors , assign zero response in different channels of different pcus @xcite . combining different detectors involves weighting different detector responses , and this introduces unmodelled scatter in to the data . in total , 64 event files were analysed with a total exposure of 167 ksec , which provided sufficient signal to noise for phase - resolved spectroscopy using 16 phase bins . the photon arrival times in each event file were corrected to the solar system barycenter with the ftool faxbary .
the photon phases were determined by using a high - order polynomial to `` whiten '' the phase residuals and increase the signal to noise of the pulsations . additionally , the arrival time of photons was corrected for the orbital motion of the binary using a first - order approximation as outlined , for example , in @xcite .
a folded pulse profile was created for each separate event file , and fitted with a simple constant offset plus a sinusoid . during the slow - decay ,
pulse profiles have been shown to be stable and nearly sinusoidal @xcite . by fitting each event file with the best fitting sinusoidal function
, an offset could be determined for each event file to accurately align the different files in phase for subsequent summation .
custom code was written to select the photons from pcu2 only and produce separate fits files containing events from each of the 16 different phase bins for each event file .
the ftool seextrct was then used to produce pha spectra for each phase bin .
background files were created using the ftool pcabackest , and extracted using saextrct over contemporaneous good time intervals created using the ftool maketime .
the ftool mathpha was used to sum pha files from each respective phase bin and sum the background spectra .
a weighted pcu2 response file was created ( due to any variations in the response over time ) from all observations using pcarsp and addrmf .
care was taken to adjust the exposure of the phase - resolved spectra to reflect the fact that each spectrum only contained 1/16th of the total exposure .
phase averaged pcu2 spectra were also obtained from the same observations . the pre - processed high energy x - ray timing experiment ( hexte ) cluster a spectra from the same observations
were summed together to provide some constraints on the phase - averaged power - law spectral component at higher energies . throughout this work , spectral
fitting was performed using xspec version 12.5.1 @xcite with a 0.5% systematic error for pca and hexte data during fitting .
it should be noted that the pcu 2 data has zero response in channel 10 , but these counts are redistributed to other channels , hence the gap in the data in figure [ fig : ironlineratio ] .
phase - averaged spectra of sax j1808.4 - 3658 from the slow - decay phase of the 2002 outburst have been fitted before @xcite .
it is therefore known that a simple power - law fit shows residuals that are consistent with compton reflection and blackbody hotspot emission . in both the 1998 and 2002 outbursts ,
thermal comptonisation models have typically shown photon indices of @xmath4 with blackbody temperatures around 0.7 kev and hot - spot areas in the range 20 - 110 @xmath5 @xcite . as well as
fitting pca data over 3.0 to 25.0 kev , the hexte spectra are also fitted simultaneously over the range 15.0 - 150 kev to constrain the power - law continuum and cut - off energy .
the continuum was best - modelled using the thermal comptonisation model nthcomp @xcite , and reflection was modelled using pexriv @xcite and gaussian components convolved with rdblur .
only the reflection component of the pexriv model was included . a thermal comptonisation model is chosen since the hot electrons in the accretion shock are thought to compton up - scatter the seed photons from the hot - spot , whilst the blurred gaussian is chosen to model the iron - line feature .
all parameters were tied between the two instruments apart from a normalising constant to allow for a calibration offset .
the hotspot emission is modelled using a bbodyrad component .
the bbodyrad model is a single temperature blackbody spectrum parameterised by the temperature and the normalisation @xmath6 , where @xmath7 is the source radius in km and @xmath8 is the distance to the source in units of 10 kpc .
the nthcomp model is parameterised by the asymptotic power - law photon index , the electron temperature , the seed photon temperature and the normalisation ( the photon flux defined at 1 kev in units of photons @xmath9s@xmath10kev@xmath10 ) .
the pexriv model describes an exponentially cut - off power - law spectrum reflected from ionised material and is parameterised by the power - law photon index , the power - law cut - off energy , the disc temperature , the disc ionization parameter @xcite , the normalisation ( photon flux at 1 kev ) of the power - law only ( without reflection ) and a dimensionless reflection scaling factor .
rdblur is a convolution model taking account of the blurring due to relativistic effects from an accretion disc around a non - rotating black hole , and is parameterised by the power - law index of emissivity , the disc inner radius , the disc outer radius and the inclination .
full parameters of the fit are shown in table [ tab : phaseavfit ] and the unfolded spectrum is included in figure [ fig : eeufspecphav ] . the spectral fits reflect the average parameters over the 167 ksec of data we analysed from the slow - decay phase of the outburst .
the parameters of this ` average ' fit are consistent with previous work , though it should be noted that as the mass accretion rate drops throughout the outburst , the disc inner edge likely recedes @xcite , causing changes in the fit parameters .
the fits were insensitive to @xmath11 , the ionisation parameter , and the disc temperature so these were frozen at values of 100 erg cm @xmath12 and @xmath13 k respectively .
the low value of @xmath11 is consistent with the line energy which was constrained to lie between 6.4 and 6.9 kev and pegged at the lower limit during fitting . in order to avoid the pexriv power - law cut - off energy becoming unphysical ,
this parameter was constrained to be equal to three times the value of the electron temperature . freeing the seed photon temperature for comptonisation from the blackbody temperature
had no statistically significant effect , consistent with the notion that the seed photons for comptonisation come from the hotspot emission .
it should also be noted that the pexriv normalisation and reflection fraction are degenerate parameters in these fits hence the large absolute values of reflection fraction quoted in table [ tab : phaseavfit ] .
the relatively poor chi - squared value of the model quoted in table [ tab : phaseavfit ] is due to the higher energy hexte data . removing the hexte data and fitting only pca data results in a @xmath14/d.o.f of 16.56/18 , but
it was considered important to include the hexte data during fitting to better model the electron cut - off and thermal comptonisation continuum .
grouping the hexte data in to bins of 4 channels each in this fit improves the @xmath14/dof to 38.44/28 , a perfectly acceptable fit given the low number of degrees of freedom .
previous pulse - profile modeling @xcite has derived a lower limit of @xmath15 for the inclination of this source , but since no eclipses or absorption dips are seen , a value of @xmath16 is assumed throughout this work consistent with previous analysis @xcite .
any difference in inclination only introduced a systematic shift in the normalisation of spectral components , and therefore did not effect the relative phases of any pulsations . .
l | r parameter & value + @xmath17(kev ) & @xmath18 + [ 1ex ] @xmath19 & @xmath20 + [ 1ex ] @xmath21 & @xmath22 + [ 1ex ] nthcomp @xmath23 ( kev ) & @xmath24 + [ 1ex ] @xmath25 & @xmath26 + [ 1ex ] folde(kev ) & @xmath27 + [ 1ex ] @xmath28 & @xmath29 + [ 1ex ] @xmath30 & @xmath31 + [ 1ex ] @xmath32 & @xmath33 + [ 1ex ] @xmath34 & @xmath35 + [ 1ex ] @xmath36 @xmath37 & @xmath38 + [ 1ex ] emissivity & @xmath39 + [ 1ex ] @xmath40 ( kev ) & @xmath41 + [ 1ex ] @xmath42 ( kev ) & @xmath43 + [ 1ex ] @xmath44 @xmath45 & @xmath46 + [ 1ex ] @xmath14 / d.o.f & 97.18 / 62 + [ 1ex ] inclination was fixed at @xmath47 , @xmath48 fixed at 1000 @xmath49 , nh fixed at @xmath50 and solar abundances assumed .
upper / lower error limits labelled with an asterisk show that the fit parameters were pegged , at times indicating that the fit was not particularly sensitive to this parameter .
phabs@xmath51(nthcomp+bbodyrad+rdblur@xmath51(pexriv+gaussian ) ) model .
bottom : data to model ratio plot of this fit.,width=317 ] a simple phenomenological fit was performed to the 16-bin phase - resolved data consisting of a phabs*(pow+bbodyrad ) model .
the absorption was fixed at @xmath52 as in @xcite ( obtained from the headas tool _ nh _ ) and the fit performed over the region 3.0 - 25.0 kev whilst ignoring the iron - line region of 5.5 - 10.0 kev .
component normalisations were allowed to be free for each of the 16 different spectra .
having fitted a model excluding the data in the iron - line region , by then including the data in this region one can plot the data - to - model residuals as in figure [ fig : ironlineratio ] .
the plot of the residuals demonstrates a broad iron - line feature and hints at a changing iron - line equivalent width with phase .
.,width=317 ] in the following sections , we shall explore different models to explain this apparent pulsation of the iron - line with phase .
a fit was performed with the model phabs*(cutoffpl+bbodyrad+diskline ) , using the 16 phase - resolved spectra over 3.0 to 25.0 kev and the phase - averaged _ hexte _ spectrum over 15.0 to 150.0 kev .
the iron line energy was constrained to lie within 6.4 and 6.9 kev and the inclination fixed at @xmath47 as in @xcite and @xcite . with the cut - off power - law and
blackbody normalisations free to vary between phase - bins ( untied ) , and the diskline normalisation constrained to its best fitting value across all phase bins ( tied ) , the fit gave a @xmath14/dof of 756.52/467 .
the best fitting diskline @xmath53 was 11.9 @xmath54 and the low value ( 1.63 ) of power - law photon index ( compared to the phase - averaged fits ) points to some degeneracy between the power - law and blackbody contributions at soft energies in this simple fit .
untying the diskline normalisation between phase - bins improved the @xmath14/dof to 694.46/451 and the spectral components demonstrate the pulsations shown in figure [ fig : phenomlineuntied ] . untying
the diskline normalisation simply allows the flux from the line to change with phase , rather than assuming the flux is constant as a function of phase .
if the blackbody temperature is untied between phase - bins instead of the diskline normalisation , the @xmath14/dof improves to 691.38/451 and the spectral components this time pulsate as shown in figure [ fig : phenomktuntied ] . untying the blackbody temperature produces an @xmath55 phase - shift in the pulsation of the blackbody normalisation compared to figure [ fig : phenomlineuntied ] , suggesting that there is some offset between the doppler shift of blackbody photons ( that affects the temperature measured by an observer at infinity and is dependent on the maximum line of sight velocity ) and the projected area of the hot - spot which dominates the blackbody normalisation .
it is also interesting to note that the diskline normalisation and kt are in phase in figures [ fig : phenomlineuntied ] and [ fig : phenomktuntied ] respectively .
again , it should be pointed out that the poor @xmath14/dof is due to the higher energy hexte data used to constrain the parameters in this fit . if a fit is performed excluding the hexte data ( fixing the now unconstrained high energy cut off to the phase - averaged values ) the @xmath14/dof improves to 473.58/407 .
although the fits obtained with this simple phenomenological model are formally unacceptable , they do provide some physical insight in to the nature of the pulsations in this source which are discussed below .
the detailed reflection models explored in section [ reflection ] produce much more acceptable values of reduced @xmath14 .
this preliminary analysis leads to two explanations for the spectral pulsation , which are equally likely on statistical grounds alone . in the first case ,
the line normalisation is pulsating due to solid angle changes of the line emitting region and in the second case the line normalisation itself is constant , but the changing spectral shape of the blackbody component ( governed by temperature changes , possibly due to doppler effects ) introduces a modulation at the spin frequency .
it is important to point out here the physical situation that these alternative models are describing . in the first case , which invokes a varying line normalisation , the visible area of line emission changes with phase
this could , for instance , be due to the neutron star self - shielding the line emission from the disc as the system rotates , as shown in figure [ fig : shielding ] , or due to solid angle changes of the emitting area with phase . in the second scenario , where the line emission is constant with phase ,
the alignment of the spin and magnetic poles are such that the illumination of the disc by the accretion shock is more or less constant with phase as depicted in figure [ fig : discillum ] .
the amount of observed iron - line emission is therefore fairly constant with phase , and the modulation can be explained solely in terms of the spectral changes in the blackbody component occurring in the same energy channels as the iron - line .
having established these two alternative scenarios using the phenomenological fits , the same physically motivated reflection model described in the phase - averaged fits was fitted to the phase - resolved spectra ( and _ hexte _ spectra ) . in the first fit , the normalisations of the bbodyrad , nthcomp , pexriv and gaussian components were untied between phase - bins , whereas the blackbody temperature and seed photon temperature were tied between phase - bins ( and tied to each other ) . throughout these phase - resolved fits , all parameters constrained to not vary with phase remained very close to the phase - averaged values quoted in table [ tab : phaseavfit ] .
the @xmath14/dof for this fit was 455.0/435 and the pulsation of the spectral components with phase can be seen in figure [ fig : fitoneparams ] .
it is clear from figure [ fig : fitoneparams ] that , in this model , the gaussian ( iron - line ) normalisation is systematically varying with phase , whilst the reflection normalisation shows no clear evidence of doing so .
we jointly fitted both the gaussian and pexriv normalisations shown in figure [ fig : fitoneparams ] with a simple constant offset plus a sinusoid , which forced the fractional amplitude and phase of the normalisation variations to be the same .
the fit was poor , but was significantly improved by untying the phase ( @xmath56 for one additional free parameter ) and amplitudes ( @xmath57 for one additional free parameter ) .
the likelihood of an observed improvement by chance was 0.003 and @xmath58 for untying phase and then fractional amplitude respectively .
the fractional amplitude of variation in the pexriv normalisation was close to zero i.e. consistent with a constant pexriv reflection component .
this inconsistency in the reflection components is the most compelling evidence that the `` iron line pulsation '' is actually no more than a change in the underlying blackbody spectral shape with phase .
the reflection and iron - line normalisation should be strongly correlated in amplitude and in phase with one another because they both emerge as a consequence of hard x - ray irradiation of the disc .
in fact , figure [ fig : fitoneparams ] demonstrates no discernible pulsation in the pexriv component at all . untying the blackbody temperature between phase - bins in this fit
did not significantly improve the fit .
a second fit was performed to the phase - resolved spectra using the same model described in the phase - averaged fits , but this time constraining the reflection and line normalisations to be constant with phase .
the @xmath14/dof of this fit was 492.46/467 . allowing the blackbody temperature to also vary with phase ( it was still assumed that the blackbody temperature and seed photon temperature for comptonisation were identical ) improved the @xmath14/dof to 464.07/451 with an f - test probability of 0.039 .
the variations with phase are shown in figure [ fig : fittwoparams ] . again , as with the earlier phenomenological fits , the blackbody normalisation is shifted along in phase when the blackbody / seed photon temperature is a free parameter .
figure [ fig : contour ] shows 99.9@xmath59 confidence contours for three phase bins ( 1,6 and 11 ) of blackbody temperature against blackbody normalisation and demonstrates that the temperature and normalisation are correlated .
however , the fact that the 99.9@xmath59 contours do not overlap significantly in temperature for a given normalisation ( and vice versa ) and there is no anti - correlation in the pulsations of these components in figure [ fig : fittwoparams ] , suggests that the pulsation in these components might be real .
untying the seed photon temperature for thermal comptonisation and the blackbody temperature had no significant effect on the spectral fits .
there is no concrete statistical evidence to favour blackbody temperature variations over varying reflection , but there is a good physical reason to do so as outlined above .
the best fit arises by invoking a changing blackbody temperature with phase and , although the error bars are large , the amplitude and shape of reflection variations do not match the line variations seen in figure [ fig : fitoneparams ] . in the next section
we will show that the observed variations can be explained quite simply in terms of the doppler shifts ( to produce the observed hotspot temperature variation ) and solid angle changes ( to produce the normalisation variation ) which are expected for this system .
confidence contours of blackbody normalisation versus kt corresponding to phase bins 1 , 6 and 11 in figure [ fig : fittwoparams ] . due to the coarseness of the grid used to search parameter space ( a step size of 0.02 kev and 5 for kt and normalisation respectively ) the contours do not join up.,width=317 ] table [ tab : rmsvalues ] shows the fractional rms values for the different model components of each model described above .
| c | c | c | c | c | + & fig 3&fig4&fig7&fig8 + diskline & 0.157 & n / a & n / a & n / a + [ 1ex ] bbody norm & 0.072 & 0.10 & 0.083 & 0.046 + [ 1ex ] power - law & 0.039 & 0.039 & n / a &
n / a + [ 1ex ] kt & n / a & 0.018 & n / a & 0.015 + [ 1ex ] nthcomp norm & n / a & n / a & 0.044 & 0.030 + [ 1ex ] pexriv norm & n / a & n / a & 0.020 & n / a + [ 1ex ] gauss norm & n / a & n / a & 0.155 & n / a + [ 1ex ] fractional rms values are obtained from the best - fitting constant offset plus sinusoid model . quoted value for the pexriv normalisation is an upper limit .
the rms values shown in table [ tab : rmsvalues ] for the physically motivated reflection model invoking blackbody temperature variations ( figure [ fig : fittwoparams ] ) , suggest a larger variation in the blackbody component than the comptonised component .
as shown in figure [ fig : phase ] , this is most likely due to blackbody emission arising from a flatter , more horizontally extended ` pancake ' type region compared to the comptonised emission from the accretion shock which could extend vertically from the neutron star surface . as the star rotates , the projected area of the blackbody region changes more than the accretion shock region , leading to a greater amplitude of variation .
the fractional variation from the comptonised emission might also be reduced if the accretion column is relatively optically thin ( which mitigates the effect of solid angle variation ) , or if there exists a second , constant comptonising region , e.g. an extended corona .
the pulsation in the comptonised component is significantly out of phase with the blackbody component , as seen in previous work @xcite .
however , the phase shift of @xmath1@xmath60 between these components in figure [ fig : fittwoparams ] ( when kt is a free parameter ) , is much larger than in previous work where the shift was @xmath1@xmath61 @xcite and @xmath1@xmath62 @xcite with kt fixed . by fixing the blackbody temperature ,
the offset between the blackbody and comptonised component is found to be @xmath63 , consistent with @xcite .
it would therefore appear that some of the role of changing normalisation is subsumed by the changing kt , shifting the variations in blackbody normalisation to later phases .
it is perhaps surprising how sensitive the phase - shift of the blackbody component pulsations are to the choice of model used .
if the accretion column is optically thick and vertically elongated above the neutron star surface ( as argued above ) , then the solid angle of the column could be maximised when it points away from the observer .
this would coincide with the minimum projected hotspot area and can therefore explain the @xmath1@xmath64 phase shift between the blackbody and comptonised components
. for a 15 km neutron star , the doppler shift @xmath65@xmath66@xmath67@xmath68@xmath69 , produces a large shift at the equator ( e.g. , around 13% at a rotation frequency of 401hz ) . for an accretion shock aligned at @xmath70 to the rotation axis , the upper limit obtained by @xcite
, r is reduced from 15 km at the equator to just 2.6 km and the doppler shift reduced to just 2% .
if we discount the possibility of pulsed iron line variations ( as implied by the much weaker variations in the reflection continuum ) , the phase - resolved fits suggest that some of the pulsed variation in the spectral shape can be explained by a varying doppler shift of the blackbody photons underlying the iron - line .
if we associate the blackbody temperature change with a doppler shift , figure [ fig : fittwoparams ] demonstrates a peak - to - peak change in temperature of approximately 0.035 kev which in this source , assuming an inclination of @xmath16 , translates to a radius of rotation of 3.5 km , consistent with the arguments of @xcite .
this shift is consistent with a hot - spot aligned at approximately 13.5@xmath71 ( 20@xmath71 ) to the rotation axis , for a 15 km ( 10 km ) radius neutron star .
the fact that the shock makes a small angle with the rotation axis of the neutron star implies that the whole inner disc is illuminated more or less evenly i.e. there is no preferential illumination of different parts of the disc with different line of sight velocities which would lead to a change in the iron - line shape with phase . a phase shift measured as @xmath55 can be seen in figure [ fig : fittwoparams ] between the pulsation in the blackbody temperature ( driven by doppler variations ) and the blackbody normalisation ( driven by projected area effects ) .
as dicussed in @xcite , this is consistent with a geometry where doppler variations are strongest when the hot - spot is at the limb of the star , and projected area effects are strongest @xmath0 later when the solid angle of the hot - spot is maximised
. clearly , this picture is only approximate , since we do not take account of effects such as light - bending ( e.g. @xcite ) .
our fits to these data are consistent with the idea that the iron - line shape and normalisation remain constant with phase , suggesting a geometry where the illumination of the disc by the accretion shock is fairly constant . by assuming that the _ only _ variable components are the normalisations of the blackbody and comptonised components @xcite
, one can detect an apparent pulsation in the residual iron - line component .
however , any apparent change in the iron - line normalisation with phase can be explained by taking into account the doppler boosting of the underlying blackbody continuum .
the constancy of the iron line is not really surprising , because for the small offset of the accretion column from the rotation axis that is envisaged here , the disc will not see much variation in solid angle of the accretion - shock region , and variations in doppler beaming of the accretion - shock emission toward the disc will be small . the presence of a constant power - law continuum component from a corona would also serve to reduce the variability of the iron line emission .
phase - resolved spectroscopy of the 2002 outburst of the amxp sax j1808.4 - 3658 as observed by _ rxte
_ reveals no convincing evidence for pulsations in the iron - line or changes of any line properties with phase . by allowing the blackbody spectral shape to change ( i.e. freeing the blackbody temperature during fitting ) one can allow for variable doppler boosting of the underlying blackbody continuum , which introduces a larger phase - shift between the comptonised component normalisation and the blackbody normalisation than has been seen in previous work .
the phase - shift between the blackbody normalisation and the blackbody temperature is approximately consistent with the expected @xmath72 offset between the maximum projected area and maximum doppler shifts .
the fraction of the disc illuminated by the comptonising region appears to be roughly constant with rotational phase .
tw is supported by an stfc postgraduate studentship grant , and pu is supported by an stfc advanced fellowship .
the research leading to these results has received funding from the european community s seventh framework programme ( fp7/2007 - 2013 ) under grant agreement number itn 215215 black hole universe " .
this research has made use of data obtained from the high energy astrophysics science archive ( heasarc ) , provided by nasa s goddard space flight center , and also made use of nasa s astrophysics data system .
the authors would like to thank the anonymous referee for useful comments .
arnaud k. a. , 1996 , aspc , 101 , 17 bhattacharya d. , van den heuvel e. p. j. , 1991 , phr , 203 , 1 bildsten l. , chakrabarty d. , 2001 , apj , 557 , 292 cackett e. m. , et al . , 2009 , arxiv , arxiv:0908.1098 cackett e. m. , altamirano d. , patruno a. , miller j. m. , reynolds m. , linares m. , wijnands r. , 2009 , apj , 694 , l21 chakrabarty d. , morgan e. h. , 1998 , natur , 394 , 346 done c. , mulchaey j. s. , mushotzky r. f. , arnaud k. a. , 1992 , apj , 395 , 275 fabian a. c. , rees m. j. , stella l. , white n. e. , 1989 , mnras , 238 , 729 galloway d. k. , cumming a. , 2006 , apj , 652 , 559 gierliski m. , done c. , barret d. , 2002 , mnras , 331 , 141 hartman j. m. , et al . , 2008 , apj , 675 , 1468 ibragimov a. , poutanen j. , 2009 , mnras , 1322 jahoda k. , swank j. h. , giles a. b. , stark m. j. , strohmayer t. , zhang w. , morgan e. h. , 1996 , spie , 2808 , 59 magdziarz p. ,
zdziarski a. a. , 1995 , mnras , 273 , 837 papitto a. , menna m. t. , burderi l. , di salvo t. , dantona f. , robba n. r. , 2005 , apj , 621 , l113 papitto a. , di salvo t. , da a. , iaria r. , burderi l. , riggio a. , menna m. t. , robba n. r. , 2009 , a&a , 493 , l39 patruno , a. , rea , n. , altamirano , d. , linares , m. , wijnands , r. , & van der klis , m. , 2009 , mnras , 396 , l51 poutanen j. , svensson r. , 1996 , apj , 470 , 249 poutanen j. , gierliski m. , 2003 , mnras , 343 , 1301 reynolds , c. s. , & nowak , m. a. , 2003 , physrep , 377 , 389 ross r. r. , fabian a. c. , 2005 , mnras , 358 , 211 wijnands , r. , & van der klis , m. , 1998 , nat , 394 , 344 ycki p. t. , done c. , smith d. a. , 1999 , mnras , 309 , 561 | we perform phase - resolved spectroscopy of the accreting millisecond pulsar , sax j1808.4 - 3658 , during the slow - decay phase of the 2002 outburst .
simple phenomenological fits to rxte pca data reveal a pulsation in the iron line at the spin frequency of the neutron star .
however , fitting more complex spectral models reveals a degeneracy between iron - line pulsations and changes in the underlying hotspot blackbody temperature with phase . by comparing with the variations in reflection continuum , which are much weaker than the iron line variations
, we infer that the iron - line is not pulsed .
the observed spectral variations can be explained by variations in blackbody temperature associated with rotational doppler shifts at the neutron star surface . by allowing blackbody temperature to vary in this way
, we also find a larger phase - shift between the pulsations in the comptonised and blackbody components than has been seen in previous work .
the phase - shift between the pulsation in the blackbody temperature and normalisation is consistent with a simple model where the doppler shift is maximised at the limb of the neutron star , @xmath0 prior to maximisation of the hot - spot projected area .
[ firstpage ] stars : neutron - x - rays : binaries - x - rays : individual ( sax j1808.4 - 3658 ) accretion |
superconductivity in intermetallic compounds has become of great interest lately . among the most relevant of these compounds
are the iron ( oxy)pnictides ( @xmath4 up to 55 k ) , the quaternary borocarbides ( @xmath4 up to 23 k ) and binary mgb@xmath5 ( @xmath6 k ) .
other intermetallic superconductors of the types ret@xmath5x@xmath5 and re@xmath0t@xmath1x@xmath1 ( re = rare earth , t = transition metal , x = si , ge , sn ) have also received attention because of their magnetic and superconducting phases . in particular those crystallizing in orthorhombic u@xmath0ni@xmath1si@xmath1 type of structure ( @xmath7 ) ,
of which the ce - based systems have ground states showing kondo effects and their nonmagnetic lanthanum counterparts have superconducting phases .
an example of isostructural intermetallic compounds are ce@xmath0pd@xmath1si@xmath1 and la@xmath0pd@xmath1si@xmath1 .
the former shows quantum criticality due to kondo effect and does not display superconductivity , while the latter becomes superconductive below 2.15 k. here we are concerned with the superconducting state of la@xmath0pd@xmath1si@xmath1 , which has been suggested to be described by a two - band model in a recent work on specific heat and critical fields @xcite .
this conclusion is consistent with recent electronic band - structure calculations that indicate multi - band fermi surfaces in this material @xcite .
two - gap superconductivity would be surprising in this material , since standard @xmath8-wave bcs superconducting states have been found in all other la - t - x compounds .
early dc magnetization and electrical resistivity measurements suggest that la@xmath0pd@xmath1si@xmath1 is a type ii superconductor @xcite .
in previous works it was suggested that impurity phases like lapd@xmath5si@xmath5 or la@xmath5pd@xmath0si@xmath0 may exist in polycrystalline samples of la@xmath0pd@xmath1si@xmath1 .
although it was not possible to index some extra x - ray peaks to a lapd@xmath5si@xmath5 phase in ref .
[ ] , the data clearly indicate the presence of an impurity phase . in another x - ray study @xcite ,
small extra peaks were indexed to la@xmath5pd@xmath0si@xmath0 .
the impurity in la@xmath0pd@xmath1si@xmath1 is not well established , but it seems sure that its presence is only of the order of a few percent or less . here , we intend to shed light on the energy gap structure of la@xmath0pd@xmath1si@xmath1 by measuring the magnetic penetration depth of polycrystalline samples down to 0.02@xmath4 .
the magnetic penetration depth @xmath9 is a direct response of the cooper pairs and is widely considered one of the most powerful probes for the superconducting energy gap structure .
we found that the penetration depth of la@xmath0pd@xmath1si@xmath1 may be explained by one - gap ( @xmath8-wave ) superconductivity in the presence of an impurity phase .
the la@xmath0pd@xmath1si@xmath1 sample ( @xmath10 k ) measured in this work was the same used in ref .
penetration depth measurements were carried out utilizing a 13.5 mhz tunnel diode oscillator @xcite .
the magnitude of the ac probing field was estimated to be less than 5 moe , and the dc field at the sample was reduced to around 1 moe .
the deviation of the penetration depth from the lowest measured temperature , @xmath11 , was obtained up to @xmath12 from the change in the measured resonance frequency @xmath13 : @xmath14 .
here @xmath15 is a constant that depends on the sample and coil geometries and that includes the demagnetizing factor of the sample . to within this calibration factor ,
@xmath16 are raw data .
we estimated @xmath15 by measuring a sample of known behaviour and of the same dimensions as the test sample .
the main panel of figure [ figlambda](a ) shows the normalized deviation from its zero - temperature value of the penetration depth of la@xmath0pd@xmath1si@xmath1 and pure indium ( 6n ) polycrystalline samples below 0.8@xmath4 .
the inset of figure [ figlambda](a ) depicts the same data in the entire superconducting region .
a small ( about 0.6@xmath17 of the total signal ) anomaly around 1 k is clearly observed in la@xmath0pd@xmath1si@xmath1 sample . since the indium data were taken under identical conditions and behave exactly as expected , the anomalous behaviour must comes from the la@xmath0pd@xmath1si@xmath1 sample . in ref .
[ ] it was suggested that la@xmath0pd@xmath1si@xmath1 may be described in terms of two - gap superconductivity .
however , a two - gap model does not follow the low - temperature penetration depth data of this compound , as can be seen in figure [ figlambda](b ) .
the experimental data are instead accounted for by a more general two - gap two - critical - temperature model , which in general suggest ( a ) a single phase with two independent superconducting energy bands or ( b ) two independent superconducting phases .
the first scenario is an exotic situation never seen thus far , whereas the second one points out to the existence of a superconducting impurity in a sample .
an example of the occurrence of the latter is mo@xmath0al@xmath5c @xcite .
we believe scenario ( b ) is the most likely here in view of the proven existence of an impurity phase in la@xmath0pd@xmath1si@xmath1 @xcite
. the small size of the anomaly is consistent with the very large superconducting volume assigned to the la@xmath0pd@xmath1si@xmath1 phase @xcite .
recent x - ray studies have discussed the possible presence of la@xmath5pd@xmath0si@xmath0 @xcite or lapd@xmath5si@xmath5 @xcite in la@xmath0pd@xmath1si@xmath1 samples . to the best of our knowledge , in the la - pd - si system , apart from la@xmath0pd@xmath1si@xmath1 , only three other phases are superconductive : lapdsi@xmath0 ( @xmath18 k ) @xcite , lapd@xmath5si@xmath5 ( @xmath19 k ) @xcite , and la@xmath0pd@xmath20si ( @xmath21 k ) @xcite .
interestingly , a careful analysis of the experimental data of la@xmath0pd@xmath20si in ref .
[ ] indicates that in this compound both the zero resistivity and the peak of the specific heat jump occur around 1 k. these latter parameters are more closely related to the onset of the superconducting transition in susceptibility measurements .
hence , the anomaly around 1 k in the penetration depth of la@xmath0pd@xmath1si@xmath1 may have been caused by small segregations of la@xmath0pd@xmath20si .
a re - analysis of the la@xmath0pd@xmath1si@xmath1 x - ray spectrum did not confirm the presence of la@xmath0pd@xmath20si @xcite .
however , the amount or volume fraction of the impurity in la@xmath0pd@xmath1si@xmath1 is of the order of 1@xmath17 or less , as detected here and in ref .
[ ] , and could be too small to allow a detection by x - ray diffraction .
similarly , the resolution of the specific heat probe may not be enough to detect a tiny superconducting jump originated by a small amount of impurity , but be sufficient to observe a deviation from a pure s - wave behaviour .
the present analysis suggests that la@xmath0pd@xmath1si@xmath1 is an @xmath8-wave superconductor .
an important result here is that the penetration depth of la@xmath0pd@xmath1si@xmath1 flattens out in the low - temperature region , below 0.2@xmath4 ( main panel of figure [ figlambda](a ) ) .
this unambiguously indicates , regardless of the origin of the second drop , a _ nodeless _ energy gap structure in la@xmath0pd@xmath1si@xmath1 .
moreover , it also implies that the superconducting impurity , say la@xmath0pd@xmath20si , has a nodeless energy gap too .
next , we discuss our data in more detail by means of the superfluid density .
we performed numerical calculations of the superfluid density for @xmath8-wave , two - gap , and two - phase models and compared them with the experimental data .
for this purpose , we used the zero - temperature penetration depth @xmath22 nm , as determined in ref .
[ ] and can be closely estimated from the london penetration depth @xmath23 nm and @xmath24 given in ref .
[ ] . with a ginzburg - landau parameter @xmath25 @xcite , la@xmath0pd@xmath1si@xmath1 is a local superconductor .
for such a superconductor the normalised superfluid density @xmath26 , where @xmath27 is the total density , is given by @xmath28 \,.\ ] ] here , @xmath29 is the contribution to the total superfluid density of @xmath30 band ( phase ) and is related to the density of states on the corresponding fermi surface .
@xmath31 is the fermi function .
the total energy @xmath32 , and @xmath33 is the single - particle energy measured from the fermi surface .
we use here the standard weak - coupling gap interpolation formula @xmath34 , where @xmath35 is the zero - temperature energy gap and @xmath36 is a constant related to the specific heat jump at the superconducting transition . for the local bcs model we used the standard bcs parameters . for the two - gap model
we consider that the gaps are somehow coupled through interband interactions and the expression of @xmath37 given above differs for the two gaps only in the values of @xmath35 . for this model eq .
( [ supdens ] ) reduces for two bands @xmath38 and @xmath8 to @xmath39 , where @xmath40 is the relative contribution of @xmath41 band to the total superfluid density .
for the two - phase model a second gap opens when the superconducting transition of the impurity phase takes place and in the expression for the gap interpolation formula , @xmath35 , @xmath36 and @xmath4 are different for each phase .
figure [ fig:2 ] shows the superfluid density data and the results of the numerical calculations . for comparison ,
figure [ fig:2](a ) also depicts the excellent agreement of the well established indium data and the simulation of the nonlocal bcs model ( indium is a nonlocal superconductor ) , just to confirm the performance of the experimental system .
consistent with the penetration depth data , the superfluid density of la@xmath0pd@xmath1si@xmath1 is temperature independent below 0.4 k. also , the anomaly around 1 k in the penetration depth is displayed as a smooth bend in the superfluid density . in figure [ fig:2](a )
we compare the experimental data with the simulations of local bcs and two - gap models .
the local bcs is in clear disagreement with the experimental data . for the best adjusting parameters
the two - gap model does not follow the data quite well either , as opposed to what was suggested from the specific heat measurements @xcite .
on the other hand , for the two - phase model the agreement with the data is remarkable in the whole temperature region , as seen in figure [ fig:2](b ) . in this case , @xmath42 is the normalized density of states of la@xmath0pd@xmath1si@xmath1 .
the best values for @xmath43 and @xmath36 are given in table [ table ] .
although in la@xmath0pd@xmath1si@xmath1 the value of @xmath43 is the one expected in the weak - coupling approximation , the parameter @xmath36 is much higher than expected in this approximation .
the impurity phase would also be a conventional @xmath8-wave superconductor but in the strong - coupling limit , as indicated by the high values of @xmath43 and @xmath36 . in ref .
[ ] la@xmath0pd@xmath20si was suggested to be a weak - coupling superconductor .
.[table ] adjusting values of the model based on two superconducting phases . [ cols="^,^,^ " , ] our results suggest that la@xmath0pd@xmath1si@xmath1 is a conventional @xmath8-wave superconductors as all la - based intermetallic compounds , leaving exotic superconductivity to the ce - based materials .
in summary , we performed magnetic penetration depth measurements in polycrystalline samples of la@xmath0pd@xmath1si@xmath1 . at the lowest temperatures of our measurements the superfluid density is temperature independent , indicating that the gap structure of la@xmath0pd@xmath1si@xmath1 is nodeless .
we found a second diamagnetic drop in the penetration depth that displays a smooth kink in the superfluid density .
we suggest that the second drop maybe due to a superconducting impurity ( possibly la@xmath0pd@xmath20si ) , and not necessarily a manifestation of two - band superconductivity .
we appreciate conversations with c. rojas and g. gonzlez .
we also thank a. argotte of the electron microscopy facility at instituto venezolano de investigaciones cientficas ( ivic ) .
this work was supported by ivic project no .
441 and by austrian fwf grant no . | we measured the temperature dependence of the magnetic penetration depth of la@xmath0pd@xmath1si@xmath1 down to @xmath2 .
we observe a temperature - independent behaviour below @xmath3 , which is a firm evidence for a nodeless superconducting gap in this material .
the data display a very small anomaly around 1 k which we attribute to the possible presence of a superconducting impurity phase .
the superfluid density is well described by a two - phase model , considering la@xmath0pd@xmath1si@xmath1 and the impurity phase .
the present analysis suggests that the superconducting energy gap of la@xmath0pd@xmath1si@xmath1 is isotropic , as expected for conventional bcs superconductors . |
the nuclear level scheme of the thorium-229 isotope is expected to feature a long - lived isomer state , @xmath3th , extremely close to the nuclear ground state .
the most recent value for the isomer energy @xmath4 , @xmath0ev , obtained from indirect measurements with a nasa x - ray microcalorimeter - rays . ] @xcite , is within the reach of modern optical laser spectroscopy and could serve as a `` nuclear frequency standard '' @xcite .
this standard could reach an uncertainty level of @xmath5 @xcite , and provide a new powerful instrument for testing the stability of fundamental constants @xcite .
it has been shown that an ensemble of thorium nuclei doped into a transparent crystal may demonstrate superradiance with a non - trivial emission dynamics @xcite , and may be used for building an ultraviolet ( uv ) laser @xcite .
finally , the frequency shifts and broadenings produced by such a crystal lattice environment might be used in studies of material properties , as is commonly done in mssbauer spectroscopy @xcite .
the necessary step towards all of these exciting applications is a direct observation and precise determination of the isomer state energy .
the existence of the low - energy state in the @xmath6th nucleus was first conjectured by kroger and reich based on studies of the @xmath1-ray spectrum following the @xmath2-decay of uranium-233 @xcite .
they concluded that this nucleus has a @xmath7 isomer level lying within @xmath8ev above the @xmath9 ground state level .
the development of high quality germanium detectors ( resolution from 300 to 900ev ) allowed helmer and reich to measure more precise @xmath1-energies in 1989 1993 and to predict the energy of the nuclear transition to be @xmath10ev , placing it into the range of optical frequencies @xcite . the decay pattern and combinations of transitions used by helmer and reich
are presented in figure [ fig : f1 ] ( a ) .
this unnaturally low value of @xmath4 triggered a multitude of investigations , both theoretical and experimental , trying to determine the transition energy precisely , and to specify other properties of the @xmath7 excited state ( such as lifetime and magnetic moment ) .
however , searches for direct photon emission from the low - lying excited state performed in the late 90 s @xcite have failed to observe a signal @xcite . in 2005 ,
guimraes - filho and helene re - analysed the data of helmer and reich , taking into account new information about the nuclear decay pattern @xcite .
they derived @xmath11ev . in 2007 , a cryogenic nasa x - ray microcalorimeter with instrumental resolution @xmath12 from 26 to 30ev ( fwhm ) allowed beck et al .
@xcite to perform a new indirect measurement of @xmath4 , involving lower energy nuclear states , as depicted in figure [ fig : f1 ] ( b ) . in this measurement ,
the obtained transition energy ( 7ev ) was corrected by accounting for the theoretical branching ratios @xmath13 estimated as 1/13 , and @xmath14 estimated as 2% in @xcite .
this correction yields the currently most accepted value @xmath15ev , now placing the transition into the vacuum uv range ( @xmath16 nm ) . in the experiments decribed above ( @xcite ) the isomer transition energy @xmath4
is not measured directly but is derived from the spectrum of higher - energy ( kev ) @xmath1-radiation of a spontaneously decaying @xmath17u source .
we will refer to these measurements as _
indirect passive_. possible alternatives are _ direct passive _ and _ active _ approaches . in the _ direct passive _ schemes ( figure [ fig : f1 ] ( d ) )
, the aim is to perform spectroscopy of the ultraviolet radiation emitted from the isomer appearing in the @xmath2-decay of @xmath17u ( 2% of the nuclei decay is expected to lead into the isomer state ) .
this method has two main difficulties : a relatively high false count rate caused by the uranium sample radioactivity , and a high probability of non - radiative decay ( quenching ) of the isomer state in neutral thorium atoms ( up to @xmath18 times higher than the radiative decay rate @xcite ) . to overcome these problems ,
it was proposed in @xcite to extract @xmath2-recoil thorium ions ejected from an uranium sample , and collect them in a small spot on a mgf@xmath19 coated surface to minimize the quenching rate .
vacuum ultraviolet spectroscopy of the emitted fluorescence radiation may then allow to measure the isomer transition energy . on the contrary , in _ active approaches _ , thorium nuclei ( in the ground state )
will be illuminated by tunable radiation to excite them to the isomer state . in the _ solid - state approach _ a macroscopic ( @xmath20 ) number of thorium ions doped into uv transparent crystals
can be excited , for example , by synchrotron radiation , and the emerging fluorescence signal can be studied @xcite
. apparent advantage of this approach is the huge number of simultaneously excited nuclei . at the same time
, crystal fluorescence can cause difficulties in identifying the thorium isomer transition , and various crystal effects can hamper the precise determination of @xmath4 .
another approach is the _ spectroscopy of trapped thorium ions_. at ptb , germany , work is under way to excite nuclei of th@xmath21 ions into the isomer state using a two - photon scheme , exploiting the electronic bridge mechanism @xcite . in georgia university of technology ,
usa , the laser manipulation of th@xmath22 ions is under investigation @xcite .
detection of the excitation of the thorium into the isomer state may be based on a change of the electronic hyperfine structure @xcite .
studies of the hyperfine structure of thorium are also performed at the igisol facility in jyvskyl , finland , in collaboration with a group of the university of mainz , germany @xcite .
we should also mention a number of studies aimed to measure the lifetime of the isomer state without a determination of @xmath4 . in @xcite ,
the half life of the isomer state for a _ bare nucleus _ was derived theoretically based on the calculations of the matrix element of the nuclear magnetic moment and on the experimental data concerning transitions at higher energies .
they predict a half - life of @xmath23 for the isomer transition , where @xmath24 is given in ev , which yields @xmath25min for @xmath26ev .
direct measurements of this lifetime were performed in several groups @xcite .
the obtained results vary from @xmath27 min @xcite to 6 hours @xcite .
this discrepancy may be explained either by an incorrect interpretation of the observations @xcite or by a difference in chemical composition of the thorium resulting in different internal conversion rates .
all active approaches and eventually all nuclear spectroscopy applications require irradiation of the sample with some external narrow - band tunable radiation , and study of the emerging fluorescence .
the estimation of the error @xmath28ev on the isomer energy presented in @xcite corresponds to one standard deviation , therefore it is necessary to scan the excitation source over 2ev ( @xmath29 ) to find the transition with 95% probability , or over 3ev ( @xmath30 ) to find the transition with 99.7% probability .
sakharov re - estimated the influence of the uncertainty of the 29.39kev peak on the isomer energy derivation in @xcite and obtained an error of 1.31.5ev @xcite .
moreover , an analysis of more recent experimental data led him to claim that the energy of the isomer state can be anywhere in the range 015ev , if the isomer state exists at all .
we believe it will be technically difficult , if not impossible , to cover such a broad energy range with a tunable narrow - band source of ultraviolet radiation in a reasonable time . we therefore propose to first increase the energy resolution on @xmath4 by an improved indirect measurement compared to @xcite .
as we show below , it appears possible to resolve the 29.18kev doublet @xcite presented in figure [ fig : f1 ] ( c ) with todays state - of - the - art x - ray spectrometers .
resolving this doublet would significantly increase confidence in the existence of the isomer state .
moreover , the isomer energy would be measured without additional theoretical input parameters like branching ratios etc .
the aim of the present study is to investigate the possibility of resolving the 29.18kev line clearly as a doublet over a broad range of values for the isomer energy splitting and the branching ratio , and to analyze the precision that can be obtained on @xmath4 depending on the relevant experimental parameters .
the operation principle of high - resolution x - ray microcalorimeters is to detect the heat deposited by an x- or @xmath1-ray interacting with an absorber , using a very sensitive thermometer .
interaction with the absorber material mainly proceeds through the photoeffect .
the energy of the produced photoelectron as well as auger electrons together with their thermalization cascade should be effectively deposited within the volume of the absorber @xcite . on the other hand
, the absorber should have a small heat capacity @xmath31 for good instrumental energy resolution @xmath12 .
various microcalorimeters differ in geometry , absorber material , sensor , etc . , which leads to different energy resolutions , stopping powers , total detector surfaces etc .
many of these parameters are connected and can not be optimized independently .
for example , increasing the size of absorber increases the solid angle and/or stopping power but degrades the instrumental energy resolution . finally , we note that after a detection event , dissipation of the deposited heat leads to a detector - specific dead time , during which the energy of a successive photon can not be measured correctly .
therefore it is impossible to improve the precision of the measurements infinitely simply by using a more active sample , or by placing the sample very close to the detector .
the total count rate can be reduced using a designed filter which will primarily absorb photons outside the 29.18kev region of interest .
the present study aims to answer two questions : how does the possibility to resolve the 29.18kev peak as a doublet depend on the experimental parameters , and how to attain the most precise determination of the isomer transition energy @xmath4 .
as outlined above , parameters of the experimental setup can be controlled to some extent only . for a proper design of suitable detectors and experimental configurations ,
we analyze how the key parameters affect these two points . the model employed for the statistical study should not contain too many parameters to make it accessible to a multi - factor analysis . on the other hand , it should be sufficiently comprehensive for a realistic feasibility study . for the sake of convenience ,
we assume a fixed total measurement time of @xmath32s , approximately 11 days , which corresponds to the total time of the successfull measurement in @xcite .
also we suppose that the background count rate near the 29.18kev doublet is flat and symmetric , and that the monoenergetic line has gaussian shape with full width at half maximum equal to @xmath12 ( see section [ sec : issues ] for a discussion of these approximations ) . the considered total energy interval ( 0 - 70ev ) is subdivided to a set of 0.4ev bins ( approximately a factor 10 below the expected instrumental energy resolution ) .
the number of counts in the @xmath33th energy bin is a poissonian random number @xmath34 with a mean value @xmath35 equal to @xmath36+d\cdot r_{\mathrm{bg}}\cdot t , \label{ee:1 } \end{split}\ ] ] where @xmath37 is the signal count rate , @xmath38 is a specific rate of background counts per 1ev energy interval , @xmath39 , @xmath40 are the centers of lines of the components of the 29.18kev doublet , @xmath41 , @xmath42ev is one energy bin , @xmath43 is the mean energy of the @xmath33th bin , and @xmath44 is the branching ratio .
the set @xmath45 of experimental data can be represented as a position vector @xmath46 in an @xmath47-dimensional `` sample space '' ( @xmath48 ) . for a given sample @xmath46
, we perform a nonlinear regression fit by a vector function @xmath49 with @xmath47 components @xmath50+d\cdot \tilde{r}_{bg}\cdot t. \label{ee:2 } \end{split}\ ] ] this fit has 6 free parameters @xmath51 . for the estimation of these parameters
, we use the maximum likelihood method .
the likelihood function @xmath52 is the probability for realizing the set @xmath46 , if true mean values @xmath35 are equal to @xmath53 .
we also introduce the logarithmic likelihood function @xmath54 .
\label{ee:3}\ ] ] as it was outlined above , resolving the 29.18kev line as a doublet would significantly reduce the doubts @xcite in the existence of the isomer state in the @xmath6th nucleus .
our first aim is to discuss the feasability of such an identification depending on experimental parametes and the ( yet unknown ) values of the isomer energy splitting @xmath55 and the branching ratio @xmath44 .
+ + + + to check whether a given set of experimental data corresponds to a single line or to a doublet , one can apply the likelihood ratio test @xcite .
the essence of this test and the method to estimate the significance level is described in appendix a. here we define the significance level as the probability to identify incorrectly the single peak as a doublet or the doublet as a single peak using the likelihood ratio test in the situation when we have either a single peak or a doublet whose parameters are specified .
figures [ ff:2 ] , [ ff:3 ] show level curves for the signal count rates @xmath37 which are necessary to attain a significance level of 1% for various values of the instrumental resolution @xmath12 and the @xmath56 ratio , in the @xmath57 plane .
it is interesting to note that the optimal energy resolution is attained when @xmath58 and not for equally strong components of the doublet .
this is explained by the fact that such a branching ratio leads to a noticeable asymmetry of the peak which facilitates the identification of a second component , increasing @xmath44 leads to a reduction of the main peak at fixed total signal count rate @xmath37 .
the aim of the proposed spectroscopy study is not only to resolve the 29.18kev line in the @xmath1-spectrum of @xmath17u as a doublet but to determine the energy splitting with maximum precision .
we study the standard deviation @xmath59 of the isomer energy as a characteristic measure of precision ( in the following we call @xmath60 the _ uncertainty of @xmath4 _ ) . here and below ,
angular brackets denote expectation values , @xmath4 and @xmath61 denote `` true '' and measured values of the isomer transition energy respectively . for the sake of brevity , we suppose that the true values of the energy splitting @xmath4 and of the branching ratio @xmath44 are equal to 7.8ev @xcite and 1/14 @xcite respectively has 8% error , therefore the value @xmath62 can be considered as a consistent but slightly pessimistic ( from the point of view of spectral resolution ) estimation . ] . as before
, we assume a total measurement time @xmath32s . scaling the results to other values of @xmath4 ,
@xmath44 , or @xmath63 is straightforward . to investigate the dependence of @xmath60 on the instrumental energy resolution @xmath12 , the signal count rate @xmath37 , and the specific background count rate @xmath38
, we perform a monte - carlo study of @xmath60 .
for any set of parameters , we simulate the sample @xmath46 as shown in figure [ fig : f1b ] , and estimate the parameters @xmath64 maximizing the sum ( [ ee:3 ] ) .
we repeat this procedure @xmath65 times and calculate @xmath60 according to ( [ ee:4 ] ) .
+ + in figure [ fig : f4 ] we present curves of constant level of @xmath60 in the plane ( @xmath66 , @xmath38 ) obtained from the monte - carlo simulation .
finally , in figure [ fig : f5 ] we present the curves of constant @xmath60 in the plane @xmath67 for fixed ratios @xmath56 .
one can see that improving the instrumental resolution by 1ev increases the precision on the determination of @xmath60 by the same amount as doubling the 29.18kev signal count rate , or doubling the measurement time @xmath63 .
in this section we estimate the attainable count rates and the resulting precision in a measurement of the isomer energy @xmath4 that can be achieved with a state - of - the - art high - resolution microcalorimeter .
in particular we consider the metallic magnetic microcalorimeter maxs-20 as described in ref .
the aim is to demonstrate , that valuable results can be obtained with currently available technology . in section [ sec : developments ] we describe ongoing work towards a more refined , dedicated detector setup .
the maxs-20 microcalorimeter consists of 8 detector elements ( pixels ) each of which has an absorber plate for incoming radiation ( @xmath68@xmath69m@xmath70 au plate ) connected to a @xmath71@xmath69m@xmath70 paramagnetic temperature sensor ( er - doped au ) through 24 gold stems ( 10@xmath69 m diameter and 5@xmath69 m height each ) .
each sensor is connected to a thermal bath , the system is installed in a cryostat operating at a temperature of about 30mk .
energy deposited into the absorber plate heats the paramagnetic sensor and causes a change of its magnetization in an external magnetic field . measuring this change in magnetization using squids , it is possible to determine the amount of absorbed energy and hence the energy of the incoming x- or @xmath1-rays .
note that after the detection of an x- or @xmath1-ray , the individual pixel can only detect again after a certain relaxation time of the order of 100 ms
. therefore the total count rate @xmath72 should not be too high . as a sample ,
we consider 1mci of @xmath17u electrodeposited as a film onto a metal planchet with a radius of @xmath73 mm .
we assume the sample to be situated 40 mm from the detector ( outside the cryostat ) with a total surface @xmath74mm@xmath75 .
also we suppose the presence of additional material related to the cryostat vacuum system ( sealing , input window of the cryostat , some other intentionally positioned shielding etc . ) , which we refer to as _
filters_. according to the nudat 2.6 database @xcite , each single decay of @xmath17u is accompanied ( on average ) by one @xmath2 particle with an energy from 4.309 to 4.824mev , 0.213 conversion electrons with energies from 2.3 to 600kev ( 97.7% of the electrons have energies below 50kev ) , and 0.0544 photons most of which ( 0.052 per decay ) are @xmath76-shell x - rays with a mean energy of 13kev .
as no individual @xmath76 x - rays are listed in nudat 2.6 ( only average energy and total intensity ) , we have taken the lacking data from the x - ray data booklet @xcite . to estimate the detector count rates ,
we suppose that all @xmath2 particles and electrons emitted from the uranium sample are stopped by the sample itself or by the filter materials , therefore , we consider only x- and @xmath1-rays . also it is supposed that all secondary electrons and photons generated in the filters are absorbed in the material locally .
this assumption is correct for relatively thick filters made from light materials like aluminium .
we take into account absorption of the photons within the sample itself , the filters , and the detector .
the total count rate @xmath77 is : @xmath78 here the sum is taken over all photon energies @xmath79 , @xmath80 is the activity of the uranium sample , @xmath81 is the relative intensity ( quantum output ) of photons with the energy @xmath43 per single decay event , @xmath82 and @xmath83 are the thicknessess of the gold absorber and filters respectively
. linear absorption coefficients @xmath84 ( @xmath85 ) were taken from the xcom photon cross sections database @xcite .
the count rate @xmath37 of signal photons is : @xmath86 , \label{ee:6 } \end{split}\ ] ] where @xmath87 corresponds to the two components of the doublet , @xmath88 is the au @xmath76 shell fluorescence yield @xcite , @xmath89 is the probability that an energy of a fluorescence photon emitted by a au atom is equal to @xmath90 , and @xmath91 is a probability that an incoming @xmath1-quant with energy @xmath43 will be absorbed , and an @xmath92-ray photon following this absorption leaves the absorber ( escape line ) .
supposing an isotropic spatial distribution of these secondary photons , we obtain @xmath93\sin \theta d\theta\right .
\\ & \left.+\int\limits_{\pi/2}^{\pi } \exp\left [ \frac{a_x x}{\cos\theta}\right ] \sin \theta d\theta \right ) dx , \label{ee:7 } \end{split}\ ] ] where @xmath94 , @xmath95 , @xmath96 .
also we suppose that the @xmath92-ray photon is emitted from the @xmath76 shell , i.e. the deepest shell that is accessible by energy conservation , and the probability @xmath89 for emission of the photon is the relative intensity tabulated in @xcite normalized to the sum of relative intensities from the @xmath76 shell . evaluating the expressions
( [ ee:5 ] ) ( [ ee:7 ] ) yields a total detector count rate @xmath97hz and a signal count rate of @xmath98mhz for a 1.3 mm thick aluminium filter . without any filter ,
the count rates for the same parameters are : @xmath99hz , and @xmath100mhz .
we see that the aluminium filter absorbs approximately 92% of all photons emitted from the sample , but only about 34% of the signal photons .
we conclude that filtering is an effective method to decrease the total count rate @xmath72 , caused mainly by low - energy thorium @xmath76 shell x - ray .
the background count rate @xmath38 is caused by the escape of some fraction of the dissipated energy of @xmath1-quants absorbed in the detector . in @xcite
, the number of background counts close to the 29.18kev doublet was about 30 40 events per 3ev bin whereas the total number of counts in the 29.18kev peak was about @xmath101 events .
this yields the ratio @xmath102ev .
assuming that a similar ratio can be realized with the maxs-20 detector , we find that the uncertainty @xmath60 on the measured isomer transition energy @xmath4 will be equal to 0.06ev for an instrumental resolution of @xmath103ev , signal count rate @xmath98mhz and total measurement time @xmath32 s. therefore the proposed experiment to determine the isomer energy @xmath4 is expected to be almost one order of magnitude more precise than the results obtained in the previous experiment @xcite . reducing the experimental resolution to 6ev and 9ev yields @xmath104ev and @xmath105ev respectively .
increasing the total measurement time to @xmath106s , 1 month , we can measure the isomer transition energy with an uncertainty @xmath107ev for @xmath103ev , @xmath108ev for @xmath109ev , or @xmath110ev for @xmath111ev .
we are aware of certain simplifications and assumptions in the above analysis . here
we briefly resume some additional issues that could arrise , a detailed discussion of these points is beyond the scope of this work .
first , the shape of signal peaks can deviate from gaussian .
for example , a long low energy tail on the spectral lines may lead to the appearance of a noticeable step in the backgound count rate ( see , for example , figure 2 ( a ) in @xcite ) .
we believe that in the work of ref .
@xcite this effect is caused mainly by the escape of energy from the absorber material , for example in the form of athermal phonons @xcite .
the yield of compton scattering is not sufficient to explain this step , see appendix b. to take this effect into account correctly , we will have to modify our model of the background .
a more difficult situation arises when the escaping energy is relatively small , which would lead to an asymmetry of the line rather than the appearance of a tail . in this case
, it would be useful to study an isolated single peak separated from the doublet of interest but intense enough to give good statistics , and/or to perform an independent study with another @xmath1 source , e.g. @xmath112am .
another issue that may appear is a slow time - dependent fluctuation of the response function caused by an uncontrollable drift of ambient magnetic fields and/or cryostat temperature over the duration of the measurement .
we believe we can suppress such drifts below 10ev by temperature stabilization and mu - metal shielding of the setup .
additionally , we will monitor the position of a series of reference x- and @xmath1-ray lines for a correct tracing of this drift , realizing a time - dependent calibration of the detector .
note that x - ray lines generally have a much broader linewidth than @xmath1 lines @xcite which simplifies the identification .
an auxiliary calibration source , for example @xmath112am , can help to enhance the quality of this callibration .
also we should mention possible interference of the 29.18kev doublet with coincidence and escape lines of x- and @xmath1-rays of @xmath113u and other elements present in the sample .
we plan to study the composition of the sample using `` ordinary '' low - precision @xmath1-spectrometry .
we are currently developing a dedicated new detector for the measurement described above , to some extend interpolating between the maxs-20 ( 0 - 20kev ) and the maxs-200 ( 0 - 200kev ) series @xcite .
it will feature a linear array of magnetic calorimeters , each with an active area of @xmath114@xmath69m@xmath75 .
we will increase the absorber thickness by a factor of 2 to 3 in comparison with 5@xmath69 m in the maxs-20 , leading to a stopping power of about 50% at 30kev .
we will operate the detectors in a dry 3he/4he - dilution refrigerator at about 20 mk .
in this situation , the intrinsic energy resolution of the detector caused by thermal noises of all kinds is expected to be below 2ev ( fwhm ) . according to the calorimetric detection principle of metallic magnetic calorimeters @xcite ,
this resolution is independent of energy as long as the total gain ( including operational temperature and external magnetic fields ) is stable , the dependence of the detector response on the event position in the absorber is negligible , and the photon energy @xmath24 is still small enough to be within the range of linear detector response @xmath115 . also we expect that the minimal time between two correctly measurable counts in a single detector element ( pixel ) will be about 100 ms .
so far , we have achieved resolving powers up to about @xmath116 ( corresponding to 1.6ev ( fwhm ) at 5.9kev ) with our maxs-20 devices ( unpublished ) , being limited by a combination of both , instabilities of the operating temperature and a position dependence .
we believe that we can improve the short - term stability of the total gain and keep the position dependence of absorbtion events small enough to allow resolving powers beyond @xmath65 in the planned experiment . also , the response of the present maxs-20 detector to photon energies below 60kev has a small quadratic deviation from a linear behavior , see figure [ fig:7 ] . at an energy of 30kev
this deviation is only about 3% , which yields a 6% degradation of the intrinsic energy resolution compared to the low - energy signals , i.e. below 2.12ev on an absolute scale .
the sample should ideally consist of isotopically pure @xmath17u to avoid a too high count rate not carrying relevant information and possible interference with the 29.18kev thorium doublet signal .
for this project , we have 560 mg ( about 5mci activity ) raw sample material available ( in oxide powder form ) .
the origin and preparation procedure of this material is unknown , from the @xmath1-spectrum we suspect that originally , @xmath117th has been activated in a high - flux neutron reactor and the @xmath17u has been separated chemically .
a mass spectrum of the raw material , produced by an in - house icp - ms can be seen in figure [ fig:8 ] .
the raw material contains @xmath118% @xmath17u , together with traces of @xmath117u , @xmath119u , @xmath120u , @xmath121u , and the decay product @xmath6th .
further daughter product of the @xmath17u chain have not been detected . to further purify the sample , we will perform a purex uranium extraction procedure .
we have also observed an efficient additional element separation in the electrodeposition process . in - house analysis using @xmath1- and @xmath2-spectroscopy , icp - ms , and neutron activation analysis will allow us to quantify the success of this procedures and finally know the exact composition of the final measurement sample .
the sample will be produced by electroplating @xmath17u from a liquid solution onto a stainless steel or aluminium planchets .
the target activity of 1mci corresponds to 104 mg of pure @xmath17u or 118 mg of uo@xmath19 .
producing correspondingly thick films ( 15 - 20@xmath69 m ) turned out to be difficult in electroplating @xcite .
we have therefore developed a process to deposit up to 20 mg uranium onto stainless steel or aluminium foils of only 10 - 50@xmath69 m thickness .
these samples can easily be stacked to realize the target activity without the carrier foils significantly reducing the count rate in the 29.18kev peak .
we have analyzed the feasability of an indirect measurement of the low - energy isomer state in @xmath6th using a high - resolution magnetic microcalorimeter .
we propose to resolve the 29.18kev doublet in the @xmath1-radiation spectrum following the @xmath2-decay of @xmath17u .
such a measurement would provide a strong indication for the existence of the isomer state and improve the accuracy on the energy measurement significantly .
the measurement appears feasable with currently available detector technology and samples .
we thank c. streli , p. wobrauschek , and j. seres for discussions and calculations concerning @xmath1-ray filters .
we thank s. smolek and m. gollowitzer for preliminary tests and characterizations of @xmath17u samples .
this work was supported by the erc project 258604-nac , the fwf project y481 , the wpi thematic program `` tailored quantum materials '' , and the fwf project m1272-n16 `` theonac '' .
here we describe the essence of likelihood ratio tests for regression models in the simple case of normally distributed observables with known dispersions , and the method which we actually used for the estimation of the significance level attainable in the experiment .
let us have @xmath47 experimental observables @xmath123 normally distributed around their ( unknown ) expectation values @xmath35 with known dispersions @xmath124 . without lost of generality , we can set @xmath125 for all @xmath33 . also , we have 2 regression models , one of which ( _ short model _ ) is a particular case of another one ( _ long model _ ) . in the long model
, it is supposed that the expectation values of observations @xmath123 are some known functions @xmath53 of the @xmath126-dimensional parameter @xmath127 . in the short model , it is supposed also that the parameters @xmath128 has some additional restrictions , and the short model has @xmath129 degrees of freedom .
it is supposed that the long hypothesis anyhow is correct .
the question we want to answer is whether the short hypothesis is correct ? or , more precisely : how plausible ( or unplausible ) is it to obtain the set of observables @xmath123 , if the short model is correct ?
the likelihood ratio test is a powerful method to answer this question . to illustrate the essence of this test ,
let us represent @xmath47 observations @xmath123 as a point @xmath130 in the @xmath47-dimensional euclidean sample space .
@xmath47 functions @xmath131 form an @xmath132-dimensional surface @xmath133 corresponding the long hypothesis . in turn , these functions with additional conditions corresponding the short hypothesis form the @xmath134-dimensional surface @xmath135 within @xmath133 , see figure [ fig : a1 ] .
we suppose that these surfaces are sufficiently smooth . in the case of linear regression models , the functions @xmath53 are linear , and the surfaces @xmath133 and @xmath135 are just hyperplanes . for normally distributed observables
@xmath123 with zero mean and unit dispersion and for some set of parametes @xmath128 , the logarithmic likelihood function is @xmath136 therefore the maximization of @xmath137 is equivalent to a minimization of the distance between the points @xmath138 and @xmath130 in the sample space .
the square of this distance we denote as @xmath139 .
let @xmath140 and @xmath141 , where @xmath142 and @xmath143 are the best likelihood estimations of parameters @xmath128 within the long and short hypotheses respectively .
it is easy to see that @xmath144 which yields @xmath145 . in turn , @xmath146 is a sum of squares of @xmath147 normally distributed random values @xmath148 with unit dispersion and different means @xmath149 .
therefore , @xmath146 is a non - central @xmath150 random value with @xmath147 degrees of freedom and non - centrality parameter @xmath151 . for the sake of brevity
, we denote this random value as @xmath152 .
it is easy to see that @xmath122 is just a distance between the short hypothesis surface @xmath135 and the point @xmath153 corresponing to the true value @xmath154 of parameters @xmath128 , see figure [ fig : a1 ] .
if the short hypothesis is true ( it corresponds the situation when the surface @xmath135 coincided with @xmath155 in figure [ fig : a1 ] ) , then @xmath156 and @xmath146 is just a `` usual '' @xmath157 random value with @xmath147 degrees of freedom . to test whether the short hypothesis is true or
not , one should choose some desirable significance level @xmath2 , and compare the value @xmath158 with some critical value @xmath159 such , that the probability @xmath160 if @xmath161 , the short hypothesis is rejected on significance level @xmath2 , otherwise it is accepted .
the probability to reject the short hypothesis incorrectly is equal to @xmath2 . on the other hand ,
the probability to accept the short hypothesis incorrectly is @xmath162 .
therefore if the non - centrality parameter @xmath163 will be larger than @xmath164 such that @xmath165=\alpha , \label{ee : a4}\ ] ] the probability to accept the short hypothesis falsely will be less than @xmath2 . in our case , the observables @xmath34 are not normal random values with known dispersion but poissonian random values . to estimate the possibility to identify the 29.18kev line as a doublet , we consider@xmath166 as observables and approximate their distribution function by a normal distribution with mean @xmath167 and dispersion @xmath168 .
this approximation is not very precise ( the bias is about 20% for @xmath169 ) , but it seems to be applicable for a coarse estimation .
then , for specific values of @xmath44 , @xmath4 @xmath63 , @xmath12 , @xmath37 , and @xmath38 we calculate the set @xmath170 according ( [ ee:1 ] ) , and fit it by a vector function whose components are given by the square root of ( [ ee:2 ] ) with additional restriction @xmath171 . then we calculate the non - centrality parameter @xmath172 our short hypothesis ( single peak fit ) has 4 degrees of freedom whereas the long hypothesis ( double peak fit ) has 6 degrees of freedom which yields @xmath173 . for significance level @xmath174 , conditions ( [ ee : a3 ] ) and ( [ ee : a4 ] ) are fulfilled at @xmath175 , and @xmath176 .
the values of @xmath37 necessary to attain @xmath177 for given values of the other parameters were calculated numerically and presented in figures [ fig : f4 ] and [ fig : f5 ] .
@xcite the 29.18kev line shows a low - energy tail , causing a noticeable step of about 10 counts per 3ev bin ( see figure 2 ( a ) in @xcite ) .
this step can be caused either by the absorbtion of photons which had 29.18kev energy initially , but lost some fraction of it in the source or filter material , or by the escape of some energy fraction from the absorber .
the first scenario could be explained by the compton `` almost forward '' scattering , when a photon loses tiny parts of its energy , about a few ev .
the klein - nishina differential cross - section of the compton scattering into the elementary solid angle is @xmath178 , \label{eq : a1}\ ] ] where @xmath179 and @xmath180 are the energies of the photon before and after the compton scattering respectively , @xmath181 is the classical electronic radius , and @xmath128 is the scattering angle . using a well - known relation between @xmath180 and @xmath128 , it is easy to express the differential cross - section in units of @xmath180 : @xmath182 .
\label{eq : a2}\ ] ] note that @xmath183ev corresponds to @xmath184 . for such small @xmath128 , ( [ eq : a2 ] )
yields approximately @xmath185 the differential probability for the 29.18kev photon to be scattered into the 1ev range can be coarsely estimated as @xmath186 , \label{eq : a4}\ ] ] where @xmath187 , @xmath34 , and @xmath188 are the nuclear charge number , the atomic density , and the thickness of the uranium ( @xmath189 ) or filter ( @xmath190 ) material respectively . in the experiment @xcite ,
the uranium activity of one 19 mm diameter planchet was about 0.02mci which corresponds to @xmath191@xmath69 m .
it was covered by a titanium foil ( `` filter '' ) with @xmath192@xmath69 m .
an estimation following ( [ eq : a4 ] ) yields @xmath193ev@xmath194 .
the peak corresponding to the absorption of 29.18kev photons has gaussian shape with about @xmath195 events per 3ev bin height in the maximum , and 26ev fwhm , see figure 2(a ) in @xcite .
this corresponds to a total number of counts forming this peak of about @xmath196 .
therefore , compton scattering of 29.18kev photons in the source or filter material produces only 0.053 events per 1ev interval , or 0.16 events per 3ev bin .
this value is significantly lower than the observed step of 10 counts per 3ev bin .
a similar estimation for a @xmath197 m layer of uranium and @xmath198 mm of aluminium filter ( parameters of the planned experiment ) leads to @xmath199ev@xmath194 .
note that in ( [ eq : a4 ] ) we treat all electrons as free .
a more accurate estimation requires the substitution of the _ incoherent scattering function _ instead of @xmath200 but
this function does not exceed @xmath200 , see @xcite for details .
therefore , a more accurate calculations can only decrease the contribution of compton scattering to the step of the background count rate .
00 b. r. beck , j. a. becker , p. beiersdorfer , g. v. brown , k. j. moody , j. b. wilhelmy , f. s. porter , c. a. kilbourne , and r. l. kelley , _ energy splitting of the ground - state doublet in the nucleus @xmath6th_. phys .
lett . * 98 * , 142501 ( 2007 ) .
b. r. beck , c. y. wu , p. beiersdorfer , g. v. brown j. a. becker , k. j. moody , j. b. wilhelmy , f. s. porter , c. a. kilbourne , r. l. kelley , _ improved value for the energy splitting of the ground - state doublet in the nucleus @xmath6th_. 12th international conference on nuclear reaction mechanisms , varenna , italy , llnl - proc-415170 ( 2009 ) . j. c. berengut , v. a. dzuba , v. v. flambaum , and s. g. porsev , _ proposed experimental method to determine @xmath2 sensitivity of splitting between ground and 7.6 ev isomeric states in @xmath6th_. phys .
lett . * 102 * , 210801 ( 2009 ) d. s. richardson , d. m. benton , d. e. evans , j. a. r. griffith , g. tungate , _ ultraviolet photon emission observed in the search for the decay of the @xmath6th isomer_. phys .
lett . * 80 * , 3206 - 3208 ( 1998 ) s. b. utter , p. beiersdorfer , a. barnes , r. w. lougheed , j. r. crespo lpez - urrutia , j. a. becker , m. s. weiss , _ reexamination of the optical gamma ray decay in @xmath6th_. phys . rev . lett . * 82 * , 505 - 508 ( 1999 ) w. g. rellergert , d. demille , r. r. greco , m. p. hehlen , j. r. torgerson , and e. r. hudson , _ constraining the evolution of the fundamental constants with a solid - state optical frequency reference based on the @xmath6th nucleus .
lett . * 104 * 200802 ( 2010 ) g. a. kazakov , a. n. litvinov , v. i. romanenko , l. p. yatsenko , a. v. romanenko , m. schreitl , g. winkler , and t. schumm , _ performance of a @xmath6thorium solid - state nuclear clock .
_ new journal of physics * 14 * , 083019 ( 2012 ) m. p. hehlen , r. r. greco , w. g. rellergert , s. t. sullivan , d. demille , r. a. jackson , e. r. hudson , j. r. torgerson _ optical spectroscopy of an atomic nucleus : progress toward direct observation of the @xmath6th isomer transition
. _ journal of luminescence * 133 * , 91 - 95 ( 2013 ) o. a. herrera - sancho , m. v. okhapkin , k. zimmermann , chr .
tamm , and e. peik , _ two - photon laser excitation of trapped @xmath117th@xmath202 ions via the 402-nm resonance line _ phys .
a * 85 * , 033402 ( 2012 ) e. ruchowska , w. a. pciennik , j. ylicz , h. mach , j. kvasil , a. algora , n. amzal , t. bck , m. g. borge , r. boutami , p. a. butler , j. cederkll , b. cederwall , b. fogelberg , l. m. fraile , h. o. u. fynbo , e. hageb , p. hoff , h. gausemel , a. jungclaus , r. kaczarowski , a. kerek , w. kurcewicz , k. lagergren , e. nacher , b. rubio , a. syntfeld , o. tengblad , a. a. wasilewski , and l. weissman .
_ nuclear structure of @xmath204th_. phys .
c * 73 * , 044326 ( 2006 ) .
t. t. inamura and h. haba _
search for a `` 3.5-ev isomer '' @xmath6th in a hollow - cathode electric discharge .
c * 79 * , 034313 ( 2009 ) .
h. kikunaga , y. kasamutsu , h. haba , t. mitsugashira , m. hara , k. takamiya , t. ohtsuki , a. yokoyama , a. nakanishi , and a. shinohara , _ half - life estimation of the first excited state of @xmath6th by using @xmath2-particle spectrometry _ phys . rev .
c * 80 * , 034315 ( 2009 ) .
x. zhao , y. n. m. de escobar , r. rundberg , e. m. bond , a. moody , d. j. vieira , _ observation of the deexcitation of the @xmath3th nuclear isomer .
lett . * 109 * , 160801 ( 2012 ) .
e. peik , k. zimmerman , _ comment on `` observation of the deexcitation of the @xmath3th nuclear isomer '' .
lett . * 111 * , 018901 ( 2013 ) .
s. l. sakharov _ on the energy of the 3.5-ev level in @xmath6th .
_ physics of atomic nuclei * 73 * , 1 - 8 ( 2010 ) f. s. porter , g. v. brown , k. r. boyce , r. l. kelley , c. a. kilbourne , p. beiersdorfer , h. chen , s. terracol , s. m. kahn , a. e. szymkowiak _ the astro - e2 x - ray spectrometer / ebit microcalorimeter x - ray spectrometer . _ rev .
* 75 * , 3772 - 3774 ( 2004 ) c. pies , s. schfer , s. heuser , s. kempf , a. pabinger , j .-
porst , p. ranitsch , n. foerster , d. hengstler , a. kampktter , t. wolf , l. gastaldo , a. fleischmann , c. enss , _ maxs : microcalorimeter arrays for high - resolution x - ray spectroscopy at gsi / fair .
phys . * 167 * , 269 - 279 ( 2012 ) a. c. thompson , d. t. attwood , e. m. gullikson , m. r. howells , j. b. kortright , a. l. robinson , j. h. underwood kwang - je kim , j. kirz , i. lindau , p. pianetta , h. winick , g. p. williams , j. h. scofield , _ x - ray data booklet_. lawrence berkeley national laboratory . univ . of california ,
berkeley , ca 94720 ( 2009 ) j. h. hubbel , p. n. trehan , n. singh , b. chand , d. metha , m. l. garg , r. r. garg , s. singh , s. puri , _ a review , bibliography and tabulation of @xmath205 , @xmath76 and higher atomic shell x - ray fluorescence yields . _
chem . ref .
data * 23 * , 339 - 364 ( 1994 ) p. a. raboud , j. cl .
dousse , j. hoszowska , and i. savoy , _ @xmath206 to @xmath207 atomic level widths of thorium and uranium as inferred from measurements of l and m x - ray spectra_. phys .
a * 61 * , 012507 ( 1999 ) .
j. h. hubbell , wm .
j. veigele , e. a. briggs , r. t. brown , d. t. cromer , and r. j. howerton , _ atomic form factors , incoherent scattering functions , and photon scattering cross sections_. j. phys .
data * 4 * , 471 - 538 ( 1975 ) | the thorium-229 isotope features a nuclear isomer state with an extremely low energy .
the currently most accepted energy value , @xmath0ev , was obtained from an indirect measurement using a nasa x - ray microcalorimeter with an instrumental resolution 26ev .
we study , how state - of - the - art magnetic metallic microcalorimeters with an energy resolution down to a few ev can be used to measure the isomer energy .
in particular , resolving the 29.18kev doublet in the @xmath1-spectrum following the @xmath2-decay of uranium-233 , corresponding to the decay into the ground and isomer state , allows to measure the isomer transition energy without additional theoretical input parameters , and increase the energy accuracy .
we study the possibility of resolving the 29.18kev line as a doublet and the dependence of the attainable precision of the energy measurement on the signal and background count rates and the instrumental resolution .
thorium-229 , isomer energy , gamma spectroscopy , design of experiment |
ideally , a good cosmological @xmath1-body simulation should resolve in some detail the internal structure of individual galaxies ( on mass scales as low as @xmath2 to @xmath3 and on length scales of a few kiloparsecs or less ) while at the same time representing the growth of density perturbations on the scale of clusters and even superclusters of galaxies .
galaxy morphologies are observed to correlate with their membership in clusters .
the virgo cluster is estimated to be an appreciable , if not dominant , source of tides on the local group 15 to 20 mpc away . and
when studying the nonlinear growth of galaxy - sized perturbations , one should not discount the possible coupling to modes with wavelengths as large as 50 mpc ( @xcite ) . achieving
the required dynamic range is a difficult task given the limitations of present - day computers .
generally , numerical codes have been able to obtain a large dynamic range either in mass or in length , but not in both .
particle - particle ( pp ) methods ( @xcite ) and tree codes ( appel 1981 , 1985 ; @xcite ; @xcite ) have essentially unlimited length resolution , bounded only by the need to soften two - body encounters when modeling a collisionless physical system , but their use for simulations with more than a few tens , respectively hundreds , of thousands of particles is currently restricted to special - purpose hardware ( such as grape chips ) or massively parallel computers ( @xcite ) .
particle - mesh ( pm ) methods ( @xcite , hereafter he ; @xcite ) , by contrast , can integrate the orbits of millions of interacting particles even on a more modest single - processor mainframe or high - end workstation .
the forces are computed by solving poisson s equation on a cartesian grid . barring complications such as a memory latency that increases with problem size , the computational cost per time step is linear in the number of particles and of order @xmath4 in the number of grid cells ( which is typically of the same order as the number of particles ) .
the dynamic range in length , however , is limited by the maximum size of the grid , typically 256 or 512 cells on a side in three dimensions on a current single - processor computer .
parallel processing can naturally push this limit towards higher values . one can circumvent this obstacle by decomposing the inter - particle forces into a long - range part , calculated by the pm technique , and a small - range part which can be evaluated in a number of different ways . for some time
, a popular approach has been to compute the short - range forces by direct summation over all sufficiently close pairs of particles .
this technique , known as the particle - particle , particle - mesh ( p@xmath5 m ) method ( he 8) , was originally applied to plasma simulations where the electrostatic repulsion between charges of the same sign makes it difficult for large density contrasts to develop .
its use with gravitating systems suffers from the tendency for ever more particles to condense into small volumes , causing the cost of the pp summation to become prohibitive as the system evolves .
increasing the total number of particles and resolving smaller scales ( on which density perturbations become nonlinear at earlier times , at least in the presently favored `` bottom - up '' scenario ) both exacerbate the problem .
we are unaware of any p@xmath5 m simulations using more than about @xmath6 particles
. one may be able to raise this limit by using a tree method to compute the short - range forces , as in xu s ( 1995 ) tpm , but in principle the difficulty remains .
most other attempts to enhance the spatial resolution of pm codes rely on the introduction of local grid refinements .
this idea lies at the root of couchman s ( 1991 ) adaptive p@xmath5 m ( ap@xmath5 m ) ( which however still uses direct summation where the number of particles is small enough ) , of numerous hierarchical pm codes ( @xcite ; @xcite ; @xcite ; @xcite , hereafter anc ; @xcite ; @xcite , hereafter jdc ; @xcite ) and of various related approaches ( @xcite ; @xcite ) . as pointed out by various authors , and most recently by suisalu and saar ( 1995b ) ,
the choice between pm , p@xmath5 m , and tree codes is not only a question of the spatial and mass resolution that can be achieved , but also of the behavior of each method with respect to two - body gravitational collisions .
these are to be avoided when the physical system being modeled is essentially collisionless , as is the case for example for cosmological dark matter .
pm codes usually prove to be less collisional , although it should be emphasized that this is at least in part a consequence of their spatial resolution being weaker than their mass resolution , and need not carry over to hierarchical pm methods unless precautions are taken to ensure that the mass granularity remains adequate at all times .
the purpose of this article is to document in some detail our implementation of a dynamically adaptive multiple - mesh code and the tests we performed to validate it .
so far , only a brief report of an earlier stage of development ( @xcite ) and a description of the methodology but not of the tests ( @xcite ) have appeared .
our code , like that of jdc , is able to track the formation and subsequent motion of density concentrations by dynamically adjusting the number , nesting depth and location of the subgrids . in our case ,
the entire grid structure is chosen afresh on every step . since our first intended application was a cosmological problem the formation of the local group that requires the freedom to specify external tidal fields other than would be implied by periodic image charges , our code presents the relatively uncommon combination of an expanding system of coordinates with isolating ( more properly non - periodic ) boundary conditions .
this introduces a number of complications not usually discussed in the literature on pm codes , and of which we shall give an account here .
( our method is also applicable to non - cosmological problems , for which isolating boundary conditions are an asset and the aforementioned complications do not arise . )
a future article ( @xcite ) will detail the results of our simulations of group formation .
the general outline of this paper is as follows .
section [ s : method ] describes in detail the method used .
the tests and their results are presented in section [ s : tests ] .
we conclude , in section [ s : conclusions ] , with our assessment of the strengths and limitations of this code .
our code integrates the equations of motion for a set of gravitating particles .
these equations can be derived from the lagrangian @xmath7 here the indices @xmath8 and @xmath9 span the set of particles in the system , @xmath10 is the mass of particle @xmath8 , @xmath11 its position at time @xmath12 , and @xmath13 describes the law of pairwise interaction between particles
. ideally @xmath14 ( we choose units such that @xmath15 ) but the numerical method forces us to use an approximation which , due to adaptive grid refinement , can depend explicitly on time and on the individual coordinates @xmath16 and @xmath17 , not just on their difference .
we could have included the term proportional to @xmath18 ( where @xmath18 is an arbitrary constant ) within the external potential @xmath19 , but for cosmological applications it is useful to show it explicitly .
it is often convenient to apply a time - dependent rescaling of the coordinate system , with new coordinates @xmath20 defined by @xmath21 .
an equivalent lagrangian is then @xmath22 for notational convenience , we define @xmath23 note that in principle @xmath24 is arbitrary , reflecting our freedom to choose the coordinate system . for cosmological applications
one normally chooses @xmath24 to be the solution of friedman s equation for some cosmological model , in which case @xmath25 is the background density and @xmath26 the hubble constant for that model . later in this paper
we shall impose the additional restriction @xmath27 , which holds for cosmological models in the matter - dominated era
. the second term in the lagrangian ( [ q : lag - x ] ) can be regarded as involving an effective potential @xmath28 which satisfies @xmath29 .
similarly , in the continuum limit and for a coulomb interaction ( @xmath30 ) , the third term satisfies @xmath31 , where the proper density @xmath32 is given by @xmath33 and @xmath34 ( respectively @xmath35 ) denotes differentiation with respect to the first ( respectively the second ) of the two position vectors on which @xmath36 depends . the usual derivation of the equations of motion from the lagrangian ( [ q : lag - x ] ) yields : @xmath37 for later convenience
, we define @xmath38 this allows the equation of motion ( [ q : motion ] ) to be written more compactly as @xmath39 it is also useful to recast the equation in terms of a generalized time coordinate @xmath40 : @xmath41 with @xmath42 a common choice is the so - called conformal time , @xmath43 ; another ( @xcite ) is @xmath44 for some constant @xmath45 .
naturally , any sufficiently differentiable monotonic function of @xmath12 is acceptable , and one is free to construct such a function _ ad hoc _ for each individual simulation .
we avail ourselves of this freedom .
we lay down a rectangular grid of @xmath46 by @xmath47 by @xmath48 nodes with uniform spacings @xmath49 , @xmath50 and @xmath51 .
the values of @xmath46 , @xmath47 and @xmath48 must be acceptable to the fast fourier transform ( fft ) routines used .
( all our tests were done with @xmath52 a power of 2 , and with @xmath53 , as this is the most common situation in practice and the only one we needed for our applications . ) at every step , a density is assigned to each grid point using the cloud - in - cell ( cic ) algorithm .
ffts are then used to solve for the potential on the same grid , and an acceleration is obtained for each particle by differentiating the grid potential , then using cic interpolation .
we use a two - point finite - difference formula to compute the derivatives of the potential at grid nodes .
this choice means that truncation errors in the force law scale with the square of the grid spacing ( he , 5 - 4 ) .
the discrete green s function we use is obtained by sampling @xmath54 at grid points ( and fourier transforming the result ) .
this differs from the more common approach in cosmological pm codes ( @xcite ) of basing the green s function on the seven - point finite difference approximation to the laplacian ; our choice has the merit of guaranteeing the truncation of the interaction law at large separations , as required for a correct implementation of isolated boundary conditions ( @xcite ) . unlike in the p@xmath5 m method , where the grid - based part of the force must possess a good degree of translational invariance to avoid introducing terms in the direct particle - particle contribution that depend explicitly on the position of each pair relative to the grid , we are not compelled to soften the interaction at small scales .
to do so would cost us precious spatial resolution ; already with our choice of green s function the effective force turns out to be softened at separations @xmath55 grid cells .
a consequence of our decision not to soften the force law any further is that the @xmath56-minimization procedure of he 8 - 3 - 3 , which requires the reference force to have no harmonics on scales smaller than the grid spacing , is not appropriate for us .
the choice of the shape of the interaction law at small separations and of the charge assignment and force interpolation scheme is a matter of compromise between accuracy and computational cost .
our choices could undoubtedly be improved upon , but we preferred to concentrate our efforts on the more innovative aspects of our method .
isolated boundary conditions are implemented by conceptually doubling the size of the grid in each dimension , and padding the density array with an appropriate constant value ( normally zero ) outside the principal octant , which contains the particles ( he , 6 - 5 - 4 ) .
if the green s function is truncated so that the interaction falls to zero at separations larger than the system ( before doubling ) , this completely suppresses any interaction between the system and its periodic images .
the computed potential , however , will only match that of an isolated system within the principal octant . since we need to apply a gradient operator to the potential in order to compute the forces
, there is a layer ( one cell thick for our approximation to the gradient ) in which we would not be able to compute them .
we therefore allow no particles in this outer layer .
( we are free to increase the thickness of this layer .
it has been convenient to do so in some of the tests discussed below , to achieve an integral ratio of cell counts in comparisons between different grid resolutions . )
we do not use james ( 1977 ) method to impose isolated boundary conditions without doubling the grid by calculating appropriate screening charges on the surface , since strictly speaking the procedure is only justified when a finite - difference approximation to the laplacian is used in solving poisson s equation .
the fft technique does not satisfy this condition , and would require screening charges throughout the volume of the box . in cosmological applications ,
the space surrounding the system is not to be thought of as empty , but rather as containing a background of uniform density @xmath57 .
( departures from uniformity can be represented through an appropriate choice for the tidal potential @xmath19 . )
we must therefore add this background , convolved with the same ( cic ) charge assignment function that is used for the particles .
its contribution to the density at any node on the boundary of the region where particles are allowed is then proportional to the number of cells adjacent to this node that lie outside the particle region : @xmath58 for nodes on a face , @xmath59 on an edge , @xmath60 at a vertex .
we implement the @xmath61 term of the equation of motion ( equation [ q : motion ] ) by subtracting @xmath62 from the density at all points before solving for the potential .
if @xmath62 coincides with the background cosmological density , the source function for the potential outside the principal octant is exactly zero .
this is the most common case , and the one that presents the fewest conceptual and practical difficulties . with periodic boundary conditions it would also be the only possible case since the mean density outside the simulation box must always equal that inside it . with isolated conditions a mismatch is permissible , and
could be used for example in applying an expanding or contracting grid to follow the evolution of an isolated system without a cosmological density background .
the varying @xmath24 would then be adjusted to match the expansion or contraction of the simulated system , providing better resolution at lower cost during the collapse phase .
a very significant difference between periodic and isolated boundary conditions is that the latter allow the exchange of mass , momentum , energy , angular momentum between the particles and the exterior . with periodic boundary conditions
, there is effectively no exterior .
particle flows are a significant concern in cosmological applications , where matter may both leave and enter the computational box during the simulation . in a typical cosmological model ,
the mass variance is of order unity in a sphere of radius @xmath63 mpc .
one expects particles at the edge of a sphere of radius @xmath64 to be displaced with respect to the center of mass of the sphere by about @xmath65 , a significant length when compared to the size of a group or cluster of galaxies .
a simulation of characteristic size @xmath64 may reasonably be expected to exhibit a twofold increase or decrease in the total mass within the box during a run . on smaller scales
the variance is even larger , at least for the currently fashionable `` bottom - up '' scenarios of structure formation . exiting particles
are easily handled by removing them from the simulation , but incoming particles have to be injected according to some prescription that fits the physical problem at hand . in the cosmological case , as long as the density perturbations grow linearly on the scale of the box , a reasonable prescription can be based on extrapolations from the linear theory . at late times , in the strongly non - linear regime , simply extrapolating from the linear solution causes large amounts of material to be injected which in reality would collapse into bound objects outside the box and remain outside . in other words , the zeldovich approximation is clearly inappropriate beyond the time at which caustics form .
it would lead to a gross overestimate of the total mass flow into the box .
an adequate model of the inflow in the nonlinear regime therefore requires an actual simulation of the mass flows in a larger region .
this has become normal practice in simulations with tree codes , and hierarchical grid methods such as ours also lend themselves well to this approach .
the main difficulty is that in order to keep the total number of particles manageable , one must essentially run a preliminary simulation simply to find out where the mass that flows into the region of interest originated and sample it with a finer granularity in the initial conditions .
this may lead to problems of contamination by more massive `` background '' particles if the small - scale structure that is not resolved by the preliminary simulation turns out to have a significant impact on the dynamics .
furthermore , the presence of particles of different masses in the system could lead to spurious mass segregation effects .
for these reasons , we attempted to keep the overall size of the computational box as small as possible , handling the tidal fields on larger scales , as well as any mass flows ( as long as they remained moderate ) as externally imposed boundary conditions .
this turned out not to be a particularly successful design choice , and we can not recommend it to others .
injecting too many , or too few , particles can have a destabilizing effect .
the first term on the right hand side of equation [ q : motion ] has , in the usual case @xmath66 , the effect of accelerating particles towards the boundary of the box .
this is balanced by the second term , which represents the attraction between the gravitating particles .
if more particles leave the box than are injected , the first term will tend to dominate and cause even more particles to be ejected ; conversely , if too many particles are added the material will tend to collapse towards the center .
this may be taken to represent a physical effect : a void is expected to expand faster than the universal average , an over - density more slowly .
but unless the algorithm for replenishing the box with particles is well thought out , a runaway instability may occur .
in practice we have found it expedient to couple the injection of particles to the outflow so as to maintain a constant mass within the system .
we then monitor the cumulative mass of particles so recycled and compare it with the total mass in the box . if the ratio becomes too large ( greater than 10% or so ) , we take it as an indication that one needs to simulate a larger volume of space .
as the volume that needs to be simulated grows larger , our original motivation for using isolated boundary conditions becomes weaker .
it turns out that the method we described in this paragraph works much better if the system is located in a region of lower density than the cosmic mean , as the mass fraction that undergoes reinjection is smaller in that case .
this allows us to simulate a number of interesting systems , but is not suitable for statistical studies involving a random selection of initial conditions .
momentum , angular momentum , and energy can be carried by the particles that flow in and out of the system and by direct interaction with the external potential @xmath67 and the uniform density we impose in the space that surrounds the particle region .
the latter unfortunately has the cubic symmetry of the computational box rather than the more convenient spherical symmetry that would be required for an exact cancellation of the induced tides .
consequently , these tides are always present except in the pure non - cosmological case with @xmath68 a constant , where the background density vanishes . whether this is a serious drawback depends on the physical problem being studied .
if necessary , one can enlarge the computational box ( this may be required in any case to keep mass flows under check ) , include a compensating term in @xmath19 , or both .
the idea of adding subgrids in regions where better spatial resolution is required is not a new one .
past approaches have differed on whether to add mass resolution at the same time by splitting the particles into a larger number of less massive ones , on how to minimize errors at the interface between the finer and the coarser grid , and on whether the solution on the coarser grid should be modified to take into account the results from the finer grid . unlike villumsen ( 1989 ) and splinter ( 1996 )
, we do not automatically introduce a new set of less massive particles on each subgrid .
we take the view that our initial mass granularity already matches the resolution we wish to achieve , and simply increase the force resolution ( by adding subgrids ) when and where the particle density is high enough to make this permissible and worthwhile .
this allows the decision of where to place subgrids to be made on a step by step basis , and spares us the need to perform a first simulation without subgrids to find out where particles of smaller mass should be placed in the initial conditions .
an additional advantage of having a single set of particles of equal mass is that we need not worry about mass segregation effects .
a drawback is that maintaining equivalent resolution within a larger computational volume requires more particles .
of course our code does support multiple particle masses , and as will become clear below the criteria for subgrid placement can be tuned to favor tracking the lower - mass particles ; we may therefore decide to experiment with particles of different masses in future . in our scheme
the various grids are merely devices , introduced independently on each integration step , to compute inter - particle forces . in this respect
we are closer in spirit to couchman s ap@xmath5 m than to most other multiple - grid approaches .
ours is effectively ( in the terminology of anc ) a particle - multiple - mesh ( pm@xmath0 ) scheme .
this distinction will have important consequences below . since there is only one set of particles that exist independently of any subgrids , the equivalent of a back - reaction from the subgrid solution to the parent grid
is automatically included : it is mediated by the particles themselves .
we typically decrease the grid spacing by a factor of two for each additional level of subgrids .
each subgrid thus covers a substantial number of cells of its parent grid .
other integral refinement factors are allowed by our formulation ( subject to the condition that the boundaries of any subgrid must coincide with cell boundaries of the parent grid ) but become increasingly inefficient in terms of volume coverage and we have not tried them in practice .
they would also exacerbate any `` ringing '' at the interface between a subgrid and its parent grid ; since this is one of the most delicate aspects of any hierarchical grid scheme , it is probably not advisable to use refinement factors larger than two under any circumstances .
a particle passing from a region of low resolution to one of higher resolution effectively undergoes a sudden change in its spatial extent .
( in pm codes , particles are best thought of as extending over about one grid cell or more , depending on the shape of the interaction law and on the charge assignment scheme . ) in its current state , our code takes no particular precautions against any transients that may result .
our comparison tests between runs that use adaptive subgridding and runs with uniform resolution show that such transients are not important enough to produce significant discrepancies in the results .
this is fortunate , since otherwise we would have had to associate a time - dependent smoothing length with every particle , which would complicate the method . that the smoothing length would have to be associated with the particle rather than with the location relative to a subgrid follows both from our choice to treat the particles as primary and the grids as auxiliary objects and from the fact that new subgrids may be added , removed or relocated to follow the flow at every step .
our strategy for computing the forces between subgrid particles differs slightly from that of villumsen .
his approach was to recalculate the potential on the entire parent grid after having set the density in the volume occupied by the subgrid to a constant value equal to the mean density within the subgrid , and use this solution as the tidal field on the subgrid particles from the rest of the system .
the forces between subgrid particles were then computed in the usual way by solving for the potential on the subgrid .
instead , we first compute forces for all particles on the parent grid , then generate the density array on the subgrid ( whose nodes must be aligned with those of the parent grid ) as well as on a coarser grid that has the same spacing as the parent grid and nodes in the same locations but covers only the volume of the subgrid . since coordinates on the subgrid
are affected by the same expansion factor @xmath24 as on the parent grid , we must subtract the same @xmath62 .
we solve for the forces on all subgrid particles from both these density arrays , and add the difference to the forces we had previously calculated on the parent grid .
this avoids double counting and means in effect that pairwise forces between particles that both lie in the subgrid are always evaluated at the higher of the two resolutions .
the coarse fft on the subgrid involves far fewer grid points , so our approach is more efficient than villumsen s .
it is also more flexible , in that we only need to store the accumulated force for every particle and the density potential for a single subgrid at a time , independently of the number and nesting depth of subgrids .
the cost of actually storing a force vector for every particle may seem high when one knows that in a traditional pm code with leapfrog time stepping the particle accelerations can be interpolated when the velocities are updated and need not be kept afterwards ; but saving them allows us to use them to adjust the time step and to compute subgrid corrections to the energy conservation test . on the other hand , we could have supported independent time steps had we chosen to save the potential on each subgrid instead .
this , however , would have caused the storage requirements to scale with the subgrid nesting depth .
let @xmath69 be the set of particles in the subgrid , @xmath70 its complement .
schematically , the acceleration on particle @xmath8 is computed as follows . for @xmath71 , @xmath72 for @xmath73 , @xmath74 where @xmath75 is the approximation to the coulomb potential on the parent grid , @xmath76 the approximation on the subgrid , and @xmath77 that on the subset of the parent grid that covers the subgrid .
( in reality , the sums are evaluated by fft and include the @xmath62 term . )
@xmath78 to a very good degree , provided only that the grid nodes are in the same locations .
( this is an essential requirement : we have found that violating it has deleterious effects on the solution . )
the two rightmost terms in the display of equation ( [ q : acc - sgr-1 ] ) therefore cancel each other .
note that we add together force contributions from the various grids rather than the potentials ; this allows the arbitrary additive constant in the potential to differ from grid to grid without affecting our solution .
we impose isolated boundary conditions on the subgrid , with an external background density ( @xmath62 ) equal to that used on the parent grid .
it may seem more appropriate to use the mean density within the subgrid , or
better yet the mean density in the immediate vicinity of the subgrid ; but one should note that some ringing is to be expected no matter what the exact scheme adopted since the density at the boundary of the subgrid always has some short - wavelength component that is resolved on the subgrid but not on the parent grid . the usual approach in codes of this type is to try to limit the ringing by introducing around the subgrid a buffer region , in which particles contribute to the potential but are not affected by it , and/or a transition region in which the potentials of the parent grid and of the subgrid are blended linearly .
the transition region implements the variable smoothing length we alluded to at the end of the previous subsection . as already noted
, it is of limited usefulness in our case since it does nt guard against transients induced by the sudden activation of a new subgrid .
the possible motivation for introducing a buffer region can be understood by considering the case of a particle @xmath79 located on the subgrid but near its edge , and subject to the forces of two nearby equidistant particles @xmath80 , also on the subgrid , and @xmath81 just outside it . in the exact solution , @xmath80 and @xmath81 exert forces of equal magnitude on @xmath79 . without a buffer region ,
the calculated force from @xmath80 is stronger
. this can create spurious small - wavelength perturbations as the pair @xmath82 tends to collapse on its own rather than as a triplet @xmath83 .
the introduction of a buffer region would allow the balance to be preserved between the attractions of @xmath80 and @xmath81 .
our implementation of the buffer region is as follows . extending the notation introduced in the previous subsection , we have split @xmath70 into the buffer region @xmath84 and a distant exterior @xmath85 .
equation [ q : acc - sgr-1 ] is modified in that the sum over @xmath86 becomes a sum over @xmath87 and the sums over @xmath88 become sums over @xmath89 .
we normally set the thickness of this buffer region to be equivalent to three cells of the parent grid : the grid - based force does not differ significantly from the exact coulomb force at separations this large , so that @xmath90 .
since the thickness of the buffer region is fixed in units of the spacing of the parent grid , large refinement factors result in the buffer region `` eating away '' most of the volume of the subgrid .
this is one of the reasons why we only experimented with a refinement factor of 2 at each level .
an essential feature of this buffer region is that it enables us to use multiple adjacent subgrids to cover extended structures that are worthy of higher resolution but do not fit within a single grid .
this could occur either because of their size or because they are located too close to other high - density objects already covered by their own subgrids .
( we do not allow arbitrary overlap of subgrids , for both simplicity and efficiency . ) a problem that might arise when the boundary between two adjacent subgrids cuts through the middle of a high density structure is that the force between two nearby particles separated by the boundary would be calculated with the ( poorer ) resolution of the parent grid .
thanks to the buffer region around each subgrid , this difficulty is avoided : the force exerted by either particle on the other will be computed at subgrid resolution by virtue of each particle lying in the buffer region of the other s subgrid . in the overlap region between adjacent subgrids
, it matters little whether the border region is implemented as described above or in the following alternative way , by subjecting particles in @xmath84 to the subgrid forces without including them in the subgrid potential .
this corresponds to applying equation [ q : acc - pgr-1 ] to particles @xmath91 only , and equation [ q : acc - sgr-1 ] to particles @xmath92 .
our first implementation used the latter approach , and comparison tests show only minute differences between the results of both variants .
a difficulty arises wherever the buffer region has finite width and does not overlap with an adjacent subgrid
. then the balancing of the forces by @xmath80 and @xmath81 over @xmath79 comes at the cost of a violation of newton s second law : the force exerted by @xmath79 on @xmath81 , being computed on the parent grid , does not balance the force by @xmath81 on @xmath79 .
the net result is that @xmath81 will be accelerated outwards , away from the subgrid .
( if the alternative implementation of the buffer region is adopted , the net effect has opposite sign and tends to push the particles inwards . ) the consequences for the orbit of a bound clump can be spectacular , as the following worst - case example , illustrated by figure [ f : tf4 g ] , shows .
we launch a 4096-particle realization of a truncated isothermal sphere ( with a half - mass radius of 0.02 units , corresponding roughly to half the cell width on the top grid ) with a mean velocity of 0.1 units towards a region covered by a fixed subgrid ( the inner bounds of which are indicated by dotted lines in the figure ) with refinement factor of two and a buffer region of width 0 ( solid curve ) or 2 ( dashed curve ) parent grid cells . for this example we adopted the alternative implementation of the buffer region , which causes the particles to be accelerated inwards .
the initial velocity is along the @xmath93 axis , perpendicular to the faces of the subgrid .
the figure shows the @xmath93 coordinate of the center of mass of the sphere as a function of time @xmath12 . when the buffer region is suppressed ( width 0 , solid curve )
the clump enters and exits the subgrid without changes to its bulk velocity .
( the apparent turnaround occurs when the clump reaches the edge of the computational volume and some particles leave the box .
the solid curve reflects the mean velocity of only those particles that remain in the box . )
the presence of a finite buffer region , by contrast , causes the clump to accelerate as it enters the subgrid , and to bounce back as it tries to exit on the other side .
the reason it does nt simply lose the momentum it gained when entering the subgrid is that the momentum change depends on the difference between the forces calculated on the subgrid and on the parent grid , and that the passage through the more highly resolved subgrid has allowed the clump core to relax to a more sharply peaked density profile , increasing the force mismatch .
the main features of this behavior are quite generic : a more carefully constructed clump with a larger number of particles that was allowed to settle into a numerical equilibrium before being launched through the subgrid underwent similar , if slightly milder , accelerations and rebounds .
clearly it is better for us to suppress the buffer region where there is no adjacent subgrid .
the violation of momentum conservation that a finite border region implies can have a drastic influence on the orbit of a bound clump .
this would not be much of a concern if we could guarantee that bound clumps are always entirely covered by subgrids at the same resolution ; the problem does not arise for smooth distributions of matter where short - range forces do nt predominate .
( in section [ s : gunn - gott - test ] we demonstrate this fact through a test of smooth infall onto a localized seed mass perturbation . ) in principle , an adaptive algorithm for subgrid placement based on the density should tend to ensure this .
however , one may want to place additional restrictions on which regions are followed at higher resolution , in which case the guarantee does not hold and we must take the precaution of suppressing the buffer region in the absence of an adjacent subgrid . note that the problem is less severe when expanding coordinates are used , as in cosmological applications , since the peculiar velocity acquired while crossing a subgrid interface will be damped away .
why has this same difficulty not been recognized by other authors , notably anc and splinter ( 1996 ) ?
the reason may be traced to a fundamental difference in philosophy between our code and theirs .
the very idea of a particle on a subgrid interacting symmetrically with a particle on the parent grid is a consequence of our decision to regard the particles , rather than the density field on the grids , as primary . in a scheme where every particle is associated with only one grid , and
particularly when the dynamics of the parent grid particles are unaffected by the subgrid ( the so - called `` one - way interface '' schemes ) , momentum conservation should be examined separately on each grid : on the parent grid , no violation is expected , while on the subgrid the effects of the parent grid particles are treated as an external field and the buffer and transition regions are used to smoothly bring any particles exiting the subgrid under the control of the parent grid field alone , as test particles , rather than reflecting them back into the subgrid . thus , the apparent contradiction between these authors use of a buffer region two cells wide and our restriction of the buffer region to the sole case of adjacent subgrids does not signal an error either on our part or on theirs , but is merely a manifestation of our different view of the respective roles of particles and grids .
a distinguishing feature of our code is that the activation of subgrids is entirely automated : the user need only set a few parameters ( maximum depth , and various optional thresholds ) , after which the code decides at every step where to place subgrids .
tuning the criteria for subgrid placement is not an easy task ; our current choices can undoubtedly be improved upon .
here we shall describe some of the criteria we have implemented .
not all of them have proved very useful in practice .
our subgrids have fixed shape and size , and may not overlap .
consequently , we found it simplest to introduce a uniform tiling of subgrids covering the entire volume of the parent grid
. each subgrid may be either active or inactive .
a subgrid is activated whenever our requirements for subgridding are satisfied within its volume .
we are free to choose the origin of the tiling independently for every parent grid at every step .
we make use of this freedom to reduce the number of active subgrids .
typically we try eight different possible origins , and adopt one that requires the activation of fewest subgrids .
the first criterion for subgrid activation is that the particle number density be sufficiently high in at least part of the covered region : there is no point in the grid ever being much finer than the smallest distance between particles . to activate a subgrid
, we require that among the corresponding cells of the parent grid either ( a ) at least one cell contains more than @xmath94 particles ; or ( b ) one cell contains at least @xmath95 particles and its 26 immediate neighbors together contain a further @xmath96 particles or more .
@xmath94 , @xmath95 and @xmath97 are specified as input parameters to the code .
the main reason for considering cubes of @xmath98 cells as well as single cells is that one needs to detect the collapse of a high - density region before it is too advanced : the additional resolution is already required during the later stages of collapse .
we do nt try to detect larger , lower - contrast potentially collapsing structures on larger scales since for separations larger than about three cells the forces are computed accurately on the parent grid . one way of choosing @xmath95 and @xmath97 is the following .
let @xmath99 be a measure of the mean number of particles per cell in some larger region , such as the entire candidate subgrid .
( this measure need not be unbiased ; in fact , if the mean particle number density is very low we found it necessary to artificially increase @xmath99 by setting it to the value 1 .
otherwise our subgrid placement criterion would be satisfied far too easily in low - density regions . )
if the particle counts were distributed according to poisson statistics , one would expect a group of @xmath100 cells to contain on average @xmath101 particles , with a standard deviation @xmath102 around that mean . by setting @xmath103 and @xmath104 , where @xmath105 is a tunable parameter
, we try to ensure that only particularly significant deviations from homogeneity trigger subgrid activation .
we have found that this approach works reasonably well during the early stages of nonlinear evolution , where it causes higher resolution to be applied to the regions we are most interested in .
however , as the simulation progresses we have needed to increase @xmath105 in order to limit the proliferation of subgrids , as our goal was merely to achieve high resolution in a few large peaks near the center of the computational box . clearly the criteria can and
should be tuned according to the needs of individual applications ; no single choice is optimal for all cases .
we follow the usual practice in pm codes of advancing the positions and velocities of the particles in leapfrog fashion : @xmath106 an alternative formula applies on a starting step , when both the positions and the velocities are known at the same time @xmath107 : @xmath108 likewise , one can synchronize positions and velocities at the final step by advancing the velocities from @xmath109 to @xmath110 then updating the positions as @xmath111 in large simulations , it is particularly important to avoid taking more time steps than required to obtain accurate results .
the time step should therefore be allowed to vary . in the standard leapfrog formulation ,
the cancellation of second - order terms in equation ( [ q : lf - v ] ) depends on @xmath112 being evaluated at the midpoint of the velocity step .
the cancellation of the second - order term in the formula to update the positions is less important since the @xmath113 are known ; however , it contributes to lowering the operation count for the method .
various approaches can be used to change the time step in the middle of a simulation .
one is to use a different midpoint in equation ( [ q : lf - v ] ) and calculate the corresponding @xmath114 . the other , which is the one we adopted , is to keep @xmath115 constant by an appropriate choice of the function @xmath40 .
we construct this function and its first and second derivatives , which are needed to evaluate @xmath112 and @xmath84 according to equations ( [ q : a ] ) and ( [ q : b ] ) , step by step as the simulation progresses .
( in practice , we do not allow @xmath116 to vary too quickly , since that tends to spoil the results .
we shrink @xmath117 by at most 25% on each step , and let it grow even more slowly .
situations in which this forces a larger @xmath117 than desired have fortunately been very rare in our runs . )
our preferred criterion for determining the time step is the courant condition : @xmath118 where the index @xmath9 spans the set of all particles , @xmath119 is the grid spacing of the finest grid used to compute the forces on particle @xmath9 , and @xmath120 is a constant coefficient . @xmath121 should be taken as being the mean velocity over the time step , and is a function of both the velocity before the step is taken and of the acceleration .
the latter can be important if the initial velocity is uncharacteristically small , for example when all the particles are initially at rest .
the equation for the maximum acceptable @xmath117 for each particle is actually a quartic .
we do not solve it exactly , preferring to estimate a close lower bound on the solution for the particles with the tightest @xmath117 constraints .
this way of taking the accelerations into account fulfills the same function as other authors use of the maximum density on a subgrid to estimate a local dynamical time . in some cosmological simulations with a large number of particles , we found that adopting the smallest @xmath117 found in this fashion was leading to time steps much shorter than warranted by our physical understanding of the dynamical time scales involved ( based on the maximum density ) .
this turned out to be due to a small number of high velocity particles : the tail of the velocity distribution produced by the violent relaxation of a newly collapsed object , which is naturally better sampled as the number of particles in the simulation increases . since these particles represent a small fraction of the mass in the typically high - density regions in which they are found , their motion has almost no impact on the mean gravitational field , which evolves on much longer time scales .
this led us to modify the time step criterion for these simulations by pretending that the velocity of each particle before the time step is negligible and basing the choice of @xmath117 solely on the accelerations .
this is not unreasonable for cosmological applications since in the linear regime the velocities and peculiar accelerations are related and in virialized clumps they are a better indication of the dynamical time scale than the velocity of the fastest particle ( which is not the same as the local velocity dispersion ) .
moreover , expanding coordinates have the property that peculiar velocities are damped away and need to be continually regenerated by the forces . as a further precaution
, we have analyzed the velocities of the particles at a late stage in a simulation run with this relaxed time step criterion and found , as intended , that the number of particles that violated the stricter courant condition was small .
the relaxed criterion is known to fail , however , for highly ordered collapse ( such as that of a homogeneous sphere ) . in that case
the largest acceleration does not provide a good estimate of the time scale over which the mean field evolves : based as it is on the more extended distribution at the beginning of the step , it systematically overestimates the maximum allowable time increment .
accordingly , we only resort to the relaxed criterion , with some reluctance , for runs in expanding coordinates , with large numbers of particles , and where different mass shells collapse at different times .
our tests on smaller cosmological runs have shown the results to be substantially identical , but with the relaxed condition requiring a much smaller number of steps .
we use a single value of @xmath117 for all particles on all subgrids at any given step . a far better approach in principle , but much more complicated to implement , would be to adopt multiple time steps and to distinguish between the time steps used for advancing individual particles and those for updating the potentials on subgrids .
independent time steps for the particles would require a way of estimating the accelerations at positions and times slightly different from those for which the potential has been calculated . for this , it may be necessary to calculate the time derivative of the subgrid potential as well as the potential itself .
in addition to the constraints resulting from the rate of change of the subgrid density , the frequency with which the subgrid potential is recalculated can also be affected by particles flowing out of the region in which forces can be interpolated from the subgrid potential : in our scheme , such events change the mapping between particles and subgrids and require recalculation of all the subgrid potentials involved if the total momentum is to be conserved .
these are our main reasons for adopting a single time step for all grids .
the other main reason is that a single time step allows us to process subgrids one by one , accumulating the force contributions on each particle without needing to store the solutions for the potential on many nested subgrids simultaneously .
almost any other time step scheme requires these subgrid potentials to be available simultaneously .
the coefficients in the leapfrog formulae ( equations [ q : lf - v ] and [ q : lf - x ] ) can depend on time . for good accuracy , their variation in a time step
must be small .
this constraint is clearly unrelated to the grid spacing , and must be imposed separately .
it is equivalent to the requirement that @xmath122 .
conservation laws provide an important diagnostic of how accurately the equations of motion have been integrated in a given simulation .
good conservation of physical invariants is not a sufficient condition for a good integration , but it is a necessary one . in the absence of coordinate expansion , subgrids , and external fields , a pm code such as ours would be expected to conserve total momentum algebraically , energy a little less well , and angular momentum only in the limit where the interparticle separation is much larger than the grid spacing ( he 76 ) .
cosmic expansion causes the usual equation of conservation of the total energy , @xmath123 ( where @xmath124 is the kinetic , @xmath125 the potential energy ) , to be replaced with the layzer - irvine ( @xcite ; @xcite ) equation , which can be derived from the well - known property of the hamiltonian @xmath126 : @xmath127 computing the canonical momentum from equation [ q : lag - x ] , @xmath128 it follows in the usual way that @xmath129 we define the kinetic and potential energies as : @xmath130 then @xmath131 our exclusion of the external potential @xmath19 from the definition of @xmath125 is somewhat arbitrary . in practice , one wants to separate the terms associated with the mutual interactions of particles , which do measure the accuracy of the integrator , from the correction terms that merely account for known , external sources and sinks of energy . in applying equation [ q : dham - dt - law ] ,
one notes that in @xmath132 the last term does not vanish if @xmath133 depends explicitly on time ( as may occur in practice when a new subgrid is activated and the force resolution increases locally ) , or if @xmath134 is not constant . for simplicity , we shall assume @xmath135 from now on , since that is the only case of actual practical interest to us .
straightforward algebra leads to the two equations : @xmath136 @xmath137 which although physically equivalent are evaluated numerically in slightly different ways when @xmath24 is not constant .
the corresponding conserved quantities are : @xmath138 @xmath139 in proper - length coordinates ( @xmath140 ) , the first integral term ( @xmath141 or @xmath142 ) vanishes and both criteria reduce to the usual law of conservation of energy .
the last two terms represent respectively the work done by external fields and time dependence in the interaction law .
one may question the appropriateness of including the latter effect , since its origin lies in the numerical method used rather than in the underlying physics
. should nt contributions from the @xmath143 term be counted as violations of energy conservation by the code ?
in reply we point out that much of that term may result from fluctuations in the zero level of the gravitational potential on subgrids , which we did not attempt to suppress since they have no bearing on the computed forces . and in any case the @xmath143 contribution
should be evaluated so that its magnitude relative to the other terms may be assessed .
it is unrealistic to expect a pm code to conserve @xmath70 and @xmath144 arbitrarily well when gravitational instability causes the individual terms ( @xmath125 , @xmath145 ) to grow by orders of magnitude . like
many other authors before us , we deem @xmath70 and @xmath144 adequately conserved if their variation is a small fraction ( typically 13% ) of the change in @xmath125 or @xmath146 , respectively . in our definition of @xmath125 ( equation [ q : w ] ) we excluded the self - energy terms @xmath147 .
but the natural way of calculating the potential in a pm code effectively includes these terms ( the restriction @xmath148 can be ignored since the scheme avoids self - forces ) ; we subtract them explicitly as part of our energy conservation test .
we adopted a momentum - conserving scheme for calculating the forces .
the total momentum should therefore be unaffected by particle - particle interactions on any given grid , and ( given the precautions we have taken in our handling of the buffer region surrounding each subgrid ) be only weakly modified by interactions across the boundary between adjacent subgrids .
however , it is still affected by all other interactions between particles and external fields , including the constant - density background outside the computational box .
( this field gives rise to forces only when isolated boundary conditions are used , which is the reason why it is nt normally discussed in the literature . ) the canonical momentum @xmath149 of particle @xmath8 satisfies @xmath150 summation over all particles and time integration yield a conserved quantity @xmath151 in the special case when ( i ) @xmath19 has no spatial dependence , ( ii ) @xmath152 or @xmath153 , and ( iii ) @xmath154 , the total momentum is constant . if one s aim is to measure the deviation of the results from those expected when @xmath154 ( such a deviation can occur in the presence of subgrids ) , one should naturally omit the corresponding term from equation [ q : pc ] .
likewise , the conservation of angular momentum @xmath155 amounts to the constancy of @xmath156 angular momentum is only conserved if ( i ) @xmath157 and ( ii ) @xmath19 is a function of @xmath158 and @xmath12 alone .
the exact coulomb force law satisfies the first condition , but grid - based approximations to it do nt . as in the analogous case for the momentum ,
if the aim is to measure deviations from the ideal behavior where ( i ) is satisfied , then one should compare the magnitude of the corresponding term to @xmath159 itself , and to the magnitude of the external torque term when present .
our first tests are meant to see how accurately the code calculates the accelerations of particles given a known mass configuration .
the simplest case is that of the forces due to a single point mass , which can be compared to the value predicted by coulomb s law .
figure [ f : tf1b.dff ] illustrates this comparison .
it shows the accelerations induced by a single particle at the center of a @xmath160 grid with isolated boundary conditions .
a single @xmath160 subgrid , twice as fine as the top grid , was centered on the massive particle .
two sets of points are readily distinguished , corresponding to test particles inside and outside the subgrid .
as expected , the acceleration is systematically underestimated since the effective potential is softer than that of a point mass . also as expected , the subgrid extends the range over which the force is accurate to within a few percent ; the accuracy threshold ( a little over 3% in this test ) could be lowered by increasing the size of the subgrid , the spacing being equal .
our second test checks that the accelerations computed on a set of adjacent subgrids agree with those one would obtain on a single , larger grid with the same spacing as the subgrids .
we compare forces on individual particles randomly chosen to sample a uniform distribution .
the long - range forces , which are adequately resolved on the parent grid , tend to cancel , so that the system is dominated by short - range interactions that will be most affected by the grid refinement .
our reference forces are from a non - subgridded run with @xmath161 cells and @xmath161 particles .
we compare these to those obtained from three different runs with only @xmath162 cells on the parent grid , in which all possible first - level subgrids are activated .
the only difference between the three runs is in the width of the buffer region where adjacent subgrids overlap .
the results are illustrated in figure [ f : df ] .
the magnitude of the force difference is plotted against that of the reference force for a randomly selected 1% of the particles ( to avoid overcrowding the plot ) .
panels ( a ) , ( b ) , and ( c ) correspond respectively to a buffer width of 0 , 1 , and 2 parent grid cells .
the solid diagonal line represents @xmath163 ; the last panel also contains a dotted line where @xmath164 .
the results show that when the buffer is sufficiently wide ( at least two cells ; we like to use three cells in production runs with larger grids ) , the forces computed by the subgrid method are mostly within 1% of the forces from a single grid of equivalent resolution .
whenever the buffer region is suppressed , an appreciable fraction of the particles exhibits force errors larger than 10% .
here it is essential to understand what we mean by `` errors '' .
figure [ f : df ] only compares the forces computed using a set of abutting subgrids of equal resolution with those from a single , larger grid with the same spacing , and shows that some overlap is both required and sufficient to avoid loss of resolution at the boundary between such abutting subgrids .
it does _ not _ compare forces computed at different grid spacings , and does not address the issue of convergence to a `` perfect '' solution in the limit of an infinitely fine grid .
we know that at the interface between a subgrid and a coarser parent grid the accuracy of the forces within the outer two or three cell layers of the subgrid will be less than that of a uniform fine grid regardless of the thickness of the buffer region .
( it will , however , be no worse than on the parent grid alone . )
we compensate for this by making the subgrids cover a slightly larger volume than the region where the higher resolution is required .
the collapse of a homogeneous self - gravitating pressure - less ellipsoid initially at rest can be described by a small set of coupled ordinary differential equations for the lengths of the principal axes , which can be integrated in a straightforward way ( @xcite ) .
this solution provides a convenient test of our code .
the test corresponds to that of collapse to a sheet or filament in a cosmological code with periodic boundary conditions , but is more appropriate to our use of isolated boundary conditions . in particular , the plane - wave test of efstathiou et al .
( 1985 ) is only easy to interpret and compare with a semi - analytic solution when periodic boundary conditions are used , and these lie outside the scope of the present article .
another test , in section [ s : voidtest ] below , shows the behavior of our code when integrating the collapse of a moving sheetlike ridge in the presence of cosmological expansion .
we followed the collapse of a homogeneous ellipsoid with a @xmath165 ratio between the lengths of the principal axes .
the experiment was repeated for two different choices of the grid spacing in order to verify convergence towards the correct solution . for the higher resolution , we performed two runs , one using a single grid of @xmath166 cells and one with a top - level grid of only @xmath160 cells but up to two levels of subgrids .
each subgrid also has only @xmath160 cells ; when necessary , multiple adjacent subgrids were automatically generated by the code to cover the entire ellipsoid .
the number of particles is the same in all three runs , about @xmath167 .
figure [ f : ts7-overview ] shows the short axis @xmath168 of the ellipsoid as a function of time @xmath12 .
the solid curve represents the analytic solution , the dot - dash curve the solution computed on the coarse grid , while the remaining two dashed curves show the results of the two high - resolution runs .
the horizontal dotted line corresponds to the grid spacing in the high resolution runs ; that of the low resolution grid is four times larger .
figure [ f : ts7-blowup ] presents a blown - up view of the same results around the time of collapse of the ellipsoid to a spindle .
one sees from these figures that increasing the resolution does indeed yield a better agreement between the @xmath1-body and analytic solutions .
furthermore , the results with subgridding are nearly identical to those obtained with a single grid of equivalent resolution . in the high - resolution runs ,
the minimum radius of the spindle is comparable to the grid spacing . at the lower resolution ,
it is significantly less than the grid spacing .
we did not investigate the reasons for this in detail , but it is likely that the convergence is also affected by the number of particles used to represent the ellipsoid . in these tests
we have kept this number constant .
the two runs without subgrids also used a small , constant value of the time step @xmath169 .
collapse occurs after 1524 and 1644 steps respectively .
the subgridded run , by contrast , used the courant condition to adjust the time step ; only 125 steps were required to reach the point of collapse .
the good agreement between the high - resolution results obtained with both choices of time step confirms the validity of our implementation of adaptive time steps .
we have also repeated the low - resolution run using our adaptive time step scheme ; on figure [ f : ts7-blowup ] the results would be indistinguishable from those obtained with a fixed time step .
an interesting test of the behavior of mass flows into a subgrid is the case of accretion from a uniform - density background onto a point mass ( or any other compact , spherically symmetric density profile )
. a regular lattice of @xmath161 particles with zero initial peculiar velocity was laid down in a computational box of unit side to be evolved with isolated boundary conditions in an @xmath170 , @xmath171 cosmological model .
we added at the center of the box a single particle of mass @xmath172 , where @xmath173 is the mean density of the other particles and of the external background , @xmath174 , and @xmath175 .
we followed the collapse to a time @xmath176 , corresponding to an expansion factor @xmath177 .
a scaling solution is available ( @xcite ; @xcite ; @xcite ; @xcite ) : @xmath178 .
this simple law is expected to hold only for those shells that enclose a mass much larger than that of the initial seed and that have evolved for a few crossing times after their initial turnaround .
the crossing time for a shell is of the same order as its turnaround time , and its proper radius is a factor @xmath179 times the turnaround radius . using a linear theory approximation ( in which the density perturbation grows in place ) until the turnaround at a mean enclosed overdensity @xmath180
, we find that for a given initial enclosed overdensity @xmath181 the turnaround occurs at @xmath182 , the subsequent virialization at @xmath183 , and the final comoving radius of the shell is @xmath184 , where @xmath185 is the initial comoving radius . requiring @xmath186 , only shells with @xmath187 are expected to show self - similar behavior , and then only for @xmath188 .
one can also evaluate bertschinger s ( 1985 ) equation ( 2.7 ) with @xmath189 , @xmath190 , @xmath191 ( these values lead to @xmath192 ) and make direct use of his numerically determined similarity solution .
a twist of our simulations is that shells with @xmath193 are prevented from falling in by the fact that they are not entirely included within the simulation volume .
accordingly , the accretion will be starved for @xmath194 , and the last complete shell should virialize at @xmath195 .
our main reason for performing this test was to compare the solutions obtained with various treatments of the border region around a subgrid .
we performed four runs : one with a uniform @xmath196 grid , the other three with a @xmath160 grid and a nested @xmath160 subgrid with a linear refinement factor of 2 .
( after subtracting edge cells as discussed in [ s : bc ] , the computational box was covered respectively by @xmath161 and @xmath162 cells . ) in one run the border region was suppressed , as suggested by our tests with a bound clump crossing the boundary , while the other two both used a border width of two parent grid cells , once with the particles in the border region contributing to the subgrid potential without feeling its effects and once with the border particles being subjected to the subgrid potential without contributing to it . in the last two cases , the net effect of the force asymmetries on overdense regions results in an acceleration pointing respectively away from the subgrid and towards it .
figure [ f : tf6d - rho ] shows the logarithmic density profiles for the runs we just described .
the differences between the results of all these runs are minute ; we find it difficult to ascribe any significance to such small deviations .
this is reassuring : it suggests that the exact way in which subgrid borders are treated does not unduly affect the profiles of collapsed peaks .
the dotted line connects open triangles that correspond to the values given by bertschinger ( 1985 ) in his table 4 . at small radii
, his solution tends towards the expected @xmath197 , and ours is in good agreement with his .
the first few caustics can easily be identified .
peebles ( 1987 ; see also @xcite , hereafter pmhj ) has proposed and used the following test for cosmological pm codes : integrating the evolution of a spherically symmetric under - dense region surrounding by a compensating over - dense shell .
the interior of such a void expands faster than the universe as a whole , and in so doing compresses the surrounding shell , making its profile higher and narrower .
the evolution can be computed analytically until the time at which the first density cusp appears . for peebles profile , @xmath198 with @xmath199 , this occurs at @xmath200 in a flat @xmath170 universe .
( here @xmath201 is the initial value of the expansion factor @xmath68 . )
we integrated such a void , realized with @xmath196 particles , on a @xmath160 grid and compared our results to the analytic prediction at @xmath202 .
figure [ f : ts6aa ] can be directly compared to the corresponding figure 4 of pmhj .
a significant difference between our code and theirs lies in their use of a staggered grid for the force ( @xcite ) .
the staggered grid increases the spatial resolution twofold ( which is why these authors used it ) at the cost of inducing self - forces on the particles ( for which reason we did not follow their example ) . because of this difference , our results obtained with @xmath196 particles on a @xmath196 grid correspond to their results with the same number of particles on a @xmath160 grid .
a comparison reveals differences of detail , particularly in the structure of the outer parts of the shell , but our results match the analytic solution as closely as theirs .
we take this occasion to illustrate the energy conservation properties of the code .
figure [ f : ts6aa - econs ] shows the evolution with time of the conserved quantities @xmath70 and @xmath144 of equations [ q : c ] and [ q : cprime ] ( top ) , as well as that of the ratio @xmath203 of the changes in @xmath70 and in the total potential energy @xmath125 from the start of the run ( bottom ) .
apart from an initial transient due to rounding in the print - out from which the plot was made , an examination of the ratio shows the energy to be conserved to about 12% accuracy throughout the run .
we have developed a useful and versatile tool for the simulation of collisionless systems of gravitating particles .
our code is particularly suited to situations that call for both fine - grained mass resolution everywhere and high spatial resolution in selected regions of high density the exact locations of which need not be known in advance .
our method is a development of the well - known particle - mesh technique .
tests indicate that in comparable conditions our code performs substantially like similar codes described in the published literature . when subgrids are introduced to increase the spatial resolution locally , the results in the regions covered by the subgrids are in very good agreement with those of a conventional pm code with a single , finer grid .
if the number of subgridded regions is sufficiently small , our approach requires less computing time and less memory than the equivalent single - grid approach .
furthermore , by not increasing the resolution in regions where the particle number density is low , we avoid making the behavior of the code collisional , which would be undesirable for applications to collisionless systems .
a significant advantage of pm ( and its variants p@xmath5 m , ap@xmath5 m , etc . ) for cosmological applications is that comoving coordinates and periodic boundary conditions can be supported in a natural way .
however , periodic boundary conditions are not always a physically appropriate approximation .
in particular , if one s interest is in systems not much smaller than the size of the computational cube , periodic boundary conditions introduce a much stronger coupling between the external tidal field and the internal dynamics of the system than one would expect to occur in the aperiodic real universe .
periodic boundary conditions are only appropriate when the individual structures of interest are much smaller than the computational volume , in which case the tides due to periodic images are small ; but then the relatively small dynamic range of conventional pm methods severely limits the resolution that can be achieved . by introducing hierarchical subgrids , we are able to extend that dynamic range ; however , it is also tempting to construct the simulation in such a way that the system of interest fills as much of the computational box as possible .
this is what led us to implement isolated boundary conditions . in principle
, the additional cost of imposing isolated boundary conditions ( which is relatively small when hierarchical grids are used since isolated conditions have to be imposed on the subgrids in any case ) is offset by the greater freedom one has to apply an arbitrary time - dependent external tidal field and to adopt a system of expanding coordinates that matches the evolution of the simulated object without necessarily coinciding with the expansion of the global cosmological model .
of course our method can also be used in a more conventional way , with periodic boundary conditions on the top grid ; in fact , numerous simplifications occur when this is done .
in addition , our code should be well - suited to non - cosmological dynamical simulations , _
e.g. _ , of brief interactions between already formed galaxies , star clusters , and so on .
our approach also has a few limitations .
the most important is shared by all particle simulation methods : it is generally impossible to improve the spatial resolution without a corresponding refinement of the time resolution : more time steps need to be taken . other difficulties of note are the non - radial nature of forces at small separations ( which could be cured , at some cost in computing time , by softening the interaction law and increasing the depth of subgridding to compensate for this softening ) , the complexity and difficulty of tuning the subgrid placement algorithm , and the fact that subgrids of fixed shape will almost inevitably also cover some regions where the particle number density is low .
the behavior of the code can become collisional in such regions ; this should not affect the computed structure of high - density condensations , but may well be a source of high - velocity particles that lead to shorter time steps according to the courant condition .
an enhancement to our current code could be to refrain from applying the refined forces to such low - density cells .
alternatively , it could prove useful to introduce individual time - varying softening lengths associated with individual particles , to represent the fact that particles in this method are to be thought of as possessing a finite extent that depends on the local number density .
another desirable improvement is the adoption of different time steps at different levels of grid refinement ; the savings that can be achieved from this depend on the detailed structure of the tree of subgrids , and will be greatest for deep trees with most of their branches at the finer resolution levels .
this work was partially supported by grants nsf - ast-86 - 57647 , nsf - ast-91 - 19475 , and nasa - nagw-2224 .
some calculations were conducted using the resources of the cornell theory center , which receives major funding from the u.s . national science foundation and the state of new york , with additional support from the advanced research projects agency , the national center for research resources at the national institutes of health , ibm corporation , and other members of the center s corporate partnership program .
we thank adrian melott , richard james , and the referee randall splinter for interesting and valuable remarks on an earlier version of this paper . | this article describes a new , fully adaptive particle - multiple - mesh ( pm@xmath0 ) numerical simulation code developed primarily for cosmological applications .
the code integrates the equations of motion of a set of particles subject to their mutual gravitational interaction and to an optional , arbitrary external field .
the interactions between particles are computed using a hierarchy of nested grids constructed anew at each integration step to enhance the spatial resolution in high - density regions of interest . as the code is aimed at simulations of relatively small volumes of space ( not much larger than a single group of galaxies ) with independent control over the external tidal fields ,
significant effort has gone into supporting isolated boundary conditions at the top grid level .
this makes our method also applicable to non - cosmological problems , at the cost of some complications which we discuss .
we point out the implications of some differences between our approach and those of other authors of similar codes , in particular with respect to the handling of the interface between regions of different spatial resolution .
we present a selection of tests performed to verify the correctness and performance of our implementation .
the conclusion suggests possible further improvements in the areas of independent time steps and particle softening lengths . |
mapping the properties of a fermionic system to a bosonic gas localised in an artificially created complex potential landscape is a powerful tool for probing new condensed matter phenomena .
this approach is known as quantum simulation @xcite attracting vast scientific interest in recent years since it allows to use well - controllable bosonic quantum systems to study far less controlled fermionic properties .
quantum simulation is quite new due to the fact that different bosonic systems , such as ultracold atomic gases @xcite , arrays of photonic waveguides @xcite , and exciton - polariton ( later , _ polariton _ ) gases @xcite , have just become available experimentally .
the latter system has certain objective advantages over the others due to the mixed light
matter nature of exciton polaritons ( eps ) .
indeed , on one hand , it became relatively easy and cheap to design and implement arbitrary linear periodic potentials for the photons , utilizing photonic crystals or metamaterial structures , thus particles can be localized and conveniently investigated . on the other hand , due to the excitonic component
, the gases of eps are characterized by strong nonlinear properties resulting from exciton - exciton kerr - like scattering which represents a superposition of dipole - dipole and exchange interaction .
it is important to note , that beside mentioned properties inherited from excitons and photons separately , eps as hybrid modes with peculiar dispersion , may form quasi - bose
einstein condensate ( bec ) in the steady state at sufficient background pumping intensities .
polariton superlattices ( later , _ lattices _ ) of different geometry have been recently extensively studied both theoretically and experimentally @xcite with most of attention paid to honeycomb and the kagome lattices . while at the moment most of the theoretical efforts are dedicated to the linear dispersion features and topological properties of the lattices , it becomes clear that the nonlinear phenomena of such systems should also be substantially affected by the linear spectrum and chirality of polaritons in the hexagonal lattice . in a recent work @xcite it
is theoretically shown that chiral edge bogoliubov states can be excited in a weakly interacting polariton gas in a kagome lattice potential .
spin - orbit coupling originating from the anisotropy of the hopping coefficient has also been studied both theoretically @xcite and experimentally @xcite .
moreover , similar structures have been proven to serve as topological insulators supporting chiral edge states @xcite .
it is well known that ultrafast relaxation ( the order of femtoseconds ) of hot electrons in graphene is dominated by the intraband coulomb scattering @xcite .
moreover , it has been shown in @xcite that the intraband electron - electron scattering in graphene is mediated by the collinear processes , which counter - intuitively play significant role in the scattering dynamics , even though they appear along the one - dimensional lines on the dispersion surface and thus , for the first glance , should give zero contribution to the behavior of two - dimensional dirac fermions @xcite .
it seems feasible to expect that a bosonic system interacting via a short - range potential ( analogously to the screened coulomb potential for electrons in graphene ) and characterized by a linear dispersion should also demonstrate similar ultrafast interaction with induced relaxation to the dirac points in a certain frequency range . in this article
, we show that the polariton - polariton interaction can be substantially enhanced in the vicinity of the dirac points .
such a behaviour can be expected if we imply the mentioned quantum simulation mapping between the bosonic and fermionic systems , recalling that the graphene is characterized by ultrafast electron relaxation caused by particle - particle interaction . speeding up the polariton scattering to the ground state
can be used in various situations @xcite .
one of the straightforward possible applications is a polariton laser . in this device ,
coherent radiation from the system occurs due to spontaneous emission from a macroscopically occupied single - particle ground state .
since no population inversion is required , such laser is characterized by smaller threshold than conventional stripe edge - emitting or vertical cavity surface emitting lasers .
polariton lasers are usually pumped nonresonantly , optically or via electric current injection @xcite .
the threshold here depends on the ratio of particle lifetime and the complex effective scattering time , the time required for the polaritons to reach the ground state from the reservoir of excited states . using our proposal ,
it becomes possible to further decrease the scattering time and thus decrease the lasing threshold .
we consider a structure with geometry similar to one described in work @xcite , see fig .
[ fig_1 ] .
it represents a two - dimensional array of micropillar cavities , a design , which can be routinely produced by etching a regular planar semiconductor microcavity .
we assume that each micropillar hosts one exciton - polariton mode , and photons are allowed to hop between the neigbouring pillars only .
a unit cell of the lattice consists of two pillars , thus we can denote two sublattices as a and b. we note that in ref .
@xcite and @xcite , the authors experimentlly demonstrate existance of the dirac cones in the dispersion of such structure and polariton bec in the vicinity of the dirac point .
the part of the system hamiltonian corresponding to linear coupling between the subsystems a and b within the tight - binding model reads @xmath0 where @xmath1 , @xmath2 and @xmath3 are the annihilation operators of cavity photons and excitons within the sublattice @xmath4 , and @xmath5 is exciton - photon coupling strength in each pillar .
the neighbouring cavities are coupled via the light components of eps only with the coupling strength @xmath6 . here
we neglect the spin - orbit coupling originating from the te - tm splitting of the cavity modes which has been addressed in @xcite . taking the fourier transform of eq .
( [ ham0 ] ) , we obtain the matrix form of the hamiltonian in the basis @xmath7 corresponding to the photon and exciton mode in the @xmath8 and @xmath9 sublattice , respectively which takes the form @xmath10 where @xmath11 .
the eigenenergies of this hamiltonian can be found by diagonalization the @xmath12 matrix in ( [ eqhammatrix ] ) .
similarly to the case of graphene @xcite , the brilloin zone here contains two inequivalent special points @xmath13 , where @xmath14 is the distance between @xmath8 and @xmath9 sublattices . however , in our case there are two dirac points with frequencies @xmath15 located at @xmath16 and @xmath17 .
the resulting band structure is shown in fig .
[ fig_2](a ) .
figure [ fig_2](b ) demonstrates the dispersion in the vicinity of the upper dirac point @xmath16 .
further we expand the hamiltonian in the vicinity of @xmath16 and yield @xmath18 where @xmath19 is the fermi velocity of the photonic mode , and @xmath20 .
we then apply the following unitary basis transformation @xmath21 which allows us to rewrite the hamiltonian ( [ ham2 ] ) as @xmath22 we see that there are two chiral polariton branches in the vicinity of energy @xmath23 and two in the vicinity of @xmath24 .
it is also clear from the matrix above , that these modes are coupled .
however , if we put ourself in the vicinity of the dirac points , such that @xmath25 , we can omit the coupling of upper and lower dirac cones since the energy correction due to these couplings is proportional to @xmath26 .
then the hamiltonian takes a block - diagonal form and splits on two @xmath27 matrices , and we can describe each of the dirac cones separately .
hence the @xmath27 hamiltonian of the two low - polariton chiral branches is given by @xmath28 with eigenenergies @xmath29
in order to account for the polariton - polariton scattering , we consider another term of the hamiltonian , @xmath30 ^ 2 , \label{intham}\ ] ] where @xmath31 is a polariton field operator .
thus we investigate contact like polariton polariton interaction .
for simplicity , we assume that the hopfield coefficients for excitonic and photonic fractions are the same , @xmath32 , hidden in the coupling strength @xmath33 . our main goal is to find the analytic formula for the polariton scattering time ( inverse scattering rate ) .
let us assume that the system is homogeneous at the initial moment of time with zero number of particles .
then it is excited with energy of external pump above the dirac cone .
a polariton with initial momentum @xmath34 can be scattered towards the state @xmath35 with the transferred momentum @xmath36 .
the momentum and energy conservation laws imply @xmath37 where @xmath38 , @xmath39 are the angles between vectors @xmath36 and @xmath34 , @xmath36 and @xmath35 , respectively . the plus sign in eq .
( [ eq : enconserv ] ) corresponds to the intraband scattering , and the minus sign to the interband one .
it can be easily shown that in the case of interband process , only the collinear scattering @xmath40 is allowed .
this means that in the case of broad in @xmath36 initial distribution , the rate of the interband processes is negligible as compared to the interband scattering rate .
even the account of finite polariton lifetime does not lift this reduciton .
therefore , in what follows , we assume that eps occupy only the positive energy states .
we now use the diagrammatic approach in order to find the polariton scattering lifetime .
it is well known , that the feynman diagram of the first order in the polariton interaction constant @xmath33 [ see fig .
[ fig3](a ) ] does not contribute to the scattering time .
in the second order in @xmath33 , the self - energy operator is given by the diagram depicted in fig .
[ fig3](b ) , @xmath41 where @xmath42 are bosonic frequencies , @xmath43 is a green function of the upper - branch polaritons , and @xmath44 $ ] is their chirality form - factor .
note , here and below we measure the energy of the polariton upper branch from @xmath45 and assume @xmath46 , where we denote @xmath47 for convenience . in this case
we will obtain the bosonic distribution with the chemical potential @xmath48 connected with the total polariton density @xmath49 . in the lowest order in @xmath33 , the function @xmath50 is given by the single - bubble diagram , @xmath51 where @xmath52}-1\}^{-1}$ ] is the bose distribution , @xmath48 is the chemical potential .
it can be easily demonstrated that the direct calculation via the formula ( [ singlebubble ] ) results in a nonphysical behavior of the scattering time at low polariton energies .
to find the correct way to go , we take into account all the bubble diagrams , as shown in fig .
[ fig3]c , and yield an analytic formula for the retarded self - energy in the form @xmath53 where @xmath54 is the result of summation of the bubble diagrams by the dyson equation , @xmath55 where @xmath56 . in the limit of low polariton energy and momenta , @xmath57 , @xmath58 , the scattering time is given by the imaginary part of the polariton self - energy : @xmath59 .
the polarization operator , representing a single - bubble diagram , at small @xmath60 can be written as @xmath61 where @xmath62 is a unity vector along the direction of @xmath63 .
further we evaluate the integrals and find @xmath64\right ) \\
\textrm{im}\,\pi^r(\omega,\textbf{k})&=&\pi_0\frac{\omega}{\sqrt{v_0 ^ 2k^2-\omega^2}}\theta[v_0 ^ 2k^2-\omega^2 ] , \label{imrepoloper } \end{aligned}\ ] ] where @xmath65.\end{aligned}\ ] ]
if we put @xmath66 in eq .
( [ selfretard ] ) and expand the bose distribution functions near @xmath67 , we find @xmath68 substituting eq .
( [ poloper1 ] ) and ( [ imrepoloper ] ) in eq .
( [ imselfen2 ] ) , we find that the imaginary part of the polariton self - energy reads @xmath69}\int\limits_1^\infty \frac{dx}{\sqrt{x-1}}\frac{1}{e^{xz-\mu / t}-1}. \label{imselfen3}\end{aligned}\ ] ] at small energies @xmath70 , when @xmath71 , the integral @xmath72 can be approximated as @xmath73 and the polariton scattering time acquires the following energy dependence : @xmath74 this dependence is the central result of our work .
if we compare our case of dirac - like dispersion ( at small @xmath36 ) with the case of usual parabolic dispersion of eps at @xmath75 , we immediately find that in the latter case @xmath76 .
that is , at @xmath57 , @xmath77 decreases slower than @xmath78 , and hence linear dispersion is more profitable from the point of view of efficient scattering towards the ground state .
we illustrate it by plotting the scattering lifetime in the vicinity of the ground state for two different temperatures in the case of linear and parabolic spectrum at fig .
[ fig_1 ] .
the chemical potential is @xmath79 mev , the effective mass for the parabolic spectrum is set to @xmath80 kg , the exciton binding energy - @xmath81 mev , bohr radius - @xmath82 nm , and fermi velocity for the linear case - @xmath83 m / s .
we can see that for both cryogenic and room temperatures , the scattering lifetimes are at least one order of magnitude smaller than for the parabolic case , which corresponds to far faster relaxation to the ground state .
it should be additionally stressed , that eps remain good quasiparticles in the vicinity of the dirac point , @xmath84 , since @xcite @xmath85
in this article , we used the feynman - dyson diagram approach to calculate polariton - polariton interaction self - energy in structures with linear dirac - like dispersion .
we explicitely demostrated that the polariton scattering time caused by the nonlinear interaction between particles is dramatically decreased in comparison with conventional two - dimensional planar structures with parabolic dispersion of polaritons .
it happens since the scattering lifetime of polaritons is proportional to the power @xmath86 of the energy rather than power @xmath87 which becomes significant at small energies .
the effect which we considered can , for instance , be used to construct polariton lasers with significantly reduced thresholds .
99 schneider c , rahimi - iman a , kim n y , fischer j , savenko i g , amthor m , lermer m , wolf a , worschech l , kulakovskii v d , shelykh i a , kamp m , reitzenstein s , forchel a , yamamoto y and hfling s 2013 _ nature ( london ) _ * 497 * 348 | using the feynman - dyson diagram technique , we study nonlinear polariton - polariton scattering in a two - dimensional micropillar - based optical superlattice with hexagonal symmetry . we demonstrate that both the emerging polariton chirality and the loop feynman diagrams up to infinite order should be strictly accounted for in the evaluation of the self - energy of the system .
further , we explicitly show that in such a design the time of polariton scattering towards the dirac points can be drastically decreased which can be used , for instance , in engineering novel classes of polariton lasers with substantially reduced thresholds .
november 2015 |
10 y. maeno , t. m. rice , and m. sigrist , phys . today . *
54 * , 42 ( 2001 ) . a. p.
mackenzie and y. maeno , rev .
phys . * 75 * , 657 ( 2003 ) .
y. maeno et al . , j. phys .
. jpn . * 81 * , 011009 ( 2012 ) .
y. maeno , h. hashimoto , k. yoshida , s. nishizaki , t. fujita , j. g. bednorz and f. lichtenberg , nature ( london ) * 372 * , 532 ( 1994 ) .
z. q. mao , y. maeno and h. fukazawa , materials research bulletin * 35 * , 1813 ( 2000 ) .
j. g. bednorz and k. a. muller , z. phys .
b * 64 * , 189 ( 1986 )
. t. m. rice and m. sigrist , j. phys .
condens . matter .
* 7 * , l643 ( 1995 ) . c. honerkamp and t. m. rice ,
phys . * 131 * , 159 ( 2003 ) . g. m. luke , y. fudamoto , k. m. kojima , m. i. larkin , j. merrin , b. nachumi , y. j. uemura , y. maeno , z. q. mao , y. mori , n. nakamura and m. sigrist , nature ( london ) * 394 * , 558 ( 1998 ) .
k. ishida , h. mukuda , y. kitaoka , k. asayama , z. q. mao , y. mori and y. maeno , nature ( london ) * 396 * , 658 ( 1998 ) .
k. d. nelson , z. q. mao , y. maeno and y. liu , science * 306 * , 1151 ( 2004 ) .
j. xia , y. maeno , p. t. beyersdorf , m. m. fejer , a kapitulnik , phys .
* 97 * , 167002 ( 2006 ) .
f. kidwingira , j. d. strand , d. j. van harlingen and y. maeno , science * 314 * , 1267 ( 2006 ) . c. bergemann , a. p. mackenzie ,
s. r. julian , d. forsythe and f. ohmichi , adv .
phys . * 52 * , 639 ( 2003 ) .
x. l. qi and s. c. zhang , rev .
* 83 * , 1057 ( 2011 ) .
s. d. sarma , c. nayak and s. tewari , phys .
b * 73 * , 220502(r)(2006 ) .
t. oguchi , phys .
b * 51 * , 1385 ( 1995 ) .
d. j. singh , phys .
b * 52 * , 1358 ( 1995 ) . i. i. mazin and d. j. singh , phys . rev
. lett . * 79 * , 723 ( 1997 ) .
i. i. mazin and d. j. singh , phys .
* 82 * , 4324 ( 1999 ) .
a. p. mackenzie , s. r. julian , a. j. diver , g. j. mcmullan , m. p. ray , g. g. lonzarich , y. maeno , s. nishizaki , and t. fujita , phys .
lett . * 76 * , 3786 ( 1996 ) . c. bergemann , s. r. julian , a. p. mackenzie , s. nishizaki and y. maeno , phys
84 * , 2662 ( 2000 ) .
a. damascelli , z. hussain and z. x. shen , rev .
phys . * 75 * , 473 ( 2003 ) d. h. lu , m. schmidt , t. r. cummins and s. schuppler , phys . rev
. lett . * 76 * , 4845 ( 1996 ) .
t. yokoya , a. chainani , t. takahashi , h. katayama - yoshida , m. kasai , and y. tokura , phys .
76 * , 3009 91996 ) ; t. yokoya , a. chainani , t. takahashi , h. ding , j. c. campuzano , h. katayama - yoshida , m .
kasai and y. tokura , phys .
b * 54 * , 13311 ( 1996 )
. a. v. puchkov , z. x. shen , t. kimura and y. tokura , phys .
b * 58 * , 13322(r ) ( 1998 ) . a. p. mackenzie , s. r. julian , g. g. lonzarich , y. maeno and t. fujita , phys .
lett . * 78 * , 2271 ( 1997 ) .
r. matzdorf , z. fang , ismail , jiandi zhang , t. kimura , y. tokura , k. terakura and e. w. plummer , science * 289 * , 746 ( 2000 ) .
a. damascelli , d. h. lu , k. m. shen , n. p. armitage , f. ronning , d. l. feng , c. kim and z .- x .
shen , phys .
* 85 * , 5194 ( 2000 ) .
k. m. shen , a. damascelli , d. h. lu , n. p. armitage , f. ronning , d. l. feng , c. kim , z .- x .
shen , d. j. singh , i. i. mazin , s. nakatsuji , z. q. mao , y. maeno , t. kimura and y. tokura , phys .
b * 64 * , 180502(r ) ( 2001 ) .
s. kashiwaya , h. kashiwaya , h. kambara , t. furuta , h. yaguchi , y. tanaka and y. maeno , phys . rev
. lett . * 107 * , 077003 ( 2011 ) .
h. iwasawa , y. aiura , t. saitoh , i. hase , s. i. ikeda , y. yoshida , h. bando , m. higashiguchi , y. miura , x. y. cui , k. shimada , h. namatame and m. taniguchi , phys .
b * 72 * , 104514 ( 2005 ) .
n. j. c. ingle , k. m. shen , f. baumberger , w. meevasana , d. h. lu , z .- x .
shen , a. damascelli , s. nakatsuji , z. q. mao , y. maeno , t. kimura and y. tokura , phys .
b * 72 * , 205114 ( 2005 ) .
t. e. kidd , t. valla , a.v .
fedorov , p. d. johnson , r. j. cava and m. k. haas , phys .
* 94 * , 107003 ( 2005 ) .
y. pennec , n. j. c. ingle , i. s. elfimov , e. varene , y. maeno , a. damascelli , and j. v. barth , phys .
* 101 * , 216103 ( 2008 ) .
g. d. liu et al . , rev .
instrum . * 79 * , 023105 ( 2008 ) .
m. p. seah et al .
, surface and interface analysis * 1 * , 2 ( 1979 ) . | high resolution angle - resolved photoemission measurements are carried out to systematically investigate the effect of cleaving temperature on the electronic structure and fermi surface of sr@xmath0ruo@xmath1 .
different from previous reports that high cleaving temperature can suppress surface fermi surface , we find that the surface fermi surface remains obvious and strong in sr@xmath0ruo@xmath1 cleaved at high temperature , even at room temperature .
this indicates that cleaving temperature is not a key effective factor in suppressing the surface bands . on the other hand , in the aged surface of sr@xmath0ruo@xmath1 that is cleaved and held for a long time ,
the bulk bands can be enhanced .
we have also carried out laser arpes measurements on sr@xmath0ruo@xmath1 by using vacuum ultra - violet laser ( photon energy at 6.994 ev ) and found an obvious enhancement of bulk bands even for samples cleaved at low temperature .
these information are important in realizing an effective approach in manipulating and detecting the surface and bulk electronic structure of sr@xmath0ruo@xmath1 . in particular
, the enhancement of bulk sensitivity , together with its super - high instrumental resolution of vuv laser arpes , will be advantageous in investigating fine electronic structure and superconducting properties of sr@xmath0ruo@xmath1 in the future .
sr@xmath0ruo@xmath1 is an unconventional superconductor that has attracted much attention@xcite since the discovery of superconductivity in 1994@xcite with a transition temperature t@xmath2 at 1.5 k@xcite .
so far , this is the only superconductor without copper that has a layered perovskite crystal structure identical to that of copper - oxide ( la , sr)@xmath0cuo@xmath1 high temperature superconductor@xcite .
theoretical expectations@xcite and experimental measurements @xcite indicate that sr@xmath0ruo@xmath1 is an unconventional superconductor with a spin - triplet pairing and a possible p@xmath3+ip@xmath4 superconducting order parameter .
its normal state above t@xmath2 exhibits fermi - liquid like behaviors@xcite .
most recently , sr@xmath0ruo@xmath1 is classified as a topological superconductor which is closely related with majorana fermions and non - abelian statistics@xcite .
the electronic structure and physical properties of sr@xmath0ruo@xmath1 are dictated by the ru 4d orbitals@xcite .
the low - lying electronic states originate from the hybridization between the ru 4@xmath5@xmath6 , 4@xmath5@xmath7 , 4@xmath5@xmath8 orbitals and the oxygen 2@xmath9 orbitals .
the quasi - two - dimensional nature of the crystal structure gives rise to nearly two - dimensional fermi surface sheets . among them ,
the @xmath5@xmath6 orbital produces an almost cylindrical fermi surface sheet around the center of the brillouin zone ( usually denoted as @xmath10 sheet , fig .
1b ) .
on the other hand , the nearly one - dimensional nature of the @xmath5@xmath7 and @xmath5@xmath8 orbitals produces nearly planar fermi surface sheets which run perpendicular to x - axis and y - axis , respectively .
the hybridization between the @xmath5@xmath7 and @xmath5@xmath8 orbital - generated bands gives rise to a hole - like fermi surface sheet enclosed around m(@xmath11,@xmath11 ) point ( denoted as @xmath12 sheet ) and an electron - like fermi surface sheet around the @xmath13(0,0 ) ( denoted as @xmath14 sheet , fig .
1b ) .
such an electronic structure of sr@xmath0ruo@xmath1 has been well described by band structure calculations@xcite and experimentally confirmed by the quantum oscillation measurements@xcite .
angle - resolved photoemission spectroscopy ( arpes ) is a powerful tool to directly determine the electronic structure and fermi surface of materials@xcite .
however , the arpes measurements on sr@xmath0ruo@xmath1 experienced a complicated and controversial process .
earlier arpes results on sr@xmath0ruo@xmath1@xcite were not compatible with the band structure calculations and quantum oscillation measurements , rendering people doubt whether arpes , as a surface sensitive technique , is suitable for studying bulk electronic properties@xcite .
it turned out that sr@xmath0ruo@xmath1 exhibits a @xmath15@xmath16@xmath15 surface reconstruction arising from the rotation of the ruo@xmath17 octahedra in the surface layer@xcite .
this surface reconstruction made it possible to resolve the controversy surrounding the electronic structure of sr@xmath0ruo@xmath1@xcite .
the occurrence of surface reconstruction produces two effects .
first , in addition to the original one set of bulk fermi surface sheets ( @xmath18 , @xmath19 and @xmath20 ) , it produces an additional set of surface fermi surface sheets ( @xmath21 , @xmath22 and @xmath23 ) .
second , the @xmath15@xmath16@xmath15 surface reconstruction produces two additional sets of fermi surface sheets : surface shadow bands ( s@xmath21 , s@xmath22 and s@xmath23 ) formed by inverting the original surface fermi surface sheets ( @xmath21 , @xmath22 and @xmath23 ) with respect to the ( @xmath11,0)@xmath24(0,@xmath11 ) line , and bulk shadow bands ( s@xmath18 , s@xmath19 and s@xmath20 ) formed also by inverting the original bulk fermi surface sheets ( @xmath18 , @xmath19 and @xmath20 ) with respect to the ( @xmath11,0)@xmath24(0,@xmath11 ) line .
considering the four sets of fermi surface sheets shown in fig .
1d , the bulk fermi surface picked up from the arpes measurements@xcite becomes consistent with the band structure calculations and quantum oscillation measurements .
ruo@xmath1 . repeated three - slab of sr@xmath0ruo@xmath1
is used with the middle slab representing the bulk electronic structure while the top and bottom ones representing surface .
( a ) .
calculated band structure without considering any surface constructions .
( b ) .
fermi surface corresponding to band structure of ( a ) in one quadrant .
( c ) . band structure with the top and bottom surface ru - o octahedra rotating by 6 degrees .
( d ) fermi surface corresponding to band structure of ( c ) .
note that here because of the @xmath15@xmath16@xmath15 surface reconstruction , the real brillouin zone is reduced by half . ]
the understanding of the surface crystal structure and surface electronic structure of sr@xmath0ruo@xmath1 is crucial for both fundamental studies of sr@xmath0ruo@xmath1 and its potential applications as a topological material in quantum computing@xcite .
first , as mentioned above , the uncovering of surface reconstruction@xcite is crucial in understanding the arpes - measured fermi surface of sr@xmath0ruo@xmath1@xcite ; second , it remains unclear whether sr@xmath0ruo@xmath1 can support surface magnetism , as expected from theory@xcite and examined by arpes measrement@xcite .
third , the detection of possible edge state@xcite associated with sr@xmath0ruo@xmath1 as a topological superconductor asks for a comprehensive understanding and control of its surface properties . in many cases to investigate the canonical fermi liquid behavior or many - body effects in sr@xmath0ruo@xmath1 , it is preferable to suppress the surface state to make the bulk state dominant@xcite .
one way used is to cleave the sr@xmath0ruo@xmath1 sample at high temperature ( such as 180 k as in @xcite ) which was found to cause suppression of the surface fermi surface sheets .
however , whether the cleaving temperature is a key controlling factor and the mechanism of such a surface band suppression remain unclear@xcite . in this paper , we report systematic investigation on the influence of cleaving temperature on the electronic structure and fermi surface of sr@xmath0ruo@xmath1 .
we find that the surface fermi surface remains obvious and strong in samples cleaved at high temperature , even at room temperature .
this is different from the previous claim@xcite .
our results indicate that cleaving temperature is not a key effective factor to suppress the surface bands . on the other hand , in the aged surface of sr@xmath0ruo@xmath1 cleaved and held for a long time , the bulk bands can be strongly enhanced .
we also carried out laser arpes measurements on sr@xmath0ruo@xmath1 by using vacuum ultra - violet laser ( photon energy at 6.994 ev)@xcite and found an obvious enhancement of bulk bands even for samples cleaved at low temperature .
these results are important in finding out an effective way to manipulate and detect the surface and bulk electronic structure of sr@xmath0ruo@xmath1 .
angle - resolved photoemission measurements on sr@xmath0ruo@xmath1 were carried out on our lab arpes system equipped with a r4000 electron energy analyzer together with both a helium discharge lamp ( @xmath25=21.218ev ) and a vacuum ultraviolet ( vuv ) laser ( @xmath25=6.994 ev ) as light sources@xcite .
the energy resolution was set at 10 mev and 2 mev for helium lamp arpes and vuv laser arpes measurements , respectively .
the angular resolution is @xmath260.3@xmath27 .
the fermi level is determined by referencing the fermi edge of a clean polycrystalline gold which is electronically connected to the sample .
the sr@xmath0ruo@xmath1 sample was grown by traveling solvent floating zone method and has a superconducting transition at t@xmath2 = 1.5 k with a sharp transition width of @xmath260.1 k. all samples were cleaved _ in situ _ and measured in vacuum with a base pressure better than 5@xmath1610@xmath28 torr .
ruo@xmath1 at 20 k. ( a ) .
the sample was cleaved at 20 k and then immediately measured at 20 k ; ( b ) .
the sample was first cleaved at 180 k and then cooled down immediately to low temperature and measured at 20 k ; ( c ) .
the sample was first cleaved at 300 k and then measured immediately at 20 k ; ( d ) .
the sample was first cleaved at 20 k and measured after nearly 17 hours at 20 k. during this period , the sample was slowly warmed up to 100 k and then cooled down .
the red lines in ( c ) shows the location of momentum cuts for the bands in fig .
3 . ]
ruo@xmath1 along @xmath13(0,0)-m(@xmath11,@xmath11)-x(0,@xmath11)-@xmath13(0,0 ) high symmetry cuts , as shown by red lines in fig .
2c .
( a ) , ( b ) , ( c ) and ( d ) show band structure for the sr@xmath0ruo@xmath1 samples cleaved at 20 k , 180 k , 300 k and at 20 k but aged , respectively
. the corresponding momentum distribution curves ( mdcs ) are shown in ( e ) .
all the samples were measured at 20 k. ] figure 2 shows arpes - measured fermi surface of sr@xmath0ruo@xmath1 samples treated with different conditions ; the corresponding energy bands along several high symmetry lines are shown in fig .
3 . there are two signatures to qualitatively compare the contributions of bulk and surface components .
one is the relative intensities of the bulk @xmath20 band and surface @xmath23 band . as seen in fig .
1d , the bulk @xmath20 band intersects with the @xmath13(0,0)@xmath24x(0,@xmath11 ) line while the surface @xmath23 band intersects with the x(0,@xmath11)@xmath24m(@xmath11,@xmath11 ) line ; their relative intensity ratio thus provides a measure on the contributions of the bulk and surface bands .
another signature is the relative intensities of the shadow bands and their original bands .
because the shadow bands from the surface fermi surface sheets are expected to be much stronger than those from the bulk fermi surface sheets , their presence and relative intensity can give another measure on the contributions of the bulk and surface bands .
it is clear that the surface components are dominant in the measured fermi surface of the sr@xmath0ruo@xmath1 sample freshly cleaved at low temperature ( 20 k , fig .
2a and fig .
3a ) , judging from both the relative intensity between the @xmath20 band and the @xmath23 band ( fig .
3e ) , and the relative intensity between the s@xmath29 band(s ) and the @xmath20 band ( fig .
3e ) .
the surface components remain dominant for the sr@xmath0ruo@xmath1 samples cleaved at high temperatures ( 180 k and 300 k , figs . 2 and 3 ) . the surface @xmath23 remains strong and the bulk @xmath20 remains weak for the samples cleaved at high temperatures ( fig .
3e ) . in addition
, the existence of s@xmath29 is clear .
these indicate that cleaving sr@xmath0ruo@xmath1 at high temperatures , even at room temperature , does not result in significant suppression of the surface bands , contrary to the previous report@xcite .
therefore , the cleaving temperature is not a key factor to effectively suppress the surface bands ; its effect in the previous work@xcite might be due to other factors such as a particular gas component in the measurement chamber . for samples cleaved at high temperatures ,
the overall signal gets weak .
moreover , the measured fermi surface sheets become blurred with some fine features clearly observed for low temperature cleaved sample ( fig .
2a ) not resolvable for high temperature cleaved ones ( figs .
2b and 2c ) .
these are additional disadvantages caused by high temperature cleaving .
sample ageing , i.e. , holding the freshly cleaved sample for a long period of time , can help suppress the surface bands . as shown in fig .
2d and figs .
3d and 3e , for the sr@xmath0ruo@xmath1 sample cleaved at 20 k and held in ultra - high vacuum for 17 hours , the bulk bands are obviously enhanced .
this can be judged by comparing the relative intensity of the bulk @xmath20 band and surface @xmath23 band in the mdc curve ( fig .
3e ) for the aged sample which gets larger than that of a freshly cleaved sample .
further increase of the surface holding time may further enhance the bulk bands although we note that the overall signal is also significantly reduced in the aged samples .
it is known that arpes is a surface sensitive technique . for the helium discharge lamp with a photon energy at 21.2 ev ,
the estimated photoelectron escape depth is on the order of 5@xmath2610 @xmath30 that is just the top one or two layers near surface@xcite . as the photon energy decreases
, the photoelectron escape depth is expected to increase to nearly 30 @xmath30 for the vuv laser with a photon energy of 6.994 ev@xcite .
this prompts us to investigate the electronic structure of sr@xmath0ruo@xmath1 by utilizing our vuv laser arpes .
fig.4 shows the fermi surface measured with the vuv laser , and the band structure along several momentum cuts .
it demonstrates that laser arpes is a powerful tool to study sr@xmath0ruo@xmath1 .
first , our vuv laser arpes has the capability to reach the x(@xmath11,0 ) point . as shown in fig .
4 , in the measured momentum range , we clearly resolve @xmath12 fermi surface sheet and @xmath10 fermi surface sheet .
second , the maximum intensity of the bulk @xmath20 sheet is comparable to that of the surface @xmath20 sheet ( figs .
4c3 and 4c4 ) .
compared with fig .
3 , it is clear that , even for the sample cleaved at 20 k , the bulk bands are strongly enhanced in our vuv laser - based arpes measurements .
ruo@xmath1 measured by using @xmath31=6.994 ev laser light source .
the sample was cleaved at 20 k and measured at 20 k. ( a ) .
fermi surface mapping ; ( b1-b4 ) show band structure for the four representative momentum cuts shown in ( a ) .
( c1-c4 ) shows the corresponding mdcs at the fermi level . ] in summary , we have systematically investigated various approaches to enhance the bulk components of the electronic structure in sr@xmath0ruo@xmath1 .
we find that the cleaving temperature is not a key factor in suppressing the surface bands , as claimed before .
sample aging can help enhance bulk bands .
vuv laser arpes is an effective way to enhance the bulk bands due to the increase of photoelectron escape depth .
these information are important in manipulating and detecting the surface properties of sr@xmath0ruo@xmath1 .
in particular , the enhancement of bulk sensitivity , together with its super - high instrumental resolution of vuv laser arpes , will be advantageous in investigating fine electronic structure and superconducting properties of sr@xmath0ruo@xmath1 in the future .
* acknowledgments * + we thank financial support from the nsfc ( 10734120 and 91021006 ) and the most of china ( 973 program no : 2011cb921703 and 2011cb605903 ) . @xmath32corresponding author : [email protected] |
in this work we intend to study the effects of an electric field on the collapse of a massive star .
we perform direct relativistic simulations assuming spherical symmetry and integrating the einstein - maxwell equations .
we studied the collapse of several stars with different values of the total electric charge and determined the limits at which the collapse is avoided by the effect of the repulsive electric field .
the electric charge is carried by the particles that compose the collapsing fluid . the electromagnetic field and the internal energy of the gas contributes to the total mass - energy of the star , so it is not clear whether it is possible to overcharge a collapsing configuration .
the analysis and conclusions drawn for collapsing charged shells can not be directly extrapolated for this case , and a complete relativistic calculation performed in a self - consistent way is needed to know the outcome of the charged collapse . in particular , we found the formation of a bubble that was not predicted before .
although its theoretical explanation is quite natural , the formation of the bubble depends on the initial conditions and its evolution is far from obvious due to the non - linearities of the einstein - maxwell equations .
numerical solutions for the problem of charged stellar collapse were found only recently @xcite , .
it is commonly assumed that the stars are nearly neutrally charged due to a mechanism of selective accretion of charges from the surrounding interstellar medium .
therefore , if a star is initially positively ( negatively ) charged , accretion of negative ( positive ) charges from the surrounding gas will tend to neutralize the total net charge .
the same reasoning is applied to black holes @xcite .
we observe that the effect of selective accretion were never fully calculated to our knowledge . on the other hand ,
an star can be electrically neutral while possessing a huge internal electric field .
the internal electric field of an astronomical object can be very large in some special scenarios @xcite , @xcite , @xcite , @xcite .
in particular , the charge separation inside a spherical compact star , in hydrostatic equilibrium , can be very large when one of the plasma components is a degenerate gas while the other is a maxwell - boltzmann gas , i.e. , like the gas of degenerated electrons and the gas of nuclei in a white dwarf @xcite .
the calculations presented in this paper corresponds to the case in which the star posses a net electric charge . however the results can be taken as an approximation to the more complicated problem of an star with total zero charge and with a non zero internal electric field .
we leave this point aside and for clarity we will concentrate here on the dynamics of an star with a total positive charge .
we ask whether an electrically charged star can collapse to form a charged black hole .
would the coulomb repulsion avoid the collapse of the star ?
, and moreover , are there physical differences with the physics of uncharged collapse ? .
moreover , from pure analytic analysis is not clear if it is possible to form an overcharged black hole ( with @xmath3 ) .
the dynamics of a collapsing star could be quite different from the collapse of a charged shell onto an already formed reissner - nordstrm black hole , mainly due to backreaction effects .
in the present paper , we will try to give the answer to these questions .
extremely charged black holes ( @xmath4 ) represent an extreme limit in the context of the cosmic censorship hypothesis , since bodies with charge equal or higher than extremal are undressed by event horizons and constitutes naked singularities @xcite , @xcite . on the other hand , the formation of reissner - nordstrm ( rn ) black holes , is in connection with the third law of black hole thermodynamics @xcite , @xcite , @xcite , because an extreme rn black hole has zero temperature .
so , there is an interest on the formation of charged black holes from a pure theoretical point of view . in this paper
we are not concerned with the mechanism that produce charge or an internal electric field .
neither magnetic fields nor rotation are taken into account , although they could increase the magnitude of the electric fields . without loss of generality ,
the density of charge was chosen proportional to the rest mass density .
the interior solutions found can be matched smoothly with the reissner - nordstrm exact solution for the vacuum space - time @xcite . therefore , in the cases in which the star collapses the result is the formation of a reissner - nordstrm black hole .
the paper is organized as follows . in section
ii , we discuss the problem of charged collapse analyzing the order of magnitude of the physical quantities . in the section iii are described the general relativistic equations that govern the dynamics of the stellar core collapse . in section iv , are described the numerical techniques used to integrate the general relativistic equations and some caveats on its applications .
section v brings a discussion and interpretation of the numerical results . in the subsection
a , we show the calculation of the maximum mass formula for newtonian charged stars , and in subsection b we calculate the optical depth for the neutrinos emitted during the collapse .
we also discuss the implications for core collapse supernova .
we end in section vi presenting some final remarks .
in this section we show some order of magnitude estimates , in order to put in a clear perspective the problem of a self - gravitating charged fluid sphere .
the calculations in the present section are valid only in the non - relativistic regime .
the formation of a charged star , is possible when the gravitational attraction overwhelms the electrostatic repulsion on each single particle of the gas that is collapsing , i.e. , @xmath5 or equivalently @xmath6 where @xmath7 and @xmath8 are the charge and mass of the star and @xmath9 and @xmath10 are the charge and mass of the particles . in this section
we use a newtonian approach , the full relativistic case could be different .
assume a star that has a total charge to mass ratio @xmath11 .
according to eq .
( [ cargas ] ) any particle with a charge to mass ratio @xmath12 can be added to the star if @xmath13 where the equality gives the maximal charge to mass ratio @xmath14 for example , a proton has a charge to mass ratio . using the eq .
( [ limit1 ] ) the proton can be added to the charged star if @xmath15 thus , a star with higher charge to mass ratio can not be assembled from protons alone . on the other hand , if the infalling particles are dust particles with a larger charge to mass ratio , the maximum limit for the charge to mass ratio of the charged star can be much higher .
an example of this is a self - gravitating ball of fei nucleus , of mass @xmath16 and charge @xmath17 . as the charge to mass
ratio for fei is @xmath18 , the eq .
( [ limit1 ] ) for fe gives @xmath19 and we see that in this case @xmath20 ( greater than for protons alone , see eq .
( [ qmprotons ] ) above ) . for `` charged dust '' formed by particles with @xmath21 , a self - gravitating sphere with mass @xmath22 and charge @xmath23 , can be in hydrostatic equilibrium if : @xmath24
so : @xmath25 .
let us estimate the amount of charge needed to have an extremely charged @xmath26 star . in this case @xmath27
, so @xmath28 therefore , it is needed to have an excess of @xmath29 in the star , where @xmath30 is the charge of one proton .
there is roughly @xmath31 baryons in a @xmath32 star .
so , there must be one charged particle on @xmath33 neutral ones in order to have a maximally charged compact object .
this is compatible with the limits given above .
therefore , in the extremal case , there are a small charge imbalance equal to : @xmath34 in the star . a tiny amount of charge from a microscopic point of view , but with a huge total sum ( see ref .
@xcite ) .
it is possible to make the objection that the nucleus would disintegrate , or suffer nuclear fission , before assembling a charged star .
however the energy locally available ( in the center of mass of the particles ) from the collapse is not enough to unbind a nucleus ( bounded by the strong force ) .
this was calculated for a compact charged neutron star .
an exception is a nearly extremal compact object , in this case is energetically favorable for the nucleus to disintegrate @xcite .
the conclusion is that we can approach to the formation of a nearly extreme " object , although it is not possible to reach the extremal value , the particles will fission first ( see @xcite for details ) . studying theories with extra dimensions ,
mosquera - cuesta and co - workers @xcite found that particles located in the brane can leak out to the bulk space .
it results that electrons can leak out more easily than baryons , producing a charge asymmetry that can be very large in very old stellar systems .
this is suitable for type ii supernova progenitors and neutron stars @xcite .
the mechanism produce a charge imbalance of @xcite : @xmath35 this is several orders of magnitude higher than the needs for producing an extremal star .
therefore , charged old stars must be considered and studied as a theoretical possibility .
we want to solve the einstein - maxwell equations ( see ref .
@xcite ) : @xmath36 where @xmath37 is the ricci tensor , @xmath38 is the metric tensor , @xmath39 is the scalar curvature , and @xmath40 is the energy momentum - tensor , composed by two parts : @xmath41 the perfect fluid energy - momentum tensor @xmath42 and the maxwell energy - momentum tensor @xmath43 .
the electromagnetic energy - momentum tensor is given by : @xmath44 here @xmath45 is the maxwell electromagnetic tensor , which can be written in terms of a potential @xmath46 : @xmath47 in the last equation and in the rest of the paper , we use semi - colon to denote covariant derivative . the perfect fluid energy - momentum tensor is : @xmath48 where @xmath49 is the density of mass - energy and @xmath50 is the pressure , given in @xmath51}$ ] , and @xmath52 is the @xmath53-velocity of the observers co - moving with the fluid @xcite , @xcite , @xcite . the electromagnetic tensor must satisfy the maxwell equations @xcite , @xcite , @xcite : @xmath54 from eq .
( [ vecpot ] ) and using the bianchi identities , we have : @xmath55}=0\,.\ ] ] we will use the common split of the energy : @xmath56 where @xmath57 is the rest mass density , and @xmath58 is the internal energy density @xcite , @xcite . the fluid must satisfy the energy - momentum conservation equation : @xmath59 and the baryon conservation equation : @xmath60 we will write the einstein - maxwell equations in coordinates and gauge choice appropriated for its numeric implementation @xcite . considering a spherical star , the line - element in co - moving coordinates
is given by @xmath61 the 4-velocity of an observer co - moving with the fluid is @xmath62\,,$ ] and satisfies @xmath63 , where @xmath64 is the speed of light .
the coordinate @xmath65 was gauged to be the rest mass enclosed by co - moving observers standing on spherical layers of the star .
each layer has its own constant value of @xmath65 .
in particular , observers at the surface of the star will be designated with a coordinate @xmath66 .
we define component of the co - moving observer s 4-velocity in schwarzschild coordinates , see ref .
@xcite for details . ]
@xmath67 so , the equation of motion , obtained from the einstein - maxwell equations @xcite is @xmath68\,,\end{aligned}\ ] ] where @xmath69 is the relativistic specific enthalpy , @xmath57 is the rest mass density , @xmath70 is the internal energy per unit mass , @xmath50 is the pressure , @xmath71 is the gravitational constant , @xmath72 is the total charge integrated from the origin of coordinates up to the spherical layer with coordinate @xmath65 , and @xmath73 is the total mass - energy defined below .
we use the notation @xmath74 , with @xmath75 , throughout the paper .
we note that the eq .
( [ motion ] ) resembles a newtonian equation of motion plus relativistic corrections . in the eq .
( [ motion ] ) the last two terms are pure relativistic , and the full equation is equivalent to its newtonian counterpart when @xmath76 . in the eq .
( [ motion ] ) to obtain the tolman - oppenheimer - volkov equation for charged stars @xcite . ]
the factor @xmath77 is a generalization of the special - relativistic @xmath78 factor for general relativity and is given by @xmath79 this factor also verifies the equality @xmath80 ( see refs .
@xcite , @xcite , @xcite ) .
the equation for the metric component @xmath81 is obtained from the energy - momentum conservation equation @xmath82 , and is given by @xcite @xmath83\,.\ ] ] the other independent component @xmath84 gives @xmath85 which is identical to the non - relativistic adiabatic energy conservation equation , and constitutes the first law of thermodynamics .
we observe that the do not contain any electromagnetic term .
this is due to the symmetry of the problem , although an electromagnetic term arise in the total energy ( see the eq .
[ totalmass ] below ) . from the conservation of the number of baryons
, we obtain the equation for the metric component @xmath86 @xcite , @xmath87 the lagrangian coordinate @xmath65 can be chosen to be @xcite @xmath88 this is the total rest mass enclosed by a sphere of circumference @xmath89 .
thus , the collapsing ball is divided in layers of constant rest mass @xmath65 and each co - moving observer is at rest in each of this layers .
the equations for the total mass and the mass conservation are , respectively @xcite : @xmath90 @xmath91 @xmath92 the charge 4-current @xmath93 is the product of a scalar electric charge density @xmath94 times @xmath95 ; @xmath96 .
it can be shown that the charge and the rest mass are conserved in shells co - moving with the fluid @xcite , @xcite .
so , the charge increment between layers can be written as being proportional to @xmath97 , @xmath98 where .
we observe that eq .
( [ difcharge ] ) is a non - linear equation , since @xmath77 depends on the integral value of the charge @xmath72 ( see eq .
[ gamma ] ) , so the set of equations above must be solved self - consistently .
the equations ( [ motion])-([difcharge ] ) above are solved numerically with the following boundary conditions : @xmath99 where @xmath100 is the mass coordinate at the surface of the star . the boundary condition ( [ bcp ] )
can be derived from the matching condition between the interior and the exterior solution @xcite .
equation ( [ bca ] ) express our coordinate freedom to choose the time synchronized with a co - moving observer moving with the surface of the star .
( [ bmu ] ) means that the ball is initially at rest ( at the initial time @xmath101 ) , and eq .
( [ bcr ] ) is the impenetrability condition at the origin of coordinates ( this condition can be stated equivalently as @xmath102 at @xmath103 ) . in the next section
we will present the initial conditions . in this work
we choose @xmath104 for simplicity and without loss of generality .
the equations given in the section above are written in a form closely paralleling the equations of may and white @xcite , @xcite , for the non - charged case @xcite , @xcite , @xcite .
we built the numerical code to integrate the eqs .
( [ motion])-([difcharge ] ) . the eq .
( [ eqaw ] ) is re - written as @xcite : @xmath105\ , .
\end{aligned}\ ] ] this equation is integrated numerically from the surface of the star toward the center , using the boundary condition , where @xmath106 and @xmath107 are given at the surface of the star ( with coordinate @xmath100 ) , note that we already choose @xmath108 ( see eq .
[ bca ] ) . the uses a leapfrog method plus a predictor corrector step , and iterates with a crank - nicholson algorithm .
the method is second order accurate in time and space .
we use a numerical viscosity to resolve the strong shock waves formed @xcite .
it is assumed a charge density proportional to the rest mass density through the star : @xmath109 . from eqs .
( [ masscoord ] ) and ( [ difcharge ] ) we see that @xmath110 , @xmath111 .
we assumed a polytropic equation of state , , where @xmath112 is the adiabatic coefficient .
shocks see @xcite , pag 128 . ]
the star has initially a uniform mass density , @xmath113 and a uniform distribution of electric charge density , in all the models studied .
the initial specific internal energy @xmath70 is varied in the different simulations .
the initial setup imitates a massive star of @xmath114 that had exhausted its nuclear fuel .
its fate depends on the internal energy and charge of the initial configuration .
this is explored with our simulations .
the initial radius of the star is @xmath115 cm , in all the simulations .
also we consider that the electric field is always much lower than the schwinger field limit for pair creation ( @xmath116 volt / cm , in _ vacuo _ and in flat space - time @xcite ) .
for example , in the case of extremal charge @xmath117 , the total charge will be @xmath118 stat - coulomb . in this case , the maximum electric field strength is @xmath119 stat - volt / cm or @xmath120 volt / cm .
however , during the collapse the field can strengthen and take higher values , so quantum effects would be taken into account in a more realistic approach .
we are not considering magnetic fields in the present simulations .
, with @xmath121 the non - radial component of the 3-velocity of the fluid . in the present case , @xmath122 . ]
the simulations converge to the same solution when the resolution or the viscosity parameter are changed .
the parameters of the simulation were chosen for accuracy and efficiency when running on a single processor .
we performed the simulations using an intel pentium iv 2.6 ghz processor ( 32 bits ) and compiled the code with v6.1 @xcite for 32 bits ; also we used an amd athlon fx 2.6 ghz processor ( 64 bits ) and compiled with @xcite for 64 bits , running under linux operating system .
the code also runs under the windows operating system and compile with the fortran 90 power station .
we did nt find essential differences between the different runs with the exception of the speed of the simulation , which is faster with the amd processor .
in addition we set : a ) the number of integration points : 1000 ; b ) viscosity parameter : @xmath123 ; c ) initial time - step : @xmath124sec .
we also performed tests simulations with 300 and 1500 points to be sure of the physical results , and tested the code varying the viscosity over a wide range of parameters . in the present simulations we tried to minimized the effects of the viscosity while keeping the ability of the code to resolve shock fronts .
the percentage of numerical error in the hamiltonian constraint , expressed by the conservation of the total mass- , is given by : @xmath125 where is the exact value of the total mass energy ( constant in time ) and is the numerical value .
the momentum constraint is build in " the code and exactly conserved . in the present simulations : @xmath126 when the ah forms , using @xmath127 points , and it is @xmath128 with @xmath129 points , with no essential qualitative differences between the solutions .
this values of @xmath130 are for the worst case of a large binding energy and with very strong shocks , so it can be considered as the superior bound of the error .
after crossing the ah the error grows to : @xmath131 ( using @xmath127 points ) , and the simulation is stopped .
thus in every simulation we set a maximum error tolerance of @xmath132 .
we checked that using a higher number of points the errors are reduced .
we emphasize that the numerical errors ( mainly @xmath130 , not the roundoff errors ) and the intrinsic difficulty on the simulation is the formation of the black hole region , because near it several numerical operators blows up . for each layer of matter with lagrangian coordinate @xmath65
it is possible to define an internal and an external horizon , @xmath133 which can be interpreted as a generalization of the schwarzschild radius for charged stars . during the simulation an apparent horizon forms when some layer of matter cross its external horizon , i.e. when : @xmath134 @xcite .
in addition to the formation of the apparent horizon , in some cases a coordinate singularity develops , i.e. : @xmath135 and @xmath136 ( for some coordinate value @xmath137 ) . note that at the coordinate @xmath103 is located the physical timelike singularity of the vacuum reissner - nordstrm space - time .
the simulation must be ended before any observer @xmath65 can meet its internal horizon @xmath138 .
hence , we can not say if the star collapses directly to a singularity or if it pass through a wormhole to another asymptotically flat universe . in this paper
, we leave this question open .
the formation of an apparent horizon ( ah ) is a sufficient condition ( although not necessary ) for the formation of an event horizon and a black hole region @xcite .
moreover , the interior solution matches with the exterior reissner - nordstrm spacetime @xcite . in this sense
, we can say that the outcome of the simulations in which an ah forms is the formation of a `` reissner - nordstrm black hole '' .
of course , if not all the matter collapse to the ah the result at the moment the simulation is stopped will be an electric black hole plus a surrounding gas with a vacuum reissner - nordstrm exterior spacetime .
we observe that there must exist other formulations of the problem in which it is possible to get closer to the singularity while avoiding the ah .
we will explore this subject in future works .
we performed a set of simulations on the collapse of a @xmath139 star varying the initial conditions .
we found that for certain cases a shell - like structure of higher mass density and charge is formed in the most external part of the star .
this shell enlarge with time until it reaches the center of the star .
the table i summarize the results of the simulations performed . in each model studied are different amounts of total electric charge ; initial internal energy or binding energy .
the fate of each model depends on this conditions .
the formation of the shell is sensible to the internal energy , or the binding energy , of the star ( see table i ) . when stronger is the bound of the star higher is the wall of the shell . in the fig .
( [ fig : rho3048 ] ) , we show the profiles of the mass density _ versus _ the lagrangian mass @xmath65 for the model @xmath140 of the table i , in this case the shell forms only mildly .
however , there is a strong shock wave formed . the fig .
( [ fig : speed3048 ] ) shows the velocity profiles , on which it is seen a strong shock propagating outwards .
we observe that although the shock is propagating outwards the sign of the fluid velocity is negative everywhere indicating that the star is always collapsing .
there is an exception in which the shock velocity acquires a positive sign , although it lasts a very brief period of time .
this is shown on the inset of the fig .
( [ fig : speed3048 ] ) : the shell is reaching the coordinate origin , and an instant later it impinges the center .
after that a shock wave is formed and start to propagate outwards . during a brief lapse of time the velocity is positive , although the shock is not strong enough to stop the total collapse .
the figure ( [ fig : a3048 ] ) contains the snapshots of the metric component @xmath141 : it can be seen that concordant with the shell impinging the center , @xmath142 acquires a value greater than one , indicating a blueshift with respect to the infalling matter .
the inset of the fig .
( [ fig : a3048 ] ) shows the details of the jumps produced on the value of @xmath141 at the shell and at the shock positions , as well as the mentioned blueshift .
figure ( [ fig : rho3029 ] ) contains the snapshots of the mass density for the model @xmath143 of the table i. in this case there is an strong shell formed . on this figure
we can see that the shell propagates inwards , to lower values of the lagrangian coordinate @xmath65 , until it impinges the center . at this moment , and place , it is formed an strong shock that starts its outward propagation .
the inset box of the figure ( [ fig : rho3029 ] ) is a zoom of the dotted box , showing the details of the collapsing shell and the shock formed .
( [ fig : r3029 ] ) shows the verification that the collapse produce a black hole . in spherical symmetry the apparent horizon
is formed simply when an observer cross his / her own schwarzschild radius . the inset box of the fig .
( [ fig : r3029 ] ) shows the radius profile at the moment that the curve touch the schwarzschild radius profile at a point .
that point indicates an observer crossing the schwarzchild radius and is the of the apparent horizon .
after then the apparent horizon will evolve approaching the event horizon at infinite coordinate time or after complete collapse ( not calculated ) .
figure ( [ fig : speed3029 ] ) , shows the velocity profiles for model 11 , there is an strong - shell and strong - shock in this case .
however , the velocity has a negative sign and the matter collapse directly to a black hole .
figure ( [ fig : speed3029zoom ] ) shows the speed profile at the moment of time in which the shell impinges the center , and the formation and propagation of the shock wave .
the inset box shows a close up of the details .
there is a short lapse of time in which the shock has positive velocity . the fig .
( [ fig : a3029 ] ) shows the metric coefficient @xmath141 going to zero . in this case @xmath142 never becomes greater than one , and the inner matter is always redshifted respect to observers standing at the surface or outside the star .
figure ( [ fig : rho00 ] ) shows the density snapshots for the collapse of an uncharged star ( model @xmath144 ) .
it is possible to compare this simulation with the former simulations , we see that for the charged collapse the density profiles are flatter all the way through the black hole formation .
figure ( [ fig : speed00 ] ) shows the velocity profiles ( compare it with the figs .
[ fig : speed3048 ] and [ fig : speed3029 ] ) .
the figure ( [ fig : a00 ] ) shows the collapse of @xmath141 ( compare it with the figs .
[ fig : a3048 ] and [ fig : a3029 ] ) . figure ( [ fig : r00 ] ) shows the profiles of the metric function @xmath39 , the bold line is the schwarzschild radius at the moment the apparent horizon forms .
the inset box shows the detail of the radius profile touching the schwarzschild profile at one point , signaling the formation of an apparent horizon at the point indicated with the vertical arrow , at the time indicated on the legend box .
the simulation is continued beyond that event , as seen in this figure .
it is possible to compare the profile of the mass - energy function for the charged collapse with very strong shocks ( model 3 ) fig .
( [ fig : massa3048 ] ) , with the case of uncharged collapse ( model 13 ) fig .
( [ fig : massa00 ] ) .
the inset box of the fig .
( [ fig : massa3048 ] ) shows a zoom of the right end of the curves which indicates the level of energy conservation in the simulation .
all curves must reach the same point for perfect energy conservation . at the moment of the apparent horizon formation
the mass - energy is conserved with a precision of @xmath145 , which is an excellent value taking into account that this is the worst case .
after the apparent horizon formation , the good conserving property is lost as in any relativistic code , and the simulation is ended with a @xmath146 error .
we emphasize that this is the physical quantity we use to check the convergence properties of our code , since the momentum constraint is conserved exactly by the algorithm . as was said before the simulations
were performed using @xmath127 points , and the conservation can be better using a larger number of points .
( [ fig : massa00 ] ) represents the energy conservation for the case of uncharged collapse .
the precision is roughly the same as in the former case , although using only @xmath147 points .
we checked that using @xmath127 points the errors are lowered by a factor of @xmath148 in this case , i.e. : @xmath149 precision at the apparent horizon formation .
the reason for better conservation of the energy constraint is that there are not very strong shocks like in the former simulations .
the explanation of the shell formation can be grasp from the equation of motion ( eq . [ motion ] ) .
this effect can also happen in newtonian physics , but its evolution in the strong field regime is highly non - linear and far from obvious .
it seems difficult to make a complete analytic description of its evolution .
however , starting with a ball of constant rest mass density and charge , it is easy to see that a shell must be formed .
the relativistic term @xmath150 , in the eq .
( [ motion ] ) , produce a repulsion of the matter on the initial homogeneous sphere , which in turn produce a charge gradient @xmath151
. the term @xmath152 in the equation of motion can be important in supporting the weight of the star ( see eq . [ motion ] ) .
near the surface of the star , the pressure gradient term @xmath153 cancels the charge gradient term @xmath154 at a certain point we call @xmath155 . since at this point the two terms cancels out , all the matter outside the layer of coordinate @xmath155 is almost in free fall , producing an accumulation of matter and giving rise to the shell , while at coordinates with @xmath156 the matter is supported by a positive net gradient .
we took care to check that this is not a numerical artifact , changing the boundary conditions at the surface , the amount of charge , and the initial conditions .
hence , it results that the point @xmath155 is roughly the _ locus in quo _ the term @xmath157 and the term @xmath158 sum zero , then the gas is compressed and a shell of matter forms .
the non - linearity in the strong field regime comes from the fact that @xmath77 is a function of the charge and the mass and multiplies the two gradient terms , enhancing the effect ( see eq . [ motion ] ) . from all the numerical experiments we performed , we observed that the shell formation effect arise more clearly when the initial density profile is flat .
for the case of charged neutron star collapse the effect is negligible @xcite .
the point @xmath155 shift to lower values with time , and this brings the shell to enlarge until @xmath159 , when it reaches the origin and rebounds forming the shock wave mentioned above .
it is observed that the shock do not form in the uncharged case , using the same set of initial conditions .
the density contrast between the shell and its interior is higher for greater total charge ( or equivalently higher internal electric field and zero total charge ) .
this is because the coulomb repulsion inside the shell will be higher , and this gives a higher compression of the matter at the shell . the coulomb repulsion term together with the pure relativistic term @xmath160 in the eq . ( [ motion ] ) , are responsible for the positive mass density gradient near the coordinate origin .
the charged collapse is slightly delayed respect to the uncharged one , i.e.:@xmath161 sec for model 1 respect to the uncharged case that lasts @xmath162 sec . at the end of the simulation
, we observe that the density profile looks globally flatter and the maximum density value is lower than in the uncharged case . the case 14 in table i is special because it represents stars with a charge greater or equal than the extremal value .
we verified that in this cases the star expands forever " , i.e. : we follow its expansion until the density gives a machine underflow . in any of these cases
we expect a re - collapse of the matter : the binding energy per nucleon tends to zero when the charge tends to the mass ( see @xcite ) , and is negative for a charge greater than mass .
for charged stars there must exist a maximum mass limit like for uncharged compact stars .
for white dwarfs the limit is known as chandrasekhar mass limit , while for neutron stars is the oppenheimer - volkoff limit , etc . in this section ,
we calculate the mass limit for newtonian charged compact stars .
we assume , for simplicity , an equation of state dominated by electrons , @xmath163 for a spherical configuration , the newtonian hydrostatic equilibrium gives : @xmath164 simplifying the two equations , for @xmath165 and @xmath166 we get @xmath167 . from this equations ,
we obtain the limiting mass @xmath168 . assuming the charge is a fraction @xmath169 of the mass , i.e. : @xmath170 , @xmath171 which reduces to the known chandrasekhar mass formula when @xmath172 ( no charge ) , modulo some geometric factor depending on the density profile . for the extreme case @xmath173 , the configuration disperse to infinity .
we checked that this also happens in the relativistic case ( see model 14 of table i ) .
the quantity of energy deposition behind the formed shock wave is critical to produce a successful supernova explosion , like in the proposed mechanisms for core collapse supernovae explosions @xcite , @xcite , @xcite . in the cases we studied the shock wave is not fast enough to reach the surface of the star before the apparent horizon forms ahead of it , and the shock ends trapped by the black hole region .
however , when including neutrino transport in our simulations the shell formation could have a dramatic effect .
this provide a new mechanism for a successful core collapse supernova explosion .
although , the calculations performed in the present work are far from a complete supernova calculation , it is simple to show that in the charged collapse there are a more efficient energy deposition to revive the supernova shock wave . as the density profile is flatter than in the uncharged case , the neutrino - sphere and gain radius for neutrino deposition must be different . in order to show this , we consider the change in the radius @xmath174 , of the neutrino sphere , which is defined by @xmath175 , where the optical depth of the uncharged matter for the neutrinos is @xcite @xmath176 according to the numerical results presented , we make the approximation @xmath177 , for the charged case , in order to obtain an analytic expression for the optical depth for neutrinos . in this case , the optical depth of the charged matter is @xmath178 it can be seen that @xmath179 .
assuming typical values @xcite : @xmath180 mev and @xmath181 , in the uncharged case the position of the neutrino - sphere is at : @xmath182 km , and in the charged case km , at which @xmath183 , for the particular case @xmath184 . ]
it is : @xmath185 km .
this implies a more efficient neutrino trapping during the charged collapse , and hence a more efficient energy deposition . [
cols=">,^,^,^,^,^,^,^,^ " , ]
we have performed relativistic numeric simulations for the collapse of spherical symmetric stars with a polytropic equation of state and possessing a uniform distribution of electric charge .
we have not studied in this paper how the matter acquires a high internal electric field before or during the collapse process . in the present model we studied the essential features of the collapse of stars with a total electric charge .
the simulations are an approximation to the more realistic problem of the temporal evolution of an star with an strong internal electric field and total charge zero .
this model is under consideration by the authors . in the cases
were the stars formed an apparent horizon it is unavoidable the formation of a reissner - nordstrm black hole .
this is guaranteed by the singularity theorems and by the birkhoff theorem . for low values of the charge to mass ratio @xmath186 , no difference with the collapse of an uncharged star was found .
the value of the total charge that prevent the collapse of the star with any initial condition is given by a charge to mass ratio @xmath1 .
that means that stars with @xmath1 spreads and do not collapse to form black holes nor stars in hydrostatic equilibrium .
it is observed a dramatically different physical behavior when @xmath187 . in this case , the collapsing matter forms a shell - like structure , or bubble , surrounding an interior region of lower density and charge .
the effect is due to the competence between the coulomb electrostatic repulsion and the attraction of gravity .
this effect must occur in non - relativistic physics as well .
the relativistic case is more interesting due to the non - linearity of the einstein equations .
the effect is more important when lower is the internal energy of the star ( see table i ) . in some of the experiments we observe a blueshift produced at the bouncing shock wave ( see fig . [
fig : a3048 ] ) , because the strong shock wave acquires a positive velocity over a small lapse of time .
for all the cases studied , the density profile is globally flatter than in the uncharged collapse .
the optical depth for neutrinos is much higher in the case of the charged collapse , hence , the neutrino trapping must be more efficient . in conclusion , if the internal electric field increase during the stellar collapse the formation of the shell must be taken seriously in supernova simulations .
in addition , we obtained the mass limit formula for newtonian charged stars , which clearly precludes the formation of naked singularities .
the mass limit can not be naively extrapolated for the relativistic case , because charge and pressure regeneration effects can change the maximum charge to mass ratio . however , in the present work , we checked that it is not possible to form an star from matter with a charge to mass ratio greater or equal than one in agreement with our mass formula and with the cosmic censorship conjecture .
a more complete simulation including neutrino transport , and other quantum effects , is being considered by the authors .
the authors acknowledge the brazilian agency fapesp . this work was benefited from useful discussions with dr .
samuel r. oliveira and dr .
eduardo gueron .
we acknowledge gian machado de castro for a critical reading of the manuscript .
the simulations were performed at the parallel computer lab .
, department of applied mathematics , unicamp ( campinas state university ) . c. r. ghezzi and p. s. letelier in `` the tenth marcel grossmann meeting : on recent developments in theoretical and experimental general relativity , gravitation and relativistic field theories '' , edited by mrio novello , santiago prez bergliaffa and remo ruffini ( world scientific , singapore , 2006 ) ; c. r. ghezzi and p. s. letelier , gr - qc/0312090 . | the time evolution of a set of @xmath0 unstable charged stars that collapse is computed integrating the einstein - maxwell equations .
the model simulate the collapse of an spherical star that had exhausted its nuclear fuel and have or acquires a net electric charge in its core while collapsing .
when the charge to mass ratio is @xmath1 the star do not collapse and spreads . on the other hand ,
it is observed a different physical behavior with a charge to mass ratio @xmath2 . in this case , the collapsing matter forms a bubble enclosing a lower density core .
we discuss an immediate astrophysical consequence of these results that is a more efficient neutrino trapping during the stellar collapse and an alternative mechanism for powerful supernova explosions .
the outer space - time of the star is the reissner - nordstrm solution that match smoothly with our interior numerical solution , thus the collapsing models forms reissner - nordstrm black holes . |
since the isolation of monolayer graphene almost a decade ago @xcite , there has been a high interest in the low energy transport and optical properties of not only the monolayer @xcite but also a few layer graphene systems .
@xcite these properties are determined by the electronic band structure near the @xmath0 point .
the undoped monolayer graphene ( mlg ) is characterized by the universal optical conductivity , @xmath1 .
this implies that the transmittance depends solely on the fine structure constant and originates the quantized visible opacity of suspended monolayer graphene @xcite .
as far as doped graphene is concerned , there are several effects that arise in the optical properties , related to the restrictions introduced on the interband transitions by the state filling and also to the onset of intraband transitions @xcite . the latter correspond to plasmons and give rise to the interesting and promising field of graphene plasmonics @xcite .
graphene multilayers offer a new ingredient to the interesting physics and potential applications .
the relatively weak interlayer coupling , on the one hand , implies that they should inherit some properties of the parent material @xcite , on the other hand , it introduces a new energy scale , of the order of few tenths of the electron - volt , that should yield some new properties .
it has been shown @xcite that there is also a universal optical conductivity in a undoped @xmath2-layer graphene , equal to @xmath3 , that is reached in undoped graphene in the low frequency limit . at the same time , the optical response of doped bilayer graphene reveals intense strongly doping - dependent features in the mid - infrared ( around 0.4 ev ) @xcite .
the origin of these experimentally observed features has been considered theoretically in these works and also , in more detail , in ref .
@xcite , where the band structure and the optical conductivity of bilayer graphene were calculated .
some novel plasmonic effects in bernal - stacked bilayer graphene were predicted in the recent work @xcite . as the number of layers increases beyond @xmath4 , the band structure and
the optical conductivity become dependent not only on @xmath2 but also upon the stacking arrangement .
there are three distinct planar projections of the honeycomb lattice ( usually denoted a , b and c ) and , consequently , @xmath5 distinct @xmath6layer sequences @xcite .
in particular , the stacking of three layers in a graphene trilayer can be either aba ( also called bernal - type ) and abc ( rhombohedral ) @xcite .
these two different stacking arrangements lead to strikingly different electronic band strucures @xcite . for instance , it was found that undoped graphene abc trilayer shows many - body correlations with an energy gap , while the bernal - type stacking ( taking place in graphite ) does not lead to a gap @xcite .
non - bernal stacked multiple graphene layers have attracted a considerable attention related to the prospect of further enhancement of capabilities of graphene - based optoelectronic devices , in particular , thz and ir photodiodes @xcite .
the optical conductivity of abc trilayers has been considered in a number of works @xcite , in particular , the effect of doping has been analysed @xcite .
even though trilayer graphene contains , on average , regions of abc and aba stacking in an 15:85 ratio @xcite , apparently the optoelectronic properties of the latter attracted less attention and we are aware of only one work @xcite devoted to this topic .
this is notwithstanding the possibility of using far - infrared ( fir ) spectroscopy , along with the common raman scattering technique @xcite in order to distinguish different trilayer graphene species @xcite .
therefore , the main purpose of this article is to analyse the spectral characteristics of the optical conductivity of intrinsic and doped aba stacked trilayers .
the mirror - symmetric bernal stacking is the most common in graphene multilayers that can be exfoliated from natural graphite since it shares its crystalline structure @xcite .
intuitively , one can expect that the electrons in aba trilayers can have a monolayer- or bilayer - like character @xcite .
we shall present the analytical dispersion relation for the electrons near the dirac point , from which it follows that the band structure of the aba trilayer indeed looks like a superposition of those characteristic of a monolayer and a bilayer , although the latter corrseponds to an effective interlayer hopping constant @xmath7 times larger than the true one .
based on this band structure , we calculate both the interband and intraband ( drude ) contributions to the optical conductivity .
the most interesting result is that the optical absorption of a doped aba trilayer is dominated by a narrow resonant peak at the frequency corresponding to this effective interlayer hopping energy ( approximately 0.56 ev ) .
it is caused by the fact that the dispersion curves corresponding to two bands are nearly parallel for a considerable range of wavevectors near the dirac point .
this effect sometimes is called `` band nesting '' @xcite in order to distinguish from van hove singularities in the single - particle density of states .
we will show that the intensity of this resonant absorption is approximately proportional to the fermi energy and , therefore , can be controlled by adjusting the gate voltage applied to the graphene layer .
the tight - binding hamiltonian for non - interacting electrons in the aba stacked trilayer involves three a - type and three b - type sites and includes the essential in - plane ( @xmath8ev ) and interlayer hoppings ( @xmath9ev ) as shown in fig .
[ fig : f1 ] . these two parameters , connecting atoms that are right on top of each other in adjacent layers , are sufficient to describe the main features of the band structure , such as the type of dispersion of the energy bands and their separation , as confirmed by recent dft calculations @xcite .
keeping only @xmath10 and @xmath11 hoppings permits to obtain simple formulae for band gaps , effective masses and the fermi velocity .
then the hamiltonian is given by @xcite : @xmath12 here @xmath13 is the composite index , which determines the atomic positions in the lattice , @xmath14 , with @xmath15 , @xmath16 being the lattice vectors .
the three vectors that connect the @xmath17 atom to its three nearest neighbors are @xmath18 , @xmath19 , and @xmath20 , where @xmath21 is the c - c interatomic distance .
the positions of the @xmath22 atoms relative to the @xmath17 atoms in each of the three layers are shown in fig .
[ fig : f1 ] .
the energy spectrum of the hamiltonian ( [ tb - hamiltonian ] ) is composed of six energy bands given by ( see appendix a for details ) : @xmath23 @xmath24 @xmath25 where @xmath26\ : .
\label { eq : phi}\ ] ] the band structure is depicted in fig .
[ fig : f2 ] . note that the gap between the bands @xmath27 and the dirac point is @xmath28 , while its counterpart in graphene bilayer is just equal to @xmath29 .
@xcite in the vicinity of @xmath30 points [ dirac points , @xmath31 the band energies can be approximated in the following way : @xmath32 where @xmath33 .
note that the effective masses at the bottom of the bands 1 and 3 are equal within the present model , @xmath34 , and the fermi velocity of the dirac - type band coincides with that of monolayer graphene , @xmath35 .
dirac point ( right ) , where @xmath36 . the @xmath37 axis is scaled by @xmath38 , @xmath39 is the lattice constant . after intersecting
the @xmath40 points the dispersion curves continue along the @xmath41 direction .
, width=10 ] the finite frequency ( optical ) conductivity is calculated through a standard procedure using the kubo formula @xcite , @xmath42-n_{f}\left[e_{l^\prime}(\vec k)\right]}{\hbar\omega - e_{l}(\vec k)+e_{l^\prime}(\vec k)+i\gamma } \ : , \label { kubo}\end{aligned}\ ] ] where @xmath43 is the electron charge , @xmath44 is the unit cell area , @xmath45 is the fermi function , @xmath46 is the eignestate corresponding to energy @xmath47 ( @xmath48 are the band indices ) , @xmath49 is a damping parameter and @xmath50 denotes the @xmath51-component of in - plane velocity operator , defined by @xmath52\:.\end{aligned}\ ] ] the summation over @xmath53 in eq .
( [ kubo ] ) involves wavevectors in the first brilluion zone , which is the usual hexagon with side @xmath54 . in practice , this summation is replaced by a 2d integration over the triangle formed by the points @xmath55 @xmath56 , @xmath0 @xmath57 , and @xmath58 @xmath59 , and the result is multiplied by twelve . for doped graphene ( @xmath60 )
it is necessary to include in the optical conductivity also the term related to intraband transitions , often referred to as drude term .
although it can be calculated through the kubo formula , it is easier to derive this term using the boltzmann transport equation @xcite .
thus , the electric current is written as @xmath61 where @xmath62 is the deviation of the carriers distribution from the equilibrium fermi dirac function , @xmath63 $ ] .
the former is readily obtained from the boltzmann equation , @xmath64}{\partial e_l(\vec k ) } \right ) \ : , \label{eq : n_k}\ ] ] where @xmath65 is the group velocity of the charge carriers in the @xmath66-th band and @xmath67 denotes the carrier relaxation time .
the factor of 4 in eq .
( [ eq : semi - classical_current ] ) is due to the spin and valley degeneracy .
substitution of eq .
( [ eq : n_k ] ) into eq .
( [ eq : semi - classical_current ] ) yields the drude conductivity and for zero temperature we have : @xmath68\ : , \label { drude}\end{aligned}\ ] ] where the damping parameters @xmath69 and @xmath49 are defined as the inverse of the corresponding relaxation time at the fermi level ( we make no distinction between bands 1 and 2 ) .
this drude term has to be added to the optical conductivity ( [ kubo ] ) .
note that the first term is precisely three times the drude conductivity of monolayer graphene .
we shall now concentrate on the frequency dependence of the real part of the derived optical conductivity , @xmath70 , that determines the absorption , for different values of the chemical potential , @xmath71 .
the spectra of real , @xmath70 , and imaginary , @xmath72 , parts of the conductivity for three different values of @xmath71 ( conveniently expressed in units of @xmath11 ) are presented in figs .
[ fig : f3 ] and [ fig : f4 ] , respectively .
note that the first value corresponds to undoped graphene ( @xmath73 ) , the second one is @xmath74 and the last @xmath75 . for intrinsic trilayer graphene , either aba or abc stacked
, @xmath70 tends to @xmath76 ( @xcite ) .
we clearly see the effects of doping present as the fermi step at @xmath77 ( for @xmath78 ) , known in semiconductors as burstein moss effect @xcite and familiar in monolayer graphene @xcite .
the feature characteristic of trilayer graphene , located at @xmath79 , splits into two for @xmath78 ( compare red and blue curves to the black one in fig . [
fig : f3 ] ) .
but the most impressive effect of the doping is the onset of the resonant peak at @xmath80 , whose intensity grows strongly with @xmath71 . in order to understand these spectral changes caused by doping
, we analyzed all possible optical transitions listed in table 1 ( note that none of the velocity matrix elements vanishes , they are all allowed ! ) . as it can be seen from fig .
[ fig : f2 ] , there are two dirac - type and four approximately parabolic bands that arise from six atoms in the unit cell of the aba trilayer .
let us consider the allowed transitions that correspond to the features of the optical conductivity .
for the undoped graphene ( @xmath73 ) , we have nine possible transitions , with energy conservation restrictions imposed on some of them ( see table 1 ) . there are four transitions allowed without any restriction for all frequencies , namely , @xmath81 to @xmath82 ( denoted as @xmath83 , @xmath84 , @xmath85 and @xmath86 .
the onset of transitons involving bands @xmath27 is at @xmath87 and @xmath88 .
that is why there are four possible transitions including @xmath89 , @xmath90 , @xmath91 and @xmath92 , with a threshold frequency @xmath93 .
still another possible transition is from @xmath94 to @xmath95 , with a threshold at @xmath96 , although it is less pronounced as clearly seen in fig .
[ fig : f3 ] ( bold black curve ) . with doping the system
so as @xmath97 , the behavior of the optical conductivity is altered , caused by the changes in the allowed transitions with applying more state filling restrictions as well as the onset of new transitions . .
thin dashed line indicates the value @xmath98 .
the inset shows the real part of the optical conductivity of an intrinsic aba stacked trilayer graphene computed using _ ab - initio _ methods @xcite .
the peak is located where the tight - binding calculation predicts it .
, width=377 ] .,width=10 ] transition ( @xmath99).,width=10 ] transition _ versus _ fermi energy , calculated analytically within parabolic approximation and numerically.,width=10 ] the general expression for the threshold energy for the @xmath100 transition ( for the case @xmath101 , @xmath102 ) is given by : @xmath103 where @xmath104 is the fermi wavevector in the @xmath66-th band ( the root of equation @xmath105 , or @xmath106 , if the equation does not have a real root ) .
at the same time , for the case where @xmath107 the expression for the threshold energy is given by @xmath108 .
\nonumber\end{aligned}\ ] ] there is also an upper cut - off frequency in this case , given by @xmath109 , \nonumber\end{aligned}\ ] ] so that the allowed transition frequencies lie inside the domain @xmath110 .
it is possible to obtain simple expressions for the threshold energies using the approximation ( [ eq : parab ] ) .
thus , we obtain the following values : @xmath111 @xmath112 @xmath113 several transitions can occur only for nonzero @xmath71 : @xmath114 .optical transitions and threshold frequencies in units of @xmath11 .
[ cols="^,^,^,>,>,>,>,>",options="header " , ] [ table : nonlin ] the numerical values of the threshold energies ( in units of @xmath11 ) are given in table [ table : nonlin ] where the numerical values correspond to the exact band structure ( see appendix b ) because the parabolic / linear approximation becomes to fail for higher values of @xmath71 .
it becomes clear from table [ table : nonlin ] that almost only the @xmath115 and @xmath116 transitions ( not possible for @xmath73 ) are responsible for the onset of the resonant peak at @xmath117 .
we calculated these contributions to the optical conductivity and found that the intensity of the @xmath115 transition exceeds by far that of the other one .
it can be understood by the fact that the dispersion curves @xmath118 and @xmath119 are nearly parallel for a broad range of @xmath120 values ( band nesting ) and therefore the joint density of states for this transition is large , as can be seen clearly from fig . [
fig : f2 ] .
this figure takes into account only the @xmath115 transition and shows that it is responsible for the intensity of the resonant absorption band .
the increasing intensity of the @xmath115 transition as @xmath71 grows is related to the growing number of occupied states that are depopulated by absorbing electromagnetic radiation .
this is the principal spectral feature observed in ref .
@xcite ( where it was denoted `` band c '' ) , which can be considered as characteristic of aba graphene @xcite and we return to it below . for heavily doped @xmath121type layers studied in ref .
@xcite , two weak absorption bands ( denoted there as a and b bands ) were observed at lower frequencies and attributed to the @xmath122 and @xmath123 transitions . in our consideration of @xmath124type layers
these transitions correspond to @xmath125 and @xmath116 , respectively , both involving the dirac - type band .
indeed , they are allowed ( even though only for doped samples ) but , according to our results , their intensity is rather low ( for instance , we can not see any feature at @xmath126 , the threshold frequency of @xmath116 transition for @xmath127 ( see table [ table : nonlin ] ) .
possibly some further effects can enhance transitions to and from the dirac - type electronic bands .
the approximately triangular shaped feature in the spectrum for @xmath73 ( @xmath79 ) ( that could be anticipated institutively @xcite ) splits into two sub - bands corresponding to the @xmath89 and @xmath91 transitions . at @xmath73
, they have the same energy @xmath128 but for @xmath78 the former is shifted to higher energy ( see table 1 ) .
note that , by chance , in figs .
[ fig : f3 ] and [ fig : f4 ] the features related to the @xmath91 transition for @xmath129 and to the @xmath89 one for @xmath127 appear nearly at the same frequency , @xmath130 .
such a feature was not observed in ref .
@xcite ( whose experimental situation qualitatively corresponds to our case of @xmath129 ) , possibly because it was hindered by fabri - perot interference in the substrate .
the intensity of the 1@xmath1313 band as function of @xmath71 can be evaluated analytically if we assume that the matrix element does not depend on @xmath120 : @xmath132-n_f \left[e_3(\vec k)\right]\right\}\ : .
\nonumber\end{aligned}\ ] ] changing from sum to integration and using the parabolic approximation ( [ eq : parab ] ) near the dirac point , in the limit of low energies for @xmath133 and @xmath134 we obtain : @xmath135- \theta\left[\mu - e_1(\vec k)- \hbar\omega\right]\right\}\ : .
\label{intensity}\end{aligned}\ ] ] here @xmath136 denotes a lorentzian replacing the @xmath137 function broadened because of natural reasons .
the integral in eq .
( [ intensity ] ) depends on the value of @xmath71 , that is , @xmath138 for @xmath139 and @xmath140 for @xmath141 .
this analytical result is compared to the numerical data in fig .
[ fig : f5 ] .
higher order hopping parameters , in particular , those connecting more distant atoms in adjacent layers and usually denoted @xmath142 and @xmath143 @xcite , are known to cause a distortion ( trigonal warping ) of the low - energy bands and a small electron - hole asymmetry @xcite . therefore they can slightly affect the intensity and shape of the resonant absorption band but its position , to a good approximation , is determined by the @xmath11 hopping @xcite .
this is confirmed by the results of our _ ab - initio _ dft calculations for intrinsic graphene , shown in the inset of fig .
[ fig : f3 ] .
in summary , we calculated the spectral dependence of the real and imaginary parts of the optical conductivity of bernal - stacked trilayer graphene .
even though the energy bands of this material look like a superposition of those of a monolayer and a bilayer ( with a larger gap between the dirac point and higher parabolic band ) , the optical spectra are rich and interesting , especially in the case of gated ( doped ) graphene because all kinds of interband transitions are allowed . in particular , there is a strong and narrow resonant band cause by transitions between two `` nested '' parabolic bands .
the intensity of this band is controlled by the fermi level position and attains a maximum for @xmath144 ( fig .
[ fig : f5 ] ) , providing a strong light - matter coupling in the atomically thin layer @xcite .
although the physical origin of the resonant absorption is clear and it could be predicted by just inspecting the band structure of the aba stacked graphene , its dependence on the fermi level is not evident without calculations . it would be interesting to show experimentally that the intensity of the resonant absorption attains its maximum for the fermi energy of @xmath145 ev .
this can be achieved by combining usual doping and application of gate voltage @xcite .
the application of a large gate voltage can slightly modify the band structure but it can be taken into account by adding appropriate ( unequal ) constant potentials in the hamiltonian ( [ ham ] ) for atoms belonging to the different monolayers @xcite .
the modulation of the peak absorption can be interesting for optoelectronic devices such as resonant photodetectors with adjustable sensitivity and it is robust with respect to the temperature .
we performed calculations also for @xmath146 k and the spectra are very similar to those presented in the figures , except for the obvious broadening of the fermi steps seen in fig .
[ fig : f3 ] , i.e. the resonant band remains narrow .
compared to bilayer graphene , the resonant band occurs at a ( 1.4 times ) higher frequency , which broadens the range of applications of bernal - stacked multilayer graphene materials that can be directly exfoliated from natural graphite .
also , we would like to point out that the drude conductivity is three times higher than for monolayer graphene with same doping level , accordingly , the surface plasmon frequency would be also higher ( by a factor of @xmath147 ) @xcite , extending the spectral range of possible applications of graphene plasmonics .
financial support from the compete programme ( feder ) and the portuguese foundation for science and technology ( fct ) through projects pest - c / fis / ui0607/2013 and ptdc / fis/113199/2009 is gratefully acknowledged .
zr thanks the hospitality of the physics center of minho university during her stay in portugal .
yvb , nmrp , and miv acknowledge support by the ec under graphene flagship ( contract no .
cnect - ict-604391 ) .
after transforming the hamiltonian ( [ ham ] ) to the momentum space [ by representing @xmath148 , @xmath149 , @xmath150 and @xmath2 is the number of unit cells ] , its matrix form is : @xmath151 where @xmath152 which is given explicitly by eq .
( [ eq : phi ] ) .
note that the relation of this @xmath153 to the frequently defined @xcite quantity @xmath154 is the following : @xmath155 .
the hamiltonian ( [ tb ] ) is written in the basis of the atomic orbital eigenfunctions : @xmath156 the aba trilayer is mirror symmetric with respect to the middle layer , therefore one can transform the hamiltonian into a tridiagonal form using a unitary transformation @xcite @xmath157 in the new basis the hamiltonian reads @xmath158 this matrix is composed of two blocks containing only intralayer hoppings and an effective interlayer one , @xmath159 .
it can easily be diagonalised yielding eqs .
( [ eq : e1 ] - [ eq : e3 ] ) .
here we derive the expressions for the characteristic energies presented in table 1 , beyond the parabolic ( linear ) approximation for the band spectra .
first we consider the case where @xmath160 , i.e. lies below the bottom of the band @xmath161 . among the first three possible transitions from the @xmath162 to the upper bands
, the @xmath163 transition threshold remains the same [ see eq .
( [ eq : trans_m1_to_3 ] ) ] and similar to the case of @xmath73 .
transition from @xmath164 become possible if @xmath165 , then for @xmath160 the allowed energies are @xmath166 . the domain of allowed energies for the the transition @xmath167 can be obtained as @xmath168 . since @xmath169 , we obtain @xmath170 the number corresponds to @xmath160 . for the second group of transitions from the @xmath171 band ,
the @xmath90 threshold remains as given by eq .
( [ eq : trans_m2_to_3 ] ) .
the next one , @xmath172 , can occur for energies higher than @xmath173 , namely @xmath166 .
transition @xmath174 is allowed if @xmath175 . since @xmath176 , by solving this equation for @xmath177 we obtain the relation : @xmath178 finally , for the last group of three possible transitions from @xmath179 to the three upper bands , for the @xmath180 one eq .
( [ eq : trans_m3_to_3 ] ) ] holds beyond the parabolic approximation , @xmath181 is allowed if @xmath182 ( then @xmath183 ) , and the @xmath184 transition is allowed if @xmath185 the non - zero chemical potential gives rise to new transitions that are not possible in the intrinsic material .
the @xmath186 transition is allowed for the frequencies in the vicinity of @xmath187 .
the @xmath188 one occurs in the domain @xmath189 for @xmath160 this corresponds to @xmath190 the situation is similar for the @xmath191 transition because , in addition to a threshold energy , there is also a cut - off : @xmath192 where the expression in the left hand side is equal to @xmath193 for @xmath160 .
the case of large chemical potential , ( @xmath194 ) , above the bottom of the band @xmath195 corresponds to the situation where the third band is partially filled . for @xmath165 ,
the @xmath83 transition becomes possible .
the frequency domain for transition @xmath84 is similar to the previous case [ eq . ( [ eq:2 ] ) ] but for @xmath194 we have @xmath196 .
transition from @xmath197 to @xmath95 is allowed for @xmath198 . from the dispersion relation @xmath199
we obtain @xmath200 , and @xmath201 .
notice that this equation is exactly the same as eq .
( [ eq : trans_m1_to_3 ] ) , obtained within the parabolic approximation . among the next three possible transitions ,
the energy domain for @xmath85 is given by eq .
( [ eq:1 ] ) .
for @xmath194 we have @xmath202 .
the @xmath203 transition is possible when @xmath204 . for @xmath205 , from the dispersion relation
@xmath199 we have @xmath200 . hence , @xmath206 .
the transitions @xmath91 and @xmath207 are allowed if @xmath208 ( similar to previous case ) and @xmath204 , respectively .
transition from @xmath94 to @xmath209 is described by eq .
( [ eq:9 ] ) , then @xmath210 .
finally , @xmath211 is described by eq .
( [ eq:2a ] ) ( then @xmath212 ) ; @xmath213 becomes possible for @xmath214 and @xmath215 for the finite frequency domain defined by @xmath216 .
for @xmath194 it means @xmath217 | the band structure and the optical conductivity of an aba ( bernal - type ) stacked graphene trilayer are calculated .
it is shown that , under appropriate doping , a strong resonant peak develops in the optical conductivity , located at the frequency corresponding to approximately 1.4 times the interlayer hopping energy and caused by the `` nesting '' of two nearly parabolic bands in the electronic spectrum .
the intensity of this resonant absorption can be controlled by adjusting the gate voltage .
the effect is robust with respect to increasing temperature . |
in the raman spectra of a single wall carbon nanotube ( swnt ) , the two in - plane optical phonon modes , that is , the longitudinal optic ( lo ) and transverse optic ( to ) phonon modes at the @xmath0 point in the two - dimensional brillouin zone ( 2d bz ) , which are degenerate in graphite and graphene , split into two peaks , @xmath1 and @xmath2 peaks , respectively .
@xcite the splitting of the two peaks for swnts is inversely proportional to the square of the diameter @xmath3 of swnts due to the curvature effect , in which @xmath1 does not change with changing @xmath3 , but the @xmath2 frequency decreases with decreasing @xmath3 .
@xcite in particular , for metallic swnts , the @xmath2 peaks appear at a lower frequency than the @xmath2 peaks for semiconducting swnts with a similar diameter .
@xcite the spectra of @xmath2 for metallic swnts show a much larger spectral width than that for semiconducting swnts .
further , the spectral @xmath2 feature shows an asymmetric lineshape as a function of frequency which is known as the breit - wigner - fano ( bwf ) lineshape .
@xcite the origin of the bwf lineshape is considered to be due to the interaction of discrete phonon states with continuous free electron states .
it has been widely accepted that the frequency shift of the @xmath4-band is produced by the electron - phonon ( el - ph ) interaction .
@xcite an optical phonon changes into an electron - hole pair as an intermediate state by the el - ph interaction .
this process gives the phonon a self - energy .
the phonon self - energy is sensitive to the fermi energy , @xmath5 . in the case of graphite intercalation compounds in which the charge transfer of an electron from a dopant to the graphite layer can be controlled by the doping atom and its concentration , eklund _
et al . _ observed a shift of the @xmath4-band frequency with an increase of the spectral width .
@xcite in this case the frequency shifted spectra show that not only the lo mode but also the to mode are shifted in the same fashion by a dopant . for a graphene monolayer , lazzeri _ et al .
_ calculated the @xmath5 dependence of the shift of the g - band frequency .
@xcite the lo mode softening in metallic swnts was shown by dubay _
et al . _ , @xcite on the basis of density functional theory .
recently nguyen _ et al . _
@xcite and farhat _ et al . _
@xcite observed the phonon softening effect of swnts as a function of @xmath5 by field effect doping and electro - chemical doping , respectively , and their results clearly show that the lo phonon modes become soft as a function of @xmath5 .
ando discussed the phonon softening for metallic swnts as a function of the @xmath5 position , in which the phonon softening occurs for the lo phonon mode and for a special range of @xmath5 , that is , for @xmath6 .
@xcite in this paper , we show that the @xmath0 point to phonon mode becomes hard when @xmath7 and has a considerable broadening for metallic zigzag nanotubes .
the occurrence of the phonon hardening for the to mode is due to the curvature effect , a special character of the electron - phonon coupling , and a basic consequence of second - order perturbation theory .
we show using a gauge symmetry argument that the electrons completely decouple from the to mode near the @xmath0 point . besides
, we show that for a chiral nanotube , both the lo and to modes are softened due to the fact that the direction of the to phonon vibration is not parallel to the nanotube circumferential direction .
@xcite another interest of ours is the mixing of lo and to phonons to form degenerate phonon frequencies .
when the lo phonon mode becomes soft , a crossing of the lo mode with the to mode occurs at a certain value of @xmath5 .
for such a mode crossing , we should generally consider the electron - phonon coupling for degenerate phonon modes to sense the crossing or anti - crossing of the two phonon frequencies as a function of @xmath5 .
we will show by an analytical calculation that there is no mixing between the lo and to phonon modes for any case due to time reversal symmetry .
the organization of the paper is as follows . in sec .
[ sec : ps ] we show our method of calculation and present the results for armchair and metallic zigzag swnts . in sec .
[ sec : chi ] using effective mass theory , we show how the el - ph interaction depends on the chiral angles of swnts , and in sec .
[ sec : dis ] a discussion based on gauge symmetry and time reversal symmetry for the el - ph coupling is given . in sec .
[ sec : sum ] , a comparison with the experiments and summary are given .
the frequency shift of the @xmath0-point lo and to phonon modes for metallic swnts is calculated by second - order perturbation theory .
the phonon energy including the el - ph interaction becomes @xmath8 ( @xmath9 ) where @xmath10 is the original phonon frequency without the el - ph interaction and @xmath11 is given by @xmath12 the factor 2 in eq .
( [ eq : omega_2 ] ) comes from spin degeneracy .
@xmath11 is the quantum correction to the phonon energy due to the electron - hole pair creation as shown in fig .
[ fig : process](a ) . in eq .
( [ eq : omega_2 ] ) , @xmath13 is the matrix element for creating an electron - hole pair at momentum @xmath14 by the el - ph interaction , @xmath15 , @xmath16 ( @xmath17 ) is the electron ( hole ) energy and @xmath18 is the decay width . in fig .
[ fig : process](a ) , an intermediate electron - hole pair state that has the energy of @xmath19 is shown .
we need to sum ( @xmath20 ) over all possible intermediate electron - hole pair states in eq .
( [ eq : omega_2 ] ) , which can have a much larger energy than the phonon ( @xmath21 ) . since @xmath13 is a smooth function of @xmath19 in the denominator of eq .
( [ eq : omega_2 ] ) , the contribution to @xmath11 from an electron - hole pair depends on its energy . in fig . [
fig : process](b ) , we plot the real part and imaginary part of the denominator of eq .
( [ eq : omega_2 ] ) , @xmath22 as a function of @xmath23 in the case of @xmath24 ev and @xmath25 mev . here
@xmath26 is a positive ( negative ) value when @xmath27 ( @xmath28 ) and the lower ( higher ) energy electron - hole pair makes a positive ( negative ) contribution to @xmath11 .
moreover , an electron - hole pair satisfying @xmath29 can not contribute to the energy shift ( shaded region in fig .
[ fig : process](a ) and ( b ) ) because of the fermi distribution function @xmath30 in eq .
( [ eq : omega_2 ] ) .
thus , the quantum correction to the phonon energy by an intermediate electron - hole pair can be controlled by changing the fermi energy , @xmath5 .
for example , when @xmath31 , then @xmath11 takes a minimum value at zero temperature since all positive contributions to @xmath11 are suppressed in eq .
( [ eq : omega_2 ] ) . since @xmath32 for @xmath33 , all high energy intermediate states contribute to phonon softening if we include all the electronic states in the system . here
we introduce a cut - off energy at @xmath34 ev as @xmath35 in order to avoid such a large energy shift in eq .
( [ eq : omega_2 ] ) . the energy shift due to the high - energy intermediate states ( @xmath36 ) can be neglected by renormalizing @xmath37 so as to reproduce the experimental results of raman spectra @xcite since the contribution from @xmath38 just gives a constant energy shift to @xmath39 .
we have checked that the present results do not depend on the selection of the cut - off energy since @xmath40 is much larger than @xmath37 .
@xmath41 is nonzero only very close to @xmath42 , which shows that the phonon can resonantly decay into an electron - hole pair with the same energy .
it is noted that when @xmath43 , @xmath44 at zero temperature while @xmath18 may take a finite value at a finite temperature . in this paper , we calculate @xmath18 self - consistently by calculating @xmath45 in eq .
( [ eq : omega_2 ] ) . is forbidden at zero temperature by the pauli principle .
( b ) the energy correction to the phonon energy by an intermediate electron - hole pair state , especially the sign of the correction , depends on the energy of the intermediate state as @xmath46 . ] in fig .
[ fig : arm ] , we show calculated results for @xmath47 as a function of @xmath5 for a @xmath48 armchair nanotube . here
we take 1595 @xmath49 and 1610 @xmath49 for @xmath50 of the to and lo modes , respectively .
the energy bars denote @xmath18 values .
we have used the extended tight - binding scheme to calculate @xmath16 , @xmath17 , and the electron wavefunction for @xmath51 .
@xcite as for the el - ph matrix element , @xcite we adopted the deformation potential derived on the basis of density - functional theory by porezag _
et al_. @xcite we show the resulting @xmath47 as a function of @xmath5 at the room temperature ( @xmath52 k ) and @xmath53 k in fig .
[ fig : arm](a ) and ( b ) , respectively .
it is shown that the to mode does not exhibit any energy change while the lo mode shows an energy shift and broadening .
as we have mentioned above , the minimum energy is realized at @xmath31 ( @xmath54 ev ) .
there is a local maximum for the spectral peak at @xmath55 .
the broadening for the lo mode occurs within @xmath56 for the lower temperature while the broadening has a tail at room temperature for @xmath57 .
dependence of the lo ( red curve ) and to ( black curve ) phonon energy in the case of the @xmath48 armchair nanotube .
( a ) is taken at room temperature and ( b ) is at 10 k. only the energy of the lo mode is shifted , with the to mode frequency being independent of @xmath5 . the decay width ( @xmath18 )
is plotted as an error - bar . ] a continuous model for electrons in a carbon nanotube is adopted in this paper to explain the lack of an energy shift of the to modes for armchair nanotubes . as we will show in sec .
[ sec : chi ] , the el - ph matrix element for the electron - hole pair creation by the lo and to modes is given by @xmath58 where @xmath59 is the phonon amplitude , and @xmath60 is the el - ph coupling constant . here
@xmath61 is the angle for the polar coordinate around the k ( or k ) point in the 2d bz , in which a @xmath62 point on a cutting line for a metallic energy subband is taken .
the @xmath63 ( @xmath64 ) axis is taken in the direction of the nanotube circumferential ( axis ) direction ( see fig . [
fig : scattering ] ) .
equation ( [ eq : coupling ] ) shows that @xmath13 depends only on @xmath61 but not on @xmath65 , which means that the dependence of this matrix element on @xmath23 is negligible .
since the armchair nanotube is free from the curvature effect , @xcite the cutting line for its metallic energy band lies on the @xmath64 axis .
thus , we have @xmath66 ( @xmath67 ) for the metallic energy subband which has @xmath68 and @xmath69 ( @xmath70 ) .
then , eq . ( [ eq : coupling ] ) tells us that only the lo mode couples to an electron - hole pair and the to mode is not coupled to an electron - hole pair for armchair swnts .
( @xmath64 ) axis is selected as the nanotube circumferential ( axis ) direction .
the amplitude for an electron - hole pair creation depends strongly on the relative position of the cutting line from the k - point .
( b ) if the cutting line crosses the k - point , then the angle @xmath61 ( @xmath71 ) takes @xmath72 ( @xmath67 ) values for @xmath73 ( @xmath74 ) . in this case , the lo mode strongly couples to an electron - hole pair , while the to mode is decoupled from the electron - hole pair according to eq .
( [ eq : coupling ] ) . ] in fig .
[ fig : zig](a ) , we show calculated results for @xmath47 as a function of @xmath5 for a @xmath75 metallic zigzag nanotube . in the case of zigzag nanotubes , not only the lo mode but also the to mode couples with electron - hole pairs .
the spectrum peak position for the to mode becomes harder for @xmath76 , since @xmath26 for @xmath77 contributes to a positive frequency shift .
it has been shown theoretically @xcite and experimentally @xcite that even for `` metallic '' zigzag nanotubes a finite curvature opens a small energy gap .
when the curvature effect is taken into account , the cutting line does not lie on the k - point , but is shifted from the @xmath64 axis . in this case , @xmath78 is nonzero for the lower energy intermediate electron - hole pair states due to @xmath79 .
thus , the to mode can couple to the low energy electron - hole pair which makes a positive energy contribution to the phonon energy shift .
the high energy electron - hole pair still decouples to the to mode since @xmath80 for @xmath81 . therefore , when @xmath82 , then @xmath83 increases by a larger amount than @xmath84 .
the to mode for the small diameter zigzag nanotubes couples strongly with an electron - hole pair because of the stronger curvature effect . in fig .
[ fig : zig](b ) , we show the diameter ( @xmath3 ) dependence of the @xmath47 of zigzag nanotubes for @xmath76 not only for metallic swnts but also for semiconducting swnts . in the case of the semiconducting nanotubes ,
the lo ( to ) mode appears around 1600 ( 1560 ) @xmath49 without any broadening .
only the metallic zigzag nanotubes show an energy shift , and the energy of the lo ( to ) mode decreases ( increases ) as compared to the semiconducting tubes . in the lower part of fig .
[ fig : zig](b ) , we show a curvature - induced energy gap @xmath85 as a function of @xmath3 .
the results show that higher ( lower ) energy electron - hole pairs contribute effectively to the lo ( to ) mode softening ( hardening ) in metallic nanotubes . in the case of semiconducting nanotubes
, we may expect that there is a softening for the lo and to modes according to eq .
( [ eq : coupling ] ) . however , the softening is small as compared with that of the metallic nanotubes because the energy of intermediate electron - hole pair states is much larger than @xmath50 in this case .
dependence of the lo ( red curve ) and to ( black curve ) phonon frequency for a @xmath75 zigzag nanotube .
not only the frequency of the lo mode but also that of the to mode is shifted due to the curvature effect .
( b ) the diameter @xmath3 dependence of the phonon frequency for zigzag nanotubes , including zigzag semiconducting tubes .
@xmath85 denotes the curvature - induced mini energy gap . ]
here , we derive eq . ( [ eq : coupling ] ) on the basis of a continuous model for @xmath86-electrons near the k - point in graphene . in a continuous model
, the local modulation of the hopping integral due to lattice vibrations appears as a deformation - induced gauge field , @xmath87 , in the weyl equation .
@xcite the weyl equation for @xmath86-electrons with energy eigenvalue @xmath23 is written by @xmath88 where @xmath89 ( @xmath90 ) is the fermi velocity and @xmath91 ( @xmath92ev ) is the nearest - neighbor hopping integral , @xmath93 , @xmath94 is the momentum operator , and @xmath95 is the pauli matrix .
@xmath96 is given in eq .
( 3 ) of ref . by a small change @xmath97 ( @xmath98 ) of the hopping integral from @xmath99 ( see fig . [
fig : graphene ] ) , as @xmath100 here @xmath101 for the lo and to modes is given by @xmath102 where @xmath103 denotes the nearest - neighbor vectors ( fig .
[ fig : graphene ] ) and @xmath104 is the relative displacement vector of a b site from an a site ( @xmath105 ) and @xmath60 is the el - ph coupling constant .
we rewrite eq .
( [ eq : a ] ) as @xmath106 where @xmath107 , ( @xmath108 ) , and @xmath109 and @xmath110 have been used ( see fig .
[ fig : graphene ] ) .
then , the el - ph interaction for an in - plane lattice distortion @xmath104 can be rewritten as the vector product of @xmath111 and @xmath104 , @xcite @xmath112 ( @xmath98 ) are vectors pointing to the nearest - neighbor b sites from an a site ( @xmath113 , @xmath114 , and @xmath115 ) .
local modulations of the hopping integral are defined by @xmath116 ( @xmath98 ) .
the modulation is given by optical phonon modes as @xmath117 where @xmath104 ( @xmath118 ) is a relative displacement vector of a b site relative to the nearest a site . ]
we consider the lo and to phonon modes with @xmath119 ( i.e. , @xmath0-point ) .
then , an electron - hole pair is excited by a constant @xmath120 . the el - ph matrix element for the electron - hole pair generation
is given by @xmath121 where @xmath122 ( @xmath123 ) denotes an energy eigenstate of @xmath124 in the conduction ( valence ) energy band with energy eigenvalue @xmath125 ( @xmath126 ) , @xmath127 where @xmath128 denotes the total area of the system , @xmath129 .
we first consider the case of a zigzag nanotube in fig .
[ fig : graphene ] .
then , we denote @xmath130 ( @xmath131 ) as a coordinate around ( along ) the axis ( so that @xmath132 ) , and @xmath133 ( @xmath134 ) are assigned to the to ( lo ) phonon mode .
the corresponding @xmath96 for @xmath133 and @xmath134 is @xmath135 and @xmath136 , respectively . by calculating eq .
( [ eq : mat_elph ] ) for the to mode with @xmath137 and for the lo mode with @xmath138 , we get eq .
( [ eq : coupling ] ) .
next , we consider the case of an armchair nanotube . then , @xmath130 ( @xmath131 ) is the coordinate along ( around ) the axis ( so that @xmath139 ) , and @xmath133 ( @xmath134 ) is assigned to the lo ( to ) phonon mode .
the direction of the gauge field @xmath96 is perpendicular to the phonon eigenvector @xmath104 and the lo mode shifts the wavevector around the tube axis , which explains how the lo mode may induce a dynamical energy band - gap in metallic nanotubes .
@xcite by calculating eq .
( [ eq : mat_elph ] ) for the to mode with @xmath138 and for the lo mode with @xmath137 , we get eq .
( [ eq : coupling ] ) , too . equation ( [ eq : coupling ] ) is valid regardless of the tube chirality if the phonon eigenvector of the lo ( to ) phonon mode is in the direction along ( around ) the tube axis .
this is because @xmath140 and @xmath104 are transformed in the same way as we change the chiral angle . as a result
, there would be no chiral angle dependence for the el - ph coupling in eq .
( [ eq : coupling ] ) .
however , the phonon eigenvector depends on the chiral angle .
_ reported that , for a chiral nanotube , atoms vibrate along the direction of the carbon - carbon bonds and not along the axis or the circumference .
@xcite in the case of a chiral nanotube , the phonon eigenvector may be written as @xmath141 where @xmath142 ( @xmath143 ) is in the direction around ( along ) a chiral tube axis , and @xmath144 is the angle difference between the axis and the vibration .
this modifies eq .
( [ eq : coupling ] ) as @xmath145 in fig .
[ fig : chi ] , we show numerical results for @xmath47 as a function of @xmath5 for a @xmath146 metallic chiral nanotube for @xmath52 k. the energy difference between the minimum point at @xmath147 and at @xmath148 shows that the lo mode couples more strongly to the low energy electron - hole pair than the to mode . since the curvature - induced mini energy gap for @xmath146 and @xmath75 tubes are almost the same ( @xmath149 ev ) , we may expect a similar energy shift of the lo and to modes .
however , the results of the @xmath75 zigzag tube show that the to mode couples more strongly for the low energy electron - hole pair than the lo mode .
this is explained not by eq .
( [ eq : coupling ] ) but by eq .
( [ eq : coupling_chi ] ) with appropriate @xmath144 ( @xmath150 ) , which is given by the phonon eigenvector calculation .
since the chiral angle for @xmath146 is @xmath151 , @xmath144 is not directly related to the chiral angle .
the identification of @xmath144 in eq .
( [ eq : coupling_chi ] ) as a function of chirality would be useful to compare theoretical results and experiments , which will be explored in the future .
dependence of the lo ( red curve ) and to ( black curve ) phonon energy for the @xmath146 metallic chiral nanotube .
@xmath85 denotes the curvature - induced mini energy gap which has the same value as that for the @xmath75 tube in fig .
[ fig : zig](a ) .
the difference between the behavior of the @xmath146 and @xmath75 nanotubes comes from the fact that the phonon eigenvector of the lo ( to ) mode is not pointing along the tube axis in the case of the @xmath146 tube . ]
the phonon frequencies with some broadening as shown in fig .
[ fig : arm ] are not directly related to the raman spectra of the g - band but relate to the phonon density of states of the lo and to phonons at @xmath152 . in previous papers ,
@xcite we have shown that the g - band intensity depends on chiral angle in which for zigzag nanotubes the lo ( to ) phonon mode is strong ( suppressed ) while for armchair nanotubes the to ( lo ) phonon mode is strong ( comparable ) . this chirality dependence of the g - band intensity is also observed in single nanotube raman spectroscopy .
@xcite this chirality dependence of the g - band mode intensity is exactly opposite to the chirality dependence of the phonon - softening in which the to phonon mode is suppressed in armchair nanotubes .
these observations clearly show that the corresponding el - ph interaction for phonon - softening and raman processes are independent of each other . in the raman process
, the phonon is emitted by scattering an electron ( or an exciton ) from one conduction band state to another conduction band state , while in the phonon softening , the electron is scattered ( or excited ) from a valence band state to a conduction band state .
thus the corresponding matrix element for the raman process is expressed by substituting @xmath122 in eq .
( [ eq : mat_etoe ] ) for @xmath123 in eq .
( [ eq : mat_elph ] ) as follows , @xmath153 then the electron - phonon matrix element for the raman scattering process becomes @xmath154 it is stressed again that the el - ph interactions of eqs .
( [ eq : coupling ] ) and ( [ eq : coupling2 ] ) are for phonon - softening and raman intensity , respectively and that they are independent of each other .
when we consider the raman scattering processes for carbon nanotubes , the @xmath14 vector which is relevant to the resonance raman process is the @xmath14 vector at the van hove singular point , @xmath155 , which corresponds to the touching points of the equi - energy surface to the cutting line ( one dimensional brillouin zone plotted in the 2d bz ) @xcite which are shown in fig .
[ fig : tri](a ) and ( b ) for zigzag and armchair nanotubes , respectively . value defined from the @xmath63 axis in fig . [ fig : graphene ] is not zero while in the case of ( b ) zigzag nanotubes , @xmath156 is always zero , which explains the chirality dependence of the relative raman intensity of lo and to phonon modes . ] in the case of the zigzag nanotubes , the @xmath155 point corresponds to @xmath157 which is not the chiral angle , but is defined in fig .
[ fig : scattering](b ) .
thus from eq .
( [ eq : coupling2 ] ) , we see that the to mode is not excited . on the other hand , in the case of the armchair nanotube ,
the value of @xmath156 is not zero because of the trigonally warped equi - energy surface as shown in fig .
[ fig : tri](b ) .
@xcite thus not only the lo but also the to phonon modes can be observed , which is consistent with the previous experimental results @xcite and the theories .
@xcite thus when we see the g - band raman spectra for metallic nanotubes , we can expect the following features : for armchair nanotubes , we can see both lo and to in which the lo mode appears at a lower frequency region with some spectral width while the to mode appears without spectral broadening because of the absence of phonon softening . for zigzag nanotubes , we can see only a broadened lo phonon mode but we can not see a to phonon mode . for a general chiral nanotube , we generally observe both lo and to phonon modes with some broadening as a function of @xmath156 .
further , we expect some phonon hardening effect for the to phonon mode . the relative intensity between the lo and to phonon modes is determined by the chirality dependent raman intensity and spectral width .
the detailed raman spectral features for all metallic @xmath158 nanotubes will be presented elsewhere .
the gauge field descriptions for the lattice deformation ( eq . ( [ eq : a ] ) ) and for the el - ph interaction of the lo and to modes ( eq . ( [ eq : optigauge ] ) ) are useful to show the appearance of the curvature - induced mini energy gap in metallic zigzag carbon nanotubes and the decoupling between the to mode with a finite wavevector and the electrons , as shown in the following . for a zigzag nanotube , we have @xmath159 and @xmath160 from the rotational symmetry around the tube axis ( see fig .
[ fig : graphene ] ) .
then , eq . ( [ eq : a ] ) shows that for @xmath161 and @xmath162 , the cutting line of @xmath163 for the metallic zigzag nanotube is shifted by a finite constant value of @xmath164 because of the aharanov - bohm effect for the lattice distortion - induced gauge field @xmath165 .
this explains the appearance of the curvature - induced mini energy gap in metallic zigzag carbon nanotubes @xcite and of the phonon broadening for the to mode as a function of @xmath5 .
the to phonon mode with @xmath166 does not change the area of the hexagonal lattice but instead gives rise to a shear deformation .
thus , the to mode ( @xmath167 ) satisfies @xmath168 using eqs .
( [ eq : optigauge ] ) and ( [ eq : to ] ) , we see that the to mode does not yield the deformation - induced magnetic field ( @xmath169 ) , but the divergence of @xmath96 instead because @xmath170 thus , we can set @xmath171 where @xmath172 is a scalar function . since we can set @xmath173 in eq .
( [ eq : weyl ] ) by a redefinition of the phase of the wavefunction ( a local gauge symmetry ) as @xmath174 , @xcite and thus the @xmath96 in eq .
( [ eq : weyl ] ) disappears for the to mode with @xmath166 .
this explains why the to mode with @xmath166 completely decouples from the electrons and that only the to mode with @xmath175 couples with electrons .
this conclusion is valid even when the graphene sheet has a static surface deformation . in this sense ,
the to phonon mode at the @xmath0-point is anomalous since the el - ph interaction for the to mode can not be eliminated by a phase of the wavefunction
. it may be difficult to include the local gauge symmetry in a numerical analysis .
on the other hand , the lo phonon mode with @xmath166 changes the area of the hexagonal lattice while it does not give rise to a shear deformation .
thus , the lo mode ( @xmath176 ) satisfies @xmath177 using eqs .
( [ eq : optigauge ] ) and ( [ eq : lo ] ) , we see that the lo mode gives rise to a deformation - induced magnetic field as @xmath178 since a magnetic field changes the energy band structure of electrons , the lo mode ( with @xmath166 ) couples strongly to the electrons . in the case of 2d graphene , eq .
( [ eq : coupling ] ) tells us that the @xmath0 point to and lo modes give the same energy shift because the integral over @xmath61 gives the same @xmath11 in eq .
( [ eq : omega_2 ] ) for both to and lo modes .
this explains why no @xmath4-band splitting is observed in a single layer of graphene .
@xcite even when we consider the to and lo modes near the @xmath0 point , we do not expect any splitting between the lo and to phonon energies because the to mode with @xmath166 is completely decoupled from the electrons .
thus , for @xmath166 , only the lo mode contributes to the @xmath4-band . in our numerical results for @xmath47
, we see a crossing between the lo and to mode at some values of @xmath5 .
if there is a finite transition amplitude between the lo and to phonon modes , we need to diagonalize the following matrix according to second - order perturbation theory for a degenerate case , @xmath179 where @xmath180 is the effective hamiltonian in second - order perturbation theory given by @xmath181 using eq .
( [ eq : coupling ] ) , we see that @xmath182 which is an odd function of @xmath64 appears in the numerator for the diagonal terms in eq .
( [ eq : d2nd ] ) .
thus , the mode coupling through the electron - hole pair at @xmath64 is canceled by that at @xmath183 , i.e. , due to the time - reversal symmetry of the system .
thus , even though the lo and to modes cross each other , there is no - mixing between the lo and to phonon modes for any chirality .
if there is a breaking of time reversal symmetry for @xmath64 , we expect some mixing between the lo and to phonon modes .
the first comparison we make is to experimental results in the literature . here
we try to assign the chirality of the nanotube that is given in fig . 2 in ref .
this figure shows two strong intensity peaks with different phonon frequencies .
the higher frequency peak does not depend on the gate voltage and the lower frequency peak shows a frequency shift with a strong broadening near the dirac point .
the existence of a flat intensity peak as a function of fermi energy for a metallic swnt indicates that the electron - phonon coupling is very weak for the phonon mode .
thus , the nanotube can be thought of as an armchair swnt ( or close to an armchair swnt in the chiral angle ) since the el - ph coupling between the to mode and electron - hole pairs is negligible in this case .
further investigation comparing theory with experiment is strongly necessary for tubes of different chiralities . in fig .
[ fig : jingkong - data](a ) and ( b ) , we show experimental results of the g - band intensity as a function of applied gate voltage for two different isolated swnts . for the given excitation energy of 1.91ev and the observed radial breathing mode ( rbm ) frequencies 196 and 193 @xmath184 for the sample of fig . [
fig : jingkong - data](a ) and ( b ) , respectively , we can assign @xmath158 by using the conventional assignment technique that we used for single nanotube raman spectroscopy.@xcite as a result , we assign the @xmath158 values as ( 12,6 ) for fig . [
fig : jingkong - data ] ( a ) and ( 15,3 ) , ( 16,1 ) , or ( 11,8 ) for fig .
[ fig : jingkong - data](b ) . although
( 12,6 ) is not so close to the chiral angle for the armchair nanotube , we expect that figure [ fig : jingkong - data](a ) exhibits a similar behavior to that of fig . 2 in ref . .
thus the assignment of ( 12,6 ) from the rbm spectra is consistent with the present phonon spectra . in fig .
[ fig : jingkong - data](b ) , the intensity near the dirac point of the to ( lo ) mode is weaker ( stronger ) than shown in ( a ) .
this indicates that the tube is not an armchair swnt but close to a metallic zigzag swnt ( or a metallic chiral swnt ) .
thus we can assign this swnt either to ( 15,3 ) or to ( 16,1 ) for the swnt for fig .
[ fig : jingkong - data](b ) and we can exclude @xmath185 .
a further comparison with more experimental data will be required , which will be carried out in the future .
a chirality dependent raman intensity was observed for metallic swnts as shown at fig . 1 in ref . . although the @xmath5 positions for the observed metallic swnts are unclear , the results are consistent with our calculations in the following way .
for example , the to mode in a @xmath186 swnt gives a sharp lineshape and the lo mode shows a broad feature for a @xmath187 swnt .
it is noted that the raman intensity is proportional to the el - ph coupling for an optically excited electron via eq .
( [ eq : coupling2 ] ) .
since this el - ph coupling depends on the chiral angle due to the trigonal wrapping effect , one of the two optical modes may be invisible in the raman intensity .
@xcite in summary , we have calculated the @xmath5 dependence of @xmath47 ( @xmath188 and to ) for metallic carbon nanotubes .
the results show a strong dependence of the phonon frequency shift on the chirality of single - walled carbon nanotubes because of the curvature - induced shift of the wavevector around the tube axis .
this is explained by the general property of second - order perturbation theory and the characteristic electron - phonon coupling for the @xmath0 point lo and to phonon modes ( eqs .
( [ eq : coupling ] ) and ( [ eq : h_int ] ) ) . for the lo and
to phonon modes near the @xmath0 point , we showed the absence of an electron - phonon coupling for the to mode for @xmath166 due to a local gauge symmetry , and that the lo mode works as a deformation - induced magnetic field .
the phonon mixing between lo and to phonon modes is absent in second - order perturbation theory due to time - reversal symmetry .
r. s. acknowledges a grant - in - aid ( no .
16076201 ) from mext . mit authors acknowledge support under nsf grant dmr 07 - 04197 .
r. saito , a. grneis , ge . g. samsonidze , v. w. brar , g. dresselhaus , m. s. dresselhaus , a. jorio , l. g. canado , c. fantini , m. a. pimenta , and a. g. souza filho , new journal of physics * 5 * , 157.1157.15 ( 2003 ) .
r. saito , c. fantini , and j. jiang . in _
carbon nanotubes : advanced topics in the synthesis , structure , properties and applications _ , edited by ado jorio , m. s. dresselhaus , and g. dresselhaus , pages 233266 , springer - verlag , berlin , 2007 .
springer series on topics in applied physics , vol.111 . | the quantum corrections to the frequencies of the @xmath0 point longitudinal optical ( lo ) and transverse optical ( to ) phonon modes in carbon nanotubes are investigated theoretically . the frequency shift and broadening of the to phonon mode
strongly depend on the curvature effect due to a special electron - phonon coupling in carbon nanotubes , which is shown by the fermi energy dependence of the frequency shift for different nanotube chiralities .
it is also shown that the to mode near the @xmath0 point decouples from electrons due to local gauge symmetry and that a phonon mixing between lo and to modes is absent due to time - reversal symmetry .
some comparison between theory and experiment is presented . |
solar magnetic activity is widely believed to be associated with dynamo action somewhere in the solar convective shell .
identification of particular details of solar dynamo with surface manifestations of solar dynamo accessible for observations remains however a disputable problem .
the point is that many important details of dynamo action being hidden in solar interior can not be observed directly , and we have to learn about them basing on indirect tracers .
the link between a particular parameter important for a solar dynamo model and observable tracers of solar activity may be quite complicated , so adding new tests for comparison between concepts of solar dynamo and observations is a very attractive and simultaneously highly nontrivial undertaking .
recently a new physical entity solar small - scale magnetic field was suggested for observational verification of solar dynamo concepts .
solar magnetic field obviously contains some small - scale details which hardly can be included in the global solar magnetic field produced by traditional mean - field dynamo models . of course
, more detailed dynamo models based on direct numerical simulations of non - averaged mhd - equations give dynamo - driven magnetic configurations much more complicated than those produced by the mean - field models .
this fact is in agreement with theoretical expectations from mean - field models because certain terms in mean - field equations appear as a result of averaging of magnetic fluctuations .
the point however is that , apart from the mean - field ( or global ) dynamo based on a joint action of solar differential rotation and mirror - asymmetric convection , the dynamo theory ( in the framework of average description in terms of correlation tensor , see , e.g. , a review by zeldovich et al .
( 1990 ) , as well as in direct numerical simulations , see , e.g. , a review by brandenburg et al .
( 2012 ) predicts an additional mechanism of magnetic field self - excitation , so - called small - scale dynamo which produces small - scale magnetic field , i.e. , magnetic fluctuations . in other words
, the dynamo theory suggests that the total dynamo - driven magnetic field @xmath0 can contain the following contributions : mean magnetic field @xmath1 , small - scale magnetic field @xmath2 connected with the mean field @xmath1 , and small - scale magnetic field @xmath3 generated independently of @xmath1 .
theoretical distinction between @xmath2 and @xmath3 was known quite a long time ago ( see , e.g. , brandenburg & subramanian , 2005 ) .
however a general presumption was that it is more or less hopeless to distinguish between @xmath2 and @xmath3 observationally .
as a result , the relationship between @xmath2 and @xmath3 was very rare addressed in dynamo theories .
a naive theoretical expectation was that both sources of small - scale magnetic field contribute somehow in magnetic fluctuations ( as soon as fluctuations are ubiquitous in cosmic phenomena ) , however any attempt to distinguish between them hardly can be a fruitful undertaking .
modern progress in solar magnetic field observations ( e.g. , ishikawa et al .
, 2007 , lites et al .
, 2008 , a review by de wijn et al . , 2009 ,
abramenko et al . , 2010 , martinez gonzalez et al . , 2012 , to mention a few ) makes it possible to explore in detail various solar small - scale magnetic structures which often look quite specific in comparison with mean solar magnetic field .
a natural naive expectation appears that we clearly see imprints of the small - scale dynamo action , i.e. the field @xmath3 .
the fundamental conceptual progress here occurs due to the papers by stenflo ( 2012 , 2013 ) .
he critically examined `` the relative contributions of these two qualitatively different dynamos to the small - scale magnetic flux '' with the following conclusion : `` the local dynamo does not play a significant role at any of the spatially resolved scales , nearly all the small - scale flux , including the flux revealed by hinode , is supplied by the global dynamo '' .
we appreciate the importance of the clear and convincing statement of the problem formulated by stenflo ( 2012 , 2013 ) .
the point however is that , generally speaking , the small - scale dynamo can act in the solar interior and contribute very little into surface observables . on the contrary
, certain versions of the dynamo theory ( e.g. , cattaneo & tobias , 2014 ) predict that `` large - scale dynamo action can only be observed if there is a mechanism that suppresses the small - scale fluctuations '' .
so , we deal with a fundamental physical problem and it looks reasonable to spend efforts in order to find tiny surface spores of the small - scale dynamo action in the solar interior . in this paper , we suggest that an imprint of the small - scale dynamo action in solar interior can be hidden in statistics of sunspot groups which violate the hale polarity law .
the law states that in odd , for example , cycles , bipolar groups in the northern ( southern ) hemisphere have a positive ( negative ) magnetic polarity of the leading sunspot .
we refer to the groups , which violate this law as anti - hale groups .
the key issue of the suggested test is that observations ( as well as direct numerical simulations ) deal with the total magnetic field @xmath0 while the underlying physical problem is formulated in terms of the statistical quantities originated from the mean - field approach . and it is far from straightforward how to make inferences about the later in terms of the observed @xmath0 .
in other words , we need an explicit description of how we separate various contributions to the total magnetic field . in statistical studies ,
such a description is referred as a probabilistic model .
the start point of the observational test suggested is as follows .
it is well - known that through a solar cycle , the polarity of sunspot groups follows the so - called hale polarity law , i.e. , leading sunspots have opposite polarities in northern and southern hemispheres .
besides , from cycle to cycle , the leading - spot polarity changes .
the hale polarity rule reflects symmetry of the mean solar magnetic field .
we recall that in the framework of dynamo studies the number of sunspot groups is believed to be related to the mean field strength . of course , there are few sunspot groups which do not follow the hale polarity rule .
suppose that anti - hale groups can be considered as a result of some magnetic fluctuations which break symmetry of the mean magnetic field .
suppose also that the amplitude of fluctuations is governed by the mean field strength , as it should be if the small - scale dynamo does not work .
then we have to expect that the number of sunspot groups which follow the hale polarity law should be proportional to the number of anti - hale groups , and the relative number ( the percentage ) of the anti - hale groups should be cycle independent .
in contrast , if there is a substantial source of magnetic fluctuations independent on the mean field , then the number of anti - hale groups has to contain a cycle - independent component . in this case
, the relative number of anti - hale groups has to be enhanced during each solar minimum .
this expectation admits an observational verification ( at least in principle ) and looks reasonable at least in the first sight .
the point however is that realization of the above scheme is not a trivial task .
below we present the key features of our approach in the most simple form as a toy model of relationship between the sunspots polarity and two contributors to magnetic fluctuations .
let us consider a given solar hemisphere in a given solar cycle and assume that according to the hale polarity law toroidal component of the mean magnetic field ( which determines polarity of sunspot groups ) is directed , say , westwards .
let the magnetic field be organized in tubes of two types : the first type are tubes produced by the mean field , and the second type tubes are associated with the small - scale dynamo action . for the sake of simplicity , we adopt that the tubes are directed * longitudinally * , i.e. , oriented along the equator .
let the tubes of the first type contain the magnetic field of two components .
the fist component , @xmath4 , is a non - random flux directed westwards .
the number density of the tubes is denoted as @xmath5 , so that @xmath6 is the mean field @xmath1 .
the number density @xmath5 is modulated by 11-year cycle , which results in the 11-year modulation of @xmath1 .
at the same time , @xmath7 is time - independent .
we presume that sunspot formation is a threshold phenomenon with a threshold magnetic field strength , @xmath8 , slightly exceeding the magnitude of @xmath4 .
the other magnetic field component is a random ( say , gaussian ) magnetic field @xmath2 which is directed * longitudinally * as well , however its mean value vanishes , and the field is directed with the equal probabilities west- or eastwards .
the r.m.s .
value of this field , @xmath9 , is proportional to @xmath10 .
if @xmath2 is directed westwards and @xmath11 exceeds @xmath12 , then a magnetic tube arises on the solar surface and creates a sunspot group which follows the hale polarity law .
if , however , the field @xmath2 is directed eastwards and , by chance , it is so strong that @xmath13 exceeds @xmath8 , then the tube arises as well , but the newcomer violates the hale polarity law .
a simple estimates ( khlystova & sokoloff , 2009 ; sokoloff & khlystova , 2010 ) show that if @xmath14 , the number of anti - hale groups should be a few percent of the total number of sunspot groups .
note that in order to explain why sunspots arise at solar minima ( when the mean magnetic field is low ) , we have to suppose that @xmath15 is time - independent and the mean field cyclic modulation comes from the modulation of @xmath5 .
the tubes of the second type contain a random ( say , gaussian ) magnetic field @xmath3 with a cycle - independent r.m.s .
value @xmath16 and zero mean value .
they do not contribute to the mean magnetic field .
when @xmath17 exceeds @xmath18 , a tube arises , and the polarity orientation depends on the occasional tube s orientation : a hale polarity law group appears if @xmath3 was directed westward , and an anti - hale group appears if @xmath3 was directed eastward . obviously , the _ number _ of anti - hale groups originated from the first - type tubes depends on the cycle ( as soon as the total number of sunspots diminishes at solar minima ) , while that originated from the second - type tubes is cycle - independent .
on the contrary , the _ percentage _ of anti - hale groups originated from the first - type tubes does not depend on the cycle , while the percentage of that produced by the second - type tubes should reach its maximum value at the solar minima because the total number of sunspots is lowered at that times .
thus , exploring a presence / absence of cyclic modulation of the anti - hale groups , we can shed light on a relationship between the small - scale field @xmath3 , originated from the small - scale dynamo , and the small - scale field @xmath2 , produced by the mean - field dynamo action .
the above model is an obvious oversimplification .
we only include the features that are ultimately required to illustrate our observational test .
further steps will be to admit that the magnetic tubes might be directed not exactly * longitudinally * , to include temporal growth of magnetic field in a given tube due to the dynamo action , etc .
the mount wilson observatory ( mwo ) has carried out magnetic observations of sunspot groups throughout a century . on the basis of those observations ,
the catalog of magnetic classes of sunspot groups was compiled , which contains the information on number of anti - hale sunspot groups within certain time intervals .
unfortunately , the digital format of the complete mwo catalog does not exist yet .
monthly catalogs for 19201958 , which include the magnetic classification of sunspot groups , had been published in each issue of the _ _ publications of the astronomical society of the pacific _ _ starting with a paper titled : `` summary of mount wilson magnetic observations of sun - spots for may and june '' ( 1920 ) . from 1962 to 2009 , the catalog data had been published in the _ solar - geophysical data _
, however there were gaps in the information on anti - hale sunspot groups .
there is a digital mwo catalog covering the time interval from 1962 to 2004 .
this one is available at the website of the _ _ national geophysical data center _ _ , however the information on anti - hale groups is presented for the 19892004 time interval only . note that in the mwo catalog , normal and anti - hale sunspot groups are determined by the sign of magnetic field in the leading sunspots and by the group s tilt relative to the e w direction .
this method works well for a majority of bipolar groups .
however , uncertainties can appear in some cases .
thus , the tilt of the sunspot axis can be misaligned with the tilt of magnetic polarities , and as a result , a normal group with a rare unusual tilt can be classified as an anti - hale one ; groups of a singular sunspot can not be classified ; besides , in some cases , the identification of a group as an anti - hale one is ambiguous because the group can change the polarity orientation during its evolution . to frame a group from its neighbors sometimes is not a unique procedure ( see , for example , the magnetic complex ar noaa 09393/09394 ) .
we acknowledge the above shortcomings , however we decided to use the catalog data .
richardson ( 1948 ) studied in detail anti - hale sunspot groups from 1917 to 1945 using the mwo catalog data published in hale & nicholson ( 1938 ) and in _ publications of the astronomical society of the pacific_. he examined the anti - hale sunspot groups on the original records .
table 5 in his paper gives yearly numbers of normal and anti - hale sunspot groups in the northern and southern hemispheres separately for the 15th , 16th and 17th solar cycles .
richardson normalized the data : the number of sunspot groups observed at mwo during a year was multiplied by 365 and divided by the number of observation days in a year , which resulted in non - integer numbers of anti - hale groups in his table 5 .
the data from this table for 19171945 were used in the present study . during periods of cycles overlapping ,
the high - latitude groups are thought to belong to new cycle , and therefore , they have the opposite ( to the old cycle ) magnetic polarity of the leading sunspots . these groups must be marked as normal , not anti - hale ones .
having this in mind , richardson introduced the cycle separation of groups , so that during the cycles overlapping , the low - latitude groups belong to the old cycle , whereas the high - latitude groups belong to the new cycle .
the mark `` anti - hale '' was attributed depending on the cycle .
figure 1 of richardson ( 1948 ) demonstrates this for anti - hale sunspot groups .
we follow this rule in the present study . to obtain the annual number of normal ( anti - hale ) groups from table 5 from richardson ( 1948 ) , we combined the normal ( anti - hale ) groups of both hemispheres . besides , for years of cycles overlapping ( 19221924 , 19331935 , 19431945 ) , we combined low - latitude normal ( anti - hale ) groups of old cycle with high - latitude normal ( anti - hale ) groups of new cycle . as a result of these efforts , our first data set , named hereinafter as the 19171945 data set , was compiled .
our second data set was compiled on the basis of the above mentioned 19892004 mwo data . in detail , these data were observed from january 1 , 1989 to august 31 , 2004 .
they are available at the website of the _ national geophysical data center_. there , for anti - hale groups , the sign `` + '' is added to the magnetic class , according to the catalog description .
total 354 groups were marked as anti - hale groups .
( note , that during the time period under study , the data have gaps , namely , november 1995 , december 1998 , and january to march 2002 . )
we scrutinized each ( out of 354 ) group marked as `` anti - hale '' . using the data about the magnetic polarity in sunspots , we determined the polarity of leading spots .
the time - latitude diagram was compiled ( figure 1 , top frame ) . following the method used by mcclintock et al .
( 2014 ) , we separated wings of activity waves to determine the boundaries between cycles ( dashes lines in figure 1 ) .
the diagram shows that there are certain mistakes in attributing the `` anti - hale '' property ( e.g. , black dots are present in the midst of white , and visa versa ; two high - altitude sunspot groups of the new cycle were wrongly marked as anti - hale groups in 1996 ) .
this forced us to re - examine all groups on the diagram by means of full disk magnetograms of different observatories available at the dpd website ( gyri et al . , 2011 ) .
we found that about one - half of identifications were erroneous .
the mis - identifications are associated predominantly with long - lived spots in decaying active regions with formation of small spots or pores of opposite polarity to the west from the main spot , or mis - identifications of boundaries of sunspot groups in decaying activity complexes .
the corrected data set 19892004 was used to explore the anti - hale statistics in the next section ( figure 1 , bottom frame ) . in our study
, we also consider the results of mcclintock et al .
the authors analyzed magnetic tilt angles of sunspot groups along with information on the magnetic polarities of leading spots .
their statistics of anti - hale bipolar sunspot regions for 1974 - 2012 was obtained from the data of li & ulrich ( 2012 ) , who determined sunspot magnetic tilt angles from mwo sunspot records and daily averaged magnetograms , as well as from soho / mdi magnetograms .
the results for the 19171945 data set are shown in figure 2 .
the upper panel shows the annual average of sunspot groups and demonstrates 11-year solar activity cycle .
the lower panel represents the annual average percentage of anti - hale groups .
the percentage of anti - hale groups tends to be enhanced during the cycles minima .
the similar result was obtained for the 19892004 data set , figure 3 .
we found that the observed enhancement of the percentage becomes less pronounced when we average data over time intervals longer than 1 year . on the contrary , averaging over shorter time intervals seem to be interesting , and it is considered below .
.,scaledwidth=50.0% ] thus , figure 4 presents results for the 19892004 data set as derived from three different temporal averaging routines .
namely , the two top frames ( a ) show the outcome from one - month averaging , the two middle frames ( b ) present the result of two - month averaging , and , finally , the two bottom frames ( c ) refer to the three - month averaging outcome .
data gaps in november 1995 , in december 1998 and from january till march 2002 are visible .
the data show that during the time intervals of active sun , the percentage of anti - hale groups predominantly does not exceed the 7% level ( the dashed horizontal lines on the percentage frames ) , however , during the minimum between the 22nd and 23rd cycles , the value of percentage reaches up to 1525 % for different averaging intervals . from figure 4
one can conclude that the relative number of anti - hale groups overcomes the 7%-level when the total number of sunspot groups becomes low ( below approximately 20 , the level marked with the dashed line on the top frame of figure 4 ) .
we compare the above results with findings of mcclintock et al .
( 2014 ) who also reported an enhancement of percentage of anti - hale groups near the ends of solar activity cycles ( figure 5 ) .
slight differences between our results and that reported by mcclintock and colleagues are visible , however we presume that they might be due to different applied routines to outline a sunspot group .
note , that the local peak in the percentage of anti - hale groups in 1995 is present in both studies , compare figure 3 ( bottom frame ) and figure 5 ( top frame ) .
thus , an enhancement of the relative number of anti - hale groups during the solar minima can be regarded as a solid result .
we demonstrate that the relative number ( the percentage ) of sunspot groups which violate the hale polarity law ( anti - hale groups ) do increase during the minima of 11-year solar activity cycles . in accordance with the probabilistic model for the magnetic fields ,
suggested in section 2 , an increase of the relative number of anti - hale groups during the solar minima implies the small - scale dynamo at work , because the small - scale dynamo provides a cycle - independent income in the number of anti - hale groups and the total number of groups becomes lower during the solar minimum .
in other words , the observed statistics of anti - hale groups give a hint that small - scale dynamo is active in the solar interior .
this conclusion is compatible with the inferences made by stenflo ( 2012 , 2013 ) , who considers the small - scale dynamo as a very negligible contributor to the total solar magnetic flux because the anti - hale groups are associated with a tiny part of solar magnetic flux only ( see table 1 in wang & sheeley , 1989 ) .
we stress however that a verification of this hint in the framework of other approaches remains highly desirable .
the point is that the link between small - scale magnetic field in solar interior and statistics of anti - hale groups is very far from straightforward , and it is problematic to exclude firmly alternative interpretations of the result obtained . in particular , in mcclintock et al .
( 2014 ) , it is supposed that the enhancement of the relative number of anti - hale groups could be associated with the solar activity at low latitudes via interaction across the equator .
we appreciate this option and its importance for solar dynamo , however we suppose that it is insufficient to explain the observed effect because , during solar minima , anti - hale groups were observed at both low and intermediate / high latitudes ( see figure 1 and time - latitude diagrams in sokoloff & khlystova 2010 ; mcclintock et al . , 2014 ) .
d.s . and a.kh . are grateful for the rfbr financial support under the grant 15 - 02 - 01407 .
efforts of v.a . were supported by the program of the presidium of russian academy of sciences no .
abramenko , v. , yurchyshyn , v. , goode , p. , kilcik , a. , statistical distribution of size and lifetime of bright points observed with the new solar telescope , the astrophysical journal letters , volume 725 , issue 1 , pp .
l101-l105 , 2010 .
lites , b. w. , kubo , m. , socas - navarro , h. , berger , t. , frank , z. , shine , r. , tarbell , t. , title , a. , ichimoto , k. , katsukawa , y. , tsuneta , s. , suematsu , y. , shimizu , t. , nagata , s. , 2008 , apj , 672 , 1237 | in order to clarify a possible role of small - scale dynamo in formation of solar magnetic field , we suggest an observational test for small - scale dynamo action based on statistics of anti - hale sunspot groups . as we have shown , according to theoretical expectations the small - scale dynamo action has to provide a population of sunspot groups which do not follow the hale polarity law , and the density of such groups on the time - latitude diagram is expected to be independent on the phase of the solar cycle .
correspondingly , a percentage of the anti - hale groups is expected to reach its maximum values during solar minima . for several solar cycles , we considered statistics of anti - hale groups obtained by several scientific teams , including ours , to find that the percentage of anti - hale groups becomes indeed maximal during a solar minimum .
our interpretation is that this fact may be explained by the small - scale dynamo action inside the solar convective zone .
[ firstpage ] magnetic fields sun : activity sun : magnetic fields |
the underlying event plays an important rle in the simulation of particle collisions at hadron colliders .
the most important input for the simulation of the underlying event is the usage of a multiple partonic interaction ( mpi ) model that is capable to produce jets at large transverse momenutum via the simulation of multiple semi - hard partonic scatters @xcite . in herwig
, the mpi model has been implemented following the jimmy model @xcite and extending it with a soft component @xcite . here
, the soft component adds additional gluons with low transverse momentum according to a gaussian in @xmath0 that is also found in experimental data on the transverse momentum of very soft particles , i.e. particles with transverse momenta lower that a few gev .
these soft gluons clearly have to be seen as a technical vehicle to simulate soft physics rather than perturbative quanta of strong interactions .
they help us to connect the event on the parton level to the hadronization stage of the event . at this point ,
the colour structure of the event , given by the connections of individual gluons to the colour lines of partons from other events or the event remnants is most important for the further evolution of the event . given the lack of a first - principles approach to these soft interactions they were modeled as a random assignment of either a colour connection as in a hard event , i.e. with colour exchange , or as a peculiar event where the produced gluons are disrupted from the colour structure of the remainder of the event .
the weight towards one or the other structure was left as a tuning parameter and fixed with a rather large probability to produce colour disrupted events .
this modeling of soft events was intended as a smooth interpolation of the hard events into the soft region with the main motivation to avoid unphysical boundary effects in the simulation of mpi events , serving as the underying event in hard events . at the same time it also allowed for a glimpse into the regime of minimum - bias ( mb ) events where no hard scatter was present .
the model has turned out to be fairly stable whenever a not too soft selection as a minimum transverse momentum of at least 500mev or the requirement of a handful of charged particles in the event has been made . at the same time , the model obviously failed when it was applied to less restricted minimum - bias events .
this is easily explained as a result of i ) the rather poor modeling of soft particle production and ii ) the overall lack of a description of diffractive events @xcite .
these two points have now been addressed in our recent work @xcite that we report on at this conference . a particularly clear failure can be observed in fig .
[ fig : bump ] .
the observable @xmath1 describes the largest rapidity gap in the detector from any given track towards the end of the detector @xcite .
this gap will be large for diffractive events where the final state of one system disappears into the beam pipe while the other system leaves tracks in the detector .
the observed bump in the herwig simulation is a result of e.g. events with disrupted colour connections that result in large rapidity gaps
. this failure led us to further studies of the simulation of soft physics with our event generator with the conclusion that we need to replace the production of soft particles with a new , more sophisticated model and that we need a model for diffractive final states to complement this soft model .
the latter model is , despite the hadronic nature of diffractive interactions , formulated as a model of ( non - perturbative ) parton production that can be used to produce low - mass diffractive final states by means of the cluster hadronization model in herwig .
we make use of a regge factorized ansatz for the diffractive cross section with an exponentially falling @xmath2 dependence and a power - like @xmath3 dependence , where @xmath4 are the commonly known momentum transfer and diffractive mass of the system . once given these kinematics per event
, we can generate the production of a quark - diquark system with the given @xmath4 that will then hadronize into the desired diffractive final state .
for very low masses we directly produce a @xmath5 resonance .
the model for soft particle production is much more detailed and is based on the model for so - called multi - peripheral particle production@xcite .
again , we translate the model into our space of partonic degrees of freedom , where we produce soft gluons approximately flat in rapidity , i.e. according to the multi - peripheral kinematics . in order to be able to hadronize this final state and to avoid high - mass clusters in the system we have to add a quark pair to the system and then produce the soft gluons ordered in rapidiy as depicted in fig .
[ fig : colourconn ] .
one such ladder is the result of one soft interaction according to our mpi model .
i.e where we previously produced a single pair of gluons we now produce a whole ladder of soft gluons cf .
[ fig : ladderkin ] .
the number of gluons depends on the available phase space in rapidity and is chosen to keep the number of soft gluons approximately constant per rapidity interval .
the transverse momenta are chosen according to a gaussian , as in the previous model for soft interactions .
a detailed description of the model is given in @xcite .
we have tuned the single parameter of the new model to minimum - bias data at different centre - of - mass energies .
the embedding of the model takes place entirely within the framework of the old mpi model , i.e. we still have a mildly @xmath6 dependent minimum transverse momentum and a radius paremeter of the proton that both determine the number of soft interactions , i.e. the number of ladders .
the number of particles within the single ladder depends on the available rapidity span and needs a single normalization factor that will as well mildly rise with @xmath6 .
furthermore we have to adjust the fraction of diffractive events .
some results of these tunes are shown in fig .
[ fig : etamb ] .
we find that the rapidity distribution of charged particles in minimum - bias events is very well reproduced also for events with a significant fraction of diffraction , the data is taken from atlas @xcite .
these observables have been used to tune the model parameters . in fig .
[ fig : ptmb ] we show transverse momentum distributions of charged particles in minimum - bias events that are also very well described in the regime of soft particle production as measured by atlas @xcite .
our tune has been focused on this regime and the extension of this plot into the harder regime has to be properly adjusted for in the semi - hard events from the mpi that have not been adjusted in our study .
once again , the results turn out very nicely and show that we have much improved the description of this observable .
last but not least , we consider the rapidity gap fraction measurement that has initially motivated the extension of our model .
with this observable we have adjusted the relative weight of the diffractive vs.non-diffractive contributions to the total cross section . in fig .
[ fig : bumpnew ] we show the result in comparison to cms data @xcite .
we find that with the new model the bump disappears , as we do not have any more artificial rapidity gaps as a result of unphysical colour assignments to the soft final states . in fig .
[ fig : bumpcontribs ] we also show the individual contributions from the diffractive matrix elements vs the hard and soft mpi contribution , called minbias in the plot .
the mpi contribution shows the exponential - like fall off for small rapidity gaps which is hardened a bit by the mpi interactions .
the diffractive contribution is fairly flat throughout the considered range in @xmath1 .
we have presented the new model for soft interactions in herwig .
the model adds diffractive final states to the simulation of minimum - bias events and allows for the production of rapidity ordered gluons with very small transverse momentum .
we can improve all considered observables related to minimum - bias and diffractive events significantly with this model .
this model is only the first of a number of steps to improve the modeling of soft physics in herwig . | we introduce a new model for soft interactions in the monte carlo event generator herwig . we add a new model for the simulation of diffractive final states , based on the cluster hadronization model in herwig .
the soft component of the multiple partonic interaction model is replaced by a refined model for soft gluon production . with these two components we are for the first time able to give a full simulation of minimum - bias events at hadron colliders .
we briefly describe the models and present a few results obtained with it . |
over the years , the x - ray observatory has revealed a significant number of x - ray bright , kiloparsec - scale jets in active galactic nuclei ( agn ; e.g. , * ? ? ?
despite notable progress in understanding these objects , it is still unclear what is the composition of the plasma carrying the energy , how and where the jet particles are accelerated , and how relativistic the jets are in terms of their bulk velocity , @xmath6 , and doppler beaming factors , @xmath7^{-1}$ ] , with @xmath8 being the jet inclination and @xmath9 the jet bulk lorentz factor .
radiative models for the broad - band emission of large - scale quasar jets have also remained a matter of debate .
the two main candidate mechanisms for producing the jet x - ray emission are synchrotron and inverse compton scattering of the cosmic microwave background ( ic / cmb ) .
a robust determination of the jet energetics requires constraints on the relative importance of these two processes .
the ic / cmb scenario implies particle - dominated and highly relativistic outflows ( @xmath10 ) , which do not suffer severe deceleration or energy dissipation between sub - pc and kpc scales ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , while the synchrotron interpretation can be reconciled with highly - magnetized and possibly slower jets on kpc distances from the quasar cores ( e.g. , * ? ? ?
* ; * ? ? ?
the ic / cmb model predicts an increase in the x - ray to radio flux ratio with redshift , due to the amplification of the cmb energy density ( e.g. , * ? ? ?
* ; * ? ? ?
* ) ; neglecting a weak dependance on the spectral slope of the non - thermal continuum , @xmath11_x}{[\nu f_{\nu}]_r } \propto \left(\frac{u'_{\rm cmb}}{u'_{\rm b}}\right)\,\left(\frac{\delta}{\gamma}\right)^2 \propto(1+z)^{4}\,\left(\frac{\delta}{b'}\right)^2\ , , \ ] ] where @xmath12 is the observed energy flux spectral density , @xmath13erg@xmath14 is the cmb energy density in the jet rest frame ( denoted hereafter by primes ) , and @xmath15 is the comoving energy density of the jet magnetic field . studying high - redshift jets can thus provide important clues on the mechanism responsible for their observed x - ray emission .
however , very few high - redshift x - ray jets are known to date @xcite .
therefore , variations of the jet beaming , @xmath16 , and jet magnetic field strength , @xmath17 , both from system to system and between different knots along the same jet , can introduce a scatter that may mask the expected @xmath18 increase even when the ic / cmb model is the dominant emission mechanism ( see * ? ? ?
increasing the sample of high - redshift quasar jets with good - quality x - ray data will not only allow us to distinguish between competing radiative models ; by combining in - depth studies of the cosmological evolution of the jet properties from the epoch of the quasar formation up to the present day with our current understanding of black hole growth and the evolution of black hole spin ( e.g. , * ? ? ?
* ) , we can shed new light on the jet launching mechanism , as well as on the evolution of the intergalactic medium that the jets propagate through . here
, we report the discovery of an extended , x - ray bright , radio faint black hole jet associated with the quasar b30727 + 409 . based on the sloan digital sky survey ( sdss ) data release 9 @xcite
, the spectroscopic redshift of this source is @xmath19 . assuming a @xmath20cdm cosmology with @xmath21 , @xmath22 , and @xmath23kms@xmath24mpc@xmath24 , the redshift of b30727 + 409 corresponds to the luminosity distance of @xmath25gpc and the conversion scale @xmath26 kpc/@xmath27 .
[ [ section ] ] the x - ray jet associated with b30727 + 409 was discovered in a relatively short ( 20 ks ) observation from 2014 december 15 ( obsid 17167 ) , targeting the nearby galaxy cluster abell585 ( @xmath28 ; see * ? ? ?
the brightest cluster galaxy of abell585 is located @xmath29 northwest of the quasar , so that both were observed in the @xmath30 field - of - view of the advanced ccd imaging spectrometer ( acis ) s3 chip .
we reprocessed the standard level 1 event lists produced by the pipeline in the standard manner , using the ciao ( version 4.7 ) software package to include the appropriate gain maps and updated calibration products .
bad pixels were removed and standard grade selections applied .
the information available in vfaint mode was used to improve the rejection of cosmic ray events .
periods of anomalously high background were excluded by examining the light - curve of the observation in the @xmath31kev energy band using the standard time binning methods recommended by the x - ray center .
the net exposure time after cleaning is 19ks . inspecting the very large array ( vla ) images of b30727 + 409 previously presented in @xcite that probe a range of spatial scales @xmath32
, we did not find any significant radio emission associated with the x - ray jet .
we therefore analyzed an additional set of vla data from the nrao archive from 2007 august 6 , when it was observed as a calibrator in program al696 .
the observations were collected in the most extended a - configuration and consisted of five short scans centered at 1.43ghz and three scans at 4.86ghz .
the data were calibrated with aips @xcite following standard procedures with amplitude calibration utilizing a scan of 3c286 at 1.43ghz and 3c147 at 4.86ghz .
after editing , the total exposure times for b30727 + 409 were approximately 9 and 4 minutes at the respective frequencies .
the ( @xmath33 ) data were imported into difmap @xcite for phase and amplitude self - calibration .
clean images were convolved with circular gaussians with full - width half maxima matched to the natural weighted beam sizes of @xmath34 at 4.86ghz and @xmath35 at 1.43ghz ( 0.09 and 0.14 mjy beam@xmath24 off - source rms , respectively ) .
we also reimaged the vla 4.71 ghz b - array dataset from @xcite to provide an image with the same @xmath35 beam as the 1.43 ghz image . )
.,scaledwidth=50.0% ] [ imgx ]
figure[imgx ] shows a raw count map of the region around b30727 + 409 in the @xmath36kev energy band . because the jet is relatively bright and compact ,
no vignetting or background corrections were applied .
the image reveals a clear extension of @xmath37 northwest of the quasar s core , corresponding to a projected length of @xmath2kpc at @xmath0 .
note that there is no visible charge - transfer streak in the acis - s3 read - out direction ( offset from the x - ray jet by @xmath38 clockwise ) that would indicate significant pile - up of counts from the bright core . in our 4.86ghz vla @xmath34 resolution map ,
we confirm the detection of a radio feature @xmath39 from the quasar core found by @xcite using the same data .
we find no significant additional emission coinciding with the rest of the x - ray jet .
figure[imgxr ] shows a image of the jet , rebinned to 0.1 acis pixels and smoothed with a @xmath40 gaussian filter , overlaid with 4.86ghz radio contours .
note that very long baseline interferometry maps show apparent superluminal motion up to @xmath41 in a one - sided milliarcsecond - scale jet oriented toward the extended structure seen in x - rays , indicating relativistic beaming on small scales @xcite . for the position on the detector corresponding to the quasar core
, the 90% enclosed - counts fraction aperture of the point - spread function ( psf ) has a radius of @xmath42 .
we extracted x - ray spectra from a @xmath43-long rectangular region with a width of @xmath44 , starting at a minimum distance of @xmath45 and extending out to @xmath46 from the quasar core .
we fit the resulting spectrum in the 0.67.5 kev energy range with an absorbed power - law model , with the hydrogen column density fixed to the galactic value , @xmath47@xmath48 @xcite , assuming no intrinsic absorption .
the background was estimated from an annulus centered on the x - ray peak , with inner and outer radii of @xmath49 and @xmath50 .
the best - fit power - law index is @xmath51 , with a corresponding flux density at 1 kev of @xmath52 njy .
this extraction region does not include the radio knot at @xmath39 , which is only marginally resolved from the quasar core with ( section [ sect_knot ] ) .
the extended jet is thus remarkably radio faint , with no visible emission in both vla images . to estimate upper limits
, we used the @xmath35-beam images and defined four adjacent @xmath53 apertures elongated along the x - ray jet direction ( position angle @xmath54 ) , centered 3.1 , 5.4 , 7.7 , and 10.0 from the quasar . using the aips task uvsub to subtract the cores from the ( @xmath33 ) data , as well as the modeled gaussian representing the @xmath4 knot in the 1.43 ghz data ( below ) , the @xmath55 point source upper limits in each of the respective apertures were @xmath56 , @xmath56 , @xmath56 , and @xmath57mjy at 1.43ghz , and @xmath58 ( larger due to contamination from the adjacent @xmath59 knot ) , @xmath60 , @xmath57 , and @xmath60mjy at 4.71ghz . assuming the radio jet is composed of a series of four unresolved radio knots within the defined apertures , the total radio emission co - spatial with the x - ray jet visible beyond the @xmath4 knot is thus @xmath61 and @xmath62 mjy , corresponding to @xmath63_x/[\nu f_{\nu}]_r>205 $ ] and @xmath64 at 1.43 and 4.71ghz , respectively .
in addition , we detect excess x - ray emission ( a total of 10 counts ) located @xmath3 from the quasar core ( 280 kpc , projected ) , along the same position angle as the jet ( figure [ imgxr ] ) . the hypothesis that this excess of counts is solely due to poisson fluctuations around the average background level estimated from a neighbouring region
is ruled out at the @xmath65-value of @xmath66 ; since it lacks any sdss counterpart , this x - ray feature may therefore likely represent the terminal hotspot of the b30727 + 409 jet .
the corresponding @xmath55 point source upper limits at the location of this x - ray hotspot are @xmath56 mjy ( 1.43 ghz ) and @xmath60 mjy ( 4.71 ghz ) . by measuring the count rates in four partial annuli with opening angles of @xmath67 and spanning @xmath68 from the quasar core , we estimate that the azimuth corresponding to the radio knot contains @xmath69 x - ray counts above the surface brightness level determined from the other three azimuths .
assuming this emission follows the same power - law index as the extended jet , this implies the knot s flux density at 1 kev is @xmath70 njy , and the 210 kev luminosity of the entire jet ( summing the knot and extended jet ) is @xmath71ergs@xmath24 . from the vla data , we measured a flux density for the knot of 1.5mjy at 4.86ghz by fitting a circular gaussian component in the ( @xmath33 ) plane with the modelfit program in difmap .
though at the edge of the beam of the bright core in the lower - resolution 1.43ghz image , that same knot is clearly visible in the clean components , and we measured a flux density of 4.5mjy . in the presence of the bright core
, we estimate the uncertainties in the flux densities are @xmath72 at both frequencies .
the resultant spectral index of the knot is @xmath73 .
the broad - band emission from the knot therefore corresponds to the flux density ratio @xmath63_x/[\nu f_{\nu}]_r\simeq15 $ ] , with very similar values at both frequencies . [ cols="^,^,^,^,^",options="header " , ] [ tab_flux ] we also extracted spectra from a circular region of radius @xmath45 centered around the quasar core ; subtracting the local background as described above and fitting the resulting spectrum with an absorbed power - law , we obtain a best - fit index @xmath74 and a 210 kev rest frame luminosity of @xmath75 ergs@xmath24 . no intrinsic absorption in addition to the galactic @xmath76
is required by the fit .
all our x - ray and radio flux measurements are summarised in table [ tab_flux ] . , smoothed with a @xmath40 gaussian filter .
vla 4.86ghz contours at 0.29 , 0.58 , and 0.87 mjy beam@xmath24 are shown in green .
we have corrected for a slight offset between the x - ray and radio contours due to the astrometric error.,scaledwidth=50.0% ]
a non - thermal origin provides the most natural explanation for the jet x - ray emission .
we modeled the radio to x - ray spectrum of the large - scale jet in b30727 + 409 with the ic / cmb scenario assuming a broken power - law shape of the electron energy distribution , @xmath77 for @xmath78 , and @xmath79 $ ] for @xmath80 , where @xmath81 is the electron lorentz factor .
this parametrization takes into account the expected break in the electron spectrum resulting from radiative cooling .
we also assume pressure equipartition between the emitting electrons and the jet magnetic field , and allow for a heavy jet content with one @xmath82 pair per 110 @xmath83 pairs ( see * ? ? ?
finally , we assume a spherical geometry for the @xmath4 knot with a @xmath84kpc radius ( consistent with the knot s radio extent in the vla 4.86ghz image ) , and a cylindrical geometry for the extended jet with the same radius and a length of @xmath85kpc .
knot , along with the corresponding synchrotron plus ic / cmb model curves.,scaledwidth=50.0% ] the broad - band jet spectrum can be described by the model shown in figure [ sed ] , returning reasonable parameters : the jet inclination @xmath86 , the jet bulk lorentz factor @xmath87 ( meaning @xmath88 ) , the jet magnetic field decreasing slowly along the outflow from @xmath89@xmath90 g down to @xmath91@xmath90 g , a `` standard '' form of the electron energy distribution with @xmath92 , @xmath93 , @xmath94 , and the cooling break decreasing from @xmath95 at the position of the @xmath4 knot down to @xmath96 at the position of the outer jet .
although all parameters are currently only weakly constrained by the data , this model seems to favour a high total jet kinetic power of the order of @xmath97ergs@xmath24 ( depending on the exact proton content ) , and highly relativistic jet bulk velocities maintained over very large scales of order the de - projected source size , @xmath98kpc .
however , the estimated jet kinetic power further depends on various assumptions in our model : for a purely leptonic jet , @xmath99 would be smaller than cited here , while if the magnetic field energy density is below the equipartition value , @xmath99 would increase , while the required relativistic beaming would decrease .
note also that , alternatively , a synchrotron origin of the x - ray emission is not excluded by the data ( however , see the discussion at the end of section [ cosmo ] ) . on the other hand , a thermal origin of the x - ray emission from the extended jet region is implausible : the best - fit thermal model ( employing the apec model with a metallicity fixed at 0.2 solar ) implies a best - fit temperature of @xmath100kev , which would be unusually high for any truly diffuse structure at @xmath0 . also the corresponding electron density , @xmath101 , and total mass would be exceptional : assuming the same cylindrical geometry as above , the best - fit emission measure of the apec model translates to @xmath102@xmath14 and @xmath103 , with @xmath104 denoting the filling factor of the gas ( with respect to the assumed cylindrical geometry ) .
such large amounts of very hot gas aligned with radio jets have not been observed in luminous quasars , although see @xcite for the @xmath105 radio galaxy pks1138 - 262 . in the past
, several claims have been made for distinct jet properties of high - redshift quasars when compared with their low - redshift analogs .
for example , @xcite argued for a decrease of the average bulk lorentz factor of high - redshift jets , because of an apparent deficit of luminous sdss radio - loud quasars above @xmath106 with respect to the model predictions based on the swift / bat ( burst alert telescope ) hard x - ray survey .
additionally , @xcite demonstrated that the `` radio - loudness '' ( i.e. , the ratio of the 5ghz core radio flux to the @xmath17-band core flux ) of the sdss@xmath107first ( faint images of the radio sky at twenty cm ) quasar population increases with redshift .
b30727 + 409 appears particularly interesting in this context for several reasons .
first , the x - ray luminosity of its large - scale jet is only 56 times lower than the x - ray luminosity of the core .
this is in contrast to the lower - redshift quasars targeted with , for which the jet to core x - ray luminosity ratio is typically @xmath108 , or lower ( see * ? ? ?
thus , the active nucleus of b30727 + 409 seems _ under - luminous _ in x - rays with respect to the emission of its large - scale outflow .
second , the unresolved core of b30727 + 409 appears particularly radio - loud for its accretion rate . using the sdss spectrum ,
@xcite estimate the mass of the central black hole to be between @xmath109 ( from the mgii line ) and @xmath110 ( based on the civ line ) .
the bolometric luminosity of the accreting matter estimated from the mgii line is @xmath111ergs@xmath24 , meaning the accretion rate is @xmath112 in eddington units , for the standard @xmath113 radiative efficiency of the accretion disk .
given this rate , b30727 + 409 appears to be characterized by a surprisingly large radio - loudness parameter of @xmath114 , at least 100 times larger than local quasars with comparable @xmath115 ratios ( see * ? ? ?
note that our ic / cmb modeling implies moreover @xmath116 , which is consistent with a very high ( maximum ) efficiency of the jet production in high-@xmath117 sources @xcite . _
x/[\nu f_{\nu}]_r$ ] ratio vs. redshift for x - ray quasar jets detected with ( adapted from * ? ? ?
* with the addition of two subsequently detected @xmath118 examples ) .
b30727 + 409 is shown in red ; the triangle represents the 3@xmath119 lower limit for the part of the jet lacking radio detection .
curves indicate the expected energy flux ratio in the framework of the ic / cmb scenario for given combinations of @xmath120 and @xmath16 ( see equation1 ) .
different knots from the same jet are connected by thin vertical lines.,scaledwidth=45.0% ] finally , in figure [ nufnu ] , we compare the measurements of @xmath63_{x(1\:kev)}/[\nu f_{\nu}]_{r(1.4\:ghz)}$ ] for b30727 + 409 to those of previously published large - scale x - ray jets detected with .
the x - ray to radio luminosity ratios for our target appear consistent with the amplification expected from the ic / cmb model in the case of large jet bulk velocities .
this seems to disfavour a synchrotron origin of the jet x - ray emission in b30727 + 409 . nonetheless , the statistical uncertainties are large due to the relatively short exposure time , and the redshift distribution of jets at @xmath121 is still rather sparsely sampled .
deeper x - ray and radio data will provide significantly improved constraints on the jet emission model for this quasar .
we report on the serendipitous discovery of an extended ( @xmath37 , or @xmath2kpc projected ) x - ray jet associated with the @xmath0 quasar b30727 + 409 , for which the archival vla maps reveal no radio counterpart ( except for a single knot @xmath39 from the quasar core ) .
a possible x - ray hot spot is identified @xmath3 ( 280kpc ) from the quasar .
the remarkable x - ray luminosity of the structure , @xmath122erg@xmath123 , implies a very efficient production of non - thermal x - ray photons by ultrarelativistic jet electrons .
the large x - ray to radio luminosity ratio supports the scenario in which this is the inverse - comptonization of the cmb photons which dominates radiative outputs of large - scale jets in the x - ray domain ( at least in core - dominated quasars , since in the cases of sources observed at larger jet viewing angles , i.e. lobe - dominated quasars and frii radio galaxies , the situation may be more complex ; see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
if correct , this would further imply a highly relativistic bulk velocity ( @xmath124 ) maintained on hundreds - of - kpc scales , even at high redshifts , and a very high efficiency of the jet production ( @xmath125 ) already during the epoch of quasar formation .
the serendipitous discovery of such an object may signal the existence of an entire population of similar systems with bright x - ray and faint radio jets at high redshift , which will have been missed by the current observing strategies that mostly focus on follow - up of known bright radio jets .
similar predictions for the ubiquity of x - ray bright , radio faint _ lobes _ at @xmath126 were put forward by e.g. @xcite and @xcite .
the seemingly different properties of this source compared to local quasars suggest that this selection bias must be accounted for when drawing any conclusions about the redshift evolution of jet properties and indeed about the cosmological evolution of supermassive black holes in general .
.s . and m.j .
were supported by polish nsc grants dec-2012/04/a
/ st9/00083 and dec-2013/09/b / st9/00599 , respectively . y.i . acknowledges a grant - in - aid for japan society for the promotion of science ( jsps ) fellows .
c.c.c . was supported at nrl by nasa dpr s-15633-y .
a.sie . was supported by nasa contract nas8 - 03060 to the x - ray center . n.w .
acknowledges nasa grant go5 - 16127x . , m. , perlman , e. s. , uchiyama , y. , cheung , c. c. , coppi , p. s. , georganopoulos , m. , worrall , d. m. , birkinshaw , m. , sparks , w. b. , marshall , h. l. , stawarz , l. , begelman , m. c. , odea , c. p. , & baum , s. a. 2013 , , 773 , 186 | a recent observation of the nearby galaxy cluster abell585 has led to the discovery of an extended x - ray jet associated with the high - redshift background quasar b30727 + 409 , a luminous radio source at redshift @xmath0 .
this is one of only few examples of high - redshift x - ray jets known to date .
it has a clear extension of about @xmath1 , corresponding to a projected length of @xmath2kpc , with a possible hot spot located @xmath3 from the quasar .
the archival high resolution vla maps surprisingly reveal no extended jet emission , except for one knot about @xmath4 from the quasar .
the high x - ray to radio luminosity ratio for this source appears consistent with the @xmath5 amplification expected from the inverse compton radiative model .
this serendipitous discovery may signal the existence of an entire population of similar systems with bright x - ray and faint radio jets at high redshift , a selection bias which must be accounted for when drawing any conclusions about the redshift evolution of jet properties and indeed about the cosmological evolution of supermassive black holes and active galactic nuclei in general . |
there are two families of interesting plane bipartite graphs , hexagonal systems @xcite and polyomino graphs @xcite , which often arise in some real - world problems .
a hexagonal system with a perfect matching is viewed as the carbon - skeleton of a benzenoid hydrocarbon @xcite .
the dimer problem in crystal physics is to count perfect matchings of polyomino graphs @xcite .
consider an @xmath3 chessboard @xmath4 ( i.e. , one special type of polyomino graphs ) with @xmath5 even . in how many ways can this board be covered by dominoes ( @xmath6 rectangles or dimers ) ?
in graph theoretics terms , this problem is equivalent to the counting of perfect matching of its inner dual graph ( for the definition of inner dual graph , see @xcite ) .
a _ domino pattern _ of @xmath4 is obtained by paving or tiling dominoes .
there is a one - to - one correspondence between the domino patterns of @xmath4 and perfect matchings of the inner dual of @xmath4 .
kasteleyn @xcite developed a so - called `` pfaffian method '' and derived the explicit expression of the number of ways for covering @xmath4 by dominoes .
in general , perfect matching existence @xcite , elementary components @xcite , matching forcing number @xcite and maximal resonance @xcite of polyomino graphs have been investigated .
in addition , polyomino graphs are also models of many interesting combinatorial subjects , such as hypergraphs @xcite , domination problem @xcite , rook polynomials @xcite , etc .
the concept of resonance originates from the conjugated circuits method @xcite .
conjugated or resonant circuits ( i.e. alternating cycles in mathematics @xcite ) appeared in clar s aromatic sextet theory @xcite and randi s conjugated circuit model @xcite .
klein @xcite emphasized the connection of clar s ideas with the conjugated circuits method . in the clar s aromatic sextet theory @xcite
, clar found that an aromatic hydrocarbon molecule with larger number of mutually resonant hexagons is more stable .
let @xmath7 be a plane bipartite graph with a _ perfect matching _ ( or kekul structure in chemistry ) @xmath8 ( i.e. a set of independent edges covering all vertices of @xmath7 ) .
a cycle of @xmath7 is called an @xmath8-_alternating cycle _ if its edges appear alternately in @xmath8 and off @xmath8 . a face @xmath9 is said to be @xmath8-_resonant _ or -_alternating _ if its boundary is an @xmath8-alternating cycle .
let @xmath10 be a set of finite faces ( the intersection is allowed ) of @xmath7 .
@xmath10 is called an @xmath8-_alternating set _
if all faces in @xmath10 are @xmath8-resonant .
further , @xmath10 is called an @xmath8-_resonant set _ of @xmath7 if the faces in @xmath10 are mutually disjoint and @xmath10 is an @xmath8-alternating set . simply , @xmath10 is a _ resonant set _ and alternating set of @xmath7 if @xmath7 has a perfect matching @xmath8 such that @xmath10 is an @xmath8-resonant set and @xmath8-alternating set respectively .
the cardinality of a maximum resonant set of @xmath7 is called the _ clar number _ of @xmath7 , denoted by @xmath11 . in 1985
, zheng and chen @xcite gave an important property for a maximum resonant set of a hexagonal system .
[ thm-1 - 1 ] _
@xcite _ let @xmath0 be a hexagonal system and @xmath1 a maximum resonant set of @xmath0 .
then @xmath2 has a unique perfect matching .
a _ forcing set _ of a perfect matching @xmath8 of a graph @xmath7 is a subset @xmath12 such that @xmath13 is not contained in any other perfect matching of @xmath7 . the _ forcing number _ of a perfect matching @xmath8 , denoted by @xmath14 ,
is the cardinality of a minimum forcing set of @xmath8 .
the _ maximum forcing number _ of @xmath7 is the maximum value of forcing numbers of all perfect matchings of @xmath7 , denoted by @xmath15 .
the concept of forcing number of graphs was originally introduced for benzenoid systems by harary et al . @xcite .
the same idea appeared in an earlier paper @xcite of klein and randi by the name `` innate degree of freedom '' .
most known results on forcing number are referred to @xcite .
pachter and kim revealed a minimax result that connects the forcing number of a perfect matching and its alternating cycles as follows .
[ thm-1 - 2 ] _ @xcite _ let @xmath7 be a plane bipartite graph with a perfect matching . then for any perfect matching @xmath8 of @xmath7 , @xmath16 , where @xmath17 denote the maximum number of disjoint @xmath8-alternating cycles in @xmath7 . by combining theorems [ thm-1 - 1 ] and [ thm-1 - 2 ] , xu et al .
@xcite obtained a relation between the forcing number and clar number of a hexagonal system as follows .
[ thm-1 - 3 ] _
@xcite _ let @xmath0 be an elementary hexagonal system
. then @xmath18 . an alternating set of a graph @xmath7
is called _ maximal _ if it is not properly contained in another alternating set of @xmath7 . in 2006 , salem and abeledo obtained the following result . _
let @xmath0 be a hexagonal system and @xmath1 a maximal alternating set of @xmath0 .
then @xmath2 has a unique perfect matching . motivated by the above work
, we will naturally investigate polyomino graphs .
this paper is mainly concerned with a maximum resonant set of a polyomino graph . by applying zheng and chen
s approach @xcite , we prove that if @xmath1 is a maximum resonant set of a polyomino graph @xmath7 , then @xmath19 has a unique perfect matching . for a maximal alternating set of @xmath7 , this property still holds . as a corollary
, we have that the maximum forcing number of a polyomino graph is equal to its clar number .
based on these results , it can be shown that the maximum forcing number of a polyomino graph can be computed in a polynomial time , and thus conforms the conjecture proposed by xu et al .
a _ polyomino graph _ is a connected finite subgraph of the infinite plane grid such that each interior face is surrounded by a regular square of side length one and each edge belongs to at least one square @xcite . for a polyomino graph @xmath0 , the boundary of the infinite face of @xmath0
is called the _ boundary _ of @xmath0 , denoted by @xmath20 , and each edge on the boundary is called a _ boundary edge _ of @xmath0 .
it is well known that polyomino graphs are bipartite graphs . for convenience
, we always place a polyomino graph considered on a plane so that one of the two edge directions is horizontal and the other is vertical .
two squares are _ adjacent _ if they have an edge in common .
a vertex of @xmath0 lying on the boundary of @xmath0 is called an _ external vertex _ , and a vertex not being external is called an _ internal vertex_. a square of @xmath0 with external vertices is called an _ external square _ , and a square with no external vertices
is called an _ internal square_. in what follows , we always restrict our attention to polyomino graphs with perfect matchings .
let @xmath7 be a graph with a perfect matching @xmath8 and an @xmath8-alternating cycle @xmath21 .
then @xmath22 is also a perfect matching of @xmath7 and @xmath21 is an @xmath23-alternating cycle of @xmath7 @xcite .
let @xmath8 and @xmath24 be two perfect matchings of a graph @xmath7 .
the _ symmetric difference _ of @xmath8 and @xmath24 , denoted by @xmath25 , is the set of edges contained in either @xmath8 or @xmath24 , but not in both , i.e. , @xmath26 .
an @xmath27-alternating cycle of @xmath7 is a cycle whose edges are in @xmath8 and @xmath24 alternately .
it is well known that the symmetric difference of two perfect matchings @xmath8 and @xmath24 of @xmath7 is a disjoint union of @xmath27-alternating cycles .
we now state our main result as follows .
[ thm-2 - 1 ] let @xmath0 be a polyomino graph with a perfect matching , and @xmath1 be a maximum resonant set of @xmath0 .
then @xmath2 has a unique perfect matching . before proving the main theorem
, we will deduce the following crucial lemma .
let @xmath7 be a plane bipartite graph , @xmath1 a set of finite faces and @xmath0 a subgraph of @xmath7 .
by @xmath28 we always mean the intersection of @xmath1 and the set of faces of @xmath0 .
[ lem-2 - 2 ] let @xmath0 be a @xmath29-connected polyomino graph with a perfect matching , @xmath1 a resonant set consisting of internal squares of @xmath0 .
if @xmath30 has a perfect matching or is an empty graph , then @xmath1 is not a maximum resonant set . and @xmath31 , and edges @xmath32 and @xmath33 with @xmath34 , n(1)=4 , n(2)=3 , n(3)=1 , n(4)=2 , n(5)=0 , n(6)=1
. and @xmath35.,height=315 ] we denote by @xmath8 a perfect matching of @xmath30 , and decompose the edges of @xmath20 into two perfect matchings @xmath36 and @xmath37 of @xmath20 since @xmath20 is an even cycle . then it is clear that @xmath38 and @xmath39 are two perfect matchings of @xmath2 .
let @xmath40 be a perfect matching of @xmath1 such that each edge of @xmath40 is vertical .
then @xmath41 is a perfect matching of @xmath42 .
moreover , @xmath43 and @xmath44 are two perfect matchings of @xmath0 .
suppose to the contrary that @xmath1 is a maximum resonant set of @xmath0 .
adopting the notations of @xcite , we can take a series of external squares @xmath45 which satisfy that neither square @xmath46 nor square @xmath47 is contained in @xmath0 as shown in fig .
[ fig : polyomino-01 ] .
we denote edges , if any , by @xmath32 , @xmath33 , @xmath48 and @xmath49 , and denote the square with edge @xmath33 which is adjacent to @xmath50 , if any , by @xmath31 , @xmath48 and @xmath49 , as shown in fig .
[ fig : polyomino-01 ] .
we first prove the following claims .
0.2 cm * claim 1 . * for a pair of parallel edges @xmath51 and @xmath52 of a square @xmath53 of @xmath0 , they do not lie simultaneously on the boundary of @xmath0 .
0.2 cm if @xmath51 and @xmath52 lie on the boundary of @xmath0 , then @xmath54 or @xmath37 , say @xmath36 .
so the square @xmath53 is @xmath36-alternating , and @xmath55 is a resonant set of @xmath0 , which contradicts that @xmath1 is a maximum resonant set of @xmath0 .
hence claim @xmath56 holds .
0.2 cm * claim 2 . *
@xmath57 is even , @xmath58 , the square @xmath59 , @xmath60 and @xmath61 .
if @xmath62 , then @xmath63 for all @xmath64 , @xmath49 .
0.2 cm claim 1 implies that @xmath57 , @xmath58 , and @xmath63 for all @xmath49 .
it remains to show that @xmath65 is even , @xmath59 , @xmath60 and @xmath61 . since @xmath1 is a maximum resonant set of @xmath0 , @xmath66 for all @xmath49 , @xmath48
. otherwise , @xmath67 is a resonant set of @xmath0 since the square @xmath50 is either @xmath68-alternating or @xmath69-alternating , a contradiction .
so @xmath70 for all @xmath71 .
first , we show that @xmath65 is even .
suppose to the contrary that @xmath65 is odd with @xmath72 ( see fig .
[ fig : polyomino-02 ] ) .
we use @xmath73 to denote the subgraph of @xmath0 formed by squares @xmath74 .
then we can see that the restriction of @xmath69 on @xmath73 is a perfect matching of @xmath73 .
let @xmath75 .
then @xmath76 is a perfect matching of @xmath0 such that each member in the set @xmath77 is an @xmath76-alternating square .
note that the set @xmath78 with cardinality @xmath79 , whereas @xmath80 .
hence , @xmath81 is a resonant set of @xmath0 larger than @xmath1 .
this contradicts that @xmath1 is a maximum resonant set of @xmath0 .
: @xmath65 is odd.,height=141 ] next , we show that @xmath59 .
suppose to the contrary that @xmath82 .
. we use @xmath84 to denote the subgraph of @xmath0 formed by squares @xmath85 .
then the restriction of @xmath68 on @xmath84 is a perfect matching of @xmath84 .
note that @xmath86 are @xmath68-alternating squares and @xmath87 .
hence , we can see that @xmath88 is a resonant set of @xmath0 larger than @xmath1 , a contradiction .
similarly , @xmath89 and @xmath90 . moreover , @xmath61 .
so we complete the proof of claim @xmath29 .
let @xmath91 be an integer with @xmath92 such that @xmath93 is even , and @xmath94 is odd for all @xmath95 .
it follows from @xmath96 , @xmath66 that @xmath97 for all @xmath98 and @xmath64 , @xmath99 , @xmath49 .
we now need to distinguish the following two cases .
: @xmath100 and @xmath101.,height=185 ] : @xmath102 , @xmath103.,height=229 ] * case 1 . * @xmath104 . in this case
we have @xmath105 for all @xmath106 .
_ subcase 1.1 : _
@xmath100 ( see fig .
[ fig : polyomino-03 ] ) .
let @xmath107 denote the subgraph of @xmath0 formed by squares in the set @xmath108 then the restriction of @xmath69 on @xmath107 is a perfect matching of @xmath107 .
let @xmath109 @xmath110 then @xmath111 is a perfect matching of @xmath0 such that each member of @xmath112 is an @xmath111-alternating square .
note that @xmath113 , whereas @xmath114 @xmath115 hence , @xmath112 is a resonant set of @xmath0 larger than @xmath1 , a contradiction . _ subcase 1.2 : _
@xmath116 ( see fig .
[ fig : polyomino-04 ] ) .
let @xmath117 denote the subgraph of @xmath0 formed by squares in @xmath118 then the restriction of @xmath69 on @xmath117 is a perfect matching of @xmath117 .
let @xmath119 @xmath120 then @xmath121 is a perfect matching of @xmath0 such that each member of @xmath122 is an @xmath121-alternating square .
note that @xmath123 , whereas @xmath124 @xmath125 hence , @xmath122 is a resonant set of @xmath0 larger than @xmath1 , a contradiction . : @xmath126 , @xmath127.,height=229 ] * case 2 .
* @xmath127 ( see fig . [
fig : polyomino-05 ] ) .
let @xmath128 denote the subgraph of @xmath0 formed by squares @xmath129 .
note that the left side vertical edge of the square @xmath130 belongs to @xmath37 , and moreover each square in @xmath128 is @xmath69-alternating .
thus we can see that @xmath131 is a resonant set of @xmath0 larger than @xmath1 , a contradiction .
now the entire proof of the lemma is complete .
0.2 cm * proof of theorem [ thm-2 - 1 ] .
* suppose to the contrary that @xmath2 has two perfect matchings @xmath8 and @xmath40 .
then @xmath132 contains an @xmath133-alternating cycle @xmath21 .
let @xmath134 $ ] denote the subgraph of @xmath0 consisting of @xmath21 together with its interior .
put @xmath135 $ ] .
then @xmath136 is not a maximum resonant set of @xmath134 $ ] since @xmath134 $ ] and @xmath136 satisfy the condition of lemma [ lem-2 - 2 ] .
moreover @xmath1 is also not a maximum resonant set of @xmath0 , which contradicts the assumption that @xmath1 is a maximum resonant set of @xmath0 .
for a maximal alternating set of polyomino graphs , we can obtain the following results .
[ lem-3 - 1 ] let @xmath0 be a @xmath29-connected polyomino graph with a perfect matching , @xmath1 an alternating set consisting of internal squares and @xmath20 the boundary of @xmath0 .
if @xmath30 has a perfect matching or is an empty graph , then @xmath1 is not a maximal alternating set .
we adopt the substructure and notations used in the proof of lemma [ lem-2 - 2 ] ( see fig . [
fig : polyomino-01 ] ) .
let @xmath8 be a perfect matching of @xmath42 such that all squares in @xmath1 are @xmath8-alternating , and let @xmath36 and @xmath37 be two perfect matchings of @xmath20
. then @xmath138 and @xmath139 are two perfect matchings of @xmath0 .
suppose to the contrary that @xmath1 is a maximal alternating set of @xmath0 .
the following claim 1 and its proof are the same as claim 1 of lemma [ lem-2 - 2 ] .
0.2 cm * claim 1 . * for a pair of parallel edges @xmath51 and @xmath52 of a square @xmath53 of @xmath0 , they do not lie simultaneously on the boundary of @xmath0 .
0.2 cm * claim 2 . *
@xmath57 , @xmath58 , the square @xmath59 , @xmath60 and @xmath61 .
@xmath66 for all @xmath49 , @xmath48 .
moreover , @xmath140 for all @xmath71 .
0.2 cm claim 1 implies that @xmath57 , @xmath58 , and @xmath63 for all @xmath49 , @xmath48 . since @xmath1 is a maximal alternating set of @xmath0 , @xmath66 for all @xmath49 , @xmath48 .
otherwise , @xmath67 is an alternating set of @xmath0 since the square @xmath50 is either @xmath68-alternating or @xmath69-alternating , a contradiction .
so @xmath140 for all @xmath71 .
now we show that @xmath59 .
suppose to the contrary that @xmath82 .
then @xmath83 and @xmath141 is @xmath68-alternating .
so @xmath142 is an alternating set of @xmath0 , a contradiction .
symmetrically , @xmath89 and @xmath60 .
so @xmath61 .
hence claim @xmath29 is proved
. 0.2 cm let @xmath91 be an integer with @xmath143 such that @xmath127 , and @xmath144 for all @xmath95 .
it follows from @xmath145 , @xmath66 that @xmath146 for all @xmath98 and @xmath64 , @xmath99 , @xmath49 .
note that the left vertical edge of the square @xmath130 belongs to @xmath36 or @xmath37 , say @xmath36 .
so @xmath130 is @xmath68-alternating and @xmath147 is an alternating set of @xmath0 , which contradicts the assumption that @xmath1 is a maximal alternating set of @xmath0 . the lemma is proved .
[ thm-3 - 2 ] let @xmath0 be a polyomino graph with a perfect matching , and @xmath1 be a maximal alternating set of @xmath0 .
then @xmath2 has a unique perfect matching .
suppose to the contrary that @xmath2 has two perfect matchings @xmath8 and @xmath40 .
then @xmath132 contains an @xmath133-alternating cycle @xmath21 .
put @xmath135 $ ] .
then @xmath136 is not a maximal alternating set of @xmath134 $ ] since @xmath134 $ ] and @xmath136 satisfy the condition of lemma [ lem-3 - 1 ] .
so @xmath1 is also not a maximal alternating set of @xmath0 , a contradiction .
motivated by theorem [ thm-1 - 3 ] , it is natural to ask the following question : when does the maximum forcing number of a plane bipartite graph equal its clar number ? in the following , we shall give a sufficient condition .
let @xmath7 be a plane graph with a perfect matching .
a cycle @xmath21 of @xmath7 is said to be _ nice _ if @xmath7 has a perfect matching @xmath8 such that @xmath21 is an @xmath8-alternating cycle .
denote by @xmath134 $ ] the subgraph of @xmath7 consisting of @xmath21 together with its interior . a cycle @xmath21 of @xmath7
is called a _ face cycle _ if it is the boundary of some finite face of @xmath7 . for convenience
, we do not distinguish a face cycle with its finite face .
[ thm-4 - 1 ] let @xmath7 be a connected plane bipartite graph with perfect matchings . if for each nice cycle @xmath21 of @xmath7 and a maximum resonant set @xmath1 of @xmath134 $ ] , @xmath134-k$ ] has a unique perfect matching , then @xmath148 .
let @xmath149 . by the definition of clar number and theorem [ thm-1 - 2 ] , we can see that @xmath150 . in the following we show @xmath151 .
define @xmath152 as the set of perfect matchings of @xmath7 whose forcing numbers equal @xmath153 . by theorem [ thm-1 - 2 ] , for any @xmath154 ,
there exist @xmath153 pairwise disjoint @xmath8-alternating cycles in @xmath7 .
we choose a perfect matching @xmath68 in @xmath152 such that @xmath153 disjoint @xmath68-alternating cycles @xmath155 of @xmath7 have face cycles as many as possible . put @xmath156 .
it suffices to show that all cycles in @xmath157 are face cycles .
otherwise , @xmath157 has a non - face cycle member and its interior contains only face cycle members of @xmath157 . without loss of generality , let @xmath158 denote such a non - face cycle member of @xmath157 and @xmath159 are all the face cycles in @xmath157 contained in the interior of @xmath158 for some @xmath98 , @xmath160 .
then the restriction of @xmath68 on @xmath161 $ ] is also a perfect matching of @xmath161 $ ] , denoted by @xmath162 . by the assumption
, @xmath163 is a non - maximum resonant set of @xmath161 $ ] .
let @xmath13 be a maximum resonant set of @xmath161 $ ] .
then @xmath164 . let @xmath165 be a perfect matching of @xmath161 $ ] such that all faces in @xmath13 are @xmath165-resonant .
let @xmath166 and @xmath167 .
then @xmath69 is a perfect matching of @xmath7 and each member of @xmath168 is an @xmath69-alternating cycle . note that @xmath169 and @xmath168 contain more face cycles than @xmath157 .
this contradicts the choices of @xmath68 and @xmath170 .
combining theorems [ thm-4 - 1 ] with [ thm-2 - 1 ] and [ thm-1 - 1 ] , we immediately obtain the following results .
[ cor-4 - 2 ] _ @xcite _ let @xmath0 be a hexagonal system with a perfect matching
. then @xmath18 .
[ cor-4 - 3 ] let @xmath4 be a polyomino graph with a perfect matching
. then @xmath171 .
we now give a weakly elementary property of such graphs that satisfies the conditions of theorem [ thm-4 - 1 ] .
let @xmath7 be a connected plane bipartite graph with a perfect matching .
an edge of @xmath7 is called _ allowed _ if it lies in some perfect matching of @xmath7 and _ forbidden _ otherwise .
@xmath7 is called _ elementary _ if each edge of @xmath7 is allowed .
@xmath7 is said to be _ weakly elementary _ if for each nice cycle @xmath21 of @xmath7 the interior of @xmath21 has at least one allowed edge of @xmath7 that is incident with a vertex of @xmath21 whenever the interior of @xmath21 contains an edge of @xmath7 @xcite .
a face @xmath9 of @xmath7 is said to be a _ boundary face _ if the boundaries of @xmath9 and @xmath172 have a vertex in common .
let @xmath7 be a connected plane bipartite graph with perfect matchings .
if for each nice cycle @xmath21 of @xmath7 and a maximum resonant set @xmath1 of @xmath134 $ ] , @xmath134-k$ ] has a unique perfect matching , then @xmath7 is weakly elementary .
suppose to the contrary that @xmath7 is not weakly elementary .
then there exists a nice non - face cycle @xmath21 of @xmath7 such that the interior of @xmath21 has no allowed edges of @xmath7 incident with vertices of @xmath21 .
it follows that the interior of @xmath21 has no allowed edges of @xmath134 $ ] are incident with vertices of @xmath21 .
so , for every perfect matching @xmath8 of @xmath134 $ ] , @xmath21 is @xmath8-alternating and any maximum resonant set @xmath1 of @xmath134 $ ] contains no boundary faces of @xmath134 $ ] .
so @xmath134-k$ ] has at least two perfect matchings , a contradiction .
xu et al . ever gave a conjecture as follows , which can be now confirmed .
[ conj-1 ] _
@xcite _ let @xmath7 be an elementary polyomino graph . then the maximum forcing number of @xmath7 can be computed in polynomial time .
abeledo and atkinson obatained the following result .
[ thm-4 - 6 ] _
@xcite _ let @xmath7 be a @xmath29-connected plane bipartite graph .
then the clar number of @xmath7 can be computed in polynomial time using linear programming methods .
a polyomino graph @xmath4 with forbidden edges can be decomposed into some elementary components @xcite .
a polynomial time algorithm to accomplish this decomposition .
such elementary components are elementary polyominoes and thus 2-connected .
note that the forcing number of any perfect matching @xmath8 of @xmath4 equals the sum of forcing numbers of the restrictions of @xmath8 on its elementary components .
so theorem [ thm-4 - 6 ] and corollary [ cor-4 - 3 ] imply the following result , which confirms conjecture [ conj-1 ] .
[ thm-4 - 7 ] let @xmath7 be a polyomino graph with a perfect matching . then the maximum forcing number of @xmath7 can be computed in polynomial time .
theorem [ thm-2 - 1 ] does not hold for general plane bipartite graphs even for plane elementary bipartite graphs .
let us see two elementary bipartite graphs @xmath7 and @xmath173 as shown in fig .
[ fig : weakly - elementary ] , where @xmath173 is a subgraph of @xmath7 bounded by a nice cicle of @xmath7 .
since there exist at most @xmath174 ( resp .
@xmath175 ) pairwise disjoint finite faces in @xmath7 ( resp .
@xmath173 ) and the faces with labels @xmath56 ( resp .
@xmath29 ) form a resonant set of @xmath7 ( resp .
@xmath173 ) , we have that @xmath176 and @xmath177 .
there is a maximum resonant set @xmath178 ( the faces with labels @xmath29 ) of @xmath173 such that @xmath179 has two perfect matchings . for any maximum resonant set @xmath13 of @xmath7
, @xmath180 is empty .
this example shows that theorem [ thm-2 - 1 ] holds for a graph @xmath7 does not imply that it holds for each such subgraph of @xmath7 . since @xmath173 has a perfect matching @xmath8 such that the faces with labels @xmath29 and the infinite face of @xmath173 are @xmath8-resonant , by theorem [ thm-1 - 2 ] we have that @xmath181 .
since there exist no @xmath182 pairwise disjoint cycles in @xmath173 , we have that @xmath183 .
so @xmath184 and @xmath185 . but @xmath186 . | a polyomino graph @xmath0 is a connected finite subgraph of the infinite plane grid such that each finite face is surrounded by a regular square of side length one and each edge belongs to at least one square . in this paper , we show that if @xmath1 is a maximum resonant set of @xmath0 , then @xmath2 has a unique perfect matching .
we further prove that the maximum forcing number of a polyomino graph is equal to its clar number .
based on this result , we have that the maximum forcing number of a polyomino graph can be computed in polynomial time .
we also show that if @xmath1 is a maximal alternating set of @xmath0 , then @xmath2 has a unique perfect matching .
* keywords : * polyomino graph ; resonant set ; clar number ; forcing number ; alternating set . *
ams 2000 subject classification :* 05c70 , 05c90 , 92e10 |
zero - range processes are stochastic particle systems with no restriction on the number of particles per site and with jump rates that depend only on the occupation of the departure site .
this simple zero - range interaction leads to a product structure of the stationary distributions @xcite .
these processes have been a focus of recent research interest since they can exhibit a condensation transition .
this is the case for space homogeneous jump rates @xmath0 that decay asymptotically with the number of particles @xmath1 .
a prototypical model with @xmath2 has been introduced in @xcite , where condensation occurs for parameter values @xmath3 ,
@xmath4 or @xmath5 , @xmath6 .
if the particle density @xmath7 exceeds a critical value @xmath8 , the system phase separates into a homogeneous background with density @xmath8 and a condensate , where the excess particles accumulate on a single randomly located lattice site .
this transition has been established on a rigorous level in a series of papers , in the thermodynamic limit @xcite , as well as on a finite system as the total number of particles diverges @xcite .
dynamic aspects of the transition such as equilibration and coarsening @xcite and the stationary dynamics of the condensate @xcite are well understood heuristically , for the latter first rigorous results have been achieved recently @xcite .
findings for the zero - range process could be applied to understand condensation phenomena in a variety of nonequilibrium systems ( see @xcite and references therein ) , as well as providing a generic model of domain wall dynamics and a criterion for phase separation using a mapping to one - dimensional exclusion systems @xcite .
the process continues to be of interest , recent work on variations of the model includes mechanisms leading to more than one condensate @xcite , or the effects of memory in the dynamics @xcite . while most of the results so far consider the thermodynamic limit , finite size effects in the model with jump rates @xmath0 and @xmath5 have been investigated in @xcite using saddle point methods , and in @xcite for a variant of the model with a single defect site . in the condensed phase region , finite systems are found to exhibit a large overshoot of the stationary current above its value in the thermodynamic limit . in this paper
we examine this phenomenon in detail for all possible values of the parameter @xmath9 $ ] .
we find the leading order finite size effects that describe the current overshoot by continuing the homogeneous ( fluid ) phase above the critical density and characterizing the condensed phase by a current matching argument .
for @xmath10 , the main focus of this article , finite systems exhibit a metastable switching behaviour between the two phases , which is prevalent in monte carlo simulations for a wide range of parameters . to capture this phenomenon we examine the system in a scaling limit and derive a rate function , which exhibits a double well structure , describing the distribution over the bulk density ( the bulk density serves as an order parameter to distinguish the fluid and condensed phases ) .
this way we rigorously establish the discontinuous behaviour on the critical scale , even though the bulk density is a continuous function in the thermodynamic limit .
this is shown to be in agreement with recent results on the condensation transition at the critical density @xcite .
based on the exact scaling limit , we can also predict the lifetime of the metastable phases for large finite systems by a heuristic random walk argument . in general , finite system size can lead to effective long - range interactions and non - convexity of thermodynamic potentials , as has been observed for various models ( see for example @xcite and references therein ) .
for the zero - range process it has been shown that for simple size - dependent jump rates metastability effects can be manifested even in the thermodynamic limit @xcite , and large crossover effects in these systems have already been observed in @xcite .
the finite - size behaviour of the zero - range process considered here also exhibits the above non - convexity , with the additional feature of a sharp crossover between a putative fluid and condensed phase in the non - convex part .
this leads to a metastable switching behaviour which disappears in the thermodynamic limit , intriguingly contradicting the usual expectation of finite systems to behave in a smoother fashion than the limiting prediction .
the onset of phase coexistence at criticality is a classical question of general interest in phase separating systems , see for example @xcite for the formation of equilibrium droplets in the ising model which also form suddenly on a critical scale .
rigorous results on metastability regarding the dynamics of the condensate in zero - range processes have also been a subject of recent research interest @xcite , and our work provides a contribution in that direction and new insight in the mechanisms of condensate dynamics on finite systems .
this is explained in more detail in the discussion .
a proper understanding of the metastability phenomenon exhibited by zero - range processes on finite systems is also of particular importance for recent applications with moderate system sizes .
clustering phenomena in granular media can be described by zero - range processes ( see @xcite and references therein ) , and metastable switching between homogeneous and condensed states has been observed experimentally @xcite . in the spirit of the mapping introduced in @xcite the zero - range process with jump rates
@xmath0 has also been applied as a simplified traffic model @xcite , where condensation corresponds to the occurance of a traffic jam .
a key feature of traffic models is the existence of a broad range of densities over which metastability between free flowing and jammed states is observed ( see for example @xcite and references therein ) .
the relevance of this study in application serves as a motivation , but the aim of the paper is a general understanding of finite size effects and their implications for a generic class of zero - range processes , rather than a detailed analysis of particular cases . the paper is organized as follows : in section 2 we introduce the model and summarise previous results on stationary distributions and the thermodynamic limit . in section 3
we present relevant observations on finite systems which we later study in detail . in section 4
we give preliminary results and a heuristic description of the finite size effects .
this motivates our main results presented in section 5 , where we rigorously establish metastability in a suitable scaling limit . in section 6
we connect these results to the lifetimes of the metastable phases , and end with a short discussion in section 7 .
we consider a one dimensional lattice of l sites @xmath11 with periodic boundary conditions .
let @xmath12 be the number of particles on site @xmath13 .
the state of the system is described by @xmath14 belonging to the state space of all particle configurations @xmath15 .
particles jump on the lattice at a rate that depends only on the occupation number of the departure site .
a particle jumps off site @xmath16 after an exponential waiting time with rate @xmath17 and moves to a target site @xmath18 according to the probability distribution @xmath19 .
we assume that @xmath20 is of finite range , i.e. @xmath21 if @xmath22 for some @xmath23 , normalised and irreducible on @xmath24 .
the main results of the paper focus on the jump rates @xmath25 of the form @xmath26 with @xmath27 and @xmath4
. a discussion on extending these results to the case @xmath28 and @xmath29 and other lattice geometries can be found in section [ sec : discuss ] .
these jump rates were first introduced by evans @xcite and represent a fairly general class of functions of interest for the condensation transition .
the infinitesimal generator of the process acting on suitable test functions @xmath30 is given by @xmath31 where @xmath32 and @xmath33 is the kronecker delta @xcite .
the process conserves the total number of particles in the system , so @xmath34 can be partitioned into invariant subsets @xmath35 on which the zero - range process is a finite state irreducible markov process .
the process can also be defined on an infinite lattice under certain constraints , for details see @xcite .
the following summarises well known results on stationary measures of the zero - range process , for details see @xcite .
the zero - range process with generator ( [ eq : infgen ] ) has a family of stationary homogeneous product measures on @xmath36 which we refer to as the * grand canonical ensemble*. these measures are parameterized by a fugacity @xmath37 and are of the form , @xmath38=\prod_{x\in\lambda_l}\nu_{\phi}\left[\eta_x\right ] \quad \textrm{where } \quad \nu_\phi\left[n\right ] = \frac{1}{z(\phi)}w(n)\phi^{n } .\end{aligned}\ ] ] these exist for all @xmath39 where @xmath40 is the radius of convergence of the ( single site ) partition function @xmath41 the stationary weights @xmath42 are given by @xmath43 and @xmath44 in the grand canonical ensemble the expected particle density is a function of @xmath37 and is given by , @xmath45 } } = \sum_{k=0}^{\infty}k\nu_\phi(k ) = \phi\,\partial_\phi\log z(\phi)\ , \end{aligned}\ ] ] which is strictly increasing and @xmath46 .
the critical density is defined by @xmath47 $ ] , and condensation occurs if @xmath48 , as explained below .
the expected jump rate off a site is proportional to the average stationary current or , in case the first moment @xmath49 vanishes , to the diffusivity .
therefore for simplicity , in the rest of this paper , * current * will refer to the average jump rate off a site , which is clearly site independent under a stationary distribution .
in the grand canonical ensemble the current is simply given by the fugacity @xmath37 , @xmath50 } } & = \frac{1}{z(\phi)}\sum_{n=0}^{\infty}g(n ) w(n ) \phi^{n}=\phi\ , \end{aligned}\ ] ] which follows directly from the form of the stationary weights @xmath51 eq . ( [ eq : wn ] ) .
for fixed @xmath52 and @xmath53 the process restricted to @xmath54 is ergodic , the corresponding unique stationary measures belong to the * canonical ensemble * and are given by @xmath55:=\nu_\phi\left[{{\boldsymbol{\eta}}}| s_l({{\boldsymbol{\eta } } } ) = n\right ] \quad \textrm{where } \quad s_l({{\boldsymbol{\eta } } } ) = \sum_{x\in \lambda_l } \eta_x.\end{aligned}\ ] ] these measures are independent of @xmath37 and are given explicitly by @xmath55=\frac{1}{z(l , n)}\prod_{x\in\lambda_l}w(\eta_x)\delta\big(\sum_x \eta_x , n\big)\ , \end{aligned}\ ] ] where the canonical partition function is @xmath56 in the canonical ensemble the form of the stationary weights , eq .
( [ eq : wn ] ) , imply that the average current is given by a ratio of partition functions , @xmath57 } } = \frac{z(l , n-1)}{z(l , n)}.\end{aligned}\ ] ] for the jump rates ( [ eq : rates ] ) the single site weights asymptotically decay as a stretched exponential @xmath58 throughout the paper we use ` @xmath59 ' to mean asymptotically proportional and ` @xmath60 ' to mean asymptotically equal .
it follows that @xmath61 and the critical density and the variance are finite , @xmath62 } } - \rho_c^2 < \infty\ .\end{aligned}\ ] ] @xmath63 is strictly increasing , so invertible on @xmath64=[0,1]$ ] and we denote its inverse by @xmath65 in this way we can parameterise the grand canonical measures by densities @xmath66 $ ] . in this section
we summarise known results on the condensation transition in the zero - range process .
it has been established in @xcite and rigorously in @xcite as a continuous phase transition in the thermodynamic limit , as particle number @xmath53 and lattice size @xmath52 tend to infinity such that @xmath67 .
it was shown that all finite dimensional marginals of the canonical measure @xmath68 converge to the grand canonical measure with density @xmath7 if @xmath69 . if @xmath70 then there is no grand canonical measure with density @xmath7 and all finite dimensional marginals of @xmath71 converge to the grand canonical measure with density @xmath8 .
this result on the equivalence of ensembles holds in terms of weak convergence ( for details see @xcite ) and can be summarised as @xmath72 where @xmath73 this implies that for @xmath70 the excess particles in the system condense on a set of vanishing volume fraction .
the result has been strengthened in @xcite ( with partial results already in @xcite ) , showing that the condensate typically resides on a single lattice site . denoting the size of the maximum component @xmath74
this can be written as @xmath75 here the notation denotes converges in probability with respect to @xmath71 , @xmath76 \to 0 \quad \textrm{for all } \quad \epsilon>0.\end{aligned}\ ] ] it will often be useful to express this result also in terms of the behaviour of the bulk of the system .
we denote the total number of particles outside of the maximally occupied site by @xmath77 .
the background density converges as @xmath78 a similar result holds for the behaviour of the canonical current in the thermodynamic limit . with eq .
( [ eq : gccur ] ) the convergence in ( [ eq : weak ] ) , ( [ eq : phi ] ) implies that @xmath79 so the current and the background density are both continuous with respect to the total system density in the thermodynamic limit . also both are strictly increasing up to @xmath8 and constant for @xmath70 .
if @xmath80 the system is said to be in the * fluid * phase region and if @xmath81 the system is in the * condensed * phase region .
the thermodynamic entropy is defined by the legendre transform @xmath82 where @xmath83 is given by eq .
( [ eq : phi ] ) .
it has been shown in @xcite that the canonical partition function converges to the thermodynamic entropy , i.e. @xmath84
finite size effects and current overshoot for @xmath85 and @xmath86 . left : the canonical current for various system sizes as a function of the density @xmath87 are plotted .
the dashed black line shows the thermodynamic current as a function of the system density @xmath7 .
right : the overshoot region for @xmath88 showing the two distinct currents measured from monte carlo simulations for various densities near the maximum current.,title="fig:",scaledwidth=48.0% ] finite size effects and current overshoot for @xmath85 and @xmath86 . left : the canonical current for various system sizes as a function of the density @xmath87 are plotted .
the dashed black line shows the thermodynamic current as a function of the system density @xmath7 .
right : the overshoot region for @xmath88 showing the two distinct currents measured from monte carlo simulations for various densities near the maximum current.,title="fig:",scaledwidth=49.0% ] in this section we present results obtained from exact numerics and monte carlo simulations in the canonical ensemble .
we can calculate the canonical current given by eq .
( [ eq : cancur ] ) by making use of the following recursion relation for the canonical partition functions , @xmath89 similarly we can calculate the canonical distribution of the maximum site occupation @xmath90 . to this end
we define the cut - off canonical partition function which counts configurations for which @xmath91 , @xmath92 this allows us to calculate @xmath93=\frac{q(l , n , m)-q(l , n , m-1)}{z(l , n)}\ , \ ] ] for all @xmath94 ( the case @xmath95 is trivial ) .
we often consider the equivalent formulation using background densities @xmath96 = \pi_{l , n}\left [ \frac{{s_l^{\textrm{bg}}}}{l-1 } = { \rho_{\textrm{bg}}}\right ] \quad \textrm { where } \quad { \rho_{\textrm{bg}}}= \frac{n - m}{l-1}\ .
\nonumber\end{aligned}\ ] ] this is more useful for illustrations and is more intuitive , since the background density characterises all but a single lattice site , while the formulation involving @xmath90 is more convenient for computations since it avoids issues with non - integer numbers . therefore keeping both formulations in parallel is the best option for a concise presentation of our results .
finite size effects for the background density for @xmath85 , @xmath86 and @xmath88 . left : the expected value of the background density .
the dashed black line shows the thermodynamic limit result .
right : distribution of the background density at three system densities shown by corresponding dash - dotted lines on the left , calculated exactly using eq .
( [ eq : canmax ] ) .
the position of local maxima of the distributions are marked on both plots ( for @xmath97 , @xmath98 for @xmath99 , for @xmath100 ) .
the high background density maximum at @xmath100 occurs with extremely low probability and is off the scale.,title="fig:",scaledwidth=48.0% ] finite size effects for the background density for @xmath85 , @xmath86 and @xmath88 .
left : the expected value of the background density .
the dashed black line shows the thermodynamic limit result .
right : distribution of the background density at three system densities shown by corresponding dash - dotted lines on the left , calculated exactly using eq .
( [ eq : canmax ] ) .
the position of local maxima of the distributions are marked on both plots ( for @xmath97 , @xmath98 for @xmath99 , for @xmath100 ) .
the high background density maximum at @xmath100 occurs with extremely low probability and is off the scale.,title="fig:",scaledwidth=49.0% ] on large finite systems we observe significant finite size effects above the critical density @xmath8 , fig .
[ fig : cur ] shows the typical behaviour of the canonical current .
below @xmath8 the current is very close to the thermodynamic limit result even on relatively small systems ( @xmath101 ) and the leading order finite size effects can be understood immediately from the proof of the thermodynamic limit result @xcite .
however the canonical current and background density significantly overshoot their critical values see fig .
[ fig : cur ] and fig .
[ fig : background ] .
an overshoot has been observed before for systems containing a single site defect @xcite for @xmath102 .
for @xmath103 the current increases monotonically with density @xmath104 , in a way that appears to vary only very slightly with system size , up to some size dependent maximum current at @xmath105 .
for @xmath106 the background density in the system is typically very close to @xmath7 ( fig .
[ fig : background ] ) . for a fixed system size
@xmath52 we associate the region @xmath107 with a * putative fluid phase*. close to @xmath108 there is an abrupt decrease in the current and background density . for @xmath109
the current and background density are both decreasing in @xmath7 and tend to their respective critical value as @xmath110 .
we associate the region @xmath109 with a * putative condensed phase * since the increasing density is entirely taken up by a large number of particles condensing on a single lattice site .
the size of the effective fluid overshoot increases with increasing @xmath111 and with decreasing @xmath112 , and is already very pronounced at @xmath113 , @xmath86 . switching dynamics for @xmath114 , @xmath86 and @xmath88 taken at @xmath115 . left : ( bottom ) current against time from monte carlo simulations , calculated by taking the average number of jumps in the system over small time windows .
the black line indicates the transition between the two putative phases .
( top ) the location of the maximum does not change until the system is fluid .
right : cumulative tail of the distribution of waiting times in the putative fluid phase on a log linear scale .
the solid line shows an exponential fit.,title="fig:",scaledwidth=48.0% ] switching dynamics for @xmath114 , @xmath86 and @xmath88 taken at @xmath115 .
left : ( bottom ) current against time from monte carlo simulations , calculated by taking the average number of jumps in the system over small time windows .
the black line indicates the transition between the two putative phases .
( top ) the location of the maximum does not change until the system is fluid .
right : cumulative tail of the distribution of waiting times in the putative fluid phase on a log linear scale .
the solid line shows an exponential fit.,title="fig:",scaledwidth=48.0% ] for @xmath87 close to @xmath105 we observe that the canonical distribution over background densities has two maxima of similar magnitude ( fig .
[ fig : background ] ) and monte carlo simulations show that the system switches between the two putative phases . fig .
[ fig : switch ] shows the typical behaviour of the current close to @xmath105 as a function of time , the putative phases can be clearly distinguished by the current . for @xmath116 in the condensed phase the location of the condensate does not change , while its position fluctuates heavily in the fluid phase , supporting the fact that the particles are distributed homogeneously .
the empirical distribution of waiting times in the two putative phases is very close to an exponential ( fig .
[ fig : switch ] right ) .
this suggests that the switching process is approximately markovian over the range of parameters and jump distributions observed , and constitutes a genuine metastability phenomenon .
the rate of the switching depends on the parameters @xmath111 and @xmath112 as well as the jump distribution @xmath117 , which will be discussed in section [ sec : dynamics ] in more detail .
as suggested by the form of the current overshoot in fig 1 , our approach is to approximate the putative fluid phase by extending the grand canonical current above @xmath40 , which we achieve by means of a cut - off grand canonical measure . since the total number of particles
is fixed canonically to @xmath53 the system can not explore states where any single site contains more than @xmath53 particles .
we therefore expect the distribution of particles in the background under the canonical measure will always be closer ( in any reasonable sense ) to a grand canonical measure , with some suitably chosen cut - off and fugacitiy chosen to fix the correct density of particles , than it is to the unconditioned distribution . for cut - off @xmath94
the * cut - off grand canonical measures * are defined by single site marginals with support on @xmath118 , @xmath119&:=\nu_{\phi}\left[\eta_x= n|\eta_x\leq m \right ] \nonumber\\ & = \frac{1}{z_{m}(\phi)}w(n)\phi^{n } \quad \textrm{for } \quad n\leq m\end{aligned}\ ] ] where the normalisation is given by the finite sum @xmath120 these measures are well defined for all @xmath121 and the current is given by @xmath122 } } & = \phi\left ( 1-\frac{w(m)\phi^{m}}{z_{m}(\phi)}\right).\end{aligned}\ ] ] also the average density in the cut - off ensemble @xmath123}}=\frac{1}{z_{m}(\phi)}\sum_{k=0}^{m}k w(k ) \end{aligned}\ ] ] is a strictly increasing function from @xmath124 onto @xmath124 and so we denote its inverse @xmath125 to a first approximation we estimate the current in the putative fluid phase using eq .
( [ eq : cutcur ] ) , @xmath126 under the assumption that @xmath127 for values of @xmath53 under consideration .
this approximation is shown in fig .
[ fig : approxs ] by the blue dashed line for @xmath128 , @xmath86 and a system size of @xmath88 .
the approximation is extremely close to the canonical current for @xmath107 and for @xmath116 is in good agreement with the empirically measured fluid currents from monte carlo simulations .
we demonstrate in the section [ sec : rigour ] that this estimate can be improved by choosing the cut - off more carefully .
current estimates above the critical point @xmath70 and @xmath129 , for @xmath88 , @xmath85 and @xmath86 .
the fluid approximation ( [ eq : fluidcur ] ) and the current matching ( [ eq : curmatch ] ) agree well with canonical numerics and the metastable branches from monte carlo simulations ( cf . figs .
[ fig : cur ] and [ fig : switch]).,scaledwidth=70.0% ] we approximate the current in the putative condensed phase by a current matching argument , a similar heuristic argument has been given before in @xcite .
the existence of a stable condensate implies that the average rate of particles exiting the condensate must be equal to the average rate of particles entering it .
conditioned on the occupation of the condensate @xmath130 , the exit rate is simply @xmath131 while the entry rate is well approximated by the stationary current in the background .
this is assumed to be in a putative fluid phase described by @xmath132 , which leads to the current matching condition .
@xmath133 the lower branch of solutions to this equation define the condensed current approximation ( fig .
[ fig : approxs ] red dashed line ) .
these are extremely close to the canonical current for @xmath134 and for @xmath116 are in good agreement with the empirically measured condensed currents from monte carlo simulations . in the thermodynamic limit
we expect the existence of a rate function @xmath135 that describes the asymptotic probability of observing a background density @xmath136 .
precisely , we will consider the limit @xmath137 and expect that @xmath138\sim e^{-l i_{\rho } ( { \rho_{\textrm{bg } } } ) } \ .\ ] ] to derive this we define the following function for finite systems @xmath139\ , \end{aligned}\ ] ] which can be written as@xmath140 { + } o(e^{-\frac{-b}{1-\gamma}m^{1-\gamma}})\ . \ ] ] this is derived in appendix [ app : deriv ] and is valid for any @xmath141 .
the first line in eq .
( [ eq : il ] ) resembles the thermodynamic entropy of the background ( cf .
( [ eq : thermoent ] ) ) , the first term on the second line is the contribution due to the maximum occupied site . the second term on the second line is the canonical normalisation and is independent of @xmath142 .
all the other terms will vanish in the limit ( [ eq : mlimit ] ) and we get the following result .
[ thm : first ] in the limit ( [ eq : mlimit ] ) @xmath143 converges for all @xmath144 as @xmath145 where @xmath146 is the thermodynamic entropy given by eq .
( [ eq : thermoent ] ) .
so for @xmath147 ( supercritical case ) , @xmath148 for each @xmath149 ( see fig . [
fig : thermo ] ) .
thermodynamic limit rate function for a fixed system density @xmath150 with @xmath85 and @xmath86 .
@xmath151 is shown by a solid black line .
the dashed lines show @xmath152 against @xmath153 with @xmath154 , for several values of @xmath52 calculated by exact numerics using eq .
( [ eq : canmax ] ) .
note convergence to @xmath151 ( solid black ) is slow and initially non - monotonic above @xmath8.,scaledwidth=70.0% ] the proof follows directly from previous results on the equivalence of ensembles , for details see @xcite .
it has been shown that @xmath155 we use eq .
( [ eq : il ] ) and choose @xmath156 as defined by eq .
( [ eq : phi ] ) .
the cutoff @xmath157 diverges linearly in @xmath52 for all @xmath158 .
it follows that for @xmath159 we have @xmath160 , and therefore with eq .
( [ eq : thermoent ] ) @xmath161 it follows from the asymptotic behaviour of the single site weights eq .
( [ eq : asymwk ] ) that @xmath162 , and the terms @xmath163 and @xmath164 also vanish in the limit eq .
( [ eq : mlimit ] ) . it remains to show that @xmath165\to 0 $ ] .
in the case @xmath166 we have @xmath167 and the first and second moments of @xmath168 converge to those of @xmath169 .
the local limit theorem for triangular arrays ( thm . 1.2 in @xcite ) covers the sum , @xmath170 , of independent random variables whose distribution depends on the number of terms @xmath52 , which is the case here via the cut - off @xmath142 .
convergence of the first two moments then implies @xmath171\sim \frac{1}{\sqrt{l}}\ , \end{aligned}\ ] ] and the result follows immediately .
the case @xmath172 can be reduced to the previous case following @xcite , by arranging the excess mass @xmath173 in the background among a finite number of sites so that each site contains at most @xmath142 particles .
this provides a sub - exponential lower bound for @xmath174 $ ] which completes the proof .
previous results in the thermodynamic limit imply that for @xmath175 the background density converges to @xmath8 ( see section [ sec : thermo ] eq .
( [ eq : thermobg ] ) ) .
naturally , the thermodynamic limit result gives no indication of the sharp transition from fluid to condensed putative phases or the metastable switching observed on large finite systems .
for @xmath70 the system will appear to be condensed for sufficiently large @xmath52 , since @xmath176 as @xmath177 ( see fig . [
fig : cur ] ) .
therefore the leading order finite size effects of thm [ thm : first ] do not describe the sharp transition or the metastability , as can be seen in fig .
[ fig : thermo ] .
the apparent double well structure of @xmath178 is less pronounced for higher @xmath52 , and the location of the global minimum shifts towards @xmath8 . for @xmath179 ,
i.e. @xmath180 , the limit of @xmath178 actually depends on the precise scaling of m , which we do not discuss here ( see [ 4 ] for more details ) . to capture the sharp transition and the metastability we will replace the limit eq .
( [ eq : mlimit ] ) by an appropriate scaling limit in section [ sec : rigour ] . current matching on a finite system with @xmath88 and @xmath99 close to @xmath181 for @xmath85 , @xmath86 . left : the current in the background @xmath182 and current out of the maximum @xmath131 plotted against @xmath153 .
right : exact numerics of @xmath178 using ( [ eq : canmax ] ) shown in solid black and approximation ( [ eq : finitediff ] ) in dashed green .
vertical lines show the correspondence between current matching and critical points of @xmath178 .
the first local minimum corresponds to the lower stable branch of the current matching curve in fig .
[ fig : approxs ] and the local maximum corresponds to the upper unstable branch .
, title="fig:",scaledwidth=48.0% ] current matching on a finite system with @xmath88 and @xmath99 close to @xmath181 for @xmath85 , @xmath86 .
left : the current in the background @xmath182 and current out of the maximum @xmath131 plotted against @xmath153 .
right : exact numerics of @xmath178 using ( [ eq : canmax ] ) shown in solid black and approximation ( [ eq : finitediff ] ) in dashed green .
vertical lines show the correspondence between current matching and critical points of @xmath178 .
the first local minimum corresponds to the lower stable branch of the current matching curve in fig .
[ fig : approxs ] and the local maximum corresponds to the upper unstable branch .
, title="fig:",scaledwidth=49.0% ] we can understand the metastability on a heuristic level in terms of a simple current matching argument .
@xmath152 can be calculated efficiently for system sizes @xmath183 using exact numerics ( see eq .
( [ eq : canmax ] ) ) . in fig .
[ fig : thermo ] we see a clearly defined double well structure close to @xmath105 which accounts for the observed switching behaviour .
we may choose @xmath184 instead of @xmath156 in eq .
( [ eq : il ] ) so that the mean under the cut - off grand - canonical distribution is @xmath185 .
the gradient of @xmath152 is given by a one step finite difference .
if we assume for large @xmath52 that @xmath186\sim\frac{1}{\sqrt{2\pi l \sigma_c^2}}$ ] ( which is valid in the scaling limit under certain conditions according to a local limit for triangular arrays ( thm .
1.2 in @xcite ) , see theorem [ thm : main ] for details ) and replace @xmath187 by @xmath142 , then a straightforward calculation shows @xmath188 where @xmath153 .
the first term on the right hand side follows in direct analogy with the thermodynamic limit for which @xmath189 ( cf .
@xcite ) , and the second term is the finite difference of the maximum site contribution using eq .
( [ eq : wn ] ) .
this result holds rigorously in scaling limit and even for relatively small systems ( @xmath190 ) @xmath178 is well approximated by eq .
( [ eq : finitediff ] ) as is shown in fig .
[ fig : curmatch ] .
the approximation breaks down for background densities close to @xmath7 since the cut off @xmath142 becomes small and so the approximation @xmath191 is no longer valid . by the above argument we expect that solutions to the current matching equation ( [ eq : curmatch ] ) correspond to local maxima or minima of the rate function . whilst the current out of the maximum occupied site is greater than the current in the background we expect the background density to increase and vice versa .
therefore the first point that the two currents cross is locally stable and so a local minimum of the rate function . the next point they cross is a local maximum by the same argument , and there is another local minimum at a point close to @xmath7 associated with fluid configurations , which the current matching argument predicts to be at the boundary .
this is shown in fig .
[ fig : curmatch ] .
in this section we explore the finite size effects by examining the leading order behaviour at the critical scale on which metastability persists . as @xmath177
this scale is given by @xmath192 for suitable @xmath193 .
we may equivalently express it as @xmath194 the correct scaling exponent @xmath195 to capture the canonical overshoot and metastability can be determined heuristically by our previous current matching argument eq .
( [ eq : curmatch ] ) . if we assume that close to @xmath8 the fluid current eq .
( [ eq : fluidcur ] ) can be approximated by the first term in the taylor expansion around @xmath8 , then @xmath196 .
the current matching equation ( [ eq : curmatch ] ) then implies @xmath197 which leads to @xmath198 . in order to make this argument rigorous , also to find the transition point and describe the metastability , we study the canonical distribution of the background density ( or equivalently the maximally occupied site ) in the scaling limit . we examine the asymptotic behaviour of @xmath199 introduced in ( [ eq : idef ] ) .
we will see in theorem [ thm : main ] that there exists a unique @xmath200 such that @xmath201 is finite and non - zero at the critical scale , we define @xmath202 where @xmath53 and @xmath142 are given by eq .
( [ eq : critscale ] ) . since @xmath148 for @xmath203 ( see sec . 4.2 )
this definition implies @xmath204\sim e^{-l^\beta i^{(2)}_{\delta\rho } \left(\delta{\rho_{\textrm{bg}}}\right)}\end{aligned}\ ] ] as @xmath177 , we recall that if @xmath205 the background density is @xmath206 and @xmath207 and @xmath208 are given by eq .
( [ eq : critscale ] ) .
we begin by describing the scaling limit for the current in the fluid bulk and out of the maximally occupied site .
[ thm : curscale ] for the scaling limit ( [ eq : critscale ] ) , with @xmath198 , the current out of the most occupied site and the average current in the fluid background are asymptotically given by @xmath209 provided @xmath210 .
the result for the current out of the condensate @xmath131 is immediate since @xmath211 , @xmath212 and @xmath213 .
the proof of the fluid result follows from a taylor expansion of the truncated density function @xmath214 introduced in ( [ eq : cutdens ] ) , details on the expansion are contained in lemmas [ taylor ] and [ convmoms ] in the appendix . for convenience
we write @xmath215 , then @xmath216 for all @xmath217 .
the result follows directly since @xmath218 and choosing @xmath219 so that @xmath220 . although the proofs in this section are restricted to @xmath221 this includes the point of intersection of the two currents ( cf .
[ fig : curmatch ] ) and we discuss a possible extension to the whole region @xmath222 at the end of this subsection .
theorem [ thm : curscale ] states that in the scaling limit the average current in the fluid background and the jump rate out of the condensate converge to the critical current @xmath61 on the same scale as the density @xmath223 converges to @xmath8 .
so for @xmath53 and @xmath142 at the critical scale , given by eq .
( [ eq : critscale ] ) , @xmath224 and @xmath225 converge to unique functions of the scaled variables @xmath207 and @xmath208 .
we now show that in the scaling limit the rate function is exactly the integral of the difference in the two limiting currents ( up to normalisation ) .
[ thm : main ] in the scaling limit ( [ eq : critscale ] ) , with @xmath198 , the rescaled rate function ( [ eq : scalingi ] ) with @xmath226 converges as @xmath227 provided @xmath210 .
firstly consider the unnormalised measure @xmath228 for which it follows from definitions in section [ subsec : statmeasure ] and eq .
( [ eq : idef ] ) @xmath229 = -l^{1-\beta}i_{l , n}(m ) + \frac{1}{l^{\beta}}\log z(l , n ) - l^{1-\beta}\log z(1)\ . \end{aligned}\ ] ] re - writing @xmath178 using eq .
( [ eq : il ] ) we find @xmath230 = & -\frac{(l-1)}{l^{\beta}}\big(\log z_{m}(\phi_m({\rho_{\textrm{bg}}}^l))- { \rho_{\textrm{bg}}}^l\log \phi_m({\rho_{\textrm{bg}}}^l ) \big)+\nonumber \\ & + l^{1-\beta}\log z(1)-\frac{1}{l^{\beta}}\log
w(m ) - \frac{1}{l^{\beta}}\log l+ \nonumber\\ & - \frac{1}{l^{\beta}}\log \nu_{\phi_m({\rho_{\textrm{bg}}}^l),m}^{l-1}\left[s_{l-1}{=}n - m \right ] { - } o(1)\ , \end{aligned}\ ] ] where we used the shorthand @xmath231 . by applying lemma [ convmoms ] , given in the appendix
, the variance under the cut - off distribution @xmath232 converges to the variance of the critical measure for @xmath233 . by the local limit theorem for triangular arrays ( theorem @xmath234 in @xcite ) we have @xmath235 \sim \frac{1}{\sqrt{2\pi
l \sigma_c^2 } } \end{aligned}\ ] ] since @xmath236 and @xmath237 } } \to \rho_c$ ] as @xmath177 with the scaling ( [ eq : critscale ] ) . so the contribution of this term to ( [ eq : scalingi ] ) vanishes . also since @xmath238 we have @xmath239 as @xmath177 .
the condensate contribution to eq .
( [ eq : scalingi ] ) is given by @xmath240 where we have used the asymptotic behaviour of the single site weights ( eq . ( [ eq : asymwk ] ) ) and @xmath241 . by applying theorem [ thm : curscale ] to @xmath242 and taking taylor expansion of @xmath243 according to lemma [ convmoms ] ( appendix ) , @xmath244 and @xmath245 since @xmath246 .
combining equations ( [ eq : asymnu ] ) to ( [ eq : asymlogz ] ) in ( [ eq : scalingi ] ) implies , @xmath247 = \frac{\delta{\rho_{\textrm{bg}}}^2}{2\sigma_c^2}+ \frac{b}{1-\gamma}(\delta\rho-\delta{\rho_{\textrm{bg}}})^{1-\gamma}+o(1 ) \end{aligned}\ ] ] in the scaling limit ( [ eq : critscale ] ) as @xmath177 , provided @xmath210 . by definition @xmath248 .
\nonumber \end{aligned}\ ] ] so we may bound @xmath249 as follows , @xmath250\ } \leq z(l , n ) \leq n z^l(1)\max_{m}\{\mu_{l , n}\left [ m_l = m\right]\}. \nonumber \end{aligned}\ ] ] applying eq .
( [ eq : asymmu ] ) the asymptotic behaviour of @xmath251 in the scaling limit is given by @xmath252 provided that the global minimum of @xmath253 $ ] is attained for @xmath210 in the scaling limit .
this follows from results in @xcite . combining equations ( [ eq : asymmu ] ) and ( [ eq : zln ] ) the result follows from definitions eq .
( [ eq : il ] ) and eq .
( [ eq : mu ] ) @xmath254 the result may break down for @xmath255 due to the large probability of a second sub - condensate forming on the same scale as the condensate .
the conditions under which this occurs could be found by conditioning on a maximum site occupation and using the same methods as for theorem [ thm : main ] to find the distribution over the second highest occupied site . by using this method and iterating
when necessary , the results here could be extended to the entire region @xmath256 .
however our results already cover the relevant critical points of the rate function , the fluid minimum , condensed minimum and local maximum that constitutes a potential barrier .
theorem [ thm : main ] together with theorem [ thm : curscale ] imply that in the scaling limit given by ( [ eq : critscale ] ) the derivative of the scaling rate function @xmath257 is given by the difference in the current out of the condensate and the average current in the background .
this is made precise in the following result , which is a rigorous version of the current matching argument of section [ sec : heur ] .
[ cor : curmatch ] @xmath258 and the two terms on the right - hand side are exactly the limiting current curves from theorem [ thm : curscale ] above the critical point .
this implies that @xmath259 so that @xmath260 always exhibits a boundary minimum at @xmath261 .
depending on the value of @xmath207 the rescaled rate function has one of three qualitative forms characterised by the number of its extreme points , or equivalently the number of roots of ( [ eq : curmatch ] ) , and the position of the global minimum . with the threshold @xmath262 eq .
( [ eq : curmatch ] ) has no real roots for @xmath263 , exactly one for @xmath264 , and two for @xmath265 .
the latter correspond to a local minimum at @xmath266 and a local maximum at @xmath267 , with @xmath268 ( cf .
[ fig : limit ] on the right ) . as usual , minima of @xmath260 correspond to metastable phases , and the depth of the local minimum at @xmath269 as compared to the one at the boundary determines which of the phases is stable ( i.e. corresponds to the global minimum ) .
so @xmath270 marks the threshold above which ( for @xmath265 ) there exists a metastable condensed phase .
stability of the phases changes when both minima have the same depth , the density where this is the case is given by @xmath271 introduced in section 3 . in the scaling limit ( [ eq : critscale ] )
this behaves as @xmath272 and from theorem [ thm : main ] and corollary [ cor : curmatch ] we get the explicit expression @xmath273 note that both @xmath270 and @xmath274 are increasing with @xmath112 and @xmath111 .
their ratio simplifies to @xmath275 which also increases monotonically with @xmath112 .
in particular , @xmath276 for all @xmath277 which implies existence of an extended metastability region .
we can now summarise the critical behaviour of the system at scale ( [ eq : critscale ] ) in three cases , which are also illustrated in fig .
[ fig : limit ] ( right ) .
there exists a unique minimum of the scaling rate function at @xmath261 , @xmath278 corresponding to a stable fluid phase .
the most likely background density under the canonical measure is @xmath279 .
there are two local minima of the scaling rate function @xmath280 corresponding to a stable fluid phase with @xmath261 and a metastable condensed phase with @xmath281 .
the most likely background density under the canonical measure is still @xmath279 .
there are two local minima of the scaling rate function @xmath282 corresponding to a metastable fluid phase with @xmath261 and a stable condensed phase with @xmath281 .
the most likely background density under the canonical measure is now @xmath283 and a finite fraction of the excess mass condenses on a single lattice site with @xmath284 . the value of @xmath285 , and for @xmath286 the position of the global minimum , agree with recent results on the distribution of the maximum at the critical scale @xcite .
the expected value of the maximum site occupation ( equivalently background density ) under the canonical measure can be interpreted in terms of the stable solution of the current matching argument as described above . scaling limit for @xmath85 and @xmath86 .
left : the thick black line shows the scaling limit result including metastable branches ( dashed ) .
thin coloured lines show the rescaled canonical currents from exact numerics , slowly approaching the scaling limit as @xmath52 increases .
the transition point is given by @xmath285 .
right : @xmath257 for various values of @xmath207 .
solid black corresponds to @xmath287 where the depth of the two local minima are equal .
the green dashed - dot - dot curve shows @xmath257 for @xmath263 , blue dot - dashed curve corresponds to @xmath288 and red dashed @xmath289 .
, title="fig:",scaledwidth=48.0% ] scaling limit for @xmath85 and @xmath86 . left : the thick black line shows the scaling limit result including metastable branches ( dashed ) .
thin coloured lines show the rescaled canonical currents from exact numerics , slowly approaching the scaling limit as @xmath52 increases .
the transition point is given by @xmath285 .
right : @xmath257 for various values of @xmath207 .
solid black corresponds to @xmath287 where the depth of the two local minima are equal .
the green dashed - dot - dot curve shows @xmath257 for @xmath263 , blue dot - dashed curve corresponds to @xmath288 and red dashed @xmath289 .
, title="fig:",scaledwidth=48.0% ] [ cor : cancur ] under the conditions of theorem [ thm : main ] , for @xmath290 @xmath291 and for @xmath289 @xmath292 follows directly by expressing @xmath293 in terms of the ratio of canonical partition functions eq .
( [ eq : cancur ] ) and applying eq .
( [ eq : zln ] ) .
this result is illustrated in fig .
[ fig : limit ] ( left ) . at the transition point
@xmath294 we must examine higher order terms to demonstrate that the system will asymptotically be in a condensed phase , see @xcite and section [ sec : dynamics ] .
having established the existence of metastable phases we can use our previous results to get an estimate of their lifetime .
previous studies include results on the dynamics of fluctuations in a finite system @xcite and the dynamics of the condensate @xcite .
our approach here follows mostly the one in @xcite where metastability in a different zero - range process has been studied , and @xcite where a random walk argument was used to find the characteristic time of the motion of the condensate . in this section we consider the dynamics in the scaling limit defined by eq .
( [ eq : critscale ] ) . in section 5.2
we have seen that for @xmath295 the system exhibits two metastable phases , we assume in the following that we are in this region .
so for sufficiently large @xmath52 we know @xmath152 exhibits a local maximum between the fluid and condensed minima .
let @xmath296 be the local maximum of @xmath178 which is well approximated by the closest integer to the largest root of eq .
( [ eq : curmatch ] ) ( see section [ sec : heur ] ) . in the scaling limit @xmath297
( see section [ sec : rigour ] ) .
we define the metastable fluid ( condensed ) phase as all states for which the maximum site occupation is below ( above ) @xmath298 , i.e. @xmath299 \nonumber\\ \pi^{\textrm{cond}}_{l , n } & : = \pi_{l , n}[\cdot | m_l({{\boldsymbol{\eta } } } ) > m^{*}]\ .\end{aligned}\ ] ] since the process spends very little time close to the maximum of @xmath178 , the precise choice of @xmath296 is not relevant for results in the scaling limit ( cf .
also @xcite for a slightly different approach ) .
the lifetimes of the two metastable phases can be expressed in terms of the following hitting time , @xmath300 we take @xmath301}}\quad\mbox{and}\quad t_{\textrm{cond } } : = 2{{\mathbb{e}_{\pi^{\textrm{cond}}_{l , n}e^{{\mathcal{l}}t } } [ \tau_l]}}\ , \end{aligned}\ ] ] where the expectation is with respect to the dynamics eq .
( [ eq : infgen ] ) with initial distribution given by @xmath302 or @xmath303 .
the factor of 2 comes from the fact that once @xmath296 is reached the process can still return to the same metastable phase with probability @xmath304 . by ergodicity of the process @xmath305 on @xmath54
the ratio of lifetimes is directly related to the stationary distribution , @xmath306}{\pi_{l , n } [ m_l({{\boldsymbol{\eta } } } ) > m^ { * } ] } = \frac{\sum_{k=0}^{m^*}e^{-l i_{l , n}(k)}}{\sum_{k = m^*+1}^{n}e^{-l i_{l , n}(k)}}\ .\end{aligned}\ ] ] applying the scaling limit result of theorem [ thm : main ] , we get for large @xmath52 @xmath307 where the second line follows by a saddle point approximation of the denominator at the local minimum @xmath269 that corresponds to condensed configurations , and by expanding @xmath308 in the numerator at the boundary minimum ( corresponding to fluid configurations ) , keeping the leading order singular term . to leading order in the exponent
this result implies that the ratio of the lifetimes is given by the relative depth of the minima , as expected .
the lower order term @xmath309 in the denominator implies that at the transition density @xmath285 in the scaling limit the condensed phase is stable and the fluid phase is metastable , in accordance with results in @xcite .
however the scaling with @xmath52 is difficult to verify in monte carlo simulations , since system sizes have to be relatively small ( at most of the order of @xmath88 ) to get good statistics on switching times . in this regime higher order
finite size effects still play a role and can affect the location of @xmath105 and therefore the relative depth of the minima of @xmath178 . .[tab:1 ] lifetime of metastable phases for @xmath88 and @xmath310 , with @xmath85 and @xmath86 . using approximation ( [ eq : lifetimes ] ) compared to measurements from monte carlo simulations on a fully connected lattice ( mf ) , a one dimensional lattice with @xmath311 and with @xmath312 . [ cols="<,^,^,^,^ " , ] we estimate the two lifetimes by approximating the number of particles in the condensate by a continuous time random walk .
since it is difficult to validate the scaling with @xmath52 , as discussed above , we will demonstrate the validity of our approach by estimating the lifetimes directly for finite systems .
the actual lifetimes are not only related to the potential barrier given by the maximum of @xmath178 , but also depend on the underlying dynamics .
the occupation of the maximum @xmath313 is a non - markovian , stationary , ergodic process with state space @xmath314 and stationary distribution @xmath315 .
since the process exhibits only single steps , it can be approximated by a continuous time random walk on @xmath316 ( the validity of this markovian assumption is discussed later ) where a particle leaves the maximum ( @xmath317 ) with rate @xmath131 .
the rates corresponding to @xmath318 are then fixed by the stationary distribution @xmath319 . in the scaling limit
this implies that particles enter the condensate with rate given by the current in the background , see corollary [ cor : curmatch ] .
the lifetime of the metastable fluid phase is approximated by the mean first passage time of the random walk starting in state @xmath320 to reach @xmath298 .
for the metastable condensed phase it is given by the mean first passage time starting at @xmath53 , and both have a factor of 2 in front due to the possibility of reaching @xmath296 and not actually switching phase .
these can be calculated using standard techniques ( see e.g. @xcite for the discrete time analog ) , @xmath321 these two approximations can be calculated numerically for reasonably large systems ( up to @xmath322 ) since we can calculate @xmath315 exactly using eq .
( [ eq : canmax ] ) .
applying the scaling limit result and using saddle point approximations we recover an arrhenius estimate with a prefactor for the lifetime of the two metastable phases , @xmath323 here @xmath324 is the location of the local maximum of the rate function , @xmath269 is the local minimum corresponding to condensed configurations and @xmath207 is the boundary minimum corresponding to fluid configurations . to leading order the lifetime of the metastable fluid ( condensed ) phase
is given by the height of the local maximum of the rate function ( potential barrier ) above the fluid ( condensed ) minimum .
this result agrees with the ratio of the lifetimes already discussed .
however as with the previous results the asymptotic form is difficult to validate due to higher order finite size effects , the exact expression eq .
( [ eq : lifetimes ] ) is therefore more appropriate for relevant system sizes .
we expect the above random walk approximation to be accurate if particles exiting the maximum equilibrate in the bulk before returning to the condensate .
this condition is best fulfilled for a fully connected lattice ( mean - field geometry ) and we expect it to be a reasonably good approximation for one dimensional totally asymmetric systems since particles have to pass through the whole fluid bulk before returning . for partial asymmetry and in higher dimensions a return without penetrating the fluid bulk is possible .
to a first approximation the lifetimes have to be multiplied by the inverse probability of the event that particles escape into the bulk , since only such particles have a chance to equilibrate and contribute on the right scale .
this pre - factor can be estimated using another random walk argument .
if a particle jumps to the right with probability @xmath20 and to the left with probability @xmath325 where @xmath326 and @xmath327 then the probability of a particle reaching some macroscopic distance before returning to the condensate is asymptotically @xmath328 .
we therefore expect the lifetime to increase by a factor of @xmath329 for partial asymmetry . as a special case , for symmetric systems in one dimension
this leads to an increase of lifetimes by a factor of @xmath52 by the same argument . for a detailed investigation and validation of this markovian ansatz
see @xcite .
we find that this argument gives a good approximation for fully connected lattices , and totally asymmetric jumps in one dimension , see table [ tab:1 ] .
we observe that the one dimensional totally asymmetric case is a factor of approximately @xmath330 faster than the mean - field case , which is due to internal structure in the fluid background , and has also been observed and discussed in @xcite . for partial asymmetry in one dimension with @xmath312
the process is slower than the totally asymmetric case by a factor of approximately @xmath331 supporting the arguments above which predict an increase by a factor of @xmath332 .
the characteristic time for the motion of the condensate in the thermodynamic limit , @xmath333 , can be approximated by considering eq .
( [ eq : critscale ] ) with a sequence of @xmath207 increasing like @xmath334 . the corresponding rescaled location of the condensed minimum @xmath335 , since the condensed phase has limiting background density @xmath8 . our rigorous results
do not technically hold in this limit , however our estimates of the lifetimes of the metastable phases agree with the characteristic times found by godreche and luck @xcite .
our results demonstrate that close to the transition point in finite systems ( i.e. @xmath336 ) the condensate typically moves via the system entering the fluid phase ( cf .
[ fig : switch ] ) , and therefore its new position is expected to be chosen uniformly at random on @xmath24 . for @xmath207 large , however , the lifetime of the fluid phase becomes small and the condensate can re - locate whilst the system remains in the condensed phase , via the mechanism described in @xcite .
a sub - condensate starts to grow in the fluid bulk , and the potential barrier the process has to cross for condensate motion is associated with the probability of the excess mass shared equally between two sites .
this is the relevant mechanism in the thermodynamic limit . for finite systems one can estimate which of these re - location processes dominates , by considering the distribution of the second largest site using the same techniques as in the previous section .
it has recently been shown that the second mechanism via a sub - condensate can lead to a non - uniform relocation of the condensate @xcite .
we have shown that a prototypical class of zero - range processes , that are known to undergo a condensation transition , exhibit large finite size effects including a metastable switching phenomenon for a large range of system parameters .
we have characterized this behaviour using the background density as an order parameter , by rigorously deriving a large deviation rate function for its distribution in an appropriate scaling limit , which shows a double well structure .
these results agree with recent work on the zero - range process at the critical scale @xcite , which we extend by establishing metastable fluid and condensed phases .
our methods give rise to a simple interpretation of these results in terms of a stationary current balance between the fluid background and the condensate .
all results presented here , except those in section [ sec : dynamics ] , concern properties of the canonical stationary distribution and are therefore independent of the geometry or dimension of the lattice , so long as it permits homogeneous stationary distributions .
our results allow us to estimate the lifetime of the two metastable phases using a heuristic random walk argument .
the estimates agree with previous studies of the dynamics of the condensate in the thermodynamic limit @xcite .
we show that in the critical scaling limit the lifetime of the two phases is growing with system size ( as a stretched exponential ) and derive the appropriate arrhenius law including prefactor .
we have also demonstrated that on finite systems the re - location dynamics of the condensate is dominated by switching to the metastable fluid phase for a large parameter range , rather than the mechanism via a second sub - condensate assumed in previous results , which is relevant in the thermodynamic limit . the current overshoot in the scaling limit for @xmath5 and @xmath86 .
the thick black line shows the scaling limit result and the rescaled canonical currents are shown for various system sizes by thin coloured lines , they slowly approaching the scaling limit as @xmath52 increases .
the transition point is given by @xmath337 which is given in @xcite .
, scaledwidth=48.0% ] the zero - range process provides an interesting example where the thermodynamic limit fails to capture essential parts of the dynamics , and contradicts the usual expectation that finite systems should behave in a smoother fashion than the limiting prediction .
although the aim of this paper is not to discuss particular applications in detail , the finite size effects described here are particularly important for a proper understanding of various real condensation phenomena . to name two relevant examples , the condensing zero - range process has recently been applied as a simplified traffic model as well as to describe clustering phenomena in granular media @xcite .
both of these applications exhibit metastability phenomena , and typical system sizes are of order @xmath338 , the region where metastability effects in our analysis are most relevant .
since our results hold for a large generic class of zero - range models , they suggest that the apparent metastability in many applications is not necessarily the result of a particular choice of the jump rates , but rather a generic finite size phenomenon .
while the results in the scaling limit hold for all zero - range processes with the same tail behaviour as the rates ( [ eq : rates ] ) , the strength of the effect for relevant system sizes will of course depend on the details of the particular application .
traffic modelling is an example where metastability is particularly pronounced @xcite , and it is an interesting question to investigate in detail the applicability of the zero - range process and our results in that area .
using the methods discussed in this paper , a direct application of the findings in @xcite allows us to extend our results on the current overshoot to zero - range processes with rates ( [ eq : rates ] ) where @xmath28 and @xmath339 , see fig .
[ fig : gammaone ] .
the critical scale in this case turns out to be of order @xmath340 , and the system also shows a discontinuous current overshoot in the scaling limit .
however , as can be seen in fig .
[ fig : gammaone ] , convergence is very slow and there are no metastable phases on the critical scale .
although metastability might occur on higher order scales , it will hardly be relevant in any finite size simulation or application .
the scaling limit rate function in theorem [ thm : main ] could also be derived using a slightly modified version of the saddle point methods applied in @xcite
. however , the direct approximation of the fluid and condensed metastable phases used in this paper seems more intuitive for our purpose , and can be made rigorous with less effort than a saddle point computation .
the authors are grateful to hugo touchette , gunter m. schtz , michalis loulakis and robin c. ball for valuable discussions .
this work was supported by the engineering and physical sciences research council , uk , grant no .
ep / e501311/1 .
in eq . ( [ eq : idef ] ) we introduce @xmath341 $ ] for the canonical probability that the maximum site occupation is @xmath142 .
the following upper bound over counts configurations in which more than one site contains @xmath142 particles @xmath342 \leq \frac{l w(m ) \sum_{{{\boldsymbol{\eta}}}\in\hat{x}}\prod_{x=1}^{l-1}w(\eta_x)}{z(l , n ) } , \end{aligned}\ ] ] where @xmath343 .
the following lower bound does not count any configurations in which more than one site contains @xmath142 particles , @xmath342 \geq \frac{l w(m ) \sum_{{{\boldsymbol{\eta}}}\in\check{x}}\prod_{x=1}^{l-1}w(\eta_x)}{z(l , n ) } , \end{aligned}\ ] ] where @xmath344 .
it follows that @xmath342 = \frac{l w(m ) \sum_{{{\boldsymbol{\eta}}}\in\hat{x}}\prod_{x=1}^{l-1}w(\eta_x)}{z(l , n)}\left(1-o(le^{-\frac{-b}{1-\gamma}m^{1-\gamma}})\right)\end{aligned}\ ] ] as @xmath345 .
we may write the last term in the numerator in terms of the cut - off grand canonical measure and get @xmath346 = \nonumber\\ & = \frac{l w(m ) z_{m}^{l{-}1}(\phi)\phi^{{-}(n{-}m)}\nu_{\phi , m}^{l{-}1}\left [ s_{l{-}1}{= } n - m \right]}{z(l , n)}\left(1{-}o(le^{-\frac{{-}b}{1{-}\gamma}m^{1{-}\gamma}})\right)\ , \end{aligned}\ ] ] which holds for all @xmath121 ( cf . section 4.1 ) . finally , taking logarithm , @xmath347 = & -\frac{(l-1)}{l}\big(\log z_{m}(\phi)- \frac{n - m}{l-1}\log \phi \big)+ \nonumber\\ & -\frac{1}{l}\log w(m ) + \frac{1}{l}\log z(l , n)- \frac{1}{l}\log l+ \nonumber\\ & - \frac{1}{l}\log \nu_{\phi , m}^{l-1}\left[s_{l-1}{=}n - m\right ] { + } o(e^{-\frac{-b}{1-\gamma}m^{1-\gamma}}).\end{aligned}\ ] ] from the asymptotic behaviour of @xmath357 eq .
( [ eq : asymwk ] ) we know there exists a @xmath358 such that @xmath359 for all @xmath360 .
so , @xmath361 fix @xmath362
. then there exists @xmath363 such that @xmath364 for @xmath365 .
it follows that @xmath366 for all @xmath367 and @xmath365 .
by applying this upper bound on @xmath368 we can bound the sum above uniformly in @xmath369 as follows , @xmath370 [ convmoms ] under the assumptions of lemma [ taylor ] we may bound the remainder in the taylor expansion of each moment of the cut - off measure @xmath371 introduced in ( [ eq : cutmeas ] ) . writing @xmath372 we have @xmath373}}&= { { \mathbb{e}_{\nu_{\mu_c } } [ \eta_1^i ] } } + \mu_n\partial_{\mu}{{\mathbb{e}_{\nu_{\mu_c } } [ \eta_1^i ] } } + o(\mu_n).\end{aligned}\ ] ] for each @xmath374 the @xmath375 moment of the cut - off ensemble @xmath376 is given by a finite sum .
we let , @xmath377 then @xmath378 . by taylors theorem expanding around @xmath379 , @xmath380 where the remainder term may be expressed as @xmath381 .
\end{aligned}\ ] ] @xmath382 is non - negative and increasing and so the remainder is bounded by @xmath383 by lemma [ taylor ] , we know that there exists @xmath384 such that @xmath385 for all @xmath374 .
it follows that , @xmath386 the result then follows by considering @xmath387 which is @xmath388 since @xmath389 , so the tail of the sum converges as a stretched exponential , this allows us to replace the expected values under the cut - off with @xmath379 with the expected values under the true critical measure . | we study zero - range processes which are known to exhibit a condensation transition , where above a critical density a non - zero fraction of all particles accumulates on a single lattice site .
this phenomenon has been a subject of recent research interest and is well understood in the thermodynamic limit .
the system shows large finite size effects , and we observe a switching between metastable fluid and condensed phases close to the critical point , in contrast to the continuous limiting behaviour of relevant observables .
we describe the leading order finite size effects and establish a discontinuity near criticality in a rigorous scaling limit .
we also characterise the metastable phases using a current matching argument and an extension of the fluid phase to supercritical densities .
this constitutes an interesting example where the thermodynamic limit fails to capture essential parts of the dynamics , which are particularly relevant in applications with moderate system sizes such as traffic flow or granular clustering .
keywords : zero range process ; condensation ; metastability ; finite size effects ; large deviations + |
.inss identification status . the columns
give the name , the year of the proposed identification , the used telescope and its aperture , the magnitude ( @xmath0-band when available ) , the distance , the interstellar absorption @xmath1 and the identification evidence . [
cols="<,<,<,<,<,<,<,<",options="header " , ] caraveo , p.a .
et al . , 2000 ,
proc . of a decade of hst science , eds .
m. livio , k. noll , and m. stiavelli , p.9 gull , t. et al .
2004 , proc .
iau symposium 218 `` young neutron stars and their environments '' , eds f. camilo and b. m. gaensler kern , b. et al . , 2003 apj 597 , 1049 hester , j. j. et al .
2002 , apj , 577 , l49 mignani , r.p .
2004 , proc .
218 `` young neutron stars and their environments '' , eds f. camilo and b. m. gaensler ( astro - ph/0311468 ) mignani , r.p . , et al .
2004a , submitted to a&a gelman , h. b , koch - miramond , l. , auriere , m. 1989 , apj , 342 , 83 shearer , a. et al .
1997 , a&a , 487 , l181 shearer , a. et al .
1998 , apj 335 , l21 wagner , s.j .
& seifert , w. , 2000 , proc . of iau coll .
177 `` pulsar astronomy : 2000 and beyond'',asp conference series , vol .
202 , p. 315
m. kramer , n. wex , and n. wielebinski | the results of optical studies of isolated neutron stars ( inss ) , their pulsar - wind nebulae ( pwne ) and pulsar bow shocks are reviewed and discussed .
pulsars , isolated neutron stars , optical |
consider a mixed integer linear program , and the optimal tableau of the linear programming relaxation .
we select @xmath2 rows of the tableau , relative to @xmath2 basic integer variables @xmath3 .
let @xmath4 denote the nonbasic variables .
let @xmath5 be the value of @xmath6 in the basic solution associated with the tableau , @xmath7 , and suppose @xmath8 .
the tableau restricted to these @xmath2 rows is of the form @xmath9 where @xmath10 , @xmath11 , and @xmath12 denotes the set of integer nonbasic variables . an important question in integer programming
is to derive valid inequalities for , cutting off the current infeasible solution @xmath13 , @xmath14 .
we will consider a simplified model where the integrality conditions are relaxed on all nonbasic variables . on the other hand
, we can present our results in a more general context , where the constraints @xmath15 , @xmath16 , are replaced by constraints @xmath17 , where @xmath0 is the set of integral points in some given rational polyhedron such that @xmath18 , i.e. @xmath0 contains @xmath19 affinely independent points . recall that a polyhedron @xmath20 is _ rational _ if the matrix @xmath21 and vector @xmath22 have rational entries .
so we study the following model , introduced by johnson @xcite .
@xmath23 where @xmath24 . note that every inequality cutting off the point @xmath25 can be expressed in terms of the nonbasic variables @xmath26 only , and can therefore be written in the form @xmath27 . in this paper
we are interested in `` formulas '' for deriving such inequalities .
more formally , we are interested in functions @xmath28 such that the inequality @xmath29 is valid for for every choice of @xmath30 and vectors @xmath31 . we refer to such functions @xmath32 as _ valid functions _ ( with respect to @xmath33 and @xmath0 ) .
note that , if @xmath32 is a valid function and @xmath34 is a function such that @xmath35 , then @xmath36 is also valid , and the inequality @xmath37 is implied by @xmath38 .
therefore we only need to investigate ( pointwise ) minimal valid functions .
andersen , louveaux , weismantel , wolsey @xcite characterize minimal valid functions for the case @xmath39 , @xmath40 .
borozan and cornujols @xcite extend this result to @xmath41 for any @xmath2 .
these papers and a result of zambelli @xcite show a one - to - one correspondence between minimal valid functions and maximal lattice - free convex sets with @xmath33 in the interior .
these results have been further generalized in @xcite .
minimal valid functions for the case @xmath41 are intersection cuts @xcite . our interest in model arose from a recent paper of dey and wolsey @xcite .
they introduce the notion of _ @xmath0-free convex set _ as a convex set without points of @xmath0 in its interior , and show the connection between valid functions and @xmath0-free convex sets with @xmath33 in their interior .
a class of valid functions can be defined as follows .
a function @xmath32 is _ positively homogeneous _
if @xmath42 for every @xmath43 and every @xmath44 , and it is _ subadditive _ if @xmath45 for all @xmath46 .
a function @xmath32 is _ sublinear _ if it is positively homogeneous and subadditive .
it is easy to observe that sublinear functions are also convex .
assume that @xmath32 is a sublinear function such that the set @xmath47 is @xmath0-free .
note that @xmath48 is closed and convex because @xmath32 is convex . since @xmath32 is positively homogeneous , @xmath49 ,
thus @xmath33 is in the interior of @xmath48 .
we claim that @xmath32 is a valid function .
indeed , given any solution @xmath50 to , we have @xmath51 where the first inequality follows from sublinearity and the last one follows from the fact that @xmath52 is not in the interior of @xmath48 .
dey and wolsey @xcite show that every minimal valid function @xmath32 is sublinear and @xmath48 is an @xmath0-free convex set with @xmath33 in its interior . in this paper , we prove that if @xmath32 is a minimal valid function , then @xmath48 is a _
maximal @xmath0-free convex set_. in section [ sec : s - free ] , we show that maximal @xmath0-free convex sets are polyhedra .
therefore a maximal @xmath0-free convex set @xmath53 containing @xmath33 in its interior can be uniquely written in the form @xmath54 .
let @xmath55 be the function defined by @xmath56 it is easy to observe that the above function is sublinear and @xmath57 . in section [ sec : rf ] we will prove that every minimal valid function is of the form @xmath58 for some maximal @xmath0-free convex set @xmath59 containing @xmath33 in its interior .
conversely , if @xmath59 is a maximal @xmath0-free convex set containing @xmath33 in its interior , then @xmath58 is a minimal valid function .
let @xmath60 be the set of integral points in some rational polyhedron of @xmath1 .
we say that @xmath61 is an _
@xmath0-free convex set _ if @xmath59 is convex and does not contain any point of @xmath0 in its interior .
we say that @xmath59 is a _
maximal @xmath0-free convex set _ if it is an @xmath0-free convex set and it is not properly contained in any @xmath0-free convex set .
it follows from zorn s lemma that every @xmath0-free convex set is contained in a maximal @xmath0-free convex set .
when @xmath41 , an @xmath0-free convex set is called a _ lattice - free convex set_. the following theorem of lovsz characterizes maximal lattice - free convex sets .
a linear subspace or cone in @xmath1 is _ rational _ if it can be generated by rational vectors , i.e. vectors with rational coordinates .
[ thm : lattice - free ] _
( lovsz @xcite ) _ a set @xmath62 is a maximal lattice - free convex set if and only if one of the following holds : * @xmath59 is a polyhedron of the form @xmath63 where @xmath64 is a polytope , @xmath65 is a rational linear space , @xmath66 , @xmath59 does not contain any integral point in its interior and there is an integral point in the relative interior of each facet of @xmath59 ; * @xmath59 is a hyperplane of @xmath1 that is not rational .
lovsz only gives a sketch of the proof .
a complete proof can be found in @xcite .
the next theorem is an extension of lovsz theorem to maximal @xmath0-free convex sets .
given a convex set @xmath67 , we denote by @xmath68 its recession cone and by @xmath69 its lineality space . given a set @xmath70 , we denote by @xmath71 the linear space generated by @xmath72 . given a @xmath73-dimensional linear space @xmath74 and a subset @xmath75 of @xmath74
, we say that @xmath75 is a _ lattice of @xmath74 _ if there exists a linear bijection @xmath76 such that @xmath77 .
[ thm : s - free ] let @xmath0 be the set of integral points in some rational polyhedron of @xmath1 such that @xmath18 .
a set @xmath62 is a maximal @xmath0-free convex set if and only if one of the following holds : * @xmath59 is a polyhedron such that @xmath78 has nonempty interior , @xmath59 does not contain any point of @xmath0 in its interior and there is a point of @xmath0 in the relative interior of each of its facets . *
@xmath59 is a half - space of @xmath1 such that @xmath78 has empty interior and the boundary of @xmath59 is a supporting hyperplane of @xmath79 .
* @xmath59 is a hyperplane of @xmath1 such that @xmath80 is not rational .
furthermore , if ( i ) holds , the recession cone of @xmath81 is rational and it is contained in the lineality space of @xmath59 .
we illustrate case ( i ) of the theorem in the plane in figure [ fig : s - free ] .
the question of the polyhedrality of maximal @xmath0-free convex sets was raised by dey and wolsey @xcite .
they proved that this is the case for a maximal @xmath0-free convex set @xmath59 , under the assumptions that @xmath78 has nonempty interior and that the recession cone of @xmath81 is finitely generated and rational .
theorem [ thm : s - free ] settles the question in general .
[ fig : s - free ] -free sets in the plane ( case ( i ) of theorem [ thm : s - free ] ) .
the light gray region indicates @xmath79 and the dark grey regions illustrate the @xmath0-free sets .
a jagged line indicates that the region extends to infinity.,title="fig:",height=264 ] [ fig : s - free ] to prove theorem [ thm : s - free ] we will need the following lemmas . the first one is proved in @xcite and is an easy consequence of dirichlet s theorem .
[ lemma : half - line ] let @xmath82 and @xmath83 .
for every @xmath84 and @xmath85 , there exists an integral point at distance less than @xmath86 from the half line @xmath87 [ lemma : lineality ] let @xmath59 be an @xmath0-free convex set such that @xmath78 has nonempty interior . for every @xmath88 , @xmath89 is @xmath0-free .
let @xmath90 and @xmath91 .
suppose by contradiction that there exists @xmath92 .
we show that @xmath93 . if not , @xmath94 , which implies that there is a hyperplane @xmath95 separating the line @xmath96 and @xmath97 , a contradiction .
thus there exists @xmath98 such that @xmath99 , i.e. there exists @xmath100 such that @xmath59 contains the open ball @xmath101 of radius @xmath86 centered at @xmath102 .
since @xmath103 , it follows that @xmath104 . since @xmath105 , by lemma [ lemma : half - line ] there exists @xmath106 at distance less than @xmath86 from the half line @xmath107 .
thus @xmath108 , hence @xmath109 .
note that the half - line @xmath107 is in @xmath79 , since @xmath110 and @xmath111 . since @xmath79 is a rational polyhedron , for @xmath100 sufficiently small
every integral point at distance at most @xmath86 from @xmath79 is in @xmath79 .
therefore @xmath112 , a contradiction .
the proof of the `` if '' part is standard , and it is similar to the proof for the lattice - free case ( see @xcite ) .
we show the `` only if '' part .
let @xmath59 be a maximal @xmath0-free convex set .
if @xmath113 , then @xmath59 is contained in some affine hyperplane @xmath114 . since @xmath114 has empty interior , @xmath114 is @xmath0-free , thus @xmath115 by maximality of @xmath59 .
next we show that @xmath80 is not rational .
suppose not .
then the linear subspace @xmath116 is rational .
therefore the projection @xmath75 of @xmath117 onto @xmath118 is a lattice of @xmath118 ( see , for example , barvinok @xcite p 284 problem 3 ) .
the projection @xmath119 of @xmath0 onto @xmath118 is a subset of @xmath75 .
let @xmath120 be the projection of @xmath59 onto @xmath118 .
then @xmath121 is the projection of @xmath81 onto @xmath118 . since @xmath59 is a hyperplane , @xmath122 .
this implies that @xmath121 is bounded : otherwise there is an unbounded direction @xmath123 in @xmath124 and so @xmath125 for some @xmath126 . since @xmath127
, this would imply that @xmath128 which is a contradiction .
fix @xmath129 .
since @xmath75 is a lattice and @xmath130 , there is a finite number of points at distance less than @xmath131 from the bounded set @xmath121 in @xmath118 .
it follows that there exists @xmath100 such that every point of @xmath119 has distance at least @xmath86 from @xmath121 .
let @xmath132 .
the set @xmath133 is @xmath0-free by the choice of @xmath86 , but @xmath133 strictly contains @xmath59 , contradicting the maximality of @xmath59 .
therefore @xmath134 holds when @xmath113 .
hence we may assume @xmath135 .
if @xmath78 has empty interior , then there exists a hyperplane separating @xmath59 and @xmath79 which is supporting for @xmath79 . by maximality of @xmath59 case @xmath136
follows .
therefore we may assume that @xmath78 has nonempty interior .
we show that @xmath59 satisfies @xmath137 .
[ claim : p ] there exists a rational polyhedron @xmath64 such that : + @xmath138 @xmath139 , + @xmath140 the set @xmath141 is lattice - free , + @xmath142 for every facet @xmath143 of @xmath64 , @xmath144 is a facet of @xmath114 , + @xmath145 for every facet @xmath143 of @xmath64 , @xmath144 contains an integral point in its relative interior . since @xmath79 is a rational polyhedron ,
there exist integral @xmath21 and @xmath22 such that @xmath146 .
the set @xmath147 satisfies @xmath138 .
the set @xmath148 is lattice - free since @xmath59 is @xmath0-free and @xmath149 does not contain any point in @xmath150 , thus @xmath149 also satisfies @xmath140 .
let @xmath151 be the system containing all inequalities of @xmath152 that define facets of @xmath148 .
let @xmath153 .
then @xmath154 satisfies @xmath155 .
see figure [ fig : proof ] for an illustration .
[ fig : proof ] .,title="fig : " ] it will be more convenient to write @xmath154 as intersection of the half - spaces defining the facets of @xmath154 , @xmath156 .
we construct a sequence of rational polyhedra @xmath157 such that @xmath158 satisfies @xmath159 , @xmath160 , and such that @xmath161 satisfies @xmath145 . given @xmath158 , we construct @xmath162 as follows .
let @xmath163 , where @xmath164 is the set of half spaces defining facets of @xmath158 .
let @xmath165 be a half - space in @xmath166 defining a facet of @xmath167 that does not contain an integral point in its relative interior ; if no such @xmath168 exists , then @xmath158 satisfies @xmath145 and we are done .
if @xmath169 does not contain any integral point in its interior , let @xmath170 .
otherwise , since @xmath158 is rational , among all integral points in the interior of @xmath169 there exists one , say @xmath52 , at minimum distance from @xmath165 .
let @xmath171 be the half - space containing @xmath165 with @xmath52 on its boundary .
let @xmath172 .
observe that @xmath171 defines a facet of @xmath162 since @xmath52 is in the interior of @xmath173 and it is on the boundary of @xmath171 .
so @xmath159 are satisfied and @xmath162 has fewer facets that violate @xmath145 than @xmath158.@xmath174 let @xmath175 be a maximal lattice - free convex set containing the set @xmath114 defined in claim [ claim : p ] . as remarked earlier , such a set @xmath175 exists . by theorem [ thm : lattice - free ]
, @xmath175 is a polyhedron with an integral point in the relative interior of each of its facets .
let @xmath95 be a hyperplane that defines a facet of @xmath64 .
since @xmath176 is a facet of @xmath114 with an integral point in its relative interior , it follows that @xmath95 defines a facet of @xmath175 .
this implies that @xmath177 .
therefore we can write @xmath175 as @xmath178 where @xmath179 are halfspaces .
let @xmath180 , @xmath181 .
[ claim : b - poly ] @xmath59 is a polyhedron .
we first show that , for @xmath182 , @xmath183 .
consider @xmath184 . since @xmath185 and @xmath114 is contained in @xmath175 , @xmath186 .
since @xmath141 and @xmath187 , it follows that @xmath188 .
hence @xmath189 because @xmath190 .
thus , for @xmath182 , there exists a hyperplane separating @xmath59 and @xmath191 .
hence there exists a halfspace @xmath192 such that @xmath193 and @xmath191 is disjoint from the interior of @xmath192 .
we claim that the set @xmath194 is @xmath0-free .
indeed , let @xmath110 .
then @xmath195 is not interior of @xmath175 . since @xmath196 and @xmath190
, @xmath195 is in the interior of @xmath64 .
hence , by , there exists @xmath197 such that @xmath195 is not in the interior of @xmath179 .
thus @xmath198 .
by construction , @xmath195 is not in the interior of @xmath192 , hence @xmath195 is not in the interior of @xmath120 .
thus @xmath120 is an @xmath0-free convex set containing @xmath59 . since @xmath59 is maximal , @xmath199.@xmath174 [ claim : lin - k ] @xmath200 .
let @xmath201 .
we show @xmath202 . by lemma [ lemma : lineality ]
applied to @xmath203 , @xmath204 is lattice - free .
we observe that @xmath89 is @xmath0-free .
if not , let @xmath205 .
since @xmath206 , @xmath207 , hence @xmath208 , a contradiction .
hence , by maximality of @xmath59 , @xmath209 .
thus @xmath210 .
suppose that @xmath211 .
then there exists a facet @xmath143 of @xmath64 that is not parallel to @xmath212 . by construction
, @xmath144 is a facet of @xmath114 containing an integral point @xmath52 in its relative interior .
the point @xmath52 is then in the interior of @xmath204 , a contradiction.@xmath174 [ claim : lin - k - rat ] @xmath69 is rational .
consider the maximal lattice - free convex set @xmath175 containing @xmath114 considered earlier . by theorem [ thm : lattice - free ] ,
@xmath213 , and @xmath214 is rational .
clearly @xmath215 .
hence , if the claim does not hold , there exists a rational vector @xmath216 . by , @xmath217 . since @xmath141 , @xmath218 . hence @xmath219 .
we will show that @xmath89 is @xmath0-free , contradicting the maximality of @xmath59 .
suppose there exists @xmath205 . since @xmath190 , @xmath220 .
therefore @xmath221 .
since @xmath222 , then @xmath223 where the last equality follows from @xmath224 .
this contradicts the fact that @xmath175 is lattice - free .
@xmath174 by lemma [ lemma : lineality ] and by the maximality of @xmath59 , @xmath225 .
[ claim : rational - cone ] @xmath80 is rational .
since @xmath69 and @xmath226 are both rational , we only need to show @xmath227 .
the `` @xmath228 '' direction follows from @xmath229 . for
the other direction , note that , since @xmath230 , we have @xmath231 , hence @xmath232.@xmath174 [ claim : facets ] every facet of @xmath59 contains a point of @xmath0 in its relative interior .
let @xmath65 be the linear space generated by @xmath80 . by claim [ claim : rational - cone ]
, @xmath65 is rational .
let @xmath120 , @xmath119 , @xmath75 be the projections of @xmath59 , @xmath0 , @xmath203 , respectively , onto @xmath118 . since @xmath65 is rational , @xmath75 is a lattice of @xmath118 and @xmath233 .
also , @xmath120 is a maximal @xmath119-free convex set of @xmath118 , since for any @xmath119-free set @xmath234 of @xmath118 , @xmath235 is @xmath0-free .
note that @xmath225 implies that @xmath121 is bounded .
otherwise there is an unbounded direction @xmath123 in @xmath124 and so @xmath236 for some @xmath126 . since @xmath237
, this would imply that @xmath128 which is a contradiction .
let @xmath238 . given @xmath100 ,
let @xmath239 .
the polyhedron @xmath240 is a polytope since it has the same recession cone as @xmath241 .
the polytope @xmath240 contains points of @xmath119 in its interior by the maximality of @xmath120 .
since @xmath75 is a lattice of @xmath118 , @xmath242 has a finite number of points in @xmath119 , hence there exists one minimizing @xmath243 , say @xmath244 . by construction
, the polyhedron @xmath245 does not contain any point of @xmath119 in its interior and contains @xmath120 . by the maximality of @xmath120 , @xmath246
hence @xmath120 contains @xmath244 in its relative interior , and @xmath59 contains a point of @xmath0 in its relative interior . for every maximal @xmath0-free convex set @xmath59
there exists a maximal lattice - free convex set @xmath114 such that , for every facet @xmath143 of @xmath59 , @xmath144 is a facet of @xmath114 .
let @xmath114 be defined as in claim [ claim : p ] in the proof of theorem [ thm : s - free ] .
it follows from the proof that @xmath114 is a maximal lattice - free convex set with the desired properties .
in this section we study minimal valid functions . we find it convenient to state our results in terms of an infinite model introduced by dey and wolsey @xcite . throughout this section ,
@xmath60 is a set of integral points in some rational polyhedron of @xmath1 such that @xmath18 , and @xmath33 is a point in @xmath247 .
let @xmath248 be the set of all infinite dimensional vectors @xmath249 such that @xmath250 where @xmath26 has _ finite support
_ means that @xmath251 is zero for all but a finite number of @xmath43 .
a function @xmath28 is _ valid _ ( with respect to @xmath33 and @xmath0 ) if the linear inequality @xmath252 is satisfied by every @xmath253 .
note that this definition coincides with the one we gave in the introduction
. given two functions @xmath254 we say that @xmath34 _ dominates _
@xmath32 if @xmath255 for all @xmath43
. a valid function @xmath32 is _ minimal _ if there is no valid function @xmath256 that dominates @xmath32 .
[ thm : min - ineq - intr ] for every minimal valid function @xmath32 , there exists a maximal @xmath0-free convex set @xmath59 with @xmath33 in its interior such that @xmath58 dominates @xmath32 .
furthermore , if @xmath59 is a maximal @xmath0-free convex set containing @xmath33 in its interior , then @xmath58 is a minimal valid function . given a valid function @xmath32 ,
define the following function @xmath257 . for all @xmath258 , let @xmath259 .
following the proof of lemma 18 in @xcite one can show that @xmath257 is a valid sublinear function that dominates @xmath32 .
given a valid sublinear function @xmath32 , the set @xmath260 is closed , convex , and contains @xmath33 in its interior . since @xmath32 is a valid function
, @xmath48 is @xmath0-free .
indeed the interior of @xmath48 is @xmath261 .
its boundary is @xmath262 , and its recession cone is @xmath263 . before proving theorem [ thm : min - ineq - intr ] , we need the following general theorem about sublinear functions .
let @xmath114 be a closed , convex set in @xmath1 with the origin in its interior .
the _ polar _ of @xmath114 is the set @xmath264 .
clearly @xmath265 is closed and convex , and since @xmath266 , it is well known that @xmath265 is bounded . in particular , @xmath265 is a compact set . also , since @xmath267 , @xmath268 .
let@xmath269 note that @xmath270 is contained in the relative boundary of @xmath265 .
let @xmath271 be defined by @xmath272 it is easy to show that @xmath273 is sublinear .
[ thm : sublinear ] let @xmath67 be a closed convex set containing the origin in its interior .
then @xmath274 . furthermore , for every sublinear function @xmath275 such that @xmath276
, we have @xmath277 for every @xmath83 .
the polar of @xmath114 is @xmath281 ( see theorem 9.1 in schrijver @xcite ) .
furthermore , @xmath270 is the union of all the facets of @xmath265 that do not contain the origin , therefore @xmath282 for all @xmath83 .
let @xmath253 .
then @xmath284 is in @xmath0 , therefore @xmath285 because @xmath59 is @xmath0-free . by theorem [ thm : sublinear ] ,
thus @xmath287 where the second inequality follows from the sublinearity of @xmath273 .
[ lemma : max - s - free ] let @xmath288 be a closed @xmath0-free convex set containing @xmath33 in its interior , and let @xmath289 . there exists a maximal @xmath0-free convex set @xmath290 such that @xmath291 for @xmath182 .
since @xmath288 is an @xmath0-free convex set , it is contained in some maximal @xmath0-free convex set @xmath175 .
the set @xmath175 satisfies one of the statements ( i)-(iii ) of theorem [ thm : s - free ] . by assumption , @xmath292 and @xmath33
is in the interior of @xmath288 .
since @xmath18 , @xmath79 is a full dimensional polyhedron , thus @xmath293 .
this implies that @xmath294 , hence case ( i ) applies .
since @xmath175 is a polyhedron containing @xmath33 in its interior , there exists @xmath296 and @xmath297 such that @xmath298 , @xmath160 , and @xmath299 . without loss of generality
, we may assume that @xmath300 where @xmath301 denotes the @xmath302th row of @xmath234 , @xmath303 .
by our assumption , @xmath304 .
therefore @xmath305 , since @xmath306 for all @xmath307 .
furthermore @xmath308 , since @xmath304 .
let @xmath309 .
note that @xmath310 . by our choice of @xmath175 ,
@xmath311 is not @xmath0-free for any @xmath312 , otherwise @xmath64 would be contained in a maximal @xmath0-free convex set whose lineality space contains @xmath313 , a contradiction .
we observe next that we may assume that @xmath317 is bounded . indeed , let @xmath318 , @xmath319 , @xmath75 be the projections onto @xmath118 of @xmath64 , @xmath0 , and @xmath203 , respectively . since @xmath65 is a rational space , @xmath75 is a lattice of @xmath118 and @xmath320 .
note that @xmath321 is bounded , since @xmath322 .
if we are given a maximal @xmath319-free convex set @xmath323 in @xmath118 such that @xmath324 and @xmath325 for @xmath326 , then @xmath327 is the set @xmath328 . since @xmath323 contains a point of @xmath319 in the relative interior of each of its facets
, @xmath59 contains a point of @xmath0 in the relative interior of each of its facets , thus @xmath59 is a maximal @xmath0-free convex set .
thus we assume that @xmath317 is bounded , so @xmath329 . if all facets of @xmath64 contain a point of @xmath0 in their relative interior , then @xmath64 is a maximal @xmath0-free convex set , thus the statement of the lemma holds .
otherwise we describe a procedure that replaces one of the inequalities defining a facet of @xmath64 without any point of @xmath0 in its relative interior with an inequality which is a convex combination of the inequalities of @xmath330 , such that the new polyhedron thus obtained is @xmath0-free and has one fewer facet without points of @xmath0 in its relative interior . more formally ,
suppose the facet of @xmath64 defined by @xmath331 does not contain any point of @xmath0 in its relative interior .
given @xmath332 $ ] , let @xmath333(x - f)\leq 1,\quad d_i(x - f)\leq 1\;\ ; i=2,\ldots , t\}.\ ] ] let @xmath339 be generators of @xmath226 .
note that , since @xmath317 is bounded , for every @xmath340 there exists @xmath341 such that @xmath342 .
let @xmath343 be the generators of @xmath226 satisfying @xmath344 note that , if no such generators exist , then @xmath345 is bounded .
otherwise @xmath346 is bounded if and only if , for @xmath347 @xmath348r_j>0.\ ] ] this is the case if and only if @xmath349 , where @xmath350 let @xmath351 be one of the vectors @xmath343 attaining the maximum in the previous equation . then @xmath352 .
thus @xmath353 contains a point of @xmath0 in its interior .
that is , there exists a point @xmath356 such that @xmath357(\bar x - f)<1 $ ] and @xmath358 for @xmath359 . since @xmath64 is @xmath0-free , @xmath360 .
thus there exists @xmath361 such that @xmath362(\bar x - f)=1 $ ] .
note that , since @xmath363 is bounded , there is a finite number of points of @xmath0 in the interior of @xmath364 .
so we may choose @xmath52 such that @xmath98 is maximum .
thus @xmath365 is @xmath0-free and @xmath52 is in the relative interior of the facet of @xmath365 defined by @xmath362(x - f)\leq 1 $ ] .
note that , for @xmath366 , if @xmath367 defines a facet of @xmath64 with a point of @xmath0 in its relative interior , then it also defines a facet of @xmath364 with a point of @xmath0 in its relative interior , because @xmath368 . thus repeating the above construction at most @xmath369 times
, we obtain a set @xmath59 satisfying the lemma .
[ rmk : not - full ] let @xmath288 and @xmath114 be as in lemma [ lemma : max - s - free ] .
given any maximal @xmath0-free convex set @xmath290 containing @xmath288 , then @xmath370 . if @xmath371 is not full dimensional , then the origin is not an extreme point of @xmath265 .
since all extreme points of @xmath265 are contained in @xmath372 , in this case @xmath373 .
therefore , when @xmath371 is not full dimensional , every maximal @xmath0-free convex set containing @xmath288 satisfies the statement of lemma [ lemma : max - s - free ] .
let @xmath32 be a valid function .
by lemma [ lemma : sublinear ] , we may assume that @xmath32 is sublinear .
let @xmath375 , and let @xmath270 be defined as in ( [ eq : set - t ] ) . note that @xmath376 .
thus , by remark [ rmk : valid ] , @xmath377 is valid for @xmath248 .
since @xmath32 is sublinear , it follows from theorem [ thm : sublinear ] that @xmath378 for every @xmath83 . to complete the proof of the theorem
, we need to show that , given a maximal @xmath0-free convex set @xmath59 , the function @xmath58 is minimal .
consider any valid function @xmath32 dominating @xmath58 .
then @xmath380 and @xmath48 is @xmath0-free . by maximality of @xmath59 , @xmath381 . by theorem [ thm : sublinear ] and remark [ rmk : k - poly ] , @xmath382 for all @xmath83 , proving @xmath383 .
[ example : ex1 ] we illustrate the ideas behind the proof in the following two - dimensional example .
consider @xmath384 and @xmath385 .
see figure [ fig : facet - tilting ] .
then the function @xmath386 is a valid linear inequality for @xmath248 .
the corresponding @xmath48 is @xmath387 .
note that @xmath48 is not a maximal @xmath0-free convex set and it corresponds to @xmath288 in lemma [ lemma : max - s - free ] .
following the procedure outlined in the proof , we obtain the maximal @xmath0-free convex set @xmath388
. then , @xmath389 and @xmath58 dominates @xmath32 .
[ rmk : psi>0 ] note that @xmath32 is nonnegative if and only if @xmath390 is not full - dimensional .
it follows from remark [ rmk : not - full ] that , for every maximal @xmath0-free convex set @xmath59 containing @xmath48 , we have @xmath382 for every @xmath43 when @xmath32 is nonnegative .
a statement similar to the one of theorem [ thm : min - ineq - intr ] was shown by borozan - cornujols @xcite for a model similar to when @xmath41 and the vectors @xmath26 are elements of @xmath391 . in this case , it is easy to show that , for every valid inequality @xmath392 , the function @xmath393 is nonnegative .
remark [ rmk : psi>0 ] explains why in this context it is much easier to prove that minimal inequalities arise from maximal lattice - free convex sets .
k. andersen , q. louveaux , r. weismantel , l. a. wolsey , cutting planes from two rows of a simplex tableau , _ proceedings of ipco xii _ ,
ithaca , new york ( june 2007 ) , lecture notes in computer science 4513 , 1 - 15 . | we show that maximal @xmath0-free convex sets are polyhedra when @xmath0 is the set of integral points in some rational polyhedron of @xmath1 .
this result extends a theorem of lovsz characterizing maximal lattice - free convex sets .
our theorem has implications in integer programming .
in particular , we show that maximal @xmath0-free convex sets are in one - to - one correspondence with minimal inequalities . |
the invariant @xmath0 , or equivalently @xmath2 , is a concordance invariant defined by the author and hom @xcite , and by ozsvth - stipsicz - szab @xcite based on rasmussen s local @xmath3 invariant @xcite .
it gives a lower bound on the @xmath1-ball genus of knots and can get arbitrarily better than the bounds from ozsvth - szab @xmath4 invariant . in this paper
, we prove a cabling formula for @xmath0 .
the main result is : [ main ] for @xmath5 and the cable knot @xmath6 , we have @xmath7 when @xmath8 . as an application of the cabling formula , one can use @xmath0 to bound the @xmath1-ball genus of cable knots ; in certain special cases , @xmath0 determines the @xmath1-ball genus precisely .
[ fourball ] suppose @xmath9 is a knot such that @xmath10
. then @xmath11 for @xmath12 .
take @xmath13 , for example .
it is known that @xmath14 . using corollary [ fourball ]
, we can determine the @xmath1-ball genus of any cable knots @xmath6 when @xmath15 .
this generalizes ( * ? ? ?
* proposition 3.5 ) .
regarding the behavior of @xmath4 invariant under knot cabling , the question was well - studied @xcite@xcite@xcite , culminating in hom s explicit formula in @xcite .
in contrast to the rather explicit computational approach used in these papers , our method of study is based on a special relationship between @xmath0 and surgery of knots , and thus avoids the potential difficulty associated to the computation of the knot floer complex @xmath16 . in order to carry out our proposed method
, we need to compute the correction terms on both sides of the reducible surgery ( [ redeq ] ) , which we shall describe in section 3 .
the most technical part of the argument is to identify the projection map of the @xmath17 structures in the reducible surgery , and this is discussed in section 4 .
the proof of the main theorem follows in section 5
. * acknowledgements .
* we like to thank yi ni for a helpful discussion .
the author was partially supported by a grant from the research grants council of the hong kong special administrative region , china ( project no .
cuhk 2191056 ) .
in this section , we review the definition and properties of the @xmath0 invariant from @xcite and relevant backgrounds in heegaard floer theory . heegaard floer homology is a collection of invariants for closed three - manifolds @xmath18 in the form of homology theories @xmath19 , @xmath20 , @xmath21 , @xmath22 and @xmath23 . in ozsvth - szab @xcite and rasmussen @xcite ,
a closely related invariant is defined for null - homologous knots @xmath24 , taking the form of an induced filtration on the heegaard floer complex of @xmath18 . in particular , let @xmath25 denote the knot floer complex of @xmath26 . consider the quotient complexes @xmath27 where @xmath28 and @xmath29 refer to the two filtrations .
the complex @xmath30 is isomorphic to @xmath31 .
associated to each @xmath32 , there is a graded , module map @xmath33 defined by projection and another map @xmath34 defined by projection to @xmath35 , followed by shifting to @xmath36 via the @xmath37-action , and concluding with a chain homotopy equivalence between @xmath36 and @xmath38 .
finally , the @xmath0 invariant is defined as @xmath39 here , @xmath40 denotes the lowest graded generator of the non - torsion class in the homology of the complex , and we abuse our notations by identifying @xmath41 and @xmath31 with their homologies . recall that in the large @xmath42 surgery , @xmath43 corresponds to the maps induced on @xmath44 by the two handle cobordism from @xmath45 to @xmath46 ( * ? ? ?
* theorem 4.4 ) .
this allows one to extract @xmath1-ball genus bound from functorial properties of the cobordism map .
we list below some additional properties of @xmath0 , all of which can be found in @xcite .
* @xmath47 is a smooth concordance invariant , taking nonnegative integer value .
* @xmath48 .
( see @xcite for the definition of @xmath4 ) * for a quasi - alternating knot @xmath9 , @xmath49 * for a strongly quasi - positive knot @xmath9 , @xmath50 for a rational homology @xmath51sphere @xmath18 with a spin@xmath52 structure @xmath53 , @xmath54 is the direct sum of two groups : the first group is the image of @xmath55 $ ] in @xmath54 , which is isomorphic to @xmath56/u{\mathbb{f}}[u]$ ] , and its minimal absolute @xmath57grading is an invariant of @xmath58 , denoted by @xmath59 , the _ correction term _ @xcite ; the second group is the quotient modulo the above image and is denoted by @xmath60 .
altogether , we have @xmath61 using this splitting , we can associate for each integer @xmath32 and the knot @xmath9 a non - negative integer @xmath62 that equals the @xmath37-exponent of @xmath63 restricted to with its homology ] @xmath64 . this sequence of @xmath65 is non - increasing , i.e. , @xmath66 , and stabilizes at 0 for large @xmath32 .
observe that the minimum @xmath32 for which @xmath67 is the same as @xmath47 defined in ( [ def1 ] ) .
this enables us to reinterpret the @xmath0 invariant in the following more concise way .
@xmath68 in addition , the sequence @xmath65 completely determines the correction terms of manifolds obtained from knot surgery .
this can be seen from the surgery formula ( * ? ? ?
* proposition 1.6 ) .
@xmath69 for @xmath5 and @xmath70 .
we will explain this formula in greater detail in section 4 .
we conclude this section by mentioning that an invariant equivalent to @xmath0 , denoted @xmath2 by ozsvth and szab , was formulated in terms of the chain complex @xmath71 in @xcite . that invariant played an important role to establish a @xmath1-ball genus bound for a one - parameter concordance invariant @xmath72 defined in the same reference . for our purpose in the rest of the paper , we will not elaborate on that definition .
recall that the @xmath73 cable of a knot @xmath9 , denoted @xmath6 , is a knot supported on the boundary of a tubular neighborhood of @xmath9 with slope @xmath74 with respect to the standard framing of this torus .
a well - known fact in low - dimensional topology states that the @xmath75-surgery on @xmath6 results in a reducible @xmath51-manifold .
[ redsurg ] @xmath76 the above homeomorphism is exhibited in many references ( cf @xcite ) . for self - containedness , we include a proof of proposition [ redsurg ] below . not only is this reducible surgery a key ingredient of establishing our main result theorem [ main ] , the geometric description of the homeomorphism is also crucial for justifying lemma [ spinproj ]
. denote @xmath77 the tubular neighborhood of @xmath9 and @xmath78 its complement , and let @xmath79 be the boundary torus of @xmath77 .
the cable @xmath6 is embedded in @xmath79 as a curve of slope @xmath74 .
consider the tubular neighborhood @xmath80 of the cable .
the solid torus @xmath80 intersects @xmath81 at an annular neighborhood @xmath82 , and the boundary of this annulus consists of two parallel copies of @xmath6 , denoted by @xmath83 and @xmath84 , each of which have linking number @xmath75 with @xmath6 .
therefore , the surgery slope of coefficient @xmath75 is given by @xmath83 ( or equivalently , @xmath84 ) , and the @xmath75-surgery on @xmath6 is performed by gluing a solid torus to the knot complement @xmath85 in such a way that the meridian is identified with a curve isotopic to @xmath83 .
-handle @xmath86 to @xmath87 along @xmath6 .
the two disk - ends of @xmath86 are identified with the corresponding disk ends of the other @xmath88-handle @xmath89 that is attached to @xmath77 . ] on the other hand , one can think of the above gluing as attaching a pair of @xmath88-handles @xmath86 , @xmath89 to @xmath85 .
see figure [ handle ] .
since the exterior of @xmath6 is homeomorphic to @xmath90 its @xmath75-surgery can be decomposed as @xmath91\cup [ n(k)\cup h_2].\ ] ] as the @xmath88-handles are attached along essential curves on @xmath79 ( isotopic to @xmath83 ) , @xmath92 and @xmath93 end up having a common boundary homeomorphic to @xmath94 .
this proves that @xmath95 is a reducible manifold . to further identify the two pieces of the reducible manifold , note that the attaching curve is isotopic to @xmath83 , which has slope @xmath74 on @xmath79 .
it follows that @xmath96 from the perspective of @xmath87 , the curve @xmath83 has slope @xmath97 .
thus , the other piece is @xmath98 this completes the proof .
let us take a closer look at the surgery formula ( [ corr ] ) , in which there is an implicit identification of @xmath17 structure @xmath99 for simplicity , we use an integer @xmath100 to denote the @xmath17 structure @xmath101)$ ] , when @xmath102\in { \mathbb{z}}/p{\mathbb{z}}$ ] is the congruence class of @xmath28 modulo @xmath103 .
the identification can be made explicit by the procedure in ozsvth and szab ( * ? ? ?
* section 4,7 ) .
in particular , it is independent of the knot @xmath9 on which the surgery is applied ) may be interpreted as comparing the correction terms of the `` same '' @xmath17 structure of surgery on different knots . ]
; and it is affine : @xmath104-\sigma[i]=[k']\in h_1(s^3_{p / q}(k))\cong { \mathrm{spin}^c}(s^3_{p / q}(k))\ ] ] where @xmath105 is the dual knot of the surgery on @xmath9 , and @xmath17 structures are affinely identified with the first homology . moreover , the conjugation map @xmath106 on @xmath17 structures can be expressed as @xmath107))=\sigma([p+q-1-i])\ ] ] ( cf ( * ? ? ? * lemma 2.2 ) ) .
we will use these identifications throughout this paper . in ( * ? ? ?
* proposition 4.8 ) , ozsvth and szab made an identification of @xmath17 structures on lens spaces through their standard genus 1 heegaard diagram , which coincide with the above identification through surgery ( on the unknot ) .
they also proved the following recursive formula for the correction terms of lens spaces @xmath108 for positive integers @xmath109 and @xmath110 , where @xmath111 and @xmath29 are the reduction module @xmath112 of @xmath103 and @xmath28 , respectively . substituting in @xmath113
, one sees : @xmath114 for a reducible manifold @xmath115 , there are projections from @xmath116 to the @xmath17 structure of the two factors @xmath117 and @xmath118 .
particularly , this applies to the case of the reducible surgery @xmath119 . in terms of the canonical identification above
, we write @xmath120 and @xmath121 for the two projections .
these two maps are independent of the knot @xmath9 , which we determine explicitly in the next lemma . with the above notations and identifications of @xmath17 structures understood , we apply the surgery formula ( [ corr ] ) to both sides of the reducible manifold @xmath119 and deduce @xmath122 for all @xmath123 . here
we used the fact that @xmath124 when @xmath123 , as @xmath65 is a non - increasing sequence .
when @xmath9 is the unknot , ( [ longeq ] ) simplifies to @xmath125 as all @xmath126 s are 0 for the unknot .
for the rest of the section , assume @xmath127 , @xmath128 and @xmath129 , and denote @xmath130 and @xmath131 the dual knots of the surgery on @xmath9 and @xmath6 , respectively .
[ spinproj ] the projection maps of the @xmath17 structure @xmath132 and @xmath133 are given by : @xmath134 @xmath135 since the projection maps are affine , we assume @xmath136 @xmath137 note that the maps @xmath132 , @xmath133 are generally not homomorphisms .
nevertheless , we claim that @xmath138 . under ozsvth - szab
s canonical identification @xmath139 , we have @xmath140-\sigma[i]=[k_{p , q}']\in h_1(y)\cong { \mathbb{z}}/pq{\mathbb{z}}$ ] .
thus , it amounts to show that @xmath141=[k']$ ] under the projection of @xmath18 into the first factor @xmath142 . this can be seen from the geometric description of the reducible surgery in last section : the dual knot @xmath143 , isotopic to the closed black curve on the right of figure [ handle ] , projects to an arc in @xmath92 on the left of figure [ handle ] .
this arc is closed up in @xmath142 by connecting it to a simple arc in @xmath144 .
since the curve intersects @xmath83 once , it must represent . ]
@xmath145\in h_1(y_1)$ ] .
hence @xmath146 from which we see @xmath147 .
a similar argument proves @xmath148 . to determine @xmath149 , note that the projection @xmath132 commutes with the conjugation @xmath106 as operations on @xmath17 structures . after substituting the equation @xmath150 into @xmath151 and applying ( [ conjspin ] )
, we get @xmath152 so @xmath153 where the identity is understood modulo @xmath112 as before .
similar arguments also imply : @xmath154 we argue that @xmath155 .
this is evidently true when both @xmath103 and @xmath112 are odd integers .
when @xmath103 is even and @xmath112 is odd , we want to exclude the possibility @xmath156 and @xmath157 using the method of proof by contradiction .
a similar argument will address the case for which @xmath103 is odd and @xmath112 is even , and thus completes the proof .
we derive a contradiction by comparing the correction terms computed in two ways . from equation ( [ lenscor ] ) and
( [ lenscor2 ] ) , we have @xmath158 for @xmath159 . on the other hand
, it follows from ( [ eq1 ] ) that @xmath160 for @xmath161 . here
, we used the fact that @xmath162 for all @xmath163 ( since @xmath164 ) and the assumptions @xmath165 and @xmath166 . comparing the above two identities , we obtain @xmath167 recall from lee - lipshitz ( * ? ? ?
* corollary 5.2 ) that correction terms of lens spaces also satisfy the identity @xmath168 for @xmath169 .
it follows @xmath170 yet , according to ( [ contraidentity ] ) , @xmath171 we reach a contradiction ! this completes the proof of the lemma . as a quick check of lemma [ spinproj ] ,
let us look at the surgery @xmath172 .
the correction terms of the three lens spaces with @xmath17 structure @xmath28 are computed using ( [ lenscor ] ) and summarized in table [ table1 ] below .
.table of correction terms for lens spaces [ cols="^,^,^,^,^,^,^,^,^",options="header " , ] .
in this section , we prove theorem [ main ] . first , note the following relationship between the sequences @xmath173 and @xmath174 if we compare ( [ longeq ] ) and ( [ eq1 ] ) . in order to evaluate @xmath176 from equation ( [ def2 ] ) ,
it is enough to determine the minimum @xmath28 such that @xmath177 . since @xmath178 when @xmath179 , we only need to consider @xmath180 by ( [ cabcorrect ] ) . when @xmath181 , we have @xmath182 , so the condition @xmath123 in lemma [ cablemma ] is satisfied for all @xmath28 in the range @xmath183 .
equation ( [ cabcorrect ] ) simplifies to @xmath184 as @xmath162 and @xmath185 for @xmath165 . as @xmath186 if and only if @xmath187 , it is easy to see from ( [ last ] ) and lemma [ spinproj ] that the minimum @xmath28 such that @xmath177 is @xmath188 . hence , @xmath189 .
the @xmath1-ball genus of the cable knot @xmath192 since on can construct a slice surface for @xmath6 from @xmath103 parallel copies of a slice surface for @xmath9 together with @xmath193 half - twisted bands .
using theorem [ main ] , we see @xmath194 from which corollary [ fourball ] follows immediately . | we prove a cabling formula for the concordance invariant @xmath0 , defined by the author and hom in @xcite .
this gives rise to a simple and effective @xmath1-ball genus bound for many cable knots . |
cold dark matter ( cdm ) dominated cosmological models have received strong support from a wide range of observations ( e.g. , spergel et al .
2003 ) . in cdm models ,
the structures of dark matter are formed through the merger and accretion of smaller structures ( white & rees 1978 ) .
dark matter halos are high - density structures that are a few hundred times of the mean background density and are in approximate dynamical equilibrium according to the virial theorem .
galaxies are expected to condense in dark matter halos due to dissipative cooling processes .
the observable properties of the galaxies are thus strongly influenced by the hierarchical growth of their host halos . during the hierarchical assembly of the halos , smaller halos merged into bigger ones are found to be long - lived even though their outer part may be stripped by the tidal interactions ( e.g. , tormen 1997 ; klypin et al .
1999 ; moore et al .
some of these subhalos are believed to be the hosts of the satellite galaxies in galactic - sized halos ( such as the milky way ) and of the galaxies in clusters of galaxies , therefore it is essential to understand the evolution and dynamics of these subhalos in galaxy formation .
indeed , it has been demonstrated that it is important to resolve the merger of subhalos in order to accurately predict the cluster luminosity function of galaxies ( springel et al .
2001 ) and the luminosity function of field galaxies ( kang et al .
the presence of the subhalos was used to interpret the anomalous flux ratios of lensed quasar images ( e.g. , mao & schneider 1998 ; kochanek & dalal 2004 ) , though it is still unclear if the amount of subhalos predicted by the cdm models is in quantitative agreement with the lensing observations ( mao 2004 ; mao et al .
2004 ; zentner et al .
in addition , the subhalos have important effects on the heating of galactic disks ( benson et al . 2004 ) and possible @xmath0-ray emission from annihilation of dark matter particles ( taylor & silk 2003 ; stoehr et al . 2003 ; see bertone et al .
2004 for a review . ) because of the importance of the subhalos in cosmological studies , there have been many high - resolution n - body studies of subhalos since moore et al .
( 1999 ) who discovered that subhalos can survive within bigger halos for a significant period of time .
the mass function of subhalos was shown to be proportional to @xmath1 ( e.g. , klypin et al .
1999 ; moore et al .
1999 ; ghigna et al .
2000 ) , and the radial distribution of subhalos above a given mass is much flatter than that of the dark matter in the host halo ( e.g. , gao et al .
2004a ; diemand et al .
2004 ; mao et al .
the velocity dispersion of the subhalos is larger than that of the underlying dark matter ( positive velocity bias ) , except in the innermost region where the velocity anti - bias is found for the subhalos ( ghigna et al .
1998 ; okamoto & habe 1999 ; colin et al .
2000 ; klypin et al .
1999 ; springel et al . 2001 ; diemand et al .
2004 ) . in their recent work ,
gao et al .
( 2004a ) showed that most present - day subhalos identified in their simulations have been accreted into their host halos very recently .
this indicates that the subhalos accreted earlier , even if they were massive enough to host bright galaxies , may have been disrupted in the current generation of high - resolution simulations ( gao et al .
2004b ) due to physical ( such as tidal stripping ) and numerical effects ( e.g. , limited resolution ) .
it is therefore very demanding resolution - wise to correctly model the dynamics , mergers and evolution of subhalos and satellites in the densest regions ( kang et al .
2004 ) , and thus an analytical theory would be preferred to follow the evolution of subhalos within their hosts .
indeed , parallel to the numerical studies , significant progress has been made in understanding the evolution and the distribution of subhalos with analytical models ( e.g. , van den bosch et al .
1999 ; klypin et al . 1999 ; taylor & babul 2002 , 2004 ; sheth 2003 ; zentner & bullock 2003 ; ougri & lee 2004 ; van den bosch , tormen & giocoli 2005 ) .
the two principal physical processes , dynamical friction and tidal stripping , which determine the evolution of subhalos , are relatively well understood . however , as noted by benson ( 2005 ) , the initial conditions ( velocities and positions ) of the infalling subhalos adopted in these studies are not from the prediction of hierarchical cdm models . although some of these studies predicted subhalos with physical properties in reasonable agreement with those found in simulations , their success depends in part on the fine - tuning of the model parameters , such as the coulomb logarithm function in the dynamical friction formula ( e.g. , oguri & lee 2004 ) . obviously , in order to make such an analytical approach more useful for studying the evolution of subhalos and satellite galaxies , it is a prerequisite to use the correct initial phase space information of the infalling subhalos as found in the cdm cosmological model .
several authors have already studied the orbital parameters of the subhalos at the time of their mergers with the host halo ( tormen 1997 ; vitvitska et al .
2002 ; khochfar & burkert 2004 ; benson 2005 ) .
tormen ( 1997 ) and khochfar & burkert ( 2004 ) used high - resolution re - simulations of halos and identified the progenitors of these halos .
the orbital distribution is then measured from the progenitors that are about to merge with the main progenitor .
in contrast , vitvitska et al .
( 2002 ) and benson ( 2005 ) identified pairs of halos that are about to merge , and measured the orbital distribution of these pairs . among these studies ,
benson ( 2005 ) used a large set of cosmological simulations provided by the virgo consortium ( jenkins et al .
2001 ) , and formed a large sample of such orbital pairs . from this
, he presented fitting formulae for the distribution of the initial infall velocity of the subhalos at the virial radius of their host halo .
although the work of benson ( 2005 ) is an important step forward to determine the initial condition of infall subhalos , there is an important question yet to be examined quantitatively , i.e. , whether the mergers are isotropic in the position space or there is some preferential direction for the mergers .
earlier studies suggest the accretion is anisotropic ( e.g. , tormen 1997 ) .
for example , aubert , pichon & colombi 2004 ) found the infall takes place preferentially in the plane perpendicular to the direction defined by the spin of the halo . in this paper , we use a cosmological n - body simulation of @xmath2 particles ( jing & suto 2002 ; kang et al .
2004 ) to investigate this problem in more detail .
as we will show , the mergers are preferentially along the major axis of the host halos ; we will quantify this anisotropic accretion as a function of the subhalo mass ( see also libeskind et al .
compared with the simulations used by benson ( 2005 ) , our simulation has higher force and mass resolutions , and the subhalos are better resolved .
this makes it easier and more reliable to quantify the orbital distribution of subhalos .
this is particularly important for small subhalos as they may lose their identities in low resolution n - body simulations when they are close to to the boundary of bigger halos .
it is difficult to account for this population of missing subhalos if only one snapshot of the simulation output is used .
the paper is structured as follows . in 2 , we briefly describe the numerical simulations we use , and how the subhalos are identified . in 3 , we study the physical properties of subhalos and present simple empirical fitting formulae to these properties . in
4 we summarize our main results and discuss areas for future improvement .
the simulation used in this paper is a @xmath3 cosmological simulation of @xmath2 particles in a box of @xmath4 .
the cosmological model is the standard concordance model with the density parameter @xmath5 , the cosmological constant @xmath6 and the hubble constant @xmath7 .
the initial density field is assumed to be gaussian with a harrison - zeldovich primordial power spectrum and with an amplitude specified by @xmath8 , where @xmath9 is the r.m.s .
fluctuation of the linearly evolved density field in a sphere of radius @xmath10 .
this simulation , which started at redshift @xmath11 , is evolved by 5000 time steps to the present day ( @xmath12 ) with our vectorized parallel @xmath3 code ( jing & suto 2002 ) .
the force softening length @xmath13 ( s2 type , hockney & eastwood 1981 ) is @xmath14 comoving , and the particle mass @xmath15 . because these simulation parameters are very similar to those adopted in many high - resolution re - simulations of individual cluster halos ( moore et al .
1999 , jing & suto 2000 , fukushige & makino 2001 , 2003 , power et al .
2003 , diemand et al 2004 ) , we have achieved a resolution that can resolve subhalos within massive host halos . the dark matter halos are identified at redshift @xmath12 in the simulation described above using the friends - of - friends method ( fof ) with a linking length equal to 0.2 of the mean particle separation .
the subhalos are then identified within the fof halos with the subfind routine ( springel et al .
2001 ) . in kang
2004 ) , the mass function of the subhalos was examined , and was found to be in good agreement with the subhalo mass functions obtained in previous halo re - simulations ( springel et al .
2001 ) down to a mass of about @xmath16 ( 50 particles ) .
this indicates that the subhalos with more than @xmath17 particles are resolved in the simulation .
considering the fact that subhalos can survive more easily in the outskirt than in the inner dense region of a host halo , we relax the lower limit to 20 particles for the subhalo mass . table [ table : haloselection ] lists the number of particles we use to identify the subhalos and halos in several different combinations .
and we only use the simulation output at z@xmath18 to calculate the distribution .
we first consider the density profile of host halos as a sphere and search for subhalos within a spherical shell with radius @xmath19 between @xmath20 and @xmath21 , where @xmath19 is the distance between the centres of the subhalo and the host halo in units of the virial radius @xmath22 of the host halo .
we identified the centres " of the halos as the lowest - potential particles using the subfind routine ( springel et al .
2001 ; kang et al . 2004 ) ; we also checked that the centre " identified this way is very close to the centre of mass for most host halos .
we normally take the thickness of the shell ( @xmath23 ) to be 0.1 or 0.2 ( see table [ table : haloselection ] ) .
the virial radius is determined according to the spherical collapse model ( kitayama & suto 1996 ; bryan & norman 1999 ) . in this section ,
we consider only those subhalos with an inward directed velocity .
we treat the subhalos and host halos as point - mass particles , and determine the velocity of the subhalos at the time when their orbits first cross the virial radius of the host halo under gravity .
we also compute the time @xmath24 for each subhalo to cross from @xmath21 to @xmath20 according to their trajectories
. subhalos with a larger @xmath24 will stay longer in the shell of @xmath25 than those with smaller @xmath24 . during a time interval @xmath26
, the chance of observing this subhalo within the spherical shell will be @xmath27 .
thus we weight each subhalo by @xmath28 when we compute the distributions of position , radial and tangential velocities .
we will excise the subhalos that do not pass through one or both of the radial limits .
the distributions of the radial ( @xmath29 ) and tangential ( @xmath30 ) velocities of these subhalos are presented in fig . 1 ; both velocities are in units of the circular velocity at the virial radius , @xmath31 . for this exercise ,
we only retain host halos with more than 500 particles and subhalos with more than 20 particles , corresponding to the selection parameter set a in table [ table : haloselection ] .
the @xmath29 distribution peaks around the virial velocity , while @xmath30 peaks at a slightly smaller value , around 0.7 . both distributions drop essentially to zero beyond 1.5 virial velocity .
the shape of these distributions qualitatively agrees with that obtained by benson ( 2005 ) .
quantitatively , however , both distributions appear slightly broader than those in benson ( 2005 )
. we do not know the exact reason for the difference , but it may be attributed to the different ways of selecting the subhalos . for low resolution simulations ,
subhalos close to the inner shell @xmath32 may be difficult to identify with the friends - of - friends or the spherical overdensity methods .
benson ( 2005 ) attempted to correct for this effect by considering a minimum radius at which the subhalo can still be identified using his algorithm . with
the subfind routine ( springel et al . 2001 ) , we are able to resolve the subhalos even when they are quite close to the host halos , so we do not need to make this complicated correction .
instead we simply consider all the subhalos within the spherical shell between radius @xmath20 and @xmath21 . a related quantity to @xmath29 and @xmath30 is the infall angle , defined as [ eq : theta ] = v_r ( v_r^2+v_^2)^1/2 . for a radially infalling subhalo @xmath33 .
the distribution of @xmath34 is shown in the bottom left panel .
the infall angle has a peak around @xmath35 , and a full - width - at - half - maximum of about @xmath36 .
it is well known that halos accrete matter along the large - scale filaments that connect the halos , this implies that the merger of the subhalos into the host halos may be anisotropic ( lee , jing , & suto 2005 ) . since the halos are generally triaxial ( jing & suto 2002 ) , there could be a correlation between the direction of the subhalo merger and the shape of the host halo .
we therefore determine the three principal axes for each halo from its inertial tensor within the virial radius . we define @xmath37 to be the angle between the major axis of the host halo and the vector from the host halo centre to the centre of a subhalo , and @xmath38 as the other polar angle ( @xmath39 ) . because of the limited sample size
, we will focus on the spatial distribution as a function of @xmath40 @xmath41 , despite of the fact the distribution also depends on @xmath38 but more weakly than on @xmath37 .
the probability distribution of @xmath40 , @xmath42 , for the subhalos is shown in the lower right panel of fig .
if the subhalos merge into the host halo isotropically , then we expect @xmath42 to be unity .
in contrast to this expectation , @xmath43 of the subhalos increases strongly with @xmath40 , implying that subhalos are accreted more preferentially along the major axis of the host halos . since dark halos are triaxial ( jing & suto 2002 ) , the dark matter density at the spherical virial radius is expected to be higher in the direction of the halo major axis .
if the spatial distribution of subhalos follows that of the dark matter , we would expect a higher @xmath43 in the direction of @xmath44 .
the question is whether the increase of @xmath45 with @xmath40 can be fully explained by the shape of host halos . in the following subsection
, we examine this question , i.e. , consider the phase space distribution in the ellipsoidal coordinate system in the more realistic triaxial halo model . for each host halo ,
we determine the axial ratios @xmath46 and @xmath47 from their inertia tensor within its virial radius @xmath22 , where @xmath48 , @xmath49 , and @xmath50 are the minor , median , and major axes of the halo respectively .
if we rotate the coordinate ( @xmath51 ) so that the new @xmath52 , @xmath53 and @xmath54 coordinate axes are parallel to the minor , median , and major axes of the halo , the isodensity surfaces of a halo are approximately described by ( jing & suto 2002 ) @xmath55 we define the boundary of an ellipsoidal halo @xmath56 such that the total mass within @xmath56 is equal to the virial mass of the halo in the spherical model .
we then examine the orbital parameters and density of the subhalos at this surface . in analogy with the previous subsection , we take a shell of the upper and lower ellipsoidal radii @xmath57 in units of @xmath56 , and compute the distributions of the normal and tangential velocities ( @xmath58 and @xmath59 ) relative to the ellipsoidal surface as well as the number density distribution . in this exercise
, we only include the subhalos that have an inward directed velocity .
the infall angle @xmath34 can be similarly defined as in eq .
( [ eq : theta ] ) with @xmath29 and @xmath30 replaced by @xmath58 and @xmath59 .
the results are given in fig .
the distributions of the normal and tangential velocities are very similar to those in spherical shells in the previous subsection .
the subhalo number density @xmath60 , which is the density averaged over @xmath38 directions for a fixed @xmath40 on the surface @xmath56 , is shown as a function of @xmath40 in the middle right panel .
notice that in the triaxial halo model , a flat distribution in @xmath40 implies that the subhalos follow the idealised ellipsoidal shape .
as expected , the dependence on @xmath40 becomes significantly weaker in the ellipsoidal shells , but there is still a preference for the subhalos to merge into their host halos along the major axis . to facilitate comparisons with previous studies ,
we also show in fig . 2 ( bottom two panels ) the distributions of the total velocity @xmath61 , and the orbital ` circularity ' @xmath62 , where @xmath63 is angular momentum and @xmath64 is the angular momentum of a circular orbit with the same energy .
the orbital circularity has a broad peak around 0.5 , indicating the infall is in neither purely radial nor tangential orbits .
this result is in good agreement with previous studies ( e.g. , tormen 1997 , fig .
4 ; ghigna et al .
1998 , fig .
next we examine the distribution of the velocities as a function of the angular positions of the subhalos relative to the major axis of the ellipsoidal host halo . to do this
, we divide the ellipsoid into three equal bins in @xmath65 . in fig .
3 , we show the distributions of the normal and tangential velocities for these three bins . the distributions are quite similar .
the bottom two panels show the mean total velocity and the mean infall angle as a function of @xmath40 .
again these two quantities have little dependence on @xmath40 .
therefore we can use the distributions as shown in figure 2 to more accurately describe the distribution of the subhalo velocities .
we have also studied the distributions as a function of halo mass . for this purpose , we increase the lower limit of the host halos to 5000 particles ( @xmath66 ) which corresponds to the halo selection parameter set c in table 1 . the results are compared with those of the host halos with parameter set b , where we included host halos with more than 500 particles . to increase the number of subhalos in the case of c , we have adopted @xmath67 . as shown in the two bottom panels of figure 3 , we found the statistical distributions change very little with reasonable changes of @xmath23 .
figure 4 shows the results .
no significant difference can be found for the two velocity distributions and the infall angle distribution .
however , the number density dependence on @xmath40 becomes weaker when we increase the lower limit of the host halo mass and hence include more subhalos with smaller @xmath68 .
if the smaller subhalos follow more closely the triaxial density profile , then the trend seen in fig .
4 can be understood because the mass density should follow the triaxial density model . to see the preferential accretion of more massive subhalos along the major axis , in fig .
5 we divide the subhalos in six bins of @xmath68 ( with roughly equal numbers ) and show their distributions of @xmath40 . the mass bins are listed in table 3 .
the lowest mass - bin subhalos have only a weak dependence on @xmath40 , indicating that they are roughly consistent with the triaxial shape
. however , the highest mass subhalos ( top left panel ) with @xmath69 shows dramatic deviation from from the ellipsoidal accretion , with a strong peak around @xmath44 ( i.e. , @xmath70 ) .
5 clearly illustrates that more massive subhalos are accreted with a much stronger preference along the major axis of the host halo .
this may have observable consequences for the satellite galaxies we see today ; we return to this important point briefly in the discussion .
we have shown that the infall of subhalos are preferentially along the major axis of the parent halo determined by the moment of inertia .
an alternative way of defining the orientation of a halo is using its spin axis , which can be defined by the total angular momentum of the particles within the virial radius .
aubert , pichon & colombi ( 2004 ) found that the infall takes place preferentially in the plane perpendicular to the direction defined by the spin of the halo .
[ fig : spin ] shows the distribution of the angle between the spin axis and the line connecting the centres of the parent halo and the satellite in our simulation . clearly there is a ( weak ) preference for satellites to be perpendicular to the spin axis , in agreement with the findings of aubert et al .
( 2004 , see their fig .
8 , which is based on a more sophisticated spherical harmonics analysis ) .
this is to be expected as the spin axis is statistically perpendicular to the major axis of the halo , as shown by previous studies ( e.g. warren et al .
1992 , dubinski 1992 ) .
many analytical models of the subhalo population require an accurate knowledge of the initial conditions for the subhalos .
in the following , we provide empirical fitting formulae found in our numerical simulations . to facilitate comparisons with benson ( 2005 ) ,
we adopt the same functional forms to fit the velocity distribution , specifically , @xmath71\ ] ] where @xmath72,\ ] ] @xmath73.\ ] ] the fitted curves are shown in fig . 2 , and
the fit parameters are listed in table [ table : fit ] .
we also use the following function @xmath74 to describe the angular distribution of the subhalo accretion .
the function has two parameters , @xmath75 and @xmath76 , and @xmath77 is chosen such that the function is properly normalised when @xmath40 is integrated from 0 to unity .
the fit parameters are listed in table [ table : dfdmu ] .
as can be seen , the fitting functions match our simulation results quite well .
we have used a high - resolution simulation to study the initial conditions of subhalos when they merge into their host halos .
most of our results are in good agreement with benson ( 2005 ) .
one finding of our study is that massive subhalos are accreted more preferentially along the major axis of the host halos than the less massive ones .
our subhalos are identified at present day ( @xmath12 ) , but the same trend for massive subhalos should apply at high redshifts ( kravtsov et al .
if the more massive subhalos house satellite galaxies and they survive until the present day , then the satellites should show a more planar geometry along the major axis .
interestingly , the satellite galaxies in the milky way appear to lie in a great disk ( kroupa et al . 2005 ; see also willman et al .
2004 ) , almost perpendicular to the stellar disk .
if the major axis of the dark matter halo is perpendicular to the stellar disk in the milky way , then such a distribution , while puzzling at first , is naturally expected in the cdm ( kang et al .
2005 ; libeskind et al . 2005 ; zentner et al .
2005b ) .
our result that more massive subhalos are more preferentially accreted along the major axis of the host halo can be understood in the cosmic web theory ( bond , kofman , & pogosyan 1996 ; lee , jing , & suto 2005 ) . in this theory , the coherence of the initial tidal field forms one dimensional filaments .
the halos are bridged by the filaments , and the merger of halos is expected to occur along the filaments .
if the major axis of host halos is determined by the neighboring massive filaments , we would expect subhalos to merge preferentially along the major axis .
a quantitative computation is possible for the distribution of @xmath45 following lee et al .
( 2005 ) which will be explored in a future work .
the initial conditions we found can be used in analytical formalisms ( e.g. , oguri & lee 2004 ; taylor & babul 2004 ) to predict the subsequent evolutions of subhalos and compare these with observations .
our results make it straightforward to generate monte carlo realizations of the subhalo population .
the extended press & schechter formalism accurately predict the mass function of the subhalos . for a given halo
, the triaxial halo shape can be sampled using monte carlo with the fitting formulae given by jing & suto ( 2002 ) .
the velocities of the halo can then be obtained using eq .
( [ eq : vr - vtheta ] ) . to describe the angular position of the satellite , the polar angles @xmath78 are needed .
to the first order , the polar angle ( @xmath38 ) is approximately uniformly distributed between 0 to @xmath79 .
the angle @xmath37 can be generated using eq .
( [ eq : mu ] ) .
the mass dependence of the angle @xmath37 can be properly taken into account by choosing different fit parameters ( in table 3 ) depending on the value of @xmath68 . in this paper
we have analyzed the phase - space information of subhalos located in a thin shell close to the virial radius . in practice
, we found that about 30% of subhalos within the shell have outgoing velocities .
some of these subhalos and perhaps even some inward moving subhalos could have been inside the host halo already , thus some of these subhalos could have evolved dynamically within the host halo and their phase - space information can not be regarded as `` initial '' conditions .
the number of secondary infall halos can be crudely estimated as follows .
let us approximate the parent halo as spherical , and assume that there are @xmath80 subhalos falling into the parent halo for the first time .
due to symmetry , we can assume all subhalos move along the positive @xmath81-axis , and a fraction @xmath82 of the subhalos will emerge as outgoing subhalos on the opposite side of the halo
. presumably all these @xmath83 subhalos will be accreted again in a second infall , and some fraction ( presumably @xmath84 ) of these subhalos will emerge as outgoing ones on the positive @xmath81 side .
so in total there are @xmath85 infall subhalos , and @xmath86 outgoing subhalos .
the ratio of infalling and outgoing subhalos is therefore @xmath87 , implying @xmath88 .
hence the fraction of the secondary infall subhalos is about @xmath89 .
gill et al .
( 2005 ) investigated satellite galaxies in the outerskirts of galaxy clusters from their high - resolution simulations ( see also gill et al .
2004a , b ) . they found that approximately one half of the galaxies with current cluster - centric distance in the interval 1 - 2 virial radii of the host are ` backsplash ' galaxies that once penetrated the cluster potential .
if one half of these ` backsplash ' galaxies have negative radial velocities , then this suggests that the fraction of subhaloes undergoing secondary infall is about @xmath90 , consistent with our estimate above . to more properly account for this population of evolved subhalos in detail
, one must track the subhalos as a function of redshift .
this will also be useful for studying the time evolution of initial conditions .
another issue not studied in detail here is the @xmath38 dependence of the subhalo spatial distribution .
this quantity is taken to be uniform between 0 to @xmath79 , while approximately correct to the first order , a more realistic treatment is desirable .
we plan to return to these issues in future works .
we thank the referee for a helpful report which improved the paper .
the research is supported by nkbrsf ( g19990754 ) , nsfc ( nos .
10125314 , 10373012 ) , and shanghai key projects in basic research ( no . 04jc14079 ) .
sm acknowledges the financial support of chinese academy of sciences and the european community s sixth framework marie curie research training network programme , contract no .
mrtn - ct-2004 - 505183 angles " .
hyw and sm wish to thank the hospitality of shanghai astronomical observatory during several scientific visits .
aubert d. pichon c. colombi s. , 2004 , mnras , 352 , 376 benson a. j. , 2005 , , 358 , 551 benson a. j. , lacey c. g. , frenk c. s. , baugh c. m. , cole , s. , 2002 , 351 , 1215 bertone g. , hooper d. , silk j. , 2004 , phys . rep .
405 , 279 bond j. , r. , kofman l. , pogosyan d. , 1996 , nature , 380 , 603 bryan g. l. , norman m. l. , 1998 , apj , 495 , 80 coln p. , klypin a. a. , kravtsov a. v. , 2000 , , 539 , 561 diemand j. , moore b. , stadel j. 2004 , , 352 , 535 dubinski j. , 1992 , , 401 , 441 gao l. , de lucia g. , white s. d. m. , jenkins a. , 2004a , , 352 , l1 gao l. , white s. d. m. , jenkins a. , stoehr f. , springel v. , 2004b , , 355 , 819 ghigna s. , moore b. , governato f. , lake g. , quinn t. , stadel j. , 1998 , , 300 , 146 ghigna s. , moore b. , governato f. , lake g. , quinn t. , stadel j. , 2000 , , 544 , 616 gill s. p. d. , knebe a. , gibson b. k. 2004a , , 351 , 399 gill s. p. d. , knebe a. , gibson b. k. , dopita a. d. , 2004b , , 351 , 410 gill s. p. d. , knebe a. , gibson b. k. , 2005 , , 356 , 1327 hockney r. w. , eastwood j. w. , 1981 , computer simulation using particles ( new york : mcgraw - hill ) jing y. p. , suto y. , 2002 , apj , 574 , 538 kang x. , jing y. p. , mo h. j. , brner g. , 2004 , apj , in press ( astro - ph/0408475 ) kang x. , mao s. , gao l. , jing y. p. , 2005
, a&a , 437 , 383 kitayama t. , suto y. , 1996 , , 469 , 480 klypin a. , kravtsov a. v. , valenzuela o. , 1999 , , 522 , 82 kravtsov a. v. , gnedin o. y. , klypin a. a. , 2004 , apj , 609 , 482 khochfar s. , burkert a. , 2004 , , submitted ( astro - ph/0309611 ) kroupa p. , theis c. , boily c. m. , 2005 , a&a , 431 , 517 kochanek c.s . , dalal n. , 2004 , , 610 , 69 lee j. , jing y.p .
, suto y. , 2005 , apj , in press libeskind n. i. , frenk c. s. , cole s. , helly j. c. , jenkins a. , navarro j. f. , power c. , 2005 , astro - ph/0503400 mao s. , 2004 , iau 220 , eds .
m. a. walker , & k. c. freeman .
asp 237 , p85 mao s. , schneider p. , 1998
, , 295 , 587 mao s. , jing y. p. , ostriker j. p. , weller j. 2004 , , 604 , l5 moore b. , ghigna s. , governato f. et al .
, 1999 , , 524 , l19 oguri m. , lee j. , 2004 , , 355 , 120 okamoto t. , habe a. , 1999 , , 516 , 591 sheth r. k. , 2003 , mnras , 345 , 1200 springel v. , white s. d. m. , tormen g. , kauffmann g. , 2001 , , 328 , 726 spergel d.n.s .
et al . , 2003 , , 148 , 175 stoehr f. , white s. d. m. , springel v. , tormen g. , yoshida n. 2003 , , 345 , 1313 taylor j. e. , babul a. , 2001 , , 559 , 716 taylor j. e. , babul a. , 2004 , , 348 , 811 taylor j. e. , silk j. , 2003 , , 339 , 505 tormen g. , 1997 , , 290 , 411 vitvitska m. , anatoly a. , kravtsov a. v. , wechsler r. h. , primack j. r. , bullock j. s. , 2002 , , 581 , 799 warren m. s. , quinn p. j. , salmon j. k. , zurek w. h. , 1992 , , 399 , 405 willman b. , governato f. , dalcanton j. j. , reed d. , quinn t. 2004 , , 353 , 639 van den bosch f. c. , lewis g. f. , lake g. , stadel j. , 1999 , 515 , 50 van den bosch f. c. , tormen g. , giocoli c. , 2005 , mnras , 359 , 1029 white s. d. m. , rees m. j. , 1978 , , 183 , 341 zentner a. r. , bullock j. s. , 2003 , 598 , 49 zentner a. r. , berlind a. a. , bullock j. s. , kravtsov a. v. , wechsler , r. h. , 2005a , apj , 624 , 505 zentner a. r. , kravtsov a. v. , gnedin o. y. , klypin a. a. 2005b , astro - ph/0502496 .halo and subhalo selection parameters : @xmath91 and @xmath92 are the lower limits of the particle number for the host halos and subhalos respectively . the subhalos are selected within a spherical shell from @xmath20 to @xmath21 in units of the virial radius .
the numbers of selected subhalos are listed in the last row .
[ cols="^,^,^,^",options="header " , ] | we use high - resolution numerical simulations to study the physical properties of subhalos when they merge into their host halos .
an improved algorithm is used to identify the subhalos .
we then examine their spatial and velocity distributions in spherical and triaxial halo models .
we find that the accretion of satellites preferentially occurs along the major axis and perpendicular to the spin axis of the host halo .
furthermore , the massive subhalos show a stronger preference to be accreted along the major axis of the host halo than the low - mass ones . approximate
fitting formulae are provided for the physical properties of subhalos .
combined with analytical and semi - analytic techniques , these empirical formulae provide a useful basis for studying the subsequent evolution of subhalos and satellite galaxies in their hosts .
future studies should however account for satellites that may not be undergoing the first infall in their evolution .
large - scale structure of universe - cosmology : theory - dark matter |
wireless sensor networks ( wsn ) consists of energy constrained sensor nodes having limited sensing range , communication range , processing power and battery .
the sensors generally follow different hop - by - hop ad - hoc data gathering protocols to gather data and communicate .
the sensors can sense the information whichever lies in its sensing range using rf - id ( radio frequency identification ) .
the sensed data can be communicated to another sensor node which lies within communication range of the sender .
the gathered data finally reaches the base station which may be hops apart from any sensor .
since sensor transceivers are omni - directional , we assume the sensing and communication ranges as spheres of certain radii .
the network is called homogeneous when all the sensors have the same radii and heterogeneous otherwise .
coverage of a certain foi and deployment of sensors are an issue of research where the aim is to make energy efficient networks .
sensor deployment and coverage in 2d requires simpler strategies and protocols as compared to 3d .
3d sensor network is used generally for underwater sensor surveillance , floating lightweight sensors in air and space , air and water pollution monitoring , forest monitoring , any other possible 3d deployments etc .
real life applications of wsns are mostly confined to 3d environments .
the term @xmath1-coverage in 3d is used to describe a scenario in which the sensors are deployed in such a way that @xmath1 sensors cover a common region .
more precisely , a point in 3d is said to be @xmath1-covered if it lies in the region that is common to the sensing spheres of @xmath1-sensors , @xmath1 being termed as degree of coverage .
indeed , the ultimate aim of this project is to * come up with a deployment strategy for sensors that guarantees @xmath1-coverage of a given 3-dimensional field of interest ( foi ) for large values of @xmath1*. a first step to address this issue is to come up with a 3-dimensional convex body ( tile ) that is guaranteed to be @xmath1-covered by a certain arrangement of @xmath1 ( or more ) sensors , and then fill the foi with non overlapping copies of that shape by repeating the same arrangement . the term sixsoid has been coined in this paper to signify a geometrical shape that resembles a super - ellipsoid @xcite .
sixsoid is created by the intersection of six sensors , each having the same sensing radius , which are placed on the six face centers of a cube of side length @xmath2 where @xmath2 is the radius of the sensing spheres .
we compare the implications of this convex body with the previously proposed model on 3d @xmath1-coverage based on reuleaux tetrahedron .
recall that the reuleaux tetrahedron , is created by the intersection of four spheres placed on the vertices of a regular tetrahedron of side length @xmath2 . in an attempt to guarantee 4-coverage of the given field
, @xcite considers a scenario in which four sensors are placed on the vertices of a regular tetrahedron of side length equal to the sensing radius @xmath2 .
it is well known that the volume of of a reuleaux tetrahedron constructed out of a regular tetrahedron of side length @xmath2 is approximately @xmath3 @xcite .
unfortunately it is not possible to obtain a tiling of the 3d space with non - overlapping copies of reuleaux tetrahedron @xcite .
in fact , such a tiling is not possible even with a tetrahedron @xcite . in @xcite a plausible deployment strategy is hinted that exploits this construction by overlapping two reuleaux tetrahedrons , gluing them at a common tetrahedron s face , but this deployment does nt seem to be pragmatic . in this paper , we propose another 3d solid ( the sixsoid ) for this purpose and an extremely pragmatic deployment strategy . we show that using our deployment strategy one is guaranteed to have 6-coverage of approximately 68.5% and 4-coverage of 100% of a given 3d polycubical field of interest which is a significant improvement over the gurantees provided in @xcite .
there are relatively fewer works done to address the problem of 3d @xmath1-coverage as compared to the 2d version of the problem . in @xcite
the authors make significant progress on this problem .
they discuss the relevance of reuleaux tetrahedron in various issues dealing with connectivity and @xmath1-coverage in 3d .
prior to that the following works disscuss the 3d version ; in @xcite , authors propose a coverage optimization algorithm based on sampling for 3d underwater wsns .
@xcite proposes an optimal polynomial time algorithm based on voronoi diagram and graph search algorithms , in @xcite authors suggest algorithms to ensure @xmath1-coverage of every point in a field where sensors may have same or different sensing radii , @xcite defines the minimum number of sensors for @xmath1-coverage with a probability value , @xcite studies the effect of sensing radius on the probability of @xmath1-coverage , @xcite proposes an optimal deployment strategy to deal with full coverage and 2-connectivity , @xcite brought forward a sensor placement model based on voronoi structure to cover a 3d region , in @xcite , authors provide a study on connectivity and coverage issues in a randomly deployed 3d wsn . in this paper
, we compare our proposed sixsoid based @xmath1-coverage model with the existing reuleaux tetrahedron based model @xcite .
comparison has been done in terms of volume of @xmath1-coverage , sensor spatial density in 3d and placement and packing in 3d .
in a previous work @xcite , the authors used a convex body that resembles a super - ellipse for @xmath1-coverage in 2d homogeneous wsn . this model
@xcite showed much better efficiency in terms of area of @xmath1-coverage , energy consumption and requirement of less number of sensors as compared to the @xmath1-coverage model with reuleaux triangle @xcite .
this work done in @xcite is further extended to 3d by taking reuleaux tetrahedron .
motivated by the fact that considering a `` super - elliptical '' tile has proven to be much better than the reuleaux triangle based model @xcite , in this work we extend that idea to 3d .
the main hurdle was to compute the volume of the solid ( sixsoid ) generated out of our construction . unlike the reuleaux tetrahedron its volume was not already known .
another reason for opting our construction is its resemblence to a superellipsoid because of which it fits much better in 3d rather than a reuleaux tetrahedron .
moreover , sixsoid provides a practical and easier packing in 3d . once we fill the foi with cubes of side length @xmath4 , a sixsoid is formed by the overlapping of six sensors lying on the six face centers of a cube .
so , every cube contains a sixsoid inside it .
when we consider packing in 3d , cubes are space filling .
we propose a result in a later section that states that our deployment strategy ensures 4-coverage of the entire foi along with fact that approximately 68% of it is 6-covered .
thus , the sixsoid based model ensures at least 4-coverage of the entire foi .
packing of reuleaux tetrahedron in 3d as proposed in @xcite is harder to achieve and it is not feasible for practical deployment . in the subsequent sections , we calculate the volume of the sixsoid with the supporting theorems and compare our results with the reuleaux tetrahedron based model . in the remaining part of the paper , we discuss the computation of volume of a sixsoid in section 3 . in section 4
we discuss the deployment and packing strategy in 3d that exploits the structure of sixsoid .
section 5 is devoted to the comparison of our model with the results proposed in @xcite .
we end the paper with conclusion and potential research directions for future .
the * sixsoid * which we denote by @xmath5 is defined as the 3-d object obtained due to the intersection of six spheres placed on the centers of the faces of a cube of side - length @xmath4 .
the radius of each sphere is @xmath4 . in this section ,
we compute the volume @xmath6 of @xmath5 .
our basic strategy is to take a horizontal cross section ( via a sliding plane ) of the cube at a distance of @xmath7 from the top of the cube and perform the following integration , @xmath8 where @xmath9 is the area of the cross section and the last equality follows due to the symmetry of @xmath10 .
henceforth , we will use the following notation regarding the geometry of the objects involved in this section .
let @xmath11(top ) , @xmath12(bottom ) , @xmath13(left ) , @xmath14(right ) , @xmath15(front ) , @xmath16(back ) be the faces of the cube and we denote the respective spheres ( centered at face centers ) by @xmath17 .
we will denote the sliding plane by @xmath18 and let @xmath19 be its radius .
let @xmath20 be the circle which is obtained by the intersection of @xmath21 with @xmath22 where @xmath23 and @xmath24 . in the remainder of this section
we show the behavior of the cross - section as the plane @xmath21 slides from @xmath25 to @xmath26 . for all @xmath27 , @xmath28 and @xmath29
are concentric circles and @xmath30 . also as @xmath7 varies from @xmath31 to @xmath32
, @xmath33 monotonically decreases , @xmath34 monotonically increases and @xmath35 at @xmath36 . for @xmath27 , @xmath37 .
@xmath38 from @xmath25 to @xmath39 , where @xmath40 we observe that , @xmath9 is exactly @xmath29 as @xmath21 slides below @xmath25 .
so to compute @xmath9 we only need to compute the radius @xmath34 of @xmath29 .
this value can be computed as @xmath41 , which proves the claim regarding @xmath9 .
now inorder to know @xmath42 ( the event when the cross section changes ) , we need to find the value of @xmath7 at which @xmath29 is tangential to @xmath43 , this is also the instance at which @xmath29 is tangential to each @xmath44 and @xmath43 . to find this @xmath7 we need to solve the following equation ,
@xmath45 @xmath46 the root of the above quadratic equation which is less than @xmath32 is @xmath47 .
this proves the lemma .
@xmath48 + @xmath49 @xmath50 where @xmath51 and @xmath52 from @xmath39 to @xmath53 where @xmath54 and @xmath55 recall that above @xmath39 , the cross - section is a circle and at @xmath53 , @xmath29 is tangential to @xmath43 . as @xmath56 slides slightly below @xmath39 ,
the cross - section is a 8-sided closed region drawn in fig.[z2 ] .
let @xmath57 be the center of the square and @xmath58 , @xmath59 .
our aim is to find out the area of the region @xmath60 . from the notation used in the fig .
[ z2 ] we deduce that @xmath61 .
also it is easy to see that @xmath62 where @xmath63 and @xmath64 is the length of the line segment joining @xmath65 and @xmath66 .
we first show how to compute @xmath64 in terms of @xmath4 and @xmath7 . + * computing @xmath64 * : notice that @xmath65 and @xmath66 are points of intersection of the circles @xmath67 and @xmath29 .
let us write the equation of these circles assuming @xmath57 to be the origin and lines parallel to @xmath68 and @xmath69 axis passing though it as @xmath70 and @xmath71 axes .
the equations of @xmath67 and @xmath29 in this coordinate system would be @xmath72 and @xmath73 respectively .
subtracting the later from the former gives @xmath74 .
now if we conside the equation of @xmath67 as a quadratic equation in @xmath71 with @xmath74 then the modulus of the difference of the roots of this equation will exactly be the length of @xmath64 .
simple algebraic manipulation shows that @xmath75 the desired area can be decomposed into two parts .
( i ) 4 times the area of sector @xmath76 ( ii ) 4 times the area of region @xmath77 .
we compute each of these as follows : * area of sector @xmath76 is @xmath78 * area of the region @xmath77 = area of ( @xmath79 ) + area of the cap , which is equal to + @xmath80 + @xmath81 adding the aforementioned two values , multiplying by 4 , and replacing the values of @xmath82 , gives us the expression of @xmath9 in the lemma . + * computing @xmath83 * : to find the value of @xmath7 at which the cross - section changes from the above mentioned 8-sided region , we need to look at the instant when the circle @xmath29 circumscribes the region formed due the intersection of @xmath84 . at this instant
the following equality holds ; @xmath85 which solves to @xmath86 and @xmath87 , since @xmath27 , the only root which is of concern for us is @xmath88 . @xmath89 + 4\left[(y - \frac{r}{2})\right]^2 $ ] . where @xmath90 and @xmath91 , @xmath92 .
we notice that for @xmath93 , @xmath9 is the area which is common to @xmath94 .
fig.[z7 ] shows the geometry of @xmath9 . again
let @xmath57 be the center of the square we divide it into two parts , ( i ) the square @xmath95 ( ii ) four caps surrounding the square each of equal area .
we compute each of these areas as follows : * the area of the square is @xmath96 where @xmath97 is the side length of the square , which according to the figure is @xmath98 .
it can be verified that @xmath99 and @xmath100 .
thus area of the square is @xmath101 * the area of a cap ( say the cap @xmath102 ) can be obtained by subtracting the area of @xmath103 from the area of sector @xmath104 which is equal to @xmath105 where @xmath90 and + @xmath91 , @xmath106 putting the values of @xmath69 and @xmath107 and multiplying the area of the cap by @xmath108 gives us the desired result .
we use observation 1 , to conclude that the cross section does nt change for all values of @xmath7 between @xmath109 and @xmath32 .
thus @xmath110 . as a consequence of the previous results
, we can conclude the following : @xmath111\end{aligned}\ ] ] in order to get a good estimate on the value of the above definite integral , we use matlab to perform the numerical integration .
the value of @xmath6 turns out to be approximately @xmath112 . to compute the volume of the foi we need to compute the volume that is only 3-covered , as the remaining volume is 4-covered .
we notice that as the horizonal plane slides from @xmath113 to @xmath36 , the topology of the cross section that is only 3-covered changes twice , thus to compute the volume inside the cube that is only 3-covered have to perform an integration similar to case of the sixsoid . it can be verified
the two transitions happen at @xmath87 and @xmath114 respectively . in the following we
how @xmath115 , the area of cross section that is only 3-covered , varies between @xmath25 to @xmath116 and between @xmath114 to @xmath36 .
we omit the details of the expression when @xmath7 is between @xmath117 and @xmath114 .
( present in an extended version of the paper ) @xmath118 + 8 \left[(\frac{s^2 \cdot \theta'}{2 } - \frac{\sqrt{3}r^2}{4 } ) \ } \right ] $ ] for @xmath25 to @xmath119 where @xmath106 , @xmath120 , @xmath121 , @xmath122 _ ( sketch ) _ consider the cross - section as depicted in the figure [ z5 ] .
the area that is only 3-covered is the union of regions @xmath123 .
a simple but tideous calculation gives us the result .
@xmath115 is equal to @xmath124 for @xmath125 to @xmath126 . _
( sketch ) _ consider the cross - section as depicted in the figure [ z5 ] .
the area that is only 3-covered is the union of regions @xmath127 .
again a simple calculation gives us the result .
the volume @xmath128 inside a cube that is 4-covered is @xmath129\end{aligned}\ ] ] where @xmath130 , @xmath131 we again use numerical methods to estimate this integral , which is approximately @xmath132 .
thus the volume of region that is 4-covered is approximately @xmath133 .
consider a tessellation of a ( poly)cubical 3d field of interest ( foi ) using cubes of side length @xmath4 .
we place the sensors each of sensing radius @xmath134 on the centers of the faces of every cube .
clearly , every sensor will be shared by 2 cubes in the tiling ( except the boundary ones ) . according to the aforementioned packing every point in the foi
is 3-covered , approximately 95.2 % of foi is 4-covered and about 68.5 % of the foi is 6-covered . by construction ,
every point inside a sixsoid formed will be covered by 6 sensors .
we just have to prove that every other point will be 4-covered . according to the notations used in the construction of sixsoid in section 3 , which we borrow here , this is equivalent to proving that every point in the cross - section ( which is a sqaure of side length r ) obtained by slicing the cube with the plane @xmath21 is 3-covered for all @xmath135 . again by symmetry of @xmath136 , we only have to prove this for all values between @xmath27 . as we have noticed in section 3 that due to the nature of sixsoid the topology of the cross section changes at two values of @xmath7 , namely @xmath39 and @xmath53 .
consider the arrangement of circles formed on the square of side length @xmath4 , it can be verified that for all the ranges when the topology of the cross - section remains fixed ( i.e. from 0 to @xmath42 , @xmath42 to @xmath83 and @xmath83 to @xmath32 ) the entire cross section is 3-covered . also from the previous section the claims regarding 4 and 6
coverage follows .
recall that according to our deployment strategy every sensor is shared by 2 cubes . thus given a 3d cubical foi of volume @xmath137
the number of sensors of radius @xmath2 needed according to our deployment strategy is @xmath138 where @xmath2 is the radius of the sensing spheres .
.volume comparison between sixsoid and reuleaux tetrahedron [ cols="<,<,<",options="header " , ]
in this section we present the comparison of our model with the releaux tetrahedron model .
we compare the volume of sixsoid with the volume of reuleaux tetrahedron with varying sensing radius .
this comparison is tabulated in table [ t2 ] .
next , we evaluate the sensor spatial density for @xmath1-coverage in the sixsoid model and compare it with the reuleaux tetrahedron model . recall that the minimum sensor spatial density per unit volume needed for full @xmath1-coverage of a 3d field is defined as a function of sensing radius @xmath2 as @xmath139 where @xmath140 .
we take the similar parameter for reuleaux tetrahedron from @xcite , which is @xmath141 where @xmath142 .
table [ t3 ] shows the comparison of the minimum sensor spatial densities and we find that sixsoid requires much less sensors per unit volume to guarantee @xmath1-coverage .
in this paper , we discussed a new geometric model for addressing the problem of @xmath1-coverage in 3d homogeneous wireless sensor networks which we have named as the sixsoid .
we show in a number of ways how this model outperforms the existing model of reuleaux tetrahedron @xcite , namely the volume of the convex body involved , which in turn implies improved spatial density and a better ( in terms of coverage ) and more pragmatic deployment strategy for sensors in a given 3d field of interest . from the point of view of geometry ,
our construction of sixsoid and its volume computation might be of independent interest .
we suspect our construction might have interesting consequences for non - homogenoeus networks as well which we leave as a direction of future research .
ammari , on the problem of @xmath1-coverage in 3d wireless sensor networks : a reuleaux tetrahedron - based approach , 7th international conference on inteligent sensors , sensor networks and information processing ( 2011 ) x. bai , c. zhang , d. xuan , j. teng , w. jia , constructing low - connectivity and full - coverage three dimensional wireless sensor networks , ieee journal on selected areas in communications , vol .
28 , issue : 7 a. jaklic , a. leonardis , f. solina , superquadratics and their geometrical properties , computational imaging and vision , chap .
20 , kluwer academic publishers , dordrecht , http://lrv.fri.uni-lj.si/ franc / srsbook / geometry.pdf ( 2000 ) | coverage in 3d wireless sensor network ( wsn ) is always a very critical issue to deal with .
coming up with good coverage models implies energy efficient networks .
@xmath0-coverage is a model that ensures that every point in a given 3d field of interest ( foi ) is guaranteed to be covered by @xmath1 sensors . in the case of 3d ,
coming up with a deployment of sensors that gurantees @xmath1-coverage becomes much more complicated than in 2d .
the basic idea is to come up with a convex body that is guaranteed to be @xmath1-covered by taking a specific arrangement of sensors , and then fill the foi with non - overlapping copies of this body . in this work ,
we propose a new geometry for the 3d scenario which we call a * sixsoid*. prior to this work , the convex body which was proposed for coverage in 3d was the so called * reuleaux tetrahedron * @xcite .
our construction is motivated from a construction that can be applied to the 2d version of the problem @xcite in which it implies better guarantees over the * reuleaux triangle*. our contribution in this paper is twofold , firstly we show how sixsoid gurantees more coverage volume over reuleaux tetrahedron , secondly we show how sixsoid also guarantees a simpler and more pragmatic deployment strategy for 3d wireless sensor networks . in this paper , we show the construction of sixsoid , calculate its volume and discuss its implications on the @xmath1-coverage in 3d wsns .
wireless sensor networks , @xmath1-coverage , reuleaux tetrahedron , sixsoid . |
mass outflows of ionized gas from active galactic nuclei ( agn ) have been detected in the majority of seyfert 1 galaxies observed in the uv ( crenshaw et al .
1999 ; kriss 2002 ) and x - rays ( reynolds 1997 , george et al .
the outflows are revealed through absorption lines that are blueshifted with respect to the systemic velocities of the host galaxies .
observations at high spectral resolution ( @xmath5/@xmath6 @xmath7 10,000 ) in the uv show that the absorption tends to split into multiple kinematic components with widths in the range 20 500 km s@xmath1 .
detailed studies of the `` intrinsic '' absorbers indicate large global covering factors ( @xmath3 @xmath8 0.5 1.0 ) of the nucleus and mass - loss rates compable to the inferred mass accretion rates ( @xmath00.01 m@xmath9 yr@xmath1 ) ( crenshaw et al .
the origin of the intrinsic absorbers ( e.g. , accretion disk , torus ) and the means by which they are accelerated outward ( e.g. , thermal winds , radiatively - driven flows , hydromagnetic flows ) are currently unknown .
the large global covering factors of the uv absorbers suggest that they should contribute to at least some of the uv and optical emission that we see .
thus , it is only natural to ask if they are associated in some way with either the narrow ( emission ) line region ( nlr ) or the broad ( emission ) line region ( blr ) .
an important clue is that the highest radial velocity observed in any seyfert 1 galaxy is only@xmath22100 km s@xmath1 ( crenshaw et al .
1999 ) , which is comparable to the maximum velocity ( half - width at zero intensity ) of the nlr gas and much less than that of the blr gas , which extends out to @xmath710000 km s@xmath1 .
furthermore , the available evidence indicates that uv absorbers have transverse velocities comparable to their radial velocities ( crenshaw et al .
2003a , 2004a ) .
thus , the kinematics of the absorbers suggest that they are more closely related to the nlr than the blr .
another clue to the origin of the absorbers and their connection to the emission - line regions is their radial location with respect to the central supermassive black hole ( smbh ) , but this is a difficult parameter to obtain .
early studies of the ultraviolet spectrum of ngc 4151 with the _ international ultraviolet explorer _ found that the uv absorbers must lie outside of the broad ( emission ) line region ( blr ) , since both continuum and broad - line emission were absorbed ( bromage et al .
this was later found to be the case as well for a sample of seyfert 1 galaxies observed by the _ hubble space telescope _ ( _ hst _ ) ( crenshaw et al . 1999 ) .
various studies have placed upper or lower limits on the distances of outflowing uv or x - ray absorbers from their nuclei , and these fall in the range of tenths to tens of parsecs of the systemic velocity can often , but not always , be attributed to gas at much larger distances in the host galaxy s disk or halo ( weymann et al .
1997 ; crenshaw & kraemer 2001 ) . ]
( kriss et al . 1997 ; espey et al . 1998 ; netzer et al . 2002 ; kraemer et al .
2002 ; gabel et al . 2003a ; netzer et al .
2003 ) .
recently , we determined a reliable distance for an outflowing uv absorber ( gabel et al .
2004a ) , which is component `` 1 '' in ngc 3783 , at a radial velocity of @xmath21350 km s@xmath1 with respect to the systemic velocity of the host galaxy . from the relative strengths of the metastable c iii * @xmath51175 multiplet lines in absorption , we determined that the hydrogen number density of the absorber was @xmath10 @xmath4 4 @xmath11 10@xmath12 @xmath13 .
combined with the ionization parameter and ionizing luminosity , this yields a distance of @xmath025 pc from the central continuum source ( gabel et al .
this absorber is therefore in the inner nlr , which typically extends out to several hundred parsecs from the nucleus ( schmitt et al .
2003a , b ) . given their locations and velocities , are the uv absorbers just nlr clouds seen in absorption ?
global estimates of the covering factor of the nlr in seyfert 1 galaxies yield values of @xmath3 @xmath4 0.02 ( netzer & laor 1993 ) .
furthermore , nlr clouds tend to have a lower ionization parameter ( @xmath14 ) , defined as the ratio of ionizing photons to hydrogen atoms at the ionized face , and higher hydrogen column density ( @xmath15 ) than uv absorbers , on average ( crenshaw et al
. however , previous studies have already shown that a typical nlr is composed of a multi - phase medium ; clouds with different physical conditions ( @xmath14 , @xmath15 , @xmath10 ) coexist at the same radial locations ( kraemer & crenshaw 2000 , kraemer et al .
one possibility is that uv absorbers are associated with a high - ionization , diffuse component of the nlr , rather than the entire nlr .
this idea is similar to that of cecil et al .
( 2002 ) , who suggest that the high - velocity ( extending up to @xmath23200 km s@xmath1 ) emission - line clouds in the nlr of ngc 1068 would resemble the `` associated absorbers '' in quasars if seen against the continuum source .
however , these high - velocity clouds have a covering factor of @xmath3 @xmath16 0.02 . in this paper
, we investigate the possibility that there is a high - ionization component of the nlr with a covering factor that is sufficiently large to account for most of the intrinsic uv absorption in seyfert galaxies .
_ hst _ images of the nlr show that nearly all seyfert 1 and 2 galaxies have a bright central knot of [ o iii ] emission ( schmitt et al .
2003a ) , and space telescope imaging spectrograph ( stis ) slitless spectra of these nuclear knots show they have high velocity dispersions ( ruiz et al .
their high surface brightnesses , proximity to their nuclei , and high velocity dispersions indicate they are good candidates for uv absorbers seen in emission . in this paper
, we explore a number of their properties based on the stis spectra of ruiz et al .
( 2005 ) . to demonstrate that the nuclear emission - line knots can be responsible for the uv absorption
, it is necessary to show that their spectra show evidence for a high - ionization , high covering factor component .
we present a stis uv / optical spectrum of the nuclear [ o iii ] knot in ngc 4151 that is ideal for this purpose ( 4 ) .
first , however , we provide new measurements that explore the kinematic and positional similarities between the uv absorbers ( 2 ) and the nlr ( 3 ) .
observations at high spectral resolution ( @xmath5/@xmath6 @xmath17 10,000 ) are needed to resolve the kinematic components of intrinsic absorption in seyfert galaxies .
this requirement is met in the uv ( 1200 3200 ) by observations with the high - resolution modes of _ hst _ s stis or the first - generation goddard high - resolution spectrograph ( ghrs ) , and in the far - uv ( 900 1200 ) with the _ far ultraviolet spectroscopic explorer _ ( _ fuse _ ) .
we have compiled the available measurements of the radial - velocity centroid ( @xmath18 ) , relative to the host galaxy , and full - width at half - maximum ( fwhm ) of the absorption depth for individual absorption components observed by these instruments .
the sample contains 61 absorption components in 10 seyfert 1 galaxies : ngc 3516 ( crenshaw et al . 1999 ; kraemer et al . 2002 ) , ngc 3783 ( kraemer et al . 2001a ; gabel et al . 2003 ) , ngc 5548 ( mathur et al . 1999 ; crenshaw et al . 2003 ) , mrk 509 ( kriss et al . 2000 ; kraemer et al . 2003 ) , ngc 4051 ( collinge et al . 2001 ) , ngc 4151 ( weymann et al . 1997 ; kraemer et al .
2001b ) , mrk 279 ( scott et al .
2004 ; j.r .
gabel , 2004 , private communication ) , ngc 7469 ( kriss et al .
2003 ) , akn 564 ( crenshaw et al . 2002 ) , and ngc 4395 ( crenshaw et al .
2004b ) .
figure 1 plots the fwhm vs. @xmath18 for each intrinsic absorption component in our sample .
although the sample is rather sparse , there is no clear trend ; components at both high and low @xmath19 show a large range in fwhm .
the tendency for most low ( within @xmath20 250 km s@xmath1 ) @xmath18 components to have low ( @xmath21 100 km s@xmath1 ) fwhm reflects the finding that a large fraction arise in the seyfert host galaxies , as has been demonstrated for specific components in ngc 4151 ( weymann et al .
1997 ) , akn 564 ( crenshaw et al .
2002 ) , and ngc 4395 ( crenshaw et al .
furthermore , a galactic origin may very well explain all of the cases of positive @xmath18 .
however , some of these low @xmath18 components are likely intrinsic to the agn ; for example , a component in ngc 3516 at @xmath18 @xmath8 30 km s@xmath1 has shown strong variability in its column density , indicating that it arises close to the nucleus ( kraemer et al .
2002).15 km s@xmath1 for _ fuse _ , @xmath08 km s@xmath1 for stis ) , so that lines with the lowest fwhm ( 20 30 km s@xmath1 ) are just barely resolved .
some of the absorption lines are likely saturated , so the fwhm may overestimate their intrinsic velocity spread .
the four points with low fwhm at @xmath0 @xmath22100 km s@xmath1 are subcomponents of an absorption feature in ngc 7469 ( kriss et al .
the point with very large fwhm ( 940 km s@xmath1 ) represents the transient absorption component d@xmath22 in ngc 4151 ( kraemer et al .
2001a ) . ] for our purposes , the importance of figure 1 is that it shows the radial velocities of intrinsic uv absorbers are @xmath23 2100 km s@xmath1 in seyfert galaxies , and the majority of components ( 51 out of 61 ) have @xmath24 1000 km s@xmath1 . in a survey of low - resolution uv spectra of seyfert galaxies
, we found a similar result : the highest ( most negative ) radial velocity was @xmath22150 km s@xmath1 , for i zw 1 ( crenshaw et al .
transverse velocities of the absorbers are more difficult to obtain , but a few lower limits on the order of @xmath01000 km s@xmath1 have been determined by detecting the motion of absorbing clouds across the blr ( crenshaw et al .
2003a , and references therein )
. the only reasonably accurate transverse velocity to date is for `` component 1 '' in ngc 3783 , which lies in the range @xmath25 @xmath8 540 1430 km s@xmath1 ( crenshaw , kraemer , & gabel 2004a ) .
many other absorption components have been monitored on a yearly basis and have not shown evidence for transverse motion , so the transverse velocities could be much lower for these components .
although more work needs to be done , the available evidence indicates that the space velocities of uv absorbers in seyfert galaxies are much less than those of blr clouds , and very similar to the velocities in the nlr .
we require the high spatial resolution of _ hst _ to isolate the nuclear emission - line knots and measure their sizes , and the spectral resolution of stis to accurately measure their kinematics .
stis medium - dispersion spectra of the bright [ o iii ] @xmath55007 emission line are ideal for this purpose .
we use the stis slitless spectra of ngc 4151 ( hutchings et al .
1998 ; kaiser et al .
2000 ) and nine other seyfert galaxies ( ruiz et al . 2005
; see this paper for details on the observations and data reduction ) , obtained with the g430 m grating and an open aperture .
the spectra have a spatial resolution of 01 ( full - width at half - maximum ) ( fwhm ) and a velocity resolution of 33 km s@xmath1 ( fwhm ) in the dispersion direction .
ruiz et al .
( 2005 ) found that nine of the ten seyfert galaxies in their sample show bright , compact ( fwhm @xmath21 05 ) [ o iii ] knots at their optical continuum peaks with apparently high velocity dispersions .
the only exception is mrk 3 , which shows _ several _ bright knots with high velocity dispersions within @xmath01@xmath26 of its optical nucleus ; we do not use mrk 3 for this study . for the 10 nuclear emission - line knots , we measured the half - width at half - maximum ( hwhm ) of the [ o iii ] @xmath55007 emission in the dispersion ( radial velocity ) and cross - dispersion ( spatial ) directions .
we also measured the half - width at zero intensity ( hwzi ) in the dispersion direction .
we use these parameters rather than the full widths to compare with the uv absorber kinematics , because they are more appropriate for simple geometries . for an expanding shell , for example
, the hwhm of the intensity profile in the spatial direction can be compared to the average absorber distance , and the hwzi in the spectral direction can be compared to the maximum absorber radial velocities . however , we note that direct comparisons can only be made by assuming very specific geometric and kinematic models .
table 1 shows measurements of the nuclear knots in our sample , which includes two seyfert 1 , one seyfert 1.9 , and seven seyfert 2 galaxies at redshifts @xmath21 0.03 .
three of the nuclear knots are only marginally resolved , and we list their hwhm sizes ( in arcsecs and parsecs ) as upper limits . the hwhm s range from 7 to 67 pc , similar to estimates for the distances of uv absorbers from their continuum sources , as described in 1 .
the hwzi s range from 310 to 1060 km s@xmath1 , similar to the range of maximum absorber velocities found in most seyfert galaxies , as described in 2 .
to determine the physical conditions in a nuclear emission - line knot , we would like to have extensive uv and optical coverage to get as many emission - line diagnostics as possible . we also require the high spatial resolution of _ hst _ to isolate the knot . of the sources in table 1 ,
only ngc 4151 has low - dispersion stis spectra covering a large wavelength range ( 1150 10,270 ) .
we searched the multimission archives at the space telescope science institute ( mast ) and found two sets of low - dispersion spectra of the nucleus : one obtained on 1998 january 8/february 10 ( published in nelson et al . 2000 ) and one obtained on 2000 may 24/28 ( previously unpublished ) .
the latter set was obtained when the uv and blr fluxes were in a much lower state ( continuum flux at 1465 @xmath8 2.0 ( @xmath200.3 ) @xmath11 10@xmath27 ergs s@xmath1 @xmath28 @xmath1 ) compared to that of the 1998 observations ( 24.1 ( @xmath200.5 ) @xmath11 10@xmath27 ergs s@xmath1 @xmath28 @xmath1 ) , and is therefore ideal for deconvolving emission lines with both broad and narrow components .
the only drawback in using ngc 4151 is that it has strong intrinsic absorption lines that contaminate the narrow emission - line spectrum in the uv . in particular , absorption due to the host galaxy and/or halo ( components f and f@xmath22 in weymann et al . 1997 )
severely absorb the narrow emission components of the uv resonance lines of l@xmath29 , n v , o i , c ii , si iv , c iv , and mg ii , to the extent that their nlr contributions are not measurable even in stis echelle spectra ( see kraemer et al .
however , when ngc 4151 is in a high state , the unusual ( for seyferts ) balmer absorption lines nearly disappear ( hutchings et al .
thus , we use the high - state slitless spectra of h@xmath30 and h@xmath29 obtained on 1997 july 15 ( previously published in kaiser et al .
2000 ; hutchings et al .
2002 ) to deconvolve their narrow and broad emission components . in table 2
, we list the stis observations that we retrieved from the mast for this analysis .
the first four observations are the low - dispersion , long - slit spectra , which have a resolving power of @xmath5/@xmath6 @xmath4 1000 , and the last two are the medium - dispersion , slitless spectra of h@xmath30 ( plus [ o iii ] @xmath31 4959 , 5007 ) and h@xmath29 ( plus [ n ii ] @xmath316548 , 6584 ) , with @xmath5/@xmath6 @xmath4 9000 .
we reduced the stis low- and medium - dispersion spectra using the idl software developed at nasa s goddard space flight center for the instrument definition team .
we identified and removed cosmic - ray hits using multiple images and removed hot or warm pixels by interpolation in the dispersion direction .
we used wavelength calibration exposures obtained after each science observation to correct the wavelength scale for zero - point shifts .
for the g750l spectra , we used a contemporaneous continuum - lamp exposure in the flat - fielding process to remove the fringes at the long - wavelength end .
we extracted the spectra from the flux - calibrated ccd images using a bin height of 11 pixels ( 055 ) in the cross - dispersion direction , which corresponds to the full - width at zero - intensity ( fwzi ) of the [ o iii ] nuclear knot from the g430 m slitless spectrum .
for the flux - calibrated mama images , we used the same extraction height of 055 , which corresponds to 22 pixels on the mama detector .
we combined the low - dispersion spectra in their regions of overlap to obtain a single spectrum from 1150 10,270 . in figure 2
, we show the full low - dispersion spectrum of the nuclear emission - line knot in ngc 4151 .
narrow emission lines covering a broad range in ionization are present , from [ o i ] to [ fe xi ] ; the ionization potential needed to create the latter is ip@xmath32 @xmath8 262 ev .
we see no evidence for even higher ionization lines , like the [ s xii ] @xmath57611 line ( ip@xmath32 @xmath8 505 ev ) seen in the stis spectrum of the hot spot in ngc 1068 ( kraemer & crenshaw 2000a ) .
the broad components of the permitted lines are clearly evident but still unusually weak for ngc 4151 .
although uv resonance lines are detected in emission , their fluxes are severely affected by the intrinsic absorption lines , as discussed previously .
we measured the fluxes of isolated emission lines by direct integration above local continuum fits . for narrow lines that were blended with other broad or narrow emission
, we used the [ o iii ] @xmath55007 profile as a template to deconvolve the blends ( see crenshaw & peterson 1986 ) .
we used the medium - dispersion spectra to deconvolve the blend of narrow and broad h@xmath30 and the blend of narrow and broad h@xmath29 and the [ n ii ] lines .
as shown in figure 3 , the narrow component of h@xmath30 , which we use as our standard for line ratios , was easily separated from the broad component in this medium - dispersion spectrum .
lines that we deblended in the low - dispersion spectra include the broad and narrow components of the h and he permitted lines and the semi - forbidden lines in the uv , as well as the narrow [ s ii ] @xmath316716 , 6731 lines and h@xmath33 plus [ o iii ] @xmath54363 lines .
we determined the reddening of the the narrow emission lines from the he ii @xmath51640/@xmath54686 ratio , the galactic reddening curve of savage & mathis ( 1979 ) , and an intrinsic he ii ratio of 7.2 .
we determined uncertainties in the dereddened ratios from the sum in quadrature of the errors from three sources : photon noise , different reasonable continuum placements , and reddening .
table 3 gives the observed and dereddened narrow - line ratios , relative to h@xmath30 , as well as the uncertainties in these ratios .
the reddening determined from the he ii ratio is only e(b@xmath2v ) @xmath8 0.02 @xmath20 0.04 .
the reddening correction has a very small effect on our line ratios , as shown in table 3 , and uncertainties in the correction have a negligible effect on our results .
the reddening that we have determined can be attributed entirely to reddening in our galaxy , which is e(b@xmath2v ) @xmath8 0.03 ( schlegel et al .
thus , there is no evidence for dust in the inner nlr of ngc 4151 , although at other locations in the nlr , the observed reddening of the emission lines ranges from e(b@xmath2v ) @xmath8 0.01 to 0.36 ( kraemer et al .
2001b ) . to explore the physical conditions and covering factor of the nuclear knot in ngc 4151 , we generated photoionization models using the code cloudy ( ferland et al . 1998 ) .
we modeled the emission - line gas as single - zoned slabs , directly irradiated by the central source .
we used the same spectral energy distribution ( sed ) of the ionizing continuum radiation as we used previously for ngc 4151 ( kraemer et al .
2001b ) , which consist of broken power - laws of the form @xmath34 , where the spectral index @xmath35 for h@xmath36 @xmath21 13.6 ev , @xmath37 over the range 13.6 ev @xmath16 h@xmath36 @xmath21 1 kev , and @xmath38 above 1 kev , and the luminosity in ionizing photons is 2 x 10@xmath39 s@xmath1 .
we assumed roughly solar elemental abundances ( e.g. grevesse & anders 1989 ) and that the gas was free of cosmic dust .
the models are parameterized in terms of the ionization parameter ( @xmath14 ) , the hydrogen number density ( @xmath10 ) , and the hydrogen column density ( @xmath15 ) .
the emission - lines in table 3 span a huge range in ionization and critical density , and it is clear that multiple model components are required .
since the nuclear knot is resolved , we used the radius inside of which one - half of the knot s [ o iii ] emission is contained as a starting point for the distances of the components .
this radius is 015 , which corresponds to a projected distance of 9.5 pc from the central continuum source .
the dereddened he ii @xmath54686/h@xmath30 ratio in table 3 is relatively high for photoionized gas , which suggests either 1 ) a much harder continuum sed than we have used , or 2 ) a strong contribution from a high - ionization , matter - bound ( optically thin to hydrogen - ionizing radiation ) component .
there is no evidence for a much harder sed and , in fact , we have used the above sed to successfully model the emission - line ratios from other parts of the nlr in ngc 4151 ( kraemer et al .
furthermore , any sed that possesses a realistic euv bump would boost both the he ii and [ ne v ] emission , but what is required is a high [ ne v]/he ii ratio ( table 3 ) .
this can only be achieved with an optically thin component that has a large ne@xmath40 zone prior to the point where the gas becomes optically thick to the he ii ionizing radiation ( kraemer & crenshaw 2000 ; kraemer et al .
2000 ) .
we have therefore adopted a matter - bound component ( `` uvabs '' ) with parameters typical of most uv absorbers ( crenshaw et al .
2003 ) : @xmath41 and @xmath42 10@xmath43 @xmath28 ( with the above parameters , @xmath10 is fixed to 10@xmath44 @xmath13 ) .
the predicted ionic column densities listed in the footnotes to table 3 would result in absorption lines that are typical of most uv absorbers : strong - to - saturated ly@xmath29 , c iv @xmath31 1548.2 , 1550.8 , n v @xmath31 1238.8 , 1242.8 , and o vi @xmath311031.9 , 1036.3 , and weak or undetectable si iv @xmath311393.8 , 1402.8 . in order to fit the observed emission - line ratios , we required two additional components : a moderate density , radiation - bound component ( `` oiii '' ) to reproduce the strong [ o iii ] @xmath55007 emission , and a higher density , radiation - bound component ( `` dense '' ) to account for the strong [ o iii ] @xmath54363 line ( see table 3 ) .
we were unable to find a satisfactory single - component solution , since , for example , the [ ne iv ] @xmath5 2423 line is quenched at intermediate densities .
the oiii component was kept at a distance of 9.5 pc , consistent with measurements of the [ o iii ] @xmath55007 emission .
we kept the same ionization parameter ( @xmath45 0.01 ) for the dense component , but increased its density , which effectively moves it to a distance of 2.0 pc from the continuum source . to produce a composite model spectrum , we adjusted the weight of each component according to the fraction of h@xmath30 it contributes until we obtained a good match with the dereddened line ratios .
we give the component and composite line ratios in table 3 , along with the final model parameters , the weight of each component , and the flux of h@xmath30 at the ionized face of each component ( @xmath46 ) .
nearly all of the dereddened line ratios are fit well by the composite model .
the low critical density lines ( see osterbrock 1989 ) from low ionization species such as [ o ii ] @xmath53227 and [ s ii ] @xmath31 6716 , 6731 are underpredicted , which could be rectified by introducing an additional component outside the emission - line bicone that is irradiated by a heavily absorbed continuum , similar to the model described for the hot spot in ngc 1068 ( kraemer & crenshaw 2000a ) . also , the model significantly underpredicts the strengths of the [ fe vii ] and [ fe xi ] lines , but this is a common problem in nlr models ( oliva 1997 ) that may be due to problems with the atomic data ( s.b .
kraemer & g.j .
ferland , 2004 , in preparation ) .
table 3 shows that the inclusion of the uvabs component allows us to match the high he ii / h@xmath30 ratio .
uvabs also provides a significantly better fit to the other he ii lines and the high - ionization n iv ] , [ ne iv ] , and [ ne v ] lines ( note that he ii @xmath53204 is in a noisy region of the spectrum and has a large uncertainty ) .
this permits a column density of @xmath010@xmath47 @xmath28 within the central knot .
we then generated a model of collisionally ionized plasma with these initial conditions using cloudy90 .
the predicted emitted he ii @xmath54686 flux was 4.27 @xmath11 10@xmath48 ergs @xmath28 s@xmath1 , which is 0.08 the contribution from uvabs , or 0.028 of the total observed he ii .
hence , this component , which , of course would contribute less if the temperature were higher , can not contribute significantly to the observed he ii emission .
] even with the inclusion of uvabs , our model underpredicts the [ ne v]/h@xmath30 ratio somewhat .
this can not be corrected with a harder sed because it would lead to an overprediction of he ii / h@xmath30 , as discussed earlier in this section .
one plausible explanation is that the ionizing continuum has been modified by optically thin gas closer to the nucleus , which has absorbed some of the radiation near the he ii lyman limit but has no effect at energies @xmath17 97 ev ( the ionization potential of ne@xmath49 ) .
the result would be that the [ ne v ] would be enhanced with respect to he ii , as demonstrated in kraemer et al .
( 2000 ) . based on our model predictions
, we can constrain the covering factors of the individual components .
the extinction - corrected h@xmath30 luminosity of the central knot is 6.4 x 10@xmath50 erg s@xmath1 .
based on the relative contribution to h@xmath30 from each model and the emitted flux of h@xmath30 at the ionized face ( table 3 ) , we derive the following covering fractions : for uvabs , @xmath3 @xmath8 1.1 ; for oiii , @xmath3 @xmath8 0.02 ; and for dense , @xmath51 @xmath8 0.03 .
the global covering factor of the uvabs component is close to one , which , given our assumptions , is consistent with the derived value of @xmath3 @xmath8 0.5 1.0 for intrinsic uv absorption in seyfert galaxies ( crenshaw et al .
increasing the column density @xmath15 of uvabs would decrease @xmath3 , but this would decrease the he ii / h@xmath30 ratio significantly below the observed value.e '' in kraemer et al .
2001b ) shows evidence for partial covering in the line - of sight , which indicates partial global covering .
however , this component is much closer to the source and has a much higher column density than uvabs ; there are other uv absorption components in ngc 4151 that may fully cover the continuum source . ]
we conclude that there is strong evidence for a high - ionization , matter - bound component in the nuclear knot of ngc 4151 with a high global covering factor .
we have shown that the bright central knots of emission in the nlrs of most seyfert galaxies are likely sources of intrinsic absorption lines that we detect in their uv spectra , particularly absorbers that are outflowing with radial velocities @xmath52 km s@xmath1 .
our evidence is based on the similarities in the kinematics and locations of the nlr knots and the uv absorbers .
more work must be done to test these connections . in particular , more information is needed on the distances of the uv absorbers from their central continuum sources , based on monitoring campaigns and/or detection of metastable lines ( see crenshaw et al .
we note that intrinsic absorption lines that are within a few hundred km s@xmath1 of the systemic velocity can arise from either the agn or host galaxy ; the percentage from each is not known and deserves further study . to establish a more direct link between the emission and absorption , we modeled the narrow emission lines from the resolved central knot in ngc 4151 .
we found that the line ratios indicate a high - ionization , matter - bound component in the nuclear emission - line knot .
our characterization of this component in the form of `` uvabs '' yields a high global covering factor , which matches that found for intrinsic uv absorption in seyfert 1 galaxies ( @xmath53 ) .
given the physical conditions of uvabs , it would be detected as a typical uv absorber if seen projected against the continuum source and blr .
we note that the observed uv spectrum of ngc 4151 shows components of absorption that span a wide range in ionization and column density ( kraemer et al .
2001b ) , but not a component that specifically matches our uvabs model component for the central emission - line knot .
this is likely due to the fact that we are viewing the agn at a special angle , close to the edge of the emission - line bicone ( crenshaw et al . 2000 ) , which may explain its complex absorption spectrum ( kraemer et al .
the uvabs component could be `` hidden '' in the complex of absorption , specifically in the broad component `` d+e '' , or just not present in the line of sight due to a variation in absorption properties with polar angle ( with respect to the accretion and/or torus axes ) .
nevertheless , an important follow - up study would be to model the emission - line spectra of central knots in other seyfert galaxies , and compare these with their absorption - line properties to test the generality of our results .
a number of dynamical models invoke accretion - disk winds to explain the intrinsic absorption ( and broad - line emission ) in agn , using radiation pressure ( murray et al .
1985 ; proga et al .
2000 ) and/or centrifugal acceleration along magnetic field lines ( blandford & payne 1982 ; bottorff et al .
the absorbers in these models are located close to the blr , which is only light - days from the central continuum source ( presumably the accretion disk around the central supermassive black hole ) . however , our results suggest that a large fraction , and perhaps a majority , of uv absorbers are located in the inner nlrs of seyfert galaxies , at distances of tens of parsecs from their central continuum sources , and are therefore not directly associated with accretion - disk winds .
this does not rule out an origin in the accretion disk , but other origins for the outflowing gas , such as the torus ( krolik & kriss 1995 ) are also possible .
once again , ngc 4151 appears to be an exceptional case , presumably as a result of our special viewing angle .
the broad component `` d+e '' is at a distance of only @xmath00.03 pc from the continuum source ( kraemer et al .
2001b ) , and may be the best candidate for an accretion - disk wind in a seyfert galaxy .
there are also outflowing absorbers at large distances from the nucleus ( @xmath0700 and @xmath02100 pc ) in ngc 4151 , which may be associated with emission - line clouds further out in the nlr or in the extended nlr ( enlr ) .
some of the absorbers in other seyfert galaxies may also arise in these more distant regions .
finally , we note that x - ray absorbers , which tend to have higher column densities than uv absorbers and likely dominate the mass outflow rates , may also be associated with the nlr .
spectra at the highest possible resolutions with _ hst_/stis and the _ chandra x - ray observatory _
( _ cxo _ ) indicate that the uv and x - ray absorption in individual seyfert galaxies cover very similar ranges in radial velocity ( crenshaw et al . 2003a , and references therein ) .
this suggests that the uv and x - ray absorbers may be spatially colocated ; for example the uv absorbers could be high - density knots in an x - ray wind ( krolik & kriss 1995 , 2001 ) . also , _
cxo _ observations of ngc 4151 and ngc 1068 ( ogle et al .
2000 , 2003 ) show extended x - ray emission - line regions that are colocated with their nlrs , which would likely be seen as x - ray absorbers if projected against the continuum source .
further identification of components in both emission and absorption should lead to much tighter constraints on the physical conditions , geometry , and dynamics of the outflowing gas .
some of the data presented in this paper were obtained from the multimission archive at the space telescope science institute ( mast ) .
stsci is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 .
this research has made use of nasa s astrophysics data system .
kriss , g.a . ,
krolik , j. , grimes , j. , tzvetanov , z. , espey , b. , zheng , w. , & davidsen , a. 1997 , in emission lines in active galaxies : new methods and techniques , ed .
peterson , f .- z .
cheng , & a.s .
wilson , ( san francisco : asp ) , asp conference series , 113 , 453 lccccccc mrk 573 & 2 & 0.0172 & 0.20 & 66 & 100 & 340 + mrk 620 & 2 & 0.0061 & 0.14 & 17 & 90 & 310 + ngc 1386 & 2 & 0.0031 & 0.24 & 14 & 440 & 860 + ngc 4151 & 1 & 0.0033 & 0.11 & 7 & 150 & 1060 + ngc 3081 & 2 & 0.0079 & @xmath160.07 & @xmath1611 & 100 & 500 + ngc 3516 & 1 & 0.0083 & @xmath160.07 & @xmath1611 & 110 & 670 + ngc 5506 & 1.9 & 0.0061 & @xmath160.07 & @xmath168 & 60 & 800 + ngc 5643 & 2 & 0.0040 & 0.16 & 12 & 110 & 680 + ngc 5728 & 2 & 0.0093 & 0.20 & 36 & 140 & 730 + ngc 7212 & 2 & 0.0267 & 0.13 & 67 & 260 & 1010 + ccrcccr 2000 may 28 & mama & g140l & 1150 1717 & 52 x 0.1 & 83 & 2220 + 2000 may 24 & ccd & g230 lb & 1672 3077 & 52 x 0.1 & 83 & 2825 + 2000 may 24 & ccd & g430l & 2906 5715 & 52 x 0.1 & 83 & 780 + 2000 may 24 & ccd & g750l & 5277 10269 & 52
x 0.1 & 83 & 780 + 1997 july 15 & ccd & g430 m & 4817 5098 & 52 x 52 & 45 & 2139 + 1997 july 15 & ccd & g750 m & 6294 6862 & 52 x 52 & 45 & 1860 + lcccccc n iv ] @xmath51486 & [email protected] & [email protected] & 0.67 & 1.82 & 0.32 & 0.65 + he ii @xmath51640 & [email protected] & [email protected] & 2.22 & 6.88 & 1.58 & 1.60 + o iii ] @xmath51663 & [email protected] & [email protected] & 1.05 & 0.17 & 0.57 & 1.63 + n iii ] @xmath51750 & [email protected] & [email protected] & 0.45 & 0.15 & 0.26 & 0.67 + c iii ] @xmath51909 & [email protected] & [email protected] & 4.77 & 1.27 & 3.08 & 6.90 + o iii]/c ii ] @xmath52324 & [email protected] & [email protected] & 0.58 & & 0.34 & 0.90 + @xmath54ne iv ] @xmath52423 & [email protected] & [email protected] & 0.35 & 1.34 & 0.39 & 0.09 + o iii @xmath53133 & [email protected] & [email protected] & & & & + he ii @xmath53204 & [email protected] & [email protected] & 0.12 & 0.35 & 0.08 & 0.09 + @xmath54ne v ] @xmath53346 & [email protected] & [email protected] & 0.38 & 1.94 & 0.15 & 0.19 + @xmath54ne v ] @xmath53424 & [email protected] & [email protected] & 1.03 & 5.23 & 0.41 & 0.50 + @xmath54fe vii ] @xmath53588 & [email protected] & [email protected] & & & & + @xmath54o ii ] @xmath53727 & [email protected] & [email protected] & 0.11 & & 0.25 & 0.02 + @xmath54fe vii ] @xmath53760 & [email protected] & [email protected] & & & & + @xmath54ne iii ] @xmath53869 & [email protected] & [email protected] & 1.92 & 0.07 & 1.69 & 2.54 + @xmath54ne iii ] @xmath53967 , h@xmath55 & [email protected] & [email protected] & & & & + @xmath54s ii ] @xmath54072 & [email protected] & [email protected] & 0.26 & & 0.13 & 0.32 + h@xmath56 @xmath54100 & [email protected] & [email protected] & 0.26 & 0.26 & 0.26 & 0.26 + h@xmath33 @xmath54340 & [email protected] & [email protected] & 0.47 & 0.47 & 0.47 & 0.47 + @xmath54o iii ] @xmath54363 & [email protected] & [email protected] & 0.92 & 0.05 & 0.35 & 1.56 + he ii @xmath54686 & [email protected] & [email protected] & 0.30 & 0.89 & 0.22 & 0.21 + h@xmath30 @xmath54861 & 1.00 & 1.00 & 1.00 & 1.00 & 1.00 & 1.00 + @xmath54o iii ] @xmath54959 & [email protected] & [email protected] & 4.53 & 0.46 & 7.21 & 3.48 + @xmath54o iii ] @xmath55007 & [email protected] & [email protected] & 13.61 & 1.39 & 21.65 & 10.43 + @xmath54fe vii ] @xmath55159 & [email protected] & [email protected] & & & & + @xmath54fe vi ] @xmath55176 & [email protected] & [email protected] & & & & + @xmath54fe vii ] @xmath55721 & [email protected] & [email protected] & & & & + he i @xmath55876 & [email protected] & [email protected] & 0.11 & & 0.11 & 0.13 + @xmath54fe vii ] @xmath56087 & [email protected] & [email protected] & 0.15 & 0.29 & 0.12 & 0.14 + @xmath54o i ] @xmath56300 & [email protected] & [email protected] & 0.55 & & 0.39 & 0.80 + @xmath54o i ] , [ fe x ] @xmath56374 & [email protected] & [email protected] & 0.25 & 0.50 & 0.13 & 0.27 + @xmath54n ii ] @xmath56548 & [email protected] & [email protected] & 0.25 & & 0.37 & 0.23 + h@xmath29 @xmath56563 & [email protected] & [email protected] & 2.91 & 2.71 & 2.89 & 2.98 + @xmath54n ii ] @xmath56584 & [email protected] & [email protected] & 0.76 & & 1.11 & 0.68 + @xmath54s ii ] @xmath56716 & [email protected] & [email protected] & 0.04 & & 0.09 & 0.02 + @xmath54s ii ] @xmath56731 & [email protected] & [email protected] & 0.10 & & 0.19 & 0.05 + @xmath54ar v ] @xmath57005 & [email protected] & [email protected] & 0.03 & 0.01 & 0.03 & 0.04 + he i @xmath57065 & [email protected] & [email protected] & 0.08 & & 0.08 & 0.09 + @xmath54ar iii ] @xmath57136 & [email protected] & [email protected] & 0.23 & & 0.25 & 0.26 + @xmath54 o ii ] @xmath57325 & [email protected] & [email protected] & 0.22 & & 0.20 & 0.28 + @xmath54 ar iii ] @xmath57751 & [email protected] & [email protected] & 0.05 & & 0.06 & 0.06 + @xmath54fe xi ] @xmath57892 & 0.10 @xmath200.02 & [email protected] & 0.01 & 0.08 & & + @xmath54s iii ] @xmath59069 & [email protected] & [email protected] & 0.35 & & 0.55 & 0.29 + @xmath54s iii ] @xmath59532 & [email protected] & [email protected] & 0.88 & 0.01 & 1.37 & 0.71 + | we present evidence that the outflowing uv absorbers in seyfert 1 galaxies arise primarily in their inner narrow ( emission ) line regions ( nlrs ) , based on similarities in their locations , kinematics , and physical conditions .
1 ) _ hubble space telescope _
observations show that nearly all seyfert galaxies have bright , central knots of [ o iii ] emission in their nlrs with radii of tens of parsecs .
these sizes are consistent with most previous estimates of the distances of uv ( and x - ray ) absorbers from their central continuum sources , and a recently - obtained reliable distance of @xmath025 pc for a uv absorber in the seyfert 1 galaxy ngc 3783 .
2 ) the nuclear emission - line knots in a sample of 10 seyfert galaxies have velocity widths of 300 1100 km s@xmath1 ( half - width at zero intensity ) , similar to the radial velocities of most uv absorbers .
the highest radial velocity for a seyfert uv absorber to date is only @xmath22100 km s@xmath1 , which is much lower than typical broad - line region ( blr ) velocities .
there is also mounting evidence that the nlr clouds are outflowing from the nucleus , like the uv absorbers .
3 ) if our hypothesis is correct , then the nlr should have a component with a high global covering factor ( @xmath3 ) of the continuum source and blr , to match that found from previous surveys of uv absorbers ( @xmath3 = 0.5 1.0 ) .
using stis spectra of ngc 4151 , obtained when the continuum and blr fluxes were low , we find evidence for optically thin gas in its nuclear emission - line knot .
we are able to match the line ratios from this gas with photoionization models that include a component with @xmath3 @xmath4 1 and an ionization parameter and hydrogen column density that are typical of uv absorbers . |
the structure of exotic hadrons , such as the tetraquarks , pentaquarks , and meson - baryon molecular states for instance , has been one of the most interesting topics over decades in terms of the strongly interacting systems , governed by quantum choromodynamics ( qcd ) .
recent discoveries of those exotics can shed light on the new understanding of qcd at low energies .
the mesons consisting of four quarks , i.e. , tetraquark state , has been reported by the belle collaboration and besiii collaboration @xcite .
the lhcb collaboration observed signals for the heavy pentaquark state @xmath12 as well @xcite .
the meson - baryon molecular state for @xmath0 , rather than a simple three - quark ( @xmath13 ) one , was proposed first even before qcd was established by dalitz and tuan @xcite and recently its properties have been investigated via the unitarized chiral dynamics @xcite and supported recently by the lattice - qcd ( lqcd ) simulation by investigating the strange form factor of @xmath10 @xcite .
the lattice - qcd ( lqcd ) simulation also supports the meson - baryon molecular nature by investigating the strange form factor of @xmath10 @xcite .
in addition to the studies of structure for the @xmath14 , the relevant production mechanisms were investigated extensively as well in refs .
@xcite . in the present work
, we would like to investigate the photoproduction of @xmath10 , i.e. , @xmath15 , employing the effective lagrangian approach with the @xmath16 and @xmath17 regge trajectories at tree level .
we focus on the contributions from the nucleon resonances near the threshold , such as @xmath18 and @xmath19 , which have been reported in the particle data group ( pdg ) @xcite .
in addition to them , a few missing resonances , i.e. , @xmath20 , @xmath21 , and @xmath22 , predicted by the relativistic su@xmath23 quark model @xcite , are also taken into account .
the electromagnetic and strong couplings are determined from the presently available theoretical and experimental results .
especially , we used the chiral unitary model ( chum ) for the strong couplings for @xmath10 @xcite , since we do not have much information for @xmath24 and @xmath25 from experiments .
the couplings for @xmath26 vertex are taken from the quark - model calculations @xcite . in order to satisfy the ward - takahashi ( wt ) identity ,
we make use of the gauge - invariant prescription for the form factors in the invariant amplitude . we compute various physical observables : total @xmath27 and differential @xmath28 cross sections , @xmath2-channel momentum transfer @xmath29 , photon - beam asymmetry @xmath30 , invariant mass plot for @xmath10 @xmath31 , and so on . from the numerical results , we observe that the nucleon - resonance contributions are crucial to reproduce the experimental data from the cebaf large acceptance spectrometer at jefferson laboratory ( clas / jlab )
@xcite for the total cross section near the threshold . among the resonances , we find that @xmath18 and @xmath19 dominates the threshold region . as for the differential cross section as a function of the outgoing @xmath32 angle ( @xmath33 ) in the center - of - mass ( cm ) frame , the nucleon resonances play an important role to produce the strength of the cross sections below @xmath34 gev as expected . as the production energy increases ,
the conventional nonresonant contributions dominate and enhance the forward peaking at @xmath35 , due to the strong @xmath16-exchange contribution .
the @xmath2-channel momentum transfer is computed for @xmath36 gev , with the help of the regge approach , which can extend a simple low - energy born approximation into the higher - energy beyond the resonance region @xcite . as expected ,
the curves are obviously modified by the @xmath37 contributions near threshold . as the same time
, the photon - beam asymmetry are computed as a function of @xmath38 for different energies , and in the shape of a distorted @xmath39 , according to the competing @xmath16 and @xmath17 contributions .
assuming that the decay width of @xmath10 is sufficiently narrow @xmath40 and the interference between the @xmath10 and other nonstrange mesons decaying into @xmath41 are negligible in the dalitz process of @xmath9 , the differential cross section of @xmath42 can be obtained from a simple formula with @xmath43 and @xmath44 , which is computed previously @xcite . by doing this , the distribution as a function of invariant mass @xmath45 is drawn and compared with the data , showing a good agreement and supporting its successful application .
finally , we investigate the constituent - counting rule ( ccr ) . the ccr is a method to analyze the internal structure of the hadrons involved in the @xmath46 reaction process by dimensional considerations of the reaction amplitude in terms of the quark and gluon propagators at the large angle as well as the high energy . applying this to the present reaction process , we observe that the numerical result , i.e. , @xmath47 as a function of @xmath48 , differs clearly from the three - quark state for @xmath10 , although it does not lead to the concrete conclusion that the results supports the five - quark state for @xmath10
. the present work is organized as follows : in section ii , theoretical framework is briefly explained .
numerical results and relevant discussions are given in section iii .
the final section is devoted for summary .
reaction process .
@xmath49 , @xmath50 , @xmath51 , and @xmath10 stand for the proton resonances , @xmath52 , @xmath53 , and @xmath14 , respectively . ] in this section , we provide a brief explanation for the present theoretical framework .
basically , we employ the tree - level born approximation with effective lagrangians for the interaction vertices and the regge trajectories for the pseudo - scalar ( ps ) and vector ( v ) meson exchanges in the @xmath2 channel . in terms of the ps meson - baryon coupling scheme ,
the relevant feynman diagrams for the @xmath54 reaction process are drawn in fig .
[ fig1 ] , in which @xmath55 and @xmath56 stand for the four momenta for the incident photon and target proton , whereas @xmath57 and @xmath58 for the outgoing @xmath32 and recoiled @xmath59 , respectively . as for the @xmath2-channel
, we consider the @xmath16 and @xmath17 exchanges with their _ regge _ propagators . by doing this
, we can explore higher energy regions , which can not be probed by the simple born approximation @xcite .
the nucleon and its resonance states are taken into account for the baryon - pole diagrams in the @xmath60 channel , whereas the hyperons , @xmath52 , @xmath53 , and @xmath14 , are included for the @xmath61 channel as shown in fig .
although there are other possible hyperon contributions in the @xmath61 channel , such as @xmath62 and @xmath63 for instance , the magnetic transitions to @xmath14 have not been reported experimentally as well as theoretically .
in addition , the @xmath61-channel resonances with a higher mass do not produce significant structure in the cross section in the energy region where we are interested in .
hence , we will not take those higher - mass hyperons into account in the present calculation .
the effective lagrangians for the em interaction vertices read : @xmath64 a^\mu , \cr \mathcal l_{\gamma k k^ * } & = & g_{\gamma k k^ * } \epsilon^{\mu\nu\alpha\beta } \partial_\mu a_\nu [ ( \partial_\alpha k_\beta^{*- } ) k^+ + k^- ( \partial_\alpha k_\beta^{*+ } ) ] , \cr \mathcal l_{\gamma nn } & = & - \bar n \left [ e_n \gamma_\mu - \frac{e\kappa_n}{2m_n } \sigma_{\mu\nu}\partial^\nu \right ] a^\mu n , \cr \mathcal l_{\gamma \lambda^ * \lambda^ * } & = & \frac{e\mu_{\lambda^*}}{2m_n } \bar \lambda^ * \sigma_{\mu\nu } \partial^\nu a^\mu \lambda^ * , \cr \mathcal l_{\gamma y \lambda^ * } & = & \frac{e\mu_{\lambda^ * \to y \gamma } } { 2m_n } \bar y \gamma_5 \sigma_{\mu\nu } \partial^\nu a^\mu \lambda^ * + \mathrm{h.c . } , \label{eq : bornlag1}\end{aligned}\ ] ] where @xmath65 , @xmath16 , @xmath17 , @xmath66 , and @xmath10 indicate the fields for the photon , pseudoscalar kaon , vector kaon , nucleon , and @xmath14 , respectively .
@xmath67 corresponds to the field for the ground - state @xmath50 or @xmath51 . @xmath68 and @xmath69 stand for the mass and electric charge of the hadron @xmath70 , while @xmath71 denotes the unit electric charge .
as for the values for the coupling constants , the charged @xmath72 is calculated from the experimental data for the decay width @xmath73 , resulting in 0.254@xmath74 @xcite .
the anomalous magnetic moment of the proton is given by @xmath75 .
the transition magnetic moments between two hyperons are also necessary and given by @xmath76 .
the su(3 ) quark model gives @xmath77 @xcite which is lied within the values predicted from the chum : @xmath78 @xcite .
meanwhile , we obtain @xmath79 from an isobar model @xcite to match the @xmath80 atom data @xcite . thus , the output of the radiative dacay widths is given by @xmath81 kev and @xmath82 kev @xcite from the formula @xmath83 derived from the @xmath84 lagrangian of eq .
( [ eq : bornlag1 ] ) . here *
k * is the magnitude of the three - momentum of the hyperon @xmath67 in the rest frame of @xmath10 .
the effective lagrangians for the strong vertices are written by : @xmath85 note that the strong coupling @xmath86 is given by @xmath87 from the nijmegen soft - core potential ( nsc97a ) @xcite . since there is no sufficient experimental information on the strong coupling constants for the excited @xmath50 hyperons , we resort to theoretical results , using the chiral - unitary model ( chum ) @xcite .
averaging those theoretical values for various cases , we determine the strengths for the couplings for the numerical calculations as @xmath88 and @xmath89 .
all the values for the relevant couplings for the numerical calculations are summarized in table [ tab0 ] .
.relevant em and strong coupling constants for the numerical calculations .
[ cols="^,^,^,^,^,^,^,^,^",options="header " , ] although we would like to reproduce the data in the relatively low - energy region near the threshold , as done in the clas experiment , it is interesting to explore the higher - energy region theoretically for future experiments in the upgraded clas / jlab and other experimental facilities . for this purpose
, we employ the @xmath2-channel regge trajectories for @xmath16 and @xmath17 and follow closely ref . @xcite . in this regge approach , the feynman propagators in eq .
( [ eq : borneachamp ] ) are replaced simply by the regge ones as @xcite : @xmath90 where we set the signature factors to be unity , assuming @xmath91 .
the regge trajectories read @xcite @xmath92 and we define the slope parameter as @xmath93 . conventionally , the energy scale parameter is chosen to be @xmath94 . with this in mind , the bare invariant amplitude of @xmath2- and @xmath60- channels in eq .
( [ eq : bornamp ] ) is modified as @xmath95 , \label{eq : bornamp2}\end{aligned}\ ] ] for the present reaction process .
it is worth mentioning that the three - body ( dalitz ) reaction process , @xmath96 , can be explored approximately in terms of the two - body one , if one assumes the followings : 1 ) the decay widths for the decaying resonances are sufficiently narrow and 2 ) the interference between the different resonances in the dalitz plot is negligible .
if these conditions are fulfilled , one can write the differential cross section ( invariant - mass plot ) for the dalitz process @xmath97 as follows @xcite : @xmath98 where @xmath44 is the two - body total cross section
. here are some justifications for the usage of eq .
( [ eq : app2 ] ) : first , the decay width of @xmath99 is about @xmath100 mev , which is much smaller than the typical energy range @xmath101 mev in the present analysis .
therefore , the condition 1 ) can be assumed to be reasonable .
second , in the dalitz process @xmath9 , the @xmath10 production interferes with the @xmath102 one on the dalitz plot .
interestingly , in ref .
@xcite , it was found experimentally that the interference between the different resonant productions is almost negligible , although they focused on the different reactions process , i.e. , @xmath103 .
hence , considering these observations , the condition 1 ) and 2 ) can be justified rather safely here , and we can use eq .
( [ eq : app2 ] ) for computing the dalitz process with the two - body cross section , which is computed by eqs .
( [ eq : borneachamp ] ) and ( [ eq : reseachamp ] ) . in doing this
, the absolute value for the @xmath32 three momentum is obtained as a function of the invariant mass @xmath45 as follows : @xmath104^\frac{1}{2}.\ ] ] note that @xmath105 by construction in the cm frame .
as mentioned previously , the internal structure of @xmath10 has been one of the most interesting subjects in the nonperturbative qcd .
there have been several approaches to pin down the genuine structure of the hyperon resonance . among various theoretical studies ,
we discuss the constituent - counting rule ( ccr ) @xcite . basically , in the ccr
, the high - energy and large - angle scattering amplitudes for two - body processes , i.e. , @xmath106 , are analyzed by dimensional considerations of the quark and gluon propagators , resulting in @xmath107 where @xmath108 is the total number of the constituents of the particles involved in the scattering process .
for instance , if @xmath10 is composed of three quarks , the value of @xmath108 becomes @xmath109 , whereas @xmath110 for the five - quark system .
note that there are many uncertain theoretical ingredients , such as the distribution functions for the involved particles and so on , to apply the ccr to real problems of @xmath10 as shown in ref .
it is , however , still valuable to test the relation in eq .
( [ eq : ccr ] ) with the present results , in which the regge approach can extend the low - energy born approximation to certain high energies beyond the resonance region .
in this section , we present the numerical results with corresponding discussions .
note that , in the present work , the free parameters to be determined by a fitting process are the cutoff mass @xmath50 in eq .
( [ eq : ff ] ) and the phase angle @xmath111 in eq .
( [ eq : resamp ] ) . to avoid theoretical uncertainties and complexities , the phase angle in eq .
( [ eq : resamp ] ) is chosen to be the same for all the nucleon resonances as @xmath112 .
the cutoff masses are given as @xmath113 gev by fitting processes with the experimental data . although the full - decay widths for the nucleon resonance , which appear in the denominator of the amplitudes as shown in eq .
( [ eq : reseachamp ] ) , are not free parameters as shown in the pdg list @xcite , we fix it to be @xmath114 mev for all the resonances for brevity . before performing the detailed calculations for reproducing the clas
/ jlab experimental data , we would like to examine the effects of the nucleon - resonance contributions in the @xmath10 photoproduction in a model - independent manner .
we consider a single resonance near the reaction threshold with different spin and parity , i.e. @xmath115 for a positive integer @xmath116 . for simplicity , we choose @xmath117 mev and @xmath118 for all the resonances .
the reasons for those choices are as follows : 1 ) in the pdg list , the reported values for @xmath119s reside in @xmath120 mev with sizable uncertainties .
hence , the sort of a middle value is chosen .
2 ) since we focus on the vicinity of the threshold , the @xmath121 contribution @xmath122 turns out to be small .
therefore , we only consider the @xmath123 contribution and the values of @xmath124 are in the order of @xmath125 as shown in the pdg list . thus , similarly , the middle value is employed .
six resonances are taken into account in total , depending on the spins and parities , and @xmath126 is fitted with the total cross section from the clas / jlab data as shown in fig .
[ fign ] with the fixed @xmath123 values via the above @xmath124 . as a result ,
the branching ratios of @xmath127 are given as follows : @xmath128 as the spin of @xmath37 increases , the shapes of the angular distributions exhibit more fluctuations on their behaviors .
we find that the nucleon resonances with @xmath129= @xmath130 , @xmath131 , and @xmath132 more or less describe the clas data within reasonable values of branching ratios , where @xmath133 and the @xmath134 is calculated from the @xmath135 for each cases .
but when more than a single resonance is considered simultaneously , we reach a different conclusion as we will see soon . for the six different nucleon resonances with @xmath136 gev.,width=604
] now , let us show numerical results when employing our model . in fig .
[ fig12 ] , we demonstrate the results for the total cross section for @xmath1 as a function of the photon laboratory energy @xmath137 .
the experimental data are taken from the clas / jlab collaboration @xcite . in the left panel of the figure ,
the full result including the @xmath37 ( dot - dashed ) and non-@xmath37 ( dashed ) contributions is given in the solid line .
it is worth mentioning that the peak - like bump around @xmath138 gev is well reproduced by the nucleon resonances , whereas the non-@xmath37 contribution curve decreases rapidly , due to the regge propagators , as the energy increases .
we also verified that the non-@xmath37 contributions can not reproduce the bump by changing form - factor types or their cutoff parameters .
moreover , by comparing the strengths of the @xmath37 and total contributions , one can conclude that the constructive interference between the @xmath37 and others plays an important role . as a consequence , within the most conventional tree - level approach like the present effective model , the nucleon resonances are crucial to explain the obvious cross - section enhancement near the threshold region for @xmath14 photoproduction .
it is also interesting to see the separate effects from the different nucleon resonances . in the right panel of fig .
[ fig12 ] , we show those curves for the total cross sections . it is clearly shown that @xmath139 and @xmath140 give dominant effects on the cross section , due to their relatively larger strong and em couplings than other resonances .
in contrast , the missing resonances play only minor roles in producing the strength as shown there .
it is worth mentioning that this strong enhancement near the threshold region was already reported in the leps experiment @xcite .
the curve shape beyond the resonance region is reproduced mainly by the @xmath2-channel @xmath16-exchange contribution , in addition to the small but finite @xmath17-exchange one .
the numerical results for the angular dependence @xmath141 are shown in fig .
[ fig3 ] for different cm energies @xmath142 gev . here
, @xmath33 denotes the angle between the incident photon beam and the outgoing kaon in the cm frame . in overall ,
the full results ( solid ) show qualitatively good agreement with the data .
we also observe that the forward - scattering enhancement and becomes more obvious as the energy increases , due to the strong @xmath2-channel @xmath16 exchange . at @xmath143 gev
, it is clearly shown that the data are reproduced by the @xmath37 and non-@xmath37 contributions constructively .
this observation agrees well with that shown in the total cross section .
as the energy gets higher , the @xmath37 contribution becomes diminished and almost flat as expected . at the same time , the small but sizable backward - scattering enhancement starts appearing gradually beyond @xmath144 gev .
we verified that it originates mainly from the larger couplings in case of @xmath50 and @xmath51 exchanges , i.e. , @xmath86 and @xmath145 , in the @xmath61 channel than other hyperons , as shown in table [ tab0 ] .
note that , quantitatively , in the low - energy region @xmath146 gev , we find sizable disagreement with the data for @xmath147 .
although we anticipate that the unknown contributions beyond the present model setup can resolve it , we will not make further detailed investigation on this issue in the present work and would like to leave it for the future .
it should be mentioned that , even without our regge approach , i.e. , with the formula of eq . (
[ eq : bornamp ] ) rather than of eq .
( [ eq : bornamp2 ] ) , we obtain similar results for total and differential cross sections in the range of @xmath148 gev .
this is expected since our hybridized regge model interpolates between low and high energy regions smoothly . in fig .
[ fig4 ] , we show the numerical results for the @xmath2-channel momentum transfer @xmath149 as a function of @xmath150 for different cm energies @xmath151 gev . in the left panel ,
the curves are computed without the @xmath37 contributions , whereas the right one with them .
we note that the curves show a typical behavior of the momentum transfer .
however , due to the @xmath37 contributions , which are effective for @xmath152 gev , the momentum - transfer curves exhibit the apparent differences between the two cases in the low - energy region .
as expected , the both curves become almost identical as the energy increases .
the photon - beam asymmetry , i.e. , the analyzing power is one of the important physical quantities , observed in hadron photoproductions .
we define it as follows : @xmath153 where the subscripts @xmath154 and @xmath155 stand for the cases , when the photon polarization vector is parallel and perpendicular to the reaction plane , respectively . according to this definition in eq .
( [ eq : ba ] ) , the electric photo - coupling interaction , i.e. , the @xmath16 exchange in the @xmath2 channel gives @xmath156 , whereas the magnetic one , i.e. , the @xmath17 exchange does @xmath157 . in fig .
[ fig5 ] , we present the numerical results for it in the same manner with fig .
[ fig4 ] for different cm energies @xmath158 gev . without the @xmath37 contributions
as shown in the left panel , the beam asymmetry becomes reversed around @xmath35 , due to the competing @xmath16 and @xmath17 contributions . by including @xmath37s , the beam - asymmetry curves are changed slightly .
note that the nucleon resonances make the beam asymmetry positive in general , because of their magnetic photon couplings . in fig .
[ fig6 ] , we plot the invariant - mass distribution @xmath159 as a function of the invariant mass @xmath45 for different cm energies @xmath158 gev , using eq .
( [ eq : app2 ] ) and the numerical results for the two - body process , i.e. , @xmath1 , given in fig .
[ fig12 ] .
as shown in the figure , the theoretical curves describe the clas data @xcite qualitatively well for all the energies , as expected from eq .
( [ eq : app2 ] ) by construction , due to the appropriate breit - wigner distribution and the strength from the two - body total cross section . from this observation
, we find that the interference between the @xmath10 and other vector - meson productions , such as @xmath17 , on the dalitz plot is practically small as shown in ref .
@xcite . in the low - energy region for @xmath160 gev , being similar to the total and differential cross sections as shown in figs .
[ fig12 ] and [ fig3 ] , the nucleon resonances make considerable contributions to the invariant - mass distribution as well , as expected .
we also observe that the nucleon resonance contributions becomes almost negligible beyond @xmath161 gev .
finally , we would like to test the ccr for the present reaction process .
for this purpose , we take @xmath162 in the cm frame as done in refs .
@xcite and the numerical results are given in fig .
[ fig7 ] , where we draw @xmath47 as a function of @xmath163 .
the solid curve stands for the result with the nucleon resonances .
following the ccr , if @xmath10 consists of three quarks as usual baryons , the resulting curve becomes flat ( dotted ) , whereas it decreases with a tendency @xmath164 for the five - quark @xmath10 ( dashed ) and @xmath165 for the seven - quark @xmath10 ( dot - dashed ) .
here , we assume that the resonance region is terminated at @xmath166 gev as observed in the invariant - mass plot in fig .
[ fig6 ] . as shown in the figure
, the curves obviously exhibit a decreasing behavior with respect to @xmath167 , and the slope matches with @xmath165 , i.e. , the seven - quark system for @xmath10 . in ref .
@xcite , the clas / jlab experimental data was fitted with @xmath168 as a function of @xmath167 , where @xmath169 denotes a scattering - angle dependent function , and it showed that @xmath170 , depending on energies , indicating that @xmath10 is not a simple @xmath13-quark state .
although the seven - quark state looks supported from the numerical results , it is still difficult to pin down the genuine structure of @xmath10 from this observation , since we need to take into account more realistic information on the structure functions , energy - angular dependences , and so on in the ccr as well as in our model .
hence , we can conclude that , more or less , @xmath10 is possibly distinctive from the simple @xmath13-quark state , as far as we can reproduce the experimental data qualitatively well with the conventionally accepted model as in the present work .
we have investigated the photoproduction of @xmath0 , i.e. , @xmath1 , by employing the effective lagrangian with the reggized @xmath2-channel @xmath16 and @xmath17 exchanges at tree level .
in addition to the ground state hadrons involved in the reactions process , we took five nucleon resonances near the reaction threshold into account .
we performed numerical calculations for the energy and angular dependences of the cross section as well as the polarization observable , such as the photon beam asymmetry .
the invariant - mass distribution for @xmath9 was extracted from the two - body cross section with some reasonable assumptions .
finally , the internal structure of @xmath10 was explored by the constituent - counting rule ( ccr ) , which provides the structure information of the hadrons in a reaction process in terms of the dimensional analyses of the scattering amplitudes at a large angle in high energies .
relevant results obtained in the present study are summarized as follows : * the nucleon resonances play an important role to reproduce the recent clas experimental data @xcite near the threshold region .
especially , the so - called pdg resonances , @xmath18 and @xmath19 , provide considerable contributions to the total cross section @xmath44 , due to their relatively larger strong and em couplings than other resonances , as far as we resort to available experimental and theoretical information to determine the couplings . *
as expected , angular dependence , such as @xmath171 , is affected as well by the nucleon - resonance contributions in the low - energy region , whereas they get diminished as the energy increases . in overall ,
we observe strong forward - scattering enhancement , due to the @xmath2-channel @xmath16 exchange .
* the numerical results for the @xmath2-channel momentum transfer @xmath172 and photon - beam asymmetry @xmath173 are also given .
we find that those results can be explained in general by the nucleon resonances and the reggized @xmath2-channel @xmath174 exchanges in the low- and high - energy regions , respectively . * assuming some reasonable conditions , i.e. , no interferences between the hyperon and vector - meson resonances on the dalitz plot for instance , we draw the invariant - mass plot for @xmath42 in terms of the two - body cross section @xmath44 and the partial decay width @xmath43 . qualitatively , the experimental data of clas collaboration are reproduced well for the considered energy regions . from this observation , the interference between the @xmath10 and @xmath17 productions on the dalitz plot must be practically small and it is consistent to the result of ref . @xcite . * finally , we examine the internal structure of @xmath10 via the ccr . the calculated curve of @xmath175 as a function of the cm energy at @xmath176 shows a strong peak - like structure below @xmath166 gev and a rapid decrease beyond it , due to the nucleon resonances and the reggized @xmath2-channel contributions , respectively . using the most simple parameterization of the scattering amplitude via the ccr ,
the slope of the calculated curve beyond the resonance region matches roughly with that for @xmath177 , which tells that @xmath10 is a seven - quark state
. however , since there are considerable theoretical uncertainties in this analysis , we can conclude safely that @xmath10 is possibly different from the usual @xmath13-quark state as reported in ref .
@xcite , if we take out results conservatively . as a summary ,
the present effective model approach with the reggized @xmath2-channel exchanges explains the @xmath10 photoproduction qualitatively well , showing good agreements with the clas data .
as discussed above , however , it is still difficult to determine the genuine internal structure of @xmath10 within this simple model .
nonetheless , we have found a signal that @xmath10 can be distinctive from usual three - quark baryons .
more realistic studies with sophisticated form factors and regge treatments are under progress and will appear elsewhere .
s.h.k . acknowledges support from the young scientist training program at the asia pacific center for theoretical physics by the korea ministry of education , science , and technology , gyeongsangbuk - do and pohang city . the work of s.i.n .
was supported by the research grant of pukyong national university .
the work of d.j . was partly supported by a grant - in - aid for scientific research from jsps ( 25400254 ) .
h.ch.k acknowledges the support by the national research foundation of korea ( nrf ) grant funded by the korean government ( msip ) ( no .
nrf-2015r1a2a2a04007048 ) . s.i.n .
also thanks j. k. ahn , a. hosaka , and h. kohri for fruitful discussions .
from the effective lagrangians of eq .
( [ eq : reslag2 ] ) , the partial decay width of nucleon resonance @xmath178 into @xmath179 can be calculated as @xmath180 = & \frac{1}{4\pi } \frac{|{\bf q}|}{m_{n^ * } } g_{k \lambda^ * n^*}^2 ( e_{\lambda^ * } \pm m_{\lambda^ * } ) , \cr \gamma\left[n^*(3/2^\pm ) \to k\lambda^*\right]= & \frac{1}{12\pi } \frac{|{\bf q}|^3}{m_{n^ * } } \frac{g_{k \lambda^ * n^*}^2}{m_k^2 } ( e_{\lambda^ * } \mp m_{\lambda^ * } ) , \cr \gamma\left[n^*(5/2^\pm ) \to k\lambda^*\right ] = & \frac{1}{30\pi } \frac{|{\bf q}|^5}{m_{n^ * } } \frac{g_{k \lambda^ * n^*}^2}{m_k^4 } ( e_{\lambda^ * } \pm m_{\lambda^ * } ) .
\label{eq : ampl1}\end{aligned}\ ] ] here , the magnitude of the three - momentum and the energy for the @xmath16 in the rest frame of the nucleon resonance reads : @xmath181[m_{n^*}^2 - ( m_{\lambda^ * } - m_k)^2 ] } , \,\,\,\,e_{\lambda^ * } = \sqrt{m_{\lambda^*}^2 + |{\bf q}|^2}. \label{eq:3-monen}\end{aligned}\ ] ] we need to know how the effective lagrangians are related to the decay amplitudes to obtain the coupling constants .
the decay amplitude for @xmath182 can be expressed as follows @xcite : @xmath183 where @xmath184 and @xmath185 are the clebsch - gordan coefficient and spherical harmonics , respectively .
the relation between the partial wave decay amplitude @xmath186 and the decay width @xmath187 can be derived as @xmath188 the spin and parity of the nucleon resonance places constraints on the relative orbital angular momentum @xmath189 of the final state .
let us first consider the case of a @xmath190 resonance .
the relative orbital angular momentum is constrained by angular momentum and parity conservation such that @xmath191 is possible , i.e. , only @xmath60 wave is allowed . in a similar way , for the decays of @xmath192 resonances into @xmath193 ,
the final state is in the relative @xmath194 wave , and for the resonances of @xmath195 in the relative @xmath196 wave .
finally , the @xmath197 resonance is allowed only @xmath198 wave . keeping this in mind ,
we can obtain the decay amplitudes in terms of the coupling constants for the decays of @xmath199 nucleon resonances into @xmath200 final state : @xmath201 99 s. k. choi _ et al .
_ [ belle collaboration ] , phys . rev .
lett . * 91 * , 262001 ( 2003 ) .
k. abe _ et al . _ [ belle collaboration ] , phys .
lett . * 98 * , 082001 ( 2007 ) .
s. k. choi _ et al . _ [ belle collaboration ] , phys .
lett . * 100 * , 142001 ( 2008 ) .
a. bondar _ et al .
_ [ belle collaboration ] , phys .
108 * , 122001 ( 2012 ) .
m. ablikim _ et al . _
[ besiii collaboration ] , phys .
* 110 * , 252001 ( 2013 ) .
z. q. liu _ et al .
_ [ belle collaboration ] , phys .
lett . * 110 * , 252002 ( 2013 ) .
m. ablikim _ et al . _
[ besiii collaboration ] , phys .
111 * , 242001 ( 2013 ) .
r. aaij _ et al . _
[ lhcb collaboration ] , phys . rev .
* 115 * , 072001 ( 2015 ) .
dalitz and s.f .
tuan , phys .
* 2 * , 425 ( 1959 ) .
dalitz and s.f .
tuan , ann .
* 10 * , 307 ( 1960 ) .
n. kaiser , p.b .
siegel , and w. weise , nucl .
* a594 * , 325 ( 1995 ) .
e. oset and a. ramos , nucl .
a * 635 * , 99 ( 1998 ) .
j. c. nacher , e. oset , h. toki , and a. ramos , phys .
b * 455 * , 55 ( 1999 ) .
j. a. oller and ulf g. meissner , phys . lett . *
b500 * , 263 ( 2001 ) .
d. jido , e. oset , and a. ramos , phys .
c * 66 * , 055203 ( 2002 ) .
d. jido , j. a. oller , e. oset , a. ramos , and u. g. meissner , nucl .
a * 725 * , 181 ( 2003 ) .
s. i. nam , h. -ch .
kim , t. hyodo , d. jido , and a. hosaka , j. korean phys .
* 45 * , 1466 ( 2004 ) .
m. f. m. lutz and m. soyeur , nucl .
a * 748 * , 499 ( 2005 ) .
v. k. magas , e. oset , and a. ramos , phys .
lett . * 95 * , 052301 ( 2005 ) .
t. hyodo and d. jido , prog .
phys . * 67 * , 55 ( 2012 ) .
m. mai and u. g. meiner , eur .
phys . j. a * 51 * , 30 ( 2015 ) .
y. nemoto , n. nakajima , h. matsufuru , and h. suganuma , phys .
d * 68 * , 094505 ( 2003 ) . s. i. nam , j. h. park , a. hosaka , and h. -ch .
kim , j. korean phys .
* 59 * , 2676 ( 2011 ) .
s. x. nakamura and d. jido , ptep * 2014 * , 023d01 ( 2014 ) .
s. i. nam and h. k. jo , arxiv:1503.00419 [ hep - ph ] .
e. wang , j. j. xie , w. h. liang , f. k. guo , and e. oset , phys .
c * 95 * , 015205 ( 2017 ) . c. patrignani
_ et al . _
[ particle data group ] , chin .
c * 40 * , 100001 ( 2016 ) .
s. capstick , phys .
d * 46 * , 2864 ( 1992 ) . s. capstick and w. roberts , phys .
d * 58 * , 074011 ( 1998 ) .
k. p. khemchandani , a. martinez torres , h. kaneko , h. nagahiro , and a. hosaka , phys .
d * 84 * , 094018 ( 2011 ) .
k. moriya _ et al . _ [ clas collaboration ] , phys .
c * 88 * , 045201 ( 2013 ) [ addendum - ibid .
c * 88 * , 049902 ( 2013 ) ] .
a. donnachie _ et al .
_ , pomeron physics and qcd ( cambridge university press , uk , 2002 ) .
r. a. williams , c. r. ji , and s. r. cotanch , phys .
c * 43 * , 452 ( 1991 ) .
d. jido , a. hosaka , j. c. nacher , e. oset , and a. ramos , phys .
c * 66 * , 025203 ( 2002 ) .
h. burkhardt and j. lowe , phys .
c * 44 * , 607 ( 1991 ) .
d. a. whitehouse , phys .
lett . * 63 * , 1352 ( 1989 ) . v. g. j. stoks and th .
a. rijken , phys .
c * 59 * , 3009 ( 1999 ) .
k. ohta , phys .
c * 40 * , 1335 ( 1989 )
. h. haberzettl , phys .
c * 56 * , 2041 ( 1997 ) .
h. haberzettl , c. bennhold , t. mart , and t. feuster , phys .
c * 58 * , r40 ( 1998 ) .
r. m. davidson and r. workman , phys .
c * 63 * , 025210 ( 2001 ) .
y. oh , c. m. ko and , k. nakayama , phys .
c * 77 * , 045204 ( 2008 ) .
y. oh , j. korean phys .
* 59 * , 3344 ( 2011 ) .
f. a. berends , j. w. van holten , p. van nieuwenhuizen , and b. de wit , nucl .
b * 154 * , 261 ( 1979 ) .
s. h. kim , s. i. nam , y. oh , and h. -ch .
kim , phys .
d * 84 * , 114023 ( 2011 ) .
s. h. kim , s. i. nam , a. hosaka , and h. -ch .
kim , phys . rev .
d * 88 * , 054012 ( 2013 ) .
m. guidal , j. m. laget , and m. vanderhaeghen , nucl .
a * 627 * , 645 ( 1997 ) .
s. y. ryu _ et al . _
[ leps collaboration ] , phys .
lett . * 116 * , 232001 ( 2016 ) .
h. kawamura , s. kumano and t. sekihara , phys .
d * 88 * , 034010 ( 2013 ) .
w. c. chang , s. kumano , and t. sekihara , phys .
d * 93 * , 034006 ( 2016 ) .
m. niiyama _ et al .
_ , phys .
c * 78 * , 035202 ( 2008 ) .
k. moriya _ et al . _ [ clas collaboration ] , phys .
c * 87 * , 035206 ( 2013 ) . as a function of @xmath202 .
the solid , dashed , and dot - dashed lines indicate the total , non-@xmath37 , and @xmath37 contributions .
right : separate contributions from the nucleon resonances .
the experimental data are taken from ref .
@xcite.,title="fig:",width=321 ] as a function of @xmath202 .
the solid , dashed , and dot - dashed lines indicate the total , non-@xmath37 , and @xmath37 contributions .
right : separate contributions from the nucleon resonances . the experimental data are taken from ref .
@xcite.,title="fig:",width=321 ] for @xmath54 as a functions @xmath38 for different cm energies , i.e. , @xmath158 gev . the experimental data are taken from ref .
the legends are the same with those for the left panel of fig .
[ fig12].,width=604 ] curves without the @xmath37 contributions as a function of @xmath38 for different cm energies @xmath158 gev . right : those with the @xmath37 ones.,title="fig:",width=321 ] curves without the @xmath37 contributions as a function of @xmath38 for different cm energies @xmath158 gev .
right : those with the @xmath37 ones.,title="fig:",width=321 ] curves without the @xmath37 contributions as a function of @xmath38 for different cm energies @xmath158 gev .
right : those with the @xmath37 ones.,title="fig:",width=321 ] curves without the @xmath37 contributions as a function of @xmath38 for different cm energies @xmath158 gev .
right : those with the @xmath37 ones.,title="fig:",width=321 ] as a function of the invariant mass @xmath45 for different cm energies @xmath158 gev .
the experimental data are taken from ref .
the legends are the same with those for the left panel of fig .
[ fig3].,width=604 ] | we investigate the photoproduction of @xmath0 hyperon resonance , i.e. , @xmath1 , employing the effective lagrangian approach with the @xmath2-channel regge trajectories at tree level .
we extensively explore the effects from the nucleon resonances in the vicinity of the threshold @xmath3 mev , i.e. , @xmath4 , @xmath5 , @xmath6 , @xmath7 , and @xmath8 , and observe that they are of great importance to reproduce the recent clas / jlab experimental data .
total and differential cross sections are given as numerical results and compared with the experimental data , in addition to the @xmath2-channel momentum transfer and photon - beam asymmetry .
the invariant - mass plots for @xmath9 via @xmath10 are also extracted from the two - body process results , showing a qualitative agreement with the data .
we also discuss the constituent - counting rule ( ccr ) for the internal structure of @xmath10 , resulting in that @xmath10 appears to be different from a simple three - quark @xmath11 state . |
transport properties in one - dimensional interacting systems such as quantum wires or metallic carbon nanotubes have been widely studied both theoretically and experimentally . when the wire or nanotube is well connected to a metallic contact , and another tunnel contact
is placed on top of it , the current exhibits a zero bias anomaly @xcite because of the coulomb interaction effects in this one dimensional system . on the contrary an isolated wire with bad contacts displays coulomb blockade effects @xcite .
another interesting tool to characterize the transport is the noise ( current - current fluctuations ) .
it allows for instance to obtain information about the charge of the carriers flowing through the conductor .
when electrons are injected in an _ infinite length _ carbon nanotube from a scanning tunneling microscope ( stm ) tip , it has been shown@xcite that the fano factor the ratio between the zero - frequency noise and the current is equal to @xmath0 , where @xmath1 is the coulomb interaction strength .
this fano factor is tied to the charge of the collective excitations which are present within the wire . however , the presence of adiabatic fermi liquid contacts at the extremities of the nanotube modifies@xcite the fano factor and effectively erases the information about the coulomb interactions . to recover information about coulomb interactions ,
finite - frequency noise is needed .
recent experiments@xcite have shown that the current - current correlator which can be detected is in general related to the non - symmetrized noise rather than the symmetrized one .
calculations of the finite - frequency non - symmetrized noise in the stm+nanotube system showed oscillations due to reflections at the contacts@xcite whose period is related to the anomalous charges . in the above calculations , electrostatic effects such as screening between the wire and tip
were typically left out .
screening from the stm tip is likely to affect locally the electronic properties of the wire .
such effects were explicitly computed in ref . , where emphasis was put on the spectral properties of the wire .
there , we developed general expressions for the dyson equations , which allowed to derive non perturbatively the greens functions of the bosonic fields which are needed to obtain the spectral function of the wire .
the tunneling density of states was shown to be enhanced ( reduced ) for large ( weak ) coulomb interaction @xcite . while the spectral function , the tunneling and the local density of states are quantities of great interest in condensed matter physics problems , they are typically measured in the context of electron transport , by tunneling experiments . in ref .
for instance , the tunneling current between two luttinger liquids is typically expressed as a convolution of the two tunneling density of states of such materials . in the present paper , we wish for the first time to address transport and screening effects on the same footing .
we argue that because the interaction between the tip and wire mixes the fermionic degrees of freedom of both source and drain , the derivation of the tunneling current and noise has to be revisited .
as we shall see , the tunneling current depends on a new class of greens function which we call `` mixed '' green s functions .
the use of the bosonization methods allows to compute the tunneling current , as usual , to second order in the tunneling amplitude , albeit `` non perturbatively '' with respect to the coulomb interaction between the wire and tip .
this work thus suggests plainly that in the presence of screening effect , special care must be taken in the derivation of transport properties .
a generalization of our results to carbon nanotubes can be straightforwardly obtained .
the structure of the paper is the following : in sec .
ii , we present the system and the hamiltonian . in sec .
iii , we give the definitions of the tunnel current and of the non - symmetrized noise and their general expressions in term of bosonic green s functions . in sec .
iv , we calculate the mixed green s functions .
next , in sec .
v , we consider a particular form of the screening potential and we calculate the transport properties associated to an infinite length wire . we conclude in sec . vi .
the total hamiltonian we consider is @xmath2 where @xmath3 is the unperturbed hamiltonian and @xmath4 is the perturbation due to the voltage applied to the tip ( see fig . [ system ] ) .
the unperturbed hamiltonian is given by @xmath5 , where @xmath6 describes the quantum wire : @xmath7~ , \label{h_w}\end{aligned}\ ] ] with @xmath8 and @xmath9 are the non - chiral bosonic fields , @xmath10 is the repulsive coulomb interactions parameter in the charge sector , and @xmath11 is the coulomb interactions parameter in the spin sector .
@xmath12 is the velocity of the excitations in each sector @xmath13 .
we put @xmath14 in the whole paper , excepted in voltage dependent terms and in figure captions .
@xmath15 describes the stm tip : @xmath16 with @xmath17 , the chiral bosonic field of the stm tip and @xmath18 is the fermi velocity of electrons in the tip .
@xmath19 is the index which refers to the spin degree of freedom .
the @xmath20 and @xmath21 axes are defined in fig .
[ system ] . and , @xmath22 describes the screening : @xmath23 where @xmath24 is the electrostatic potential caused by the proximity between the stm tip and the quantum wire .
@xmath25 is the density operator associated to the wire and @xmath26 , the density operator associated to the tip : @xmath27 when @xmath28 , we have shown in ref . that the @xmath29contribution to the density operator is not a relevant contribution to the screening hamiltonian .
we neglect it in this work .
is the applied voltage to the tip and @xmath30 is the tunnel amplitude.[system],width=188 ] the perturbation here is caused by a voltage @xmath31 applied to the tip : electrons can by injected from the tip to the wire or , inversely from the wire to the tip , by tunnel effect : @xmath32~,\nonumber\\\end{aligned}\ ] ] where @xmath33 is the tunnel amplitude and the chirality @xmath34 refers to the right movers ( @xmath35 ) or to the left movers ( @xmath36 ) .
we assume that the electrons are injected in the wire at the position of the tip , i.e. at position @xmath37 .
@xmath38 and @xmath39 are the creation and annihilation operators attached to the tip : @xmath40 where @xmath41 is a majorana fermion associated to the tip , and @xmath42 is the ultraviolet cut - off . @xmath43 and @xmath44 are the creation and annihilation operators attached to the chirality @xmath34 and to the spin @xmath19 of the electrons in the wire : @xmath45 where @xmath46 is the majorana fermion associated to the wire , @xmath47 is the fermi wave vector .
the coefficient @xmath48 is equal to @xmath49 when @xmath50 and to @xmath19 when @xmath51 .
in this section , we give the general expressions of the tunnel current and the non - symmetrized noise as a function of the bosonic green s functions in the presence of screening .
the average tunnel current is defined as : @xmath52 where @xmath53 denotes time ordering along the keldysh contour@xcite , @xmath54 indexes the position of time on it , and @xmath55 .
the tunnel current operator is defined by : @xmath56 up to the second order of the perturbative calculation with the tunnel amplitude @xmath57 , the average tunnel current can be expressed as @xmath58 where @xmath59 .
@xmath60 is the keldysh green s function of the tip in the presence of the electrostatic coupling defined as : @xmath61 similary , @xmath62 , @xmath63,@xmath64 and @xmath65 are the keldysh green s functions of the wire in the presence of the electrostatic coupling . and finally , @xmath66 , @xmath67,@xmath68 and @xmath69 are the mixed keldysh green s functions which are non - zero only in the presence of screening .
( [ tunnel_current ] ) differs to what one obtains in the absence of screening , in the sense that mixed green s functions are not present in this case@xcite .
we call them mixed green s functions , because they mix the bosonic fields @xmath70 or @xmath71 of the wire to the bosonic field @xmath72 of the tip .
the appearance of these new correlators is due to the electrostatic coupling between the wire and the tip , and they will affect the transport properties . notice that these mixed green s functions do not have any influence on the spectral properties of the wire because in the definitions of the spectral function and of the density of states , it only appears the bosonic fields of the wire@xcite . on the contrary , in the transport properties ,
the bosonic fields of the wire and of the tip are mixed due to the term @xmath4 in the hamiltonian , which corresponds to a tunnel transfer of electrons between the two sub - systems .
the calculation of the non perturbative green s functions associated to @xmath3 : @xmath73 , @xmath74 , @xmath75,@xmath76 , and @xmath77 is presented in detail in ref . , and the calculation of the mixed green s functions is done in sec .
we define the non - symmetrized noise as @xmath78 up to the second order with the tunnel amplitude @xmath57 , the non - symmetrized noise reads @xmath79 since we consider an applied voltage constant in time , the noise depends only on the difference @xmath80 : @xmath81 , we can thus define the following fourier transform for the noise : @xmath82 in what follows , we will study both the zero - frequency noise @xmath83 and the finite - frequency noise @xmath84 .
however , we need before to determine the dyson equations of the mixed green s functions which appear in the expressions of the current and the noise .
the first step is to write the partition function in the matsubara formalism@xcite : @xmath85&=&\int { \mathcal d}\varphi_{\uparrow}{\mathcal d}\varphi_{\downarrow}{\mathcal d}\phi_c{\mathcal d}\phi_s{\mathcal d}\theta_c{\mathcal d}\theta_s e^{-\int d\tau ( l_0-l_{aux})}~,\nonumber\\\end{aligned}\ ] ] where @xmath86 is the lagrangian associated to @xmath3 and @xmath87 contains auxiliary fields @xmath88 , @xmath89 and @xmath90 needed to extract the green s functions , such as for example @xmath91 , through the relation : @xmath92 where @xmath93 is the time ordering operator . as the hamiltonian @xmath3 is quadratic with the bosonic fields @xmath8 , @xmath9 and @xmath17 , it is possible to calculate exactly the full green s function by integrating out the degrees of freedom of the tip .
we obtain @xmath94 where @xmath95 , and @xmath96 the green s function of the wire without screening .
the derivation of the partition function with other auxiliary fields gives access , in a similar way , to the dyson equations verified by the other mixed correlators : @xmath97 @xmath98 and , @xmath99 in the next section , we present the application of these results to an infinite length wire and we make the approximation , ( which is justified in ref . ) that the double derivative of the electrostatic potential is represented by a local interaction .
for the same reasons that are given in ref . , we assume an electrostatic potential of the form @xmath101 with this specific form of the electrostatic potential and for an infinite length wire , we are able to solve the dyson equations verified by the green s functions associated to @xmath3 . the keldysh green s function associated to the bosonic field @xmath72 of the tip that we need in eqs .
( [ tunnel_current ] ) and ( [ tunnel_noise ] ) reads@xcite : @xmath102~,\nonumber\\\end{aligned}\ ] ] where @xmath103 is the frequency cut - off , @xmath104 is the screening frequency , @xmath105 is the exponential integral function , and @xmath106 is the bare green s function of the tip in the absence of screening .
notice that eq .
( [ green1 ] ) was derived exactly since no restriction has been imposed to the strength of the screening potential .
the only assumption we have made is to take @xmath28 in order to neglect the @xmath29-contribution in @xmath107 .
the keldysh green s function attached to the bosonic field @xmath70 of the wire is : @xmath108~.\nonumber\\\end{aligned}\ ] ] for the spin sector ( @xmath51 ) , the green s function is not affected by the screening : @xmath109 .
the other keldysh green s functions associated to the wire are aslo unchanged : @xmath110 the mixed green s functions @xmath111 and @xmath112 are calculated in appendix a , we obtain : @xmath113~.\end{aligned}\ ] ] the two other mixed green s functions , which appear in eqs .
( [ tunnel_current ] ) and ( [ tunnel_noise ] ) , are equal to zero even in the presence of screening : @xmath114 because the hamiltonian @xmath22 couples only the @xmath70 and @xmath72 fields .
now , we have all the ingredients to calculate the transport properties of the wire in the presence of screening by the tip . , for @xmath115 .
bottom panel : tunnel current for @xmath116 . on both graphics ,
the screening strength is @xmath117 ( solid line ) , @xmath118 ( dotted line ) , and @xmath119 ( dashed line).,title="fig:",width=245 ] , for @xmath115 .
bottom panel : tunnel current for @xmath116 . on both graphics ,
the screening strength is @xmath117 ( solid line ) , @xmath118 ( dotted line ) , and @xmath119 ( dashed line).,title="fig:",width=245 ] when we report these green s functions in eq .
( [ tunnel_current ] ) , we obtain up to second order in @xmath57 : @xmath120~,\end{aligned}\ ] ] where we have performed the explicit sums over @xmath34 and over the keldysh indices : only the @xmath121 terms contribute to the integral , because for @xmath122 , the function we have to integrate is odd in @xmath123 . after some algebra ,
we obtain : @xmath124~,\nonumber\\\end{aligned}\ ] ] where @xmath125 , and where the summation over @xmath19 has been made assuming a non - magnetic stm tip . to simplify the notations , we assume identical fermi velocities in the wire and in the tip : @xmath126 . in fig .
[ figurecurrent ] the tunneling current is plotted as a function of applied voltage . for weak coulomb interactions ( top panel of fig .
[ figurecurrent ] ) and in the absence of screening , we see that the i - v characteristic is linear ( solid line ) .
when the tip is approached close to the wire , the i - v characteristic is modified : the amplitude of the current is reduced and the i - v curve follows a power law . for strong coulomb interactions ( bottom panel of fig .
[ figurecurrent ] ) , the i - v characteristic follows a power law in the absence of screening .
when the tip is approached close to the wire , the tunnel current stays a power law but is strongly reduced .
the interpretation of these results is the following : electrostatic interactions between the tip and the wire plays a similar role than coulomb interactions inside the wire : they both reduce the tunnel current and change the linear behavior of the i - v curve to a power law behavior .
the tunnel current is strictly linear only in the absence of coulomb interactions and electrostatic screening .
up to second order with the tunnel amplitude @xmath57 , the zero - frequency noise reads : @xmath127 , for @xmath115 .
bottom panel : zero - frequency noise for @xmath116 . on both graphics ,
the screening strength is @xmath117 ( solid line ) , @xmath118 ( dotted line ) , and @xmath119 ( dashed line).,title="fig:",width=245 ] , for @xmath115 .
bottom panel : zero - frequency noise for @xmath116 . on both graphics ,
the screening strength is @xmath117 ( solid line ) , @xmath118 ( dotted line ) , and @xmath119 ( dashed line).,title="fig:",width=245 ] when one reports the expressions of the green s functions , it reduces to : @xmath128 which leads to @xmath129~,\nonumber\\\end{aligned}\ ] ] where the summation over @xmath19 has been made assuming a non - magnetic stm tip . in the top panel of fig .
[ figurenoise ] , we plot the zero - frequency noise as a function of applied voltage for weak coulomb interactions .
the zero - frequency noise , which is linear in the absence of screening ( solid line ) , follows a power law in the presence of screening , similarly to what we obtain for the tunnel current . for strong coulomb interactions ,
the zero - frequency noise follows a power law whatever the screening strength is ( bottom panel of fig .
[ figurenoise ] ) .
the effect of the screening is to reduce strongly the zero - frequency noise .
an interesting quantity to consider is the fano factor defined as the ratio between the zero - frequency noise and the average current : @xmath130 this fano factor is related to the charge of the carrier which is transferred trough the conductor : for example , @xmath131 equals @xmath132 for a junction between normal metals , whereas @xmath131 equals @xmath133 for a junction between a normal metal and a superconductor@xcite . in a two - dimensional electron gas in the fractional quantum hall regime , simultaneous measurements of the current and noise at a constriction were performed and a fano factor equals to @xmath134 , which corresponds to the fractional charge , was obtained@xcite . here , the fano factor stays equal to the electron charge @xmath132 even in the presence of screening .
the reason is that only electron can tunnel from the tip to the wire and we end up with the standart schottky relation@xcite : @xmath135 at zero frequency , the non - symmetrized noise and the symmetrized noise are identical because of the time translation invariance .
it is not more the case when one studies the finite - frequency noise@xcite . in the next section ,
we focus our interest to the finite - frequency non - symmetrized noise @xmath136 , knowing that the symmetric noise can be deduced simply by taking @xmath137 . up to the second order with the tunnel amplitude @xmath57 ,
the finite - frequency non - symmetrized noise is : @xmath138 ) .
bottom panel : the same quantity in the presence of coulomb interactions ( @xmath139 ) . on both graphics ,
the screening strength is @xmath140 ( solid line ) , @xmath141 ( dotted line ) , and @xmath142 ( dashed line ) , and the voltage value is @xmath143.,title="fig:",width=264 ] ) .
bottom panel : the same quantity in the presence of coulomb interactions ( @xmath139 ) . on both graphics ,
the screening strength is @xmath140 ( solid line ) , @xmath141 ( dotted line ) , and @xmath142 ( dashed line ) , and the voltage value is @xmath143.,title="fig:",width=264 ] when one reports the expressions of the green s functions , it reduces to : @xmath144 finally , @xmath145~.\nonumber\\\end{aligned}\ ] ] in experiments , we have access to the total noise which contains many contributions , such as zero - point fluctuations , @xmath146 noise@xcite or thermal fluctuations@xcite .
if one wants to isolate the contribution due to the applied voltage ( shot noise ) , one has to look at the non - symmetrized excess noise defined as : @xmath147 in the top panel of fig .
[ figureffnoise ] , we plot the non - symmetrized excess noise , divided by the tunnel current times the electron charge , in the absence of coulomb interactions . in the unscreened case ,
the emission noise ( @xmath148 ) cancels when @xmath149 , and the absorption noise ( @xmath150 ) cancels when @xmath151 ( solid line in the top panel of fig .
[ figureffnoise ] ) . in the screened case ,
the emission noise cancels when @xmath152 , whereas the absorption noise does not cancel any more when @xmath153 ( dotted and dashed lines in the top panel of fig . [ figureffnoise ] ) . in the presence of coulomb interactions ( bottom panel of fig .
[ figureffnoise ] ) , the non - symmetrized excess noise is always asymmetric , whatever the value of the screening strength is . from these graphics , we conclude that the coulomb interactions inside the wire and the electrostatic potential due to the tip have a similar effect on the asymmetry of the finite - frequency excess noise .
moreover , the excess noise is symmetric only when both coulomb interactions in the wire and screening by the stm tip are turned off . in refs . and
, it has been shown that the difference between the absorption noise and the emission noise is related to the real part of the conductance : @xmath154~.\end{aligned}\ ] ] as a consequence , the real part of the excess conductance can easily be obtained by taking the difference between the excess noise at positive frequency and the excess noise at negative frequency : @xmath155=\frac{\delta s_t(\omega)-\delta s_t(-\omega)}{2\hbar\omega}~,\end{aligned}\ ] ] where @xmath150 . ) .
on both graphics , the coulomb interactions is @xmath156 ( solid line ) , @xmath115 ( dotted line ) , and @xmath139 ( dashed line ) , and the voltage value is @xmath143.,title="fig:",width=264 ] ) . on both graphics ,
the coulomb interactions is @xmath156 ( solid line ) , @xmath115 ( dotted line ) , and @xmath139 ( dashed line ) , and the voltage value is @xmath143.,title="fig:",width=264 ] in fig .
[ figurecond ] , we plot the real part of the excess conductance normalized by its value at zero frequency .
the top panel of the fig .
[ figurecond ] is obtained when screening is absent : we see that the excess conductance decreases rapidly and converges through its value in the absence of coulomb interactions , which is zero . indeed , in the absence of coulomb interactions ( i.e. , for @xmath157 ) , the current is linear with the voltage and as a consequence , the conductance is constant in voltage .
thus , the excess conductance is simply zero in this case .
we also see a cusp when the frequency equals the voltage , here at @xmath158 , which is attenuated when the coulomb interactions strength increases . the bottom panel of the fig .
[ figurecond ] shows the excess conductance in the presence of screening .
we observe a weaker decreasing at low frequency in comparison to the unscreened case .
more interestingly , the excess conductance becomes almost independent of the coulomb interactions strength .
this is due to the fact that the effect of screening by the stm tip dominates over the effect of the coulomb interactions in the quantum wire .
in this paper , we have calculated the transport properties of a luttinger liquid wire in the presence of electrostatic screening by a stm tip . whereas the spectral properties of the wire depend only on the green s functions of the wire@xcite , the transport properties depend also on the green s function of the tip and on new correlators that we call mixed green s functions because they mix the bosonic fields of the wire with the bosonic field of the tip .
these new correlators are non - zero due to the electrostatic tip - wire interaction . we have solved the dyson equations associated to them in the case where the double derivative of the electrostatic potential can be approximated as a local interaction and we have calculated the tunnel current , the noise and the conductance . we have shown that the tunnel current and the zero - frequency noise are strongly reduced in comparison to the unscreened situation .
in addition , the i - v characteristic , which is linear in the absence of coulomb interactions in the wire , follows a power law behavior when screening is turned on .
we can conclude that the electrostatic interactions between the wire and the tip have a similar effect on the i - v curve than the electron - electron interactions inside the wire .
next , we have studied the finite - frequency non - symmetrized excess noise and we obtain an interesting result . whereas in the absence of screening , the non - symmetrized noise associated to a non - interacting wire is symmetric in frequency , it becomes asymmetric in the presence of screening .
for an interacting wire , the non - symmetrized excess noise is always asymmetric in frequency regardless of the screening strength is .
the explanation of this asymmetric signal for a non - interacting wire is again the fact that the electrostatic potential between the wire and the tip has similar effects as electron - electron interactions .
the subtle difference here is that the effective local density - density interaction which is induced by the presence of the tip contains retardation effects : they involve the dynamics of the tip ( via its green s function ) .
a useful tool to test the asymmetry of the excess noise is to calculate the excess conductance : indeed , when the excess noise is symmetric in frequency , the excess conductance is simply zero . we have calculated the excess conductance in different situations . in the absence of coulomb interactions and screening ,
this quantity cancels . on the contrary , in the presence of coulomb interactions , or of screening
, the excess conductance does not cancel any more .
the results that we have obtained in this paper are preliminaries since we assume an infinite length wire . however , with the help of the dyson equations that we have derived , it is now possible to calculate the transport properties for more realistic situations : in particular , in the presence of electrical contacts , at the extremities of the wire , which are needed to measure transport properties .
we can estimate that current and noise will be affected by the contacts only for short length wire , due to finite size effects .
moreover , it would be interesting to calculate the current and its fluctuations in the wire , and not only the tunnel current and tunnel noise .
it has been shown that the electrical contacts induce oscillating behavior of the finite - frequency noise in the wire due to reflections at the contacts@xcite .
the question of how these oscillations are modified by the screening , in the case where the electrostatic potential between the wire and the tip has a finite length extension , would also be interesting to consider .
we thank d. bercioux , f. dolcini , h. grabert , f. hekking and i. safi for valuable discussions .
the starting point is the dyson equation given by eq .
( [ mixed_dyson_1 ] ) that verifies the mixed green s function @xmath159 in imaginary time : @xmath160 where @xmath161 is the green s function of the tip in the presence of screening , @xmath162 , and @xmath163 is the green s function of the wire in the absence of screening .
since we assume that the tunnel transfert takes place at @xmath164 , we only need @xmath165 which obeys to : @xmath166 using the time translation invariance , we perform a fourier transform and obtain : @xmath167 reporting the expression of @xmath168 which was determined in ref . :
we obtain @xmath170 whose fourier transform is : @xmath171 finally , performing an analytic continuation @xmath172 where @xmath173 , we get the keldysh mixed green s function @xmath174 next , we consider the dyson equation verified by @xmath175 which is given by eq .
( [ mixed_dyson_2 ] ) : @xmath176 the resolution of the equation is similar to the previous one , it leads to : @xmath177 now , we look at the dyson equation verified by @xmath178 which is given by eq .
( [ mixed_dyson_3 ] ) : @xmath179 at positions @xmath164 , it reduces to : @xmath180 since we have @xmath181 ( see appendix b ) , it leads to @xmath182 and as a consequence : @xmath183 finally , we look at the dyson equation for @xmath184 which is given by eq .
( [ mixed_dyson_4 ] ) : @xmath185 at positions @xmath164 , it reduces to : @xmath186 since we have @xmath187 ( see appendix b ) , it leads to @xmath188 and as a consequence : @xmath189
for an infinite length wire and in the absence of screening , the matsubara bosonic green s functions at zero temperature and at positions @xmath190 are : @xmath191 99 z. yao , h.w.ch .
postma , l. balents , and c. dekker , nature * 402 * , 273 ( 1999 ) . c. kane , l. balents , and m.p.a .
fisher , phys .
lett . * 79 * , 5086 ( 1997 ) .
s.j . tans , m.h .
devoret , r.j.a .
groeneveld , and c. dekker , nature * 394 * , 761 ( 1998 ) .
a. crpieux , r. guyon , p. devillard , and t. martin , phys .
b * 67 * , 205408 ( 2003 ) .
lebedev , a. crpieux , and t. martin , phys .
b * 71 * , 075416 ( 2005 ) .
r. deblock , e. onac , l. gurevich , and l.p .
kouwenhoven , science * 301 * , 203 ( 2003 ) .
billangeon , f. pierre , h. bouchiat , and r. deblock , phys .
* 96 * , 136804 ( 2006 ) .
m. guigou , t. martin , and a. crpieux , _ screening of a luttinger liquid wire by a scanning tunneling microscope tip : i. spectral properties _ , arxiv:0904.4019 .
m.p.a . fisher and l.i .
glazman in _ mesoscopic electron transport _ , nato asi series e , kluwer ac .
, dordrecht .
keldysh , sov .
jetp * 20 * , 1018 ( 1965 ) .
m.e . peskin and d.v .
schroeder , in _ an introduction to quantum field theory _ , west - view press ( 1995 ) .
lesovik , t. martin , and j. torrs , phys .
b * 60 * , 11935 ( 1999 ) ; x. jehl , p. payet - burin , c. baraduc , r. calemczuk , and m. sanquer , phys .
. lett . * 83 * , 1660 ( 1999 ) ; a.a .
kozhevnikov , r.j .
schoelkopf , and d.e .
prober , _ ibid .
_ * 84 * , 3398 ( 2000 ) ; r. cron , m.f .
goffman , d. esteve , and c. urbina , _ ibid . _ * 86 * , 4104 ( 2001 ) .
l. saminadayar , d.c .
glattli , y. jin , and b. etienne , phys .
lett . * 79 * , 2526 ( 1997 ) .
r. de picciotto , m. reznikov , m. heiblum , v. umansky , g. bunin , and d. malahu , nature * 389 * , 162 ( 1997 ) .
w. schottky , ann .
( leipzig ) * 57 * , 541 ( 1928 ) . c. bena and i. safi , phys .
b * 76 * , 125317 ( 2007 ) .
i. safi , c. bena , and a. crpieux , phys .
b * 78 * , 205422 ( 2008 ) . p. dutta and p.m. horn , rev .
phys . * 53 * , 497 ( 1981 ) .
johnson , phys . rev . *
32 * , 97 ( 1928 ) .
h. nyquist , phys .
rev . * 32 * , 110 ( 1928 ) .
i. safi and e.v .
sukhorukov , private communication .
m. guigou , a. popoff , t. martin , and a. crpieux , phys . rev .
b * 76 * , 045104 ( 2007 ) .
s. pugnetti , f. dolcini , d. bercioux , and h. grabert , phys .
b * 79 * , 035121 ( 2009 ) . | we study the effect of an electrostatic coupling between a scanning tunneling microscope tip and a luttinger liquid wire on the tunneling current and noise between the two . solving the dyson equations
non perturbatively for a local interaction potential , we derive the green s functions associated to the wire and to the tip .
interestingly , the electrostatic coupling leads to the existence of new correlators , which we call mixed green s functions , which are correlators between the bosonic fields of the wire and the tip .
next , we calculate the transport properties up to second order with the amplitude of the tunnel transfer : the tunnel current is strongly reduced by the presence of screening . the zero - frequency noise is modified in a similar way , but the fano factor remains unchanged .
we also consider the effect of the screening on the asymmetry of the finite - frequency non - symmetrized noise and on the conductance . |
in recent decades , the combination of large aperture telescopes and high resolution spectrographs have allowed for precision analysis of a variety of astrophysical objects .
echelle spectrographs are the instrument of choice for high resolution , and most large telescopes now have one @xcite .
echelle gratings can give spectra with high spectral resolution , with a large slit width , and a large wavelength range in a single setting @xcite .
an echelle grating disperses the spectrum into many tens of spectral orders , which are then cross dispersed by a second dispersive element so that the orders can be placed , one above the other , on a rectangular ccd detector .
it is difficult to combine the spectra from the many spectral orders of an echelle to produce a single continuous spectrum .
this difficulty arises because the response varies rapidly across each order , and at a given wavelength is usually different in different orders .
high quality relative flux calibration of echelle spectra is highly desirable in many scientific applications .
for example , accurate flux levels over a large range of wavelengths makes it much easier to place continuum levels on spectra with pervasive blended absorption , such as the lyman alpha forest absorption seen in high redshift qso spectra . in this paper
, we discuss the relative flux calibration of spectra from the hires echelle spectrograph on the keck - i telescope @xcite .
we do not discuss absolute flux calibration , as it requires additional calibration data and is not necessary for our absorption line work .
we intend that a spectrum with relative flux calibration has the correct shape over some range of wavelengths , and differs from an absolute calibration in only the normalization .
we developed the methods described in this paper to improve our measurements of the primordial deuterium to hydrogen abundance ratio in qso absorption systems , for which the bulk of our spectra come from hires .
we have found that we obtain more accurate and reliable estimates of the absorption column densities when we use spectra with accurate relative flux .
high quality flux calibration was not a major design goal for hires , and we have found that special steps must be taken to obtain the quality of calibration that we need for our work on d / h .
the usual @xcite methods of echelle flux calibration appear inadequate for reasons that we do not fully understand .
this inadequacy motivated the development of the methods we describe .
we would like to both minimize the flux errors in our spectra , and to estimate the size of errors which remain after our calibrations .
we shall find that the flux calibration errors depend on wavelength and they are correlated on various wavelength scales .
we would like to estimate the size of the errors on these different scales .
we would also like to make the error in the relative flux calibration less than the photon noise on some relevant scale .
for example , when we fit a flat continuum level to a 50 pixel segment of a hires spectrum with signal to noise ratio of 100 per pixel , the photon noise on the continuum level is 0.14% .
we found that is harder to approach a given accuracy in relative flux calibration in many places .
these places include the regions where echelle orders join , regions where spectra have lower signal to noise ratio , wavelengths in the near uv , and in general as the wavelength range increases .
it is very hard to get flux calibration errors of @xmath0% over even a wavelength range of @xmath1 within one hires order .
fortunately for our absorption line work , errors that vary smoothly over wavelength scales @xmath2 are not as serious as smaller scale errors . in this paper
, we describe a way of calibrating the relative flux in a hires spectrum using well calibrated reference spectra of the same target to transfer flux information from other spectrographs to the hires spectra .
we force each hires spectrum to have the shape of these reference spectra .
this method should correct a variety of flux calibration errors , both from variable vignetting and differences in spectrum extraction and reduction .
the method could also be applied to spectra from various spectrographs , and not just echelles .
this method is based on that introduced by @xcite and applied with improvements to in @xcite .
the paper is organized as follows : we first describe the nature of inconsistencies between repeated hires observations of a target , and how this impacts flux calibration .
second , we describe the spectra we used to illustrate our flux calibration method .
third , we describe at first in overview , and then in detail , our methods for flux calibration .
fourth , we describe how we combine the hires orders that we have flux calibrated individually . finally , we discuss the accuracy and errors of our method .
the usual methods of flux calibration appear inadequate when applied to hires spectra .
when we determine the instrument response by observing a spectrophotometric standard star , different exposures of the same star give different instrument response functions , even when we believe that the exposures were taken with the same instrument configuration .
for example , in figure [ x2 ] , we show the signal extracted from two hires integrations of g191-b2b .
the exposures were taken on consecutive nights , with similar instrument configurations .
on the second night the star was 0.07 degrees higher in the sky and the image rotator ( appendix [ rotator ] ) physical angle differed by 0.071 degrees to compensate for the change in parallactic angle .
the spectra were both extracted with tom barlow s makee software ( april 2001 version ) .
we show the raw " adu from the ccd before division by the flat field integration or any other calibration .
the differences between the two exposures shown in figure [ x2 ] are large and unexpected .
in particular , even though the exposures were taken on different nights , we did not expect to see large ( @xmath3% ) differences within a single order , even after the two exposures were normalized to have the same mean flux in that order .
the differences are largest ( @xmath4 10% ) at the ends of the order . similar differences ( both in shape and magnitude ) are present in each observed spectral order .
however , the form of the difference is not precisely the same for each order , as demonstrated in figure [ x22c ] .
@xcite also noted the possibility of systematic flux errors of 10% near order overlaps .
flux differences have also been reported for the subaru telescope hds echelle spectrograph that sometimes shows 10% changes during observations @xcite .
we have examined approximately 20 other pairs of standard star integrations from hires that have similar instrumental setups .
most show differences of order 10% , though the exact shape of the differences is not always the same .
in addition to the
u " shaped ratios ( seen in figure [ x2 ] ) , we see three other shapes for ratios : near flat ratios , tilted and ogive ( or fallen s " ) shapes . in each case , the shapes of the ratios vary gradually order - to - order in a systematic way , such that adjacent orders have similar shapes
. the shapes of the ratios vary much more between pairs of integrations than they do from order - to - order for a given pair of integrations .
approximately 30% of integration pairs give flat ratios that indicate that the instrumental response was very similar for the two integrations , which should make flux calibration simple .
a cursory examination did not show any predictors ( e.g. telescope elevation , rotator angle , position of target on the slit , seeing ) as to which pairs would be similar and which different . in figure [ x16 ] , we show hires integrations on two stars that we flux calibrated in the usual manner , each using a response function determined from a hires exposure of a standard star .
the spectra that we calibrated differ from the known flux levels by large amounts over a wide range of wavelengths .
the main deviations are systematic across each hires order .
we also see large differences at the order overlap wavelengths .
different ways of combining spectral orders leave different flux calibration errors .
if we take the mean of the signals then the flux will jump in a single pixel , by up to 10% , where each order begins and ends .
there will be approximately 70 such jumps in a complete hires spectrum with 36 orders . in the past
, we have attempted to reduce the flux errors by fitting continua to each order , and dividing by these continua , before we take the mean flux in a wavelength overlap .
this method is unsatisfactory because it is very hard to ensure that the continua that we fit to adjacent orders have both the same flux levels and the same shapes , and we loose flux information . for some unknown reason , long integrations on a qso show smaller differences than short integrations on a bright star .
a dependence on integration time might relate to some averaging , perhaps related to the target position on the slit .
a dependence on brightness might relate to the 100 times lower signal in the qso integrations , and perhaps subsequent differences in the spectrum extraction .
we still have difficulty flux calibrating qso exposures , even though they appear to be more stable than exposures on standard stars , because usual procedures still require that we determine the instrument response from a standard star exposure .
the variations we see in standard star exposures indicate that there is some instability in either keck+hires , in our data reduction processes , or in both .
we have investigated two possible origins for the instability : variable vignetting inside of hires , and inadequate the extraction of the spectra .
we do not see a clear signature of either in our spectra , but we do know that the vignetting is expected to change .
we have explored the possibility of extraction errors by varying the type of profile used during extraction , and the profile width , but found no differences from the standard extraction results from makee .
we also measured flux ratios from the raw counts in the ccd images that were similar to those in the spectra extracted by makee .
we know that the vignetting inside hires changes with the position angle of the sky image on the hires slit and with the telescope elevation ( appendix [ rotator ] ) .
when the image rotator is used there are two main options .
we can use the image rotator in vertical angle mode " to keep the vertical angle along the slit , so that the position angle varies with the position of the target in the sky .
the rotator can also be used to keep a desired position angle along the slit .
if the rotator is not used , the relevant angle is the telescope elevation .
the variation in vignetting arises from a known mechanical and optical misalignment between the keck - i telescope and hires , and the expected amount of change in the vignetting , from ray tracing kindly provided by steve vogt , is approximately 10% , consistent with our observations .
we can account for why the ratios of hires spectra have similar shapes across all orders if the variable vignetting happens after the echelle , and before leaving the cross - disperser . in this part of hires the light from the red end of each order
is separated from that in the blue end , but all orders are coincident .
we might explain the shape of the ratios , and the similarity from order - to - order , if a varying amount of light misses the top and bottom of the cross - disperser ( the grooves are vertical ) , where the red and blue ends of the orders land .
this variation could arise when the cone of light from the telescope tips up and down in the vertical plane that connects the center of the tertiary mirror and the hires slit . however , we suspect that vignetting is not the sole cause of the variations in the flux , because we have seen variations in spectra taken under apparently identical instrumental conditions ( same elevation , image rotator setting , and target location on the slit ) on consecutive nights , as we saw in figure [ x2 ]
. the changes might also come from differences in the extraction of the spectra from the ccd image , e.g. if we do not extract a fixed proportion of the flux recorded at each wavelength .
differences in the extraction of spectra , including multiple integrations on a given target , are likely whenever there are changes in conditions , such as the location of the target along the slit , the seeing , the sky brightness and the amount of signal recorded . however , extraction problems seem an unlikely explanation for standard stars that have high signal to noise ratio .
here , we introduce the spectra that we will use to illustrate our method of relative flux calibration . this is the set of spectra that we used to measure d / h towards ( @xmath5 , v=16.9 ; kirkman 2003 ) . for our d / h work ,
we were mostly interested the flux calibration in a 40 region centered on the damped ly@xmath6line near 4285 , and on the lyman limit near 3210 .
we began the development of the methods using a similar set of spectra of that we had used to make an earlier d / h measurement @xcite .
we used 5 spectra from the kast spectrograph on the lick 3-m telescope , and one from the esi echelle spectrograph on the keck - ii telescope .
we used both the kast and esi spectra separately to make independent flux calibrations of 8 integrations from hires .
further details of the observations , including the dates , the resolution , the mean s / n , and plots of the spectra are in @xcite .
all of the spectra that we used were shifted into the heliocentric frame , and placed on a logarithmic wavelength scale with a constant velocity increment per pixel , although with different increments for different spectra .
the kast double spectrograph uses a beam splitter to record blue and red spectra simultaneously in two cameras . for
we have five kast integrations , one from 1997 , and two each from 1999 and 2001 .
all integrations were obtained using the d46 dichroic that splits the spectrum near 4600 , the 830 line / mm grism blazed at 3460 for the blue side , and the 1200 line / mm grating blazed at 5000 for the red side .
we reduced all exposures with the iraf package longslit .
esi covers from 3900 11,000 in ten overlapping orders @xcite .
we have one exposure of using a 1 slit , taken in the echellette mode on january 11 , 2000 . from the same night
, we also have an exposure of the flux standard star feige 110 .
our hires spectra of all used similar instrumental setups .
the angle of the hires echelle was the same for all exposures , and placed the center of each order near the center of the ccd .
the cross - disperser angle was also similar for all exposures , except for one exposure that extended to much larger wavelengths .
the image rotator ( appendix [ rotator ] ) was used in vertical mode , to minimize slit losses from atmospheric dispersion .
we used the c5 decker , which provides an entrance aperture to the spectrograph with dimensions @xmath7 . in each case
we placed the target near the middle of the slit .
the spectra were all recorded on the engineering grade tektronix ccd with 2048 @xmath8 2048 pixels that has been used in hires since 1994 .
the hires pixel size is 2.1 km s@xmath9 .
the ccd is large enough to record beyond the free spectral range for all orders at wavelengths @xmath10 , and we placed the spectra on the ccd such that we did record this flux for all such orders .
all but one integration covered from near 3200 out to 4700 in approximately 36 orders .
these integrations were 7200 9000 seconds long .
the s / n in each integration is near 3 per 2.1 km s@xmath9 pixel near the lyman limit at @xmath4 3200 , and rises to near 35 at the peak of the ly@xmath6 emission line at @xmath4 4400 .
the hires spectra we use are the normal output from the makee software , which are the raw counts spectrum divided by spectra extracted from flat field integrations .
in addition , the ccd defects were marked with negative error values .
these spectra differ from the raw counts that we discussed in the previous section in that the flat field division has removed most of the variation across the orders due to the blaze and vignetting .
although this flat field division may introduce additional undesirable changes in the relative flux , we proceed with these spectra because it is imperative that the ccd defects have been removed .
we measured the instrument resolution of hires by fitting gaussian functions to narrow , apparently un - blended lines in a single thorium - argon arc integration taken before one of the qso integrations .
we fit hundreds of arc lines in all parts of the spectrum , and found a dispersion of @xmath11 km s@xmath9 , which corresponds to a fwhm spectral resolution of @xmath12 km s@xmath9(@xmath13 km s@xmath9 ) .
we did not detect any variation in the resolution with wavelength .
however , we did detect that the arc lines are not gaussian in shape , with more extended wings , such that the best fitting @xmath14 increases when we extend the fitting range around an individual line .
we also expect that the spectra will have slightly different fwhm from the arc spectra , because the illumination of the slit is different .
the wavelength shifts that we describe in the next section suggest that the resolution will depend in part on the guiding and seeing .
we measured wavelength shifts between the hires integrations and we shifted the spectra onto the same scale to correct these shifts .
we measured the cross - correlations between each of the 7 integrations with the eighth that we used as the reference .
an example of the shifts is shown in figure [ x20a ] .
comparisons of other pairs of integrations often show a much larger dispersion in the measured shifts . in all cases ,
the shifts measured in each order are consistent with a single shift for each integration .
the shifts had a standard deviation of 0.7 km s@xmath9 , which is 30% of one hires ccd pixel .
each shift was measured to an accuracy of 0.13 km s@xmath9 , which we determined from the scatter in the shifts that we measured separately for each order .
these shifts may arise from differences in the placement of the qso light in the hires slit , which projects to approximately 4 hires pixels .
the 0.7 km s@xmath9 rms shift is approximately 9% of the fwhm resolution , which itself is similar to the slit width .
the shifts do not correlate with hour angle or the correction that was applied for the earth s orbital motion ( @xmath15 km s@xmath9 ) and spin ( @xmath16 km s@xmath9 ) .
the shifts are also larger than we expect from wavelength scale errors .
however , we did find much larger wavelength scale errors when we did not ensure that makee used enough arc lines to determine the dispersion solution for each order .
there are three main steps in our procedure to apply relative flux calibration to hires spectra . * step 1 : flux calibrate a high quality reference spectrum of the target .
* step 2 : flux calibrate the hires echelle orders with the reference spectrum .
this imposes the long scale ( @xmath17 ) spectral shape of the reference spectrum upon each hires echelle order .
the flux information on smaller scales ( e.g. absorption line profiles ) still comes from hires . *
step 3 : combine the hires orders .
we first sum the integrations and then join the orders to give one continuous spectrum .
these procedures do not replace standard ccd spectrum extraction procedures .
rather , they should be thought of as a software patch " , applied to correct errors that remain after the spectra have been extracted , and perhaps calibrated , in the usual way .
reasonably well calibrated hires spectra are required as inputs to our method , because the flux information on scales @xmath18 will still come directly from the hires spectra .
the procedure is not unique , and we expect that other sequences might be appropriate for different spectra .
we now discuss in further detail each of the steps outlined above .
since we can not transfer information about the instrument response between exposures , we must `` self - calibrate '' each exposure we take with hires .
the first step of our method is thus to obtain a well calibrated spectrum of the target .
the reference spectrum must come from a well calibrated and stable spectrograph . in our work on q1243 + 3047 , we obtained reference spectra from both lick+kast and keck+esi , and compared them to gauge the accuracy of the final flux calibration .
to calibrate the flux in our kast spectra of , we took flux information from a model spectrum of g191 b2b for the 1997 spectrum , and a stis spectrum of bd 28 4211 for the 1999 and 2001 exposures .
we discuss the reasons behind our choice of these standard stars in appendix a. we measured and matched the resolution of the observed and reference spectra of the standard stars before we used them to calculate the kast response function .
the dip in the response at the end of blue ccd and the beginning of red the ccd is due to the changing efficiency of the beam splitter . in figure [ x9 ]
, we illustrate the flux calibration process .
the spectrum shows atmospheric ozone absorption lines @xcite as the wiggles of the raw ccd counts ( panel b ) below 3400 .
their strength depends on the temperature of the ozone layer and the effective airmass of the integration , and we have not made appropriate adjustments .
we left the wiggles un - smoothed in the response ( figure [ x9 ] , panel d ) to help partially remove their effect from the quasar spectrum . in figure [ x23 ]
, we show the accuracy of the flux calibration of a kast spectrum of another star .
the flux residuals between our calibrated kast spectrum and a stis spectrum of the same star are less than 3% .
we flux calibrate esi spectra in the same way we do those from kast . in figure [ x10 ]
, we show the steps in the flux calibration of an esi spectrum .
the esi orders overlap in wavelength , and in figure [ x6 ] , we see that the orders do not always have the same flux .
these differences increase in size as we approach the end of an order .
we have not investigated the origin of these flux differences .
we cut off most of the regions where the differences exceeds 2% ( which is the noise level in our spectrum ) without leaving any gaps in the wavelength coverage .
we do not know the size of the remaining error , especially in regions where there was no wavelength overlap , because we do not have redundant observations of bright stars . nonetheless , the esi spectra could have better relative flux calibration than hires spectra , for several reasons .
esi has fewer orders , each of which covers more wavelength range and a much larger portion of each order is sampled by adjacent orders .
esi has a fixed instrumental configuration and our esi spectra have much higher s / n than our individual hires integrations , which may change the proportion of the flux that is extracted .
the errors in the relative flux calibration of the reference spectrum are a fundamental limitation on how well we can apply relative flux calibration to the hires spectrum .
one way to explore these errors is to compare different reference spectra .
we will see that the differences increase with wavelength range and they are the main source of error in our calibration of our hires spectra of .
we found that the 5 kast spectra show two types of shape .
the two from 1999 are similar , as are the two from 2001 .
we call the sum of the two flux calibrated spectra from 1999 k99 , and the two from 2001 k01 .
ksum is our name for the sum of all five spectra .
the 1997 spectrum is similar to k01 , but has much lower s / n .
these two groups , k99 and k01 , differ in shape on the largest scales across the whole lyman-@xmath6 forest , but they do have very similar shapes across a few hundred angstroms after we normalize them to each other at those wavelengths .
these differences are best seen when we smooth them slightly by differing amounts to reduce differences in the spectral resolution .
the k01 spectra have up to 10% systematically higher flux at wavelengths @xmath19 than do the k99 spectra .
the k01 spectra had a few percent lower flux from 4000 4300 and higher flux across the ly@xmath6 emission line .
we do not know the origin of these differences .
possible origins include variation in atmospheric extinction @xcite or a variation in the qso .
we find that the esi spectra differed from the various kast spectra by typically 2% or less per kast pixel , from 4100 4400 .
the differences correlate over a few kast pixels in the lyman-@xmath6 forest , with no large scale trends , except that to the red side of the ly@xmath6 emission line the ratio of the kast to esi flux increases by approximately 5% from 4300 4400 for all five kast spectra .
we calibrated the relative flux in a hires spectrum using a calibrated reference spectrum from either kast or esi .
we divided a smoothed version of the hires spectrum by the reference spectrum to find the conversion ratio " spectrum .
we found that the smoothed hires spectrum and the reference had to have the same wavelength scale and resolution , because the lyman-@xmath6 forest absorption lines cause the flux to vary rapidly in wavelength .
we calibrated the individual orders of each hires integration using a smooth function fitted to the conversion ratio spectra , one for each order of each integration . in figure [ x17 ]
, we illustrate the calibration of the relative flux in the hires orders that we describe in the following five sub - sections . because the wavelengths from hires are more accurate , we shifted the kast spectra onto the hires wavelength scale .
we measured the shifts by cross - correlation of complete hires orders , and we confirmed the values by cross - correlating individual strong lines in the lyman-@xmath6 forest . some of the wavelength shifts may arise because the qso was not exactly centered in the slit . this is a reasonable explanation for the typical shift which was 42 km s@xmath9 , or 0.4 kast pixels , only 17% of the projected width of a 2 arcsecond wide slit .
these shifts should vary monotonically along a spectrum , and some kast spectra show this .
however , figure [ x11 ] shows that other spectra have more complex wavelength shifts . in such cases we measured the mean shift for each hires order , and we fit a low order polynomial to these mean values , similar to that used to fit the arc line wavelengths , to give a smoothly changing wavelength scale without gaps or discontinuities .
we also placed the esi spectra on the hires wavelength scale .
this wavelength scale assignment was obtained by first smoothing the hires spectra to the approximate spectral resolution of esi , and then finding the velocity shift by cross - correlation of the esi with the hires integration .
we shifted the complete esi spectrum by + 5.61 km s@xmath9 , which is 7.5% of the projected 1 arcsec slit width . as with the kast spectra , this shift is larger than we would expect from the wavelength fits to the arc calibration lines and may arise because the qso was not exactly centered in the slit .
we smooth the hires spectrum to match the resolution of the kast spectrum .
this procedure is sensitive to wavelength shifts between the two spectra , and hence it was done after the wavelength scales were matched .
in contrast , the wavelength scale matching is insensitive to the spectral resolution .
we smoothed the hires spectrum , and sampled it in the wavelength bins of the kast spectrum .
we smoothed with a gaussian function of known fwhm , truncated at @xmath20 , and normalized to unit area .
we smoothed by different fwhm to find the amount of smoothing that left the smallest residuals between the smoothed hires and kast spectra .
these residuals appeared flat across strong absorption lines , which suggests that a gaussian is an adequate approximation to the line spread function of the kast spectrum .
as with the wavelength matching , the lyman-@xmath6 forest provides additional signal for resolution matching .
the kast spectral resolution varies from spectrum to spectrum , and it can vary with wavelength in a spectrum , depending on where we chose to focus .
we took the kast resolution to be the fwhm of the smoothing applied to the hires , after subtracting the initial hires 8 km s@xmath9 fwhm in quadrature .
for example , in the 1999 spectra of , the fwhm of the kast spectrum is near 300 km s@xmath9 near 3300 and from 3800 4300 but it improves to 220 km s@xmath9 near 3500 , where the measurement error is approximately 10 30 km s@xmath9 , depending on the region of the lyman-@xmath6 forest .
we smoothed the hires spectrum by a single mean fwhm even when the resolution varied along the kast spectrum . using similar methods we found that the esi spectrum had a fwhm of 63.2 @xmath21 3.0 km s@xmath9 .
we divide the smoothed hires spectrum by the kast spectrum , to obtain the conversion ratio " ( cr ) spectrum .
the lyman-@xmath6 forest hinders our calculation of the cr , because we are very sensitive to slight remaining errors in the wavelengths and the resolution .
when we calculate the cr , we weight the flux in the individual spectra by their errors , and we assign an error to each pixel in the cr .
it is common to see increased error in the cr near strong absorption lines ( e.g. near 4285 in panel ( e ) of figure [ x17 ] ) , but this error does not have a systematic shape when the wavelengths and resolution are well matched .
we fit the cr spectra to obtain a smoothly changing function .
the original cr spectra vary from pixel to pixel because of photon noise in the reference and hires spectra , especially in strong absorption lines .
these variations can be much larger than the expected change in the flux calibration and hence we can improve the flux calibration by interpolating .
we explored various ways of smoothing and fitting , some in one dimension , along each order separately , and others in two dimensions , both along and between adjacent orders .
the best choice will increase the s / n in the cr as much as possible without changing the structure .
we found that a 4th order chebyshev polynomial fit to the cr spectrum for each hires order was a good choice .
we choose this order by trial and error .
it leaves enough freedom to fit the shapes of the cr and give a reduced @xmath22 .
other fits would be appropriate in spectra with different amount of structure and s / n . in figure [ x21b ] , we show the cr spectra for many hires orders .
we see that the cr spectra for adjacent orders are very similar in shape , but change systematically as we move across many orders .
we also found that the changes in shape from order - to - order are larger than those between integrations for a given order , except for the effects of photon noise in the cr .
we calculate the cr twice , in an iterative fashion , to improve the accuracy near the ends of the hires orders .
the first time , we stop the calculation before the order ends , where the filter , which we use to smooth the hires spectrum to the resolution of the reference spectrum , just touches the last pixel in the order .
we can not , at that time , calculate the cr for the remaining pixels because we do not know the flux from beyond the end of the order .
this flux is needed when we smooth the hires spectrum to match the reference spectrum .
we do , however , allow the fit to the cr to extrapolate to the end of the order , and we use that extrapolation to make the first estimate of the flux calibration . when we calculate the cr the second time , we know the hires flux from beyond the end of the order , because we have already calibrated the whole hires spectrum .
an example of the type of error that can occur in the cr at order ends is shown in figure [ x21f ] . in middle panel on the left ,
the cr is erroneously low in the first two pixels at the start of order 98 .
here we calculated the cr once only , and we ignored the flux from beyond the end of the order , where there happens to be a ly@xmath23 absorption line .
this absorption line lowers the flux in the reference spectrum , but not in the hires spectrum for that order .
the error in the relative flux calibration of a hires integration depends in part on the s / n in that hires integration .
the conversion ratios in figure [ x21f ] are for a second integration of that can be compared to the integration shown in figure [ x21b ] .
the two integrations were calibrated using the same kast spectrum , and hence the differences are caused by noise in the hires integration , especially at the shortest wavelengths .
we experimented with other ways of smoothing the cr to reduce the effects of photon noise .
the 4th order polynomials smooth in wavelength along an order , but they do not use any information from adjacent orders .
we made a 2-dimensional map of the cr in the coordinate frame of the hires ccd detector , and we smoothed this map in both dimensions , both using a gaussian filter and independently by fitting a 2d surface using chebyshev polynomials .
the gaussian filter did not work well because the largest width that did not change the shape of the cr surface had a @xmath24 width of only 3 orders , not enough to reduce the noise significantly . in figure [ twodfit ]
, we show a surface fit that was an improvement over the 1d polynomial fits , but we did not use this in @xcite .
the error in the cr depends on the s / n in each cr pixel , and on the fitting or smoothing that we use .
the error will correlate on the smoothing scale .
when we apply a 4th order polynomial to each order of 40 , ( the correlation length is approximately 10 ) and this can be seen when we compare figures [ x21b ] & [ x21f ] .
there are also strong correlations in the cr near the ends of orders , and near strong absorption where the errors are largest .
the contribution to the error in the cr from the s / n in the reference or hires spectra can be estimated from the number of pixels in the correlation length .
there are approximately 34 pixels in a kast spectrum per hires order , and hence 8.5 per correlation length .
an individual kast spectrum of has s / n 60 per pixel near 4250 and 20 near 3250 .
the cr will then have errors of at least 0.6% near 4250 and 1.7% near 3250 .
we divide the original hires integration , with full spectral resolution and pixels , by the 1d fits to the cr to obtain the desired flux calibrated hires spectrum .
since we have not merged the orders , the wavelength overlaps remain .
the resulting hires integration now has relative flux calibration on each order .
the final step is to combine the fluxed hires orders into a single spectrum .
we add the hires integrations that have very similar wavelength coverage , order by order , and then merge the orders .
we choose this sequence because it facilitates checks of the relative flux calibration .
we compare the flux in each order in multiple integrations before we sum the integrations .
after they are summed , the enhanced s / n makes it easier to see errors in the flux calibration in the wavelength overlap between orders .
first , we place all orders from all hires integrations on a single wavelength scale , so that pixels from orders that overlap in wavelength sample exactly the same wavelengths .
we choose a scale with a constant velocity increment of 2.1 km s@xmath9 per pixel , equivalent to constant log wavelength increment .
this choice has two advantages over constant wavelength : in velocity units , the spectral resolution of the echelle does not vary significantly with wavelength , and we use velocity units to describe the widths of absorption line systems .
we simultaneously correct errors in the wavelength scales , using information from the cross - correlation of different integrations as discussed in ( [ hireswave ] ) .
we do this correction at the same time because we want to re - bin the hires spectra only once .
we calculate the mean flux in each pixel , from all integrations , after rejecting large deviations that we identify by evaluating the @xmath25 statistic for each pixel .
we ignore the flux in a pixel from an individual integration if it is @xmath26 away from the weighted mean for all integrations , where @xmath27 are the errors on the fluxes in the individual integrations .
the value of @xmath28 was determined iteratively to remove most pixels with conspicuously deviant flux values .
for , we use @xmath29 .
the algorithm iterates , and re - derives the @xmath30 after rejecting the flux value from one integration .
if all integrations are @xmath31 away from the weighted mean , we examine the errors .
if the errors in some of the integrations have data , shown by non - zero errors returned by makee , we use the flux from the integration with the smallest error as the mean . if , however , there is no data in a particular pixel , for example because of a known ccd defect , then we use the weighted mean of the two pixels on either side , again using only those integrations that are within @xmath32 of the weighted mean for that wavelength .
if none of the above criterion are met , we use the average of the fluxes in all integrations .
finally , we verify that adjacent orders show similar flux levels where they overlap , and we used a weighted mean to combine the orders in these regions , producing a single flux calibrated spectrum . in figure
[ f344thord ] , we show order overlap for spectra of star feige 34 . this spectrum can be compared to the spectra in figures [ x16 ] that show the flux calibration using usual methods .
the ratio of the flux in adjacent orders shows most variation near 3250 : 0.92 1.08% , and decreases to 0.98 1.02 near 4300 , where the signal to noise ratio is highest in the kast spectrum that we used as a reference .
the calibration of feige 34 used the method and algorithm that we developed and applied to our spectra of , with a few exceptions . the star does not have lyman-@xmath6 forest lines , and hence we matched the wavelength scale of the kast and hires spectra using a single shift for the whole kast spectrum .
we fitted the cr with one 4th order polynomial per order . a 6th order fit would work better near 4300 where the signal to noise ratio is high .
we consistently found , from many spectra , that the remaining difference between between the orders was largest at the end of an order , where the cr is less well known .
hence , we tapered the flux from each order using a weighting function that declined linearly from one , where the overlap begins , to zero , where the order ends .
we made several different calibrations of the hires spectra of using different low resolution spectra to convey the flux information .
we label these hires spectra by the spectra that we used for the flux calibration : hence , hk99 is the hires spectrum of calibrated using the kast k99 spectrum of , hksum was calibrated using the mean of all five kast spectra and hesi was calibrated using the esi spectrum of .
we also made a spectrum , hh , which we calibrated using a hires spectrum of a standard star .
this hh calibration is not typical of the accuracy that we usually obtain , but rather it is the best that we could obtain with our spectra .
we made many calibrations using different hires integrations of standard stars , and we show the calibration that had the smoothest order joins .
this hh calibration contrasts with the calibrations of the stars that we showed in fig .
[ x16 ] that did not work as well . in figure [ x15 ]
, we show the ratios of hires spectra calibrated in these different ways .
the top panel shows hk99 spectra divided by hesi .
there is a 7% difference across the wavelength region shown .
the hk01 differs from hesi by 5% at most , while the hksum calibration differs from hesi by 4% .
the calibrations hh and hesi differ the least only 3% but show the largest jumps at order joins , e.g. near 4325 .
the four gray bands of figure [ x15 ] show the wavelength overlaps between the hires orders .
we show three complete orders and two partial ones .
we show the wavelength region centered on the strong ly@xmath6 lines that we have use to measure the h i column density that contributes to a d / h measurement .
this ly@xmath6 line has damping wings that absorb approximately 1% of the flux near 4233 and 4340 and hence accurate flux calibration of this region helps us measure the column density , although most information comes from 20 on either side of the line center .
the differences between the calibrations have three origins .
the largest differences , which are on the largest scales , come from the differences between the k99 , k01 and esi spectra .
other differences , especially near the order joins , are related to the fitting of the cr , and are sensitive to the signal to noise ratio in the kast and esi spectra .
a third type of difference occurs from pixel to pixel , and comes from the rejection of deviant pixels from among the 8 hires integrations .
the cr fits are smooth curves , and the ratios would also be smooth were it not for differences in the pixels which are rejected when we took the sum of the 8 integrations .
the numerous small 1 2% deviations arise when a pixel is not used from one integration or another .
the size of these deviations is given by the signal to noise ratio of the individual integrations .
the noise increases in the strong absorption lines where we are dividing two hires spectra , each of which has low signal to noise noise .
trends that are seen in more than one panel may come from differences of the esi spectrum from the others .
we do not know which of the spectra is the more accurate .
all of the spectra used in figure [ x15 ] were calibrated with cr spectra that were fit order by order .
we found that the 2-dimensional fits to the cr , like that shown in figure [ twodfit ] , slightly reduced the deviations near the order joins .
many factors contribute to the errors in the relative flux calibration , including : * errors in the flux reported for the standard star .
* errors in our spectra of the standard star and reference .
common error sources include extinction and absorption in the earth s atmosphere , slit losses that depend on wavelength , a dichroic in the spectrograph , variation of the target , variation with wavelength of the proportion of the flux extracted , and the s / n of the spectra .
* errors calculating the response function that we use to calibrate the reference spectrum .
errors occur matching the resolution and wavelength scales of our spectrum of the standard to the published flux information .
such errors are especially conspicuous near absorption lines in the standard .
* errors in the preparation of the hires spectrum , including the bias subtraction , flat field division and extraction .
* errors applying the flux calibration to hires spectrum , including wavelength shifts , resolution differences , the signal to noise ratio of the hires spectrum and fitting the cr . we have found that many of these factors can produce 1 10% errors in flux calibration , but it is difficult to assign typical values for these errors .
the size of the error in the relative flux calibration depends on the wavelength and the wavelength range .
we do not include errors from the cr in the usual error array because the cr errors are correlated over many pixels . in this paper
we have concentrated on scales of a few orders , or approximately 120 that are most relevant to our work on d / h . we have paid much less attention to the relative flux calibration on larger scales that are dominated by a different set of factors , such as the extinction at the time of observation .
we expect errors due to extinction to monotonically increase with scale in our reference spectrum .
an indication of the accuracy we have attained in relative flux calibration is given in figure [ f344thord ] .
we compare three spectra of the flux standard star feige 34 : a hires spectrum that we flux calibrated using a kast spectrum as the reference , the kast reference spectrum , and stis a spectrum .
the kast and stis spectra are those shown in figure [ x23 ] .
the stis spectrum was not used in the calibration , except to provide the normalization of the kast spectrum across the range 3200 4450 .
we used a one - dimensional 4th order polynomial to fit the cr . in the wavelength region
shown , the hires spectrum differs from its reference by at most 2.5% , except near the absorption line , and typically @xmath0% . at wavelengths
3400 3800 , where the signal to noise ratio is lower , the differences are twice as large .
the differences correlate on scales @xmath33 .
the flux in different hires orders joins smoothly , with no unusual structure .
the remaining differences of the hires and stis spectra come from the deviation of the kast spectrum from the stis , shown in figure [ f344thord ] .
this comparison demonstrates that the method can give errors of @xmath0% in the relative flux over approximately 200 . for feige 34
the accuracy of the flux calibration was limited by the accuracy of the reference , and not by the method itself .
we found that the distribution of the signal recorded in hires integrations differs from integration to integration .
we do not have a complete explanation for this behavior , although varying vignetting and inadequate extraction may be involved .
we found that these differences persist even when the instrument is apparently unchanged .
these changes mean that the usual methods of flux calibration are inadequate .
the methods we have described for applying relative flux calibration to a hires spectrum use three spectra : the hires spectrum of the target that we wish to calibrate , and kast spectra of both the target and a flux standard star .
we use the latter to get the kast response and calibrate the kast spectrum of the target that we name the reference spectrum .
we use reference spectrum to impose flux calibration on the hires target spectrum .
this has method has three advantages .
first , we can calibrate hires when normal calibrations using standard stars observed with hires alone are inadequate .
second , we can correct many types of error in the hires spectrum , including those from varying vignetting and inadequate extraction .
finally , we can obtain all the calibration spectra at a different time and on a different telescope .
the error in relative flux calibration , and the solution that we describe , could apply to any spectrum with inadequate relative flux calibration , whether from an unstable spectrograph or from inadequate extraction .
vignetting could vary in any spectrograph that was unstable .
instability could involve an optical misalignment , as with hires .
variable vignetting would be harder to recognize in a first order spectrum because we expect the largest changes near the largest field angles , but there is only one spectrum to show this change , and flux calibration is often harder near the ends of a single spectrum , for other reasons .
this work was funded in part by grant nag5 - 9224 and by nsf grant ast-9900842 and ast-0098731 .
the spectra were obtained from the lick and the w.m .
keck observatories . the w.m .
keck observatory is managed by a partnership among the university of california , caltech and nasa .
we are grateful to steve vogt , the pi for the keck hires instrument that enabled our work on d / h , for many invaluable discussions , along with tom barlow , who developed the makee extraction software package .
we also thank wako aoki and kunio noguchi for discussions of the hds echelle spectrograph , and the w.m .
keck observatory staff and the lick observatory staff .
we used the calibrations of the flux in stars based on stis spectra @xcite . in figure [ x4 ]
we see that the stis spectra do not show the wiggles at 3200 - 3850 that are present in the @xcite spectrum of g191 b2b .
this and other oke spectra are widely used by default in the reduction packages iraf and midas .
the differences between the @xcite and stis spectra can reach several percent , larger than our random photon noise . based on the lack of features in the spectrum and the stis data quality , we preferred the following stars for uv flux calibration near 3200 : g191 b2b , bd+28 4211 , feige 110 , and feige 34 . for g191 b2b
we have the additional choice of using a model spectrum given by @xcite .
this model spectrum fits their stis spectra to within 0.7 % in the continuum @xcite , and it simplifies flux calibration because it has full resolution .
we used the entire stis spectrum for flux calibration , including the broad balmer absorption lines .
the @xcite paper provides ab magnitudes at discrete continuum points in 5 - 50 intervals .
these points skip the balmer absorption lines , but we can not do this , because we then have insufficient information to calibrate several echelle orders , each of which is only 30 - 60 long .
our use of the flux calibration information near the balmer lines can help us avoid significant errors . in figure [ x5 ]
we show a spectrum from esi echellette order 15 that has its sensitivity peak around 4350 that coincides with balmer @xmath34 line , 4341.68 .
@xcite happens to have its ly@xmath6 emission at 4330 . in an early flux calibration of this order , poor interpolation across this balmer line
had led us to make an 8% calibration error that was three times the random error .
the vignetting in hires depends on whether or not the image rotator is used , and on the mode in which the rotator is used .
hires is fixed to the nasmyth platform of the keck - i telescope with its slit approximately parallel to the horizon .
when we look at the image of the sky on the hires slit , we see that the vertical direction in this image rotates at a rate given by the elevation ( el ) of the telescope .
this is because telescope is rotating in el while hires is fixed . if the telescope is pointing at the horizon , and looking at an arrow in the sky that is pointing towards the zenith , the image of this arrow on the slit plane is also pointing towards the vertical , which is perpendicular to the length of the slit .
as the telescope moves to higher el , the arrow rotates until it is aligned along the slit when the telescope is pointing to the zenith .
the hires image rotator allows us to rotate the image of the sky on the slit plane in any way we like .
we installed an image rotator in hires in late 1996 .
this is a large quartz prism that sits in front of the hires slit .
the light from the keck - i telescope tertiary mirror undergoes three total internal reflections in the prism before coming to a focus on the slit plane .
the prism can rotate continuously in either direction about the axis of the beam that converges on the center of the slit .
the prism is aligned so that the image of a star on the center of the slit moves by under 0.5 arc seconds when the prism is moved in or out of the beam , and when the prism is rotated .
the prism can be spun rapidly to demonstrate this alignment .
the prism does not vignette any of the beam that lands within 60 arc seconds of the center of the slit .
the image rotator has two main modes of operation : position angle and vertical angle .
the position angle mode is used when we wish to keep two stars in the slit , where as the vertical angle mode is used to keep the vertical direction in the sky parallel with the slit , as a surrogate for an atmospheric dispersion compensator . when the image rotator was used in vertical mode
, the position angle along the slit is the parallactic angle , and this varies as we track a target .
the parallactic angle is measured at the target , from the north celestial pole to the zenith , in the direction from north via east .
the parallactic angle is fixed for a given elevation and azimuth in the sky , but it changes when we track a target across the sky .
hires spectra are hard to flux calibrate in part because the vignetting can change by 10% from spectrum to spectrum .
the vignetting changes because there is a known misalignment between the beam coming from tertiary mirror and the hires optical axis .
when hires was installed , the center of the keck telescope pupil was measured to be approximately 9 mm away from the collimator center , which corresponds to a beam misalignment of 7.4 arcminute .
if a star is held at one position on the hires slit , the axis of the beam entering hires will rotate around the hires optical axis at a rate given by any change in the position angle of the sky image on the slit .
if the position angle moves through 360 degrees , the axis of the beam entering hires follows the surface of a cone with an apex angle of approximately 14.8 arcminutes .
steve vogt used ray tracing to find that this rotation causes the vignetting to vary by approximately 10% , depending on the angle . the vignetting occurs due to a dewar which forms a central obstruction in the beam near the camera s prime focus .
, h. , dodorico , s. , kaufer , a. , delabre , b. , & kotzlowski , h. 2000 , in proc .
4008 , p. 534 - 545 ,
optical and ir telescope instrumentation and detectors , masanori iye ; alan f. moorwood ; eds .
4008 , 534545 , f. , charalambous , a. , fish , a. c. , & walker , d. d. 1990 , in instrumentation in astronomy vii ; proceedings of the meeting , tucson , az , feb .
13 - 17 , 1990 ( a91 - 29601 11 - 35 ) .
bellingham , wa , society of photo - optical instrumentation engineers , 1990 , p. 562 - 576 .
anglo - australian observatory - supported research .
, vol . 1235 , 562576 , i. s. , becklin , e. e. , bendiksen , o. , brims , g. , canfield , j. , figer , d. f. , graham , j. r. , hare , j. , lacayanga , f. , larkin , j. e. , larson , s. b. , levenson , n. , magnone , n. , teplitz , h. , & wong , w. 1998 , in proc .
3354 , p. 566 - 578 , infrared astronomical instrumentation , albert m. fowler ; ed .
3354 , 566578 , a. i. , miller , j. s. , bolte , m. , & sutin , b. m. 2000 , in proc .
4008 , p. 522 - 533 ,
optical and ir telescope instrumentation and detectors , masanori iye ; alan f. moorwood ; eds .
4008 , 522533 , s. s. , allen , s. l. , bigelow , b. c. , bresee , l. , brown , b. , cantrall , t. , conrad , a. , couture , m. , delaney , c. , epps , h. w. , hilyard , d. , hilyard , d. f. , horn , e. , jern , n. , kanto , d. , keane , m. j. , kibrick , r. i. , lewis , j. w. , osborne , j. , pardeilhan , g. h. , pfister , t. , ricketts , t. , robinson , l. b. , stover , r. j. , tucker , d. , ward , j. , & wei , m. z. 1994 , in proc .
spie instrumentation in astronomy viii , david l. crawford ; eric r. craine ; eds . ,
volume 2198 , p. 362
2198 , 362 + , divided by average ratio from the ten orders with the largest wavelengths .
we show this average in the top left panel , and other panels each show one hires order .
we show the full amount of each order that fell on the ccd , which includes some wavelength overlap between adjacent orders .
we have re - binned the original 2048 pixels into 205 pixels for presentation .
[ x22c ] ] in the middle 20% of each hires order .
prior to this , the cr varied systematically by approximately a factor of two .
the thick curves show the chebyshev polynomial along each order .
these polynomials are constrained to 4th order in both the ccd pixel and echelle order directions .
[ twodfit ] ] | we describe a new method to calibrate the relative flux levels in spectra from the hires echelle spectrograph on the keck - i telescope .
standard data reduction techniques that transfer the instrument response between hires integrations leave errors in the flux of 5 10% , because the effective response varies .
the flux errors are most severe near the ends of each spectral order , where there can be discontinuous jumps .
the source of these errors is uncertain , but may include changes in the vignetting connected to the optical alignment .
our new flux calibration method uses a calibrated reference spectrum of each target to calibrate individual hires integrations .
we determine the instrument response independently for each integration , and hence we avoid the need to transfer the instrument response between hires integrations .
the procedure can be applied to any hires spectrum , or any other spectrum .
while the accuracy of the method depends upon many factors , we have been able to flux calibrate a hires spectrum to 1% over scales of 200 that include order joins .
we illustrate the method with spectra of towards which we have measured the deuterium to hydrogen abundance ratio . #
1[nb : _ # 1 _ ] # 1[dk : _ # 1 _ ] # 1[dt : _ # 1 _ ] # 1[jo : _ # 1 _ ] # 1[ns : _ # 1 _ ] |
the study of gravitational lenses has become a powerful tool to address several distinct cosmological and astrophysical questions .
these include the distribution of dark matter in galaxies ( keeton et al .
@xcite ) , studying dust extinction at redshift @xmath3 ( nadeau et al . @xcite ; jean & surdej @xcite ) , and determining the hubble parameter @xmath4 from measurements of light travel time delay between different lines of sight ( refsdal @xcite ) .
the appeal of this method lies in its complete independence of the traditional cosmic distance ladder , yielding distances for high - redshift objects in a single leap .
the uncertainties of @xmath4 estimation are mainly limited by ambiguities in modelling the deflector mass distribution , since all relevant _ measurement _ errors can be reduced to insignificance . while ` simple ' lenses such as double qsos do not strongly constrain the deflector model , multiple systems such as quadruply imaged qsos provide many additional constraints , allowing e.g. to independently test the assumed mass distribution models ( cf .
saha & williams @xcite ) . a problem with quadruple systems , however , can be excessive symmetry leading to time delays which are short compared to intrinsic radio and optical variability timescales , as in the famous ` einstein cross ' 2237@xmath5030 or in the ` clover leaf ' h 1413@xmath5117 .
in this paper we present the discovery of the new multiple qso he 0230@xmath02130 .
the object was originally identified as a high - probability qso candidate in the course of the hamburg / eso survey ( wisotzki et al .
@xcite ) and subsequently observed as part of a large imaging search for lensed qsos .
the coordinates of this object are r.a .
= 02@xmath6 32@xmath7 33@xmath8 , dec = @xmath9 17@xmath10 26@xmath11 ( j2000.0 ) , as measured in the _ digitized sky survey_. ( 180,44 ) ( 0,0)=4.35 cm ( 4,38)@xmath12 ( 45,0)=4.35 cm ( 49,38)@xmath13 ( 90,0)=4.35 cm ( 94,38)@xmath14 ( 135,0)=4.35 cm ( 139,38)@xmath15
imaging data were obtained at the ctio 1.5 m on 13 november 1998 , in johnson @xmath12 and @xmath13 , and cousins @xmath14 . in each band , three images were taken of 5min exposure time each .
the seeing was somewhat variable : @xmath16 in @xmath12 , @xmath17 in @xmath13 , and @xmath18 in @xmath14 .
pixel size was @xmath19 .
small sections of these images are reproduced in fig .
[ fig : brik ] , revealing the multiple structure of the qso already at first glance .
two almost merging images ( a1 and a2 ; see fig .
[ fig : pos ] ) of nearly equal brightness dominate the total magnitudes .
another discrete component , b , is clearly apparent with roughly similar colours .
components c and d are much less prominent , and in fact appear as two distinct sources rather than just one only in @xmath14 . on the same night
, near - infrared ( @xmath15 band , 2.02.3@xmath20 m ) images were taken at the ctio 4 m telescope with the cirim imager .
total exposure time was 18min under @xmath21 seeing .
the pixel size of the raw data was @xmath22 , but the single exposures were combined on a @xmath23 finer grid ; the resulting image is also shown in fig . [ fig : brik ] .
the centrally located component d is now almost as bright as a1 and a2 and unambiguously separated from c. a quantitative analysis of the astrometric and photometric properties is given in the next section . at the time of these observations
, he 0230@xmath02130 had been a mere qso candidate based on a digital objective prism spectrum , with an estimated redshift of @xmath24 .
although we had little doubt about its qso nature , confirmation was clearly needed .
we obtained a first low - resolution spectrum on 23 november 1998 with the eso / danish 1.5 m telescope equipped with dfosc .
a slit of @xmath25 positioned east - west gave a spectral resolution of @xmath26 fwhm .
the rather poor seeing of @xmath27 inhibited all attempts to separate different components ; and the spectrum , displayed in fig .
[ fig : spcdanish ] , represents more or less the superposition of all components .
it shows a typical qso with @xmath28 , as measured from the centroid of the mgii emission line .
an attempt to obtain individual spectra of different components was made at the eso 3.5 m ntt on 10 december 1998 , using the red arm of emmi with grism # 3 and a @xmath16 slit , at 8 resolution .
only one spectrum could be taken , with the slit oriented north - south crossing a2 and b. at a seeing of @xmath29 , the components were well separated , showing two qso spectra at the same redshifts .
unfortunately , the spectrum of b suffers from substantial slit losses : while the flux ratio a2/b is 2.3 in the red , in good agreement with the psf photometry described below , it increases to @xmath30 at the blue end around 4000 . to correct for these losses ,
the quotient spectrum was fitted with a 5th - order polynomial , and the spectrum of b was multiplied by the fit , thereby adjusting both spectra to the same global level of relative fluxes .
the result is displayed in fig .
[ fig : spcntt ] and discussed below . = 8.8 cm
to decompose the complex configuration into discrete sources , we employed the daophot ii software package as implemented within eso - midas . for the optical images ,
a numerical composite psf was built for each image using three bright stars at @xmath31 distance to the qso . in the @xmath32 band data no bright psf star was available within the small field of view , and
a purely analytical psf was estimated from the only two isolated stars in the field .
the qso was then modelled by the superposition of five point sources , with positions and psf scaling factors simultaneously optimised ( routine allstar ) .
fitting five components gave always a significantly better fit , in terms of residual @xmath33 , than with four , even in the @xmath12 band image where c and d are both very faint .
the fitting results are collected in tables [ tab : pos ] and [ tab : phot ] .
positions are measured relative to a1 , which was arbitrarily taken as reference point . for components a1 , a2 , and b , the positions measured in the four _ brik _ images are very consistent , and the scatter between the photometric bands reflects the measurement error . for c and d ,
image positions are consistent only between @xmath14 and @xmath32 ; at shorter wavelengths , the fitted centroids approach each other as illustrated in fig .
[ fig : pos ] .
this could be related to different intrinsic colours of the objects , but could also be an artefact of low s / n and the somewhat poorer seeing in the shorter wavelength data .
differential psf photometry of the qso components relative to a1 was available from the allstar analysis ( tab .
[ tab : phot ] ) .
flux calibration was established separately using simple aperture photometry ( aperture diameter was @xmath34 for _ bri _ and @xmath35 for @xmath15 ) .
standard stars from the ru149 field ( landolt @xcite ) served to obtain colour terms and photometric zeropoints in the optical , while sj9106 ( persson et al .
@xcite ) was used for the nir photometry .
the total magnitudes thus measured are listed in the first line of table [ tab : phot ] , with uncertainty estimates as given in the daophot output .
we have also determined aperture _ bri _ magnitudes for 14 nearby stars in the field that may be useful to serve as reference stars in future monitoring .
a list with these measurements is available on request .
the relative photometry confirms that a1 , a2 , and b have very similar optical - nir colours although b appears slightly redder than a1 and a2 .
c and d are much redder , on the other hand , so neither can correspond to a single unobscured fourth qso image .
because of the apparent positional shift between the bands , we computed a second model with fixed positions imposed from the @xmath14 band image , fitting only the psf scaling factors .
the resulting colours are slightly bluer for components b and c , and even much redder for d. however , inspection of the psf - subtracted images indicated that the fit quality of these restricted models was much poorer , leaving residuals significant on the 23@xmath36 level .
= 8.8 cm [email protected]@.l component & & + & & + a1 & 0&00 & 0&00 + a2 & @xmath00&68 @xmath37 0.01 & @xmath00&27 @xmath37 0.01 + b & @xmath01&17 @xmath37 0.02 & @xmath01&80 @xmath37 0.02 + c & 0&28 @xmath37 0.04 & @xmath01&98 @xmath37 0.04 + d & 0&01 @xmath37 0.02 & @xmath01&13 @xmath37 0.03 + [email protected]@[email protected]@.l component & & & & + total & 18&20 & 17&66 & 17&18 & 14&96 + & @xmath38&01 & @xmath38&01 & @xmath38&01 & @xmath38&02 + a1 & 19&27 & 18&83 & 18&49 & 16&54 + & @xmath38&02 & @xmath38&02 & @xmath38&02 & @xmath38&04 + a2 & 19&27 & 18&89 & 18&43 & 16&65 + & @xmath38&02 & @xmath38&02 & @xmath38&02 & @xmath38&04 + b & 20&06 & 19&45 & 19&00 & 16&95 + & @xmath38&02 & @xmath38&02 & @xmath38&02 & @xmath38&04 + c & 21&87 & 20&42 & 19&74 & 16&80 + & @xmath38&08 & @xmath38&04 & @xmath38&05 & @xmath38&04 + d & 22&02 & 20&62 & 19&63 & 16&65 + & @xmath38&24 & @xmath38&05 & @xmath38&04 & @xmath38&05 + [email protected]@[email protected] component & & & + a1 & 0&44 & 0&34 & 1&95 + a2 & 0&38 & 0&45 & 1&78 + b & 0&62 & 0&45 & 2&05 + c & 1&45 & 0&68 & 2&94 + d & 1&40 & 0&98 & 2&98 + while in double qsos there is always the possibility that a true binary system is being observed , a configuration like that seen in he 0230@xmath02130 is almost certainly best explained as a lensed system , even without spectroscopic evidence . although we do not yet have spectra of all components , the available data allow nevertheless to confirm the lens hypothesis beyond all reasonable doubt : \(1 ) the total spectrum ( fig . [ fig : spcdanish ] ) contains no trace of absorption features that would be expected if a1 was a star or a galaxy .
we conclude that a1 and a2 have most probably very similar spectra , given the broad - band colours .
\(2 ) a2 and b have both very similar qso spectra , apart from the slit loss effects .
figure [ fig : spcntt ] shows that the emission line centroids agree within the measurement accuracy , the line widths are equal , and also the strong ` associated ' ( @xmath39 ) civ absorption system is clearly present in both components . a curious feature , however , are the significant residuals detected in the difference spectrum a2b(scaled ) , indicating non - identical emission line profiles and/or equivalent widths . whether this might be due to differences in intervening line absorption along the lines of sight , or due to selective continuum microlensing such as proposed in he 1104@xmath01805 ( wisotzki et al .
@xcite ) remains to be explored ; the current data do not permit a more detailed analysis .
= 8.8 cm visual inspection of the available images ( see fig .
[ fig : galenv ] ) shows a number of faint galaxies in the vicinity of the qso , but most of these objects are very faint and at the limit of the present data ( @xmath40 , @xmath41 , @xmath42 ) .
the surface density of faint objects around the qso ( within @xmath43 radius ) was found to be enhanced by a factor of 23 compared to the surrounding regions . as star - galaxy separation was not reliable for such faint sources , every detected object was counted , so the overdensity is a conservative estimate .
however , the signal is significant only on the 2@xmath36 level , using poisson statistics .
most of the @xmath44 objects within @xmath45 from the qso are detected in _
( and expectedly not in @xmath12 ) .
the @xmath46 and @xmath47 colours show relatively little dispersion and are consistent with galaxy colours at low to intermediate redshifts , @xmath48 ( cf .
fukugita et al .
we conclude that there is some evidence for a cluster near the line of sight to he 0230@xmath02130 , but that this has to be confirmed by deeper imaging and spectroscopy .
( 88,44 ) ( 0,0)=4.4 cm ( 3,39)@xmath13 ( 44,0)=4.4 cm ( 47,39)@xmath15
components a1 , a2 , and b of he 0230@xmath02130 have been clearly identified , by multicolour imaging and spectroscopically , as being images of one and the same qso .
the identities of components c and d have not yet been ascertained . from their locations ,
a reasonable guess might be that c is the expected fourth image and d the lensing galaxy .
the colours of d are certainly consistent with this notion , but the case is more complicated with component c. using only positions as constraints , simple models which take c to be the 4th image give large residuals , whereas models which fit only the positions of images a1 , a2 and b are poorly constrained
. moreover , the colours colours of c are very different from the confirmed qso components .
this might indicate either heavy reddening by dust in the fourth image , or the presence of a second galaxy , or even superposition of a galaxy _ plus _ fourth image . without spectroscopic information on c
we presently can not decide between these possibilities .
if c is partly or largely due to a galaxy , it is _ a priori _ most likely ( but not granted ) that c and d are at the same redshifts , forming one common deflector plane .
if , on the other hand , c is formed by just the fourth quasar image , this would be one of the few known lenses with strongly differential dust extinction , and the first optically selected one .
furthermore , the inferred amount of differential extinction would be @xmath49 , just intermediate between the very small and highly uncertain values found for most lenses and the few cases where the light path in the optical is strongly obscured ( cf .
falco et al .
@xcite ) .
he 0230@xmath02130 is an interesting new target for photometric monitoring .
it is moderately symmetric so that the time delay is probably much shorter than a year , avoiding the considerable windowing problem of observations stretching over several years . yet
, it displays enough departure from symmetry that the time delay will not be uncomfortably low . in many respects ,
it resembles pg 1115@xmath5080 , the presently cleanest system for determination of @xmath4 ( e.g. , schechter et al .
@xcite ) . the apparent magnitude and colour of the putative lensing galaxy suggest that its redshift is not too high , @xmath50 .
falco e.e . ,
impey c.d . ,
kochanek c.s . , et al . , 1999 , apj submitted , astro - ph/9901037 fukugita m. , shimasaku k. , ichikawa t. , 1995 , pasp 107 , 945 jean c. , surdej j. , 1998 , a&a 339 , 729 keeton c.r .
, kochanek c.s . ,
falco e.e . , 1998 ,
apj 509 , 561 landolt a.u . , 1992 , aj 104 , 340 nadeau d. , yee h.k.c .
, forrest w.j . , et al . , 1991 , apj 376 , 430 persson s.e . , 1998 ,
aj 116 , 2475 refsdal s. , 1964 , mnras 128 , 295 saha p. , williams l. , 1997 , mnras 292 , 148 schechter p.l .
, bailyn c.d . ,
barr r. , et al .
, 1997 , apj 475 , l85 wisotzki l. , khler , t. , kayser r. , reimers d. , 1993 , a&a 278 , l15 wisotzki l. , khler t. , groote d. , reimers d. , 1996 , a&as 115 , 227 | we report the discovery of the new gravitational lens system he 0230@xmath02130 , a qso at redshift @xmath1 consisting of at least five distinct components .
three of these are clearly lensed images of the qso , one is most likely the lensing galaxy , while for the fifth component the identity is unclear : it could be a fourth qso image ( if so , then highly reddened ) , or another intervening galaxy , or a superposition of the two .
differential reddening seems to be important also for the first three qso images .
the surface density of faint galaxies near the qso appears to be enhanced by a factor of @xmath2 , indicating the presence of a distant cluster close to the line of sight . |
we like to think of the simplest multi - quark system , consisting of four quarks , as a bridge between a single hadron and interacting multi - hadron systems .
earlier , simulations of static four - quark systems have been performed for a general set of geometries ( * ? ? ?
* ; * ? ? ?
* references therein ) . these quenched su(2 )
calculations have produced binding energies up to 120 mev . in order to understand these results
a phenomenological model based on two - body potentials and multi - quark interaction terms has been developed .
the basis states of the model come from the three possible ways to pair four quarks into two mesons in su(2 ) .
these pairing are taken to be in the ground state ( @xmath3 ) or first excited state ( @xmath4 ) of the two - body potential , giving six states ( @xmath5 ) in total .
the basic equation of the model is @xmath6=0,\ ] ] where @xmath7 contains overlaps of any two states ( @xmath8 ) , @xmath9 has the interactions ( @xmath10 ) and @xmath11 is the four - quark energy .
the binding energy is defined as @xmath12 , i.e. as the difference of the four - quark energy and the energy of the lowest - lying two - body pairing .
a central element in the model is a multi - quark interaction term @xmath13 , defined as @xmath14 . for small distances @xmath15 and for large @xmath16 .
the potentials @xmath17 have a perturbative one - gluon exchange prefactor and lattice two - body potentials @xmath18 , giving the form @xmath19
. this expression alone would lead to unphysical van der waals forces , which are removed by the @xmath13 .
the interaction between basis states @xmath20 and @xmath21 is thus @xmath22 .
the factor @xmath13 can be parameterized as @xmath23 , where @xmath24 is the string tension and @xmath25 the minimum area bounded by the four quarks .
a simple version of the model with just states @xmath26 and @xmath15 is reproduced by perturbation theory to @xmath27@xcite .
for quarks at the corners of a regular tetrahedron binding increases with size , which suggests two - body excitations are relevant , as they are closer to the ground state for larger distances .
therefore , we have included excited basis states in our model .
this gives new @xmath13-factors @xmath28 , the former of which is parametrized like @xmath13 and the latter as @xmath29 .
the interaction matrix elements involving two - body excitations are parameterized in a manner similar to the ground state matrix elements , e.g. @xmath30 $ ] . when this model is fitted to 100 ground and excited state four - quark energies it turns out that @xmath31 is always consistent with one , suggesting that two - body excitations interact in a perturbative manner .
we are left with four independent parameters @xmath32 that fit the energies with @xmath33 .
the values of these parameters are stable when energies from smaller lattice spacings are used .
if the excited basis states are left out we get @xmath34 , which mostly comes from the maximally degenerate regular tetrahedra .
the flux distribution is measured in quenched su(2 ) using @xmath35,\ ] ] where @xmath36 is the wilson loop , or more generally , a combination of wilson loops corresponding to the ground or excited state of the multi - quark system . since extracting a signal
is not easy we use link - integration on the wilson loops . as in the calculation of energies ,
we utilize a variational approach in order to extract the ground or an excited state of the system .
the variational basis is formed of wilson loops with various levels of fuzzing .
we can check the accuracy of our flux measurement with lattice sum rules .
these sum rules relate spatial sums over fields to energies via @xmath37-functions such as @xmath38 , e.g. @xmath39 . \\\ ] ] using our earlier two - quark results @xcite for @xmath40 this relation allows us to compare the sum over the distribution to a much more accurately measured energy .
this is especially convenient when the binding is considered ; @xmath41 is a lattice artefact from quark self - energies , and is removed for the binding energy .
three - dimensional pictures of our results ( fig .
1 ) can be viewed at www.physics.helsinki.fi/~ppennane / pics/. the binding distribution is found to have approximately constant height in between the quarks , whereas the first excited state has a `` cloverleaf '' symmetry with sign changes . replacing energies with flux distributions in our model without excitated basis states
gives @xmath42 , which agrees with fits to energies @xcite .
when excited basis states are included the ground - state interaction has shorter range ; @xmath43 , suggesting that ground state potentials are important at small distances and excited ones at larger distances .
we calculate the potential between two heavy - light mesons as a function of the heavy quark separation @xmath44 in quenched su(3 ) at @xmath45 .
the heavy quarks are taken to be static and the light ones have approximately the @xmath46 mass ( @xmath47 ) with the propagators obtained as stochastic pseudofermionic estimates with maximal variance reduction @xcite .
the principal advantage of the stochastic estimates is the ability to get propagators to and from most points on the lattice ; hence the nickname `` all - to - all '' . at @xmath48 the meson - antimeson ( @xmath49 ) system
should have a mass similar to a pion , while the @xmath50 system should resemble a heavy - light baryon ; the former is indeed observed .
the preliminary results in fig .
[ fb4 ] do not show the one-@xmath51 exchange effect of deuson models which leads to differences in the @xmath50 and @xmath49 potentials at large distances . as expected , at small distances
the @xmath49 potential is more attractive .
as discussed in other papers in these proceedings @xcite , the heavy - light meson - antimeson system is relevant to string breaking .
unfortunately calculations of this system at large enough distances are very hard , but the all - to - all propagator scheme offers a way to overcome the difficulties . in our case
the @xmath49 and static @xmath52 energies become equal at @xmath53 fm @xcite ; this is reached by our ongoing calculations on a @xmath54 lattice .
a variational operator mixing analysis @xcite similar to that presented by philipsen in a higgs model is then performed . by running our code on ukqcd unquenched configurations we hope to shed light on the differences due to sea quarks in the mixing region .
a. m. green and p. pennanen , phys
b426 * , 243 ( 1998 ) , http://xxx.lanl.gov/abs/hep-lat/9709124[hep-lat/9709124 ] ; phys . rev . *
c57 * , 3384 ( 1998 ) , http://xxx.lanl.gov/abs/hep-lat/9804003[hep-lat/9804003 ] . | in order to understand the binding of four static quarks , flux distributions corresponding to these binding energies are studied in quenched su(2 ) and also related to a model for the energies .
the potential relevant to string breaking between two heavy - light mesons is measured in quenched su(3 ) using stochastic estimates of light quark propagators with the sheikholeslami - wohlert action .
epsf our goal is to understand multi - quark interactions from first principles . since `` multi - quark '' means that the system can be decomposed into more than one color singlet , the simplest such system consists of four quarks . in this work
we try to understand this simple case , relevant for meson - meson interactions , hoping that generalization to more complicated cases will be relatively straightforward .
we first give a brief introduction to a model that reproduces one hundred ground- and excited state energies of four static quarks @xcite .
these energies have been calculated earlier in su(2 ) gauge theory on @xmath0 lattices with @xmath1 fm a continuum extrapolation has been performed @xcite . the flux distribution producing four - quark
binding has been measured @xcite and we attempt to relate it to our model .
preliminary results are reported from a more realistic calculation concerning the potential between two heavy - light mesons in quenched su(3 ) .
the heavy quarks are taken to be static and the light quark propagators are calculated using stochastic estimators with an @xmath2 improved fermion action .
this work also concerns string breaking . |
in this review , i will try to describe what galactic chemical evolution models tell us about the evolution and the primordial abundances of d and @xmath0he , and what , in turn , d and @xmath0he may tell us about stellar and galactic evolution .
in particular , i wish to emphasize that the light elements should not be treated separately , but should always be considered together with the other more diffuse elements , to better constrain their evolution .
the reason why galactic chemical evolution models are required to derive the primordial d and @xmath0he abundances from the observed ones is that all the objects where the two elements are measurable are relatively young ( the oldest being the sun with an age of 4.5 gyr ) and have therefore formed , with the only exception of high - redshift clouds , from an ism whose chemical composition had been modified by the previous stellar generations . to infer the primordial abundances from these measurements
, it is thus necessary to take into account the effects of the various cycles of gas astration and gas return , and the variations of the ism chemical composition due to stellar nucleosynthesis and gas flows occurring up to the time when the observed objects have formed .
this is accomplished by chemical evolution models .
d and @xmath0he are obviously related to each other , since all the d which enters a star is immediately burnt into @xmath0he ( reeves et al . 1973 ) .
however , the problems faced when studying their galactic evolution are quite different , and i will thus treat them separately in this paper .
since d is completely destroyed inside stars already in pre - main sequence phase , if we consider the big bang nucleosynthesis as the only source of d , the amount of d which can be present in any galactic region and at any place is that contained in gas which has never been through stars . in other words ,
the fraction of primordial d surviving at any epoch and in any region is equal to the fraction of virgin gas there .
hence , in principle , to infer the primordial abundance of d from its present one it would be sufficient to know the current fraction of gas which has not entered a star yet ( steigman & tosi 1995 ) . unfortunately , we do not know the fraction of pristine gas even in the most local medium and we must therefore rely on galactic chemical evolution models to derive the d evolution . historically , there are two schools of thought on how to proceed in studying the d evolution , as sketched in fig.1 . the first one , in chronological order , can be referred to as the _
cosmological school_. the approach of this school is to start from standard big bang nucleosynthesis ( sbbn ) prescriptions , select the observational constraints on d which can be considered reliable , and infer from these sets of data what the d evolution must have been in the galaxy .
they then build models of galaxy evolution able to reproduce the inferred trend of d vs time , and the predictions of such models on the other galactic quantities are a by - product .
the other school , which i will call _ galactic _ , follows the opposite approach .
we start from chemical evolution models of our galaxy , select only those which are able to reproduce the largest set of observational constraints , and take the predictions on d only from these selected models .
the consequences of these predictions on the primordial d abundance and on cosmology are a by - product .
were we living in the best of all possible worlds , the two approaches should provide the same results . instead , their predictions are quite different from each other .
fig.2 shows the d abundances derived from all the available observations and plotted as a function of the supposed epoch of formation of the observed objects .
all the error bars are @xmath3 .
the two vertical bars at t=0 represent the ranges of values derived by songaila et al .
( 1994 , hereinafter schr94 ) and burles & tytler ( 1998 , b&t98 ) from high - redshift , low - metallicity , absorbers on the line of sight of distant qso s .
the bar at 8.5 gyr represents the value of the protosolar cloud inferred by geiss & gloeckler ( 1998 , g&g98 ) from solar system data .
the solid bar at 13 gyr shows the range of abundances derived by linsky ( 1998 , l98 ) for the local ism , while the length of the dotted bar shows the possible cloud - to - cloud variations suggested by vidal - madjar , ferlet & lemoine ( 1998 , v - m98 ) .
since the _ cosmological school _ was founded when the primordial @xmath4he was definitely supposed to be low ( [email protected] in mass fraction ) , and since sbbn predicts the primordial d to be anti - correlated with y@xmath6 , members of this school obviously thought that the only reliable measures of d in almost primeval systems , like high - redshift , low - metallicity absorbers , were those leading to high d abundances .
they thus thought that the natural evolution of deuterium with time is that connecting the schr94 value with the local ones and sketched by the solid line in fig.2 , i.e. a d destruction by one order of magnitude from the primordial to the present abundance .
to obtain such a high d destruction during the galaxy evolution , one must invoke a high star formation rate ( sfr ) , which is usually assumed to occur at the earliest epochs , because all the observational evidences are against high sfr at relatively recent times . these high sfrs ( and their related metal enrichment ) , inevitably imply a large overproduction of the heavy elements with respect to the observed stellar abundances , unless compensated by mechanisms able to reduce the excess of metals , by diluting or removing them from the galaxy .
for this reason , models with high d destruction usually invoke infall of metal poor gas and galactic winds powered by supernovae explosions , sometimes coupled with variations in the initial mass function .
there have been several attempts to find viable galactic models with strong deuterium depletion , but no scenario consistent with all the galactic data has been found .
for instance , in their pioneering work , vangioni - flam & audouze ( 1988 ) concluded that they excessively overproduced the metals , and scully et al .
( 1997 ) , in order to obtain the desired d without overproducing the metals , ended up with a present local sfr at least one order of magnitude lower than observed .
tosi et al .
( 1998 ) have tested all the possible combinations of the various parameters ( sfr , infall , winds , etc . ) and have always found significant inconsistencies in the models with high d destruction : metal overabundance with wrong galactocentric distribution , or metallicity distribution of the g - dwarfs in the solar neighbourhood completely at odds with the observed one , or abundance ratios in halo or disk stars different from the observed ones ( e.g. [ o / fe ] vs [ fe / h ] ) , or sfr inconsistent with the observed range . in no way
have we been able to find a fairly self - consistent model with high d destruction .
the _ galactic school _ works instead on chemical evolution models able to reproduce as well as possible the largest set of observed galactic features .
thanks to the improvements both on the observational and on the theoretical sides , good chemical evolution models of the milky way nowadays can reproduce the average distribution of the following list of observed features ( see e.g. tosi 1996 and 2000 , boissier & prantzos 1999 for references ) : @xmath7 current distribution with galactocentric distance of the sfr , @xmath7 current distribution with galactocentric distance of the gas and star densities , @xmath7 current distribution with galactocentric distance of element abundances as derived from hii regions and from b - stars , @xmath7 distribution with galactocentric distance of element abundances at slightly older epochs , as derived from pne ii , @xmath7 age - metallicity relation not only in the solar neighbourhood but also at other distances from the center , @xmath7 metallicity distribution of g - dwarfs in the solar neighbourhood , @xmath7 local present - day - mass - function ( pdmf ) , @xmath7 relative abundance ratios ( e.g. [ o / fe ] vs [ fe / h ] ) in disk and halo stars . when one compares with each other all the models in better agreement with these data ( e.g. tosi 1996 ) ,
the striking result is that they all predict essentially the same deuterium evolution , in spite of the fact that they are computed by different people , with different assumptions on the input parameters and with different numerical codes .
the bottom panel of fig.2 shows an updated version of the comparison : the plotted models are from galli et al .
( 1995a , short - dashed line ) , dearborn , steigman , & tosi ( 1996 , solid line ) , chiappini & matteucci ( 1996 , long - dashed line ) and boissier & prantzos ( 1999 , dotted line ) .
all the shown curves fit very well the average abundances derived for the local ism , the pre - solar nebula and the high - redshift absorbers by b&t98 .
they all show a fairly moderate ( a factor from 1.5 to 3 , at most ) d destruction during the galaxy lifetime , and therefore suggest that the primordial d abundance should be low : 2@xmath8(d / h)@xmath9 .
this homogeneity of predictions is not a chance effect , but the consequence of the circumstance that all these models fit equally well the observational data on the present sfr , gas and mass densities , and chemical abundances , which necessarily implies that they predict similar fractions of pristine gas and , therefore , of surviving primordial deuterium . our current knowledge on the galactic evolution of d
can thus be summarized as follows : models predicting high deuterium destruction can not account for all the observed galactic properties ; models able to reproduce the largest set of galactic properties all predict low deuterium destruction and , hence , low primordial d.
@xmath0he has a more complex evolution than d , because it is produced not only during the big bang but also inside stars , during the main sequence phase .
this early stellar production may be however largely compensated by further nuclear processing in subsequent phases .
standard stellar nucleosynthesis studies predict that , at the end of the star life , the @xmath0he present in the initial stellar composition is significantly destroyed in massive stars , but preserved or even strongly enhanced in lower mass stars , and that the @xmath0he net yield is a steeply decreasing function of the stellar initial mass , with a large net production in stars below 22.5 m@xmath10 ( see e.g the monothonic curves in fig.3 ) .
this behaviour was known since the late sixties ( e.g. iben 1967 ) , and already in 1976 rood , steigman , & tinsley noticed that it leads to overproduce the solar abundance . only in the mid - nineties , however , with the advent of more detailed combinations of @xmath0he yields with galactic chemical evolution models ( vangioni - flam , olive & prantzos 1994 , galli et al .
1995a , dearborn , steigman & tosi 1996 ) it became apparent that the results on @xmath0he of standard nucleosynthesis studies are definitely inconsistent both with the solar and with the ism observed abundances .
this inconsistency is found with any type of galactic evolution models , including those in agreement with all the other observational constraints ( e.g. tosi 1996 ) and was emphasized by several groups at the elba meeting on the light element abundances in 1994 ( cass & vangioni - flam 1995 , galli et al .
1995b , tosi , steigman & dearborn 1995 ) . in that occasion ,
michel cass concluded with what has been the most popular refrain on @xmath0he ever since : _
@xmath0he delendum est _ , like the city of carthago for the ancient roman m.p .
cato censor .
the most probable solution to the @xmath0he problem is less drastic than that applied to carthago by the romans and was proposed already in 1995 ( charbonnel 1995 , hogan 1995 ) .
it consists in the further @xmath0he processing into heavier elements favoured by an extra - mixing occurring in the red giant phase of low - mass stars ( see both charbonnel and sackman , this volume ) .
when low - mass stars are assumed to experience this extra - mixing and the so - called cool bottom processing ( cbp ) , galactic evolution models do not overproduce @xmath0he anymore and fit well the observed solar and hii region abundances ( tosi 1996 ) .
the question is : in what fraction of low - mass stars cbp should occur to best fit all the data , taking into account that bania , rood et al .
( this volume ) measure in a few pne a high @xmath0he perfectly consistent with the predictions of standard stellar nucleosynthesis ( fig.3 ) ?
galli et al . (
1997 ) showed that the fraction should be larger than 80@xmath2 to fit the @xmath0he abundances observed in the solar system ( geiss & gloeckler 1998 ) , in pne and in hii regions ( rood et al .
1995 and this volume ) .
fig.4 shows the predictions of the best of models tosi-1 ( see tosi 1988 , dearborn et al .
1996 ) when 0% ( dotted line in both panels ) , 90@xmath2 ( solid lines ) , or 100@xmath2 ( short - dash - dotted lines ) of stars with m @xmath11 2.5 m@xmath10 are assumed to follow sackman & boothroyd s ( 1999 ) prescriptions for cbp @xmath0he depletion . for the remaining low - mass stars , as well as for all the intermediate and high - mass ones ,
the @xmath0he yield is taken from dearborn et al .
the dotted , solid and short - dash - dotted lines correspond to models assuming as initial abundances ( d / h)@xmath12 and ( @xmath0he / h)@xmath13 .
the dashed lines show the predictions of the same model with 90@xmath2 cbp , when only the initial d is changed to ( d / h)@xmath14 , while the long - dash - dotted lines correspond to ( d / h)@xmath15 .
the vertical bars in the left hand panel represent the ranges of @xmath0he abundances ( at 2@xmath16 ) derived by geiss & gloeckler ( 1998 ) and gloeckler & geiss ( 1998 ) for the protosolar and the local interstellar clouds , here assumed to be representative of the local ism , 4.5 gyr ago and now , respectively .
the data points in the right hand panel show the @xmath0he abundances derived by rood et al .
( 1995 ) from hii region radio observations .
it is apparent that the models assuming 90@xmath2 and 100@xmath2 of low - mass stars with cbp fit quite well all the data when the initial d is sufficiently low .
the cbp depletion is however insufficient to compensate the @xmath0he overproduction if the initial d / h , subsequently turned into @xmath0he , is higher than a few 10@xmath17 , in which case , first the observed protosolar abundance , and then also the local ism one , can not be reproduced any more .
this is a further argument in favour of the low primordial deuterium resulting from the previous section and from tytler s ( this volume ) discussion of the observations at high redshift .
hence , if ( d / h)@xmath18 , the @xmath0he problem is solved if 90% of low - mass stars burn it during the extra - mixing occurring in their red giant phase . in fact , we can simultaneously reproduce the low @xmath0he abundances of the solar region and of hii regions at any galactocentric distance , and the high abundance of ngc 3252 and the other pne measured by rood et al .
, which would consequently be associated to the remaining fraction ( 10% ) of stars without deep mixing .
since the deep mixing depletes not only @xmath0he , but also the @xmath19c/@xmath20c ratio ( see charbonnel and sackman , this volume ) , it is important to check the self - consistency of the solution by comparing the model predictions with the carbon isotopic ratio .
charbonnel and do nascimento ( 1998 ) find indeed that more than 90% of 191 field and cluster red giants
present carbon ratios significantly lower than the @xmath19c/@xmath20c=25 predicted by standard nucleosynthesis .
what we also want to check are the predictions of chemical evolution models .
this has been done by palla et al .
( 2000 , hereinafter pbstg ) with a two - folding approach : a ) we have compared the available observational data on the carbon isotopic ratio with the corresponding predictions of chemical evolution models assuming the deep mixing in various percentages of low - mass stars ; b ) we have observed @xmath19c and @xmath20c in 28 pne in mm - waves and compared the derived ratios with those predicted by stellar nucleosynthesis .
fig.5 shows what model tosi-1 predicts for the carbon ratio when the @xmath19c and @xmath20c adopted yields are from boothroyd & sackman ( 1999 ) for low - mass stars with cbp , from marigo ( 2000 ) for low and intermediate - mass stars without cbp , and from limongi , chieffi & straniero ( 2000 ) for massive stars .
equivalent results are described by pbstg for stellar yields from other sources .
the dotted line shows that without extra - mixing in low - mass stars the @xmath19c/@xmath20c ratio is overpredicted with respect to both the abundances observed in the sun and in molecular clouds ( assumed to be representative of the present disk abundances ) .
vice versa a good agreement is achieved if the fraction of stars with cbp is as high as possible ( recall that one can not assume 100% because of the few pne with high @xmath21he ) . as discussed by pbstg , the amount of predicted @xmath19c and @xmath20c strongly depends not only on the extra - mixing assumptions but also ( mostly ) on the assumptions for the nucleosynthesis in intermediate - mass stars , which are the major contributors to the ism enrichment of the carbon isotopes .
however , we can safely conclude that the observed carbon ratios are always better reproduced by models adopting high percentages of low - mass stars with cbp .
this is also supported by the comparison of the @xmath19c/@xmath20c derived by pbstg for the pne where @xmath20c was actually measurable with the carbon ratio predicted for stars right before the ejection of the pn by various nucleosynthesis studies .
the left panel of fig.6 shows that most of the data points ( triangles with associated error on the progenitor mass estimate ) present carbon ratios lower than those expected from standard nucleosynthesis .
the right hand panel shows that the measured carbon ratios are consistent with the predictions of the cbp models at the end of the red giant phase ( unfortunately , no nucleosynthesis models are available yet up to the pre - pn phase , and one can not perform a more appropriate comparison with the pne observed ratios ) . hence , with deep mixing in @xmath2290% of low - mass stars one can reproduce the abundances of @xmath21he observed in the sun , in the ism and in pne and the @xmath19c/@xmath20c measured in the sun , in red giants , in the ism and in pne .
we can then conclude that this mechanism appears to be a very promising process , which needs to be further investigated , both to individuate its possible causes and to check its effects of later stellar evolution phases .
our current knowledge on the galactic evolution of @xmath21he can be summarized as follows : all its available observational abundances can be explained if a ) its primordial abundance is low , ( @xmath21he / h)@xmath23 , b ) the deterium primordial abundance is also low , and c ) deep mixing occurs in almost all low - mass stars .
most of what has been described in this review results from the collaborations with c. chiappini , d. galli , f. matteucci , f. palla , l. stanghellini and g. steigman , and i thank them all for their help .
interesting conversations with c. charbonnel and n. prantzos have also been very useful .
i thank the organizers for such an enjoyable and successful symposium .
this work has been partially supported by the italian cofin98-murst at arcetri .
boissier , s. , & prantzos , n. 1999 , , 307 , 857 boothroyd , a.i . , &
sackman , i .- j .
1999 , , 510 , 232 burles , s. , & tytler , d. 1998 , in primordial nuclei and their galactic evolution , n.prantzos , m.tosi , & r. von steiger eds , 84 , 65 , b&t98 cass , m. , & vangioni - flam , e. , 1995 , in the light element abundances , p.crane ed .
( springer , de ) , p.44 charbonnel , c. 1994 , , 282 , 811 charbonnel , c. 1995 , , 453 , l41 charbonnel , c. , & do nascimento , j.d.jr 1998 , , 336 , 915 chiappini , c. , & matteucci , f. 1996 , in from stars to galaxies : the impact of stellar physics on galaxy evolution , c. leitherer , u. fritze - von - alvensleben , & j. huchra eds , conf.ser . , 98 , 541 dearborn , d.s.p .
, steigman , g. , & tosi , m. 1996 , 465 , 887 ( dst96 ) forestini , m. , & charbonnel , c. 1997 , , 123 , 241 galli , d. , palla , f. , ferrini , f. , & penco , u. 1995a , 443 , 536 galli , d. , palla , f. , ferrini , f. , & straniero , o. 1995b , in the light element abundances , p.crane ed .
( springer , de ) , p.224 galli , d. , stanghellini , l. , tosi , m. , & palla f. 1997 , 477 , 218 geiss , j. , & gloeckler , g. 1998 , in primordial nuclei and their galactic evolution , n.prantzos , m.tosi , & r. von steiger eds , 84 , 239 , g&g98 hogan , c.j .
1995 , , 441 , l17 iben , i. 1967 , , 147 , 650 limongi , m. , chieffi , a. , straniero , o. 2000 , in the chemical evolution of the milky way : stars versus clusters , f. matteucci and f. giovannelli eds ( kluwer , holland ) in press linsky , j.l .
1998 , in primordial nuclei and their galactic evolution , n.prantzos , m.tosi , & r. von steiger eds , 84 , 285 l98 marigo , p. 2000 , in the chemical evolution of the milky way : stars versus clusters , f. giovannelli and f. matteucci eds ( kluwer , holland ) in press , astro - ph/9912341 palla , f. , bachiller , r. , stanghellini , l. , tosi , m. , & galli , d. 2000 , , in press , astro - ph/9912086 , pbstg reeves , h. , audouze , j. , fowler , w.a . ,
& schramm , d.n .
1973 , , 179 , 979 rood , r.t . , bania , t.m . ,
wilson , t.l . , & balser , d.s .
1995 , in the light elements abundances , p.crane ed .
( springer , de ) , p.201 rood , r.t .
, steigman , g. , & tinsley , b.m .
1976 , , 207 , l57 sackman , i .-
j . , & boothroyd , a.i . 1999 , , 510 , 232 scully , s. , cass , m. , olive , k.a . , & vangioni - flam , e. 1997 , , 476 , 521 songaila , a. , cowie , l.l .
, hogan , c.j . , rugers , m. 1994 , nature , 368 , 599 , schr94 steigman , g. , & tosi , m. 1995 , , 453 , 173 tosi , m. 1988 , , 197 , 33 tosi , m. 1996 , in from stars to galaxies : the impact of stellar physics on galaxy evolution , c. leitherer , u. fritze - von - alvensleben , & j. huchra eds , conf.ser . 98 , 299 tosi , m. 2000 , in the chemical evolution of the milky way : stars versus clusters , f. giovannelli and f. matteucci eds ( kluwer , holland ) in press , astro - ph/9912370 tosi , m. , steigman , g. , & dearborn , d.s.p .
1995 , in the light element abundances , p.crane ed .
( springer , de ) , p.228 tosi , m. , steigman , g. , matteucci , f. , & chiappini , c. 1998 , , 498 , 226 van den hoek , l.b . , &
groenewegen , m.a.t .
1997 , , 123 .
305 vangioni - flam , e. , & audouze , j. 1988 , , 193 , 81 vangioni - flam , e. , olive , k.a . , & prantzos , n. 1994 , , 148 , 3 vidal - madjar , a. , ferlet , r. , lemoine , m. 1998 , in primordial nuclei and their galactic evolution , n.prantzos , m.tosi , r. von steiger eds , 84 , 297 , v - m98 | the predictions of galactic chemical evolution models for d and @xmath0he are described in connection with those on the other galactic quantities for which observational constraints are available .
models in agreement with the largest set of data predict deuterium depletions from the big bang to the present epoch smaller than a factor of 3 and do not allow for d / h primordial abundances larger than @xmath1 .
models predicting higher d consumption do not reproduce other observed features of our galaxy . if both the primordial d and @xmath0he are low , models assuming that 90@xmath2 of low - mass stars experience an extra - mixing during the red giant phase reproduce all the @xmath0he observed abundances .
the same percentage allows to fit also the observed carbon isotopic ratios , thus supporting the self - consistency of the extra - mixing mechanism .
# 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in |
within the framework of hierarchical structure formation , galaxies form within virialized dark - matter ( dm ) halos . the halo properties and their time evolution can be predicted robustly within a given cold dark matter ( cdm ) model using n - body simulations or analytic approximations .
a fundamental property of a dm halo is its maximum circular velocity , which is likely to be related to the observable internal velocity of the luminous galaxy that resides in the halo . in the local universe
, it is well known that galaxies obey fundamental scaling relations , including the relation between total luminosity and internal velocity ( commonly known as tully - fisher or faber - jackson relations ) , which we will focus on here .
this lv relationship surely reflects some fundamental aspect of galaxy formation , although its origin is poorly understood .
an observational determination of the _ redshift evolution _ of this relationship would provide crucial clues as to its physical origin and would constitute an important constraint on models of galaxy formation .
however , a direct observational determination of the internal velocities of galaxies at high redshift is extremely difficult .
rotation curves have been obtained for a small number of galaxies up to redshift @xmath5 ( vogt et al .
1996 , 1997 , 2000 ) .
these results suggest that there is little evolution in the tully - fisher relation for objects similar to local large spirals out to @xmath5 ( but see mallen - ornelas et al .
1999 ) . at higher redshifts ( @xmath6 ) , no direct observations are yet available .
while galaxies of different morphological types obey different lv relations , one can define a mean lv relation for any given morphological mix .
this can be used to make a direct connection between the distribution functions of galaxies in terms of their _ internal velocity _ and _ luminosity_. clearly , if the evolution of the velocity function is predicted by theory , and the evolution of the luminosity function is determined observationally , this connection can provide a constraint on the evolution of the lv relation . determining the number density of galaxies as a function of their rest - frame luminosity is considerably easier than measuring their rotation velocities directly at high redshift . observed luminosities can be k - corrected to the rest visual frame up to @xmath7 using multi - band photometry extending into the near - ir .
it has been very difficult to obtain spectroscopic redshifts in the ` desert ' regime of @xmath8 , but photometric redshifts offer a means of filling in this gap .
these advances , combined with theoretical progress in predicting the number density and structural properties of dark matter halos , suggest the approach for constraining the redshift evolution of the lv relation that we exploit here .
the basis of our analysis is the evolution of the velocity function of dm halos as determined via a high - resolution n - body simulation of the cosmology ( sigad et al .
2000 , hereafter s00 ) .
we extend these results to other cosmologies using analytic models .
we consider a wide range of possibilities for the redshift evolution of the lv relation , including some motivated by the results of semi - analytic models ( cf . somerville & primack 1999 ) . we then present predictions of the corresponding redshift evolution of the comoving number density of bright galaxies , which may be compared directly with observations to provide quantitative constraints on how the lv relation evolves in time .
we have calculated the velocity function of dm halos using an n - body simulation of the currently popular flat model with @xmath9 , @xmath10 , hubble constant @xmath11 , and @xmath12 .
the simulation ( klypin 1999 ) used a high - resolution adaptive refinement tree code ( kravtsov , klypin & khokhlov 1997 ) to follow the evolution of structure in a periodic box of comoving side 60@xmath13 , with a force resolution of @xmath14 in the densest regions .
we identify bound , virialized halos using a spherical overdensity - based algorithm and the standard definition of virial density from the spherical top - hat collapse model ( cf .
eke , cole , & frenk 1996 ; bryan & norman 1999 ) .
details of the halo finder are described in bullock ( 2000 ; hereafter b00 ) .
although the virial velocity @xmath15^{1/2}$ ] ( where @xmath16 and @xmath17 ar the virial mass and radius ) is often used to approximate galactic circular velocities , a more relevant quantity observationally is the maximum rotation velocity of the halo @xmath18 , which can be related to @xmath19 via the halo density profile .
over relevant radii , halo density profiles are well approximated by the nfw functional form with two free parameters ( navarro , frenk & white 1996 ) : @xmath20 , where @xmath21 is the radius at which the log slope of the profile is @xmath22 .
the halo maximum circular velocity occurs at @xmath23 and is related to @xmath19 via @xmath24 , where @xmath25 $ ] .
here , @xmath26 is the `` concentration '' parameter @xmath27 .
the value of @xmath26 is closely related to the halo mass , and for a typical dm halo surrounding a bright galaxy ( @xmath28 ) at @xmath29 , our simulations yield @xmath30 , implying @xmath31 .
the concentration at a fixed mass evolves rapidly with redshift , @xmath32 ( b00 ) , implying a corresponding evolution in @xmath33 .
the halo catalogs described in b00 provide @xmath34 and @xmath19 for each halo .
a detailed study of the halo velocity function derived from these catalogs is presented by s00 .
( 0,0)(10,10 . ) ( -.2,10.)=9.cm ( 0,1 . )
= 4.5 mm fig .
1. evolution of relative comoving number density for fixed @xmath35 ( bold curves ) and @xmath36 halos in three cosmologies .
figure 1 shows the evolution with redshift of the comoving number density of simulated halos of fixed velocity , either @xmath37 or @xmath38 .
the evolution of halos of fixed @xmath34 is weaker than for halos of fixed @xmath19 due to the strong evolution of @xmath26 . a fit n= ( ) ( )
= _ * ^ ( ) , [ eq : vfit_log ] for the whole halo @xmath34 distribution is provided by s00 , who find that the comoving density at @xmath39 is constant as a function of @xmath40 ( @xmath41 in the log ) at least out to @xmath42 , while @xmath43 , where @xmath44 .
figure 1 also shows predictions based on the analytic approximation of sheth & tormen ( 1999 , hereafter st ) a modification of the press - schechter approximation ( 1974 ) which improves the agreement with simulations .
we converted the st mass to @xmath19 and then to @xmath34 using the analytic model of b00 for @xmath45 .
we see that for , the analytic model provides a good approximation to the simulated evolution both for fixed @xmath34 and fixed @xmath19 .
this allows us to extend our predictions using st to two other cosmological models : @xmath46cdm , a flat model with @xmath47 , @xmath48 and a power spectrum of shape parameter @xmath49 , and ocdm , an open geometry with @xmath50 , @xmath51 , and @xmath11 . in the case of
, we note an increase of @xmath52 in the density for constant @xmath34 between @xmath29 and @xmath53 , and a gradual decline back to today s density by @xmath54 .
the density of halos of a constant @xmath19 grows by a factor @xmath55 by @xmath53 , and then declines at higher @xmath40 .
note that a halo with a fixed @xmath19 has a decreasing mass and virial radius at increasing redshift .
compared to , the ocdm model predicts a slightly larger increase in the density of @xmath0 halos at all redshifts . in the case of @xmath46cdm ,
the peak density is obtained at @xmath56 , followed by a sharper decline , but note that even in this case the comoving density at @xmath57 is comparable to the density at @xmath29 .
armed with this knowledge about the evolution of the halo velocity function , we can now predict the evolution of the number density of galaxies brighter than a fixed absolute luminosity for a given model of lv evolution .
a possible hitch is that the halo maximum velocity @xmath34 typically occurs at a radius @xmath58 that lies outside the luminous parts of the galaxy for which the lv quantities are measured .
for example , a @xmath59 halo at @xmath29 has @xmath60kpc .
the velocity due to the halo alone obviously declines at smaller radii , but the condensed baryons contribute significantly to the total circular velocity there .
we therefore make the assumption that the rotation curve is roughly flat between where it is measured and @xmath58 , i.e. that the halo @xmath34 is the same as the velocity used in the observed lv relations .
gonzalez ( 1999 ; g99 ) used this assumption at @xmath29 to compare the velocity function of dm halos with that obtained from observed luminosity functions and lv relations .
the good agreement that they obtained is evidence that this is a reasonable approximation at low redshift .
we now assume that this relation holds at high redshift as well .
we parameterize the mean lv relation as : @xmath61 , where only the zero - point may evolve in time according to the parametric functional form @xmath62 we choose @xmath63 and @xmath64 , corresponding to the empirical local lv relation used in g99 .
the parameters @xmath65 and @xmath66 allow a wide range of shapes , where @xmath65 governs the behavior at @xmath67 , and @xmath66 the asymptotic shape at larger redshifts .
the indications for only little evolution at @xmath68 guide us to consider mostly large positive values for @xmath65 , while the value of @xmath66 is unconstrained , and even its sign is unknown . in figure 2
we consider five different representative cases for the lv zero - point evolution @xmath69 , and show the predicted evolution of the number density of galaxies with @xmath70 , relative to their number density at @xmath29 .
the case @xmath71 is the case of no evolution in the lv relation . the case @xmath72 is close to the evolution predicted for the model if the mass - to - light ratio is constant in time . in this model , because halos of fixed @xmath34 have smaller masses at high redshift , they are fainter and therefore the predicted number of bright galaxies decreases . the case @xmath73 is close to the prediction of a semi - analytic model ( the `` accelerated quiescent '' model of somerville , primack , & faber 2000 ) in which the efficiency of star formation scales with the inverse dynamical time of the halo .
this scaling is motivated by one of the empirical recipes suggested kennicutt ( 1998 ) , based on observations of nearby galaxies .
the dynamical time for a halo of a fixed velocity is smaller at high redshift ( because of the higher density ) , and so galaxies of a fixed velocity are _
brighter_. this model predicts an _ increase _ in the number density of bright galaxies out to @xmath74 . although this star formation recipe is a standard choice in semi - analytic models ( e.g. , kauffmann , white , & guiderdoni 1993 ) , note that it predicts fairly strong evolution in the lv relation even at @xmath5 .
the other two cases , @xmath75 and @xmath76 , are arbitrary choices that illustrate a wide range of possibilities while keeping the lv relation almost unchanged out to @xmath77 .
( 0,0)(11,11 . ) ( -.1,11.)=8.9 cm ( 0,2 . )
= 4.5 mm fig .
2. five parameterized models marked by @xmath78 for the evolution of the lv zero - point , @xmath79 ( eq . 1 ; larger values are brighter ) , along with the implied relative evolution of the comoving density of galaxies brighter than @xmath80 for the three cosmological models . 0.5 cm
we have argued that observed estimates of the number density of galaxies above a given luminosity as a function of redshift can be used to constrain the redshift evolution of the lv relation .
lilly et al .
( 1995 ) found only little evolution in the number density of _ bright _ galaxies in the rest - frame b - band out to @xmath5 , consistent with the direct constraints on lv evolution out to this redshift ( vogt et al .
1996 , 1997 , 2000 ) .
redshift surveys of the past decade were unable to reach the higher redshift range @xmath81 associated with the spectroscopic desert or ` high place of sacrifice ' .
the results presented by dickinson ( 2000 ) , based on nicmos ( near - ir ) imaging of the hubble deep field ( hdf ) suggest that the assembly of present - day luminous galaxies may be taking place precisely in this redshift interval . using photometric redshifts , dickinson finds that there are only one - third as many galaxies with rest - frame @xmath82 in the redshift interval @xmath83 as in the interval @xmath84 . in our fiducial cosmology ,
the comoving volumes of the hdf for these two intervals are equal , implying an overall drop in the comoving number density at @xmath85 of about a factor of three . taken at face value , this would correspond to dramatic evolution in the lv relation at @xmath6 , rather similar to the most extreme ` dimming ' model shown in figure 2 .
probably the weakest link in our calculation is our assumption that the halo maximum velocity is the same as the measured internal velocity for a galaxy .
the actual measured velocity will depend on the baryon fraction in the galaxy , the spatial distribution of the baryons , and the degree to which angular momentum is preserved in the collapse .
however , note that because we expressed the space density in figure 2 _ relative _ to that at @xmath29 , our plotted results will not change by much as long as the ratio of observed velocity to the halo maximum velocity , @xmath86 , is of order unity and roughly constant with redshift .
we have calculated the value of @xmath87 using a simple analytic model of disk formation , which assumes that a constant fraction of the halo mass condenses into a disk , and angular momentum is conserved ( mo , mao , & white 1998 ) .
we find that @xmath88 for the redshift range @xmath89 , and falls to @xmath90 by @xmath74 .
this suggests that our predictions should be reasonably robust at least out to @xmath7 .
how significant is the effect seen in the hdf ?
one possibility is that galaxies are preferentially missed at @xmath91 because of the @xmath92 cosmological surface - brightness dimming .
dickinson ( 2000 ) claims that this is unlikely to be a large effect ; when bright hdf galaxies from @xmath93 are artificially redshifted to @xmath94 with no intrinsic luminosity evolution , most of them are still easily detectable .
another possibility is that the photometric redshifts are inaccurate ; dickinson points out that there are _ no _ spectroscopic redshifts in the range @xmath95 where the deficit is seen . however
, this would require quite a dramatic failure where _ all _ of the galaxies ( with photometric redshifts ) assigned to @xmath96 would have to be reassigned to the higher redshift range .
most importantly , recall that the hdf covers a relatively small volume , and therefore random fluctuations along the line of sight are likely to be large .
the observed deficit of galaxies might simply reflect large - scale `` spikes '' in the redshift distribution . although there is some evidence for a similar deficit of bright galaxies at @xmath97 in other deep optical / ir surveys ( fontana et al .
1999 ) , secure confirmation of these results will require deep wide - field surveys with multi - band optical - ir photometry .
several such surveys are now in progress ( e.g. the cambridge - carnegie las campanas ir survey ) .
moreover , the deep survey ( davis & faber 1998 ) will obtain measured line - widths for a large number of galaxies at @xmath98 , which will provide direct constraints on the lv relation at high redshift .
finally , our results may be relevant to a recent proposal by newman & davis ( 2000 ) to use velocity - selected galaxies as a tracer population in a classical @xmath99 measure of cosmological parameters .
this analysis would be subject to similar uncertainties in relating the halo velocity to observable galaxy velocity , e.g. , the dependence on whether the halos are selected by @xmath34 or @xmath19 .
moreover , because the proposed spectroscopic surveys are magnitude limited , the indicated evolution of the lv relation could introduce severe selection biases into the analysis .
this work has been performed at the itp , santa barbara .
it has been supported by grants from the us - israel bsf , the israel sf , nasa and nsf at ucsc and nmsu .
jsb was supported by nasa ltsa grant nag5 - 3525 and nsf grant ast-9802568 .
jrp acknowledges a humboldt award at the max planck institute for physics in munich .
rss is supported by a rolling grant from pparc .
we thank marc davis and mark dickinson for helpful correspondence .
dickinson , m. , 2000 , in `` building galaxies : from the primordial universe to the present '' , eds .
hammer , f. , thuan , t.x .
, cayatte , v. , guiderdoni , b. , & tranh than van j. ( paris : ed .
frontires ) , p. 257
, astro - ph/0004027 | we present a method for constraining the evolution of the galaxy luminosity - velocity ( lv ) relation in hierarchical scenarios of structure formation .
the comoving number density of dark - matter halos with circular velocity of @xmath0 is predicted in favored cdm cosmologies to be nearly constant over the redshift range @xmath1 .
any observed evolution in the density of bright galaxies implies in turn a corresponding evolution in the lv relation .
we consider several possible forms of evolution for the zero - point of the lv relation and predict the corresponding evolution in galaxy number density .
the hubble deep field suggests a large deficit of bright ( @xmath2 ) galaxies at @xmath3 .
if taken at face value , this implies a dimming of the lv zero - point by roughly 2 magnitudes .
deep , wide - field , near - ir selected surveys will provide more secure measurements to compare with our predictions .
# 1([eq:#1 ] ) # 1equation ( [ eq:#1 ] ) # 1eq .
[ [ eq:#1 ] ] # 1[sec:#1 ] # 1figure [ fig:#1 ] # 1fig .
[ fig:#1 ] # 1figure [ fig:#1 ] + # 1#1 # 1#1 # 1#1@xmath4 # 10= -.025em0 - 0 .05em0 - 0 -.025em.0433em0 0 |
the history of galaxies is , in large part , the history of star formation .
massive stars play key roles both as highly visible tracers of star formation and as players in altering surrounding star formation and both energy and chemistry of the interstellar medium .
it is these stars which dominate the observed properties of actively star - forming galaxies .
the massive part of stellar populations is most clearly observed in the ultraviolet , where their energy distributions peak and competing light from cooler stars is minimal , as long as the foreground extinction allows escape of enough of this radiation . in such environments , ultraviolet studies of star - forming regions
have proven fruitful in understanding these populations .
the recent opening of the far - ultraviolet window , between lyman @xmath2 and the lyman limit , for deep observations , allows study of massive hot stars in a range where they fully dominate the spectrum .
this relatively narrow band contains an embarrassment of spectral riches , with numerous lines from stellar winds as well as interstellar material both atomic and molecular .
these include the strong and highly - ionized lines of o vi and the unique ability to measure cold h@xmath0 .
in addition , this piece of the spectrum is accessible for high - redshift galaxies , at least for composite samples where the lyman @xmath2 forest can be averaged adequately , allowing the possibility of direct comparisons of stellar populations over a large span of cosmic time .
the very sensitivity of the far - uv light to star formation and reddening makes it a purer probe of some properties of star - forming regions than observations at longer wavelengths .
since only short - lived stars contribute significantly in the far - uv range , the details of star - forming history should matter only for very brief bursts ( such as might be found in individual h ii regions , but are less likely on galaxy scales ) .
this makes the far - uv spectrum more sensitive to the stellar population itself than to its history .
furthermore , although the extinction is high , its differential effect across the far - uv band is modest , and paradoxically the effective reddening to the stars we see is smaller than found at longer wavelengths . in observing stars
intermingled with highly structured dust distributions , the picket - fence " effect ( heisler & ostriker 1988 ) means that most of the stars are so reddened as to make no significant contribution in the deep ultraviolet ; all the stars we see are only lightly reddened .
this also reduces the effects of the forest of h@xmath0 absorption features because of the mixing of molecular gas and dust .
observational data on nearby galaxies , before the _ far - ultraviolet spectroscopic explorer _ ( fuse ) , were limited to a handful of star - forming systems .
four starburst galaxies were observed using hut on _ astro-2 _ ( leitherer et al . 1995 , 2002 ) , largely to measure escaping radiation in the lyman continuum , which provided initial data for comparison with models based on stellar spectra dating back to _
copernicus_. the strongest features fall into two blends near 970 ( ly @xmath3 + c iii ) and 1030 ( o vi + ly @xmath4 + c ii ) .
a _ voyager 2 _ observation of m33 , with some spatial resolution in one direction , was analyzed by keel ( 1998 ) , showing that its far - uv continuum is virtually identical to those of the powerful starbursts , and that ngc 604 is bluer in this range than the overall disk .
these data also suggested significant lyman @xmath2 emission , with spatial profile suggesting an origin in the diffuse ism rather than giant h ii regions .
ironically , until the availability of fuse observations , the richest information on galaxies shortward of lyman @xmath2 came from objects at high redshift , particularly composite spectra of lyman - break galaxies ( steidel et al .
2001 , shapley et al . 2003 ) . to enable comparisons between local , well - studied star - forming systems and these powerful , young objects
, we have undertaken a series of fuse observations of star - forming regions innearby galaxies .
we present here the analysis of these spectra in the context of their stellar populations and systematic changes with metallicity . a companion paper ( keel , shapley , & steidel 2004 )
considers the comparison with composite spectra of lyman - break galaxies .
the fuse targets were selected for known high uv flux and hence low reddening , and to span a range of metallicity .
the regions targeted in m33 include the brightest h ii regions in the mid - uv , and span a radial range within the disk from 1.6 - 4.5 kpc ( see fig .
they also include examples with a single dominant star cluster and with multiple clusters or more diffuse associations ( also illustrated in fig .
1 using hst data at 1700 ) . none of the objects in m33 or m101 achieves super star cluster " ( ssc ) status , if we follow common usage in requiring not only high luminosity and stellar mass ( melnick , moles & terlevich 1985 ) but that the stars be concentrated into a single clump on 10-pc scales as noted by such studies as that of meurer et al .
while ngc 604 has the right luminosity , its stars are widely spread throughout a 100-pc region including multiple clumps ( hunter et al .
1996 ) . in the more distant systems , the fuse aperture samples multiple regions ; fig .
2 shows the bluest archival hst imagery with the aperture superimposed . in ngc 7714 , there are @xmath5 luminous sscs , comparably luminous in the near - uv , within the aperture , more clearly shown in the inset to fig .
2 . in each of these cases ,
essentially the entire star - forming region fits within the aperture , an important desideratum in comparing these observations with global measurements of distant galaxies .
for each of these objects , there are abundance measurements from the traditional optical emission lines , which we use as an widely - applicable tracer of the abundances of recently - formed stars as well .
the m33 objects span much of its disk s abundance gradient , while ngc 5461 and the nucleus of ngc 7714 were selected to sample lower and higher metallicity .
we also analyze the summed archival fuse spectrum of i zw 18 from aloisi et al .
( 2003 ) , as a comparison with the lowest known gas - phase abundances .
table 1 includes our adopted values from the literature for [ o / h ] , the best - measured of the abundances due to the optical lines from multiple ionization stages of oxygen and its importance as a coolant .
the fuse optical system and detectors are described by moos et al .
four primary mirrors are used to feed independent detectors optimized for subsections of the far - uv band ; maintaining the coalignment of these optical systems is an important operational issue .
each detector has a distinct wavelength calibration , but for our purposes , we are not pushing the resolution limit of fuse and can combine the various data segments in 0.1- pixels . as shown in figs .
1 and 2 , the fuse pointing positions were set at the midpoint of the stellar distributions from available uv images .
the large 30 " aperture was specified for the m33 and m101 h ii regions at the outset , both for flux integrity in alignment of the four instrumental channels and to include essentially all their starlight , and adopted for ngc 7714 as well when the performance penalty for using a smaller aperture became clear during the mission .
the requested center coordinates and total exposure times are listed in table 1 . of these objects ,
ngc 604 is the brightest in the far - uv range by a factor 10 , and will thus play a continued role in our knowledge of stellar populations in this spectral range . in each of these spectra ,
the useful resolution is limited by the signal - to - noise ratio , dictating the wavelength binning to detect features of interest .
for ngc 604 , we can work at 0.1- binning , for a velocity spacing typically 30 km s@xmath6 . only in this object
does it ( marginally ) matter that it is not a point source , with starlight coming from most of the aperture area .
this smearing along the dispersion axis contributes a broadening of about 60 km s@xmath6 .
the spectra were processed through the fuse pipeline ; we inspected the two - dimensional spectra to verify data quality and background subtraction , and to note such effects as the worm " , a shadowing of the detector by a repeller wire , near the red end of some spectra .
each of our targets was observed during two orbits , so we could compare each orbit s summed spectra as a check on the errors at each wavelength .
loss of flux due to the worm " does occur in some of the spectra around from 11701180 , a range which does not affect any of the features we analyze .
similarly , there are spurious emission features near 1044 and 1169 which are second - order scattered solar features , which do not seriously confuse features of interest . for clarity of illustration , we interpolate across terrestrial airglow features in ly @xmath4 , ly @xmath3 and o i at the highest resolution before further averaging and display .
we also use the summed archival spectrum of i zw 18 described by aloisi et al .
( 2003 ) , who kindly provided it in electronic form .
we compare the spectra of ngc 604 and i zw 18 at 0.1- resolution in fig .
3 , showing foreground absorption from h@xmath0 , atomic absorption lines in the foreground and in these galaxies , and prominent stellar - wind lines .
we draw on additional data from a variety of sources for these well - observed objects .
iue spectra in the large @xmath7-arcsecond apertures give mid - uv spectra in roughly matched apertures ; we have collected spectra from the archive , rejecting those with obviously discrepant pointings or flux levels from the combined spectra .
hst imaging is invaluable in examining the stellar populations , resolving the brightest stars in each of these associations .
archival wfpc2 data exist for each of the m33 regions in f170w plus longer - wavelength filters , so we can assess the stellar statistics at least into the mid - ultraviolet .
the best uv images of ngc 604 were taken in support of a slitless - spectroscopy program , using the stis nuv - mama detector working at 1800 - 2700 .
each of these stellar collections ionizes a substantial surrounding region .
h@xmath2 data ( as in bosch et al .
2002 ) show structure extending 200 - 800 pc .
all four of the m33 regions fit their description of an evolved h ii region , one in which star formation has been long enough to generate extensive filaments and loops driven by stellar energy input ; they estimate ages @xmath8 myr for all except ngc 604 , at 3 myr , for the current episode of star formation ( matching the conclusions of gonzalez delgado & perez 2000 ) .
stars which contribute strongly in the far - uv have spectral types b0 and earlier , for ages up to 15 myr .
however , the continuum slopes of the shorter - lived o stars vary only slightly across the far - uv range , so the continuum offers little signature of the recent star - forming history ( robert et al . 2001 ) . to this point ,
the far - uv line behavior is not well enough calibrated to infer the star - forming history independently .
the far - uv range is rich in narrow absorption lines from the ism , both ionic and molecular .
the spectrum of ngc 604 is especially complex , with over 100 detected and identified transitions .
foreground h@xmath0 in the milky way is so prominent in all these objects that it must be carefully accounted for in identifying intrinsic absorption features . for the strongest fe ii lines , we detect not only the m33 ism ( at about -200 km s@xmath6 ) , but the foreground high - velocity cloud structure near -370 km s@xmath6 described by wakker et al .
( 2003 ) .
numerous atomic absorption features are seen in these spectra , with stronger absorption in most cases from foreground milky way gas .
their strengths in the integrated spectra will reflect the distribution of gas toward the stellar associations weighted by contribution to the far - uv flux ; for these extensive distributions of stars , derived column densities should be regarded as characteristic values .
the lyman and werner band systems of h@xmath0 are at once an advantage and a nuisance in far - uv spectroscopy , detected in almost every line of sight crossing a significant distance in our own galaxy s ism .
a total of 56 such lines are individually detected in ngc 604 .
the h@xmath0 features in these m33 data have also been analyzed by bluhm et al .
( 2003 ) , who found that h@xmath0 absorption at the velocities appropriate for disk gas in m33 is weakly detected in ngc 588 , 592 , and 595 , at column densities n(h@xmath0)= @xmath9 @xmath10 , but only an upper limit @xmath11 could be derived for ngc 604 , despite the higher data quality and comparable column densities from atomic lines , with all the values approximate due to the lines being too weak for independent measurement of the doppler - width @xmath12 parameter .
bluhm et al . also present instructive simulations on the difficulty of deriving unique column densities when the background source consists of stars spanning a range of column density and position within the spectroscopic aperture .
for example , the nondetection in ngc 604 might result from patchy extinction , correlated with the h@xmath0 distribution , and the far - uv background flux being dominated by those stars with the shallowest h@xmath0 absorption .
this will act in addition to any physical effects related to the radiation field and shocks in the vicinity of such an active star - forming region .
these regions are certainly associated with large concentrations of molecular gas ; all three of them within the survey area of engargiola et al . ( 2003 ) are associated with molecular clouds ( ngc 592 , 595 , and 604 with their clouds 40 , 32 , and 8 , respectively ) .
their h@xmath0 masses are estimated in the range @xmath13 solar masses . in spite of these issues in physical interpretation of molecular absorption ,
the empirical results are directly comparable to what we may see in high - redshift systems , as this spectral region becomes accessible for at least composite study in lyman - break galaxies ( shapley et al .
we can use both the sets of h@xmath0 line strengths seen from galactic gas in these spectra , and synthetic absorption spectra , to ask what signatures of molecular absorption remain at modest spectral resolution .
the least confusion with other spectral features for modest h@xmath0 column densities occurs for blends of features near 1005 and 1072 .
the wavelengths and shapes of these blends depend on the spin - level populations in the gas , generally shifting to longer wavelengths for warmer gas ( mccandliss 2003 ) .
we have re - examined the ngc 604 spectrum for evidence of associated h@xmath0 absorption . the ability to see such features at low redshift
is limited largely by blending with foreground galactic features , since the spacing between multiplet members provides ample opportunity for overlap . for atomic absorption species
, we find a difference between foreground and m33 absorption typically 220 km s@xmath6 , which gives wavelength shifts very close to the spacing of low - order members of the h@xmath0 multiplets near 963 , 982 , 1002 , and 1014 . as a result
, only the bluest members of these multiplets could be securely detected from gas near the velocity of ngc 604 in the presence of the much stronger foreground gas . to guard against unrecognized atomic features
, we also require that a putative ngc 604 feature not be seen in the i zw 18 spectrum at the corresponding _ emitted _ wavelength .
compared to 56 h@xmath0 lines seen from foreground milky way gas , only two potential detections resulted for ngc 604 .
these are the transitions at rest wavelengths 1001.69 ( @xmath14 ) and 1008.38 ( @xmath15 ) at equivalent widths of roughly 0.02 and 0.08 , respectively .
the 1008 line , with better s / n ratio , appears substantially broader , with a profile compatible with that of the other narrow features .
if accurate , these detections are insufficient to derive the absorbing column density , lacking information on the intrinsic doppler broadening and level populations .
if we adopt the optimum " @xmath12-value 10 km s@xmath6 for the ngc 588 and 592 detections from bluhm et al . , and
incorporating the h@xmath0ools template models described by mccandliss 2003 , we derive n(h@xmath0)=@xmath16 @xmath10 from this single transition and assuming the same mix of level populations as in the foreground gas .
hotter molecular gas could give comparable absorption at column densities several times lower , close to the bluhm et al . limit .
lines arising in the states @xmath17 are seen in the foreground milky way gas , well below the saturated level that drove rachford et al .
( 2002 ) to use detailed profile fitting . as in their work
, we also find significant changes in derived n(h@xmath0 ) between various absorption bands .
we follow them in obtaining a characteristic kinetic temperature from the level populations in @xmath18 as @xmath19 for the ngc 604 foreground h@xmath0 spectrum , we obtained @xmath20 k , at the high end of values seen by rachford et al . for dense and self - shielded environments but reasonable for the more diffuse
intercloud " ism .
the populations in @xmath18 are taken to represent a thermal ( collisional ) temperature , with excess " absorption from higher @xmath21 representing radiative excitation , so that only the lowest levels are useful in deriving the thermal environment of the molcular gas .
similar issues affect the strongest h@xmath0 lines in i zw 18 , albeit with less overlap from milky way features because of the larger velocity shift . from the stacked spectrum , aloisi et al .
( 2003 ) set an upper limit of n(h@xmath0 ) = @xmath22 @xmath10 against the emerging continuum .
most of the stellar signature in far - uv spectra occurs in the wind lines , which dominate this spectral range thanks to the numerous resonance transitions of metals .
both the wind lines and the evolutionary tracks when winds are important should have strong metallicity dependences , since the radiation pressure driving the winds acts largely through the opacity of heavy elements .
this behavior has been shown , albeit sometimes in complex ways , upon comparison of stellar spectra from the milky way and magellanic clouds .
leitherer et al .
( 2001 ) examined wind lines in the mid - uv , notably c iv , from smc to galactic abundances , confirming a significant metallicity dependence in its strength but noting than ionization states other than the dominant ones can behave differently as secondary abundance effects shift the ionization balance in the winds ( as seen in si iv , which is stronger in lmc stars than the milky way ) .
these changes are most pronounced for the brighter luminosity classes . among the most luminous stars , those in the magellanic clouds have smaller wind velocities even for comparable depth ( e.g. kudritzki & puls 2000 )
. while photospheric lines in the far - uv range are weak enough to scale directly with abundances , they are either extremely weak or overlain by wind lines
. lamers et al .
( 1999 ) suggest that line blanketing is a more secure route to the stellar abundances . from the comparison by robert et al .
( 2001 ) , si iv @xmath23 are high - excitation photospheric lines that should serve as useful indicators of the stellar population , independent of winds . in interpreting the o vi profiles from galaxy - scale systems such as ngc 7714 , gonzalez delgado et al .
( 1998 ) caution from hut spectra that interstellar absorption from large - scale outflows can significantly overlap wind lines , specifically between ly @xmath4 and o vi @xmath24 .
even fuse - based spectral syntheses in the far - ultraviolet extend only to 1000 ( robert et al . 2003 ) , so we continue our empirical approach in comparing systems of various properties to seek differences linked to composition or history .
we can , however , be guided by available syntheses of some of the prominent blends of features , such as the ly @xmath4-o vi - c ii regions from 1025 - 1038 ( gonzalez delgado et al .
they find that ly @xmath4 and c ii absorption arise in b stars , at the cool end of far - uv contributors , while o vi p cygni profiles come from the most massive stars .
high spectral resolution is crucial to separating the interstellar absorption contributions in each case ; for ly @xmath4 , hydrogen column densities @xmath25 @xmath10 lead to blending of stellar and interstellar components .
robert et al .
( 2001 ) show that o vi is very weak at smc metallicities , and that the youngest population can be diagnosed from the presence of c ii @xmath26 and c iv / n iv @xmath27 which are specific to hot o stars . at magellanic cloud abundances ,
the s iv @xmath28 lines have a wind contribution only from supergiants , and p v @xmath29 exhibit similar behavior . to show the some of the spectral differences seen with metallicity , fig .
4 compares the spectra of ngc 604 and i zw 18 , now with the strongest galactic h@xmath0 features removed by fitting gaussians or voigt profiles ( for the stronger lines ) , and plotted in the emitted wavelength frame .
5 is a similar comparison of ngc 7714 with ngc 604 , illustrating the yet stronger features seen at near - solar metallicity .
the wind lines seen in these systems may be described as follows . for the fainter objects with short exposures ( ngc 588 , 592 , 595 , 5461 , 7714 )
the data have been smoothed by typically 0.7 in making these assessments .
ngc 604 ( fig .
4 ) : combining both members of the o vi doublet to reduce confusion , there is a broad wind absorption reaching 2800 km / s and probably blending with ly @xmath4 .
o vi @xmath30 , c iii @xmath31 , and n ii @xmath32 1083 have nearly black cores .
a distinct detached absorption may be present in c iii from 1600 - 2000 km / s
. there may be a broad wind feature in n iii to @xmath33 km / s .
p v is weak except for possibly photospheric components near zero velocity .
the emission sections of p cygni profiles are strong in o vi @xmath30 , n ii @xmath34 , and n iii @xmath35 .
ngc 588/592/595 : these spectra have wind features as much like one another as their s / n ratio can tell , both individually and when averaged .
compared to ngc 604 , the cores of c iii , o vi , and n ii are shallower , with residual intensities 10 - 40% . as noted below , the continuum level in ngc 604 is higher than in any of the other systems below @xmath36 .
ngc 5461 : a trough occurs in n iv to about 1000 km s@xmath6 .
there is a single well - defined trough in c iv to 600 km s@xmath6 .
no feature is obvious in n iii or s iii .
broad absorption is seen in both o vi lines to about 1000 km s@xmath6 .
i zw 18 : there is at most weak o vi absorption between 1700 - 2800 km s@xmath6 , seen only in the @xmath30 line .
p cygni emission is absent in o vi , n ii , and n iii by comparison with ngc 604 . essentially no blueshifted troughs occur for n iii .
two features may be present for c iii from 0 - 1200 and 1700 - 2000 km / s , but blending with saturated interstellar c ii is an issue .
p v is very weak .
ngc 7714 : this nucleus has the strongest wind features in our study . both o vi lines are strong , extending to 900 km @xmath6 .
the p v lines at @xmath37 are broad and blueshifted suggesting a wind contribution , which is plausible for non - supergiants at its metallicity .
n iii 991 may have a broad wind component blending with the interstellar o i lines at about 1000 km s@xmath6 .
the core of c iii is blended with o i but shows a wind trough to beyond 1000 km s@xmath6 . for high - redshift objects
, it will be difficult or impossible to separate wind and interstellar contributions to some of these lines . for purposes of comparison ,
we have generated simple equivalent - width values with respect to the adjacent ( pseudo)continuum for lines in clean parts of the spectrum , typically spanning a 5- line region , and tabulate these in table 2 .
airglow emission contaminated the fainter m33 regions ( combined in the table as
m33 avg " ) too strongly for a reliable measurement of n iii @xmath35 . for the n ii and n iii lines we list both the absorption and net equivalent widths ( in parentheses ) , for use when the components are not resolved .
we include values for c iv @xmath38 from summed iue low - dispersion spectra , as these values were taken with a large enough aperture to sample most of the stellar population in a way much like the fuse aperture .
errors on the fuse value are typically 0.2 for such broad features , as reflected in the limits for some lines in i zw 18 , while the c iv errors are closer to 1 .
each of these lines shows a strong trend with emission - line metallicity ( fig .
6 ) , showing that they do in fact have potential use as abundance indicators .
the wind lines measure stellar values directly , in contrast to the large regions of the interstellar medium sampled by emission - line techniques .
these relations are quite sensitive in the sub - solar regime of particular interest for the evolution of galaxies at @xmath39 .
the equivalent widths measured for n iv , c iii , and p v vary almost linearly with abundances .
our results indicate that these strong features can be used as metallicity indices for high - redshift stellar populations , in regimes for which the optical emission lines lie in infrared bands of high atmospheric emission and absorption , and can be a valuable tool for approaching the chemical evolution of galaxies in the range @xmath40 . for ease of use
, we present least - square quadratic fits to the data in fig .
6 , along with the derived constants needed to invert these fits with line equivalent widths as the independent variables .
these interpolation curves are shown in fig .
6 , and the numerical values are found in table 3 .
the o / h ratio is in solar units as found from emission - line analysis ( as cited in table 1 , updated to the new " solar abundance scale with 12 + log ( o / h)=8.69 ) , and all equivalent widths are in in the emitted frame . for each line ,
the data have been approximately fitted in the form @xmath41 .
similarly , an inverse fitting function for metallicity derived from each line is obtained from the quadratic formula , with constants tabulated as @xmath42 .
the values are based on least - squares fits with equal weights , slightly modified in two cases to keep the fit monotonic across the metallicity range of our sample . for [ n ii ]
, we adopt a flat value ew=1 for ( o / h)@xmath430.5 solar .
these forms are for interpolation purposes , and their limitations appear from the fact that they do not all approach zero line strength at zero metallicity .
crudely speaking , wind - dominated lines would have a quadratic dependence on metallicity until saturation sets in , since the mass - loss rate and fraction of the mass in the right ionic state each depend on metallicity , while mostly photospheric lines ( such as p v ) are more nearly linear in strength with the abundances
. strictly speaking , these relations apply only to galaxies with a long ( even if episodic or weak ) history of star formation , since the elements involved in these lines come from very different stellar processes . for example , oxygen should be enriched rather quickly , coming from massive stars , while carbon will take longer coming from intermediate masses , with the dominant sources nitrogen still somewhat ambiguous .
at best , these relations could be taken seriously for single elements in high - redshift galaxies , and can in fact be used to test for the differential enrichment history of such systems , in much the same way that the behavior of si iv and c iv lines with redshifts has been used to infer such differences by mehlert et al .
( 2002 ) , who derive a relation between c iv ew and metallicity consistent with our iue analysis included in fig .
6 . beyond the stellar - wind lines ,
the greatest difference among all these far - uv spectra is the flux excess in ngc 604 from about 912 - 940 , above what is seen in any of the other objects either of higher or lower metallicity .
small - number statistics in the massive stars might be expected to be more important for the less rich systems , but in m33 it is these which have the same spectral shape as the more luminous and distant systems , leaving ngc 604 as the odd one out .
previous work on its stellar content indicates that there was a strong peak in its star - forming history about 3 myr ago , recent enough to affect the mix of stars contributing in this spectral range .
we have examined additional uv data on the stellar content of these m33 h ii regions to see how ngc 604 might be different .
the stellar population in ngc 604 has drawn considerable attention , in being one of the brightest star - forming complexes in the local group while offering a dramatic structural contrast to the dominant , compact stellar cluster in 30 doradus .
bruhweiler , miskey , & smith neubig ( 2003 ) combined mid - uv wfpc2 images with a wide - slit stis spectrum to extract individual spectra of the brightest members .
they note that the ten most uv - luminous stars will dominate the integrated spectral features from ngc 604 , and that two of these are located @xmath44 magnitude above the usual 120-solar - mass limit . specifically ,
the top ten " stars contribute 46% of the total measured flux just longward of lyman @xmath2 .
mid - uv stis images ( central wavelengths 1820 - 2700 ) obtained in support of a slitless - spectroscopy program by j. mais - apellaniz show , via the finer pixel sampling , that there are additional spatially resolved companions to each of the brightest stars identified by bruhweiler et al . , although none so bright as to bias the measured colors or magnitudes .
ngc 604 also contains significant numbers of wr stars , although with a wr / o star ratio near 0.1 rather than the 0.3 seen in , for example , ngc 595 ( drissen , moffat , & shara 1993 ) . to compare the populations in these h ii regions as they affect the integrated far - uv spectra , we produced color - magnitude arrays for each based on the archival wfpc2 data in f170w and f555w ( as well as intermediate - wavelength data when available ) .
lacking imagery in the far - uv band itself , we use the mid - uv properties as proxies to at least identify the hottest luminous stars .
reddening corrections do not enter for our immediate purposes , since we need to know only how many stars contribute to the uv flux and extinction does not varying greatly between 1700 and 1100 .
bruhweiler et al .
find , for stars in ngc 604 , that the 1100- extinction is about 1.5 times that at 1700 .
likewise , we are interested here only in stars selected from uv flux , so the diagrams sample only those stars well detected in the mid - uv data .
these observational hr diagrams are shown in fig .
we include only stars without serious crowding issues , measured within 0.3 " radii .
the samples include stars to about @xmath45 on the stmag scale ( based on flux per unit wavelength ) , although the completeness varies considerably even at @xmath46 because of differences in crowding ; for ngc 604 , the cumulative counts with flux suggest statistical completeness only above @xmath47 .
the brightest stars in the rich population of ngc 604 are 1.5 - 2 magnitudes brighter than found in any of the other associations .
the color range among uv - bright stars in ngc 588 is smallest , extending from the blue envelope near @xmath48 redward only to -2.0 , while all the other regions have the diagram populated to @xmath49 , and in ngc 604 to 2.4 .
this may in part be a reddening issue , since the stellar distribution in ngc 588 is more compact , with less scope for differential reddening across the association , than the others . indeed , dust lanes are prominent in the continuum images of ngc 604 in both the optical continuum and uv / optical colors ; it is clear observationally that the uv reflects only the least - extinguished stars in this object .
the observed color distribution in ngc 604 is rich in the bluest stars , but these stars are also represented with similar color in ngc 588 and 595 .
a more important difference is that the richer population in ngc 604 includes several stars with f170w magnitude brighter than seen in any of the other clusters . while richness effects mean that the brighter star - forming regions will have brighter first - ranked stars , the studies referenced above suggest that ngc 604 has undergone a distinct burst of star formation about 3 myr ago , younger than the other regions , an event which is recent enough to leave its mark on the massive - star population .
we now focus on these brightest stars .
these data allow us to address how much the brightest stars dominate the mid - uv flux in each case , incorporating the total f170w flux within the fuse aperture .
fluctuations in the bright population could affect the integrated spectra of the lower - luminosity , sparser regions more strongly simply from statistics , but the detailed recent history of star formation will enter through aging as well .
given the temperature range of these stars ( @xmath50 k following bruhweiler et al . )
, the hottest stars should be more dominant at the shorter wavelengths of the fuse data .
it is not clear that all the far - uv flux is from direct starlight ; hill et al .
( 1995a , b ) and malamuth et al .
( 1996 ) showed evidence that a significant fraction of mid - uv light from similar systems is scattered .
the hst f300w image of ngc 604 shows reflection nebulosity , and smoothed versions of the mid - uv stis images match its morphology , indicating that scattering important at shorter wavelengths as well , as would be expected for a roughly @xmath51 rayleigh behavior . since the fuse aperture is not much larger than the stellar distribution in ngc 604 , the widths of narrow absorption features are not a sensitive test of whether scattering is important at these wavelengths .
we do see a role for scattering around some of the brightest stars in the hst stis image at 2400 , from analysis of the point - spread functions of stars .
this is easier to interpret than the distinct emission and absorption structures , since reflected continuum can be confused with emission from the weak [ o ii ] doublet near 2471 .
some of the stars match the nominal psf closely , while two of the brightest ones exhibit excess light from 0.3 - 1.0 " from the core .
this excess contains as much of 28% of the mid - uv light in the brightest case .
these data leave open the possibility that scattered light is important at the shorter far - uv wavelengths .
we therefore bracket the total uv flux between the sum of detected stars and the large - aperture sum . for larger fractions of scattered light ,
the brightest individual stars are more dominant , since the total number of stars producing the observed light is smaller .
cumulative distributions of f170w magnitude are shown in fig .
8 , along with simple geometric estimates of corrections for crowding ( which are small , but underestimate the actual effect where stars are more clumped than random within the h ii region ) .
accumulating the flux from the bright stars , we find that only about 20% of the total flux within the fuse aperture at 1700 comes directly from the stars , less than half of what bruhweiler et al .
( 2003 ) find at 1200 within a narrow slit .
this may be a sign that scattering is important , since much of the scattered light we see at mid - uv wavelengths in the hst imagery is on larger scales than this .
in contrast , as a fraction of the light _ from detected stars _ , the brightest ten ( a good approximation to the stars brighter than any found in the other h ii regions ) contribute about 40% , more in line with the bruhweiler et al . results .
this also makes sense for these stars being able to affect the overall spectrum .
however , the high temperature needed for the excess component in ngc 604 means that we are seeing a difference in history rather than simply small - number statistics in what stars appear at a given time .
in general , metallicity will be manifested in the composite spectra both directly , through photospheric and wind lines , and indirectly , as the evolutionary tracks of stars change with abundances .
effects on the initial - mass function are too small to see at the abundances found in the m33 disk , as shown by malamuth et al .
( 1996 ) for some of the same h ii regions we observed .
they suggest that the excitation trends seen in the associated ionized gas result from the different emergent ionizing fluxes for stars at various metallicities . however , different evolutionary histories are still implied by the strong changes in wind properties seen with metallicity .
comparison of the our spectra suggests that very recent events in the star - formation history do have observable impact in the far - uv , as exemplified by ngc 604 .
its mix includes light from a greater fraction of hot stars than the other systems , which fits with the other properties of this region in suggesting a burst so recent that even in the far - uv , it does not look like constant star - formation ( in this case , ages @xmath52 myr ) . in general , the timescales may be metallicity - dependent ; robert et al .
( 2003 ) show that evolved o i / iii and b i / iii stars appear later at lower abundances .
we have used fuse spectra of star - forming regions in nearby galaxies , whose gas - phase metallicities range from 0.05 - 0.8 solar , to explore the utility of far - ultraviolet spectra in measuring the abundances in star - forming galaxies , as well as to probe the massive - star populations in these galaxies .
the absorption lines from radiatively - driven winds prove to be very sensitive to metal abundance ; all six strong and unblended species ( including c iv from archival iue data ) have a strong , monotonic metallicity dependence . for n iv , c iii , and p
v , the relation between straighforward equivalent - width values and oxygen abundance from emission - line spectra is closely linear , suggesting that these lines will be useful in tracing the chemical history of galaxies from @xmath40 , beyond which the lyman @xmath2 forest makes even composite spectra progressively less informative .
the continuum of ngc 604 departs from the uniform shape of the other objects below 950 . after considering the effects of small - number statistics among the massive stars in these objects ,
we conclude that this difference probably traces to a discrete burst of star formation @xmath1 myr ago in ngc 604 .
this region had been considered by several previous studies to have hoisted such a burst , on grounds of both morphology of the gas and fitting of the h - r diagram .
andersson was helpful in understanding some of the issues in scheduling and data analysis from fuse .
we thank alessandra aloissi and her collaborators for providing their summed fuse spectrum of i zw 18 .
we also acknowledge the community service provided by steven mccandless in making his h@xmath0ools compilation of data and routines available .
dick tipping patiently explained some of the intricacies of the h@xmath0 spectrum on several occasions .
this work was supported by nasa through fuse gi grant nag5 - 8959 .
we also made use of the mast archive system in retrieving data from hst , iue , and uit .
we thank the referee , claus leitherer , for a detailed , expeditious , and helpful critique .
lcccrcl object & ra & dec & @xmath53 & exposure & o / h & source & @xmath54 & @xmath55 & km s@xmath6 & seconds & ( solar units ) & ngc 588 & 01 32 45.50 & + 30 38 55 & -174 & 5220 & 0.41 & vilchez et al .
1988 ngc 592 & 01 33 12.27 & + 30 38 49 & -162 & 3965 & 0.48 & interpolated ngc 595 & 01 33 33.60 & + 30 41 32 & -178 & 7123 & 0.56 & vilchez et al .
1988 ngc 604 & 01 34 32.50 & + 30 47 04 & -226 & 7151 & 0.66 & vilchez et al .
1988 ngc 5461 & 14 03 41.30 & + 54 19 05 & 298 & 5189 & 0.68 & luridiana et al .
2002 ngc 7714 & 23 36 14.0 & + 02 09 19 & 2798 & 6023 & 0.81 & gonzalez - delgado et al .
1995 i zw 18 & 09 34 02.30 & + 55 14 25 & 751 & 95097 & 0.05 & izotov et al .
1999 lcccccc object & n iv 955 & c iii 977 & n iii 991 & n ii 1083 & p v 1122 & c iv 1549i zw 18 & 0.27 : & 0.47 & @xmath56 ( @xmath56 ) & 1.11 ( 1.02 ) & 0.27 & 2.0:m33 avg & 1.26 & 1.19 & ... ( ... ) & 0.86 ( 0.24 ) & 0.66 & 8.4 ( 7.8 ) ngc 604 & 2.17 & 1.83 & 1.77 ( 0.59 ) & 1.50 ( 1.33 ) & 0.89 & 9.49 ( 7.97 ) ngc 5461 & 2.11 & 1.56 & 2.40 ( 1.87 ) & 2.19 ( 1.92 ) & 0.90 & 7.82 ( 6.03)ngc 7714 & 3.26 & 2.56 & 3.68 ( 2.36 ) & 2.52 ( 2.52 ) & 1.01 & 10.1 ( 10.2 ) transition & @xmath57 & @xmath58 & @xmath59 & @xmath60 & @xmath61 & @xmath62 & @xmath63 + n iv & 0.30 & -0.656 & 5.2201 & 0.063 & 0.0958 & -5.8314 & 20.8804 +
n iii & 0.05 & -1.079 & 6.5539 & 0.082 & 0.0763 & -0.0794 & 26.2157 + n ii & 1.28 & -4.029 & 7.0447 & 0.286 & 0.0710 & -19.7753 & 28.1790 + c iv & 1.14 & 19.210 & -10.8494 & 0.885 & -0.0461 & 418.6269 & -43.3975 + c iii & 0.50 & -0.502 & 3.6120 & 0.070 & 0.1384 & -6.9750 & 14.4481 + p v & 0.22 & 0.910 & 0.1037 & -4.386 & 4.8224 & 0.7356 & 0.41473 + | we present fuse spectroscopy and supporting data for star - forming regions in nearby galaxies , to examine their massive - star content and explore the use of abundance and population indicators in this spectral range for high - redshift galaxies .
new far - ultraviolet spectra are shown for four bright h ii regions in m33 ( ngc 588 , 592 , 595 , and 604 ) , the h ii region ngc 5461 in m101 , and the starburst nucleus of ngc 7714 , supplemented by the very - low - metallicity galaxy i zw 18 . in each case , we see strong milky way absorption systems from h@xmath0 , but intrinsic absorption within each galaxy is weak or undetectable , perhaps because of the uv bias " in which reddened stars which lie behind molecular - rich areas are also heavily reddened .
we see striking changes in the stellar - wind lines from these populations with metallicity , suggesting that c ii , c iii , c iv , n ii , n iii , and p v lines are potential tracers of stellar metallicity in star - forming galaxies .
three of these relations - involving n iv , c iii , and p v - are nearly linear over the range from o / h=0.050.8 solar .
the major difference in continuum shapes among these systems is that the giant h ii complex ngc 604 has a stronger continuum shortward of 950 than any other object in this sample .
small - number statistics would likely go in the other direction ; we favor this as the result of a discrete star - forming event @xmath1 myr ago , as suggested by previous studies of its stellar population . |
in many numerical simulations in general relativity one integrates einstein s field equations on a spatially compact domain with artificial timelike boundaries , effectively truncating the computational domain .
this raises the question of how to specify boundary conditions . in this article
we address this question within the context of the cauchy formulation , in which the field equations split into evolution and constraint equations .
we adopt a free evolution approach , in which the constraints are solved on the initial time slice only and the future of the initial slice is computed by integrating the evolution equations .
boundary conditions within this approach should ideally satisfy the following three requirements : ( i ) be compatible with the constraints in the sense of guaranteeing that initial data which solves the constraint equations yields constraint - satisfying solutions , ( ii ) permit to control , in some sense , the gravitational degrees of freedom at the boundary , and ( iii ) be stable in the sense of yielding a well posed initial - boundary value problem ( ibvp ) .
there are several motivations for the construction of boundary conditions satisfying the above properties .
first , recent detailed analysis of binary neutron star evolutions @xcite showed that the presence of artificial boundaries in current state of the art evolutions , which use rather ad hoc boundary conditions , dramatically affects the dynamics in the strong field region , near the stars .
while it can be argued that this effect should disappear when placing the boundaries further and further away from the strong field region , until ideally , the region of interest in the computational domain is causally disconnected from the boundaries , in practice , this would require huge computer resources , especially in the three - dimensional case . even though some kind of mesh refinement should help in placing the boundaries far away , there is in any case a minimum resolution needed in the far region in order to reasonably represent wave propagation , thus constraining the size of the computational domain for a given amount of memory .
next , an understanding of boundary conditions is important if the evolution equations include elliptic equations , since in this case effects from the boundaries can propagate with infinite speed , and so have an immediate effect on the fields being evolved .
examples of cases in which elliptic equations arise include elliptic gauge conditions ( such as maximal slicing @xcite or minimal distortion conditions @xcite ) and constraint projection @xcite methods .
finally , isolating the physical degrees of freedom at the boundaries should be important in view of cauchy - characteristic ( see @xcite for a review ) or cauchy - perturbative @xcite matching techniques , where a cauchy code is coupled to a characteristic or perturbative `` outer module '' which is well adapted to carry out the evolution in the far zone . in these cases
, it is important to communicate only the physical degrees of freedom at the boundary between the two codes , since the cauchy code and the outer module might be based on completely different formulations .
boundary conditions satisfying requirement ( i ) can be constructed by analyzing the constraint propagation system , which constitutes an evolution system for the constraint variables and is a consequence of bianchi s identities and the evolution equations .
if it can be cast into first order symmetric hyperbolic form , the imposition of homogeneous maximally dissipative boundary conditions @xcite for the constraint propagation system guarantees that a smooth enough solution of the evolution system ( if it exists ) which satisfies the constraints initially automatically satisfies the constraints at later times .
homogeneous maximally dissipative boundary conditions consist in a linear relation between the in- and outgoing characteristic fields of the system , which is chosen such that an energy estimate can be derived .
this energy estimate implies that the unique solution of the constraint propagation system with zero initial data is zero , i.e. that the constraints are preserved during evolution .
maximally dissipative boundary conditions for the constraint variables usually translate into differential conditions for the fields satisfying the main evolution equations , since the constraint variables depend on derivatives of the main variables .
this means that the resulting boundary conditions for the main system usually are not of maximally dissipative type and , as discussed below , analyzing well posedness of the corresponding ibvp becomes more difficult .
requirement ( ii ) , controlling the physical degrees of freedom , is a difficult one since there are no known _ local _
expressions for the energy or the energy flux density in general relativity .
nevertheless , one should be able to control the physical degrees of freedom in some approximate sense , as for example in the weak field regime approximation , in which one linearizes the equations around flat spacetime . in this approximation
, it might be a good idea to specify conditions through the weyl tensor , since it is invariant with respect to infinitesimal coordinate transformations of minkowski spacetime . more precisely ,
since there are two gravitational degrees of freedom , we should specify two linearly independent combinations of the components of the weyl tensor .
below we will discuss boundary conditions which involve the newman - penrose complex scalars @xmath1 and @xmath2 ( see , for instance , @xcite ) with respect to a null tetrad adapted to the time - evolution vector field and the normal to the boundary .
if the boundary is at null infinity these scalars represent the in- and outgoing gravitational radiation , respectively .
furthermore , it turns out that these scalars are invariant with respect to infinitesimal coordinate transformations and tetrad rotations for linear fluctuations of any petrov type d solution represented in an adapted background tetrad @xcite .
this class of solutions not only comprises flat spacetime but also the family of kerr solutions describing stationary rotating black holes . since in many physical situations
one is interested in modeling asymptotically flat spacetimes , such that if the outer boundary is placed sufficiently far away from the strong field region spacetime can be described by a perturbed kerr black hole , we expect the boundary condition @xmath3 to be a good approximation for a `` non - reflecting '' wave condition .
these boundary conditions are actually part of the family of conditions imposed in the formulation of ref .
@xcite , to date the only known well posed initial - boundary value formulation of the vacuum einstein equations , and were also considered in @xcite . requirement ( iii ) , the well posedness of the resulting ibvp , turns out to be a difficult problem as well : for quasilinear symmetric hyperbolic systems with maximally dissipative boundary conditions there are well - known well posedness theorems @xcite which state that a ( local in time ) solution exists in some appropriate hilbert space , that the solution is unique , and that it depends continuously on the initial and boundary data .
the proof of ref .
@xcite is based on these techniques . there , using a formulation based on tetrad fields , the authors manage to obtain a symmetric hyperbolic system by adding suitable combinations of the constraints to the evolution equations in such a way that the constraints propagate tangentially to the boundary . in this way
, the issue of preserving the constraints becomes , in some sense , trivial .
( see @xcite for a treatment in spherical symmetry in which the constraints propagate tangentially to the boundary as well . )
however , for the more commonly used metric formulations it seems difficult to achieve tangential propagation for the constraints with a symmetric or strongly hyperbolic system , and therefore , one has to deal with either constraint propagation across the boundary or systems that are not strongly or symmetric hyperbolic .
here we choose to deal with constraint propagation across the boundary and strongly or symmetric hyperbolic systems .
there has been a lot of effort in understanding such systems , both at the analytical @xcite and numerical @xcite level .
although partial proofs of well posedness have been obtained using symmetric hyperbolic systems with maximally dissipative boundary conditions @xcite , it seems that these kind of boundary conditions are not flexible enough since constraint - preserving boundary conditions usually yield differential conditions . in this article
we construct constraint - preserving boundary conditions ( cpbc ) for a family of first order strongly and symmetric hyperbolic evolution systems for einstein s equations . for definiteness , we focus on the formulation presented in ref .
@xcite , which is a generalization of the einstein - christoffel type formulations @xcite with a bona - mass type of gauge condition @xcite for the lapse . however , our approach is quite general and should also be applicable to other hyperbolic formulations of einstein s equations .
in section [ sect : shlg ] we briefly review the family of formulations considered here and recall under which conditions the main evolution equations are strongly or symmetric hyperbolic . in section [
sect : escv ] we discuss the constraint propagation system and analyze under what circumstances it is symmetric hyperbolic .
having cast this system in symmetric hyperbolic form we impose cpbc via homogeneous maximally dissipative boundary conditions for the constraint variables .
these conditions are differential boundary conditions when expressed in terms of the fields satisfying the main system .
we complete these boundary conditions in section [ sect : bc ] by imposing extra conditions which control the physical and gauge degrees of freedom .
there are several possibilities for doing so , of which we analyze the following two : 1 ) we first specify algebraic conditions in the form of a coupling between the outgoing and ingoing characteristic fields ( referred to as cpbc without weyl control later in this article ) ; 2 ) we specify boundary conditions via the weyl scalars @xmath1 and @xmath2 , as discussed above ( referred to as cpbc with weyl control later in this article )
. section [ sect : fl ] is devoted to an analysis of the well posedness of the resulting ibvp . as mentioned above ,
this is a difficult problem since our boundary conditions are not in algebraic form .
i particular , they are not in maximally dissipative form , so we can not apply the standard theorems for symmetric systems with maximally dissipative boundary conditions .
therefore , our goal here is more modest : we analyze the ibvp in the `` high frequency limit '' by considering high - frequency perturbations of smooth solutions . in this regime
the equations become linear with constant coefficients , and the domain can be taken to be a half plane . in this case solutions
can be constructed explicitly by performing a laplace transformation in time and a fourier transformation in the spatial directions that are tangential to the boundary .
this leads to the verification that a certain determinant is non - zero . if this determinant condition is violated , the system admits ill posed modes growing exponentially in time with an arbitrarily small growth time .
therefore , the determinant condition yields _ necessary _ conditions for well posedness of the resulting ibvp and allows us to discard several cases which would lead to an ill posed formulation .
_ these ill posed formulations appear even in cases in which both the main and the constraint propagation system are symmetric hyperbolic_. we also stress that the determinant condition verified in this article is a weaker version of the celebrated kreiss condition @xcite , which yields well posed ibvp for hyperbolic problems with _ algebraic _ boundary conditions @xcite .
however , the kreiss condition guarantees nothing for the present case of _ differential _ boundary conditions .
see @xcite for an example of an ibvp with differential boundary conditions which satisfies the kreiss condition but fails to be well posed in @xmath4 .
in section [ sect : nr ] we discretize the ibvp by the method of lines .
the spatial derivatives are discretized using finite difference operators that satisfy the summation by parts property and we explain in detail how we numerically implement our cpbc . then
, in sections [ sect : sim1 ] and [ sect : sim ] we perform the following @xmath0d numerical tests : first , we evolve ibvps which fail to fulfill the determinant condition and are therefore ill posed .
the goal of evolving these systems is to confirm the expected lack of numerical stability .
we do confirm such instability , finding an obvious lack of convergence : the results exhibit exponentially in time growing modes , where the exponential factor gets larger as resolution is increased , as predicted by our analytical analysis .
next , we focus on evolutions of two systems that do satisfy the determinant condition , differing only on whether they control components of the weyl tensor at the boundary or not .
finally , we compare stable evolutions of systems with cpbc with evolutions where maximally dissipative boundary conditions are given for the _ main _ evolution system , and are therefore expected to violate the constraints .
we first analyze evolutions of these four systems through a _ robust stability test _ @xcite . in this test
, random initial and boundary data is specified at different resolutions , and the growth rate in the time evolved fields is observed .
a growth rate that becomes _ larger _ as resolution is increased is a strong indication of a numerical instability , while for numerical stability the growth rate should be _ bounded _ by a constant that is independent of resolution .
we find that systems that violate the determinant condition fail to pass the robust stability test , as expected .
however , we _ also _ find that at least some systems with cpbc with weyl control which satisfy the determinant condition are numerically unstable as well , although in a somehow weaker sense ( explained in the text ) that reminds the numerical evolution of weakly hyperbolic systems @xcite .
in contrast to this , the systems with cpbc without weyl control which satisfy the determinant condition that we have evolved pass the robust stability test for the length of our simulations ( usually between @xmath5 and @xmath6 crossing times ) .
next , we concentrate on evolutions of brill waves @xcite , and confirm the expectations drawn from the robust stability test in what concerns numerical stability .
using these waves we further concentrate on a detailed comparison between the results using maximally dissipative boundary conditions for the main evolution system and stable cpbc .
our convergence tests strongly suggest that in the former case the constraint variables _ do not _ converge to zero in the limit of infinite resolution , implying that one _ does not _ obtain a solution to einstein s field equations , while in the latter case for the same resolutions the constraint variables do converge to zero .
next we concentrate on the stable cpbc case and evolve pure gauge solutions , using high order accurate finite difference operators which satisfy the summation by parts property .
the operators used are eighth order accurate in the interior points and fourth order accurate at and near the boundary points . as a final numerical experiment
, we also concentrate on the stable cpbc case and inject pulses of gravitational radiation through the boundaries of an initially flat spacetime .
we inject pulses of large enough amplitude to create very large curvature in the interior ( as measured by curvature invariants ) , showing that our cpbc are strong enough to handle very non - linear dynamics .
the order of magnitude achieved by the curvature invariant measured corresponds to being at roughly @xmath7 from the singularity , in a schwarzschild spacetime of mass one . in these simulations
this curvature is produced _ solely _ by the injected pulses .
a summary of the results and conclusions are presented in sect .
[ sect : conc ] .
technical details , like the derivation of the constraint propagation system and of the characteristic fields , and a special family of solutions to the linearized ibvp with weyl control are found in [ app : mecps ] , [ app : charfields&sh ] and [ app : sfs ] .
in this section we review the family of hyperbolic formulations of einstein s field equations constructed in @xcite , which is an extension of the einstein - christoffel type @xcite of formulations which incorporates a generalization of the bona - mass @xcite slicing conditions .
it consists of @xmath8 evolution equations for the variables @xmath9 , where @xmath10 is the lapse function , @xmath11 the three - metric , @xmath12 the extrinsic curvature , and where the extra variables @xmath13 and @xmath14 represent the first order spatial derivatives of the three - metric and of the logarithm of the lapse , respectively .
the evolution equations are obtained from the adm evolution equations in vacuum by adding suitable constraints to the right - hand side of the equations ( see [ app : mecps ] for more details ) . following the notation of ref .
@xcite , we have @xmath15 here , @xmath16 , and @xmath17 is a function that is smooth in all its arguments and that satisfies @xmath18 . for
the simulations below we shall choose @xmath19 which corresponds to time - harmonic slicing , but for our analytical results we shall leave @xmath20 unspecified for generality .
we assume that the shift vector @xmath21 is a fixed , a priori specified vector field .
the parameters @xmath22 , @xmath23 , @xmath24 , @xmath25 , @xmath26 control the dynamics off the constraint hypersurface , defined by the vanishing of the following expressions @xmath27ij}\ ; , \label{eq : clkij}\\ c^{(a)}_i & \equiv & a_i - \partial_i n / n\ ; , \label{eq : cai}\\ c^{(a)}_{ij } & \equiv & \frac{1}{n } \partial_{[i } ( n a_{j]})\ ; . \label{eq : caij}\end{aligned}\ ] ] here , @xmath28 is the hamiltonian constraint , @xmath29 the momentum one , and @xmath30 , @xmath31 , @xmath32 and @xmath33 are artificial constraints that arise as a consequence of the introduction of the extra variables @xmath13 , @xmath14 . and @xmath34 are automatically zero if the constraints @xmath30 , @xmath32 are satisfied , and so they are redundant .
however , we will need these variables in the next section in order to cast the constraint propagation system into first order form . ]
the ricci tensor @xmath35 belonging to the three - metric is written as @xmath36 where @xmath37 , @xmath38 and @xmath39 the evolution equations ( [ eq : ndot],[eq : gdot],[eq : kdot],[eq : ddot],[eq : adot ] ) have the form of a quasilinear first order system , @xmath40 where @xmath41 and the matrix - valued functions @xmath42 ,
@xmath43 , and the vector - valued function @xmath44 are smooth .
we are looking for solutions with given initial data on some three - dimensional manifold which satisfies the constraints equations . in order to guarantee the existence of solutions
we restrict the freedom in choosing the parameters @xmath45 , @xmath22 , @xmath23 , @xmath24 , @xmath25 , @xmath26 by demanding that the corresponding initial - value formulation is well posed .
this means that given smooth initial data , a ( local in time ) solution should exist in some appropriate hilbert space , be unique , and depend continuously on the initial data .
although in this article we are interested in the numerical evolution of spacetimes on a spatially compact region with boundaries , well posedness of the problem in the absence of boundaries is a necessary condition for obtaining a numerical stable and consistent evolution inside the domain of dependence of the initial slice . an easy and intuitive way of finding necessary conditions for well posedness is to look at high frequency perturbations of smooth solutions : let @xmath46 be a fixed point and @xmath47 a smooth solution in a neighborhood of @xmath48 .
perturb @xmath47 according to @xmath49 , with @xmath50 , @xmath51 real , and @xmath52 a constant one - form on @xmath53 which is normalized such that @xmath54 .
if we evaluate the evolution equations at a point near @xmath48 , divide by @xmath55 , and take first the limit @xmath56 and then the limit @xmath57 , the evolution equations reduce to @xmath58 \tilde{u } , \label{eq : frozen}\ ] ] where the matrix @xmath59 is given by @xmath60 \\ -2n_k k_{ij } + \eta\,g_{k(i}\left(k_{j)n } - n_{j ) } k \right ) + \chi\,g_{ij}\left ( k_{kn } - n_k k \right ) \\
-2\sigma n_i k + \xi\,\left ( k_{in } - n_i k \right ) \end{array } \right ) , \nonumber\\ \label{eq : principalsymb}\end{aligned}\ ] ] where we have set @xmath61 and where the index @xmath62 refers to the contraction with @xmath63 .
notice that by rescaling and rotating the coordinates @xmath64 one can always achieve that @xmath65 , and by rescaling the coordinate @xmath66 one can achieve that @xmath67 .
for this reason , we drop the entry @xmath48 in the following .
we call the system ( [ eq : frozen ] ) the associated frozen coefficient problem
. a necessary condition for the well posedness of the initial - value formulation defined by the ( non - linear ) eqs .
( [ eq : ndot],[eq : gdot],[eq : kdot],[eq : ddot],[eq : adot ] ) is the well posedness of the associated frozen coefficient problem .
if some extra smoothness properties are satisfied ( see [ app : charfields&sh ] ) this condition is also a sufficient one .
the frozen coefficient problem is well posed if the matrix @xmath68 is diagonalizable and has only real eigenvalues for each @xmath62 . clearly , this is true if and only if the matrix @xmath69 is diagonalizable and has only real eigenvalues .
as we have shown in @xcite this can be easily analyzed by taking advantage of the block structure of @xmath69 : suppose @xmath70 is an eigenvector of @xmath71 with eigenvalue @xmath72
. then @xmath73 where @xmath74 , @xmath75 , @xmath76 , and where the matrices @xmath77 and @xmath78 are read off from eq .
( [ eq : principalsymb ] ) . a sufficient condition for @xmath69 to be diagonalizable and
posses only real eigenvalues can be obtained by considering the equation @xmath79 explicitly , we have @xmath80 where the coefficients @xmath81 , @xmath82 and @xmath83 are @xmath84 we now demand that @xmath85 is diagonalizable and has only positive eigenvalues . as we have shown in @xcite this guarantees that @xmath69 is diagonalizable and has only real eigenvalues .
representing @xmath85 in an orthonormal basis @xmath86 , @xmath87 , @xmath88 such that @xmath89 , we find @xmath90 where @xmath91 and @xmath92 . from this
we immediately see that @xmath85 is diagonalizable with only positive eigenvalues if and only if @xmath93 are positive and if @xmath94 whenever @xmath95 . in [ app : charfields&sh ] we derive the characteristic fields , which are given by the projections of @xmath70 onto the eigenspaces of @xmath69 .
these fields play an important role in the construction of boundary conditions . using these fields
, we also derive in [ app : charfields&sh ] sufficient conditions for the non - linear evolution system to be strongly hyperbolic and thus yield a well posed initial value formulation .
in order to obtain a solution to einstein s equations , not only do we have to solve the evolution equations but also the constraints .
we want to follow a free evolution scheme , in which the constraints are solved initially only . for this scheme to be valid
, we have to guarantee that any solution for such initial data automatically satisfies the constraints in the computational domain everywhere and at every time . at the numerical level , since the constraints are already violated initially due to truncation or roundoff errors , we have to guarantee that the numerical solution to the evolution equations converge to a constraint - satisfying solution to the continuum equations in the limit of infinite resolution . in order to show this , the constraint propagation system , which gives the change of the constraint variables under the flux of the main evolution system , plays an important role . in this section
we show that for a suitable range of the parameters @xmath22 , @xmath23 , @xmath24 , @xmath25 , @xmath26 this system can be cast into first order symmetric hyperbolic form .
we then specify boundary conditions that guarantee that zero initial data for this system leads to zero constraint variables at later times .
the constraint propagation system is derived in [ app : mecps ] , up to lower order terms which are linear algebraic expressions in the constraint variables and whose precise form are not needed for the purpose of this article . in order to analyze under which conditions the system is symmetric hyperbolic , it is convenient to decompose @xmath96 into its trace and trace - less parts , @xmath97 where @xmath98(ij)}$ ] is trace - less with respect to all pair of indices and where @xmath99 and @xmath100 are determined by the traces @xmath101 , @xmath102s}$ ] and @xmath103 , @xmath104 ( notice that @xmath105 is symmetric trace - less while @xmath106 and @xmath107 are antisymmetric . ) in terms of the variables @xmath108 , where @xmath109 , the non - linear constraint propagation system has the form @xmath110 where the principal symbol @xmath111 is given by @xmath112 \\
\left[\frac{1}{4}(2\chi-\eta ) + \xi \right]\ , n_{[i } c_{j ] } \\ ( \eta + 3\chi)\ , n_{[i } c_{j ] } \\ \xi\ ; n_{[i } c_{j ] } \end{array } \right),\ ] ] and where the matrix @xmath113 depends on the main fields @xmath70 and their spatial derivatives , but not on @xmath114 .
the system eq .
( [ eq : evolconstr ] ) is called _ symmetric hyperbolic _ if there exists a symmetric positive definite matrix @xmath115 which may depend on @xmath70 but not on @xmath62 such that @xmath116 is symmetric for all @xmath70 and @xmath62 . from the above representation of the principal symbol it is not difficult to see that the system is symmetric hyperbolic if the following conditions hold @xmath117 notice that these conditions automatically imply that @xmath118 and @xmath119 , which are necessary conditions for the main evolution system to be strongly hyperbolic .
if the conditions ( [ eq : constrsym1],[eq : constrsym2],[eq : constrsym3 ] ) are satisfied , a symmetrizer is given by the quadratic form @xmath120 where @xmath121 the symmetrizer allows us to obtain an energy estimate for solutions to eq .
( [ eq : evolconstr ] ) on a domain @xmath122 of @xmath53 with smooth boundary @xmath123 and suitable boundary conditions on @xmath123 . defining the energy norm @xmath124 differentiating with respect to @xmath66 and using the constraint propagation system , eq .
( [ eq : evolconstr ] ) , we obtain @xmath125\ ; d^3 x \nonumber\\ & = & \int_{\partial o } u^t { \bf h } n{\cal a}_c(n ) u \ ; d^2 x \nonumber\\ & + & \int_{o } u^t \left [ n { \bf h}{\cal b } + n { \cal b}^t { \bf h } - \partial_i(n{\bf h}{\cal a}_c^i + { \bf h}\beta^i ) \right ] u\ ; d^3 x,\end{aligned}\ ] ] where here @xmath52 denotes the unit outward one - form to the boundary . in the last step
we have used the fact that @xmath126 is symmetric with respect to the scalar product defined by ( [ eq : sym ] ) and assumed that the shift is tangential to the boundary at @xmath127 .
as one can easily verify , @xmath128 where @xmath129 are the in @xmath130 and out @xmath131 going characteristic constraint fields . therefore ,
if we impose the boundary conditions @xmath132 where the matrix @xmath133 satisfies @xmath134 for all one - forms @xmath135 , we obtain the energy estimate @xmath136 where the constant only depends on bounds for @xmath137 and @xmath138 . since @xmath139 this proves that @xmath140 for all @xmath141 provided that @xmath142 . for this reason ,
we call the three conditions ( [ eq : cpbc ] ) _ constraint - preserving boundary conditions_.
in this section we consider the main evolution system , eqs .
( [ eq : ndot],[eq : gdot],[eq : kdot],[eq : ddot],[eq : adot ] ) , on a open domain @xmath122 of @xmath53 with smooth boundary @xmath127 .
we also assume that the shift vector is chosen such that it is tangential to @xmath123 at the boundary .
this means that at the boundary we have six ingoing characteristic fields , denoted by @xmath143 , @xmath144 , @xmath145 , @xmath146 ( see [ app : charfields&sh ] for their definition ; here and in the following , @xmath147 , and quantities with a hat denote trace - free two by two matrices ) , and thus we have to provide six independent boundary conditions . following the classification scheme of ref .
@xcite one can show that the first field is a gauge field , the second ones are physical fields , and the last are constraint - violating fields .
we stress that this classification scheme does only make precise sense in the linearized regime for plane waves propagating in the normal direction to the boundary ( see @xcite for a more detailed discussion about this ) .
if we forgot about the constraints , we could give data to the six ingoing fields .
the simplest possibility would be to freeze the ingoing fields to their values given by the initial data . provided the evolution system is symmetric hyperbolic this would yield a well posed ibvp .
however , in the presence of constraints , the boundary conditions have to ensure that no constraint - violating modes enter the boundary , i.e. the boundary conditions have to be compatible with the constraints .
in fact , the numerical results of section [ sect : sim ] show explicitly that freezing of the ingoing fields to their initial values does not , in general , provide a solution to einstein s equations : the constraints are violated . instead of the freezing non - constraint preserving boundary conditions just mentioned ,
we impose the three constraint - preserving boundary conditions ( [ eq : cpbc ] ) .
this fixes three of the six conditions we are allowed to specify at @xmath123 .
we complete these three conditions in the following two ways : 1 .
* cpbc without weyl control * + here we adopt the simplest possibility and impose algebraic boundary conditions on the `` gauge '' and `` physical '' fields , @xmath148 where @xmath149 and @xmath150 are constants satisfying @xmath151 , @xmath152 , and @xmath153 and @xmath154 are functions on @xmath123 describing the boundary data . the justification for the bounds on @xmath149 and @xmath150 will become clear in the next section .
notice that the choice @xmath155 results in dirichlet conditions in the sense that data is imposed on some components of the extrinsic curvature while the choice @xmath156 yields boundary conditions that impose data on combinations of spatial derivatives of the three metric ( see the definitions of @xmath143 and @xmath144 in [ app : charfields&sh ] ) . in our simulations
below , we choose @xmath157 which yields sommerfeld - like conditions .
* cpbc with weyl control * + here we replace eq .
( [ eq : maxdissmodes ] ) by a similar condition for the weyl tensor .
we impose the boundary conditions @xmath158 where @xmath159 and where @xmath160 are defined in terms of the electric ( @xmath161 ) and magnetic ( @xmath99 ) parts of the weyl tensor in the following way : let @xmath162 , @xmath163 , @xmath164 be an orthonormal triad with respect to the three metric @xmath11 such that @xmath162 coincides with the unit outward normal @xmath63 to the boundary .
then , @xmath165 , where @xmath166 is the volume element associated to @xmath11 .
if the vacuum equations hold , the electric and magnetic components can be determined by @xmath167 where one uses the evolution equation ( [ eq : kdot ] ) in order to reexpress @xmath168 in terms of spatial derivatives of the main variables .
the boundary conditions ( [ eq : maxdissweyl ] ) correspond to the conditions imposed by friedrich and nagy @xcite .
in the symmetric hyperbolic system considered in @xcite , where the components of the weyl tensor are evolved as independent fields , these boundary conditions arise naturally when analyzing the structure of the equations since they give rise to maximally dissipative boundary conditions .
in particular , a well posed initial - boundary value problem incorporating the condition ( [ eq : maxdissweyl ] ) is derived in @xcite .
in contrast to this , the boundary conditions ( [ eq : maxdissweyl ] ) are not maximally dissipative for our symmetric hyperbolic system since they are not even algebraic .
+ the conditions ( [ eq : maxdissweyl ] ) can also be expressed in terms of the newman - penrose scalars @xmath1 and @xmath2 with respect to a null tetrad which is adapted to the boundary in the following sense : let @xmath169 denote the future - directed unit normal to the @xmath170 slices .
together with the above vectors @xmath162 , @xmath163 , @xmath164 it forms a tetrad . from this , we construct the following newman - penrose null tetrad : @xmath171 then , we find @xmath172 and the boundary condition ( [ eq : maxdissweyl ] ) is simply @xmath173 where the star denotes complex conjugation ( we could generalize this boundary condition by allowing for complex values of @xmath174 ) .
notice that @xmath1 and @xmath2 are not uniquely determined by the unit normals @xmath169 and @xmath62 : with respect to a rotation of @xmath163 , @xmath164 about the angle @xmath175 , these quantities transform through @xmath176 , @xmath177 , the factor @xmath178 reflecting the spin of the graviton .
however , the boundary condition ( [ eq : maxdissweylbis ] ) is indeed invariant with respect to such transformations . in our simulations below , we shall choose @xmath179 corresponding to an outgoing radiation condition .
+ finally , as discussed in the introduction , @xmath1 and @xmath2 represent , respectively , in- and outgoing radiation when evaluated at null infinity and are gauge - invariant quantities for linearizations about a kerr background .
we now analyze the well posedness of the ibvp defined by the evolution equations ( [ eq : ndot],[eq : gdot],[eq : kdot],[eq : ddot],[eq : adot ] ) , the cpbc ( [ eq : cpbc ] ) and the boundary conditions ( [ eq : gauge ] ) and ( [ eq : maxdissmodes ] ) or ( [ eq : gauge ] ) and ( [ eq : maxdissweyl ] ) .
we derive necessary conditions for well posedness by verifying a certain determinant condition in the high frequency limit .
we assume that the parameters are such that the evolution equations are strongly hyperbolic since otherwise the problem is ill posed even in the absence of boundaries , and such that the constraint propagation system is symmetric hyperbolic .
let @xmath180 be a point on the boundary . by taking the high frequency limit
we obtain the associated frozen coefficient problem at @xmath48 .
after rescaling and rotating the coordinates if necessary , we can achieve that @xmath181 , @xmath65 and that the domain of integration is @xmath182 . in this way
, we obtain a constant coefficient problem on the half space . introducing the operator @xmath183 , it is given by @xmath184 notice that this system is equivalent to the one that one would obtain by linearizing the evolution equations around flat spacetime in a slicing with respect to which the three metric is flat , the lapse is one and the shift is constant and tangential to the boundary , but not necessarily zero .
since we have a linear constant coefficient problem on the half plane , we can solve these equations by means of a laplace transformation in time and a fourier transformation in the @xmath185 and @xmath186 directions @xcite .
that is , we write the solution as a superposition of solutions of the form @xmath187 , where @xmath188 with @xmath189 , @xmath190 , @xmath191 and @xmath192 .
( notice that for such solutions , @xmath193 . ) substituting this into eqs .
( [ eq : linkij],[eq : lindkij],[eq : linai ] ) one obtains a system of ordinary differential equations coupled to algebraic conditions . since there are six in- and six outgoing modes , there are twelve independent differential equations .
the remaining equations which are algebraic can be used in order to eliminate the characteristic fields which have zero speeds , and one ends up with a closed system of twelve linear ordinary differential equations . because the system is strongly hyperbolic we expect @xcite exactly six linearly independent solutions that decay as @xmath194 , and six solutions that blow up as @xmath194 .
since we require the solution to lie in @xmath4 we only consider the six decaying solutions .
the determinant condition consists in verifying that the boundary conditions with homogeneous boundary data annihilate these six solutions .
if the determinant condition is violated , the problem admits solutions of the form @xmath195 for some @xmath189 , where @xmath51 can be arbitrarily large , and the system is ill posed since @xmath196 as @xmath197 for each fixed @xmath66 .
thus , the determinant condition is necessary for the well posedness of the ibvp and , as we will see , will yield nontrivial conditions .
a convenient way for finding the six decaying solutions is to look at the second order equation for @xmath12 , which is a consequence of eqs .
( [ eq : linkij],[eq : lindkij],[eq : linai ] ) , @xmath198 where the coefficients @xmath81 , @xmath82 and @xmath83 are given in eqs .
( [ eq : coeffa],[eq : coeffb],[eq : coeffc ] ) .
we show below that there are exactly six solutions to this second order system which have the form @xmath199 , with @xmath189 and such that @xmath200 decays as @xmath201 . since in fourier - laplace space the operator @xmath202 is just multiplication with the nonzero factor @xmath203 , corresponding solutions to the original system
can be obtained by determining @xmath13 and @xmath14 from eqs .
( [ eq : lindkij ] ) and ( [ eq : linai ] ) , respectively .
in order to find the decaying solutions of eq .
( [ eq : secondkij ] ) we make the ansatz @xmath204 , where here and in the following we set @xmath205 for simplicity , and where @xmath206 is a complex number with negative real part to be determined .
it is also convenient to introduce a unit two - vector @xmath207 which is orthogonal to @xmath208 , and to decompose @xmath209 in the components @xmath210 , @xmath211 , @xmath212 , @xmath213 , @xmath214 and @xmath215 . using this ansatz , eq .
( [ eq : secondkij ] ) splits into the following two decoupled systems , @xmath216 where the matrices @xmath217 and @xmath218 are given by @xmath219 and @xmath220 , @xmath221 , @xmath222 are defined in eqs .
( [ eq : lam1],[eq : lam2],[eq : lam3 ] ) .
the matrix @xmath217 has the eigenvalue - eigenvector pair @xmath223 the two vectors are always linearly independent from each other since @xmath224 .
this yields the solution @xmath225 where @xmath226 , @xmath227 are two constants and where here and in the following @xmath228 , @xmath229 ( @xmath230 ) where the branch of the square root for which @xmath231 for @xmath189 is chosen .
similarly , after obtaining the eigenvalues and eigenvectors of @xmath218 one obtains the solution @xmath232 where @xmath233 therefore , we have obtained six linearly independent solutions which decay exponentially as @xmath234 and thus lie in @xmath235 .
they are parameterized by the constants ( which depend on @xmath51 and @xmath236 ) @xmath226, ... ,@xmath237 .
a necessary condition for the ibvp to be well posed is that these constants are uniquely determined by the boundary data . before checking this condition , it is instructive to have a closer look at the six - parameter family of solutions given by eqs .
( [ eq : k1 ] ) and ( [ eq : k2 ] ) .
let us first compute the momentum constraint variable @xmath238 : it has the form @xmath239 where @xmath240 where @xmath241 if @xmath242 and @xmath243 otherwise . thus , @xmath29 for this family of solutions
if and only if @xmath244 .
in other words , the three - parameter subfamily of solutions parametrized by @xmath227 , @xmath245 and @xmath246 are _ constraint - violating _ modes .
next , consider an infinitesimal coordinate transformation parametrized by a vector field @xmath247 , and assume zero shift for simplicity .
with respect to such a transformation , the linearized lapse and extrinsic curvature change according to @xmath248 on the other hand , the linearization of eq.([eq : ndot ] ) around a minkowski background for all @xmath10 , which is satisfied by the time - harmonic slicing condition adopted in our simulations .
] yields @xmath249 we see that the choice @xmath250 leaves eq .
( [ eq : ndotlin ] ) invariant and induces the transformation @xmath251 while @xmath252 and @xmath253 remain invariant .
therefore , it is possible to gauge away the solution parametrized by @xmath237 and we call this solution a _ gauge _ mode from hereon .
the remaining family of solutions parametrized by @xmath226 and @xmath254 are _ physical _ modes : they satisfy the constraints and @xmath226 and @xmath254 are gauge - invariant .
next , we verify that the integration constants @xmath226, ... ,@xmath237 are uniquely determined by the boundary conditions .
first , we notice that the expressions ( [ eq : ceta],[eq : cx],[eq : comega ] ) for the fourier - laplace transformation of the momentum constraint yield a three - parameter family of solutions for the constraint propagation system , eq .
( [ eq : evolconstr ] ) . since this system is symmetric hyperbolic and since we specify homogeneous maximally dissipative boundary conditions for it ( see eq .
( [ eq : cpbc ] ) ) , the corresponding ibvp is well posed . in particular , zero is the only solution with trivial initial data .
this implies that @xmath255,[eq : cx],[eq : comega ] ) into the fourier - laplace transformed of the cpbc ( [ eq : cpbc ] ) . ] .
we stress that such a conclusion can not be drawn if the constraint propagation is strongly but not symmetric hyperbolic , see ref .
@xcite for a counterexample . next , using eqs .
( [ eq : lindkij ] ) and ( [ eq : linai ] ) , we find the following expressions for the relevant characteristic fields at the boundary @xmath256,\\ \hat{v}^{(\pm)}_{ab } & = & \hat{k}_{ab } \mp \frac{1}{\omega z}\left [ \partial_x\hat{k}_{ab } - ( 1+\zeta)\partial_{(a } k_{b)x } \right]^{tf},\\ \hat{w}^{(\pm)}_{ab } & = & \omega z\hat{k}_{ab } \mp \left [ \partial_x\hat{k}_{ab } - \partial_{(a } k_{b)x } \right]^{tf } + \frac{1}{\omega z}\left [ -2\sigma\partial_a\partial_b k + \xi\partial_{(a } c_{b ) } \right]^{tf},\\\end{aligned}\ ] ] where @xmath257 and where @xmath258^{tf}$ ] denotes the trace - free part with respect to the metric @xmath259 . plugging into this the six - parameter family of solutions given in eqs .
( [ eq : k1 ] ) and ( [ eq : k2 ] ) , taking into account the vanishing of the constraint - violating modes , @xmath260 , and evaluating at @xmath261 , we obtain @xmath262 \sigma_6 + ( 1-\omega)\left [ \sigma_3 + \sigma_6 \mp \frac{\zeta}{\sqrt{\lambda_1}\ , z}\left ( \nu_4\sigma_3 + \nu_1\sigma_6 \right ) \right ] , \label{eq : vxx}\\ \tilde{\hat{v}}^{(\pm)}_{\omega\omega } & = & -\left [ \left(z^2 + \frac{1}{2}\right ) \pm \left ( z - \frac{\zeta}{2z } \right)\sqrt{z^2 + 1 } \right]\sigma_3 - \frac{1}{2z}\left [ z \mp \nu_1\zeta \right]\sigma_6\ ; , \\ \tilde{\hat{w}}^{(\pm)}_{\omega\omega } & = & -\omega z\left [ \left(z^2+\frac{1}{2}\right ) \pm z\sqrt{z^2 + 1 } \right]\sigma_3\ ; , \label{eq : hatwww}\\ \tilde{v}^{(\pm)}_{\omega\eta } & = & -\frac{i}{z}\left [ z\sqrt{z^2 + 1 } \pm \left ( z^2 + \frac{1}{2}(1-\zeta ) \right ) \right]\sigma_1\ ; , \\ \tilde{w}^{(\pm)}_{\omega\eta } & = & -i\omega\left [ z\sqrt{z^2 + 1 } \pm \left ( z^2 + \frac{1}{2 } \right ) \right]\sigma_1\ ; .
\label{eq : hatwwe}\end{aligned}\ ] ] in particular , the fields @xmath263 and @xmath264 are gauge - invariant .
this is to be expected since these fields are linear combinations of the components of the weyl tensor , which is a gauge - invariant quantity in the linearized regime .
the fields @xmath265 are gauge - invariant too , but the remaining fields depend on the constant @xmath237 which represents a pure gauge degree of freedom .
in particular , if the parameters of the formulation are such that one can choose @xmath266 , @xmath267 are pure gauge fields .
now it is not difficult to check the determinant condition for the different boundary conditions proposed in the previous section .
here we analyze the family of boundary conditions ( [ eq : gauge],[eq : maxdissweyl ] ) , where we set the boundary data @xmath153 and @xmath268 to zero .
we first analyze the boundary conditions @xmath269 , where @xmath270 : plugging into this the expressions from eqs .
( [ eq : hatwww],[eq : hatwwe ] ) , we obtain @xmath271\sigma_1 = 0 , \label{eq : sigma1}\\ & & \left [ ( z + \sqrt{z^2 + 1})^2 - c ( z - \sqrt{z^2 + 1})^2 \right]\sigma_3 = 0 .
\label{eq : sigma3}\end{aligned}\ ] ] since the function @xmath272 maps @xmath189 onto the _ outside _ of the unit disk minus the negative real axis , while the function @xmath273 maps @xmath189 onto the _ inside _ of the unit disk minus the negative real axis , and since @xmath270 , it follows that the gauge - invariant constants @xmath226 and @xmath254 vanish .
now it also becomes clear why we restricted the range of the parameter @xmath174 : if @xmath274 , there are nontrivial exponential growing solutions with either @xmath275 or @xmath276 , and the system is ill posed . in order to analyze the consequences of the boundary condition
@xmath277 , we assume that either @xmath266 ( in which case @xmath267 are pure gauge fields ) or @xmath278 . in both cases we obtain , using @xmath279 and eq .
( [ eq : vxx ] ) , @xmath280\left [ ( 1-a)z + ( 1+a)\sqrt{z^2+\lambda_1 } \right]\sigma_6 = 0.\ ] ] the term in the second bracket is non - vanishing for @xmath189 if and only if @xmath281 ( see lemma 1 of ref .
@xcite for a proof ) .
the term in the first bracket never vanishes provided that @xmath282 .
summarizing , the cpbc with weyl control and a choice of parameters that allows for @xmath266 or such that @xmath283 and @xmath284 does always yield an initial - boundary value formulation that satisfies the determinant condition , as long as the main evolution system is strongly hyperbolic and the constraint propagation system is symmetric hyperbolic ( see sections [ sect : shlg ] and [ sect : escv ] ) .
next , we analyze the family of boundary conditions ( [ eq : gauge],[eq : maxdissmodes ] ) where we set the boundary data @xmath153 and @xmath154 to zero . in this case , the result is less robust and depends more strongly on the choice of the parameters .
for example , assume again that @xmath266 which implies that @xmath285 .
now eq . ( [ eq : sigma3 ] ) has to be replaced by @xmath286\sigma_3 = 0.\ ] ] by taking @xmath236 real and positive and by taking the limits @xmath287 and @xmath288 one sees that the expression inside the square bracket changes its sign if @xmath289 and @xmath290 . in this case
it does not follow that @xmath254 is zero and the system suffers from the presence of ill posed modes . for simplicity , let us assume that @xmath291 and that @xmath292 .
all of the simulations below will satisfy these conditions . in those cases ,
we obtain @xmath293 = 0,\\ f_a(z)\left [ ( 1-\omega)(\sigma_3 + \sigma_6 ) + z^2\sigma_6 \right ] = 0,\end{aligned}\ ] ] where @xmath294 .
since @xmath281 , @xmath152 , @xmath295 and @xmath296 never vanish for @xmath189 and the only condition we obtain is @xmath297 .
in this section we discuss how to numerically implement the ibvp with or without weyl control . for quasilinear first order symmetric hyperbolic systems with maximally dissipative boundary conditions
there are well known methods for discretizing the problem such that numerical stability is guaranteed at the linearized level ( see @xcite for a recent application in the context of numerical relativity , and references therein for the original work ) .
unfortunately , in our case , the boundary conditions are more complicated since the cpbc ( [ eq : cpbc ] ) and the conditions ( [ eq : maxdissweyl ] ) which control the weyl tensor are not even algebraic , and so we can not apply this methods for discretizing the boundary conditions .
so we first describe our method for implementing the constraint - preserving boundary conditions , and then explain how the differential equations are discretized . for simplicity ,
we focus on the case where the coupling coefficients @xmath298 , @xmath149 , @xmath150 and @xmath174 vanish , which corresponds to sommerfeld - like conditions , as discussed in section [ sect : bc ] . the boundary conditions ( [ eq : cpbc ] ) and ( [ eq : maxdissweyl ] ) are implemented in the following way : for all points that lie on the boundary we add to the right - hand side of the evolution equations terms that are linear in the quantities @xmath299 and @xmath300 ( which are zero if the boundary conditions are satisfied ) .
more precisely , we replace the right - hand side of equations ( [ eq : kdot],[eq : ddot],[eq : adot ] ) by @xmath301 where the matrix coefficients
@xmath302 , @xmath303 , etc .
are allowed to depend on the three metric and the unit outward normal one - form @xmath52 to the boundary .
( the @xmath304 are set to zero for the case without weyl control . )
it is clear that these extra terms change the principal part of the equations .
the idea is to choose the @xmath48 s and @xmath304 in such a way that with respect to the unit outward normal @xmath52 to the boundary the ingoing characteristic fields @xmath305 , @xmath306 and @xmath307 become zero speed fields while the speed of the other fields remains unchanged .
this has the effect of eliminating the normal derivatives in the evolution equations for @xmath305 , @xmath306 and @xmath307 , and hence these equations are _ intrinsic _ to the boundary .
we will see that this requirement uniquely determines the @xmath48 s and @xmath308 s . in order to see this
we first notice that the ingoing characteristic constraint fields have the form @xmath309 + { \cal q } ( \partial_{|| } u , u ) , \\
v_a^{(+ ) } & = & -\frac{1}{2}\frac{\partial}{\partial
n } \left [ ( 1+\sqrt{\lambda_3 } ) v_{na}^{(+ ) } + ( 1-\sqrt{\lambda_3 } ) v_{na}^{(- ) } \right ] + { \cal q}_a ( \partial_{|| } u , u ) , \\
\hat{w}_{ab}^{(+ ) } & = & \frac{\partial}{\partial n}\ , \hat{v}_{ab}^{(+ ) } + { \cal q}_{ab } ( \partial_{|| } u , u),\end{aligned}\ ] ] where @xmath310 denotes the normal derivative to the boundary and where the expressions @xmath311 , @xmath312 and @xmath313 only depend on the variables @xmath70 and their derivatives @xmath314 tangential to the boundary . therefore , the principal symbol corresponding to the modified system ( [ eq : kdotmod],[eq : ddotmod],[eq : adotmod ] ) and the direction of the unit outward normal @xmath52 is given by @xmath315 where @xmath316 , ... @xmath317 and @xmath318 ,
... @xmath319 are defined in terms of @xmath320 and @xmath321 in the same way as @xmath322 , ... @xmath323 in terms of @xmath324 , and where @xmath325,\\ \tilde{v}_a^{(+ ) } & = & -\frac{1}{2}\left [ ( 1+\sqrt{\lambda_3 } ) v_{na}^{(+ ) } + ( 1-\sqrt{\lambda_3 } ) v_{na}^{(- ) } \right].\end{aligned}\ ] ] assume first that @xmath326 . in this case , we see immediately that @xmath322 , ... @xmath323 can only remain characteristic fields ( i.e. fields with respect to which the principal symbol is diagonal ) if all @xmath48 s and @xmath308 s are zero except for @xmath327 , @xmath328 and @xmath329 . then , the choice @xmath330 yields zero speeds for the variables @xmath305 , @xmath306 and @xmath307 .
the matrix coefficients @xmath302 , ...
@xmath331 are easily obtained from this by applying the inverse transformation to the one that defines the characteristic fields in terms of the main variables ( see [ app : charfields&sh ] ) . if @xmath332 or @xmath333 it is not possible to retain the fields @xmath305 , @xmath306 as characteristic fields . in this case , we replace @xmath305 , @xmath306 by @xmath334 and @xmath335 , respectively , where @xmath336 and @xmath337 are given in [ app : charfields&sh ] , and set @xmath338 and all other
@xmath48 s and @xmath308 s to zero .
this implies @xmath339 , @xmath340 , @xmath341 , so @xmath342 , @xmath343 and @xmath344 are zero speed fields while the speeds of the remaining characteristic fields are unchanged .
next , we discretize the evolution system ( [ eq : ndot],[eq : gdot],[eq : kdotmod],[eq : ddotmod],[eq : adotmod ] ) using the method of lines .
the matrices @xmath48 and @xmath308 are set to zero at interior points and are chosen as described above at boundary points .
the spatial discretization uses difference operators satisfying summation by parts ( sbp ) ( see , for example , @xcite ) . in this paper
we use two of these difference operators , which we call d2 - 1 and d8 - 4 .
these operators satisfy sbp with respect to diagonal norms ; it can be seen @xcite that the use of these kind of norms implies that the order of the operator at and near boundary points is half that one in the interior .
d2 - 1 is a simple operator which is second order accurate in the interior and first order accurate at boundaries .
if the gridpoints range from @xmath345 to @xmath10 and the gridspacing is denoted by @xmath346 , this operator is given by @xmath347 we typically add kreiss - oliger like dissipation @xcite modified at and near boundaries such that the resulting operator is negative definite with respect to the norm for which sbp holds ( much in the way kreiss - oliger dissipation is negative definite in the absence of boundaries ) .
this operator , which we call @xmath348 , is added to the right - hand side of the evolution equations with a free ( nonnegative ) multiplicative parameter @xmath349 ( see @xcite for more details ) : @xmath350 d8 - 4 is one of strand s three - parametric difference operators of order eight in the interior and four at and close to boundaries , chosen in ref .
@xcite to minimize its spectral radius and therefore its associated courant limit .
we have not been able to extend the associated kreiss - oliger dissipation for this operator in a straightforward way such as to include the presence of boundaries ( see @xcite for more details ) .
therefore , here , we simply set the dissipative operator to zero near boundaries .
however , as we will discuss later , the resulting dissipation operator does not seem to yield a stable numerical scheme in the presence of cpbc ( as opposed to maximally dissipative boundary conditions ) and a better dissipation operator is needed . for the time discretization
we use a third order runge - kutta integrator . the boundary conditions ( [ eq : gauge ] ) and
( for the case without weyl control ) ( [ eq : maxdissmodes ] ) are imposed in the following way : we first compute the vector field corresponding to the discretized right - hand side of the system ( [ eq : ndot],[eq : gdot],[eq : kdotmod],[eq : ddotmod],[eq : adotmod ] ) and represent it in the characteristic basis .
then , the fields corresponding to @xmath351 are overwritten by the time derivative of @xmath153 and ( for the case without weyl control ) the field corresponding to @xmath307 are overwritten with the time derivative of @xmath154 . finally , the resulting vector field is transformed back to the original basis . in all the simulations presented below we use the same gridspacing in all directions , @xmath352 and a courant factor @xmath353 of either @xmath354 or @xmath355 .
the code we use is based on the one described in ref .
the following initial data is used in the @xmath0d evolutions of the next section . in each case , the spatial domain is the cube @xmath356 ^ 3 \}$ ] of side length @xmath178 . here
we consider initial data corresponding to flat space and add a random perturbation to it .
therefore , initially , the fields are chosen to be @xmath357 where the different @xmath358 quantities are random numbers which are uniformly distributed in @xmath359 $ ] .
similarly , at each timestep the boundary data @xmath153 , @xmath154 and @xmath268 ( see eqs . (
[ eq : gauge ] ) , ( [ eq : maxdissmodes ] ) and ( [ eq : maxdissweyl ] ) , respectively ) for the `` gauge '' and `` physical '' degrees of freedom is set to a random number of the same order , @xmath360 . in the simulations shown below
we choose @xmath361 .
this _ robust stability test _
@xcite is designed to test the numerical stability of the scheme by finding out whether there is , at fixed time , unbounded growth as resolution is increased .
one of the advantages of the test is that it excites all frequency modes allowed by a given resolution and it is therefore useful in spotting instabilities ( if present ) that could otherwise remain hidden in some convergence tests .
however , this test is _ not _ a convergence test , since the constraints are not satisfied and since different random data is used when resolution is changed .
axisymmetric brill waves @xcite are evolved in the next section .
the corresponding initial data for the three - metric is given in cartesian coordinates @xmath362 in the form @xmath363 , \label{eqn : brillmetric}\ ] ] where @xmath364 , the function @xmath308 has the form @xmath365 and the conformal factor @xmath366 is obtained by solving the hamiltonian constraint for time - symmetric initial data . in order to do so , we use the numerical elliptic solver bam and the idbrill thorn , both publicly available from the cactus distribution @xcite .
the initial data for @xmath12 , @xmath10 , @xmath14 and @xmath13 is given by @xmath367 where @xmath368 is the d2 - 1 finite difference operator in the @xmath369 direction .
for these brill wave simulations we choose the d2 - 1 operator not only for computing the initial data for @xmath13 but also for the discretization of the right - hand side of the evolution equations .
there is no advantage in using a higher order accurate difference operator because the elliptic solver is only second order accurate . in the evolutions shown below , @xmath370 . for weak enough waves the solution bounces at the origin and disperses to infinity , while for strong waves an apparent horizon
is typically found @xcite . in the evolutions below we use very weak waves , corresponding to an amplitude @xmath371 ( while the critical solution is believed to correspond to @xmath372 @xcite ) .
as we shall see , even in these very weak field evolutions there is a clear violations of the constraints when non constraint - preserving boundary conditions are used .
even though these waves are axisymmetric , we have evolved them in full @xmath0d .
initial data for a static solution corresponding to a gauge transformation of flat spacetime can be obtained by setting @xmath12 , @xmath10 and @xmath14 to the same expressions as above [ eqs.([ki],[ni],[ai ] ) ] , while the three - metric is obtained by starting from the flat metric in spherical coordinates , @xmath373 performing a coordinates transformation @xmath374 and transforming the resulting metric to the cartesian coordinates @xmath375 .
initial data for @xmath13 is obtained by computing analytically the gradient of the resulting cartesian components of the three metric , @xmath376 .
this gauge solution has been used to compare the stability properties of different formulations of einstein s vacuum equations @xcite . in our simulations , we choose @xmath377 and @xmath378 . finally , we consider an example where we start with flat initial data , i.e. @xmath379 , @xmath380 , @xmath381 , @xmath382 , @xmath383 , but inject a pulse of gravitational radiation by choosing as boundary data @xmath384}\end{aligned}\ ] ] where @xmath385 is a `` polar coordinate at each face '' .
for example , at the @xmath386 faces @xmath387 and similarly for the other faces . in the simulations shown below we choose @xmath388 , @xmath389 and amplitudes @xmath390 and @xmath178 . in the simulations below , we choose @xmath391 corresponding to time harmonic slicing , set the shift to zero everywhere and at all times and restrict ourselves to the case in which the parameter @xmath23 in eq .
( [ eq : kdot ] ) is set to @xmath292 .
furthermore , we only consider choices of parameters for which @xmath392 and @xmath393 , and set the parameter @xmath394 defined in eq .
( [ eq : omega ] ) to zero .
this implies that the characteristic directions lie along the light cone or the normal to the @xmath170 hypersurfaces .
there are two subsets of parameter space which fulfill these requirements : 1 .
mono - parametric family ( parametrized by the nonzero value of @xmath25 ) : @xmath395 2 . bi - parametric family ( parametrized by the parameters @xmath24 and @xmath396 ) : @xmath397 one can show that for these families the evolution system ( [ eq : ndot],[eq : gdot],[eq : kdot],[eq : ddot],[eq : adot ] ) is symmetric hyperbolic .
however , notice that the mono - parametric family violates the condition ( [ eq : constrsym1 ] ) which means that the constraint propagation system is not symmetric hyperbolic .
the bi - parametric family satisfies the conditions ( [ eq : constrsym1],[eq : constrsym2],[eq : constrsym3 ] ) if and only if @xmath398 . according to the analysis in the previous section the determinant condition is satisfied for this subfamily . in this article
, we consider the following four cases : * _ cpbc without weyl control : a completely ill posed case _ + the bi - parametric family is used in this case , with parameters @xmath399 , and cpbc without weyl control .
as mentioned above , the resulting parameters _ violate _ the condition for the constraint propagation system to be symmetric hyperbolic .
therefore , there is no guarantee that the ibvp is well posed . as a matter of fact
, it turns out that the determinant condition is violated in this case and thus the problem admits ill posed modes .
all the simulations performed below confirm this fact . * _ cbpc with weyl control _
+ here we choose the bi - parametric family and cpbc with weyl control .
the parameters @xmath24 and @xmath22 are chosen to be either @xmath400 or @xmath401 ; both choices satisfy the determinant condition .
although the growth rate is smaller for the second choice , in both cases we find that the system is numerically unstable .
however , we also notice that the instability seems to be milder than in the completely ill posed case : in the weak brill wave runs , for instance , one needs to run for several crossing times or use very high resolution before noticing the lack of convergence .
a similar situation arises in the numerical evolution of weakly hyperbolic systems @xcite with periodic boundary conditions . *
_ cpbc without weyl control _ + again , we choose the bi - parametric family , but now we consider cpbc without weyl control , and choose @xmath402 .
the numerical simulations presented below suggest that the system is numerically stable . * _ maximally dissipative _ + the mono - parametric family with @xmath403 is used with maximally dissipative boundary conditions .
this system should be stable , since the evolution equations are symmetric hyperbolic and so the ibvp is well posed .
the simulations below confirm this . but
more importantly , they provide an explicit demonstration that evolutions of a system with constraint - preserving boundary conditions , when numerically stable , are more accurate than standard , maximally dissipative ones , as in the latter case the boundary conditions introduce constraint violations that do not converge to zero .
for each of the four ibvp described above we run the robust stability test and evolve weak brill waves . for the case of cpbc without weyl control ,
we also evolve the gauge solution and inject pulses of gravitational radiation from the boundary of an initially flat spacetime . for these runs ,
the courant factor is @xmath355 and the resolution varies from @xmath404 , @xmath405 to @xmath406 gridpoints .
the spatial derivatives are discretized using the d2 - 1 operator , and a dissipation parameter of @xmath407 is chosen .
recall that random initial _ and _ boundary data is given here . in figure
[ random ] we show the energy of the constraints versus light crossing time .
the energy of the constraints is defined to be the sum over all gridpoints of the sum of the square of the components of the constraint variables divided by the number of gridpoints .
as expected , in the completely ill posed case , the energy of the main variables grows at fixed time as resolution is increased , which makes the code crash in the timescale of less than a crossing time ( there is also an increase in the energy at fixed resolution as a function of time , though this does not necessarily represent a numerical instability ) .
figure [ random ] shows that there is also a similar kind of growth in the energy for the constraints .
the timescale and explosive kind of the numerical instability are similar to those found in initial value problems that are completely ill posed due to the the presence of complex eigenvalues in the principal part @xcite .
this similarity is , indeed , expected , as at the analytical level the problem admits exponentially in time growing modes , where the exponential factor increases with the frequency as shown in section [ sect : fl ] . _
recall , however , that the main evolution system is symmetric hyperbolic ; thus the boundary conditions are responsible for the instabilities . _ and dissipation @xmath408 .
the energy of the constraints is defined to be the sum over all gridpoints of the sum of the square of the components of the constraint variables divided by the number of gridpoints.,title="fig:",height=226 ] and dissipation @xmath408 .
the energy of the constraints is defined to be the sum over all gridpoints of the sum of the square of the components of the constraint variables divided by the number of gridpoints.,title="fig:",height=226 ] and dissipation @xmath408 .
the energy of the constraints is defined to be the sum over all gridpoints of the sum of the square of the components of the constraint variables divided by the number of gridpoints.,title="fig:",height=226 ] and dissipation @xmath408 .
the energy of the constraints is defined to be the sum over all gridpoints of the sum of the square of the components of the constraint variables divided by the number of gridpoints.,title="fig:",height=226 ] and dissipation @xmath408 .
the energy of the constraints is defined to be the sum over all gridpoints of the sum of the square of the components of the constraint variables divided by the number of gridpoints.,title="fig:",height=226 ] in the cpbc case with weyl control the runs also show evidence of a numerical instability both in the main and constraint variables , though the growth rate is somehow smaller than before .
what is somehow unexpected is the resolution dependent growth in the constraint variables .
after all , at the continuum , the constraints evolution is governed by a symmetric hyperbolic system with maximally dissipative boundary conditions which constitutes a well posed problem by itself .
however , it seems that the main system is unstable and that at the discrete level , the instabilities do have an effect on the constraint variables . in order to gain a better insight into this problem
we have analyzed the semi - discrete problem , where only the space derivatives are discretized . even in the simpler case of linearizations about minkowski spacetime we found that while the discrete constraints do obey a symmetric hyperbolic system it is far from obvious that the way in which we implement the constraint - preserving boundary conditions allow for a semi - discrete energy estimate .
it also seems difficult to represent the boundary conditions in a different way such that a semi - discrete energy estimate can be shown for the constraint propagation system .
we will not attempt to address this question further in this article . notice that even if we found such a discretization , the system could still suffer from the presence of more weakly ill posed gauge or physical modes that are undetected by the determinant condition .
an explicit example of such weakly ill posed modes is given in [ app : sfs ] . as opposed to the previous cases ,
in the cpbc case without weyl control the runs strongly suggest that the resulting systems is numerically stable and that the continuum problem might be well posed . at early times
the constraints energy decrease while at around @xmath5 crossing times the constraint variables start to slowly grow in time .
however , the growth rate does not become larger with increasing resolution , as opposed to the previous cases .
finally , the random runs with maximally dissipative boundary conditions also strongly suggest that the system is numerically stable , though this is expected because in this case we know that the ibvp is well posed . in these simulations we use the smaller courant factor of @xmath409 .
figure [ weak_brill ] shows weak brill wave runs for the four formulations of the ibvp described in section [ formulations ] .
the expectations based on the random data runs of the previous section are confirmed .
namely , the completely ill posed cpbc case is manifestly numerically unstable .
the cpbc with weyl control case that satisfies the determinant condition seems to be unstable as well , though in a `` weaker '' sense while the two remaining cases ( the maximally dissipative one and the cpbc without weyl control case that satisfies the determinant condition ) seem to be numerically stable . .
the expectations based on that figure regarding numerical stability ( or its lack ) are here confirmed.,title="fig:",height=226 ] .
the expectations based on that figure regarding numerical stability ( or its lack ) are here confirmed.,title="fig:",height=226 ] .
the expectations based on that figure regarding numerical stability ( or its lack ) are here confirmed.,title="fig:",height=226 ] .
the expectations based on that figure regarding numerical stability ( or its lack ) are here confirmed.,title="fig:",height=226 ] figure [ weak_brill_short ] shows the last two cases on a shorter timescale and with a higher resolution run added .
_ notice how in the cpbc case the constraint variables seem to converge to zero , while in the maximally dissipative case with the same initial data and resolution the lack of convergence to zero is evident_. here , we concentrate on the cpbc without weyl control case . since the initial data is given in analytic form we use the high order accurate finite differencing operator d8 - 4 . the results are shown in figure [ gauge ] .
notice that the high order accurate scheme results in a much faster convergence of the constraints to zero as resolution is increased . here
, the courant factor is @xmath354 and the dissipation parameter @xmath410 . , and dissipation @xmath410 , height=226 ] as in the previous gauge simulations , we use the cpbc formulation without weyl control .
however , we have found that in this case for the same resolutions used in the previous simulations a numerical instability shows up after some time .
this is probably due to the fact that the dissipative operator is zero near the boundary and is not negative semi - definite with respect to the scalar product for which sbp holds .
therefore for this case we have used the d2 - 1 operator , with courant factor @xmath411 , and dissipation parameter @xmath407 . in figure [ col_energy ]
we show the energy of the constraints as a function of time for two resolutions and the two amplitudes @xmath412 and @xmath413 .
the energy starts from zero , as the initial data consists of minkowski spacetime .
after a short time the energy quickly grows as the pulses are injected into the domain through the six boundaries , until the pulses have been completely injected at roughly one crossing time , time at which the energy stays approximately constant ( at fixed resolution ) .
figure [ col_curvature ] shows the maximum ( in the computational domain ) of the curvature invariant @xmath414 .
this curvature invariant starts at zero as well and remains small ( compared to one ) for amplitude @xmath412 while for amplitude @xmath413 it increases to very large values as the pulses are injected .
for example , at @xmath415 we have @xmath416 .
to have an idea of how strong the curvature is , recall that for the schwarzschild solution @xmath417 , where @xmath418 is the mass of the black hole and @xmath419 is the area radial coordinate .
therefore , a curvature of @xmath420 would correspond to being inside a black hole of mass @xmath421 at @xmath422 .
this curvature is being caused _
solely _ by the injection of pulses through the boundaries , showing that the latter are able to handle very non - linear dynamics .
and the dissipation parameter is @xmath410 .
the energy starts from roundoff as the initial data corresponds to flat spacetime .
the energy curves for the highest resolutions have been scaled by a factor of four in a way that overlapping with the curves for to the coarser resolution corresponds to second order convergence .
, height=226 ] .
the values achieved by this invariant in the simulations with amplitude @xmath413 are comparable to values found inside a schwarzschild black hole .
this curvature is produced _ solely _ by the injection of pulses through the boundaries .
, height=226 ]
we derived a family of outer boundary conditions for first order hyperbolic formulations of einstein s field equations with live gauges .
these boundary conditions have the property of being constraint - preserving in the sense that they guarantee that any smooth solution to the evolution equations subject to these boundary conditions automatically satisfy the constraints if so initially .
furthermore , we have discussed different possibilities for using constraint - preserving boundary conditions in order to control the gauge and physical degrees of freedom at the boundary .
one of these possibilities ( which we called cpbc with weyl control ) is attractive from a physical point of view since it provides boundary data to the weyl scalars @xmath1 and @xmath2 , which at null infinity represent the in- and outgoing radiation .
furthermore , these scalars are gauge - invariant for linearizations about the kerr metric , an approximation that should be good in many simulations provided the boundaries can be moved sufficiently far from the strong field region .
we also discussed simpler ways ( which we called cpbc without weyl control ) of controlling the physical degrees of freedom ; however , their physical interpretation is less clear .
next , we analyzed the stability of the resulting ibvp by analytical and numerical means . by considering high - frequency perturbations we obtained the corresponding frozen coefficient problem which is linear and
can be analyzed using fourier - laplace transformations . in this way
, one obtains a determinant condition which is a necessary condition for the well posedness of the problem .
the satisfaction of this condition restricts the freedom in the choice of parameters in the formulation and the boundary conditions .
in particular , we found that the violation of the determinant condition leads to the presence of ill posed constraint - violating or gauge modes .
these modes are ill posed in the sense that they grow exponentially in time with an exponential factor that is unbounded .
one could say that the boundary conditions are responsible for the presence of these modes since , in the absence of boundaries , the initial value problem is well posed because the evolution equations are strongly hyperbolic .
a further example of boundary conditions leading to an instability of an otherwise stable initial - value problem is given in [ app : sfs ] .
next , we performed three - dimensional numerical simulations . in order to do so
, we extended an earlier finite - differencing code and implemented the constraint - preserving boundary conditions .
we first performed a robust stability test which consists in specifying random initial and boundary data for different resolutions and checking that the time evolution of the fields does not exhibit unbounded resolution - dependent growth .
we found that the set of cpbc without weyl control considered in this article is robustly stable in this sense .
we also compared the results from these boundary conditions with results obtained by simply freezing the ingoing fields of the main evolution system to their initial values . while both systems are robustly stable , we found very strong evidence for the freezing boundary condition to yield _ constraint - violating _ solutions :
in contrast to the cpbc , the constraint variables seem to converge to a nonzero value for freezing boundary conditions .
we further tested the cpbc without weyl control by evolving a gauge solution and weak brill waves , and found that in both cases the constraint variables seem to converge to zero .
finally , as an application , we considered a situation in which one starts with flat initial data and injects `` pulses of gravitational radiation '' through the boundaries , of enough amplitude to create very large curvature in the interior .
we expect the cpbc without weyl control to be useful for many applications , like improving the accuracy and stability of current binary black hole and binary neutron star simulations or for a successful implementation of characteristic or perturbative matching techniques @xcite unfortunately , our numerical results for the cpbc with weyl control do not pass the robust stability test , although the instability in this case is much weaker than for the case of cpbc that violate the determinant condition .
there are several possible causes for this instability .
first , it is not clear if this instability is caused by a `` bad '' discretization of the problem the continuum ibvp being well posed or if the continuum problem is actually ill posed and thus the cause for the instability observed .
another possible source of problem is the cubical domain used in our simulations , which has edges and corners
. it might be that we violate some compatibility conditions at such points and that this causes an instability .
whatever the cause of the instability might be , it would be nice to have a better analytic insight into the problem .
although the determinant condition used here is necessary for the well posedness of the ibvp it is not sufficient , not even for the frozen coefficient problem .
when making such a statement we have to specify in what sense we expect the ibvp to be well posed . in the linear regime , we expect the problem to be well posed with respect to a hilbert space that controls the @xmath4 norm of the fields and the @xmath4 norm of the constraint variables . in this case , there are explicit examples @xcite that show that the satisfaction of the determinant condition is _ not _ sufficient for well posedness .
a different examples which applies to the ibvp considered in this article is given in [ app : sfs ] .
clearly , it would be much more satisfactory to derive cpbc with and without weyl control for which well posedness can be guaranteed .
this research was supported in part by the nsf under grants no : phy0244335 , phy0326311 , int0204937 to louisiana state university , the horace hearne jr .
institute for theoretical physics , nsf grant no .
phy-0099568 to caltech , and nsf grants no .
phy0354631 and phy0312072 to cornell university .
this research used the resources of the center for computation and technology at louisiana state university , which is supported by funding from the louisiana legislature s information technology initiative .
we are especially indebted to mark miller for many discussions , suggestions and help throughout this project .
we also thank l. buchman , c. gundlach , l. kidder , l. lehner , l. lindblom , g. nagy c. palenzuela , o. reula , m. scheel , s. teukolsky and j. york for many helpful discussions and comments .
in this appendix we re - derive the main evolution system , carefully keeping track of the constraints that are being used , and find the constraint propagation system with the help of the ( twice contracted ) bianchi identities .
we start with the @xmath423 split of the four - dimensional ricci tensor , @xmath424 where @xmath425 and @xmath426 are the ricci tensor and the covariant derivative , respectively , belonging to the three metric @xmath11 , @xmath10 is the lapse , @xmath12 the extrinsic curvature , and @xmath427 denotes the operator @xmath428 . in order to obtain first order equations we rewrite the spatial derivatives of the three - metric as @xmath429 where @xmath13 are new fields and @xmath430 are constraints .
correspondingly , we can split the christoffel symbols , @xmath431 , belonging to the three metric as @xmath432 where @xmath433 are christoffel symbols belonging to a new connection @xmath434 which is torsion - free but not necessarily metric ( we have @xmath435 , so @xmath434 is only metric if the constraints @xmath430 are satisfied ) , and where @xmath436 since the symbols @xmath437 correspond to the difference between the connections @xmath438 and @xmath434 , they are actually the components of tensor field . substituting eq .
( [ eq : splitchris ] ) into the expression for the ricci tensor , we obtain @xmath439 where @xmath35 is given by eq .
( [ eq : rij ] ) and where @xmath440 here , we have defined @xmath441ij}$ ] .
note that the hatted quantities vanish if @xmath430 .
next , we replace the gradient of the logarithm of the lapse by a new field @xmath14 minus a corresponding constraint variable , @xmath442 and rewrite @xmath443 using eqs .
( [ eq : subst1 ] ) and ( [ eq : subst2 ] ) , we rewrite eq .
( [ eq : fourricci ] ) as @xmath444 where @xmath445 groups together the four ricci tensor ( which vanishes in vacuum ) and the constraint variables .
an evolution equation for @xmath12 in vacuum can be obtained from the identity ( [ eq : kijid ] ) by setting @xmath446 to zero . however , in order to obtain a strongly hyperbolic evolution system , one needs to set @xmath446 equal to suitable combinations of the constraint variables ( see below ) .
next , in order to obtain evolution equations for the new fields @xmath13 and
@xmath14 , we apply the operator @xmath427 on eqs .
( [ eq : splitchris],[eq : splitgradnabn ] ) and use the commutation relation @xmath447 t_{i_1 i_2 ... i_r } & = & \frac{\partial_k n}{n } \partial_0 t_{i_1 i_2 ... i_r } \nonumber\\ & + & \frac{1}{n}\left ( t_{s i_2 ... i_r } \partial_k\partial_{i_1 } \beta^s + ... + t_{i_1 i_2 ...
i_{r-1 } s } \partial_k\partial_{i_r } \beta^s
\right ) , \label{eq : comrel}\end{aligned}\ ] ] for any @xmath419-rank symbol @xmath448 , where the lie derivative @xmath449 of @xmath448 is formally defined by @xmath450 using the evolution equations @xmath451 and @xmath452 for the three metric and the lapse , respectively , we obtain @xmath453 where @xmath454 finally , we rewrite the hamiltonian and momentum constraint .
let @xmath455 be the einstein tensor , and let @xmath456 denote the contraction with the vector field @xmath427 .
then , we have @xmath457 where the expressions for @xmath83 and @xmath458 are given by eqs .
( [ eq : c],[eq : ci ] ) and where @xmath459 the main evolution equations are @xmath451 , @xmath452 , and eqs .
( [ eq : kijid],[eq : dkijid],[eq : aiid ] ) where one sets the quantities @xmath460 to zero . with this information
it is not very difficult to find the constraint propagation system using the commutation relation ( [ eq : comrel ] ) and the twice contracted bianchi identities ( written in 3 + 1 form ) @xmath461 substituting @xmath462 , @xmath463 and using the equations @xmath464 , @xmath465 and @xmath466 and eqs .
( [ eq : lambdaij],[eq : lambdakij],[eq : lambdai ] ) , a lengthy but straightforward calculation yields @xmath467l } + \frac{1}{2 } c_{ijkl } - \zeta c_{k(ij)l } \right ) \nonumber\\ & - & g^{is}\partial_s c_{ij}^{(a ) } + l.o.,\\ \partial_0 c_{kij } & = & l.o.,\\ \partial_0 c_{lkij } & = & \frac{\eta}{2}\left ( g_{i[k } \partial_{l ] } c_j + g_{j[k } \partial_{l ] } c_i \right ) + \chi\ , g_{ij } \partial_{[l } c_{k ] } + l.o.,\\ \partial_0 c_k^{(a ) } & = & l.o.,\\ \partial_0 c_{ij}^{(a ) } & = & \xi\,\partial_{[l } c_{k ] } + l.o.,\end{aligned}\ ] ] where we have defined @xmath468}^{(a)})$ ] and where the lower order terms , @xmath469 , are linear algebraic expressions in the constraint variables , with coefficients that depend on the main variables and their spatial derivatives .
the hard part of the calculation is to show that no spatial derivatives of @xmath470 and @xmath471 other than the antisymmetric ones ( which can be re - expressed in terms of the constraint variables @xmath96 and @xmath472 , respectively ) enter the lower order terms .
given a first order evolution system with principal symbol @xmath69 , the characteristic fields with respect to a fixed direction @xmath52 are defined to be the projections of the main fields onto the eigenspaces of @xmath69 . in order to find the characteristic fields for the symbol defined by eq .
( [ eq : principalsymb ] ) , it is convenient to choose an orthonormal basis of three - vectors @xmath162 , @xmath163 , @xmath164 , such that @xmath473 and express the main variables @xmath12 , @xmath13 and @xmath14 with respect to this basis : @xmath474 here and in the following , we assume that @xmath52 is normalized such that @xmath475 . assuming that @xmath220 , @xmath221 and @xmath222 are positive ( which is a necessary condition for strong hyperbolicity ) and defining @xmath476 we can express the characteristic fields as ( @xmath62 refers to the the index @xmath477 and @xmath478 to the indices @xmath479 ) @xmath480 where @xmath481 + \gamma(b_n - d_n ) , \\ d_{na } & = & -\frac{1}{4}(1-\zeta)(d_{nna}-b_a ) + \frac{1}{4}(1+\zeta ) d_{ann } - \frac{1}{4}\ , d_a - \frac{1}{2}\ , a_a\ ; , \\ \hat{d}_{ab } & = & \frac{1}{2}\left [ -d_{nab } + ( 1+\zeta ) d_{(ab)n } \right ] - \frac{1}{4}\,\delta_{ab}\delta^{cd}\left [ -d_{ncd } + ( 1+\zeta ) d_{cdn } \right].\end{aligned}\ ] ] here , @xmath482 , @xmath483 and @xmath484 .
@xmath322 , @xmath485 , @xmath485 , @xmath486 have the speeds @xmath487 , @xmath488 , @xmath489 , and @xmath490 , respectively .
( these are the speeds with respect to the normal derivative operator .
the speeds with respect to the time evolution vector field are obtained from these after the transformation @xmath491 . )
the remaining characteristic fields have speeds @xmath492 . for the following representation ,
we assume that the conditions ( [ eq : constrsym1],[eq : constrsym2 ] ) , which are necessary for the constraint propagation system to be symmetrizable , hold . defining @xmath493 , the zero speed characteristic fields are @xmath494 - ( 2\sigma + \xi)\omega ( b_n - d_n ) , \\ v_a^{(0 ) } & = & a_a - \xi\omega(b_a - d_a).\end{aligned}\ ] ] having obtained the characteristic fields explicitly , it is not difficult to verify the additional smoothness requirements for the system to be strongly hyperbolic : namely , we have to construct a symmetric positive definite matrix @xmath495 which symmetrizes the principal symbol @xmath496 defined in eq .
( [ eq : principalsymb ] ) .
the system is called strongly hyperbolic if @xmath115 depends smoothly on @xmath48 , @xmath62 , @xmath70 and is such that the matrix @xmath497 is symmetric for all @xmath48 , @xmath62 , @xmath70 .
the smoothness requirements is needed for the pseudo - differential calculus while the symmetry condition allows for appropriate energy estimates .
the matrix @xmath115 can be obtained from the quadratic form which is built by summing over the square of the eigenfields : @xmath498 clearly , this quadratic form depends smoothly on @xmath52 since it can be written in terms of contractions with @xmath499 and @xmath63 .
@xmath115 is also smooth in the other variables provided that the parameters are smooth functions , and provided that the function @xmath394 can be chosen smoothly .
the latter condition is nontrivial if there are crossing points in the values swept by @xmath220 and @xmath221 . in all our simulations ,
we choose smooth parameters such that @xmath500 and set @xmath501 , so the system is strongly hyperbolic . finally , we give the inverse transformation which allows to recover the main variables from the characteristic ones .
first , compute @xmath502 from which @xmath503 is obtained .
next , compute @xmath504 next , set @xmath505 where @xmath506 notice that as a consequence of the conditions ( [ eq : constrsym1],[eq : constrsym2 ] ) , @xmath507 < 0 $ ] and the matrix @xmath508 is well defined . using this ,
compute @xmath509 + ( 2\sigma + \xi)\omega ( b_n - d_n ) , \\
a_a & = & v_a^{(0 ) } + \xi\omega(b_a - d_a ) , \\
d_{nab } & = & -2\hat{d}_{ab } + ( 1+\zeta ) d_{(ab)n } + \delta_{ab}\left [ \gamma(b_n - d_n ) - d^c_c \right ] , \\
d_{nna } & = & \zeta^{-1}\left [ 2 d_{na } - \frac{1}{2}(1+\zeta)\delta^{bc}(d_{bca } - d_{abc } ) + a_a \right ] + b_a - \frac{1}{2}\ , d_a\ ; , \\
d_{nnn } & = & 2\zeta^{-1}\left [ d_{nn } -\frac{1}{2}(1-\zeta ) b_n + \frac{1}{2}\ , d_n + a_n - \frac{\gamma}{2}\ , ( b_n - d_n ) \right].\end{aligned}\ ] ] from which one can compute the components of @xmath13 and @xmath14 with respect to the orthonormal basis .
finally , one obtains the coordinate components of the main variables by contracting with the dual basis @xmath510 , which is defined by @xmath511 , @xmath512 : @xmath513
in this appendix we show explicitly that the linearized ibvp with weyl control can not be well posed in @xmath4 if the parameter @xmath394 defined in eq .
( [ eq : omega ] ) is one ; independent on whether or not the determinant condition is satisfied . in order to see this , we consider the following family of solutions : let @xmath514 be a one - form that satisfies @xmath515 , @xmath516 , and @xmath517 , and let @xmath518 it is not difficult to check that these expressions satisfy the evolution equations ( [ eq : linkij],[eq : lindkij],[eq : linai ] ) and the constraints @xmath519 , @xmath520ij } = 0 $ ] , @xmath521 . for systems without boundaries ,
these solutions are trivial if appropriate fall off conditions on the fields are demanded since then the harmonic condition on @xmath514 implies that it must vanish
. however , if boundaries are present , @xmath514 may be nontrivial .
the electric and magnetic components of the linearized weyl tensor corresponding to the solutions ( [ eq : sol ] ) are @xmath522 and @xmath99 vanishes if the one - form @xmath514 is closed .
in particular , this is true if @xmath514 is exact , i.e. if @xmath523 for some time - independent harmonic function @xmath524 . in this case , we also have @xmath525 therefore , if @xmath266 , the family of solutions ( [ eq : sol ] ) with @xmath523 and @xmath524 harmonic and time - independent shows that the linearized ibvp with the boundary conditions ( [ eq : cpbc],[eq : gauge],[eq : maxdissweyl ] ) is not well posed in @xmath4 since then the boundary conditions are satisfied with homogeneous data and since the initial data depends only on second derivatives of @xmath524 whereas @xmath13 depends on third derivatives of @xmath524 for @xmath141 .
this results in frequency dependent growth of the solution of the form @xmath526 , where @xmath369 is a characteristic wave number of the initial data .
a. m. abrahams _ et al . _
[ binary black hole grand challenge alliance collaboration ] , gravitational wave extraction and outer boundary conditions by perturbative matching , phys .
* 80 * ( 1998 ) 18121815 .
s. frittelli and r. gomez , boundary conditions for hyperbolic formulations of the einstein equations , _ class .
grav . _ * 20 * ( 2003 ) 23792392 , einstein boundary conditions for the 3 + 1 einstein equations , _ phys .
d _ * 68 * ( 2003 ) 044014 , einstein boundary conditions in relation to constraint propagation for the initial - boundary value problem of the einstein equations , _ phys . rev .
d _ * 69 * ( 2004 ) 124020 , einstein boundary conditions for the einstein equations in the conformal - traceless decomposition , _ phys .
d _ * 70 * ( 2004 ) 064008 .
m. alcubierre , g. allen , b. brugmann , e. seidel and w. m. suen , towards an understanding of the stability properties of the 3 + 1 evolution equations in general relativity , phys .
d * 62 * ( 2000 ) 124011 .
g. calabrese , l. lehner , d. neilsen , j. pullin , o. reula , o. sarbach and m. tiglio , novel finite - differencing techniques for numerical relativity : application to black hole excision , class .
quant . grav .
* 20 * ( 2003 ) l245l252 . | outer boundary conditions for strongly and symmetric hyperbolic formulations of @xmath0d einstein s field equations with a live gauge condition are discussed .
the boundary conditions have the property that they ensure constraint propagation and control in a sense made precise in this article the physical degrees of freedom at the boundary .
we use fourier - laplace transformation techniques to find necessary conditions for the well posedness of the resulting initial - boundary value problem and integrate the resulting three - dimensional nonlinear equations using a finite - differencing code .
we obtain a set of constraint - preserving boundary conditions which pass a robust numerical stability test .
we explicitly compare these new boundary conditions to standard , maximally dissipative ones through brill wave evolutions .
our numerical results explicitly show that in the latter case the constraint variables , describing the violation of the constraints , do not converge to zero when resolution is increased while for the new boundary conditions , the constraint variables do decrease as resolution is increased . as an application
, we inject pulses of `` gravitational radiation '' through the boundaries of an initially flat spacetime domain , with enough amplitude to generate strong fields and induce large curvature scalars , showing that our boundary conditions are robust enough to handle nonlinear dynamics .
we expect our boundary conditions to be useful for improving the accuracy and stability of current binary black hole and binary neutron star simulations , for a successful implementation of characteristic or perturbative matching techniques , and other applications .
we also discuss limitations of our approach and possible future directions . |
giant molecular clouds ( gmcs ) are well known to be the birthplaces of stars . studies of various gmcs have not only revealed numerous protostellar objects of varying age , they have also detected localized maxima in dust column density that exhibit no signs of star formation and have been designated as starless cores ( e.g. , motte et al .
1998 ; andr et al .
1993 ; jrgensen et al .
2007 ; di francesco et al . 2007 ) .
some of these starless cores have been shown to be gravitationally bound and have been labeled as prestellar cores , as they are believed to be the progenitors of protostars ( e.g. , di francesco et al 2007 ; andr et al .
2007 and references therein ) .
further examination of prestellar cores can shed light on the earliest stages of star formation and , in particular , the processes that lead to the gravitational collapse of a core .
the ophiuchus complex is particularly suitable for studying star formation as it is relatively close at only 120 pc ( loinard et al . 2008 ) and is an active star forming region ( loren et al . 1990 ) .
previous studies have found multiple clumps inside the ophiuchus complex ( loren et al .
1990 , johnstone et al .
2004 ) . of particular interest is the oph a clump , the brightest of the clumps at mm wavelengths ( motte et al . 1998 ) , in which numerous starless cores and the prototypical class 0 object , vla 1623 ( andr et al . 1993 ) , are located .
class 0 objects are deeply embedded protostars that have yet to accrete the majority of their final masses ( barsony 1994 ) .
the oph a clump has previously been surveyed in millimeter wavelengths ( e.g. andr et al . 1993 ) and in numerous spectral lines including c@xmath5o ( j=1@xmath20 ) ( wilking & lada 1983 ) , @xmath3co ( j=1@xmath20 ) ( loren 1989 ) , cs ( j = 2@xmath21 ) ( liseau et al .
1995 ) , and dco@xmath1 ( j=2@xmath21 , 3@xmath22 ) ( loren et al . 1990 ) .
recently , the n@xmath0h@xmath1 j=1@xmath20 emission from the oph a clump was mapped with the bima array and the iram 30 m telescope ( di francesco et al .
2004 , hereafter dam04 ) .
dam04 discovered that the maxima of n@xmath0h@xmath1 emission are not coincident with the maxima of 1.3 mm continuum emission found by motte et al .
this result was inconsistent with results from other low - mass starless cores surveyed by tafalla et al .
( 2002 ) . within oph
a , vla 1623 was found to have very little n@xmath0h@xmath1 emission , sm2 , a previously identified starless core ( andr et al . 1993 ) , was found to have weak n@xmath0h@xmath1 emission and a new core not identified in motte et al.s ( 1998 ) millimeter continuum survey as a continuum maximum , designated oph a - n6 , was found as an emission peak in n@xmath0h@xmath1 .
dam04 suggested that the difference in n@xmath0h@xmath1 emission from these three cores was due to the cores being in different stages of development with n6 being the youngest and vla 1623 being the oldest .
no intensity peak coincident with the location of the n6 core was found in either the 850 @xmath6 m or the 450 @xmath6 m surveys of oph a conducted with the submillimeter common - user bolometric array ( scuba ) on the james clerk maxwell telescope ( wilson et al .
1999 ; johnstone et al .
these scuba data were later re - reduced and released as part of the scuba legacy catalogues ( di francesco et al . 2008 ) .
an interferometric submillimeter survey by bourke et al .
( 2009 , in preparation ) , however , detected a localized intensity maximum at the location of the n6 core .
no mid - infrared point source , down to 0.001 l@xmath7 , consistent with the presence of a protostar was found within the boundaries of the n6 core by the spitzer cores to disks ( c2d ) survey ( evans et al . 2003 , padgett et al .
2008 ) , suggesting that the n6 core most likely does not contain a protostar .
figure [ fig : dust ] shows the 850 @xmath6 m and 450 @xmath6 m emission from the oph a clump ( di francesco et al . 2008 ) , the n@xmath0h@xmath1 emission from the n6 core ( dam04 ) , the location of the submillimeter intensity maximum found by bourke et al .
( 2009 , in preparation ) and the locations of spitzer sources consistent with the presence of protostars in the oph a clump as identified by the c2d survey ( evans et al . 2003 ) .
dam04 found that the n@xmath0h@xmath1 line widths in n6 were only slightly larger than expected from thermal motions alone and that the n6 core is , to first order , in gravitational virial equilibrium .
thus , all evidence suggests that the n6 core is a prestellar core and that it is no longer strongly interacting with the turbulent flows of its surroundings .
it has been suggested that turbulent motions could be reduced in dense cores if the cores become smaller than the cut - off wavelength for magnetohydrodynamic ( mhd ) waves in the region ( e.g. , mckee et al .
1993 ; nakano 1998 ; myers 1998 ) .
since turbulent gas flows and magnetic support are likely important parts of a core s support against gravitational collapse ( caselli et al . 2002 ; myers 1999 ; mckee 1989 ) , the n6 core may be on the verge of gravitational collapse .
this makes the n6 core a valuable object for further study into the removal of support against gravitational collapse . in section [ obs ]
, we describe the details of our molecular line observations of the n6 core and the results of these observations are discussed in section [ results ] .
we also present derived dust temperatures and column densities of n@xmath0d@xmath1 in section [ results ] .
we then propose a model for the chemical and physical state of the n6 core based upon these observations in section [ discuss ] .
we present a summary of our conclusions in section [ conclusions ] .
the following observations were obtained at the james clerk maxwell telescope ( jcmt ) on mauna kea using the position - switching mode of the telescope .
data reduction was performed using a combination of the jcmt s starlink software and the class data reduction package . between 2005 february and 2005 july
, observations were made of the n@xmath0d@xmath1 j=3@xmath22 , dco@xmath1 j=3@xmath22 , dco@xmath1 j=5@xmath24 , hco@xmath1 j=3@xmath22 , and cs j=7@xmath26 transitions . between 2007 march and 2007 june , observations of the cs j=5@xmath24 , h@xmath3co@xmath1 j=3@xmath22 , and h@xmath3co@xmath1 j=4@xmath23 transitions were made .
table [ tab : obs ] gives the rest frequencies of each of the above transitions along with the front and back end instruments used , the spectral resolutions obtained , the full width at half maximum ( fwhm ) of the beams , the beam efficiencies , and the baseline rms achieved for these observations .
the critical densities of these transitions are also given .
all observations , except those of the h@xmath3co@xmath1 j=4@xmath23 transition , were made on a 5 x 5 grid with 10 spacing centered on the position of maximum n@xmath0h@xmath1 j=1@xmath20 integrated intensity found by dam04 , i.e. [ ra , decl . ] = [ @xmath8 , -@xmath9 ( j2000 ) , which we define as our standard grid .
unless otherwise stated , all positional offsets in this paper are relative to the central point of this standard grid and are given in arcseconds .
we will consider this position of maximum n@xmath0h@xmath1 integrated intensity to be the central location of the n6 core .
the h@xmath3co@xmath1 j=4@xmath23 emission was observed on a 16 x 16 grid with 7.5 spacing .
main beam efficiencies ( @xmath10 ) were obtained from observations of mars and jupiter , using the method detailed by mangum ( 1993 ) , which accounts for the coupling of a planetary disk with a gaussian beam .
observations of nh@xmath4 ( 1,1 ) and ( 2,2 ) inversion line emission , at a single pointing centered on the n@xmath0h@xmath1 maximum position , were obtained using the 100 m robert c. byrd green bank telescope ( gbt ) .
a total of ten minutes was spent on source over 2006 april 9 - 10 while the source was at @xmath11 elevation and @xmath12 k. the observations were done in frequency - switched mode using the gbt k - band ( upper ) receiver as the front end and the gbt spectrometer as the backend .
this setup allowed for the observation of both lines simultaneously in two 50mhz - wide ifs , each with 8192 spectral channels , giving a frequency resolution of 6.104khz , or 0.077kms@xmath13 at 23.694ghz .
the average telescope aperture efficiency @xmath14 and main beam efficiency @xmath10 were @xmath15 and @xmath16 respectively , and were determined through observations of the recommended calibration source 3c286 at the start of each shift .
the absolute flux accuracy is thus @xmath17% and the fwhm beam size for these observations was approximately 32 .
all data in this paper are presented in units of t@xmath18@xmath19 ( kutner & ulich 1981 ) .
however , to compare our observations to model predictions and to calculate physical conditions , we included the main beam coupling efficiency ( @xmath10 ) such that t@xmath20 = @xmath21 .
figure [ fig : allmol ] shows integrated intensity maps for the n@xmath0h@xmath1 1@xmath20 transition observed by dam04 and for the n@xmath0d@xmath1 3@xmath22 , cs 5@xmath24 , cs 7@xmath26 , dco@xmath1 3@xmath22 , dco@xmath1 5@xmath24 , h@xmath3co@xmath1 3@xmath22 , and h@xmath3co@xmath1 4@xmath23 transitions that we observed . for the cs , dco@xmath1 and h@xmath3co@xmath1 transitions , the integrated intensities were found by fitting gaussian curves to the observed spectra .
the observed n@xmath0d@xmath1 spectra are composites of at least 16 strong hyperfine components ( gerin et al .
2001 ) and these hyperfine components are badly blended ( as shown in figure [ fig : n2dpspec ] ) , so we have calculated the integrated intensity of the n@xmath0d@xmath1 3@xmath22 spectra by summing all emission from @xmath22 = 2.1 km s@xmath13 to @xmath22 = 5 km s@xmath13 after baseline subtraction .
we rejected any spectra that did not have a peak intensity at least 3x the 1@xmath23 ( rms ) noise of each transition ( see table [ tab : lineprop ] ) .
we estimated the average uncertainty in the integrated intensity of a transition by multiplying the average baseline rms by the average width of the innermost 9 spectra of that transition .
the n@xmath0d@xmath1 distribution , while showing a slight elongation in the northwest - southeast direction , appears to be less elongated than the n@xmath0h@xmath1 distribution .
while the peak integrated intensity for the n@xmath0d@xmath1 occurs at the [ -10,10 ] position , there is still a small plateau at the central [ 0,0 ] position where the n@xmath0h@xmath1 peaks in intensity .
figure [ fig : n2dpspec ] shows the n@xmath0d@xmath1 3@xmath22 spectrum at the [ -10,10 ] position .
the n@xmath0d@xmath1 integrated intensity peak occurs only 3 , or 15% of the beam size , from the location with the narrowest n@xmath0h@xmath1 line widths ( 0.19 km s@xmath13 ) as reported by dam04 .
little cs emission is seen at the positions of peak n@xmath0h@xmath1 and n@xmath0d@xmath1 integrated intensity . instead
, the cs integrated intensity maps show emission in a ring - like structure around these positions .
the dco@xmath1 integrated intensity maps also show little sign of the n6 core that is seen so clearly in the n@xmath0h@xmath1 and n@xmath0d@xmath1 maps . instead ,
the dco@xmath1 maps show decreasing brightness gradients from northwest to southeast and closely resemble the dust continuum profiles shown in figure [ fig : dust ] .
the only indication of the n6 core in the dco@xmath1 data is the presence of slight plateaus in the brightness gradient close to the [ 0,0 ] and [ -10,10 ] positions . like the dco@xmath1 ,
the h@xmath3co@xmath1 integrated intensity maps show little sign of the n6 core and are , instead , dominated by single brightness gradients .
the h@xmath3co@xmath1 brightness gradients , however , lie along west - to - east lines rather than the northwest - to - southeast direction seen in the dco@xmath1 data . the plateaus seen in the dco@xmath1 data
are also not present in the h@xmath3co@xmath1 data .
table [ tab : lineprop ] lists the average line widths and centroid velocities for the innermost nine pointings of the n@xmath0d@xmath1 3@xmath22 , cs 5@xmath24 , cs 7@xmath26 , dco@xmath1 3@xmath22 , dco@xmath1 5@xmath24 , h@xmath3co@xmath1 3@xmath22 , and h@xmath3co@xmath1 4@xmath23 transitions .
we have regridded the h@xmath3co@xmath1 4@xmath23 data onto our standard grid for this analysis and we have used class hyperfine splitting routine to determine the line widths and centroid velocities for the n@xmath0d@xmath1 data .
the n@xmath0d@xmath1 line widths tend to be the smallest while the h@xmath3co@xmath1 and cs line widths tend to be the largest .
the average n@xmath0d@xmath1 line width of 0.30 km s@xmath13 is similar to the average n@xmath0h@xmath1 1@xmath20 line width of 0.25 km s@xmath13 reported by dam04 and 0.37 km s@xmath13 reported by andr et al . ( 2007 ) , who used a 26 fwhm beam . while there appear to be significant differences in average centroid velocity between the different transitions
, we found that there were discrepancies between the rest frequencies given in the jcmt database , which we have used , the leiden atomic and molecular database ( lamda ) and the nasa jet propulsion laboratory ( jpl ) database for some of the above transitions .
these discrepancies were on the order of 0.1 to 0.2 km s@xmath13 .
consequently , we can not say whether these centroid velocity differences are significant .
the individual transitions do not show significant point - to - point variations in centroid velocity .
the spectra from cs , dco@xmath1 , and h@xmath3co@xmath1 show symmetric profiles with no self - absorption .
figure [ fig : hcop ] , however , shows that the hco@xmath1 3@xmath22 spectra have asymmetric profiles .
the hco@xmath1 spectra located near the central position of the n6 core show blue peaks around @xmath22 = 3.1 km s@xmath13 , absorption features around @xmath22 = 3.4 km s@xmath13 and smaller red peaks around @xmath22 = 3.8 km s@xmath13 .
figure [ fig : infall ] shows an example of how the optically thin dco@xmath1 3@xmath22 line appears at the same velocity as the absorption feature for the hco@xmath1 3@xmath22 spectra at the same location .
this correspondence is seen in all hco@xmath1 3@xmath22 spectra that have absorption features .
a few locations along the @xmath24 row , see figure [ fig : hcop ] , appear to show reverse asymmetry .
we have calculated the asymmetry parameter suggested by mardones et al .
( 1997 ) for the 17 spectra for which we were able to simultaneously fit gaussian profiles to both the red and blue peaks .
the asymmetry parameter , @xmath25 , is calculated as : @xmath26 where @xmath27 is the centroid velocity of the brightest peak in the optically thick spectrum ( hco@xmath1 3@xmath22 ) , @xmath28 is the centroid velocity of the optically thin spectrum ( dco@xmath1 3@xmath22 ) and @xmath29 is the line width of the optically thin spectrum .
the asymmetry parameters for these 17 spectra are listed on figure [ fig : hcop ] . for the 15 spectra that have a larger blue peak ,
every asymmetry parameter was calculated to be more negative than @xmath30 , which places these spectra into the `` significant blue shift '' category of mardones et al .
( 1997 ) , and the average asymmetry parameter for these 15 locations was calculated to be @xmath31 .
for the two spectra with larger red peaks , the [ -10,20 ] position had an asymmetry parameter that showed `` no significant shift '' while the asymmetry parameter at [ -20,20 ] indicated a `` significant red shift '' .
while both the [ 0,0 ] and [ -10,10 ] positions , the positions of peak n@xmath0h@xmath1 and n@xmath0d@xmath1 column densities respectively , show significant blue shifts , the asymmetry parameters for these two locations are not significantly different from the other 13 locations that have significant blue shifts .
mardones et al .
( 1997 ) also defined a blue excess value , e , as : @xmath32 where @xmath33 is the number of spectra with a significant blue shift , @xmath34 is the number of spectra with a significant red shift and n is the total number of spectra .
we find a blue excess value of 0.52 for the n6 core , which is greater than the significance threshold value of 0.25 of mardones et al .
interpretation of these results will be presented in section [ infall ] .
figure [ fig : nh3 ] shows the nh@xmath4 spectrum from the [ 0,0 ] position .
assuming gaussian profiles , the 18 hyperfine components of the nh@xmath4 ( 1,1 ) line shown in figure [ fig : nh3 ] were fit simultaneously using a chi - square reduction routine custom written in idl [ see friesen et al .
( 2009 ) ] .
the returned fit , overlaid on figure [ fig : nh3 ] , provide estimates of the line centroid velocity , @xmath22 , the line fwhm , @xmath35 , the opacity of the line summed over the 18 components , @xmath36 , and @xmath37 $ ] . as shown in figure [ fig : nh3 ] ,
the satellite components of the nh@xmath4 ( 2,2 ) line are visible above the rms noise of our data .
they are not detected , however , with sufficient signal - to - noise to fit the full hyperfine structure of the transition , and the main component seen in these data was consequently fit with a single gaussian .
assuming lte , we then determined the rotational temperature @xmath38 describing the relative populations of the two states , and calculated further the kinetic temperature @xmath39 of the gas following the updated conversion between @xmath38 and @xmath39 given by tafalla et al .
we find @xmath40k .
again , the fwhm resolution of these data is approximately 32 .
we have also used the 450 @xmath6 m and 850 @xmath6 m data from the scuba legacy catalogues . the effective beam for the 450 @xmath6
m data can be modeled as a combination of two gaussian beams with fwhms of 11 and 40 and relative intensities of 0.88 and 0.12 , respectively , while the effective beam for the 850 @xmath6 m data can be modeled as two gaussian beams with fwhms of 195 and 40 and relative intensities of 0.88 and 0.12 respectively ( di francesco et al .
these effective beams are different from the intrinsic beam patterns of the jcmt due to post - processing performed during the compilation of the scuba legacy catalogues .
we convolved the 850 @xmath6 m image with the two - component model beam of the 450 @xmath6 m data and we convolved the 450 @xmath6 m image with the two - component model beam of the 850 @xmath6 m data .
both data sets were then re - gridded onto our standard grid using a cubic spline interpolation routine .
johnstone et al .
( 2006 ) , for example , give the following relation for submillimeter dust continuum radiation : @xmath41 where @xmath42 is the observed flux , @xmath43 is the frequency , @xmath44 is the planck function evaluated at a temperature of @xmath45 , @xmath45 is the dust temperature and @xmath46 is the dust emissivity spectral index . we have used equation [ eqn : cont ] , the ratio of 850 @xmath6 m flux to 450 @xmath6 m flux at the center of the n6 core and the nh@xmath4 central temperature of 20.1 k to derive a value of 0.9 for @xmath46 at the center of the n6 core by assuming that the dust temperature is the same as the gas temperature .
models of the dense interiors of starless cores show that the dust and gas temperatures should be relatively similar ( e.g. hollenbach et al .
1991 ; galli et al .
a @xmath46 of 0.9 is comparable to values found in other star - forming regions ( e.g. beuther et al .
2004 ; hogerheijde & sandell 2000 ; goldsmith et al .
it is , however , lower than the commonly used value of 2.0 in the ism ( hildebrand 1983 ) and the value of 1.5 found for oph a by andr et al .
( 1993 ) . assuming that @xmath46 is constant over the n6 core , we were then able to derive dust temperatures for all points on our standard grid from the ratio of the submillimeter fluxes . figure [ fig : temp ] shows our derived dust temperature profile for the n6 core . with a @xmath46 of 0.9
, we find that the central dust temperature is 20.2 k and that the average dust temperature is 15.4 k. the dust temperature at the location of the n@xmath0d@xmath1 integrated intensity peak is 18.8 k. the scuba legacy catalogues also contain error maps for the submillimeter data .
we recalculated two average dust temperatures for the n6 core after adding the 850 @xmath6 m errors to the 850 @xmath6 m data and subtracting the 450 @xmath6 m errors from the 450 @xmath6 m data and vice versa . by averaging the absolute differences between these new average dust temperatures and our original average dust temperature
, we estimate that our continuum - derived dust temperatures are accurate to 4.5 k. since the peak dust temperature of the n6 core is 4.8 k higher than the average dust temperature of the core , we believe that the dust temperature peak at the central location of the n6 core is significant , albeit marginally . it is also unlikely that noise would produce a coherent dust temperature peak that spans multiple pixels , as seen in figure [ fig : temp ] . a multi - wavelength analysis with high quality fluxes from at least four bands ( and 350 microns in particular ) should improve significantly the dust temperature characterization ( see schnee et al .
we also consider our average dust temperature of 15.4 k to be in agreement with the average n@xmath0h@xmath1 excitation temperature derived by dam04 ( 17.4 k ) , supporting their suggestion that the n@xmath0h@xmath1 1@xmath20 line is thermalized . di francesco et al .
( 2008 ) suggested that the absolute fluxes in the 850 @xmath6 m and 450 @xmath6 m maps have uncertainties of 20% and 50% respectively .
these errors are not expected to fluctuate significantly over small spatial scales , however .
therefore , these errors would only affect our determination of @xmath46 and/or the absolute temperature scale , but would not affect the derived temperature structure of the n6 core , since the relative temperature structure is independent of the choice of @xmath46 .
the observed changes in the ratio of the submillimeter fluxes can be explained as due to either a temperature gradient , with a constant value of @xmath46 over the entire core , as assumed above , or due to a gradient in @xmath46 , with a constant temperature everywhere in the core .
if the entire n6 core were to be at a temperature of 20.1 k , @xmath46 would have to range from 0.9 at the center of the core to 0.23 at the periphery .
previous studies ( e.g. goldsmith et al . 1997 ) have found that @xmath46 decreases with increasing density and increasing grain size , the opposite of what would be needed in the n6 core for the core to be at a constant temperature . thus , we believe that the submillimeter continuum data are indicative of a temperature gradient in n6 rather than a gradient in @xmath46 . due to the blending of the n@xmath0d@xmath1 hyperfine components , we adopted the method given by caselli et al .
( 2002 ) to determine the column densities of n@xmath0d@xmath1 .
this method assumes that the n@xmath0d@xmath1 emission is optically thin and in lte .
since the n@xmath0h@xmath1 emission from the n6 core is only marginally optically thick ( dam04 ) , we believe the assumption that the n@xmath0d@xmath1 emission is optically thin is reasonable .
we used the dust temperatures calculated in section [ temp ] , to account for the decrease in temperature away from the n@xmath0h@xmath1 peak , and rejected any spectra that did not have a s / n greater than three .
the peak column density of n@xmath0d@xmath1 is @xmath47 @xmath48 and occurs at the integrated intensity peak ( at [ -10,10 ] ) .
the ratio of n@xmath0d@xmath1 to n@xmath0h@xmath1 column density changes from approximately 0.01 at the n@xmath0h@xmath1 peak to approximately 0.02 at the n@xmath0d@xmath1 peak . we have assumed that the dominant uncertainties in calculating the n@xmath0d@xmath1 column densities come from the uncertainty in the temperatures , @xmath49 k , and the uncertainty in the integrated intensities of the lines , 0.06 k km s@xmath13 .
we calculated new column densities using excitation temperatures 4.5 k greater than and less than our derived temperatures and integrated intensities 0.06 k km s@xmath13 greater than and less than our observed values . from this , we estimate the uncertainty in our column densities to be @xmath50 @xmath48 .
dam04 found that the line widths for n@xmath0h@xmath1 in the n6 core were only slightly larger than the theoretical , thermal line widths .
we have calculated the thermal line widths for the molecular transitions that we have observed using the gas temperature derived from the nh@xmath4 observations , 20.1 k , for all of the transitions observed with the jcmt and the following equation from myers et al .
( 1991 ) : @xmath51 where @xmath52 is the average thermal line width , @xmath53 is boltzmann s constant , @xmath54 is the gas temperature , and @xmath55 is the molecular mass of the species . since we are not concerned with pixel - to - pixel variations in line width for these calculations , we have chosen to use the 20.1k gas temperature derived from the nh@xmath4 data .
we also calculated the average non - thermal component of the line width for each transition by subtracting the thermal line width from the average observed line width in quadrature . for each transition , the average observed line width was calculated from the innermost 9 grid points . these results are summarized in table [ tab : linewidth ] .
at the high densities and low temperatures present at the center of starless cores , various molecules can freeze out onto dust grains and the abundances of these molecules , and their related ions , in the gas phase can become depleted ( e.g. , lger 1983 ; bergin & tafalla 2007 ) .
it has also been suggested that the varying properties of molecules , such as their binding energies to dust grains and their reactivity with co , will cause different molecules to deplete or form at different depths inside a core ( di francesco et al .
2007 and references therein ) .
this variable depletion creates shells of different chemical composition in the cloud ( tafalla et al .
2002 ; jrgensen et al .
2005 ) .
the localized emission peaks seen in the n@xmath0h@xmath1 and n@xmath0d@xmath1 integrated intensity maps , shown in figure [ fig : allmol ] , suggest that these molecules are present near the center of the core .
furthermore , the locations of the n@xmath0h@xmath1 and n@xmath0d@xmath1 emission peaks seem to suggest that the central part of the n6 core is a high - density ridge that stretches between [ 0,0 ] and [ -10,10 ] . this is supported by the position of a small submillimeter continuum peak interferometrically detected by bourke et al .
( 2009 , in preparation ) at [ 0,5 ] .
a comparison of figure [ fig : temp ] to figure [ fig : allmol ] shows that the strongest n@xmath0d@xmath1 emission occurs at a location with a colder dust temperature than the location of the peak n@xmath0h@xmath1 intensity .
this is expected since colder temperatures tend to increase the chemical abundance of deuterated species in cores ( vastel et al .
the colder temperature of the n@xmath0d@xmath1 maximum location would also explain why this was the location of dam04 s minimum observed , thermally dominated n@xmath0h@xmath1 line widths .
while the d / h ratio of 0.02 observed is smaller than those seen in the most heavily deuterated cores ( e.g. , 0.1 to 0.5 ; crapsi et al .
2005 ) , it is consistent with the increased deuterium abundance seen in other starless cores ( 0.01 to 0.44 ; see ceccarelli et al .
2007 and references therein ) and is still a high level of fractionation .
figure [ fig : allmol ] shows how the cs integrated intensity is smaller at the center of the n6 core than at its edges .
this drop in cs integrated intensity suggests that the cs column density must also be significantly reduced at the center . as cs has been observed to deplete readily in other cold , dense cores ( e.g. , bergin et al .
2001 ; lai et al .
2003 ; tafalla et al .
2006 ) , we suggest that cs is freezing out onto dust grains in the center of n6 .
thus , cs most likely only exists in the gas phase in an outer shell of material .
figure [ fig : allmol ] shows also that dco@xmath1 and h@xmath3co@xmath1 have a steady decreasing brightness gradient from the northwest edge to the center of our grid that is suggestive of a decline in the column densities of these molecules towards the center of n6 .
depletion of these molecules in cold , dense environments is also well known and has been observed in the l1498 and l1517b cores ( e.g. , tafalla et al .
the decrease in the integrated intensity of the cs lines at the center of the n6 core is more pronounced than in the dco@xmath1 or h@xmath3co@xmath1 lines and thus , we suggest that cs is more heavily depleted towards the center and occurs in a shell - like structure further from the center than the h@xmath3co@xmath1 and dco@xmath1 .
since co is known to destroy h@xmath0d@xmath1 , the depletion of co would be expected to lead to an increase in the rate of formation of deuterated species such as dco@xmath1 ( bergin & tafalla 2007 and references therein ) and this may explain the differences in the dco@xmath1 and h@xmath3co@xmath1 morphologies . since hco@xmath1 is chemically similar to dco@xmath1 and h@xmath3co@xmath1 we expect that the hco@xmath1 is depleting at a similar location to where the dco@xmath1 and h@xmath3co@xmath1 are depleting .
figure [ fig : structure ] shows a cut - away diagram of the chemical shell structure that we suggest exists in the n6 core . for molecules where
multiple transitions were observed , the combined average line width for these transitions is shown . given the chemical structure of the n6 core described in section [ structure ] , the turbulent component of the line widths , ( see section [ nontherm ] and table [ tab : linewidth ] ) , appears to increase with increasing distance from the center of the core .
the only exception to this trend is that there is no significant difference in line width between the h@xmath3co@xmath1 and cs .
the cs and h@xmath3co@xmath1 line widths seem to be consistent with the line widths of the general oph a medium ( dam04 ) , which would explain why no further increase in line width is seen between cs and h@xmath3co@xmath1 .
the slight increase in temperature towards the outer layers of the n6 core is insufficient to explain the significant increase in line width observed , as a temperature in excess of 200 k would be needed for the cs line widths to be entirely due to thermal broadening .
the difference in line width between the n@xmath0d@xmath1 and the n@xmath0h@xmath1 may be due to the difference in beam sizes used in observing these molecular transitions .
the larger 20 beam used to observe the n@xmath0d@xmath1 would have detected emission coming from more outlying , and thus more turbulent , gas than the approximately 10 x 6 beam used to observe the n@xmath0h@xmath1 would have .
this more turbulent gas would have broadened the n@xmath0d@xmath1 lines slightly .
this trend is supported by andr et al.s ( 2007 ) finding of a 0.37 km s@xmath13 n@xmath0h@xmath1 average line width in the n6 core with a 26 beam .
we also see a barely significant increase of 0.08 km s@xmath13 in average line width between the dco@xmath1 5@xmath24 transition , observed with a 14 beam , and the dco@xmath1 3@xmath22 transition , observed with a 20 beam .
we see no significant difference between the line widths of the two observed transitions of cs and h@xmath3co@xmath1 , suggesting , again , that the cs and h@xmath3co@xmath1 probe regions with the same level of turbulent motion as the general oph a medium .
these results suggest that turbulent motions have been suppressed in the center of the n6 core and that the central part of the n6 core is no longer strongly interacting with the general , turbulent oph a material .
a possible mechanism for this dissipation of turbulent motion has been proposed by various authors ( e.g. , mckee et al .
1993 ; nakano 1998 ; myers 1998 ) who suggest that turbulent motions in dense cores could be reduced if the cores become smaller than the cut - off wavelength for mhd waves in the region . myers ( 1998 ) suggested that these regions cut off from mhd waves could be on the order of 5000 au ( @xmath56 40 at the distance of the n6 core ) and dam04 described how the n6 core seems to fit well the description of a kernel as described by myers ( 1998 ) .
the suppression of turbulent motions and mhd waves in the n6 core also has a significant impact on the evolutionary state of the core , as it is believed that turbulence and magnetic fields are significant components of a core s support against gravitational collapse .
the lack of such support in the n6 core could make the core prone to gravitational collapse . as first mentioned in section [ morph
] , many of the hco@xmath1 3@xmath22 spectra are double peaked and asymmetric , with the gap between the peaks occurring at the same frequency as the optically thin dco@xmath1 3@xmath22 line .
if this dual peak structure were due to a second gas cloud in the line - of - sight , we would expect that the dco@xmath1 3@xmath22 spectra would appear at the same velocity as one of the two peaks in the hco@xmath1 spectra ( e.g. , gregersen et al . 1997 ; myers et al .
since the dco@xmath1 3@xmath22 spectral line appears between the two peaks in the hco@xmath1 spectra for all asymmetric hco@xmath1 spectra , this dual - peak structure is most likely due to self - absorption , rather than a second line - of - sight cloud .
the rest frequencies of the hco@xmath1 3@xmath22 transition and dco@xmath1 3@xmath22 transition given by the jcmt and lamda databases are consistent and thus , we do not believe that this alignment is a coincidental artifact due to uncertainty in the rest frequencies of these lines .
gregersen et al ( 1997 ) showed that the type of blue asymmetry predominately seen in the hco@xmath1 3@xmath22 spectra is indicative of infall motion .
this asymmetry is most likely not due to random motions in the core , as this would produce a blue excess value close to zero .
rotation , alone , would also produce a blue excess value close to zero ( pavlyuchenkov et al .
the prevalence of blue asymmetry suggests that the n6 core is collapsing .
models of well - shielded , starless , dense cores have shown that the central kinetic temperature will be low ( @xmath57 k ) and that temperature increases with increasing radius due to heating by an external radiation field ( e.g. , hollenbach et al .
after the onset of gravitational collapse , the central region should be warmed by the release of gravitational potential energy and the central temperature will increase ( e.g. , lee et al .
this is the behavior seen in our dust continuum temperature map ( figure [ fig : temp ] ) in which the central region ( 20.2 k ) is slightly warmer than the surrounding dust envelope ( 15.4 k ) and is consistent with the onset of collapse of the core . as the temperature increase observed in the n6 core
is still relatively small , it has probably only recently started to collapse .
gas at temperatures above 20.2 k , however , may exist in a region smaller than the size of our grid spacing .
such a collapse is also consistent with the lack of turbulent support mentioned in section [ turbulence ] .
the lack of cs emission from the center of n6 is also consistent with the core being at an early stage of collapse , since cs is predicted to evaporate off dust grains only during the later stages of gravitational collapse ( e.g. , charnley 1997 ) .
the spitzer c2d survey of the oph a region did not detect any infrared source consistent with a young stellar object ( yso ) within the boundaries of the n6 core , which further suggests that the n6 core has yet to evolve to the class 0 protostar stage .
it should be noted that the blue asymmetry in the hco@xmath1 spectra is not strongly peaked around the central parts of the n6 core but rather , is detected at multiple locations toward the periphery of the core as well as at the central locations of the core .
furthermore , since hco@xmath1 is most likely depleted in the innermost parts of the n6 core , as suggested by the depletion of h@xmath3co@xmath1 and dco@xmath1 , this infall signature is most likely only coming from the mid - to - outer layers of the cloud .
our tentative detection of a temperature increase at the center of the core , however , suggests that this infall motion does extend to the core s center .
these points suggest that the n6 core is not undergoing an inside - out collapse , as predicted by the shu ( 1977 ) model , but rather , is undergoing a global collapse along the lines of the larson ( 1969 ) and penston ( 1969 ) models .
indeed various numerical hydrodynamic simulations ( e.g. , foster & chevalier 1993 ; hennebelle et al .
2003 ) have shown that the larson - penston model provides the better approximation at early times of the collapse .
we also can not discount the possibility that we are just observing a global compression or contraction of the n6 core and not a gravitationally induced collapse .
our observations suggest that the oph a - n6 core is an extremely young prestellar core , which has been recently cut off from mhd support and begun to collapse .
evidence to support this claim comes from : 1 . _ the chemical structure of n6 _ - the n@xmath0d@xmath1 map shows a well - defined core centered at [ -10,10 ] , which is slightly offset from the peak of the n@xmath0h@xmath1 emission at [ 0,0 ] ( dam04 ) .
the n@xmath0d@xmath1 emission also appears to be slightly more compact than the n@xmath0h@xmath1 emission , which shows a small ridge extending approximately 10@xmath58 to the northwest .
the dco@xmath1 and h@xmath3co@xmath1 maps , however , show little sign of a centrally concentrated core - like structure .
rather , they mimic the dust continuum gradient ( di francesco et al .
2008 ) suggesting that these two molecules are depleted at the center of the n6 core .
the cs maps show a ring of emission around the central location of n6 , also indicating significant depletion of cs at the center . as a result ,
we believe that the n6 core is composed of shells of different chemical composition due to the different densities at which each chemical species freezes out onto dust grains . in this model , the undepleted species n@xmath0d@xmath1 and n@xmath0h@xmath1 trace the cold , high - density interior of n6 , dco@xmath1 and h@xmath3co@xmath1 trace a region which lies further from the center , and cs traces the outermost edges of the core . this type of molecular freeze out is a common feature in other dense starless cores ( e.g. tafalla et al . 2006 ) .
2 . _ the decreased line widths in the center of n6 _ - the average n@xmath0d@xmath1 and n@xmath0h@xmath1 line widths in the central region of the core are @xmath59 0.3 km s@xmath13 .
the dco@xmath1 and cs line widths , however , are 0.44 km s@xmath13 and 0.55 km s@xmath13 respectively . given our proposed chemical structure for n6 , this is consistent with the line widths increasing with increasing distance from the center .
a calculation of the non - thermal component required to match the observed line widths suggests that the non - thermal contribution to the line widths is greater at the edge of n6 than in the core center .
this supports the claim that the center of n6 has been cut off from turbulent support .
3 . _ observed infall motion _ - the hco@xmath1 spectra in n6 were found to exhibit distinct blue asymmetries across the n6 core .
the average asymmetry parameter for the hco@xmath1 spectra ( mardones et al .
1997 ) was found to be -0.71 which suggests that the mid - to - outer layers of the n6 core , which are probed by the optically thick hco@xmath1 spectra , are falling towards the center of the core .
we also tentatively detected a slight increase in temperature towards the center of the n6 core , which suggests that this infall motion extends to the center of the core if this temperature increase is due to the release of gravitational potential energy .
if the temperature is highest in the center and lower at the edges , the gradient of non - thermal motions is steeper than the line widths above suggest .
while our data have revealed a shell - like structure for the n6 core , this analysis is still restricted by the relatively low spatial resolution of the jcmt at these frequencies .
interferometric observations of n6 of these transitions would help constrain the radius at which each species begins to deplete and may confirm our proposed chemical structure .
a detailed high - resolution examination of the kinematic structure of the n6 core is currently being conducted by bourke et al .
( 2009 , in preparation ) .
an investigation into the ionization level of the n6 core ( e.g. see hezareh et al . 2009 , in preparation ; plume et al .
1998 ) may also provide further evidence for whether mhd waves are not propagating into the interior of the core and hence , whether the cut - off from mhd waves is the cause of the decreased non - thermal support in the center of oph a - n6 .
+ we would like to thank the staff of the joint astronomy centre in hilo , david gibson , melissa gurney , tyler bourke , doug johnstone , helen kirk , sarah sadavoy , andrew yau and michael weiser for their advice on this project .
ap would also like to thank the natural sciences and engineering research council of canada for providing funding for this research via an undergraduate student research award and a canada graduate scholarship .
this research has made use of nasa s astrophysics data system .
di francesco , j. , evans , n. j. , ii , caselli , p. , myers , p. c. , shirley , y. , aikawa , y. , & tafalla , m. 2007 , in protostars and planets v , ed .
b. reipurth , d. jewitt , & k. keil ( tucson : univ .
arizona press ) , 17 h@xmath1 1@xmath20 , n@xmath0d@xmath1 3@xmath22 , cs 5@xmath24 , cs 7@xmath26 , dco@xmath1 3@xmath22 , dco@xmath1 5@xmath24 , h@xmath3co@xmath1 3@xmath22 and h@xmath3co@xmath1 4@xmath23 transitions .
the cross on each map denotes the location of the submillimeter continuum peak found interferometrically by bourke et al .
( 2009 , in preparation ) .
the contour lines are separated by 2 @xmath23 on each map .
the 1 @xmath23 uncertainties for each map , in k km s@xmath13 , from top left to bottom right , are 1.4 , 0.03 , 0.07 , 0.04 , 0.06 , 0.03 , 0.06 and 0.12 respectively .
the uncertainty for the n@xmath0h@xmath1 was taken from dam04 , the uncertainties for the cs , dco@xmath1 and h@xmath3co@xmath1 were found by multiplying the rms from class baseline fit by the width of the lines and the uncertainty for the n@xmath0d@xmath1 was taken to be twice the product of the rms of class baseline fit and the width of an n@xmath0d@xmath1 line.,width=384 ] ( 1,1 ) transition from the [ 0,0 ] position in black and our best fit to the spectrum in red .
the lower spectrum shows the nh@xmath4 ( 2,2 ) transition from the [ 0,0 ] position in black and our best gaussian fit in red.,width=576 ] | we have observed the oph a - n6 prestellar core in the following transitions : n@xmath0d@xmath1 j=3@xmath22 , dco@xmath1 j=3@xmath22 and j=5@xmath24 , hco@xmath1 j=3@xmath22 , cs j=5@xmath24 and j=7@xmath26 , and h@xmath3co@xmath1 j=3@xmath22 and j=4@xmath23 , using the james clerk maxwell telescope .
we also observed the nh@xmath4 ( 1,1 ) and ( 2,2 ) inversion transitions towards the oph a - n6 peak with the green bank telescope .
we have found that the n6 core is composed of shells of different chemical composition due to the freezing out of chemical species at different densities .
the undepleted species n@xmath0d@xmath1 appears to trace the high - density interior of the core , dco@xmath1 and h@xmath3co@xmath1 trace an intermediate region , and cs traces the outermost edges of the core . a distinct blue - red spectral asymmetry , indicative of infall motion , is clearly detected in the hco@xmath1 spectra , suggesting that n6 is undergoing gravitational collapse .
this collapse was possibly initiated by a decrease in turbulent support suggested by the fact that the non - thermal contribution to the line widths is smaller for the molecular species closer to the center of the core .
we also present a temperature profile for the core .
these observations support the claim that the oph a - n6 core is an extremely young prestellar core , which may have been recently cut off from mhd support and begun to collapse . |
the ability to reliably prepare multi - particle entangled states is crucial for applications in quantum information science @xcite and quantum enhanced metrology @xcite .
dicke states @xcite constitute a particularly robust class of maximally entangled states @xcite .
their preparation has been demonstrated with photons @xcite and ions @xcite . however , these approaches are difficult to scale up to many particles . in experiments with neutral atoms dicke states
have been prepared using spin - changing collisions @xcite or cavity - based quantum non - demolition measurements @xcite . in the latter case
the successful preparation of a dicke state is heralded by light . in those experiments ,
the success probability is extremely low if high fidelity is to be achieved .
ensembles of rydberg atoms with long - range interactions offer a platform that could overcome both the problem of scalability and low success probability .
first demonstrations of the preparation of a @xmath0 state ( or first dicke state ) @xcite with rydberg atoms have been reported @xcite , however , they face the problem that due to the probabilistic loading of traps the atom number is not know precisely . since the collective rabi coupling depends on the atom number this leads to relatively low fidelity of the state preparation .
we propose a scheme in which the successful preparation of a @xmath0 state is heralded by the detection of an ion , making use of the possibility perfect rydberg blockade in small ensembles of atoms .
this scheme achieves near - unity preparation fidelity and the atom number uncertainty only deteriorates the success probability , which is still much larger than in experiments with atoms in optical cavities .
( b ) level schemes of single atoms states with laser couplings and symmetrized ensemble states ( dicke states ) . by strongly coupling pairs of collective states
the coupling laser causes ac stark shifts , which differ between different rungs of the dicke ladder .
this makes it possible to tune the the probe laser to resonance between the collective ground state @xmath1 and only one dressed excited state . ]
the proposed @xmath0 state preparation scheme is illustrated in fig .
[ fig : scheme](a ) . in brief , an ensemble of @xmath2 atoms
is confined to a volume smaller than the rydberg blockade radius , such that only one atom can be excited to a rydberg state , and all atoms are initialized in the ground state @xmath3 .
the ensemble is then excited to a collective state containing one rydberg excitation @xmath4 and one excitation in a metastable intermediate state @xmath5 .
subsequently , the number of rydberg excitations is measured by field ionization and detection of the removed ion .
the remaining @xmath6 particles are then left in the state @xmath7 .
the key point of this scheme is that the successful preparation is _ heralded _ by the detected ion
. the excitation step will not always be successful due to fluctuating atom numbers or other imperfections , but if the rydberg atom was detected , one can be sure that the @xmath0 state was prepared successfully .
otherwise the preparation scheme can be repeated until the preparation is successful .
essential for this scheme is the ability to drive rabi - oscillations between two collective states without populating any other states .
this is achieved by strongly laser coupling the intermediate and the rydberg state ( @xmath8 ) , as illustrated in fig .
[ fig : scheme](b ) , which causes ac - stark shifts of the collective states .
these shifts differ between states with different numbers of excited atoms since the effective rabi - couplings between these states are collectively enhanced for higher excited states .
this allows us to tune the probe field to the resonance between the collective ground state @xmath9 and a doubly excited state .
in the following we introduce the mathematical framework employed to simulate the preparation scheme outlined above and to assess its fidelity .
the atoms are regarded as motionless on the timescale of the experimental sequence . for
now , all three atomic levels are assumed to be non - decaying .
these assumptions and concrete experimental realizations will be discussed in secs .
[ sec : issues ] and [ sec : experiment ] .
the hamiltonian that governs the dynamics of an ensemble of @xmath2 atoms , subject to probe- and coupling lasers resonantly driving the transitions between the ground state @xmath3 and the intermediate state @xmath5 and between @xmath5 and the rydberg state @xmath4 , respectively , in rotating wave approximation , reads ( @xmath10 ) @xmath11 \label{eq : hamiltonian}\ ] ] where @xmath12 and @xmath13 and @xmath14 are the detunings of the probe laser from the @xmath15-transition and of the coupling laser from the @xmath16-transition , respectively .
the level scheme of a single atom is illustrated in fig .
[ fig : scheme](b ) . the same level scheme ,
now in the rotating reference frame is shown in fig .
[ fig : rabi_osc](a ) , where the given energies are the diagonal elements of the hamiltonian and the double arrows symbolize off - diagonal elements , or couplings , between the states . in the given case
the probe laser is blue detuned ( @xmath17 ) and the coupling laser red detuned ( @xmath18 ) with respect to the atomic transition energies .
we assume perfect rydberg blockade @xcite , which means that the rydberg - rydberg interactions are only reflected by the fact that many - body states with more than one atom in the rydberg state are excluded from the hilbert space .
thus , in the hamiltonian itself no interaction term appears .
the hamiltonian is invariant under exchange of particles .
therefore , it is useful to transform the problem to a basis of collective states that also show this symmetry , which is the dicke state basis . in the absence of decoherence effects the dynamics
is restricted to the manifold of symmetric dicke states , which tremendously simplifies its theoretical treatment @xcite .
these states are fully symmetric superpositions of all excited states with the same number of @xmath19 and @xmath20 excitations .
we label them as @xmath21 with @xmath22}$ ] .
we use the shorthand notation @xmath23 , @xmath24 and @xmath25 . note that the index @xmath26 for a perfectly blockaded ensemble .
the @xmath8 part of the hamiltonian only couples states with equal @xmath27 which means that in the case of perfect blockade , the coupling laser hamiltonian splits into two - by - two blocks as illustrated in fig .
[ fig : rabi_osc](a2 ) .
we further simplify the problem by diagonalizing the coupling laser hamiltonian and working in the resulting dressed state basis .
the resulting dressed state energies ( diagonal part of the hamiltonian in the dressed basis ) are shown in figure [ fig : rabi_osc](a3 ) .
the key idea of our proposal is now that , by adjusting the parameters @xmath13 , @xmath14 , and @xmath8 , the energy of the state @xmath28 can be tuned into resonance with the state @xmath1 while all other states are far off - resonant .
this choice will result in coherent rabi oscillations between @xmath1 and @xmath28 , while the populations of all other states stay small .
( b ) , now in the rotating frame .
the energy @xmath29 of the levels are the diagonal elements of the hamiltonian .
the coupling @xmath8 leads to a repulsion between pairs of dicke state .
we choose the parameters such that the transition between states @xmath1 and @xmath28 is resonant .
( b ) and ( c ) show the rabi oscillations between these sates .
( b ) relatively small @xmath30 .
imperfections due to coupling to singly and triply excited sates are visible .
we compare solutions of the time dependent schrdinger equation on the full state space ( black dots ) to the restriction to the lowest dressed dicke states ( solid green line ) and the two - state model with all but states @xmath1 and @xmath28 eliminated ( dashed blue line ) .
the red dotted line shows the ratio of rydberg excitation probability in states other than @xmath31 over the total rydberg probability .
parameters are @xmath32 , @xmath33mhz , @xmath34mhz , @xmath35 , @xmath36 .
( c ) different rabi frequencies and atom numbers .
solid green and dotted red line : @xmath37mhz , @xmath38mhz .
dot - dashed black and dashed blue line : @xmath39mhz , @xmath40mhz .
@xmath41 was adjusted to yield the same period of the effective rabi oscillations for each @xmath8 .
for small @xmath30 near - perfect transfer to the state @xmath28 is achieved.,title="fig : " ] + ( b ) , now in the rotating frame .
the energy @xmath29 of the levels are the diagonal elements of the hamiltonian .
the coupling @xmath8 leads to a repulsion between pairs of dicke state .
we choose the parameters such that the transition between states @xmath1 and @xmath28 is resonant .
( b ) and ( c ) show the rabi oscillations between these sates .
( b ) relatively small @xmath30 .
imperfections due to coupling to singly and triply excited sates are visible .
we compare solutions of the time dependent schrdinger equation on the full state space ( black dots ) to the restriction to the lowest dressed dicke states ( solid green line ) and the two - state model with all but states @xmath1 and @xmath28 eliminated ( dashed blue line ) .
the red dotted line shows the ratio of rydberg excitation probability in states other than @xmath31 over the total rydberg probability .
parameters are @xmath32 , @xmath33mhz , @xmath34mhz , @xmath35 , @xmath36 .
( c ) different rabi frequencies and atom numbers .
solid green and dotted red line : @xmath37mhz , @xmath38mhz .
dot - dashed black and dashed blue line : @xmath39mhz , @xmath40mhz .
@xmath41 was adjusted to yield the same period of the effective rabi oscillations for each @xmath8 . for small @xmath30 near - perfect transfer to the state @xmath28
is achieved.,title="fig : " ] +
figure [ fig : rabi_osc ] illustrates that rabi oscillations are indeed observed and perfect population transfer from @xmath1 to @xmath28 is achieved if the parameter @xmath30 is small .
thus , the excitation step of the proposed preparation scheme will be a weak probe pulse that induces complete population inversion ( @xmath42-pulse ) .
since the state @xmath28 is a superposition of the dicke states @xmath43 and @xmath31 , ideally , the induced dynamics involves only these states and the state @xmath1 .
this means if a rydberg atom is detected in the ionization step , the system was in state @xmath31 previously , and the measurement projects the remaining @xmath6 atoms onto the dicke state @xmath44 , which is the desired @xmath0 state . as shown in fig .
[ fig : rabi_osc](b ) the period of the effective rabi oscillations depends on @xmath2 , which means that the population transfer will in general be imperfect if @xmath2 fluctuates . however , if the applied laser pulse is not a precise @xmath42-pulse , only the probability of ion detection decreases , but the probability for successful @xmath0 state preparation _ if _ an ion was detected , does not .
the probability of detecting a rydberg excitation is what we call the _ success probability _ , while the probability that in the case of an ion detection a @xmath0 state was indeed prepared is termed _
the fidelity is illustrated by the red dotted line in fig .
[ fig : rabi_osc](a ) , which shows the fraction of detected ions not stemming from the state @xmath31 ( infidelity , fraction of false heralds ) .
this quantity has a wide plateau around @xmath45 , which illustrates the tolerance of the scheme with respect to excitation with pulses of imprecise length and thus to fluctuating atom numbers .
what leads to a decrease of the state preparation fidelity is the admixture of states other than @xmath1 and @xmath28 which contain rydberg excitations . in order to avoid this , the ( ac stark shifted )
dressed state energies of all other states should differ significantly from zero on the level of the typical probe - rabi - coupling , i.e. , @xmath46 [ see fig . [
fig : rabi_osc](b2 ) ] . in this parameter regime
the relevant state space can be restricted to the low excitation sector .
it turns out to be sufficient to restrict to the dressed states @xmath1 , @xmath47 , @xmath48 , @xmath28 , @xmath49 , and @xmath50 , which is demonstrated in fig . [ fig : rabi_osc](a ) by comparing to calculations on the full state space . by adiabatic elimination of all but the states
@xmath1 and @xmath28 we further obtain analytic expressions for effective rabi - frequency @xmath51 and detuning @xmath52 , which will be given below for a specific set of parameters .
in the following we study the performance of the scheme more systematically and optimize the laser parameters .
the requirement of degeneracy between the states @xmath1 to @xmath28 leaves a certain freedom in the choice of parameters . for a given coupling strength @xmath8
, we can in principle choose the detuning @xmath14 freely .
the degeneracy condition can still always be satisfied by adjusting @xmath13 .
the effective rabi frequency @xmath53 , which determines the duration of the preparation scheme , will have to be chosen such that we operate on a timescale fast compared to typical decoherence times of the system . choosing a specific @xmath53 will fix the probe rabi frequency @xmath41 .
we characterize the quality of our scheme by the two quantities mentioned above : the success probability of the state preparation corresponds to the probability of exciting the target state @xmath31 , and the infidelity , or probability of false heralds , is the probability of rydberg excitation in states other than @xmath31 divided by the total rydberg probability
. state preparation protocol .
( a ) rydberg probability @xmath54 ( success probability ) and fidelity as a function of @xmath14 .
the fidelity shows an optimum at approximately @xmath35 .
the success probability is @xmath55 at the optimal point and approaches 90% at @xmath56 .
parameters : @xmath57 , @xmath37mhz . here , as in ( c ) and
( d ) , @xmath41 was adjusted such that @xmath58s and @xmath13 is determined by the resonance condition . in panels ( b)-(d )
the condition @xmath35 was imposed .
( b ) dependence on the atom number @xmath2 for fixed laser parameters .
the laser parameters are chosen such that the success probability is maximal for @xmath59 atoms .
the dot - dashed green line shows the probability @xmath60 to prepare an ensemble of @xmath2 atoms for a poissonian distribution with a mean of @xmath59 atoms .
the protocol shows reasonable success probability and fidelity across the whole atom number distribution .
( c ) fidelity as a function of @xmath8 for different atom numbers . at @xmath61mhz , fidelities beyond 99% can be achieved .
the black line shows an estimated upper bound on the fidelity .
( d ) dependence on different sources of decoherence .
the fidelity increases linearly with spontaneous decay rate from intermediate ( @xmath62 ) and rydberg state ( @xmath63 ) , as well as with the single - atom dephasing rate @xmath64 . ]
figure [ fig : infidelity](a ) shows these two quantities as a function of the chosen detuning of the coupling laser @xmath14 .
we used @xmath65mhz , corresponding to an excitation time of @xmath58s .
we find that the minimal infidelity is reached close to @xmath35 , marked by the dotted vertical line . here
, the probability of preparing the state @xmath28 is approximately @xmath55 .
note that the qualitative behavior , especially the position of the minimum infidelity is largely independent of the choice of @xmath2 and @xmath8 . in
the following we focus on the choice @xmath35 , for which the degeneracy condition @xmath66 implies @xmath67 . for
this choice one obtains the dressed state @xmath68 .
the adiabatic elimination yields the effective rabi frequency @xmath69 and the effective detuning @xmath70 .
the effective detuning can be compensated ( approximately ) by modifying the probe detuning to @xmath71 ( as also done in fig .
[ fig : rabi_osc ] ) . in order to demonstrate that this fidelity is not decreased much by fluctuations in the atom number @xmath2 we calculated the mean fidelity for a poissonian atom number distribution with mean @xmath72 for @xmath73mhz .
the result is shown in fig .
[ fig : infidelity](b ) .
the weighted average over the poisson distribution gives an infidelity of @xmath74 instead of @xmath75 obtained for @xmath59 with an optimal @xmath42-pulse .
this also illustrates that the excitation time can be chosen much smaller than @xmath42-pulse without loss of fidelity , but only at the expense of a smaller success probability , which makes the scheme robust not only with respect to atom number fluctuations but also to imperfect control of the spatial and temporal shape of the probe laser pulse . in fig .
[ fig : infidelity](c ) we show the infidelity as a function of @xmath8 for different atoms numbers keeping @xmath65mhz fixed .
the infidelity decreases as @xmath76 and becomes smaller than @xmath77 for rabi frequencies exceeding @xmath78mhz . by estimating the off - resonant admixture of states containing rydberg excitations other than @xmath31 we find that @xmath79 presents an upper bound on the infidelity .
so far we have assumed perfectly coherent dynamics .
however , in a realistic scenario various sources of decoherence , such as spontaneous emission , finite laser line - widths , and atomic motion will be present . in the following we discuss the effect of such imperfections on the fidelity of our protocol . in order to include decoherence effects
the system is described by a density operator obeying the master equation ( me ) @xmath80+\mathcal{l}[\rho ] \,.\ ] ] decoherence effects can be included via lindblad terms @xmath81=\gamma_k\left(s_{k}\rho s_{k}^\dagger - 1/2\left\{s_{k}^\dagger s_{k},\rho\right\}\right)\ ] ] where , e.g. , @xmath82 ( @xmath83 ) for spontaneous emission from the intermediate ( rydberg ) level , respectively , and @xmath84 for single - atom dephasing of the rydberg state ( @xmath85 is the atom index ) . to account for dephasing effects that do not affect all atoms independently but the ensemble as a whole
, we consider projectors on the collective states @xmath86 .
decay of the intermediate level leads to incoherent population transfer from @xmath31 to singly excited states involving the unwanted state @xmath87 and thus decreases the fidelity .
decay from the rydberg state will harm the proposed scheme less , since such a decay leaves the system in a state without rydberg excitation and thus does not produce false heralds .
the infidelity as a function of decay rates is shown in fig .
[ fig : infidelity](d ) for @xmath57 atoms , @xmath73mhz , and @xmath65mhz
. the infidelity increases linearly with the time - integrated decay probability ( @xmath88 ) , where the proportionality factor is about a factor of five smaller for the dependence on @xmath63 compared to @xmath62 .
other decoherence effects include finite laser linewidths , atomic motion , and stray electric fields stark shifting the rydberg state .
these effects can result in single - atom or collective dephasing .
collective dephasing here means , that the coherences between the symmetric dicke state are damped but no coupling to asymmetric states is caused as the perturbing process also conserves the exchange symmetry of the atoms .
such process do not harm our scheme much .
single atom dephasing , on the other hand , leads to the distribution of population into states outside the symmetric dicke state manifold and therefore reduces the probability to prepare the symmetric state @xmath31 .
again , the infidelity increases linearly with the dephasing rate as shown in fig .
[ fig : infidelity](d ) . to avoid imperfection due to decoherence
short excitation times ( large @xmath51 ) are required . on the other hand
we have seen that in order to minimize couplings to unwanted state small ratios of @xmath89 are favorable .
thus , in order to reach the maximum fidelity of @xmath0 state preparation for given @xmath8 and given rates of the decoherence process , the effective rabi frequency ( or duration of the protocol ) can be optimized and the maximum fidelity for given atomic parameters depends crucially on the achievable coupling laser strength .
another possible issue is whether the ion and electron that might traverse the atomic ensemble after ionization would degrade the fidelity by inducing local stark shifts or collisions with the remaining atoms . in order to assess this effect we performed classical monte carlo trajectory simulations assuming that the ionizing electric field is ramped from @xmath90 to @xmath91kv / cm in @xmath92ns @xcite . as an example we use the differential @xmath93-polarizability @xmath94 of @xmath95sr clock transition @xcite to estimate the induced local phase shifts .
the ion leaves a trap of diameter @xmath96 m in about @xmath97ns . for a trap of volume @xmath96m@xmath98 containing @xmath59 atoms ,
the probability to induce a significant phase shift in an atom - ion collision turns out to be smaller than @xmath99 .
the ionizing electric field itself also does not affect the coherence between states @xmath3 and @xmath5 significantly .
the proposed scheme assumes that one operates in the regime of perfect rydberg blockade , meaning that the interaction shift is much larger than the laser line widths and coupling strengths @xmath100 .
this means that the ensemble size should not exceed a few micrometers .
at the same time the mean distance between the particles must not become too small such that effects of the interaction of the rydberg electron with the surrounding ground state atoms can be neglected .
this limit the atom numbers that can be used to a few tens or hundreds .
a more detailed account of the effects of imperfect blockade and ground - rydberg state interactions are left for future studies .
we have seen in the previous section that the most important requirement for achieving high fidelities in the proposed scheme is a small spontaneous decay rate from the intermediate level @xmath5 .
note that using , e.g. , the @xmath101 state of rubidium as intermediate state is not possible due to its fast decay rate of tens of mhz .
two possible ways to overcome this problem are outlined in the following . using rubidium or other alkali atoms ,
one possibility is to use two hyperfine ground states as @xmath3 and @xmath5 .
the coupling to the rydberg state can then either be achieved by a two - photon transition via an ( off - resonant ) optically excited state or by directly coupling to rydberg state using a laser in the uv - range @xcite .
the transition between the two lower states can either be driven by microwaves or by a two - photon raman transition .
such schemes have recently been successfully implemented experimentally @xcite .
the raman scheme might also allow for a doppler free implementation @xcite .
a second possiblity is to use alkaline earth ( like ) atoms , such as ca , mg , sr , or yb , that are promising because of their metastable triplet @xmath102-states , which have been extensively used in optical atomic clocks ( see ref .
@xcite for a recent review ) . many alkaline earth ( like ) species have been laser - cooled to microkelvin temperatures and their transitions from the @xmath103 ground state to states @xmath104 or even the doubly forbidden @xmath105 offer a wide range of linewidths ( see e.g. table v in ref .
@xcite ) . from the intermediate state
, @xmath106 rydberg states can be excited by blue or near uv - lasers .
the rydberg interaction potentials are similar to those of alkaline atoms for s - states ( near - isotropic van der waals ) @xcite .
atomic structure calculations for these states have recently been reported @xcite .
also , there have been first experiments using rydberg states of @xmath95sr @xcite .
the greatest challenge for both implementations will be to build strong enough blue lasers . here , recent efforts in the context of rydberg dressing and direct rydberg excitation from the ground state could lead to significant advances in the near future @xcite .
for example , tunable , solid - state lasers producing @xmath107 of narrow - band cw light have been developed for this wavelength regime @xcite .
we also note that most experiments the trapping potential would be anti - trapping for the rydberg states and thus has to be switched off during the preparation sequence involving the population of rydberg states . for a further processing of the prepared @xmath0 states it might be necessary to recapture the ensemble after the preparation , which may pose an additional experimental challenge .
however , if recapture or operation in a trap with magic wavelength rydberg trapping could be achieved , this would allow to repeat the preparation sequence [ fig .
[ fig : scheme](a ) ] without having to reload the trap in every cycle but by re - initializing the internal state of the atoms by optical pumping into the ground state .
this might reduce the atom number in the sample in cases where a rydberg atom was present but was not detected due to the finite efficiency of ion detection .
note , however , that current experiments reach 90% detection efficiency @xcite .
in conclusion we have proposed a scheme for the preparation of a maximally entangled @xmath0 state heralded by the detection of an ion .
this scheme could be implemented using small clouds of ultracold alkali or alkaline earth atoms in a small optical dipole trap .
so far , we restricted ourselves to the low @xmath108 part of level structure of the artificial atom created by the coupling laser .
higher excited collective states could be prepared using a sequence of probe pulses with different frequencies or by employing adiabatic preparation schemes @xcite . due to their multi - particle entanglement ,
higher excited dicke states have many applications in metrology @xcite and quantum information science @xcite .
dicke states are not spin squeezed in the conventional sense @xcite , but nevertheless their entanglement is useful for sub - shot - noise interferomentry . in general , the maximum fringe contrast that can be obtained is a spectroscopic measurement with a given input state is given by @xmath109 , where @xmath110 is the quantum fisher information @xcite
. for dicke states of @xmath2 spins with @xmath111 excitations , one obtains @xmath112 , so for a @xmath0 state @xmath113 we have @xmath114 , which means that the fringe contrast can be enhanced at most by a factor @xmath115 compared to the shot - noise limit .
thus , ideally one would like to prepare the state with @xmath116 , corresponding to @xmath117 in the notation used above .
this state enables spectroscopy at the heisenberg limit @xcite and is also useful for quantum information processing tasks , such as @xmath118 telecloning or open - destination teleportation @xcite .
furthermore , the hamiltonian of the artificial atom can be mapped to the jaynes cummings model , where the role of the @xmath19-excitation is taken by the photons and the @xmath20-state corresponds to the state of the two - level atom in the cavity . in the dicke basis
the hamiltonian reads @xmath119 thus , the probe field @xmath120 would account for a coherent external drive of the cavity while the coupling laser part resembles the hamiltonian of a single atom in a cavity , where @xmath108 is the number of cavity photons and the atomic ground ( excited ) state corresponds to the absence ( presence ) of a rydberg excitation .
one could envision an experiment , in which first a probe pulse is applied which creates a coherent superposition of the states @xmath121 ( binomial distribution ) . then switching on
the coupling laser would induce collapse and revival dynamics of the rydberg population , characteristic for the jaynes - cummings model .
another interesting question concerns the motional state of the ensembles after the removal of the rydberg atom .
can the ionization be done in a coherent way , such that a superposition of different position states is created instead of an incoherent mixture ?
for example , in a setup with high control of atomic positions as in @xcite states entangled not only in the internal , but also in the external degrees of freedom might be prepared that way .
50ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1103/revmodphys.81.865 [ * * , ( ) ] http://dx.doi.org/10.1038/nphoton.2011.35 [ * * , ( ) ] link:\doibase 10.1103/physrev.93.99 [ * * , ( ) ] link:\doibase 10.1103/physreva.91.032108 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.020504 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.020503 [ * * , ( ) ] http://dx.doi.org/10.1038/nature04279 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.155304 [ * * , ( ) ] link:\doibase 10.1126/science.1248905 [ * * , ( ) ] http://dx.doi.org/10.1038/nature14293 [ * * , ( ) ] link:\doibase 10.1103/physreva.62.062314 [ * * , ( ) ] link:\doibase 10.1038/nphys2413 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.043602 [ * * , ( ) ] @noop ( ) link:\doibase 10.1103/physrevlett.87.037901 [ * * , ( ) ] http://stacks.iop.org/1367-2630/16/i=6/a=063042 [ * * , ( ) ] link:\doibase 10.1103/physreva.89.033839 [ * * , ( ) ] link:\doibase 10.1007/s11467 - 013 - 0396 - 7 [ ( ) ] link:\doibase 10.1103/physrevlett.109.263004 [ * * , ( ) ] link:\doibase 10.1103/physreva.89.033416 [ * * , ( ) ] http://stacks.iop.org/1367-2630/16/i=8/a=083034 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.133201 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.104.010503 [ * * , ( ) ] link:\doibase 10.1103/physreva.91.012337 [ * * , ( ) ] link:\doibase 10.1393/ncr / i2013 - 10095-x [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.103601 [ * * , ( ) ] http://stacks.iop.org/0953-4075/47/i=15/a=155001 [ * * , ( ) ] http://stacks.iop.org/0953-4075/45/i=13/a=135004 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.213004 [ * * , ( ) ] link:\doibase 10.1103/physreva.87.013407 [ * * , ( ) ] link:\doibase 10.1103/physreva.87.053409 [ * * , ( ) ] http://dx.doi.org/10.1038/nphys3214 [ * * , ( ) ] link:\doibase 10.1007/s00340 - 011 - 4771 - 1 [ * * , ( ) ] link:\doibase 10.1007/s00340 - 006 - 2274 - 2 [ * * , ( ) ] link:\doibase 10.1007/s00340 - 011 - 4435 - 1 [ * * , ( ) ] @noop ( ) link:\doibase 10.1103/physreva.87.033416 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.260502 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.080504 [ * * , ( ) ] link:\doibase 10.1103/physreva.85.022322 [ * * , ( ) ] link:\doibase 10.1103/physreva.85.022321 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.173604 [ * * , ( ) ] @noop ( ) link:\doibase 10.1103/physreva.47.5138 [ * * , ( ) ] link:\doibase 10.1103/physreva.46.r6797 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.physrep.2011.08.003 [ * * , ( ) ] link:\doibase 10.1103/physreva.79.042334 [ * * , ( ) ] link:\doibase 10.1103/physreva.82.012337 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.113002 [ * * , ( ) ] link:\doibase 10.1103/physrevx.4.021034 [ * * , ( ) ] | we propose a scheme for preparing an ensemble of atoms in a maximally entangled @xmath0 state by exploiting the rydberg blockade effect .
the success of our protocol is indicated by the detection of an ion , which thus serves as a herald for successful entangled state preparation .
the fidelity of the preparation scheme is independent of the number of atoms in the ensemble .
therefore , a small cloud of atoms in a dipole trap randomly loaded from a background gas can be reliably prepared in a maximally entangled state despite of atom number fluctuations . |
characterizing old stellar populations provide important constraints on the early star formation histories of galaxies .
only the satellite galaxies of the milky way ( mw ) are sufficiently close to resolve individual stars well below the oldest main - sequence turn - offs , which is a pre - condition for accurate photometric age dating of old stellar populations .
all local group galaxies , for which adequate data exist , appear to contain stars older than 10 gyr @xcite .
this result is based on main - sequence turn - off photometry of globular clusters and field populations in galactic satellites and a few more distant local group galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , as well as the detection of horizontal branch stars ( including rr lyrae variables ) in the local group and beyond .
globular clusters are preferred as the basis for old stellar population age tracers since they are usually single - age , single - metallicity objects facilitating comparative studies .
moreover , while globular cluster systems exhibit a range of ages ( e.g. , * ? ? ?
* ) , the oldest ones may belong to the most ancient surviving stellar systems to have completed their formation in the youthful universe ( e.g. , * ? ? ?
in those nearby galaxies where relative age dating based on main - sequence photometry was carried out in comparison to the oldest globular clusters in the milky way , no age difference within the measurement accuracy was found ( e.g. , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* and references therein ) .
the relative age dating of the oldest identifiable population ii objects thus indicates a common epoch of substantial early star formation in the milky way and its companions , although information about a putative , even older population iii remains to be uncovered in these objects . a galaxy that may _ not _
share this common epoch of early star formation at least not with respect to its globular clusters ( e.g. , * ? ? ?
* ) is the small magellanic cloud ( smc ) .
the smc is one of the closest and therefore best studied dwarf galaxies orbiting our galaxy .
while the smc hosts a large number of intermediate - age and young star clusters , it only contains one old globular cluster , ngc121 , which is also the most massive star cluster .
ngc121 is located @xmath7 2.4 ( @xmath73 kpc ) west of the smc bar at ( @xmath8 , @xmath9 ) = ( @xmath10 , @xmath11 ) .
ngc121 is the only cluster in the smc that is sufficiently old to have developed an extended red horizontal branch @xcite and to contain rr lyrae stars . indeed , whether or not to call a star cluster a globular cluster is a matter of definition .
in this case we refer to @xcite who consider lindsay1 as having a stumpy red clump and not a red horizontal branch .
three rr lyrae stars were discovered in ngc121 by @xcite .
@xcite found a fourth rr lyrae variable in the cluster and an additional 75 in a 1 @xmath12 1.3 square degree field centered on ngc121 .
studies of various clusters in the lmc and in the mw showed that the presence of rr lyrae variables indicates that the parent population is as old as or older than @xmath7 10 gyr .
an important question is whether ngc121 is as old as the typical old globular clusters in the large magellanic cloud ( lmc ) and in the mw .
previous studies found ages ranging from 8 to 14 gyr for ngc121 @xcite using a variety of different techniques .
studies based on the deepest available color - magnitude diagrams from hubble space telescope ( hst ) observations with the wide field and planetary camera 2 ( wfpc2 ) indicate an age of 10 to 10.6 gyr for ngc121 , suggesting that this globular cluster is several gyr younger than the oldest globulars in other nearby galaxies and in the mw @xcite .
the capabilities of the advanced camera for surveys ( acs ) provide an improvement in both sensitivity ( depth ) as well as angular resolution , which is essential for a reliable photometric age determination in this dense star cluster . here
we present deep photometry of ngc121 obtained with acs aboard the hst .
we determine the age of ngc121 utilizing both absolute and relative methods ( e.g. , * ? ? ?
the current study is the first in a series of papers based on hst studies of rich intermediate - age and old star clusters in the smc .
in addition to ngc121 , six intermediate - age smc star clusters have been observed as part of our program : lindsay1 , kron3 , ngc339 , ngc416 , lindsay38 and ngc419
. we will derive fiducial ridgelines and fit isochrones to obtain accurate ages for each cluster using the same reduction techniques and isochrone models as described here ( see @xmath13 [ sec : obs]-[sec : age ] ) , and will present our results in future papers . in table
[ tab : journalobs ] we list the cluster identification , date of observation , passband , exposure times and location of all clusters in our hst program ( go-10396 ; principal investigator : j. s. gallagher ) . in the next section
we describe the data reduction procedure . in @xmath13 [
sec : cmd ] we present the color - magnitude diagram ( cmd ) of ngc121 and discuss its main features . in @xmath13 [ sec : age ] we describe our age derivation methods and present our results .
the smc cluster ngc121 was observed with hst s acs on 2006 march 21 as part of our program focused on star clusters and field stellar populations in the smc .
the program aims at exploring the star formation history and properties of the smc using both a number of carefully selected clusters and field regions . for ngc121 we obtained imaging in the f555w and f814w filters , which resemble the johnson v and i filters in their photometric properties @xcite .
the images were obtained using the wide field channel ( wfc ) of acs and cover an area of @xmath14 with a pixel scale of @xmath7 0.05 arcsec .
one set of exposures was taken at the nominal position of the cluster center .
four long exposures were obtained in each filter for hot pixel removal and to fill the gap between the two halves of the 4096 @xmath12 4096 pixel detector .
each pointing has an exposure time of 496 s in the f555w , and 474 s in the f814w filter .
moreover , two short exposures were taken in each filter with an exposure time of 10 s in f555w and 20 s in f814w .
ccccccc ngc121 & j96106030 & 2006@xmath1503@xmath1521 & f555w & 40.0 & @xmath16&@xmath17 + & j96106040 & & & 1984.0 & @xmath18 & @xmath19 + & j96106010 & & f814w & 20.0 & @xmath16&@xmath17 + & j96106020 & & & 1896.0 & @xmath18 & @xmath19 + lindsay1 & j96105030 & 2005@xmath1508@xmath1521 & f555w & 40.0 & @xmath20 & @xmath21 + & j96105040 & & & 1984.0 & @xmath22 & @xmath23 + & j96105010 & & f814w & 20.0 & @xmath20 & @xmath21 + & j96105020 & & & 1896.0 & @xmath22 & @xmath23 + kron3 & j96107030 & 2006@xmath1501@xmath1517 & f555w & 40.0 & @xmath24 & @xmath25 + & j96107040 & & & 1984.0 & @xmath26 & @xmath27 + & j96107010 & & f814w & 20.0 & @xmath24 & @xmath25 + & j96107020 & & & 1896.0 & @xmath26 & @xmath27 + ngc339 & j96104030 & 2005@xmath1511@xmath1528 & f555w & 40.0 & @xmath28 & @xmath29 + & j96104040 & & & 1984.0 & @xmath30 & @xmath31 + & j96104010 & & f814w & 20.0 & @xmath28 & @xmath29 + & j96104020 & & & 1896.0 & @xmath30 & @xmath31 + ngc416 & j96121030 & 2006@xmath1503@xmath1508 & f555w & 40.0 & @xmath32 & @xmath33 + & j96121040 & & & 1984.0 & @xmath34 & @xmath35 + & j96121010 & & f814w & 20.0 & @xmath32 & @xmath33 + & j96121020 & & & 1896.0 & @xmath34 & @xmath35 + lindsay38 & j96102030 & 2005@xmath1508@xmath1518 & f555w & 40.0 & @xmath36 & @xmath37 + & j96102040 & & & 1940.0 & @xmath38 & @xmath39 + & j96102010 & & f814w & 20.0 & @xmath36 & @xmath40 + & j96102020 & & & 1852.0 & @xmath38 & @xmath39 + ngc419 & j96103030 & 2006@xmath1501@xmath1505 & f555w & 40.0 & @xmath41 & @xmath42 + & j96103040 & & & 1984.0 & @xmath43 & @xmath44 + & j96103010 & & f814w & 20.0 & @xmath41 & @xmath42 + & j96103020 & & & 1896.0 & @xmath43 & @xmath44 + [ tab : journalobs ] the data set was processed adopting the standard space telescope science insitute acs calibration pipeline ( calacs ) to subtract the bias level and to apply the flat field correction . for each filter , the short and long exposures were co - added independently using the multidrizzle package @xcite . with this package the cosmic rays and hot pixels were removed and a correction for geometrical distortion was applied .
the resulting ngc121 data consist of one 40 s and one 1940 s exposure in f555w and one 20 s as well as one 1896 s exposure in f814w .
the two short exposures allowed us to measure brighter stars that are saturated in the long exposures .
the photometric reductions were carried out using the daophot package in the iraf environment .
we discarded saturated foreground stars and background galaxies using the _ source extractor _
package @xcite . due to the different crowding and signal - to - noise ratio properties of the long and the short exposure images , photometry involving point spread function ( psf ) fitting
was only performed on the long exposures .
for the short exposures we used aperture photometry , which turned out to yield smaller formal errors than psf photometry .
we ran daophot on our data and set the detection threshold at 1 @xmath45 above the local background level in order to detect even the faintest sources .
the list of stars detected in the f814w image was then used as coordinate input list to identify the stars in the f555w image and serve as our coordinate reference .
49493 sources were found to be common to both long exposure frames .
for these sources , we performed aperture photometry using an aperture radius of 3 pixels .
we then constructed a psf by combining 150 bright and isolated stars that were distributed fairly uniformlly across the image .
finally , psf photometry was carried out .
the photometric calibration was accomplished by converting the magnitudes of the individual stars to the standard acs magnitude system by using an aperture with a radius of @xmath46 ( or 10 pixels on the image ) , in combination with the aperture correction from the @xmath46 aperture radius to infinity and the synthetic zero points for the acs / wfc @xcite .
the aperture correction was derived for each frame independently .
the objects found in both images were cross - identified and merged with a software package written at the bologna observatory by p. montegriffo ( private communication ) .
altogether we were able to cover a luminosity range of @xmath7 10 magnitudes after combining the resultant photometry of the short and long exposures . in figure
[ fig : ngc121_sigma ] we show the photometric errors assigned by daophot . for stars measured on the short exposures ,
the formal photometric errors remain negligible over a wide range of magnitudes . in the long exposures ,
all the brighter stars with @xmath47 mag and @xmath48 mag are saturated . at @xmath49 19.6 mag , the short exposure ( _ blue dots _ ) and aperture photometry from the long exposure ( _ black dots _ ) samples were combined , and for stars fainter than @xmath49 22.2 mag , long exposure psf photometry ( _ red dots _ ) was used .
we chose where to cut between the aperture and psf photometry catalogues based on the @xmath50 data and adopted the same value for @xmath51 so as to avoid a color slope associated with this division .
for our study , we rejected all stars with a @xmath45 error larger than 0.2 mag and a daophot sharpness parameter @xmath52 in both filters . to obtain a superior cmd , we discarded all stars within a radius of @xmath53 from the cluster center , which excludes the very dense core of the cluster . with this selection ,
our final sample contains 17464 stars common in both filters .
the resulting color - magnitude diagram ( cmd ) of ngc 121 and its surroundings is shown in figure [ fig : ngc121_letterfinal ] .
our cmd for ngc121 reaches @xmath73.5 mag below the msto ( @xmath7 0.5 magnitudes deeper than the previous deepest available photometry ) , which allows us to carry out the most accurate age measurements obtained so far .
the cmd shows a well - populated main sequence ( ms ) , subgiant branch ( sgb ) , red giant branch ( rgb ) , horizontal branch ( hb ) , and asymptotic giant branch ( agb ) .
the gap on the rgb at @xmath54 20 mag is an artificial feature due to small number statistics resulting from our exclusion of crowded stars in the cluster center .
ngc121 appears to be a single - age population object just as one would expect for a canonical star cluster .
as expected , there is no obvious evidence for field star contamination by younger populations due to the location of ngc121 in a low - density area in the outer parts of the smc . within the field of view of the acs and at the high galactic latitude of the smc
, galactic foreground contamination is very low ( e.g. , * ? ? ?
another possible contamination source is the massive and extended galactic globular cluster 47tuc , which has a tidal radius of 42.86 arcmin @xcite and an angular distance from ngc121 of @xmath7 32 arcmin .
we visually estimated the location of the center of ngc121 on the image and selected all stars within an annulus of @xmath53 and @xmath55 to create a bona fide sample .
this cmd is displayed in figure [ fig : ngc121_lettercenter ] .
there is no evidence for a binary sequence in ngc 121 , but we can not exclude their presence , due to the photometric error .
the aforementioned traces of minor field contamination have mainly vanished . due to crowding
, incompleteness becomes significant at the faint end of the ms : this affects particularly faint stars in the cluster center .
hence in figure [ fig : ngc121_lettercenter ] the ms becomes less densely populated at fainter magnitudes .
the red hb is well populated and extends into the rr lyrae instability strip ( clementini et al .
2007 , in prep ) .
the presence of a red hb provides a circumstantial suggestion that ngc121 may be younger than old galactic and lmc globular clusters , since the hbs of the oldest globular clusters tend to extend farther into the blue ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
red hbs , however , can also be due to a second parameter
other than age affecting the hb morphology ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
since a true hb is present , an age measurement for ngc121 can be made using the @xmath56 age measurement , which we will do in @xmath13 [ sec : relage ] .
this method requires the determination of the apparent mean magnitude of the hb .
our data yield @xmath57 = @xmath58 mag for this observable , which is in agreement with the magnitudes determined by shara et al .
( 1998 ) , alves & sarajedini ( 1999 ) and dolphin et al .
( 2001 ) . at @xmath50
= @xmath59 mag we find the ngc121 rgb bump ( @xmath60 ) which is 0.06 mag brighter than the magnitude found by @xcite .
the difference in luminosity is due to the exclusion of the center stars . if we determine the @xmath60 on the entire sample , we obtain @xmath50 = @xmath61 mag , which is in excellent agreement with the magnitude found by @xcite .
this feature is predicted by stellar evolution models , which also show that the luminosity of the rgb bump is dependent on the metallicity and age of the cluster .
when the metallicity is known the difference between @xmath62 and @xmath63 can be used as an age indicator . above the ms turn - off
, @xcite found 42 candidate blue straggler stars ( bss ) .
evolved descendants of the bsss are important as possible sources of stars lying above the traditional hb ( e.g. , * ? ? ?
. in our acs study we recovered the @xcite bss sample and also found more stars in the bss region of which some ( about 20 ) turned out to be pulsating variables ( dwarf cepheids ) .
these stars will be discussed in clementini et al .
( 2007 , in preparation ) where we will present the results of an hst study of variable stars in ngc121 .
age determinations of star clusters using isochrones depend crucially on the interstellar extinction , distance , and metallicity of the cluster , as well as on the chosen isochrone models . in fitting isochrones to the cmd of ngc121 we adopted the spectroscopic metallicity measurement of [ fe / h ] = @xmath64 from da costa @xmath65 hatzidimitriou ( 1998 , see also johnson et al .
2004 ) on the metallicity scale introduced by @xcite ( zw84 ) .
the distance and the extinction were treated as free parameters .
the smc distance modulus is @xmath66 = @xmath67 mag ( 60 kpc ) ( e.g. , * ? ? ?
* ) , but due to the large depth extension of the smc along the line of sight we adjusted the distance modulus @xmath66 to produce the best isochrone fits to our cmd data . for easier comparison to the isochrones ,
we first derived a fiducial ridgeline ( table [ tab : ridgeline ] ) , which reproduces the mean location of the stellar distribution in the cmd ( exempting the hb ) . in order to determine the ridgeline
, we separated the cluster center cmd into three sections : the ms , the sgb and the rgb . on the ms we determined the mode of the color distribution in magnitude bins of 0.3 mag width . for the sgb
, we performed a linear least squares fit of a polynomial of 5th order to a hess diagram of this region in the cmd .
finally , the rgb was fit by a third - order polynomial of the mean color , again in magnitude bins with a size of 0.3 mag each .
the resulting ridgeline is shown in fig .
[ fig : ngc121_lettercenter ] as a cyan line .
we fitted our @xmath50 vs. @xmath50-@xmath51 cmd with three different isochrone models : padova isochrones ( girardi , `` private communication '' , girardi et al .
2000 ) , ] teramo isochrones @xcite , both with scaled solar isochrones ( [ @xmath4/fe ] = 0.0 ) , and dartmouth isochrones @xcite with both [ @xmath4/fe ] = 0.0 and + 0.2 .
the padova isochrone grid has an age resolution of log(t)=0.05 , the teramo isochrone grid of 0.1 myr and the dartmouth isochrone grid of 0.5 gyr .
our adopted spectroscopic metallicity of [ fe / h ] = @xmath68 corresponds most closely to z = 0.0004 in the padova models , to z = 0.0006 in the teramo models , and to [ fe / h ] = @xmath69 in the dartmouth models .
all three sets of isochrone models are available in the standard acs color system .
we fitted a large number of isochrones using different combinations of reddening , age , and distance .
for each set of models , we selected the isochrone that best matched the observed data ( fig .
[ fig : ngc121_letteriso ] , fig .
[ fig : ngc121_dart ] ) .
first we discuss figure [ fig : ngc121_letteriso ] .
our best - fit age using padova isochrones is 11.2 gyr with @xmath66 = 19.05 mag and @xmath70 = 0.05 .
the best fitting teramo isochrone yields an age of t = 11.8 gyr , @xmath66 = 19.02 mag , and @xmath70 = 0.04 . on the ms , both the teramo isochrone and the padova isochrones trace the ridge line almost perfectly . at the faint end of the ms , the padova isochrone continues further to the blue than the teramo isochrone and our derived ridge line ; however , this only becomes more apparent at magnitudes of @xmath50 @xmath7 25.5 mag and below .
both isochones also provide an excellent approximation to the sgb and to the lower rgb up to about half a magnitude below the hb . at brighter magnitudes ,
the two isochrones deviate increasingly to the blue of the observed upper rgb . here
the padova isochrone shows the strongest difference , deviating by approximately 0.38 mag in color from the observed tip of the rgb .
the isochrone shows a magnitude for the base of the red hb that is about 0.5 mag fainter than the observed one .
unlike teramo and dartmouth , the padova isochrone also models the agb and its tip , which is @xmath7 1 mag brighter than the tip of the rgb .
the teramo isochrone is too blue by about 0.23 mag at the magnitude of the tip of the rgb and indicates a magnitude for the base of the red hb that is 0.2 mag too bright .
if we had no prior knowledge of the metallicity of ngc121 and were to use the upper rgb as a metallicity indicator , a better fit would be obtained by choosing isochrones of a different metallicity or @xmath4 abundance .
the problems of various isochrone models of given metallicities in reproducing the upper red giant branches of globular clusters with the same metallicities are a well - known problem ( e.g. , * ? ? ?
* ; * ? ? ?
our figure [ fig : ngc121_letteriso ] reflect the general failure of the chosen stellar evolutionary models to simultaneously reproduce the major features of cmds @xcite in spite of the excellent fit to the lower rgb , sgb , and ms .
fortunately the latter are the most age - sensitive features of the cmd .
the isochrone model provided by @xcite with [ @xmath4/fe ] = + 0.2 yield the best fit to the cmd ( fig .
[ fig : ngc121_dart ] ) .
the best - fit isochrone has the parameters _ t _ = 10.5 gyr , @xmath66 = 18.96 mag , and @xmath70 = 0.035 , using the @xmath4-enhanced isochrones of [ @xmath4/fe ] = + 0.2 .
_ all _ the major features of the cmd are very well reproduced , including the upper rgb where the isochrone is offset slightly to the blue relative to the fiducial ridgeline .
this offset is no more than 0.01 to 0.02 on average along the entire upper rgb ; i.e. , even the _ slope _ of the rgb is very well reproduced along its entire extent .
unfortunately the stellar evolution models used here terminate at the he flash , and therefore do not fit the hb or the agb .
is our use of @xmath4-enhanced models justified ?
for ngc121 a value of @xmath71 } = + 0.24 $ ] has been measured , which is similar to the outer lmc cluster hodge 11 and to the old galactic outer halo clusters with @xmath71 } = + 0.3 $ ] @xcite .
consequently , we assume that ngc121 also is enhanced in @xmath4-elements .
we note that in this respect ngc121 differs from the general trend observed in red giant stars in the lmc and in dwarf spheroidal galaxies , where the [ @xmath4/fe ] ratios at a given [ fe / h ] tend to be lower by up to a few tenths of a dex than in the galactic halo ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
when we adopt the values for distance and reddening , but fit the cluster with an isochrone scaled solar , ngc121 gets a slightly older age of 11 gyr .
the isochrone model with [ @xmath4/fe ] = 0.0 still provides a better fit than the teramo or padova models , but has an offset of @xmath7 0.05 on average along the upper rgb .
past studies found that @xmath4-enhanced models imply a higher luminosity and temperature for the same mass than the solar scaled models and therefore an older age for the same magnitude ( e.g. , * ? ? ? * ) .
the dartmouth models show exactly the opposite behavior .
this is because in these models an increase in [ @xmath4/fe ] is accompanied by a corresponding increase of the total metallicity z , which makes the isochrones cooler at constant age and [ fe / h ] .
finally , we note that the derived reddenings agree with the extinction @xmath72 = @xmath73 from the @xcite maps .
the reddening law of @xcite is assumed . in the figures [ fig : ngc121_isocheck ] and [ fig : ngc121_isocheck_dart ]
we show a range of isochrones for the three sets of stellar evolution models in order to illustrate the age uncertainty in a given model .
the finally chosen , `` best '' isochrone is always displayed along with two younger and two older isochrones .
for the cases of the teramo and padova models , the two isochrones that are one age step younger or older than the chosen , central isochrone provide an upper or lower envelope for the location of the high - density part of the sgb , the ms turn - off , and the base of the rgb . for
the dartmouth isochrones the outermost isochrones provide this envelope .
considering the high quality of the fit of the central isochrone in this cmd region in all models and the larger deviations of the adjacent isochrones , we estimate that the resultant age uncertainty is of the order of approximately @xmath74 gyr for the teramo and dartmouth isochrones and @xmath75 gyr for the padova isochrones .
+ cccc 1.7400 & 16.6859 & 1.0000 & 19.9538 + 1.7300 & 16.6916 & 0.9900 & 20.0855 + 1.7200 & 16.7011 & 0.9800 & 20.2218 + 1.7100 & 16.7107 & 0.9700 & 20.3630 + 1.7000 & 16.7224 & 0.9600 & 20.5090 + 1.6900 & 16.7339 & 0.9500 & 20.6601 + 1.6800 & 16.7483 & 0.9400 & 20.8162 + 1.6700 & 16.7605 & 0.9300 & 20.9774 + 1.6600 & 16.7834 & 0.9200 & 21.1440 + 1.6500 & 16.8001 & 0.9100 & 21.3159 + 1.6400 & 16.8177 & 0.9000 & 21.4932 + 1.6300 & 16.8381 & 0.8900 & 21.6762 + 1.6200 & 16.8554 & 0.8800 & 21.8899 + 1.6100 & 16.8706 & 0.8700 & 22.0599 + 1.6000 & 16.8867 & 0.8600 & 22.1921 + 1.5900 & 16.9098 & 0.8600 & 22.1933 + 1.5800 & 16.9289 & 0.8500 & 22.2305 + 1.5700 & 16.9540 & 0.8400 & 22.2520 + 1.5600 & 16.9712 & 0.8300 & 22.2660 + 1.5500 & 16.9874 & 0.8200 & 22.2770 + 1.5400 & 17.0008 & 0.8100 & 22.2890 + 1.5300 & 17.0153 & 0.8000 & 22.3050 + 1.5200 & 17.0309 & 0.7900 & 22.3180 + 1.5100 & 17.0477 & 0.7800 & 22.3280 + 1.5000 & 17.0658 & 0.7700 & 22.3380 + 1.4900 & 17.0851 & 0.7600 & 22.3480 + 1.4800 & 17.1057 & 0.7500 & 22.3570 + 1.4700 & 17.1276 & 0.7400 & 22.3670 + 1.4600 & 17.1508 & 0.7300 & 22.3770 + 1.4500 & 17.1753 & 0.7200 & 22.3880 + 1.4400 & 17.2013 & 0.7100 & 22.4010 + 1.4300 & 17.2286 & 0.7000 & 22.4120 + 1.4200 & 17.2575 & 0.6900 & 22.4260 + 1.4100 & 17.2877 & 0.6800 & 22.4520 + 1.4000 & 17.3195 & 0.6700 & 22.4710 + 1.3900 & 17.3528 & 0.6600 & 22.5020 + 1.3800 & 17.3877 & 0.6500 & 22.5190 + 1.3700 & 17.4241 & 0.6400 & 22.5410 + 1.3600 & 17.4621 & 0.6300 & 22.5720 + 1.3500 & 17.5018 & 0.6200 & 22.6240 + 1.3400 & 17.5432 & 0.6100 & 22.6890 + 1.3300 & 17.5962 & 0.6000 & 22.7610 + 1.3200 & 17.6378 & 0.5950 & 22.8500 + 1.3100 & 17.6904 & 0.5950 & 23.0310 + 1.3000 & 17.7246 & 0.6125 & 23.4000 + 1.2900 & 17.7605 & 0.6670 & 24.0000 + 1.2800 & 17.7983 & 0.7400 & 24.6000 + 1.2700 & 17.8379 & 0.8400 & 25.2000 + 1.2600 & 17.8794 & 0.9700 & 25.8000 + 1.2500 & 17.9231 & 1.1000 & 26.4000 + 1.2400 & 17.9690 & & + 1.2300 & 18.0171 & & + 1.2200 & 18.0676 & & + 1.2100 & 18.1205 & & + 1.2000 & 18.1761 & & + 1.1900 & 18.2343 & & + 1.1800 & 18.2952 & & + 1.1700 & 18.3590 & & + 1.1600 & 18.4257 & & + 1.1500 & 18.4955 & & + 1.1400 & 18.5685 & & + 1.1300 & 18.6447 & & + 1.1200 & 18.7242 & & + 1.1100 & 18.8071 & & + 1.1000 & 18.8936 & & + 1.0900 & 18.9837 & & + 1.0800 & 19.0775 & & + 1.0700 & 19.1752 & & + 1.0600 & 19.2767 & & + 1.0500 & 19.3823 & & + 1.0400 & 19.4920 & & + 1.0300 & 19.6059 & & + 1.0200 & 19.7241 & & + 1.0100 & 19.8367 & & + [ tab : ridgeline ] to check the reliability of the isochrone ages , we use a reddening - independent method to derive relative ages of ngc121 .
this method is also independent of the photometric zeropoint of our data .
this vertical method
relies on the fact that the absolute magnitude of the msto depends on the age of the cluster ( e.g. , * ? ? ?
* ) , while the absolute magnitude of the hb remains approximately age - independent for clusters older than @xmath76 10 gyr ( e.g. , * ? ? ?
we measure the apparent magnitudes of the msto and hb , i.e. , @xmath77 and @xmath62 , in order to obtain the magnitude difference @xmath56 . with increasing age , a cluster has generally larger values of this parameter since the ms moves to fainter magnitudes .
unfortunately , the determination of these two points comes with significant uncertainties .
@xmath78 is hard to measure accurately both because of the width of the hb in luminosity and the msto s vertical extent in the turn - off region .
first , we calculate the magnitude difference @xmath79 between the hb and the msto , as originally described by @xcite . because @xmath50 is proportional to v and we are only interested in the magnitude difference , which we measure at constant color
, there is no need to transform the magnitudes from the acs system to _ v _ and _ i _ magnitudes @xcite .
we follow the general definition of the msto as the bluest point along the ms ; in our case represented by the bluest point on the ridgeline .
we find the msto at @xmath80 = @xmath81 mag , @xmath82 = @xmath83 mag . as described in section [ sec : cmd ] , the mean hb magnitude is @xmath84 = @xmath85 mag .
this yields a @xmath86 mag .
our result is only slightly lower than former values , e.g. , @xmath56 = 3.32 determined by @xcite , 3.33 by @xcite or 3.29 by @xcite .
@xcite published a mean value of @xmath87 mag for old galactic halo globular clusters ( gc ) .
the @xmath79 of ngc121 is 0.28 mag smaller than this value , which indicates that ngc121 is younger than most of the older _ galactic _ gcs .
@xcite ( see also buonanno et al .
1989 ) found a relation between age and @xmath56 based on a study of 41 galactic globular clusters .
adopting again a metallicity of [ fe / h ] = @xmath64 , we find with the formula @xmath88 } + 0.37 \delta v^{hb}_{to } - 0.24 $ ] an age of @xmath89 gyr for ngc121 . in order to compare our data directly with @xcite , we use the age calibration provided by @xcite as was done by these authors .
for this purpose we must adopt the metallicity fe / h ] = @xmath91 of @xcite on the metallicity scale introduced by @xcite ( cg97 ) to be consistent with @xcite s calculation .
note that elsewhere in the paper we are using metallicities on the zw84 scale which agrees reasonably well with the spectroscopically derived chemical abundances @xcite . using this method and the higher metallicity we determine an age of @xmath92 gyr for ngc121 , which is similar to the age that @xcite found .
we have to emphasize that this age is younger than the absolute ages determined in section [ sec : isochrone ] due to the different metallicity scale .
if we use our preferred zw84 metallicity scale with this method we obtain an older age of @xmath93 gyr . while the measurement of @xmath77 is affected by significant observational errors ( @xmath7 0.05 mag ) ,
the color of the msto is well - defined .
@xcite found that the usage of a point on the sbg brighter than the msto and 0.05 mag redder ( @xmath94 ) provides more precise relative ages than @xmath77 .
@xcite provide a conversion between @xmath95 and @xmath96 , @xmath97 data for a grid of five metallicities .
we choose the conversion for [ fe / h ] @xmath98 because it is closest to the metallicity of ngc121 . in our data , we measured @xmath99 = @xmath100 at @xmath101 = @xmath102 mag . to convert @xmath99 to the absolute magnitude @xmath103 we use
the distance modulus derived above .
this yields @xmath104 mag . with the modified calibration by @xcite
we obtain an age of @xmath105 gyr for ngc121 .
we have summarized all our age results in table [ tab : ages ] .
calculating @xmath106 = @xmath107 we find @xmath108 mag .
there is a general trend of increasing rgb bump brightness with decreasing age @xcite , assuming that a different second parameter is not affecting the position of the rgb bump .
@xcite presented a @xmath109 vs. [ fe / h ] diagram ( their fig.6 ) , where the brightness difference between the rgb bump and the hb is plotted against the cluster metallicity .
if we assume that the hb magnitude does not critically depend on age , then ngc121 is slightly older than 10 gyr based on this relation .
these comparative results are sensitive to the abundance of @xmath4-elements in ngc121 as compared to galactic globular clusters .
many nearby galactic globular clusters are enhanced in @xmath4-elements relative to the solar value @xcite . as mentioned earlier , spectroscopic results for ngc121
indicate that this cluster is similarly enhanced in @xmath4 elements , which means that the relative ages should not be affected as long as we confine ourselves to the comparison of globular clusters with similar @xmath4-element ratios .
the fact that ngc121 does not follow the trend of reduced [ @xmath4/fe ] ratios observed in other nearby dwarf galaxies facilitates both our relative age determinations .
ccc @xmath5 & @xmath56 & @xcite ( zw ) + @xmath6 & @xmath95 & @xcite + @xmath93 & @xmath90 & @xcite + @xmath1 & isochrones & @xcite + @xmath110 & isochrones & girardi , `` private communication '' , + & & @xcite + @xmath3 & isochrones & @xcite + [ tab : ages ] in table [ tab : comp_ages_zw ] we compare the relative age of ngc121 with those for a sample of galactic globular clusters .
while this comparison sample is located in the galactic halo , some objects may have formed outside the galaxy and might have been subsequently captured or accreted .
we list the clusters by their identification in column ( 1 ) .
the [ fe / h ] values are given in column ( 2 ) in the scale by @xcite .
column ( 3 ) shows the @xmath111 and column ( 4 ) the ages obtained by using the @xcite calibration .
finally , column ( 5 ) gives the relative age difference of these clusters to ngc121 @xmath112 .
for @xmath111 we adopted the values from @xcite , unless differently stated ( see footnotes of table [ tab : comp_ages_zw ] ) .
the clusters are listed in order of increasing @xmath111 and are divided into two groups .
the first group shows nine pure gcs with similar metallicities as ngc121 . among these nine clusters
is ngc2808 for which multiple mstos have been found @xcite .
its @xmath111 value , derived prior to the study by @xcite , is comparatively small . even though those nine clusters have all similar metallicities
, the spread in @xmath111 and therefore in age is quite large : ngc1262 shows the lowest @xmath111 = 3.24 and is similar in age to ngc121 , while ngc6656 has @xmath111 = 3.55 , which makes it @xmath73 gyr older than ngc121 .
the second group includes a subset of galactic halo clusters that appear to be significantly younger than the average of the galactic globular cluster population ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
some members of this group are listed in the last part of table [ tab : comp_ages_zw ] .
ngc362 and ngc288 are known to be a second parameter cluster pair of different ages , as reflected in their different @xmath111 ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
as ngc362 has a similar @xmath113 and [ fe / h ] as ngc121 , it should therefore be of a similarly young age . other members of this group of young halo globulars are ic 4499 @xcite , ruprecht 106 ( assumed to be 35 gyr younger than the bulk of the galactic globulars with similar metallicities ; * ? ? ?
* ; * ? ? ?
* ) , arp 2 @xcite , terzan 7 @xcite , and pal 14 @xcite .
ccccc ngc121 & -1.46 & 3.27 & 10.9 & 0 + ngc1261 & -1.32 & 3.24 & 10.4 & -0.5 + ngc5272 & -1.66 & 3.24 & 10.8 & 0.1 + ngc2808 & -1.36 & 3.25 & 10.6 & -0.3 + ngc3201 & -1.53 & 3.28 & 11.0 & 0.1 + ngc5904 & -1.38 & 3.34 & 11.4 & 0.5 + ngc6254 & -1.55 & 3.37 & 11.9 & 1.0 + ngc6218 & -1.40 & 3.48 & 12.9 & 2.0 + ngc6752 & -1.54 & 3.53 & 13.7 & 2.8 + ngc6656 & -1.41 & 3.55 & 13.7 & 2.8 + pal 12 & -0.94 & 3.17 & 9.45 & -1.3 + ig 4499 & -1.75 & 3.25 & 11 & 0.1 + ngc362 & -1.33 & 3.27 & 10.7 & -0.2 + pal 14 & -1.65 & 3.33 & 11.7 & 0.8 + ngc288 & -1.40 & 3.45 & 12.6 & 1.7 + [ tab : comp_ages_zw ] among the theories that try to explain the existence of these young objects is the model according to which they are intergalactic clusters captured by the mw @xcite , or clusters formed during interactions between the mw and the magellanic clouds assuming that they are on bound orbits @xcite .
@xcite argued that the apparent young halo globular clusters formed in dwarf galaxies that later merged with the mw .
hence the galactic globular clusters are assumed to be a mixture of objects that formed with the mw itself ( old halo group ) and others accreted from destroyed dwarf satellites ( young halo clusters ) ( see also mackey & gilmore 2004 ) .
at least six globular clusters are believed to be associated with the sagittarius dwarf galaxy ( e.g. , * ? ? ?
* and references therein ) , providing support for the accretion scenario .
note , however , that sagittarius is contributing both old ( m54 , ter 8 , arp 2 ) and `` young '' ( ter 7 , pal 12 , whiting 1 ) globular clusters to the mw .
similarly , the only other galactic dsph galaxy known to contain globular clusters , fornax , would contribute both kinds of globulars @xcite if it were to merge with our galaxy .
this also holds for the lmc ( * ? ? ?
* see also discussion in grebel , gallagher , & harbeck 2003 ) .
if we take all the ages determined in this paper into account , we find that ngc121 is consistently 23 gyr younger than the oldest galactic globular clusters and lmc globular clusters .
the age offset remains when comparing ngc121 to old galactic globular clusters in the same metallicity range ( see tab .
[ tab : comp_ages_zw ] , upper panel ) .
we also show that ngc121 is not as young as the youngest galactic and sagittarius globular clusters , some of which are @xmath7 2 gyr younger than ngc121 .
we derived ages for the old smc globular cluster ngc121 based on our high dynamic range hst / acs photometry that extends at least three magnitudes below its msto . in order to obtain absolute ages , we applied three different isochrone models .
these isochrone models yielded ages of @xmath110 gyr ( padova ) , @xmath114 gyr ( teramo ) , and @xmath3 gyr ( dartmouth ) .
we find the @xmath4-enhanced dartmouth isochrones provide the closest approximation to the ms , sgb , and rgb , whereas the other models can not reproduce the slope of the upper rgb .
high - resolution spectroscopy indicates that ngc121 is indeed @xmath4-enhanced @xcite , a property that it shares with many of the old outer galactic halo globulars .
given the proximity of ngc121 to the smc on the sky and its distance , its physical association with the smc seems well - established .
our determinations of relative ages for ngc121 are consistent with the results of our absolute age determination .
relative age estimates , when converted to an absolute age scale , are @xmath5 gyr ( @xmath111 ) , @xmath93 ( @xmath90 ) and @xmath6 gyr ( @xmath115 ) .
these numbers agree well with the absolute age derivations .
our results confirm that ngc121 is 23 gyr younger than the oldest mw and lmc clusters ( as also found in earlier wfpc2 studies ) .
ngc121 is similar in age to the youngest globular cluster in the fornax dsph @xcite , and to several of the young galactic halo clusters . on the other hand , ngc121 is not as young as some of the sgr dwarf galaxy s globular clusters or the youngest galactic globular clusters .
it is intriguing that the smc in contrast to other galactic companion dwarf galaxies with globulars does not contain any old classical globular clusters . but given the existence of only one cluster and the question of star cluster survival , this could be a result of the one survivor from the smc s epoch of globular cluster formation randomly sampling an initial distribution of star cluster ages . on the other hand , in low - mass galaxies without bulges , spiral density waves , and
shear it is much more difficult to destroy globular clusters through external effects . that this cluster is both younger than the galactic mean and enhanced in @xmath4-elements
may have interesting implications for the early development of the smc .
it also is intriguing that the only globular cluster in the smc is not very metal - poor .
the smc must have experienced substantial enrichment prior to the formation of ngc121 . in the lmc , where two main epochs of the formation of populous compact star clusters have been found ( e.g. , * ? ? ?
* ) , a few globular clusters are found that are old enough to exhibit blue hbs .
interestingly , these globular clusters , which are similarly old as the oldest galactic globulars @xcite , have a similar metallicity to ngc 121 @xcite ( e.g. , ngc1898 , ngc2019 ) , indicating very early chemical enrichment .
the mw also contains old classical globular clusters ( with blue hbs ) that have similarly high metallicities as the somewhat younger ngc 121 . evidently , the conditions for and the efficiency of star formation varied in these three galaxies at early epochs .
after ngc 121 formed there was a hiatus in surviving stars clusters and thus possibly in cluster formation activity in the smc : the second oldest smc cluster is lindsay 1 with an age of @xmath7 8 gyr ( glatt et al . 2007 , in preparation ) . since then compact populous star clusters formed fairly continuously until the present day in the smc ( e.g. , da costa 2002 ) in contrast to both the lmc and the mw . in forthcoming papers on our acs photometry of smc clusters and field populations we will explore the evolutionary history of the smc in more detail .
clearly , clues about the early star formation history of the smc will have to come from its old field populations .
we thank the anonymous referee for extremely useful suggestions to improve our paper .
we gratefully acknowledge support by the swiss national science foundation through grant number 200020 - 105260 and 200020 - 113697 .
support for program go-10396 was provided by nasa through a grant from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas 5 - 26555 .
we warmly thank paolo montegriffo to provide his software and leo girardi for the padova isochrones in the standard acs color system .
gisella clementini and monica tosi have been partially supported by prin - miur-2004 and prin - inaf-2005 , and jay gallagher also obtained helpful additional support from the university of wisconsin graduate school . , d. r. , & sarajedini , a. 1999 , , 511 , 225 , m. , fusi pecci , f. , ferraro , f. r. , galleti , s. , catelan , m. , & landsman , w. b. 2001 , , 122 , 2569 , g. , mateo , m. , chiosi , c. , & bressan , a. 1992 , , 388 , 400 , e. , & arnouts , s. 1996 , , 117 , 393 , t. m. , et al .
2007 , , 658 , l95 , r. , corsi , c. e. & fusi pecci , f. 1989 , , 216 , 80 , r. , corsi , c. e. , fusi pecci , f. , richer , h. b. , & fahlman , g .
g. 1993 , , 105 , 184 , r. , corsi , c. e. , pulone , l. , fusi pecci , f. , richer , h. b. , & fahlman , g .
g. 1995b , , 109 , 663 , r. , corsi , c. e. , fusi pecci , f. , richer , h. b. , & fahlman , g .
g. 1995a , , 109 , 650 , r. , corsi , c. , bellazzini , m. , ferraro , f. r. , & fusi pecci , f. 1997 , , 113 , 706 , r. , corsi , c. e. , zinn , r. , fusi pecci , f. , hardy , e. , & suntzeff , n. b. 1998 , , 501 , l33 , r. , corsi , c. e. , castellani , m. , marconi , g. , fusi pecci , f. , & zinn , r. 1999 , , 118 , 1671 , g. , zinn , r. , & moni bidin , c. 2007 , , 466 , 181 , e. , & gratton , r. g. 1997 , , 121 , 95 , m. , ferraro , f. r. , & rood , r. t. 2001a , , 560 , 970 , m. , bellazzini , m. , landsman , w. b. , ferraro , f. r. , fusi pecci , f. , & galleti , s. 2001b , , 122 , 3171 , m. 2005 , arxiv astrophysics e - prints , arxiv : astro - ph/0507464 , b. , demarque , p. , kernan , p. j. & krauss , l. m. 1996 , science , 271 , 957 , b. , demarque , p. , kernan , p. j. , krauss , l. m. , & sarajedini , a. 1996a , , 283 , 683 , b. , demarque , p. , & sarajedini , a. 1996b , , 459 , 558 , g. , held , e. v. , baldacci , l. , & rizzi , l. 2003 , , 588 , l85 , a. a. , et al .
2007 , , 659 , l17 , g. s. , armandroff , t. e. & norris , j. e. 1992 , , 104 , 154 , g. s. & hatzidimitriou , d. 1998 , , 115 , 1934 , f. , piotto , g. , cassisi , s. , busso , g. , recio - blanco , a. , salaris , m. , aparicio , a. , & rosenberg , a. 2005 , , 130 , 116 , a. e. , walker , a. r. , hodge , p. w. , mateo , m. , olszewski , e. w. , schommer , r. a. , & suntzeff , n. b. 2001 , , 562 , 303 , a. , chaboyer , b. , jevremovi , d. , baron , e. , ferguson , j. w. , sarajedini , a. , & anderson , j. 2007 , , 134 , 376 , i. , ferraro , f. r. , fusi pecci , f. , corsi , c. e. & buonanno , r. 1995 , , 275 , 1057 , j. p. 2002
, , 123 , 404 , f. , bellazzini , m. , cacciari , c. & ferraro f. r. 1995 , , 110 , 1664 , f. , bellazzini , m. , ferraro , f. r. , buonanno , r. , & corsi , c. e. 1996 , formation of the galactic halo ... inside and out , asp conf .
morrison , h. l. & sarajedini , a. , 221 , c. , zoccali , m. , & aparicio , a. 2005 , , 43 , 387 , l. , & salaris , m. 2001 , , 323 , 109 , l. , bressan , a. , bertelli , g. , & chiosi , c. 2000 , , 141 , 371 , j. a. 1975 , , 87 , 641 , e. k. 1997 , reviews in modern astronomy , 10 , 27 grebel , e. k. 1999 , the stellar content of the local group , iau symp . 192 , eds .
p. whitelock & r. cannon ( san francisco : asp ) , 17 , e. k. 2001 , astrophysics and space science supplement , 277 , 231 , e. k. , & gallagher , j. s. , iii 2004 , , 610 , l89 , d. , et al . 2001 , , 122 , 3092 , w. e. 1996 , , 112 , 1487 , e. v. , clementini , g. , rizzi , l. , momany , y. , saviane , i. , & di fabrizio , l. 2001 , , 562 , l39 , v. , franois , p. , spite , m. , primas , f. , & spite , f. 2000 , , 364 , l19 , i. j. , & faulkner , j. 1968 , , 153 , 101 , j. a. bolte , m. , stetson , p. b. , hesser , j. e. , & somerville , r. s. 1999 , , 527 , 199 , j. a. , bolte , m. , hesser , j. e. , ivans , i. i. , & stetson , p. b. 2004 , abundances in lmc and smc globular clusters , eds .
a. mcwilliam & m. rauch ( pasadena : carnegie observatories ) , j. a. , ivans , i. i. , & stetson , p. b. 2006 , , 640 , 801 , a. , grebel , e. k. , gilmore , g. f. , wyse , r. f. g. , kleyna , j. t. , harbeck , d. , wilkinson , m. i. , & evans , n. w. 2007 , , submitted , a. m. , fruchter , a. s. , hook , r. n. , & hack , w. 2002 , in _
hubble _ after the installation of the acs and the nicmos cooling system , eds . s. arribas , a. koekemoer & b. whitmore ( baltimore : stsci ) , 337 , l. m. , & chaboyer , b. 2003 , science , 299 , 65 , y .- w . , demarque , p. , & zinn , r. 1994 , , 423 , 248 , a. d. & gilmore , g. f. 2004 , , 352 , 153 makarova , l. n. , et al .
2002 , , 396 , 473 , k. j. , sarajedini , a. , & frenck , r. s. 1998 , , 116 , 2395 , b. , diemand , j. , madau , p. , zemp , m. , & stadel , j. 2006 , , 368 , 563 , j. e. 1994 , , 422 , 158 , k. a. g. , hodge , p. w. , mateo , m. , olszewski , e. w. , schommer , r. a. , suntzeff , n. b. , & walker , a. r. 1998 , , 300 , 665 , e. w. , suntzeff , n. b. & mateo , m. 1996 , , 34 , 511 , a. , cassisi , s. , salaris , m. , & castelli , f. 2004 , , 612 , 168 , g. , bedin , l. r. , anderson , a. , king , i. r. , cassisi , s. , milone , a. p. , villanova , s. , pietrinferni , a. & renzini , a. 2007 , , 661 , l53 , b. j. , armandroff , t. e. , jacoby , g. h. , & da costa , g. s. 2005 , , 129 , 2232 , b. j. , venn , k. a. , & irwin , m. i. 2005 , , 130 , 2140 , k. u. , & bahcall , j. n. 1985 , , 59 , 63 , a. , saviane , i. , piotto , g. , & aparicio , a. 1999 , , 118 , 2306 , g. a. , hesser , j. e. , stetson , p. b. , mateo , m. , simard , l. , bolte , m. , friel , e. d. , & copin , y. 1997 , , 109 , 883 , m. , & girardi , l. 2002 , , 337 , 332 , m. , & weiss , a. 2002 , , 388 , 492 , a. 1997 , , 113 , 682 , a. , geisler , d. , harding , p. , & schommer , r. 1998 , , 508 , l37 , a. , et al .
2002 , , 567 , 915 , d. j. , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 , m. m. , fall , s. m. , rich , r. m. , & zurek , d. 1998 , , 508 , 570 , m. d. , ct , p. , & sargent , w. l. w. 2001 , , 548 , 592 , m. , et al .
2005 , , 117 , 1049 , j. , carney , b. w. , gieren , w. p. , fouqu , p. , latham , d. w. , & fry , a. m. 2004 , , 415 , 531 , l. l. , da costa , g. s. & mould , j. r. 1985 , , 298 , 544 , a. d. 1958 , , 118 , 117 , a. 1998 , acta astronomica , 48 , 383 , d. a. 2000 , , 129 , 315 , d. a. , swenson , f. j. , rogers , f. j. , iglesias , c. a. , & alexander , d. r. 2000 , , 532 , 430 , a. r. 1989 , , 101 , 570 , a. r. 1991 , the magellanic clouds , iau symp . 148 , eds .
r. haynes & d. milne ( los angeles : asp ) , 307 , a. r. 1992 , , 390 , l81 , r. , & west , m. j. 1984 , , 55 , 45 zinn , r. 1993 , the globular cluster - galaxy connection , asp conf .
smith , g. h. & brodie , j. p. ( san francisco : asp ) , 302 | as first paper of a series devoted to study the old stellar population in clusters and fields in the small magellanic cloud , we present deep observations of ngc121 in the f555w and f814w filters , obtained with the advanced camera for surveys on the _ hubble space telescope_. the resulting color - magnitude diagram reaches @xmath0 mag below the main - sequence turn - off ; deeper than any previous data .
we derive the age of ngc121 using both absolute and relative age - dating methods .
fitting isochrones in the acs photometric system to the observed ridge line of ngc121 , gives ages of @xmath1 gyr ( teramo ) , @xmath2 gyr ( padova ) and @xmath3 gyr ( dartmouth ) .
the cluster ridge line is best approximated by the @xmath4-enhanced dartmouth isochrones . placing our relative ages on an absolute age scale , we find ages of @xmath5 gyr ( from the magnitude difference between the main - sequence turn - off and the horizontal branch ) and @xmath6 gyr ( from the absolute magnitude of the horizontal branch ) , respectively .
these five different age determinations are all lower by 2 3 gyr than the ages of the oldest galactic globular clusters of comparable metallicity .
therefore we confirm the earlier finding that the oldest globular cluster in the small magellanic cloud , ngc121 , is a few gyr younger than its oldest counterparts in the milky way and in other nearby dwarf galaxies such as the large magellanic cloud , fornax , and sagittarius .
if it were accreted into the galactic halo , ngc121 would resemble the `` young halo globulars '' , although it is not as young as the youngest globular clusters associated with the sagittarius dwarf .
the young age of ngc121 could result from delayed cluster formation in the small magellanic cloud or result from the random survival of only one example of an initially small number star clusters . |
recent spectroscopic studies have demonstrated the existence of numerous carbon - enhanced , metal - poor ( cemp ) stars that exhibit strong enhancements of their neutron - capture elements , in particular at metallicities [ fe / h]@xmath6 .
this is believed to be due to mass transfer in binary systems from former asymptotic giant - branch ( agb ) companions that underwent @xmath1-process nucleosynthesis during their lifetimes @xcite . however , a significant fraction of these stars appear to exhibit large deviations from the scaled solar @xmath1-process distribution of elemental abundance ratios , especially with regard to their enhanced eu ( e.g. , * ? ? ?
this observed discrepancy in eu for some cemp stars prompted @xcite , @xcite , @xcite , and @xcite to suggest that the large excess of eu in these stars might be due to an additional contribution from the @xmath2-process .
@xcite , in their suggested taxonomy of cemp stars , refer to this class as cemp-@xmath7 .
@xcite proposed an accretion - induced collapse ( aic ) of a white dwarf into a neutron star in a binary system as one astrophysical scenario for enrichment of the surviving companion with both @xmath2-process and @xmath1-process elements .
as an alternative , @xcite suggested that these double enhancements might be due to the explosions of degenerate cores in agb stars .
previously , @xcite and @xcite suggested massive agb stars ( @xmath8 ) , is likely to be too high for stars that undergo agb evolution .
the low metallicity of the cemp-@xmath7 stars pushes the appropriate mass range to even lower values ( @xmath9 , e.g. , * ? ? ?
] to be the origin of these double enhancements .
in this paper we explore the astrophysical scenario suggested by @xcite and @xcite . specifically , we consider the available constraints on a model in which @xmath3 stars in wide binary systems may be invoked to explain the double enhancements of @xmath2-process and @xmath1-process elements that results in the creation of cemp-@xmath7 stars . in
2 the observed abundances of cemp-@xmath7 stars are compared to the abundances from a low - metallicity agb model @xcite , which implies additional contributions of @xmath2-processed material to these stars . in 3 , the wind pollution model is examined to explain the double enhancements with @xmath2-process and @xmath1-process elements in these stars . the efficiency of wind pollution by a supernova required to be consistent with the galactic chemical evolution of @xmath2-process elements is then discussed . a brief summary of our conclusions and a discussion of future areas for theoretical and observational investigation of the cemp-@xmath7 phenomenon is presented in 4 .
table 1 lists seven cemp-@xmath7 stars with large enhancements of neutron - capture elements reported in the recent literature @xcite , all of which were further selected to have the lowest observed [ ba / eu ] = \log ( n_{\rm a}/n_{\rm b } ) - \log ( n_{\rm a}/n_{\rm b})_\odot$ ] ] ratios ( @xmath10 ) .
note that the highly @xmath2-process - enhanced star cs 22892 - 052 @xcite , which marginally qualifies as a cemp star , is added on the last line of table 1 for comparison purposes . with
the exception of this star , which exhibits a nearly solar @xmath2-process ratio of [ ba / eu ] , all of the cemp-@xmath7 stars listed in table 1 lie between the solar @xmath2-process and @xmath1-process values of [ ba / eu ] ( @xmath11 and + 1.15 , respectively , * ? ? ? * ) .
these stars have similar metallicities ( @xmath12 \le - 2.3 $ ] ) and excesses in their c and ba abundances ( @xmath13 \sim [ { \rm ba}/{\rm fe } ] \sim + 2 $ ] ) .
these stars also exhibit very high pb abundances ( @xmath14 \sim + 3 $ ] ) , and hence belong to the class of `` lead stars '' @xcite .
the over - production of pb is believed to be a consequence of the operation of an @xmath1-process with a high neutron - to - seed ratio in an agb star , owing to its low metallicity @xcite .
the similarity of the abundance patterns amongst these stars implies that all the cemp-@xmath7 stars may have been formed in similar astrophysical environments .
figure 1 compares the observed abundances of two representative stars from table 1 , he 2148 - 1247 and cs 29497 - 030 , to the abundances of elements predicted to be found in the dredge - up material of a low - metallicity agb model taken from @xcite , as shown by the thin - solid line .
the metallicity of this model , @xmath15 = - 1.3 $ ] , is clearly not as low as to be completely relevant for the stars presented here .
nevertheless , we employ this comparison , since a zero - metal agb model by @xcite appears to show a similar abundance trend .
the @xcite result predicts a slightly lower [ ba / eu ] ratio ( @xmath16 ) than the solar @xmath1-process value ( @xmath17 ) , but still significantly higher than applies to the measured stellar abundances ( @xmath18 ) listed in table 1 . as can be seen in figure 1 , the abundance curve from the low - metallicity agb model ( scaled to match the ba abundances )
does not appear to account for any of the stellar abundances of eu , gd , and dy ( as well as ho , er , yb , and hf for cs 29497 - 030 ) , although the highly enhanced pb ( and bi for cs 29497 - 030 ) can be reasonably explained .
it should be noted that the [ ba / eu ] ratio in the atmosphere of the observed stars could be expected to be lower than the dredge - up value taken here , after this material is mixed with the envelopes of the primary ( former agb ) and secondary ( post - accretion ) stars .
for example , these envelopes might already have contained @xmath2-process matter at the time of their formation ( e.g. , from supernova - induced star formation with production of @xmath2-process nuclei ; * ? ? ? * ) .
it is unlikely , however , that the mixture of envelope material will result in such low [ ba / eu ] values ( @xmath18 ) for all the stars considered here .
in fact , the cemp-@xmath7 stars account for about @xmath19 of all the cemp stars with the enhancements of @xmath1-process elements ( e.g. , * ? ? ?
on the other hand , non - cemp-@xmath7 stars with a high [ eu / fe ] ratio , relevant to those considered here ( table 1 , @xmath20 } \le + 2.0 $ ] ) are extremely rare
cs 22892 - 052 ( @xmath21 } = + 1.64 $ ] , * ? ? ?
* ) and cs 31082 - 001 ( @xmath21 } = + 1.63 $ ] , * ? ? ?
* ) are the only two such stars with published high - quality abundance analyses .
such stars account for only a few percent of the stars near @xmath22 = - 3.0 $ ] @xcite .
there exists the possibility that , under conditions of extremely high neutron density ( @xmath23 ) , with a sufficiently large exposure in an agb star , one might obtain a low [ ba / eu ] value by an `` @xmath24-process '' ( e.g. , * ? ? ?
* ) , a model that must be investigated more thoroughly in the future . with currently available data , however , the possibility of @xmath2-process contributions can not be excluded . as an exercise , the abundances of he 2148 - 1247 and cs 29497 - 030
are further compared with a simple mixture of the abundances of the low - metallicity agb model and the solar @xmath2-process abundances @xcite , which is normalized to match the [ ba / eu ] ratio in the star .
this result is shown in figure 1 by a thick - solid line , together with the solar @xmath2-process ( thick - dotted ) and @xmath1-process ( thin - dotted ) abundance curves , scaled to match the ba abundance .
good agreement can be seen for both stars , including the boosted pb abundances .
the non - detection of th in these stars ( see also * ? ? ?
* ) may not be crucial , since theoretical studies show that the th abundance can be significantly lower than the scaled solar @xmath2-process curve , even when good agreement is observed up to the third @xmath2-process - peak elements @xcite .
in addition , the upper limit on th for he 2148 - 1247 ( figure 1a ) still does not conflict with the mixture of the solar @xmath2-process and @xmath1-process ( assuming the th abundance at the birth of the star may be @xmath25 dex higher ) .
nevertheless , future detection of th ( which can not be synthesized by the @xmath1-process ) would strongly support a contribution from the @xmath2-process , although it presents an observational challenge because of severe blending with ch lines @xcite .
alternatively , measurements ( or solid upper limits ) on the abundances of elements near the third @xmath2-process peak ( os , ir , and pt ) , which are also not produced by the @xmath1-process , would be of particular importance to support a contribution from the @xmath2-process ( see figure 1b ) . another way to check for possible contamination by the @xmath2-process
may be the ( accurate ) determination of isotopic ratios of , e.g. , eu .
for instance , the ratio @xmath26eu/(@xmath26eu + @xmath27eu ) would be @xmath28 if there were a substantial contribution from the @xmath2-process @xcite .
in contrast , the ratio would be @xmath29 if the @xmath1-process dominated , as is found for some of the cemp stars with large enhancements of @xmath1-process elements , such as lp 625 - 44 and cs 31062 - 050 @xcite .
future accurate measurements of the isotopic ratio of eu ( or other elements , if possible ) for the stars listed in table 1 would be of special importance to test for @xmath2-process contributions to their abundances .
the presence of @xmath1-process elements , along with large enhancements of carbon ( @xmath13 \sim + 2 $ ] , table 1 ) suggests that a mass - transfer episode from a former agb companion in a binary system took place @xcite .
thus , one major goal is to find an astrophysical scenario , associated with an agb star in a binary system , in which the @xmath2-process might also occur .
recent nucleosynthesis studies suggest that core - collapse supernovae , which include `` neutrino winds '' @xcite and `` prompt explosions '' @xcite may be responsible for the production of the @xmath2-process elements .
it should be emphasized that all of these models suffer from severe problems that remain to be solved ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and no consensus has yet been achieved . nevertheless , remarkable agreements of the neutron - capture elements in some extremely metal - poor stars , e.g. , cs 22892 - 052 @xcite and cs 31082 - 001 @xcite , with the scaled solar @xmath2-process curve strongly support the idea that @xmath2-process elements originate from short - lived , massive stars .
@xcite suggested that the double enhancement in the cemp-@xmath7 star he 2148 - 1247 is due to the @xmath1-process occurring in an agb star member of a binary system , followed by the @xmath2-process taking place in a subsequent aic of the white dwarf remnant of the former agb . the nucleosynthetic outcome from an aic event , _ if it occurs _ , can be similar to that arising from a core - collapse supernova , although the absence of an outer envelope in the former case may cause some differences .
the rate of the occurrence of the aic process in the galaxy is highly uncertain , and perhaps no more than @xmath30 yr@xmath31 ( e.g. , * ? ? ?
* ) ( but see * ? ? ?
this rarity seems to be in conflict with the substantial fraction of cemp-@xmath7 stars ( @xmath32 ) among all the cemp-@xmath1 stars currently observed .
in addition , this scenario involves _ three _ separate mass - transfer episodes transfer of @xmath1-process elements from a former agb companion , mass accretion onto the white dwarf remnant , and subsequent pollution of the presently observed member of the system by @xmath2-process elements formed during an aic event of the white dwarf .
this may make such an event extremely rare , although the possibility can not be excluded .
furthermore , an aic is thought to only occur in close binary systems , which is in conflict with the long periods observed for some of the stars listed in table 1 ( see below ) .
note , however , that it remains possible that the explosion may change the orbital period of the binary , or even fractionate the pair into single stars .
compared to the above model , the scenario suggested by @xcite has the advantage that it involves only two mass - transfer steps transfer of @xmath1-process elements from an agb companion followed by pollution with @xmath2-process elements by a `` type 1.5 '' supernova event . the nucleosynthetic outcome from a `` type 1.5 '' supernova may be very similar to that of a type ia supernova , in which @xmath2-processing is not expected to be significant . @xcite and @xcite suggested an alternative scenario , in which the double enhancements are due to a massive agb star that may eventually collapse to be an electron - capture supernova rather than a `` type 1.5 '' supernova .
below we further examine the possibilities suggested by @xcite and @xcite , by investigation of a scenario in which an @xmath3 star with a low - mass companion ( @xmath33 ) in a wide binary system is responsible for the double enhancements of @xmath2-process and @xmath1-process elements , resulting in cemp-@xmath7 stars .
a star in this mass range is likely to undergo @xmath1-processing during its agb phase @xcite , although the amount of the @xmath1-processed material produced and its expected abundance distribution is uncertain .
subsequently , the degenerate o - ne - mg core of this star may collapse by electron capture and explode @xcite ; in such a scenario the @xmath2-process is expected to take place @xcite .
this model also involves only two mass - transfer episodes , as in @xcite .
the possibility of @xmath1-processing occurring in a @xmath34 star with an o - ne - mg core has been suggested recently by @xcite ( see also n. iwamoto et al .
2005 , in preparation , for a similar result with a @xmath35 star ) .
these authors demonstrate that the reaction @xmath36 is efficient in such high - mass stars , owing to the high temperature ( @xmath37 k ) reached at the base of the he convective shell , suggesting that the likely occurrence of the @xmath1-process for stars in this mass range .
it is clear that a more quantitative study of the @xmath1-process nucleosynthesis in this mass range is needed in the future , but the results to date are certainly encouraging .
nomoto ( 1984 ) pointed out that the final fate of @xmath38 stars could be divided into the following two cases , depending on the highly uncertain mass - loss rate .
( i ) for stars in the mass range from 8 @xmath39 to @xmath40 , mass loss results in the ejection of the entire envelope before the core mass reaches the chandrasekhar limit , thereby leaving an o - ne - mg white dwarf .
( ii ) stars in the mass range from @xmath40 to @xmath34 undergo electron - capture supernovae . here
@xmath40 denotes the upper bound mass of the white dwarf progenitors ; if only the c+o white dwarfs are considered , @xmath41 ( see , e.g. , * ? ? ?
* for the metallicity dependence of @xmath42 ) , while @xmath43 applies to the progenitors of o - ne - mg white dwarfs ( see also the more recent studies of * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
stars in the mass range @xmath3 that are found in close binary systems , on the other hand , become helium stars that expand to red giants .
subsequently , their helium envelope is lost by roche - lobe overflow and o - ne - mg dwarfs are formed @xcite .
hence , these stars do not undergo electron - capture supernovae .
currently , he 2148 - 1247 , cs 22948 - 027 , cs 29497 - 034 , cs 29526 - 110 , and cs 29497 - 030 ( table 1 ) have been found to be radial - velocity variables , indicating their binarity . for cs 22948 - 027 , cs 29497 - 034 @xcite , and cs 29497 - 030 @xcite ,
the orbital periods are estimated to be 426.5 , 4130 , and 342 days , respectively .
these stars may belong to the class of ch - star binaries with orbital periods of @xmath44 years , as found by @xcite , although no clear evidence of binarity for the other two cemp-@xmath7 stars in table 1 has been obtained to date .
this high binary frequency ( see also * ? ? ?
* ) implies that , _ if the current scenario is correct _
, an electron - capture supernova must only rarely , if ever , fractionate the pair into single stars .
this is in contrast to the neutrino - driven supernova from a more massive progenitor that may obtain a large recoil velocity ( @xmath45 km s@xmath31 ; * ? ? ?
the lack of fractionation might result if the shock arising from an o - ne - mg core is lifted too early after bounce ( @xmath46 ms , * ? ? ? * ) to obtain a large recoil by convective instability ( @xmath4 s , * ? ? ? * ) , owing to the steep density gradient of the outer edge of the core @xcite .
for wide binary systems mass transfer operates through stellar winds , rather than by roche - lobe overflow @xcite .
@xcite estimated the fraction of the mass captured by its companion ( @xmath47 ) to be @xmath48 % of the mass lost by the wind ( 15 km s@xmath31 ) from the agb star ( @xmath49 ) with a period of 895 days and an orbital velocity of 36 km s@xmath31 , using a three - dimensional , smoothed particle hydrodynamic simulation .
note that most of the cemp-@xmath7 stars under consideration here ( five stars in table 1 ) have relatively high effective temperatures ( @xmath50 k , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , which suggests that these stars are main - sequence turn - off stars . for main - sequence stars of @xmath51 with [ fe / h ] @xmath52 , the mass of the convective envelope , @xmath53 , is smaller by a factor of @xmath5 than that for a stars with solar metallicity ( a few @xmath54 , e.g. , see table 1 in * ? ? ?
as a result , the dilution of the accreted material is relatively small .
here we assume @xmath55 , which is one order of magnitude smaller than that of the sun ( @xmath56 , * ? ? ?
if we take [ c / fe ] = [ ba / fe ] = + 2 and [ fe / h ] = @xmath57 as representative of the cemp-@xmath7 stars in table 1 , the accreted masses of c and ba from the agb onto its companion are estimated to be @xmath58 and @xmath59 , respectively . assuming the mass accretion rate to be @xmath60 , according to ( * ? ? ?
although the binary system in their simulation is not completely relevant to this study ) , the masses of c and ba ejected from the agb are estimated to be @xmath61 and @xmath62 .
the former is in good agreement with the result from the stellar evolution calculation of a @xmath63 star by n. iwamoto et al .
( 2005 , in preparation ) .
for the abundance of ba , a future study of the @xmath1-process nucleosynthesis in an @xmath64 model is needed to confirm the current hypothesis .
note that the amount of c added by the subsequent supernova wind is negligible , owing to its far less efficient accretion onto the secondary compared to that from an agb star ( 3.6 ) .
there is no currently available hydrodynamic study of the efficiency of wind pollution by a supernova in a binary system .
hence , we now estimate the fraction of the ejected mass from a subsequent supernova explosion which is captured by the companion star simply to be @xmath65 , as in @xcite . here , @xmath66 is the geometric fraction of the companion , @xmath67 is the radius of the secondary , @xmath68 is the initial orbital separation , and @xmath69 is the enhancement ( or reduction ) factor .
the value of @xmath69 can be larger than unity as the result of `` gravitational focusing '' , if the ejecta velocity is decelerated below the escape velocity from the surface of the secondary .
on the other hand , @xmath69 can be smaller than unity if the ejecta velocity is large enough to strip the surface material from the secondary .
assuming the masses of the primary and secondary to be @xmath63 and @xmath70 , respectively , with @xmath71 and an initial orbital period of one year , we obtain @xmath72 ( which reduces to @xmath73 , if we change the initial orbital period to five years ) .
note that the orbital period would become as twice as large as its initial value after the explosion of the primary , owing to the reduction of its mass to @xmath74 , even if the orbital separation were unchanged . the accreted material is further diluted with the mass of the convective zone of the secondary , e.g. , @xmath75 ( see 3.4 ) .
this results in the required mass of eu produced per supernova event ( from a star of initial mass @xmath63 ) is @xmath76 , to obtain @xmath77 \sim + 2 $ ] for the secondary star with @xmath15 \sim - 2.5 $ ] .
a galactic chemical evolution study shows that the required mass of eu per supernova event in order to account for its solar value is @xmath78 ( e.g. , * ? ? ?
* ) , if _ all _ core - collapse supernovae equally contribute to its enrichment .
@xcite have suggested , however , that the supernova progenitors that contribute to the chemical evolution of eu ( as representative of @xmath2-process elements ) must be limited to a small range ( @xmath79 of all supernova events , e.g. , @xmath3 or @xmath80 , see also * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this leads to a natural explanation of the large star - to - star scatter of the @xmath2-process elements ( e.g. , eu ) relative to iron ( by more than two orders of magnitude ) that can be seen in extremely metal - poor halo stars .
if one assumes that the stars of @xmath3 ( i.e. , @xmath81 , see 3.2 ) are the dominant source of eu , the mass of eu ejected per explosive event estimated from a galactic chemical evolution study should be increased to @xmath82 @xcite .
this follows because the mass range @xmath3 accounts for about @xmath83 of all supernova events , when assuming a salpeter initial mass function .
the mass of eu per event would be @xmath84 if @xmath40 were , e.g. , @xmath85 , since the mass range of @xmath86 accounts for only about @xmath87 of all supernova events .
note that further restriction of the mass range ( e.g. , @xmath88 ) would lead to larger star - to - star scatter of [ eu / fe ] values than those observed in extremely metal - poor stars .
it should be also noted that the amount of eu from a supernova event estimated here seems reasonable from a nucleosynthetic point of view ( e.g. , from the neutrino wind scenario as a possible explanation , * ? ? ?
* ; * ? ? ?
thus , the estimate from the wind model ( 3.5 ) and the constraint from galactic chemical evolution above are in good agreement when assuming @xmath89
. however , an enhancement of the accretion by gravitational focusing is needed ( @xmath90 ) , if we assume an initial orbital period of five years pertains .
this shows that wind pollution by a supernova explosion is far less effective than that by an agb star .
in fact , the efficiency of accretion for the agb star is more than four orders of magnitude larger than that estimated by the simple geometric fraction ( i.e. , @xmath91 ) , owing to its small expansion velocity ( comparable to the orbital period of the system ) . it should be noted that the contribution of @xmath1-processed material ( e.g. , ba ) produced for stars of @xmath3 ( which are short - lived stars ) must be a negligible contributor to the galactic chemical evolution of neutron - capture elements .
this is required in order to be consistent with observations of non - cemp stars having [ ba / eu ] values very close to the solar @xmath2-process ratio , with no sign of an increase owing to the @xmath1-process for [ fe / h ] @xmath92 @xcite .
the mass of ba per supernova event ( from a star of initial mass @xmath63 ) due to the @xmath2-process can be estimated to be @xmath93 , assuming @xmath94 applied above ( 3.5 ) and the solar @xmath2-process ratio of ba / eu (= 9.29 , * ? ? ?
on the other hand , the estimated mass of ba from the @xmath1-process during the agb phase ( for a star with an initial mass of @xmath35 ) is @xmath62 ( 3.4 ) .
this is negligible compared to the @xmath2-process contribution considered here , when assuming @xmath95 .
the discussion above demonstrates that the accretion efficiency onto the secondary should not be reduced with respect to its geometric fraction , leading to @xmath96 . for supernova explosions with a typical explosion energy ( @xmath97 ergs ) ,
the velocity of the inner ejecta is expected to be a few thousand km s@xmath31 ( e.g. , * ? ? ?
this is larger than the escape velocity from the secondary ( e.g , @xmath98 km s@xmath31 for the sun ) , which may result in @xmath99 .
however , a collapsing o - ne - mg core is expected to lead to a neutrino - powered explosion with a rather low explosion energy ( a few times @xmath100 ergs ; * ? ? ?
furthermore , its innermost ejecta , in which the @xmath2-process is expected to operate , may expand rather slowly .
in fact , the core - collapse supernova from a more massive progenitor ( @xmath101 ) is considered to suffer from fallback onto the remnant @xcite , in which the expansion velocity of the inner ejecta becomes zero at some point .
if the innermost ejecta from a collapsing o - ne - mg core expands slowly ( e.g. , @xmath102 a few hundred km s@xmath31 ) without substantial fallback , the accretion can be significantly enhanced by gravitational focusing .
this may result in @xmath69 becoming larger than unity .
it is obvious , however , that a detailed hydrodynamic study will be needed in the future to estimate quantitatively the efficiency of the wind - pollution model discussed here .
the abundances of cemp stars with large enhancements of @xmath1-process elements , but with the lowest [ ba / eu ] ratios ( @xmath18 ) , disagree with the predicted elemental abundance patterns from contemporary low - metallicity agb models , and seem to require an additional @xmath2-process contribution . we have investigated a model in which these cemp-@xmath7 stars could be accounted for by an @xmath3 star in a wide binary system that is responsible for enrichment with @xmath1-process elements during its agb phase , and with @xmath2-process elements by the subsequent supernova explosion of its collapsing o - ne - mg core .
it should be cautioned , however , that the expected @xmath1-process signature resulting from the agb stage in stars in this mass range , as well as the @xmath2-process abundance signature of the subsequent core collapse of stars of this mass , are still not well known .
the estimated mass of eu ( as representative of @xmath2-process elements ) captured by the secondary , through the wind from the supernova , is in good agreement with the constraint obtained from a galactic chemical evolution study , at least when the initial orbital separation is taken to be @xmath4 year .
however , the efficiency of wind pollution from the supernova must be enhanced by a factor of @xmath5 when assuming an initial orbital separation to be @xmath103 years .
it is suggested that the expansion velocity of the supernova s innermost ejecta , in which the @xmath2-process has taken place , must be significantly slow , resulting in an enhancement of accretion efficiency by gravitational focusing .
future theoretical studies of @xmath1-process and @xmath2-process nucleosynthesis in @xmath3 stars , as well as a full hydrodynamic study of wind pollution during the supernova explosion in a wide binary system , are needed before one can draw firm conclusions .
future comprehensive spectroscopic studies of cemp-@xmath7 stars , in particular measurements of their pt - peak ( os , ir , and pt ) and th abundances , and/or isotopic eu measurements , are also of special importance to confirm the @xmath2-process contribution to these stars
. we would like to acknowledge w. aoki , s. goriely , and h. umeda for helpful discussions .
we also acknowledge the contributions of an anonymous referee , which led to clarification of a number of points in our original manuscript . this work was supported in part by a grant - in - aid for the japan - france integrated action program ( sakura ) , awarded by the japan society for the promotion of science , and scientific research ( 15204010 , 16042201 , 16540229 , 17740108 ) , and from the 21st century coe program ( quest ) from the ministry of education , culture , sports , science , and technology of japan . t.c.b .
acknowledges partial support from a series of grants awarded by the us national science foundation , most recently , ast 04 - 06784 , as well as from grant phy 02 - 16783 ; physics frontier center / joint institute for nuclear astrophysics ( jina ) .
aoki , w. , ryan , s. g. , norris , j. e. , beers , t. c. , ando , h. , & tsangarides , s. 2002 , , 580 , 1149 aoki , w. , honda , s. , beers , t. c. , & sneden , c. 2003 , , 586 , 506 aoki , w. , et al .
2003 , , 592 , l67 arlandini , c. , kppeler , f. , wisshak , k. , gallino , r. , lugaro , m. , busso , m. , & straniero , o. , 1999 , , 525 , 886 bahcall , j. n. , pinsonneault , m. h. , & wasserburg , g. j. 1995 , rev .
phys , 67 , 781 bailyn , c. d. & grindlay , j. e. 1990 , , 353 , 159 barklem , p.s .
, et al . 2005 , , 439 , 129 beers , t.c . , & christlieb , n. 2005 , araa , 43 , 531 boffin , h. m. j. & jorissen , a. 1988 , , 205 , 155 barbuy , b. , spite , m. , spite , f. , hill , v. , cayrel , r. , plez , b. & petitjean , p. 2005
, , 429 , 1031 cohen , j. g. , christlieb , n. , qian , y. -z .
& wasserburg , g. j. 2003 , , 588 , 1082 eldridge , j. j. & tout , c. a. 2004 , , 353 , 87 gallino , r. , arlandini , c. , busso , m. , lugaro , m. , travaglio , c. , straniero , o. , chieffi , a. , & limongi , m. 1998 , , 497 , 388 goriely , s. & mowlavi , n. 2000 , , 362 , 599 goriely , s. & siess , l. 2001 , , 378 , l25 habets , g. m. h. j. 1986 , , 165 , 95 hashimoto , m. , iwamoto , k. , & nomoto , k. 1993 , , 414 , l105 hill , v. , et al .
2000 , , 353 , 557 hill , v. , et al .
2002 , , 387 , 560 hillebrandt , w. , nomoto , k. , & wolff , g. 1984 , , 133 , 175 iben , i. & renzini , a. 1983 , , 21 , 271 iben , i. jr . , ritossa , c. , & garca - berro , e. 1997 , , 489 , 772 ivans , i. i. , sneden , c. , gallino , r. , cowan , j. j. , & preston , g. w. 2005 , , 627 , l145 ishimaru , y. & wanajo , s. 1999 , , 511 , l33 ishimaru , y. , wanajo , s. , aoki , w. , & ryan , s. g. 2004 , , 600 , l47 janka , h. -t . , buras , r. , kitaura joyanes , f. s. , marek , a. , rampp , m. , & scheck , l. 2005 , , 758 , 19 johnson , j. a. & bolte , m. 2001 , , 554 , 888 johnson , j. a. & bolte m. 2004 , , 605 , 462 kppeler , f. , beer , h. , & wisshak , k. 1989 , rep .
phys . , 52 , 945 lucatello , s. , tsangarides , s. , beers , t. c. , carretta , e. , gratton , r. g. , & ryan , s. g. 2005 , , 625 , 825 mcclure , r. d. & woodsworth , a. w. 1990 , , 352 , 709 nomoto , k. 1984 , , 277 , 791 nomoto , k. 1987 , , 322 , 206 nomoto , k. & hashimoto , m. 1988 , , 163 , 13 nomoto , k. , & kondo , y. 1991 , , 367 , 19 nomoto , k. , wanajo , s. , iwamoto , n. , & ishimaru , y. 2004 , in proc .
12th workshop on nuclear astrophysics ( 22 - 27 march 2004 , ringberg castle : tegernsee ) , eds .
e. muller & h .-
janka ( garching : max - planck - institut fur astrophysik ) , 123 norris , j.e . , ryan , s.g .
, & beers , t.c . 1997 , 489 , 169 podsiadlowski , p. , nomoto , k. , maeda , k. , mazzali , p. , & schmidt , b. 2002 , , 567 , 491 podsiadlowski , p. , langer , n. , poelarends , a. j. t. , rappaport , s. , heger , a. , & pfahl , e. 2004 , , 612 , 1044 qian , y. -z . & wasserburg , g. j. 2003 , , 588 , 1099 ritossa , c. , garca - berro , e. , & iben , i. jr . 1996 , , 460 , 489 ritossa , c. , garca - berro , e. , & iben , i. jr . 1999 , , 515 , 381 scheck , l. , plewa , t. , janka , h. -th . , kifonidis , k. , & mller , e. 2004 , , 92 , 011103 shigeyama , t. , suzuki , t. , kumagai , s. , nomoto , k. , saio , h. , & yamaoka , h. 1994 , , 420 , 341 sivarani , t. , et al .
2004 , , 413 , 1073 sneden , c. , cowan , j. j. , lawler , j. e. , burles , s. , beers , t. c. , & fuller , g. m. 2002 , , 566 , l25 sneden , c. , et al .
2003 , , 591 , 936 sneden , c. , preston , g. w. , & cowan , j. j. 2003 , , 592 , 504 sumiyoshi , k. , terasawa , m. , mathews , g. j. , kajino , t. , yamada , s. , & suzuki , h. 2001 , , 562 , 880 theuns , t. , boffin , h. m. j. , & jorissen , a. 1996 , , 280 , 1264 tsujimoto , t. , shigeyama , t. , & yoshii , y. 2000 , , 531 , l33 umeda , h. , nomoto , k. , yamaoka , h. , & wanajo , s. 1999 , , 513 , 861 umeda , h. & nomoto , k. 2002 , , 565 , 385 umeda , h. & nomoto , k. 2003 , , 422 , 871 van eck , s. , goriely , s. , jorissen , a. , & plez , b. 2003 , , 404 , 291 wanajo , s. , kajino , t. , mathews , g. j. , & otsuki , k. 2001 , , 554 , 578 wanajo , s. , itoh , n. , ishimaru , y. , nozawa , s. , & beers , t. c. 2002 , , 577 , 853 wanajo , s. , tamamura , m. , itoh , n. , nomoto , k. , ishimaru , i. , beers , t. c. , & nozawa , s. 2003 , , 593 , 968 wanajo , s. & ishimaru , i. 2005 , , in press woosley , s. e. , wilson , j. r. , mathews , g. j. , hoffman , r. d. , & meyer , b. s. 1994 , , 433 , 229 yoshii , y. 1981 , , 97 , 280 zijlstra , a. a. 2004 , , 348 , l23 cllllllc he 2148 - 1247 & @xmath1042.3 & + 1.91 & + 2.36 & + 1.98 & + 0.38 & + 3.12 & 1 + cs 22948 - 027 & @xmath1042.47 & + 2.43 & + 2.26 & + 1.88 & + 0.38 & + 2.72 & 2 + cs 29497 - 034 & @xmath1042.90 & + 2.63 & + 2.03 & + 1.80 & + 0.23 & + 2.95 & 2 + cs 29526 - 110 & @xmath1042.38 & + 2.2 & + 2.11 & + 1.73 & + 0.38 & + 3.3 & 3 + cs 22898 - 027 & @xmath1042.25 & + 2.2 & + 2.23 & + 1.88 & + 0.35 & + 2.84 & 3 + cs 31062 - 012 & @xmath1042.55 & + 2.1 & + 1.98 & + 1.62 & + 0.36 & + 2.4 & 3 + cs 29497 - 030 & @xmath1042.57 & + 2.30 & + 2.32 & + 1.99 & + 0.33 & + 3.65 & 4 + cs 22892 - 052 & @xmath1043.10 & + 0.95 & + 0.99 & + 1.64 & @xmath1040.65 & + 1.20 & 5 + | recent spectroscopic studies have revealed the presence of numerous carbon - enhanced , metal - poor stars with [ fe / h ] @xmath0 that exhibit strong enhancements of @xmath1-process elements .
these stars are believed to be the result of a binary mass - transfer episode from a former asymptotic giant - branch ( agb ) companion that underwent @xmath1-process nucleosynthesis . however , several such stars exhibit significantly lower ba / eu ratios than solar @xmath1-process abundances .
this might be explained if there were an additional contribution from the @xmath2-process , thereby diluting the ba / eu ratio by extra production of eu .
we propose a model in which the double enhancements of @xmath2-process and @xmath1-process elements originate from a former @xmath3 companion in a wide binary system , which may undergo @xmath1-processing during an agb phase , followed by @xmath2-processing during its subsequent supernova explosion .
the mass of eu ( as representative of @xmath2-process elements ) captured by the secondary through the wind from the supernova is estimated , which is assumed to be proportional to the geometric fraction of the secondary ( low - mass , main - sequence ) star with respect to the primary ( exploding ) star .
we find that the estimated mass is in good agreement with a constraint on the eu yield per supernova event obtained from a galactic chemical evolution study , when the initial orbital separation is taken to be @xmath4 year .
if one assumes an orbital period on the order of five years , the efficiency of wind pollution from the supernova must be enhanced by a factor of @xmath5 .
this may , in fact , be realized if the expansion velocity of the supernova s innermost ejecta , in which the @xmath2-process has taken place , is significantly slow , resulting in an enhancement of accretion efficiency by gravitational focusing . |
vacuum polarization is known to result in such nonlinear quantum electrodynamics ( qed ) effects as coherent photon scattering and photon splitting in an external electromagnetic field .
photon scattering in the electric field of atoms ( delbrck scattering ) has been investigated in detail both theoretically and experimentally @xcite .
recently , an essential progress in the understanding of the photon splitting process in an atomic electric field was achieved @xcite and the first successful observation of photon splitting was performed @xcite .
it turned out that higher orders of the perturbation theory with respect to @xmath0 play an important role and drastically modify the cross sections of delbrck scattering and photon splitting at high photon energy ( here , z is the nuclear charge number , @xmath1 is the fine - structure constant with @xmath2 being the electron charge , the system of units @xmath3 is used ) .
therefore , the theoretical and experimental investigation of photon splitting gives the possibility to verify qed in the presence of strong external field when the nonperturbative effects in the field strength ( contributions of higher - order terms ) are very important .
photon splitting in a constant and uniform electromagnetic field for arbitrary values of the field strength was considered in numerous papers because of the potential astrophysical implications of the process ( see , e.g. , refs .
the possibility of observing photon splitting in the electric fields of single crystals at high photon energies was investigated in @xcite .
photon splitting due to vacuum polarization in a strongly magnetized plasma was discussed in @xcite .
photon splitting in a laser field with arbitrary strength and frequency content has never been considered before for an arbitrary energy of the incoming photon .
the kinematics of photon splitting depends essentially on the structure of the external field . in the coulomb field , the wave vector @xmath4 of the initial photon and the wave vectors @xmath5 and @xmath6 of the final ones are , in general , not coplanar . in a constant and uniform electromagnetic field ,
the initial and final photons are almost collinear ( up to nonzero photon mass due to vacuum polarization in the external field ) . in a classical electromagnetic plane wave , due to energy - momentum conservation ,
the initial and final photons are coplanar in the frame where the laser wave propagates in the direction anti - parallel to @xmath4 . throughout the paper
we refer to this frame as the laboratory frame .
the structure of the photon splitting amplitude in a laser field depends on the value of two lorentz - invariant parameters . in the laboratory frame
they are given by @xmath7 where @xmath8 is the energy of the initial photon , @xmath9 and @xmath10 are the characteristic frequency and electric field strength of the laser pulse , @xmath11v / cm is the critical electric field and @xmath12 is the electron mass . for a monochromatic laser field , @xmath9 and @xmath10 in eq .
( [ chieta ] ) are the laser frequency and the root mean square of the laser electric field . for @xmath13 and @xmath14 ,
photon splitting in a monochromatic laser wave was considered in @xcite in the leading approximation with the help of the euler - heisenberg lagrangian .
the contribution to the amplitudes of photon splitting linear in @xmath15 but at arbitrary @xmath16 is nothing but the usual photon - photon scattering which was studied in detail theoretically ( see , e.g. , @xcite ) but never observed experimentally @xcite .
the possibility to observe `` laser - assisted '' photon - photon scattering by using the next generation of petawatt lasers has been investigated in @xcite .
other processes related to photon - photon scattering and initiated by vacuum polarization in the presence of the superposition of two laser fields , or a laser field and a coulomb field have been considered in @xcite . for the strongest optical ( @xmath17ev ) laser beam available today
the laser intensity is @xmath18 and the corresponding electric field amplitude is @xmath19 @xcite . the extreme light infrastructure ( eli ) project aims at reaching laser intensity values higher than @xmath20 corresponding to @xmath21 . the x - ray free - electron laser ( x - fel ) facilities presently being developed at desy ( hamburg , germany ) and at slac ( stanford , usa ) are intended to provide spatially coherent and highly brilliant beams of synchrotron radiation with single - photon energies up to @xmath22 kev at a maximum intensity @xmath23 corresponding to a field strength of @xmath24 @xcite .
the above mentioned projects have initiated intensive investigations of various qed processes that can be experimentally observed when using these future facilities @xcite .
on the other hand , sources of high - energy photons are nowadays available in a wide range of photon energies spanning from tens of mev up to hundreds of gev .
it will then be possible to reach experimentally unexplored regions for the values of the parameters @xmath16 and @xmath15 where vacuum polarization effects in a laser field , and photon splitting in particular , are essentially nonperturbative . in the present paper ,
we derive the amplitudes of photon splitting in a plane wave described by an arbitrary four - vector potential @xmath25 , where @xmath26 with @xmath27 , @xmath28 , and @xmath29 . at a first stage
, we do not assume that the plane wave is monochromatic or even periodic .
it is convenient , without loss of generality , to consider the process in the laboratory frame , i.e. in the frame where @xmath4 is anti - parallel to the vector @xmath30 .
the calculations are performed by using the operator technique developed in refs .
@xcite for the calculation of the polarization and mass operators in a laser field .
alternative forms of these operators were derived independently in refs .
@xcite by means of a different method .
the results of refs . @xcite and refs .
@xcite are in agreement with each other .
the operator technique allows one to find the amplitudes without using the explicit form of the green s function for the dirac equation in the external field .
this circumstance considerably simplifies the calculations .
the rest of the paper is organized as follows : in section ii we describe the general properties of the photon splitting amplitudes .
then , starting from the operator representation of the amplitudes , in sections iii and iv we calculate the explicit expressions of the amplitudes for a general vector potential of the form @xmath31 . in section v we consider a monochromatic laser field with particular emphasis on the case of a circularly polarized laser field . in section
vi we analyze various asymptotic forms of the amplitudes both for a circular and for a linearly polarized field .
the possibility of the experimental observation of photon splitting in a laser field is discussed in detail in section vii .
finally , in section viii the main conclusions of the paper are presented .
the amplitude @xmath32 of the photon splitting process of a photon with four - momentum @xmath33 into two photons with four - momenta @xmath34 and @xmath35 with @xmath36 is represented by the feynman diagram shown in fig .
the corresponding formal expression for @xmath32 in the furry representation reads @xmath37 where @xmath38 is the initial photon polarization four - vector , @xmath39 and @xmath40 are the polarization four - vectors of the final photons , @xmath41 with @xmath42 and @xmath43 being the dirac matrices .
the factor @xmath44 , contained in the external photon wave functions , has been included into the formula for the probability of the process .
the calculation of the amplitude @xmath32 is essentially simplified by using the green s function @xmath45 of the `` squared '' dirac equation : @xmath46 as shown in @xcite , the expression of @xmath32 can be written as a sum of two contributions containing either three or two green s functions @xmath47 : @xmath48 the term @xmath49 has the form @xmath50 for each photon with the four - momentum @xmath51 ( @xmath52 ) it is convenient to introduce two vectors @xmath53 , ( @xmath54 ) @xmath55 due to gauge invariance , it is possible to write the tensor @xmath56 introduced in eq .
( [ tensorm ] ) for real photons in the form ( see ref .
@xcite ) : @xmath57 for the coefficients @xmath58 we have @xmath59 so , the coefficients @xmath58 can be calculated from eqs .
( [ m2 ] ) and ( [ m3 ] ) by the substitution @xmath60 . using the coefficients @xmath58
, we write the amplitude @xmath32 of photon splitting in the laser field as @xmath61 this expression of the amplitude @xmath32 is very convenient because it automatically provides the fulfillment of the gauge invariance principle . as a consequence ,
it is not necessary to perform the subtraction in the calculation of the lowest - order photon splitting amplitude , described by the photon - photon scattering diagrams .
we perform the calculations in terms of the helicity amplitudes @xmath62 , with @xmath63 . in this case
, the polarization vector @xmath64 of each external photon with the wave vector @xmath65 and the energy @xmath66 satisfies the relations @xmath67 and @xmath68 . due to momentum conservation in the laser field ,
the final transverse momenta @xmath69 and @xmath70 ( with respect to the laser propagation direction ) are equal in modulus and opposite in direction : @xmath71 and @xmath72 .
we direct the vector @xmath73 along the vector @xmath74 , so that @xmath75 , then @xmath76 the corresponding polarization four - vectors @xmath77 have the following products with the four - vectors @xmath78 : @xmath79 these useful relations allow us to state that the amplitude @xmath62 can be obtained from the coefficient @xmath58 by the substitution in the latter : @xmath80 the energies @xmath81 and the parallel components @xmath82 of the momenta of the final photons are functions of the transverse momentum @xmath74 and of the energy @xmath83 absorbed from the laser field , @xmath84 .
it is convenient to introduce the quantities @xmath85 the angle @xmath86 coincides with the angle between the vectors @xmath5 and @xmath30 in the frame where @xmath87 , so that @xmath88
. then @xmath89 it follows from eqs .
( [ m2 ] ) and ( [ m3 ] ) that , like the amplitude @xmath32 , the coefficients @xmath58 can also be presented as the sum @xmath90 . in order to explain the application of the operator technique to the calculation of the photon splitting amplitude , we present the derivation of the contribution @xmath91 .
the calculation of @xmath92 can be performed in a similar way and only the final result is given for this contribution .
we direct the vector @xmath30 along the z - axis and pass from the variables @xmath93 and @xmath94 to the variables @xmath95 and @xmath96 . in this way
@xmath97 then , we use in eq .
( [ m2 ] ) the representation of the operator green s function derived in @xcite @xmath98\nonumber\\ & & = -i\int_0^\infty ds\,\exp(-ism^2)\,\left\{1+\frac{e\hat{\varkappa}}{2p_t } [ \hat{a}(\phi+2sp_t)-\hat{a}(\phi)]\right\}\nonumber\\ & & \times\exp\left[-i\int_0^sds'\left({\bm p}_\perp - e{\bm a}(\phi+2s'p_t)\right)^2\right]\,\exp(2ip_\phi p_t)\ , .\end{aligned}\ ] ] as before , the index @xmath99 indicates transverse quantities with respect to the laser propagation direction @xmath30 .
now , we pass in eq .
( [ m2 ] ) from the integration over the variables @xmath100 and @xmath101 to the corresponding momenta @xmath102 and @xmath103 in accordance with @xmath104 we also exploit the relations @xmath105 valid for an arbitrary function @xmath106 of the variable @xmath107 . besides , the matrix elements @xmath108 and @xmath109 become @xmath110|\bm p_\perp\rangle = ( 2\pi)^2\delta(\bm k_{2\perp}+\bm k_{3\perp})\,.\end{aligned}\ ] ] the factors ( [ delta ] ) are common to all amplitudes .
they will be omitted below in the expressions of the amplitudes and will be taken into account in the usual way in the formula for the probability of the process .
the matrix element @xmath111 also becomes the @xmath112-function containing the variable @xmath102 in its argument .
therefore , the integration over @xmath102 becomes trivial as well as the integration over @xmath113 .
following the procedure pointed out above and using the useful relations @xmath114 with @xmath115 , we arrive at the following result for @xmath91 : @xmath116\nonumber\\ & & \times\int d\phi\exp\left\{i\omega\phi + i\frac{4\omega_1\omega}{m^2}\varphi_j - i(s_1+s_2)\left[\int_0 ^ 1dy{\bm\delta}^2(y\varphi_j)- \left(\int_0 ^ 1dy{\bm\delta}(y\varphi_j)\right)^2\right]\right\}\nonumber\\ & & \times\left[-2{\bm a}_{\lambda_j}\cdot\int_0 ^ 1dy \bm{\delta}(y\varphi_j)+\frac{(s_1+s_2)^2}{2s_1s_2}{\bm a}_{\lambda_j}\cdot\bm{\delta}(\varphi_j)\right]\ , .\end{aligned}\ ] ] here we defined the quantities @xmath117\ , .
\quad\end{aligned}\ ] ] in the parametrization of the green s function we made the substitution @xmath118 .
as it has to be according to the furry theorem , the contribution ( [ s2final ] ) contains only odd powers of the external field .
the explicit result of the integration over @xmath107 depends on the form of the function @xmath31 . for a monochromatic wave , the integral over @xmath107
can be easily performed ( see section v ) .
the calculation of the contribution @xmath92 is similar to that of @xmath91 , but considerably more cumbersome . in order to bring the answer into a relatively short form
, we introduce the convenient notation @xmath119\ , , \nonumber\\ & & f_0= s{\bm f}^2-\int_0 ^ 1dy[s_1\bm { d}_1 ^ 2(y)+s_2\bm { d}_2 ^
2(y)+s_3\bm { d}_3 ^ 2(y)]\ , , \nonumber\\ & & \bm g=\frac{2s_1 } { s}\int_0 ^ 1dy\,[s_2(\bm { d}_1(y)-\bm { d}_2(y))+s_3(\bm { d}_1(y)-\bm { d}_3(y))]\ , , \end{aligned}\ ] ] and use the relation @xmath120 . before taking the trace over the @xmath121-matrices ,
we obtain the following compact representation for @xmath92 : @xmath122\nonumber\\ & & \times\bigg\{\frac{1}{4}{\mbox{tr}}\bigg[\left(-2{\bm a}_{\lambda_3}\cdot{\bm n}_3 + \frac{\hat{\varepsilon}_{3,\lambda_3}\hat{k}_3}{m}\right)\left(1+\frac{m\hat{\varkappa}\hat{d}_3}{2\omega_1}\right ) \left(-2{\bm a}_{\lambda_1}\cdot{\bm n}_1 -\frac{\hat{\varepsilon}_{1,\lambda_1}\hat{k}_1}{m}\right)\nonumber\\ & & \times\left(1+\frac{m\hat{\varkappa}\hat { d}_2}{2\omega_1}\right)\left(-2{\bm a}_{\lambda_2}\cdot{\bm n}_2+\frac{\hat{\varepsilon}_{2,\lambda_2}\hat{k}_2}{m}\right ) \left(1+\frac{m\hat{\varkappa}\hat{d}_1}{2\omega_1}\right)\bigg]\nonumber\\ & & + \frac{4i } { s}\left[\delta_{(1)}{\bm a}_{\lambda_1}\cdot({\bm n}_1-\bm v)+ \delta_{(2)}{\bm a}_{\lambda_2}\cdot({\bm n}_2+\nu_2\bm v)+ \delta_{(3)}{\bm a}_{\lambda_3}\cdot({\bm n}_3+\nu_3\bm v)\right]\bigg\}\ , \nonumber\\ & & + ( 2\leftrightarrow 3\ , , \,\ , \bm \rho\rightarrow -\bm \rho ) \,,\end{aligned}\ ] ] where @xmath123 the substitution @xmath124 in eq . ( [ s3trace ] )
results in the replacement @xmath125 , @xmath126 , and @xmath127 .
in fact , the role of the term with the substitution @xmath128 is to provide the fulfillment of the furry theorem ( @xmath92 contains only odd powers of the external field ) .
therefore , the sum of the two terms is twice the odd part of the first term with respect to the external vector potential @xmath129 .
finally , we take the trace and obtain : @xmath130\nonumber\\ & & \times\bigg<-4({\bm a}_{\lambda_1}\cdot{\bm n}_1 ) ( { \bm a}_{\lambda_2}\cdot{\bm n}_2 ) ( { \bm a}_{\lambda_3}\cdot{\bm n}_3)+4({\bm
a}_{\lambda_2}\cdot{\bm n}_2 ) ( { \bm a}_{\lambda_3}\cdot{\bm n}_3)({\bm a}_{\lambda_1}\cdot{\bm v } ) \nonumber\\ & & -4\nu_3({\bm a}_{\lambda_1}\cdot{\bm n}_1 ) ( { \bm a}_{\lambda_2}\cdot{\bm n}_2 ) ( { \bm a}_{\lambda_3}\cdot{\bm v})-4\nu_2({\bm a}_{\lambda_1}\cdot{\bm n}_1 ) ( { \bm a}_{\lambda_3}\cdot{\bm n}_3)({\bm a}_{\lambda_2}\cdot{\bm v } ) \nonumber\\ & & + ( { \bm a}_{\lambda_1}\cdot{\bm n}_1)\bigg\ { \frac{[\bm \rho\cdot\bm \rho]_{\lambda_2\lambda_3}}{2\nu_2\nu_3}+[{\bm \rho}\cdot ( \bm d_1-\bm d_2-\bm d_3 ) ] _ { \lambda_2\lambda_3}-2\nu_2\nu_3[\bm d_1\cdot(\bm d_2+\bm d_3)]_{\lambda_2\lambda_3}\bigg\ } \nonumber\\ & & -({\bm a}_{\lambda_2}\cdot{\bm n}_2)\bigg\ { \frac{[\bm \rho\cdot\bm \rho]_{\lambda_1\lambda_3}}{2\nu_3}+[{\bm \rho}\cdot ( \bm d_1+\bm d_2-\bm d_3 ) ] _ { \lambda_1\lambda_3}-2\nu_3[\bm d_3\cdot(\bm d_1+\bm d_2)]_{\lambda_1\lambda_3}\bigg\ } \nonumber\\ & & -({\bm a}_{\lambda_3}\cdot{\bm n}_3)\bigg\
{ \frac{[\bm \rho\cdot\bm \rho]_{\lambda_1\lambda_2}}{2\nu_2}+[{\bm \rho}\cdot ( \bm d_1-\bm d_2+\bm d_3 ) ] _ { \lambda_1\lambda_2}-2\nu_2[\bm d_2\cdot(\bm d_1+\bm d_3)]_{\lambda_1\lambda_2}\bigg\ } \nonumber\\ & & + ( { \bm a}_{\lambda_1}\cdot{\bm \rho})[{\bm \rho } \cdot\bm d_1]_{\lambda_2\lambda_3}-\frac{1}{\nu_2}({\bm a}_{\lambda_2 } \cdot{\bm \rho})[{\bm \rho}\cdot\bm d_3]_{\lambda_1\lambda_3}- \frac{1}{\nu_3}({\bm a}_{\lambda_3 } \cdot{\bm \rho})[{\bm \rho}\cdot\bm d_2]_{\lambda_1\lambda_2}\nonumber\\ & & -\frac{\rho^2}{2}\bigg\{[{\bm a}_{\lambda_1}\cdot\bm d_1]_{\lambda_2\lambda_3 } -\frac{1}{\nu_2}[{\bm a}_{\lambda_2}\cdot\bm d_3]_{\lambda_1\lambda_3 } -\frac{1}{\nu_3}[{\bm a}_{\lambda_3}\cdot\bm d_2]_{\lambda_1\lambda_2 } \bigg\}\nonumber\\ & & + \frac{1}{\nu_2\nu_3}(\bm a_{\lambda_1}\cdot { \bm \rho } ) ( \bm a_{\lambda_2}\cdot { \bm \rho})(\bm a_{\lambda_3}\cdot { \bm \rho } ) -\frac{\rho^2}{4\nu_2\nu_3}[\delta_{(1)}(\bm a_{\lambda_1}\cdot{\bm \rho})+\delta_{(2)}(\bm a_{\lambda_2}\cdot{\bm \rho } ) + \delta_{(3)}(\bm a_{\lambda_3}\cdot{\bm \rho})]\ , \nonumber\\ & & + \frac{2i } { s}\left[\delta_{(1)}{\bm a}_{\lambda_1}\cdot({\bm n}_1-\bm v)+ \delta_{(2)}{\bm a}_{\lambda_2}\cdot({\bm n}_2+\nu_2\bm v)+ \delta_{(3)}{\bm a}_{\lambda_3}\cdot({\bm n}_3+\nu_3\bm v)\right]\ , \nonumber\\ & & + u_1 -\nu_2\,u_2-\nu_3\,u_3 + 2\nu_2\nu_3\,u_4\bigg>\,\ , + \quad ( 2\leftrightarrow 3\ , , \,\ , \bm \rho\rightarrow -\bm \rho ) \,.\end{aligned}\ ] ] here , we introduced the operator @xmath131_{\lambda_i\lambda_j}\equiv ( \bm x\cdot\bm a_{\lambda_i } ) ( \bm y\cdot\bm a_{\lambda_j})+(\bm x\cdot\bm a_{\lambda_j } ) ( \bm y\cdot\bm a_{\lambda_i})-(\bm x\cdot\bm y ) ( \bm a_{\lambda_i}\cdot\bm a_{\lambda_j})\end{aligned}\ ] ] for any two vectors @xmath132 and @xmath133 and we used the abbreviations @xmath134_{\lambda_2\lambda_3}+ ( \bm a_{\lambda_2}\cdot { \bm \rho})[\bm d_2\cdot \bm d_3]_{\lambda_1\lambda_3}+ ( \bm a_{\lambda_3}\cdot { \bm \rho})[\bm d_2\cdot \bm d_3]_{\lambda_1\lambda_2}\nonumber\\ & & + 2(\bm a_{\lambda_1}\cdot{\bm \rho})(\bm a_{\lambda_2}\cdot \bm d_2)(\bm a_{\lambda_3}\cdot\bm d_3 ) -(\bm d_2\cdot{\bm \rho})[\bm
d_3\cdot\bm a_{\lambda_2}]_{\lambda_1\lambda_3 } -(\bm d_3\cdot{\bm \rho})[\bm d_2\cdot\bm a_{\lambda_3}]_{\lambda_1\lambda_2}\ , , \nonumber\\ & & \nonumber\\ & & u_2=(\bm a_{\lambda_1}\cdot { \bm \rho})[\bm d_1\cdot \bm d_2]_{\lambda_2\lambda_3}- ( \bm a_{\lambda_2}\cdot { \bm \rho})[\bm d_1\cdot \bm d_2]_{\lambda_1\lambda_3}+ ( \bm a_{\lambda_3}\cdot { \bm \rho})[\bm d_1\cdot \bm d_2]_{\lambda_1\lambda_2}\nonumber\\ & & + 2(\bm a_{\lambda_2}\cdot{\bm \rho})(\bm a_{\lambda_1}\cdot \bm d_2)(\bm a_{\lambda_3}\cdot\bm d_1 ) -(\bm d_1\cdot{\bm \rho})[\bm d_2\cdot\bm a_{\lambda_3}]_{\lambda_1\lambda_2 } -(\bm d_2\cdot{\bm \rho})[\bm d_1\cdot\bm a_{\lambda_1}]_{\lambda_2\lambda_3}\,,\nonumber\\ & & \nonumber\\ & & u_3=(\bm a_{\lambda_1}\cdot { \bm \rho})[\bm d_1\cdot \bm d_3]_{\lambda_2\lambda_3}+ ( \bm a_{\lambda_2}\cdot { \bm \rho})[\bm d_1\cdot \bm d_3]_{\lambda_1\lambda_3}- ( \bm a_{\lambda_3}\cdot { \bm \rho})[\bm d_1\cdot \bm d_3]_{\lambda_1\lambda_2}\nonumber\\ & & + 2(\bm a_{\lambda_3}\cdot{\bm \rho})(\bm a_{\lambda_1}\cdot \bm d_3)(\bm a_{\lambda_2}\cdot\bm d_1 ) -(\bm d_1\cdot{\bm \rho})[\bm d_3\cdot\bm a_{\lambda_2}]_{\lambda_1\lambda_3 } -(\bm d_3\cdot{\bm \rho})[\bm d_1\cdot\bm a_{\lambda_1}]_{\lambda_2\lambda_3}\,,\nonumber\\ & & \nonumber\\ & & u_4=(\bm a_{\lambda_1}\cdot \bm d_2)[\bm d_1\cdot \bm d_3]_{\lambda_2\lambda_3}+ ( \bm a_{\lambda_1}\cdot \bm d_3)[\bm d_1\cdot \bm d_2]_{\lambda_2\lambda_3}- ( \bm a_{\lambda_1}\cdot \bm d_1)[\bm d_2\cdot \bm d_3]_{\lambda_2\lambda_3}\nonumber\\ & & + 2(\bm a_{\lambda_1}\cdot \bm d_1)(\bm a_{\lambda_2}\cdot \bm d_2)(\bm a_{\lambda_3}\cdot \bm d_3 ) + ( \bm d_1\cdot \bm d_2)[\bm d_3\cdot ( \delta_{(2)}\bm a_{\lambda_2}- \delta_{(3)}\bm a_{\lambda_3 } ) ] \nonumber\\ & & -(\bm d_1\cdot \bm d_3)[\bm d_2\cdot ( \delta_{(2)}\bm a_{\lambda_2}- \delta_{(3)}\bm a_{\lambda_3})]-(\bm d_2\cdot \bm d_3)[\bm d_1\cdot ( \delta_{(2)}\bm a_{\lambda_2}+ \delta_{(3)}\bm a_{\lambda_3})]\ , .\end{aligned}\ ] ] equations ( [ s2final ] ) and ( [ s3final ] ) are valid for an arbitrary vector potential @xmath31 and they are our starting point for further consideration .
the simplest case @xmath135 , where @xmath136 is a constant vector perpendicular to @xmath30 , corresponds to the so - called crossed field with the electric field @xmath137 and the magnetic field @xmath138 , @xmath139 .
in this case there is no energy and momentum transfer from the field so that @xmath140 and @xmath141 ( the momenta @xmath5 and @xmath6 of the final photons are parallel to the momentum @xmath4 of the initial photon ) .
note that in this case the expansion of the amplitudes over @xmath142 starts with the terms of order of @xmath143 while the linear terms in @xmath142 vanish .
we have checked that for a crossed field our amplitudes agree with those found in @xcite .
below , we consider the external field to be a monochromatic plane wave with the frequency @xmath9 . in this case
, the helicity amplitude @xmath62 has the form @xmath144 after the integration over the variable @xmath107 . in the above equation
the prime means the summation over odd numbers @xmath145 and where @xmath146 denotes the invariant amplitude of the photon splitting process with absorption of @xmath145 laser photons . using the usual fermi golden rule
, we obtain for the photon splitting rate @xmath147 @xmath148 where @xmath149 , @xmath150 is defined in eq .
( [ nu ] ) , and @xmath151 is the azimuth angle of the vector @xmath74 in the plane perpendicular to @xmath152 .
we remind the reader of the fact that @xmath74 is a two - dimensional vector perpendicular to @xmath152 and @xmath4 .
besides , @xmath153 coincides with the differential of the solid angle of the vector @xmath5 in the frame where @xmath154 .
it is also useful to express the rate @xmath155 in terms of the polar angle @xmath156 ( or of the polar angle @xmath157 ) . using the expressions ( [ theta0 ] ) valid for any @xmath83
we obtain the relations @xmath158 as a result , we find from eq .
( [ w ] ) that @xmath159 ^ 2}\ , .\end{aligned}\ ] ] in the general case of an elliptically polarized and monochromatic laser wave , the vector potential reads @xmath160 where @xmath161 .
generally speaking , the vectors @xmath162 and @xmath163 are not parallel to the unit vectors @xmath73 and @xmath164 introduced in eq .
( [ varepsilon ] ) , respectively , as there is an azimuth asymmetry of the photon splitting amplitude for an elliptically polarized laser wave ( we point out that @xmath73 has been chosen to be parallel to @xmath74 ) .
however , in a circularly polarized laser field it is @xmath165 , then the photon splitting rate is symmetric with respect to the azimuth angle @xmath151 of the vector @xmath74 , and the vector @xmath162 can be directed along @xmath73 .
the photon splitting rate can be converted into a cross section according to the relation @xmath166 with @xmath167 being the laser photon flux and @xmath168 .
note that this cross section has a restricted meaning because in the presence of a strong field it still depends on the laser flux through the parameters @xmath169 . in terms of the parameters @xmath169 and @xmath16 , the parameter @xmath15 introduced in eq .
( [ chieta ] ) has the form @xmath170 .
the integral over @xmath107 in eqs .
( [ s2final ] ) and ( [ s3final ] ) can be taken analytically for an elliptically polarized laser wave
. however , the result is very cumbersome even in the case of linear laser polarization .
this is why we limit ourselves to the case of a circularly polarized laser field . in this case
the answer is essentially simpler ( but not simple ) .
thus we set @xmath171 , @xmath172 and direct @xmath162 along @xmath73 and @xmath163 along @xmath164 ( for positive helicity ( p.h . ) ) or @xmath163 along @xmath173 ( for negative helicity ( n.h . ) ) . only the case of positive helicity of the laser field
will be considered below and the amplitudes for negative helicity of the laser field polarization can be obtained by means of the relation @xmath174 , where @xmath175 denotes the helicity opposite to @xmath176 .
now , by using the substitution rule of eqs .
( [ subs ] ) and by carrying out the change of variables @xmath177 and @xmath178 in eq .
( [ s2final ] ) , we obtain for the nonzero contributions of @xmath179 the following result @xmath180\right\}\nonumber\\ & & \times\mbox{e}^{i\vartheta}\left[2i\left(\mbox{e}^{i\vartheta } - \frac{\sin\vartheta}{\vartheta}\right)+ \,\frac{\sin\vartheta}{u(1-u)}\right]\,;\quad \vartheta=2u(1-u)s\eta\nu \ , .\end{aligned}\ ] ] in order to write @xmath181 in a compact form , we introduce the following definitions @xmath182 in terms of these functions , the contribution @xmath181 has the form @xmath183\nonumber\\ & & \times \sum_{j=-3}^{3}\ , \mbox{e}^{-i(n+j)\phi_0}\,j_{n+j}(z)b_{j,\sigma_1\sigma_2 \sigma_3}\ , , \end{aligned}\ ] ] where @xmath184 are ordinary bessel functions . the coefficients @xmath185 are expressed via the functions @xmath186 introduced in eq .
( [ not ] ) and they are presented in the appendix .
the expressions ( [ m2circular ] ) and ( [ m3circular ] ) are valid for any value of the parameters @xmath16 and @xmath187 . as it is clear from eq .
( [ m3circular ] ) , the photon splitting amplitudes are not equal to zero for any number of absorbed laser photons .
this is because the angle between the two final photons is not zero ( @xmath188 ) and the conservation of the projection @xmath189 of the total angular momentum does not imply any selection rule .
note that the expansion of the amplitude @xmath190 for a fixed @xmath145 ( see eqs .
( [ m2circular ] ) and ( [ m3circular ] ) ) , contains in general all odd powers @xmath191 of the parameter @xmath15 ( or , equivalently , @xmath192 ) starting from @xmath193 . the terms with @xmath194 in @xmath195 correspond to rescattering processes of absorption and emission of laser photons with a net absorption of @xmath145 laser photons . in the next section we consider the asymptotics of the amplitudes @xmath179 and @xmath181 in some limiting cases
various limiting cases of the photon splitting amplitudes for a monochromatic plane wave are studied in this section .
we first consider the case of both small @xmath16 and @xmath15 , @xmath13 and @xmath14 .
then , the leading contribution to the amplitudes is determined by the ratio @xmath196 @xcite . for @xmath197 , the nonzero amplitudes in the leading approximation
are given by the linear terms in @xmath15 in eqs.([m2circular ] ) and ( [ m3circular ] ) and correspond to @xmath198 .
we obtain @xmath199 these amplitudes are nothing but the usual photon - photon scattering amplitudes for small mandelstam variables @xmath200 , @xmath201 and @xmath202 ( see , e.g. , @xcite ) . for @xmath203 ,
the leading contributions are proportional to @xmath204 and correspond to both @xmath198 and @xmath205 .
the nonzero amplitudes in the leading approximation are given by @xmath206 these results are in agreement with those obtained in @xcite .
we proceed now to the region of parameters @xmath13 and @xmath15 fixed . in this case
the leading terms of the amplitudes in a circularly polarized laser field contain all powers of @xmath15 but only for @xmath198 and @xmath205 they still are not equal to zero .
the reason is the following : at @xmath13 , in a frame where @xmath207 , the vectors @xmath5 and @xmath6 are almost parallel to @xmath4 and anti - parallel to @xmath30 . since the clockwise polarized laser field contains photons with projection of the total angular momentum @xmath208 then , due to the conservation of @xmath189 , only the values @xmath198 and @xmath205 are allowed and only the amplitudes @xmath209 , @xmath210 , @xmath211 and @xmath212 are different zero . for these amplitudes
we found in the leading approximation with respect to @xmath16 : @xmath213\big\ } + { \cal g}(1)+{\cal g}(-\nu_3)\ , ; \nonumber\\ & & { } \nonumber\\ & & m_{1,---}=\frac{ime^3\chi^3}{\sqrt{2}\pi^2}\int\negthickspace\negthickspace\int\limits_0^{\infty } \negthickspace\negthickspace\int { ds_1\,ds_2\,ds_3}\,\mbox{e}^{-i(s+\psi_0)}\big\{-\frac{i(z_2+z_3)}{s^2\chi^2}\nonumber\\ & & + s\left[z_1z_2z_3-\nu_2\nu_3z_1t_1 ^ 2+\nu_2z_3+\nu_3z_2-\nu_2\nu_3(1+t_1 ^ 2)\right]\big\ } + { \cal g}(-\nu_2)+{\cal g}(-\nu_3)\ , ; \nonumber\\ & & { } \nonumber\\ & & m_{3,+--}=\frac{ime^3\chi^3}{\sqrt{2}\pi^2}\int\negthickspace\negthickspace\int\limits_0^{\infty } \negthickspace\negthickspace\int { ds_1\,ds_2\,ds_3}\,\mbox{e}^{-i(s+\psi_0)}\,\nonumber\\ & & \times s\left[z_1z_2z_3-\nu_2\nu_3z_1t_1 ^ 2+\nu_2z_3t_2 ^ 2+\nu_3z_2t_3 ^ 2 + 2\nu_2\nu_3t_1t_2t_3\right]\,.\end{aligned}\ ] ] here we introduced the following quantities : @xmath214\mbox{e}^{-i(s+\psi_1)}\ , ; \nonumber\\ & & z_1=2x_1(x_2\nu_2+x_3\nu_3)+2x_2x_3 - 1\ , , \quad z_2=2x_1(-x_2\nu_2+x_3\nu_3)-2x_2x_3+\nu_2\ , , \nonumber\\ & & z_3=2x_1(x_2\nu_2-x_3\nu_3)-2x_2x_3+\nu_3\,,\quad x_j=\frac{s_j } { s}\,,\quad t_j=1 - 2x_j\,,\nonumber\\ & & \psi_0=\frac{4}{3}\chi^2s^3[x_2 ^ 2x_3 ^ 2 + 2x_1x_2x_3(\nu_2x_2+\nu_3x_3)+x_1 ^ 2(\nu_2x_2-\nu_3x_3)^2]\,,\nonumber\\ & & \psi_1=\frac{4}{3}\nu^2\chi^2u^2(1-u)^2s^3\,.\end{aligned}\ ] ] the amplitude @xmath210 is obtained from @xmath209 by the replacement @xmath215 .
equations ( [ m3as ] ) are independent of the parameter @xmath16 and are valid for arbitrary values of @xmath15 .
the linear terms of the expansion of these amplitudes with respect to @xmath15 vanish .
then , the expansion starts with the terms proportional to @xmath204 that coincide with eqs .
( [ n1 ] ) . as it is known , both the electromagnetic invariants @xmath216 and @xmath217 are equal to zero for a plane wave , @xmath218 being the tensor of the electromagnetic field .
the euler - heisenberg lagrangian depends on these two invariants but not on @xmath15 which in our case has the form @xmath219
. therefore only the terms ( [ n1 ] ) can be obtained by means of this effective lagrangian but not the total answer ( [ m3as ] ) .
thus , the applicability of the euler - heisenberg lagrangian to the calculation of the photon splitting amplitudes is restricted by the conditions @xmath13 and @xmath14 .
it is worth noting that the region of applicability of the small-@xmath15 asymptotics is very narrow .
for instance , at @xmath220 the difference between the leading terms ( [ n1 ] ) and the exact result ( [ m3as ] ) is already around @xmath221 .
we illustrate the importance of high - order terms in @xmath15 in fig .
[ fig : chi ] where we show the total photon splitting rate @xmath222 at intermediate values of the parameter @xmath15 .
the solid line is obtained by using the amplitudes ( [ m3as ] ) , the dashed line by employing eq .
( [ n1 ] ) . at @xmath223
it is very difficult to measure the energy of the final photons with the accuracy necessary to distinguish between the process of photon splitting with one - photon absorption and three - photon absorption .
however , there is another possibility to distinguish these two processes which is based on the use of the angular distribution of the final photons . at @xmath207 ,
the final photons are almost parallel to the initial one so that @xmath224 where @xmath225 and @xmath226 .
it follows from eq .
( [ w1 ] ) that @xmath227 using the expressions ( [ m3as ] ) , we have tabulated the ratio @xmath228 the result of this tabulation at @xmath229 is shown in fig .
[ angle ] .
it can be seen that @xmath230 noticeably differs from unity at angles @xmath231 such that @xmath232 .
it is interesting to compare the amplitudes of photon splitting in a circularly polarized laser field with those in a linearly polarized field . in the latter case for @xmath13 and
fixed @xmath15 , the leading terms of the amplitudes are not zero for all odd @xmath145 .
this is due to the fact that a linearly polarized laser beam contains photons with both @xmath208 and @xmath233 . in this sense
the situation reminds us of the properties of the polarization operator in the laser field @xcite . in the case of a linearly polarized laser field
it is convenient to perform the calculations of the amplitudes in terms of the linear polarization of photons .
since the vectors @xmath5 and @xmath6 are almost parallel to @xmath4 , the azimuth asymmetry disappears .
therefore , we can choose @xmath234 and we have @xmath235 with @xmath236 .
the polarization of each photon is parallel either to @xmath73 or to @xmath164 so that our amplitudes coincide with the coefficients @xmath58 in eq .
( [ rvm ] ) . note that @xmath75 and thus @xmath164 is a pseudo - vector
. therefore , the nonzero amplitudes in the collinear approximation are those which either do not contain or contain twice the indices @xmath237 , namely , @xmath238 , @xmath239 , @xmath240 , and @xmath241 . putting @xmath242 equal to zero in eqs .
( [ s2final ] ) and ( [ s3final ] ) and performing the integral over @xmath107 , we obtain the leading contributions of the amplitudes at @xmath13 for a linearly polarized laser field @xmath243\bigg\ } + { \widetilde{\cal g}}_n(1)+{\widetilde{\cal g}}_n(-\nu_2)+{\widetilde{\cal g}}_n(-\nu_3 ) \ , ; \nonumber\\ & & { } \nonumber\\ & & m_{n,122}=\frac{i^{\frac{n+3}{2}}\sqrt{2}me^3\chi^3}{\pi^2}\int\negthickspace\negthickspace\int\limits_0^{\infty } \negthickspace\negthickspace\int { ds_1\,ds_2\,ds_3}\,\mbox{e}^{-i(s+\psi_0)}\bigg\{-\frac{iz_1{\cal d}_n^{(1)}(\psi_0)}{2s^2\chi^2}\nonumber\\ & & -s{\cal d}_n^{(2)}(\psi_0)\nu_2\nu_3(1-t_1)\big[(1+t_1)z_1 + 1-t_2t_3\big]\bigg\ } + { \widetilde{\cal g}}_n(1 ) \ , ; \nonumber\\ & & { } \nonumber\\ & & m_{n,212}=\frac{i^{\frac{n+3}{2}}\sqrt{2}me^3\chi^3}{\pi^2}\int\negthickspace\negthickspace\int\limits_0^{\infty } \negthickspace\negthickspace\int { ds_1\,ds_2\,ds_3}\,\mbox{e}^{-i(s+\psi_0)}\bigg\{-\frac{iz_2{\cal d}_n^{(1)}(\psi_0)}{2s^2\chi^2}\nonumber\\ & & + s{\cal d}_n^{(2)}(\psi_0)\nu_3(1-t_3)\big[(1+t_3)z_2-\nu_2(1-t_1t_2)\big]\bigg\ } + { \widetilde{\cal g}}_n(-\nu_2 ) \,,\end{aligned}\ ] ] where @xmath244 \mbox{e}^{-i(s+\psi_1)}{\cal d}_n^{(1)}(\psi_1)\ , , \nonumber\\\end{aligned}\ ] ] and the other notations are presented in eq . ( [ psi ] ) .
the amplitudes @xmath241 are obtained from @xmath240 by the replacement @xmath215 . in fig .
4 we show the photon splitting rate @xmath245 in a linearly polarized laser field as a function of the number @xmath145 of laser photons absorbed .
the rate has been calculated from the above amplitudes ( [ ml ] ) with @xmath229 .
the figure shows that , as the number @xmath145 increases , the photon splitting rate decreases quite rapidly and in a non - perturbative way .
the leading contributions to the amplitudes at @xmath14 have the form [ see also eq .
( [ n1 ] ) ] @xmath246 note that at @xmath13 there is a simple relation between the amplitudes of the photon splitting in a crossed field and those calculated in a linearly polarized laser field @xcite .
the amplitudes in a crossed field depend only on the parameter @xmath15 while those in a laser field depend only on @xmath15 in the leading approximation for @xmath13 . in this case
, it can easily be shown that @xmath247 therefore , due to parseval s identity , we have latexmath:[\[\sum_n|m_{n,\lambda_1\lambda_2\lambda_3}^{(\text{lin})}(\chi)|^2=\int_0^{2\pi}\frac{d\phi}{2\pi } for the same reason , in the case of a circularly polarized laser field the following relation holds latexmath:[\[\label{circritus } \sum_{n,\,\sigma_i}|m_{n,\sigma_1\sigma_2\sigma_3}^{(\text{circ})}(\chi)|^2=\sum_{\sigma_i } indices @xmath176 denote the helicities of the external photons .
we emphasize that eq .
( [ circritus ] ) is not valid without the summation over @xmath176 .
this is because the rotational symmetry around the vector @xmath152 is absent in a crossed field ( see discussion below eq .
( [ u ] ) ) , as is the conservation of the projection of the total angular momentum @xmath189 .
therefore , it is impossible to obtain the amplitudes @xmath249 in a circularly polarized field from those in a crossed field by using any fourier transformation as it was done in eq .
( [ mn ] ) .
now , at @xmath250 the asymptotics ( [ m3as ] ) obtained for the circularly polarized field read @xmath251+\frac{8\gamma^2(1/3)}{5\gamma(2/3)}(1-\nu_3^{1/3})\bigg\}\ , ; \nonumber\\ & & { } \nonumber\\ & & m_{1,---}={\cal n}_c\bigg\ { \int_0 ^ 1du\,u\int_0 ^ 1\frac{dw}{{\tilde \psi}^{4/3}}\big [ 4(z_2+z_3){\tilde \psi } + z_1z_2z_3-\nu_2\nu_3z_1t_1 ^ 2 \nonumber\\ & & + \nu_2z_3+\nu_3z_2-\nu_2\nu_3(1+t_1 ^ 2)\big ] -\frac{8\gamma^2(1/3)}{5\gamma(2/3)}(\nu_2^{1/3}+\nu_3^{1/3})\bigg\}\ , ; \nonumber\\ & & { } \nonumber\\ & & m_{3,+--}={\cal n}_c\int_0 ^ 1du\,u\int_0 ^ 1\frac{dw}{{\tilde \psi}^{4/3 } } \big(z_1z_2z_3-\nu_2\nu_3z_1t_1 ^ 2+\nu_2z_3t_2 ^ 2+\nu_3z_2t_3 ^ 2 + 2\nu_2\nu_3t_1t_2t_3\big)\,,\end{aligned}\ ] ] where @xmath252 and @xmath253 and @xmath254 are expressed via @xmath255 as in eq .
( [ psi ] ) .
analogously , at @xmath250 we find from eq .
( [ ml ] ) for a linearly polarized laser field @xmath256 + \frac{8\gamma^2(1/3)}{5\gamma(2/3)}(1-\nu_2^{1/3}-\nu_3^{1/3})\bigg\rangle \ , ; \nonumber\\ & & { } \nonumber\\ & & m_{n,122}={\cal n}_n\bigg\langle \int_0 ^ 1du\,u\int_0 ^ 1\frac{dw}{{\tilde \psi}^{4/3 } } \big\{4z_1{\tilde\psi}-\nu_2\nu_3(1-t_1)\big[(1+t_1)z_1 + 1-t_2t_3\big]\big\ } + \frac{8\gamma^2(1/3)}{5\gamma(2/3)}\bigg\rangle \ , ; \nonumber\\ & & { } \nonumber\\ & & m_{n,212}={\cal n}_n\bigg\langle \int_0 ^ 1du\,u\int_0 ^ 1\frac{dw}{{\tilde \psi}^{4/3}}\big\{4z_2{\tilde\psi } + \nu_3(1-t_3)\big[(1+t_3)z_2-\nu_2(1-t_1t_2)\big]\big\ } \nonumber\\ & & -\frac{8\gamma^2(1/3)}{5\gamma(2/3)}\nu_2^{1/3}\bigg\rangle \ , , \quad { \cal n}_n=\frac{me^3(6\chi)^{1/3}\mbox{e}^{i\pi/3}\gamma(1/6)\gamma(n/2 - 1/6)}{2^{5/6}72 \pi^{5/2}\gamma(n/2 + 7/6)}\ , .\end{aligned}\ ] ] the dependence on @xmath145 in these expressions is contained in the coefficients @xmath257 which at large @xmath145 rapidly decrease as @xmath258 . instead , the dependence of the amplitudes on the parameter @xmath15 , both for a linearly and a circularly polarized laser field , is weak and proportional to @xmath259 .
the structure of the photon splitting amplitudes at large @xmath16 and fixed @xmath187 is rather interesting because the function @xmath260 introduced in eq .
( [ m2circular ] ) and which enters the contributions @xmath261 has the form @xmath262 - 4\right.\\ & \left.-4i\lambda\int_0^\infty d\vartheta\frac{\sin\vartheta\ , \mbox{e}^{i\lambda\vartheta}}{\vartheta}\log\left[\frac{1+\xi^2}{1+\xi^2(1-\sin^2\vartheta/\vartheta^2)}\right]\right\}\ , , \end{split}\ ] ] with @xmath263 , and therefore contains different powers of @xmath264 as well as some functions of @xmath265 . however , when we add the corresponding asymptotics of @xmath266 which is the leading term in the limiting case under discussion , any dependence on @xmath16 and @xmath192 is cancelled out and we obtain the high - energy asymptotics of photon - photon scattering in vacuum @xcite . therefore , analogously to the case of photon - photon scattering in vacuum ,
the high - energy asymptotics of the amplitudes of photon splitting in a laser field is independent of the energy of the initial photon as well as of the parameter @xmath192 .
more precisely , these statements are valid for any value of @xmath192 which satisfies the inequality @xmath267 .
we now discuss the possibility of experimental observation of photon splitting in a laser field .
it is natural to consider the principal scheme of the experiment which was used in the first successful observation of photon splitting in a coulomb field @xcite . in this experiment
a tagged photon beam was used and the energy of each final photon was measured by a calorimeter .
it was necessary to confirm two - photon events , and the distance @xmath268 between photons in the calorimeter was about one centimeter . assuming a distance @xmath269 between the calorimeter and the interaction region between tagged photons and laser field , then the angles @xmath225 and @xmath270 have to be larger than @xmath271 . at @xmath207
the final photons are almost parallel to the initial one so that @xmath224 .
it follows from eq .
( [ theta23 ] ) that in this limit @xmath272 if , for instance , @xmath273 , then for @xmath198 @xmath274 for the planned photon energy of the x - fel at desy @xcite , with @xmath275 kev , the latter condition becomes @xmath276 gev and is not restrictive .
then , we can essentially reduce @xmath277 or @xmath8 , or increase @xmath278 .
the planned intensity of the x - fel is about @xmath279 corresponding to @xmath280 and @xmath281 . in this case , the amplitudes of photon splitting coincide with those of photon - photon scattering .
therefore , it is convenient to choose @xmath282 mev for which @xmath283 .
this value of @xmath16 corresponds to the threshold of electron - positron pair production and therefore the background initiated by this process will be strongly suppressed .
besides , at @xmath283 the cross section of photon - photon scattering is almost maximal and is given by @xmath284 ( see e. g. @xcite on pg .
now , from eq .
( [ flux ] ) we find @xmath285 .
the x - fel pulses will have a duration of about 100 fs and the interval between pulses will be about 93 ns .
if we take , for example , a flux of @xmath286 tagged photons per second @xcite , we obtain about two photon splitting events per hour .
however , the very small laser beam - size makes this estimation essentially smaller . in consequence ,
the possibility of experimental observation of photon splitting in a laser field becomes problematic when using the x - fel and the scheme of ref .
@xcite .
we discuss now the possibility of using a very strong optical laser as that described in @xcite . in that case @xmath287 , and the laser intensity will be of order of @xmath288 corresponding to @xmath289 . for a value of the angle @xmath290
we obtain , e. g. for @xmath291 , approximately @xmath292 as the upper limit for @xmath8 . for these parameters
we find @xmath293 and @xmath294 .
therefore , we have to use eq .
( [ m3as ] ) but not the asymptotics ( [ n1 ] ) because high - order corrections in @xmath15 are important .
using the numerical results shown in fig .
[ fig : chi ] we obtain @xmath295 at @xmath294 . though the value of @xmath222 is very large , the duration of each pulse is @xmath296 fs and the laser repetition rate is only @xmath297 hz @xcite . for a total flux of @xmath286 photons per second @xcite and for an electron bunch revolution frequency equal to @xmath297 mhz in the accelerator , we obtain an average number of tagged photons per pulse equal to @xmath298 . if all the photons pass through the laser pulse , we obtain again approximately two events per hour . however , the transverse beam size of a focused optical laser is even smaller than that of an x - fel and the reduction due to the tagged and laser beams overlapping is larger .
our negative conclusion about the possibility of observing photon splitting in a laser field when using the scheme discussed is mainly due to the short duration and small size of the laser beam as well as to the small flux of tagged photons . on the one hand ,
short and focused laser beams are closely coupled with large intensity of the laser pulses . on the other hand ,
the use of tagged photons , which essentially restricts the statistics , is needed to suppress the large background due to many other processes ( double compton scattering with the experimental equipment like collimators , bremsstrahlung accompanied by @xmath299 pair production and so on ) . of course
, we can not exclude the existence of other schemes of the experiment that avoid the difficulties pointed out in this section .
in conclusion , we have derived compact expressions for the amplitudes of the photon splitting process in a plane wave of arbitrary form , intensity and polarization ( see eqs .
( [ s2final ] ) and ( [ s3final ] ) ) . in the simpler case of a monochromatic circularly polarized laser field ,
the amplitudes have the form of three - fold integrals ( see eqs .
( [ m2circular ] ) and ( [ m3circular ] ) ) .
moreover , we analyzed in detail various asymptotics of the amplitudes in the case of a monochromatic wave with linear or circular polarization .
we have demonstrated that the angular distribution of the probabilities corresponding to different numbers of absorbed photons is noticeably different ; this permits us the distinction between these processes for the case where the laser photon energy is less than the accuracy of the final photon energy measurement . the photon splitting amplitudes depend on the two lorentz - invariant parameters @xmath16 and @xmath15 .
for arbitrary values of these parameters , the amplitudes both in the case of a circularly and a linearly polarized laser field are not equal to zero for any ( odd ) number @xmath145 of absorbed laser photons .
this is because the angle between the two final photons is not zero and the conservation of the projection @xmath189 of the total angular momentum ( in the case of a circularly polarized laser field ) does not imply any restriction .
however , we have shown that in the limit of small @xmath16 the leading contribution to the amplitudes for a circularly polarized laser comes from the terms with @xmath198 and @xmath205 , while for a linearly polarized laser field the leading contributions contain the terms with any odd @xmath145 .
finally , we have demonstrated that the scheme previously employed in the experiment of photon splitting in an atomic field together with new generation x - fel sources and strong optical lasers is not suitable for the corresponding experiment in a laser field .
therefore , it is necessary to find another more appropriate experimental configuration to observe photon splitting in a laser field .
we would like to thank k. z. hatsagortsyan for valuable discussions .
a. i. m. gratefully acknowledges the great hospitality and the financial support he has received during his visit at max - planck - institute for nuclear physics . the work was supported in part by the rfbr under grants 0502 - 16079 and 06 - 02 - 04018 .
here we present the coefficients @xmath185 used in eq .
( [ m3circular ] ) .
we introduce the abbreviations @xmath300 then , the nonzero coefficients for the contribution @xmath301 are @xmath302\nonumber\\ & & -4\rho_n^2\xi x_1\left(\frac{x_2 g_2}{\nu_3}+\frac{x_3 g_3}{\nu_2}+i d_1\right)-\frac{4i\xi } { s}(g_2+g_3)\ , , \nonumber\\ & & \nonumber\\ b_{0,+++}&=&-4\rho_n\xi^2\bigg[g_1^*\left(\frac{x_2g_2}{\nu_3}+\frac{x_3g_3}{\nu_2}\right ) + ig_1^*d_1 + h^*d_1+\left(\frac{\nu_2x_2}{\nu_3}+\frac{\nu_3x_3}{\nu_2}\right)|h|^2\bigg ] \nonumber\\ & & + \frac{\rho_n^3x_1(4x_2x_3 - 1)}{\nu_2\nu_3}+\frac{4i\rho_n } { s}\left(\frac{x_2}{\nu_3}+\frac{x_3}{\nu_2}\right)\ , , \nonumber\\ & & \nonumber\\ b_{-1,+++}&=&\frac{\rho_n^2\xi}{\nu_2\nu_3}\left[(4x_2x_3 - 1)g_1^*+ih^*\right ] \ , , \end{aligned}\ ] ] the nonzero coefficients for the contribution @xmath303 are @xmath304\ , , \nonumber\\ & & \nonumber\\ b_{0,++-}&=&4\rho_n\xi^2\left[g_2\left(\frac{x_2 g_1^*}{\nu_3}-x_1g_3^*\right ) -ig_2d_3^ * -\nu_2hd_3^*+\nu_2\left(\frac{x_2}{\nu_3}+x_1\nu_3\right)|h|^2\right]\nonumber\\ & & -\frac{\rho_n^3x_3(4x_1x_2 - 1)}{\nu_2\nu_3}+\frac{4i\rho_n } { s}\left(x_1-\frac{x_2}{\nu_3}\right)\ , , \nonumber\\ & & \nonumber\\ b_{-1,++-}&=&4\xi^3\left[-g_1^*g_2g_3^*+\nu_2(\nu_3g_1^*-g_3^*)|h|^2+\nu_3g_2d_3^{*2}-i\nu_2\nu_3h(h^{*2}+d_3^{*2})\right ] \nonumber\\ & & + \frac{4\rho_n^2\xi x_3}{\nu_2 } \left(x_1
g_3^*-\frac{x_2}{\nu_3}g_1^*+id_3^*\right ) + \frac{4i\xi } { s}(g_1^*+g_3^*)\ , , \nonumber\\ & & \nonumber\\ b_{-2,++-}&=&\frac{4\rho_n\xi^2x_3}{\nu_2}\left(g_1^*g_3^*-\nu_3d_3^{*2}\right)\ , .\end{aligned}\ ] ] finally , the nonzero coefficients for the contribution @xmath305 are @xmath306 \,,\nonumber\\ & & \nonumber\\ b_{-2,+--}&=&4\rho_n\xi^2\bigg [ x_1g_2^*g_3^*- g_1^*\left(\frac{x_2}{\nu_3}g_2^*+\frac{x_3}{\nu_2}g_3^*\right)+ig_1^*d_1^*+ig_2^*d_3^ * + ig_3^*d_2^*\nonumber\\ & & + x_1\nu_2\nu_3d_1^{*2}+\frac{x_2\nu_2}{\nu_3}d_2^{*2}+\frac{x_3\nu_3}{\nu_2}d_3^{*2}+\nu_2d_1^*d_2^ * + \nu_3d_1^*d_3^ * -d_2^*d_3^ * \bigg]\ , , \nonumber\\ & & \nonumber\\ b_{-3,+--}&=&4\xi^3 ( g_1^*g_2^*g_3^*+\nu_2\nu_3g_1^*d_1^{*2}-\nu_3g_2^*d_3^{*2}-\nu_2g_3^*d_2^{*2 } + 2i\nu_2\nu_3d_1^*d_2^*d_3^ * ) \ , .\end{aligned}\ ] ] the coefficients for the amplitude @xmath307 can be obtained from @xmath303 by the replacement @xmath308 .
the coefficients for the amplitudes @xmath309 , @xmath310 and @xmath311 can be obtained with the help of the substitutions @xmath312 where the replacement @xmath313 means that after complex conjugation , it is also necessary to change the sign of the terms containing @xmath314 in eqs .
( [ + + + ] ) , ( [ + + - ] ) and ( [ + ] ) .
feynman diagram for the photon splitting amplitude given by eq .
( [ mg ] ) .
this diagram corresponds to the furry representation and the thick line denotes the electron propagator ( green s function ) in the laser field .
the wavy lines symbolize the external photons .
the diagram with the permutation @xmath315 has to be added.,width=302 ] in a circularly polarized laser field in units @xmath316 as a function of @xmath15 at @xmath13 .
the solid line is obtained by using the amplitudes ( [ m3as ] ) , the dashed line is based on the small-@xmath15 asymptotics ( [ n1 ] ) . the rate is averaged over the polarization of the initial photon , summed up over the polarizations of the final ones and integrated over @xmath317 and @xmath151.,width=377 ] in a linearly polarized laser field in units @xmath316 as a function of the number @xmath145 of laser photons absorbed .
the rate is calculated from the amplitudes in eqs .
( [ ml ] ) with @xmath229 , is averaged over the polarization of the initial photon , summed up over the polarizations of the final ones and integrated over @xmath317 and @xmath151.,width=377 ] v. n. baier , a. i. milstein , and r. zh .
shaisultanov , zh .
fiz . * 90 * , 1141 ( 1986 ) [ jetp 63 , 665 ( 1986 ) ] ; phys .
* 77 * 1691 ( 1996 ) ; zh .
111 * , 52 ( 1997 ) [ jetp 84 , 29 ( 1997 ) ] .
g. mourou , t. tajima , and s. v. bulanov , rev .
phys . * 78 * , 309 ( 2006 ) . v. n. baier , a. i. milstein , and v. m. strakhovenko , zh .
. fiz . * 69 * , 1893 ( 1975 ) [ sov . phys .
jetp * 42 * , 961 ( 1976 ) ] . | photon splitting due to vacuum polarization in a laser field is considered . using an operator technique
, we derive the amplitudes for arbitrary strength , spectral content and polarization of the laser field .
the case of a monochromatic circularly polarized laser field is studied in detail and the amplitudes are obtained as three - fold integrals .
the asymptotic behavior of the amplitudes for various limits of interest are investigated also in the case of a linearly polarized laser field . using the obtained results ,
the possibility of experimental observation of the process is discussed . |
the high energy transient explorer ( space.mit.edu/hete/ ) hete-2 is dedicated to the study of gamma - ray bursts ( grbs ) .
hete-2 is currently the only grb detector capable of localizing and disseminating grb coordinates in near real - time .
low - energy emission during and shortly after a grb ( @xmath1 hr ) potentially holds the key to significant progress in understanding the central engine of grbs and could provide valuable clues to their progenitors @xcite .
the hete-2 detection of grb 010921 together with data from the interplanetary network ( ipn ) provided the first hete-2 localization which has resulted in the detection of an optical afterglow .
although the afterglow was relatively bright , early observations of the error box failed to reveal any candidate afterglows because of source confusion with its bright host galaxy ( @xmath2 ) @xcite .
spectroscopy of the host galaxy performed with the palomar 200-in telescope four weeks after the burst indicates a redshift of @xmath3 @xmath4 @xcite .
on 21 september 2001 at 05:15:50.56 ut ( september 21.21934 ut ) the fregate instrument detected a bright grb ( trigger 1761 ) .
grb 010921 had a duration of @xmath5 s in the 8 - 85 kev band , a peak flux of @xmath6 and a fluence of @xmath7 @xcite .
for a spatially flat frw cosmology ( @xmath8 = 0.3 , @xmath9 = 0.7 , and @xmath10 = 65 km s@xmath11 mpc@xmath11 ) the measured redshift implies an equivalent isotropic energy release of @xmath12 erg . this inferred energy would be within the range suggested by @xcite for the standard energy reservoir , @xmath13 erg , suggesting the opening angle of the gamma - ray jet in grb 010921 is modest .
the burst was also detected in the wide field x - ray monitor wxm x detector but was outside the field - of - view of the wxm y detector @xcite .
because the grb was not well localized in the y direction , a location was not distributed with the real - time trigger .
approximately five hours after the grb , ground analysis gave a long narrow error box ( @xmath14 ) centered at @xmath15 ( j2000.0 ) @xcite .
_ ulysses _ and _ bepposax _ also detected grb010921 .
these detections resulted in an ipn triangulation approximately 15 hours after the burst .
the ipn location for the burst is an annulus centered at @xmath16 ( j2000.0 ) , with radius of @xmath17 .
the combined hete-2/ipn data resulted in an error box with an area of a @xmath18 box with corners at @xmath19 , @xmath20 , @xmath21 , @xmath22 , ( j2000.0 ) @xcite .
although no coordinates were distributed with the real - time hete-2 trigger , the @xmath23 , 0.11-m lotis telescope @xcite observed the position of the error box during routine sky patrol on september 21 ut at 21.255 ut and 21.417 ut , only 52 min and 4.75 hours after the grb .
the lotis telescope consists of four lens / camera systems directed toward a common @xmath24 field - of - view .
two of the lotis cameras are equipped with clear filters , one with a standard cousins @xmath25 filter , and one with the standard johnson @xmath26 filter . during sky patrol observations the lotis telescope obtained 50 s exposures at each position on the sky .
grb010921 was also observed with the sloan digital sky survey s ( sdss ; york et al .
2000 ) photometric telescope ( pt ) , which is located at apache point observatory ( apo ) in sunspot , new mexico .
the pt is an @xmath27 , 0.5-m telescope equipped with @xmath28 filters .
the sdss is designed to be on the @xmath28 photometric system described in fukugita et al .
( 1996 ) which is an @xmath29 system where flat spectrum objects ( @xmath30 ) have zero colors .
however , the current photometric calibration of the pt may differ from this system by at most a few percent ( stoughton et al . 2002 ) , and we therefore denote the pt photometric magnitudes by @xmath31 . the single site @xmath32 ccd camera has a @xmath33 field - of - view . on september 22 ut beginning at 22.108 ut ( 21.33 hours after the grb ) 200 s exposures were taken in each of the @xmath28 filters , and on september 23 ut , beginning at 23.224 ut 200 s and 400 s exposures were taken in each of the same filters ( lamb et al .
observations on both nights covered the entire improved hete-2 error box ( ricker et al .
2001 , hurley et al . 2001 ) .
since the grb exposures were unusually long , the afterglow was near the sky level in most of the images , and conditions were cloudy at the time , magnitude errors are uncharacteristically large . the center of the ipn error box for grb 010921 was added to the super - lotis sky patrol table and observations of the location began shortly after nightfall on september 22.128 ut .
super - lotis is an @xmath34 , 0.6-m boller&chivens telescope at kitt peak near tucson , arizona .
its focal plane array is a loral @xmath32 ccd camera covering @xmath35 field - of - view .
the telescope system is fully automated and is capable of responding to a real - time grb trigger within 30 seconds .
a total of twenty 50-second exposures were obtained during the first epoch . super - lotis re - observed the field on september 22 and 23 ut beginning at 22.267 , 23.128 , and 23.269 ut .
each of these observations also consisted of twenty 50-second exposures .
for these observations , no astronomical filter was used . on september 22.781
the tautenburg schmidt telescope began observations of the error box in the johnson @xmath36-band .
this telescope , located at tautenburg , germany , is an f/2 , 1.34-m aperture telescope equipped with a site @xmath32 ccd as the focal plane array .
its field - of - view is @xmath37 .
thirty - eight 120-second exposures were acquired .
the same field was re - observed on october 25 and eight 120-second exposures were acquired . during these exposures ,
the telescope was dithered to allow a better treatment of bad pixels . on november 17 ut ( 56 days after the burst ) ,
the usnofs 1.0-m telescope at flagstaff , arizona , observed the grb010921 area to determine the brightness of the host galaxy in @xmath38 and @xmath39-bands .
the usnofs is an f/7.3 telescope equipped with @xmath40 filters ; its focal plane arrays are a site / tektronix @xmath32 ccd with @xmath41 field - of - view . for these observations , eight 600-second exposures per filter were acquired and coadded .
early observations of the error box of grb 010921 with large aperture telescopes found no evidence of an optical afterglow to the limit of the dposs plates , @xmath42 @xcite .
however , follow - up observations by @xcite resulted in the report of a fading source which displayed the characteristic decay behavior of a grb afterglow .
the coordinates of the suspected afterglow are @xmath43 ( j2000.0 ) .
following the reported detection of an optical afterglow candidate , we searched our images for the optical afterglow . in the lotis clear- and @xmath26-band images taken 52 min after the burst we found no optical transient ( ot ) brighter than @xmath44 @xmath45 and @xmath26 @xmath46 .
the shutter for the @xmath25-band camera failed to open . in the second epoch of lotis clear- and @xmath26-band images taken at september 21.417 ut ( 4.75 hours after the burst ) we found no ot to the same limiting magnitude .
analysis of the sdss pt images obtained on september 22 ut reveals the optical afterglow at @xmath47 , and @xmath48 , and gives @xmath49 upper limits of @xmath50 and @xmath51 .
analysis of the sdss pt images obtained on september 23 ut reveals the optical afterglow at @xmath52 and @xmath53 , and gives @xmath49 upper limits of @xmath54 , @xmath55 , and @xmath56 .
the completeness upper limits are the magnitudes at which the probability of detection is 100% . because of variable extinction over the field due to clouds , the 50% detection limits are more than a magnitude fainter than the completeness limits .
the top panel of figure 1 ( @xmath57 to @xmath58 ) shows the sdss pt images in @xmath28 obtained on september 22 ut of the grb010921 afterglow area ( marked by a circle ) reported by price et al .
( 2001 ) .
we searched for the ot in the super - lotis data by co - adding twenty images obtained during each epoch , two epochs per night . in the first co - added image , which was taken on september 22 ut beginning at 22.128 ut , we detect the optical afterglow at @xmath59 @xmath60 at 15 @xmath61 above noise level .
the same analysis applied to the second epoch data ( september 22.267 ut ) detects the optical afterglow at @xmath59 @xmath62 at 12 @xmath61 above the noise level .
panels @xmath63 and @xmath64 of figure [ fig : images ] show the co - added images with the afterglow indicated by an arrow .
we also searched for an ot in the data set obtained on september 23 ut and find no afterglow to the 10 @xmath61 noise limit of @xmath59 @xmath65 .
the tautenburg data taken on september 22 ut and october 25 ut were co - added to search for the grb 010921 afterglow .
panel @xmath66 of figure [ fig : images ] show the location of the afterglow indicated by an arrow .
calibrating the afterglow against secondary standards @xcite , the brightness of the afterglow is estimated as @xmath67 .
the host galaxy is marginally detected at the 5 @xmath61 level in the second epoch image ( october 25 ut at 25.770 ) at @xmath68 ( panel @xmath69 of figure [ fig : images ] ) .
the usnofs deep imaging data was analyzed to obtain the brightness of the host galaxy .
the host galaxy was clearly visible in the coadded image 56 days after the burst and we determine its brightness to be @xmath70 and @xmath71 .
table 1 summarizes the magnitudes and upper limits in the various filters from all the above observations .
we transformed the fluxes measured in various filters to the @xmath25 filter by normalizing a @xmath72 power - law spectrum @xcite to the effective wavelength and flux of each data point , and then plotting the point at the effective wavelength of the @xmath25-band .
this calculated values are listed in the last column of table 1 .
figure [ fig : lightcurve ] shows the resulting light curve of the optical afterglow of grb 010921 .
we fit the data with a power law decay plus a constant host galaxy flux , @xmath73 .
we obtain a best fit decay index of @xmath74 applying a host galaxy magnitude of @xmath75 from the usno measurements performed 56 days after the burst .
this is a typical value for an optical afterglow prior to the jet break , which is predicted to take place @xmath76 130 days after the grb based on the observed energetics of this burst @xcite .
the optical afterglow reported by @xcite was @xmath77 on september 22 and significantly fain.ter on september 23 ( we place a limit on @xmath25 of roughly 20.5 ) .
we attempt to constrain the early - time power law decay by extrapolating the best - fit power - law decay model back to the lotis upper limit of @xmath78 on september 21.256 .
figure [ fig : lightcurve ] shows that the early - time lotis upper limits are inconsistent with an unchanging decay index from @xmath79 min to @xmath80 h. this may suggest that the optical emission peaked at a magnitude fainter than the lotis limiting magnitude , ( perhaps similar to the afterglow of grb 970508 ) or that the slope changed between the observations , as suggested in the case of grb 991208 @xcite .
complex light curve shapes at very early times have been observed ( e.g. , grb970805 ) and can be explained in terms of a distribution of lorentz factors produced by the central engine @xcite and the complex evolution of multi - component shock emission in a relativistic fireball @xcite .
our early time observations show that afterglow behavior can change quickly within hours after the burst . with rapid localizations we can probe the transition from the prompt emission phase to the subsequent unfolding of the canonical afterglow phase .
grb010921 only provided an upper limit 52 minutes after outburst , but hete-2 triggers should eventually allow us to obtain simultaneous flux measurements .
robotic telescopes like lotis and super - lotis are well suited to finding this early - time emission in response to a near real - time localization of the grb by hete-2 .
support for lotis and super - lotis is provided by nasa ( s-03975 g and s-57797f ) under the auspices of the u. s. department of energy by university of california lawrence livermore national laboratory ( w-7405-eng-48 ) .
the sloan digital sky survey ( sdss ) is a joint project of the university of chicago , fermilab , the institute for advanced study , the japan participation group , the johns hopkins university , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , princeton university , the u.s .
naval observatory , and the university of washington .
apache point observatory , site of the sdss telescopes , is operated by the astrophysical research consortium ( arc ) .
funding has been provided by the alfred p. sloan foundation , the sdss member institutions , nasa , the nsf , u.s .
department of energy , monbukogabusho , and the max planck society .
k. hurley is grateful for _ ulysses _ support ( jpl 958059 ) and for hete support ( mit - sc - r-293291 ) .
castro - tirado , a. j. , et al 2001 , a&a , 370 , 398 .
djorgovski , s. g. et al . 2001 , grb circular network , 1108 fox , d. w. et al .
2001 , grb circular network , 1099 frail , d. a. et al .
2001 , , 562 , l55 fukugita , m. , ichikawa , t. , gunn , j. e. , doi , m. , shimasaku , k. , & schneider , d. p. 1996
, , 111 , 1748 henden , a. et al . 2001a , grb circular network , 1100 henden , a. et al . 2001b , grb circular network , 1101 hurley , k. et al .
2001 , grb circular network , 1097 kobayashi , s. 2000 , , 545 , 807 kulkarni , s. r. et al .
2001 , in preparation .
lamb , d. q. et al .
2001 , grb circular network , 1125 m ' esz ' aros , p. 2001 , science , 291 , 79 park , h. , ables , e. , barthelmy , s. d. , bionta , r. m. , ott , l. l. , parker , e. l. , & williams , g. g. 1998 , , 3355 , 658 price , p. a. et al .
2001 , grb circular network , 1107 rees , m. j. , @xmath81 meszaros , p. 1998
, , 496 , l1 ricker , g. et al . 2001 , grb circular network , 1096 ricker , g. et al . 2001 , in preparation .
stoughton , c. et al .
2002 , , submitted and accepted .
york , d. g. et al .
2000 , , 120 , 1579 ccccccc sep 21.219 & 0.000 & hete-2 & grb010921 & & & + sep 21.255 & 0.036 & lotis & clear & @xmath82 s & @xmath83 & @xmath84 + sep 21.255 & 0.036 & lotis & @xmath26 & @xmath82 s & @xmath85 & @xmath86 + sep 21.417 & 0.198 & lotis & clear & @xmath82 s & @xmath83 & @xmath84 + sep 21.255 & 0.198 & lotis & @xmath26 & @xmath82 s & @xmath85 & @xmath86 + sep 22.108 & 0.889 & sdss pt & @xmath87 & @xmath88 s & @xmath89 & @xmath90 + sep 22.112 & 0.893 & sdss pt & @xmath91 & @xmath92 s & @xmath93 & 19.85 + sep 22.115 & 0.896 & sdss pt & @xmath94 & @xmath92 s & @xmath95 & 19.44 + sep 22.118 & 0.899 & sdss pt & @xmath96 & @xmath92 s & @xmath97 & 19.49 + sep 22.121 & 0.902 & sdss pt & @xmath98 & @xmath92 s & @xmath99 & @xmath100 + sep 22.128 & 0.909 & super - lotis & clear & @xmath101 s & @xmath102 & 19.24 + sep 22.267 & 1.048 & super - lotis & clear & @xmath101 s & @xmath103 & 19.54 + sep 22.781 & 1.562 & tautenburg & @xmath39 & @xmath104 s & @xmath105 & 20.05 + sep 23.128 & 1.909 & super - lotis & clear & @xmath101 s & @xmath106 & @xmath107 + sep 23.244 & 2.025 & sdss pt & @xmath87 & @xmath88 s & @xmath108 & @xmath109 + sep 23.249 & 2.030 & sdss pt & @xmath91 & @xmath110 s & @xmath111 & 21.45 + sep 23.255 & 2.036 & sdss pt & @xmath94 & @xmath110 s & @xmath112 & 21.44 + sep 23.260 & 2.041 & sdss pt & @xmath96 & @xmath110 s & @xmath113 & @xmath114 + sep 23.266 & 2.047 & sdss pt & @xmath98 & @xmath110 s & @xmath115 & @xmath116 + sep 23.269 & 2.050 & super - lotis & clear & @xmath101 s & @xmath117 & @xmath107 + oct 25.770 & 34.551 & tautenburg & @xmath39 & @xmath118 s & @xmath119 & 21.67 + nov 17.7 & 56.700 & usnofs & @xmath38 & @xmath120 s & @xmath121 & 21.93 + nov 17.7 & 56.700 & usnofs & @xmath39 & @xmath120 s & @xmath122 & 21.78 + | we present multi - instrument optical observations of the high energy transient explorer ( hete-2)/interplanetary network ( ipn ) error box of grb 010921 .
this event was the first gamma ray burst ( grb ) localized by hete-2 which has resulted in the detection of an optical afterglow . in this paper
we report the earliest known observations of the grb010921 field , taken with the 0.11-m livermore optical transient imaging system ( lotis ) telescope , and the earliest known detection of the grb010921 optical afterglow , using the 0.5-m sloan digital sky survey photometric telescope ( sdss pt ) .
observations with the lotis telescope began during a routine sky patrol 52 minutes after the burst .
observations were made with the sdss pt , the 0.6-m super - lotis telescope , and the 1.34-m tautenburg schmidt telescope at 21.3 , 21.8 , and 37.5 hours after the grb , respectively .
in addition , the host galaxy was observed with the usnofs 1.0-m telescope 56 days after the burst .
we find that at later times ( t > 1 day after the burst ) , the optical afterglow exhibited a power - law decline with a slope of @xmath0 .
however , our earliest observations show that this power - law decline can not have extended to early times ( t < 0.035 day ) . |
the insight that clustered star formation may be the dominant mode for star formation has grown over the last years .
the form of the true distribution of stellar masses within these clusters , of the stellar initial mass function ( imf ) , has been a subject of debate for a long time .
the evolution of the stars , unresolved binaries , and the dynamical evolution of the clusters complicates the observational efforts to extract the imf .
unfortunately the most promising objects , very young stellar clusters ( age @xmath3 3 myr ) , are often still embedded in their natal cloud - again aggravating observations .
nevertheless the distribution of stars in young clusters seems to be fairly well described by a multi power - law function with a slope or index ( @xmath4 ) of 2.35 ( the so - called salpeter value ) for stars with a mass larger than @xmath5 @xcite . the initial mass function ( imf ) , @xmath6 where @xmath7 is the number of stars in the mass interval @xmath8 .
several observations find the salpeter value ( @xmath9 ) for a large variety of conditions @xcite .
it is therefore useful to describe the _ stellar imf _ with an invariant , multi - power law form @xcite , @xmath10 we refer to this form as the standard or canonical stellar imf because this form fits the luminosity function of galactic - field and cluster stars below 1 @xmath11 and also represents young populations above 1 @xmath11 @xcite . as pointed out by @xcite
though , significant uncertainties remain in the determination of the imf to the point that the case can also be made that a single form of the imf may not exist . in view of this , the ansatz made here and elsewhere is to propose the hypothesis of an invariant standard or canonical imf ( eq .
[ kroupa - imf ] ) and to test if the variation of the observed imf can be understood to be the result of astrophysical effects ( obscuration , stellar evolution , stellar multiplicity ) , dynamical effects ( mass segregation , stellar evaporation and ejections ) , stochastic effects ( finite @xmath12-sampling from the imf ) and the construction of composite populations ( addition of many different clusters ) .
similarly , the embedded cluster mass function ( ecmf ) has been found to be well - described by at least one power - law , @xmath13 where @xmath14 is the number of embedded clusters in the mass interval @xmath15 , @xmath16 and @xmath15 is the cluster mass in stars .
the observational evidence points to a possibly universal form of the ecmf : @xcite find a slope @xmath17 in the solar neighbourhood for clusters with masses between 50 and 1000 @xmath11 , while @xcite find @xmath18 for @xmath19 in the smc and lmc , and @xcite find @xmath20 for @xmath21 in the antennae galaxies .
@xcite discovered that @xmath22 best reproduces the observed correlation between the brightest young cluster and the galaxy - wide star - formation - rate for a large sample of late - type galaxies .
as already mentioned by @xcite and discussed in more detail by @xcite , the composite or integrated galaxial stellar imf ( igimf ) is obtained by summing up the stellar imfs contributed by all the star clusters that formed over the age of a galaxy , @xmath23 where @xmath24 is the ecmf and @xmath25 is the stellar imf in a particular cluster within which the maximal mass of a star is @xmath26 .
@xmath27 (= @xmath28 , taurus - auriga - type `` clusters '' ) is the minimal cluster mass , while the maximal cluster mass , @xmath29 , depends on the galaxy - wide star - formation rate , sfr @xcite .
a critical function entering this description is thus @xmath30 . assuming the stellar imf to be a continuous distribution function , this mass of the most massive star in an embedded cluster with the total mass @xmath15 in stars is given by @xmath31 with @xmath32 since there exists exactly one most massive star in each cluster , and neglecting statistical variations . here
@xmath33 = 0.01
@xmath11 is the minimal fragmentation mass and @xmath34 is the measured maximal stellar mass limit @xcite . on combining eqs .
[ eq : mmmc ] and [ eq : meclmc ] the function @xmath35 is quantified by @xcite and in
[ sec : maxstmass ] .
this is the analytical ( `` ana '' ) maximum - stellar - mass cluster - mass relation which incorporates the fundamental stellar upper mass limit ( noted by the leading superscript `` l '' ) of @xmath36 . later - on other maximum - stellar - mass cluster - mass relations are indicated by different superscripts : `` ran '' , `` con '' and `` sort '' for the different monte - carlo sampling methods ( see [ se : mc ] ) and also `` u '' for the case without a fundamental stellar upper mass limit .
@xcite infer that a fundamental upper stellar mass limit , @xmath37 , appears to exist above which stars do not occur , unless @xmath38 , in which case no conclusions can be drawn based on the expected number of massive stars . as reviewed by @xcite , the existence of such a stellar upper mass limit has been further substantiated by @xcite and @xcite for a range of star clusters and ob associations .
we thus have , for each @xmath15 , the maximal stellar mass , @xmath39 , and with this information eq .
[ eq : igimf ] can be evaluated to compute the igimf .
@xcite find the igimf , when evaluated to the highest cluster masses , to be significantly steeper than the stellar imf , and @xcite extend the analysis to a time varying ecmf by noting that @xmath40 increases with the star - formation rate of a galaxy .
they show the igimf to be not only steeper than the stellar imf , but also to depend on galaxy type .
the implications of these findings are rather significant for the supernova rate @xcite and for the chemical evolution of galaxies @xcite .
but these results remain not without a challenge .
@xcite argues that there is no evidence of a relation @xmath41 .
this relation implies that many small , low - mass , star - forming events will not have the same combined imf as one major sf event of the same mass .
thus , according to @xcite , @xmath42 clusters each with a mass of @xmath43 would provide a combined imf that differs from that of one cluster with a mass of @xmath44 by being underrepresented in stars with a mass above about @xmath45 .
the contrary , often voiced view is that stellar masses sample the imf purely statistically such that @xmath42 clusters containing 50 stars ( on average 20 @xmath11 ) will give the same combined imf as one cluster containing @xmath46 stars @xcite . with this contribution
we demonstrate conclusively that the purely statistical notion is false , and that the stellar imf is sampled to a maximum stellar mass that correlates with the cluster mass . therewith we affirm the results obtained by @xcite and @xcite , and we also attain useful insights into the process of star - cluster formation . in section [ sec : maxstmass ] our monte - carlo procedure
is described and the maximal - star - mass cluster - mass relation is derived , while in [ sec : montec ] the monte - carlo experiment is applied to the igimf and the results are presented .
a discussion with conclusions is available in [ sec : concs ] .
over the past 20 years several studies investigated a possible connection of the maximum stellar mass in a cluster and the mass of the cluster because such a relation , if it were to exist , would allow important insights into the star - formation process .
@xcite compared the properties of molecular clouds with the spatial distribution of the associated stellar population .
he found that more massive and dense clouds favour the formation of massive stars and fitted the following formula to the observations , @xmath47 he re - evaluated @xcite this equation with more recent data and applied it directly to the cluster mass instead of the cloud mass , @xmath48 this correlation is plotted as a _ dotted line _ in figs .
[ fig : mvsm3 ] and [ fig : mvsm2 ] .
@xcite proposed a model for the origin of bound galactic clusters where the luminosity of the stars from a @xcite imf overcomes the binding energy of a molecular cloud core .
the star - formation efficiency then discriminates between bound clusters and ob associations .
he derived an analytical estimate for a relation between the maximal star mass and the cluster mass from statistical considerations regarding the appearance of stars with various masses from the miller - scalo - imf , @xmath49}{1-f[a_{1}(\log m_{\rm max } -c_{2})]},\ ] ] with @xmath50 and @xmath51 , @xmath52 , @xmath53 , @xmath54 and @xmath55 . in figs .
[ fig : mvsm3 ] and [ fig : mvsm2 ] this relation is shown as a _ short - dashed line_. on the other hand , using a single power - law imf with a @xcite slope , @xcite solved eqs .
[ eq : mmmc ] and [ eq : meclmc ] , assuming @xmath56 , with the result , @xmath57 which is shown as a _ long - dashed line _ in figs .
[ fig : mvsm3 ] and [ fig : mvsm2 ] .
@xcite and @xcite numerically studied star - formation in clusters using their smooth - particle - hydrodynamics ( sph ) code . here a turbulent molecular cloud fragments hierarchically into smaller subunits .
when the density of a clump gets higher than a critical value , it is replaced by a so - called sink particle which only lets matter in but not out .
these sink particles form the final stellar cluster by interactions and mergings and are called stars at the end of the simulation .
this hierarchical cluster formation scenario leads to the relation , @xmath58 and is shown as a _ short - dashed - dotted line _ in figs .
[ fig : mvsm3 ] and [ fig : mvsm2 ] .
there eq . [ eq : mvmb ] is normalised to @xmath59 for a cluster of @xmath60 @xcite .
it should be noted here that these simulations do not include magnetic fields , stellar feedback and stellar mergers , all of which are believed to be of great importance for star - formation .
@xcite started with similar assumptions as @xcite but included a physical upper limit for the stellar mass , @xmath61 , while solving eqs .
[ eq : mmmc ] and [ eq : meclmc ] . due to @xmath62 and using the standard multi - power law imf ( eq .
[ kroupa - imf ] ) , the result can not be written out analytically but the equations have to be solved numerically .
the result is plotted as a _
thick solid line _ in figs .
[ fig : mvsm3 ] and [ fig : mvsm2 ] . in a broader statistical analysis ,
@xcite calculated the probabilities that the observed upper mass limits in several clusters and ob associations in the mw , lmc and smc come from a sample with a fundamental upper mass limit , @xmath63 , or not .
they conclude that an upper mass limit between 120 and 200 @xmath11 is the most likely explanation for the observed maximum masses . in oder to do so , they calculated the expectation value for the maximum mass if a number @xmath12 of stars is randomly sampled from an imf .
this yields the equation , @xmath64^{n}dm_{\rm ecl}.\ ] ] integrating this numerically yields the _ long - dash - short - dashed line _ in figs .
[ fig : mvsm3 ] and [ fig : mvsm2 ] assuming @xmath65 .
, in dependence of the cluster mass , @xmath15 , both in logarithmic units . the @xcite - relation ( eq .
[ eq : la03 ] ) is shown as a _ dotted line _ , while the @xcite - relation ( eq .
[ eq : el83 ] ) is the _ short - dashed line_. the @xcite - relation ( eq .
[ eq : el00 ] ) is the _ long - dashed line_. the result ( eq . [ eq : mvmb ] ) from numerical sph star - formation simulations @xcite is plotted as a _ short - dashed - dotted line _ and as a _ triangle _ for a specific model from @xcite , while the _ long - dashed - short - dashed line _ marks the expectation values ( eq . [ eq : oc05 ] ) from @xcite and the _ thick solid line _ shows the semi - analytical model ( eq . [ eq : mmaxvsmecl ] ) of @xcite for the standard imf ( eq . [ kroupa - imf ] ) .
thin solid line _ shows the identity relation , were a `` cluster '' consists only of one star .
[ fig : mvsm3],width=302 ] all the above mentioned investigations suggest that the cluster mass indeed appears to have a limiting influence on the stellar masses within it .
however , it would be undisputed that a stochastic sampling effect from the imf must be present when stars form . to investigate the possible existence of a maximal stellar mass in clusters statistically and to confirm or rule out
if such a relation is purely the result of the random selection from an imf , three monte - carlo experiments are conducted : * pure random sampling ( random sampling ) , * mass - constrained random sampling ( constrained sampling ) , * mass - constrained random sampling with sorted adding ( sorted sampling ) . for each ,
two possibilities are probed : stars are sampled from the imf without a maximal mass ( @xmath66 is adopted to be @xmath67 in the unlimited case . ] ) , or their masses are limited by @xmath68 . for the random sampling experiment , @xmath69 clusters are randomly taken from a cluster distribution with a power - law index @xmath70 .
the clusters contain between 12 stars and @xmath71 stars .
the relevant distribution function is the embedded - cluster star - number function ( ecsnf ) , @xmath72 which is the number of clusters containing @xmath73 $ ] stars .
each cluster is then filled with @xmath12 stars randomly from the standard imf ( eq .
[ kroupa - imf ] ) without a mass limit , or by imposing the physical stellar mass limit , @xmath74 .
the stellar masses are then summed to get the cluster mass , @xmath75 .
note that such a cluster distribution gives an embedded cluster mass function ( ecmf ) that is virtually identical to eq .
[ eq : ecmf ] ( fig .
[ fig : ecmf ] ) .
the resulting distribution of maximum stellar masses is plotted in the @xmath26 , @xmath76 plane ( fig .
[ fig : contourran ] ) as contour lines , to show the overall distribution ( for a more detailed discussion see
[ se : comp ] ) . as in this method
the higher cluster masses are only very scarcely sampled , a second method is used to evaluate the mean of maximal masses in more detail . in order to do so the cluster star numbers @xmath77 12 to @xmath78 are divided into 10 logarithmically equally - spaced values .
each of these numbers is then filled 10000 times with stars form the imf , keeping only the mass of the most massive star for each cluster .
the mean maximum mass for every bin then defines , @xmath79 with a limit of @xmath68 and @xmath80 in the unlimited case .
clusters using eq .
[ eq : ecmf ] ( _ dashed line _ ) and as described in
[ se : rasa ] ( _ solid line _ ) .
the slopes are virtually the same ( @xmath81 ) . only for very small cluster masses
does the _ solid line _ deviate due to the underlying standard imf because small-@xmath12 clusters can nevertheless have masses @xmath82 if their constituent stars happen to be massive .
this accounts for the turn down below 10 @xmath11 and the surplus in the range 10 - 150 @xmath11 .
[ fig : ecmf],width=302 ] in this case @xmath69 clusters are randomly taken from the ecmf ( eq . [ eq : ecmf ] ) between 5 @xmath11 ( the minimal , taurus - auriga - type , star - forming `` cluster '' counting @xmath83 15 stars ) and @xmath2 ( an approximate maximum mass for a single stellar population that consists of one metallicity and age , * ? ? ?
* ; * ? ? ?
* ) and again with @xmath84 .
note that @xmath85 because the ecsnf and the ecmf only differ by a nearly - constant average stellar mass ( fig .
[ fig : ecmf ] ) . then stars are taken randomly from the standard imf and added until they reach or just surpass the respective cluster mass , @xmath15 .
afterwards the clusters are searched for their maximum stellar mass ( plotted as contours in fig .
[ fig : contourcon ] , see [ se : comp ] for discussion ) .
again the sampling of high cluster masses is very poor .
therefore , the cluster masses from @xmath28 to @xmath2 are divided into 10 logarithmically equally - spaced values . then each of these 10 cluster masses is filled 10000 times with stars form the imf until their combined mass reach or just surpass the cluster mass and only the maximum star mass is recorded .
the average @xmath26 values for each of the 10 cluster masses define the relation , @xmath86 in the limited case and @xmath87 in the unlimited case . for the third method again @xmath69 clusters are randomly taken from the ecmf ( eq . [ eq : ecmf ] ) between 5 @xmath11 and @xmath2 and with @xmath84 . however , this time the number @xmath12 of stars which are to populate the cluster is estimated from @xmath88 / @xmath89 , where @xmath90 is the average stellar mass for the standard imf ( eq . [ kroupa - imf ] ) between 0.01 @xmath11 and 150 @xmath11 .
these stars are added to give @xmath91 , @xmath92 such that @xmath93 . if @xmath94 in this first step , an additional number of stars , @xmath95 , is picked randomly from the imf , where @xmath95 = @xmath96 ) / @xmath97 .
again these stars are added to obtain an improved estimate of the desired cluster mass , @xmath98 again with @xmath93 .
when @xmath99 surpasses @xmath15 , it is checked whether the sum is closer to @xmath75 when the last star ( the most massive one ) is subtracted or not . if @xmath99 is closer to @xmath15 with the last star this one is the most massive one for the cluster , otherwise it is the second last star . ,
@xmath100 and @xmath101 , then for @xmath102 , @xmath26 would be 10 @xmath11 , as 52 is closer to 50 than 42 .
if @xmath15 would be 45 @xmath11 , @xmath26 would be 5 @xmath11 .
] this procedure is followed and repeated until all clusters from the ecmf are filled. the contour plots of the most massive star for each cluster are shown in fig . [
fig : contoursort ] and discussed in [ se : comp ] .
again , for a more detailed analysis , 10 cluster masses are generated like in [ se : cosa ] but filled with stars in the sorted way described above .
the mean over every of the 10 cluster masses then yields the relation @xmath103 in order to exemplify the differences between the three sampling methods the following gedanken experiment may be considered : + a sample of 10 stars consists of 9 stars with 5 @xmath11 and one with 11 @xmath11 . for the random sampling , this would be a cluster with @xmath104 = 56 @xmath11 , with @xmath105 = 11 @xmath11 .
if , for the sorted sampling , the aimed - at cluster mass is 50 @xmath11 , the actual cluster mass would be 45 @xmath11 ( @xmath106 ) , because 45 is closer to 50 than 56 , and @xmath107 would be 5 @xmath11 . in the case of
constrained sampling the order would be important .
if the aimed - at cluster mass is 50 @xmath11 and the 11 @xmath11 star is among the first 9 stars , @xmath108 would be 51 ( @xmath109 ) and @xmath110 = 11 @xmath11 .
but if the 11 @xmath11 star is the tenth star , @xmath108 would be 45 @xmath11 and @xmath110 = 5 @xmath11 as in the sorted - sampling case because , 45 , rather than 56 , is closer to 50 @xmath11 . + figs .
[ fig : contourran ] , [ fig : contourcon ] and [ fig : contoursort ] plot the contour lines of the most massive stars of all the simulated clusters for random sampling ( fig .
[ fig : contourran ] ) , constrained sampling ( fig . [ fig : contourcon ] ) and sorted sampling ( fig .
[ fig : contoursort ] ) , all with the physical limit @xmath111 . as can be seen from fig .
[ fig : contourran ] , the random sample occupies the whole parameter space between the two extremes , which are either that nearly the whole cluster consists only of low - mass stars , or a single star accounts for the entire cluster mass ( @xmath112 ) .
such one - star - clusters would correspond to freak star - formation events such as is envisaged for a variable gas - equation - of - state - star - formation theory @xcite . for
sorted sampling the clusters are shifted more towards smaller stellar masses and never touch this line . while the constrained sampling lies in between these two extremes and still populates the parameter space up to the identity relation . ) , eg . @xmath113 .
the inner - most one ( near @xmath114 ) stands for @xmath115 clusters with this maximal star mass .
the lines in - between are in steps of 0.5 dex . here
the _ contour lines _ are the result of the monte - carlo simulations with random sampling .
the semi - analytic result ( eq . [ eq : mmaxvsmecl ] ) , @xmath116 , is the _ thick solid line_. mean values are shown as the _ thick dash - dotted line _ for random sampling ( @xmath117 , eq .
[ eq : mvsmranl ] ) , as the _ thin dashed line _ for constrained sampling ( @xmath118 , eq .
[ eq : mvsmconl ] ) and as the _ thin dotted line _ for sorted sampling ( @xmath119 , eq .
[ eq : mvsmsortl ] ) .
the identity relation @xmath120 is plotted as a _ thin solid line_. [ fig : contourran],width=340 ] but this time the _ contour lines _ are the result of the monte - carlo simulations with mass - constrained sampling , and @xmath121 is shown as a _ thick line _ , while @xmath122 and @xmath123 are _
@xmath124 is the _ thick solid line_. [ fig : contourcon],width=340 ] and [ fig : contourcon ] but this time the _ contour lines _ are the result of the monte - carlo simulations with sorted sampling with @xmath125 as a
_ thick line_. [ fig : contoursort],width=340 ] note that @xmath123 , shown in figs .
[ fig : contourran ] to [ fig : contoursort ] ( _ thick dotted line _ in fig .
[ fig : contoursort ] , _ thin _ in figs .
[ fig : contourran ] and [ fig : contourcon ] ) , is nearly identical to the semi - analytical estimate @xmath126 ( eq .
[ eq : mmaxvsmecl ] , _ thick solid line _ in figs .
[ fig : contourran ] to [ fig : contoursort ] ) .
the slight deviations are probably due to the stochastic element in the process to decide which is the most massive star in the sorted monte - carlo experiment ( see [ se : sosa ] ) .
the non - smoothness of the contour lines in the upper right edge of the figures is an effect of low - number sampling . here
only the outer - most contour line is present - indicating that there are only single events in this region of parameter space obtained with @xmath69 clusters , which have together about @xmath127 stars .
this number of clusters comprises the computational limit of the available hardware .
the agreement of the mean value @xmath128 and the semi - analytic result of @xcite , @xmath129 , is in principal not surprising .
the method of sorted adding of stellar masses in order to get the cluster mass corresponds to a monte - carlo integration of eqs .
[ eq : mmmc ] and [ eq : meclmc ] .
therefore the result should agree with the numerically integrated ( semi - analytical ) one .
another difference between the samplings lies in the average mean stellar masses in the clusters . to determine these ,
cluster masses from @xmath28 to @xmath2 are divided in 10 logarithmically equally - spaced values and each is filled 10000 times with stars from the imf in the three different ways described in [ se : rasa ] , [ se : cosa ] and [ se : sosa ] .
the mean stellar mass for each cluster is calculated by @xmath130 where @xmath12 is the number of stars in each cluster .
then for each cluster mass , @xmath131 ( @xmath132 ) , the mean of means is computed by @xmath133 with @xmath134 = 10000 .
the different average mean stellar masses are shown in fig .
[ fig : mean ] . for the random sampling
the mean is always about @xmath135 , as expected for the canonical imf between 0.01 and @xmath136 .
the other sampling methods have lower means for light clusters which rise up to @xmath135 for more massive ones .
this is a result of the limit which low - mass clusters impose on their stellar content . for random sampling
only the number of stars determines the cluster mass , in contrast to reality where the natal cloud - mass and the star - formation - efficiency rule the final cluster mass .
also shown in fig .
[ fig : mean ] is the average mean stellar mass for the analytical model ( eq . [ eq : mmaxvsmecl ] ) , which is given by @xmath137 where @xmath138 is eq .
[ kroupa - imf ] . ) . for random sampling ( _ dotted line _ ) the value is constant within the numerical noise around the expected value of @xmath135 . in the case of constrained sampling ( _ dashed - dotted line _ ) and sorted sampling ( _ long - dashed line _ ) it starts rather low for low mass clusters and rises up to expected value of @xmath135 .
the _ solid line _ shows the relation described by eq .
[ eq : mofmana ] .
note that the _ long - dashed line _ lies above the _ solid curve _ because the maximal stellar mass in the sorted - sampling algorithm is systematically higher than the analytical result ( fig .
[ fig : mvsm2 ] to [ fig : mvsmage ] below ) for cluster masses below @xmath139 .
[ fig : mean],width=302 ] table [ tab : clusters ] shows a non - exhaustive compilation of cluster masses and upper stellar masses for a number of mw and lmc clusters ( see appendix [ appendixa ] for more details ) .
these data are shown as dots with error - bars in figs .
[ fig : mvsm2 ] , [ fig : mvsm ] and [ fig : mvsmage ] .
the result of a cluster formation simulation by @xcite is shown as a large triangle .
.[tab : clusters ] empirical cluster masses , maximal star masses within these clusters and cluster ages from the literature .
[ cols="^,^,^,^,^,^ " , ] for the maximal stellar masses in these clusters the values within the papers are used whenever possible , which is for ngc2244 , ngc2264 , ngc6611 , m42 , @xmath140 ophiuchi , @xmath141 orionis , monoceros r2 , berkley 86 , trumpler 14/16 , arches and r136 . in the other cases ( ngc1333 , ngc2024 , ngc6530 , ic348 , serpens svs2 , taurus - auriga ) mass estimates are derived from the spectral types of the most luminous members using the spectral - type mass - relation from @xcite
because the @xcite single stellar evolution ( sse ) package is only calibrated for stellar models up to @xmath142 , additional fitting formulae have been developed for more - massive stars .
based on the @xcite models for 60 , 85 and 120 @xmath11 , functions for @xmath143 , @xmath144 and @xmath145 have been obtained : as long as the age , @xmath146 ( in myr ) , of the star is below @xmath147 , the main - sequence life - time , the mass - evolution can be described according to , @xmath148 for @xmath149 @xmath150 for both the parameters are @xmath151 and constants @xmath152 when @xmath146 is larger than @xmath153 the star is considered dead .
no remnant mass , @xmath154 or luminosity is assigned .
the resulting curves for a 120 , 85 , 60 and 50 @xmath11 star ( _ solid lines _ ) in comparison with the model data ( _ dotted lines _ ) are plotted in fig .
[ fig : fitmass ] .
-star is shown as a _
dashed line_. [ fig : fitmass],width=302 ] for @xmath155 the equation @xmath156 adequately captures the evolution . for @xmath149 , @xmath157 is used .
+ the parameters are @xmath158 and the constants @xmath159 the resulting @xmath154 fitting curves are plotted in fig .
[ fig : fitteff ] for the same masses as in fig .
[ fig : fitmass ] . .
as the @xmath154 for massive stars are rather similar , the lines in the plot have been shifted upwards as indicated in the plot .
[ fig : fitteff],width=302 ] the luminosity evolution is divided into three parts . for @xmath155 , @xmath160 for @xmath161 @xmath162 and finally for @xmath163 @xmath164 here the parameters are @xmath165 + l_{\rm jc},\\ \tau_{\rm b}&=&\frac{1\,\cdot\,10^{-6}}{\tau_{\rm bb } - ( \frac{\tau_{\rm bm}}{m_{\rm ini}})},\\ l_{\rm peak}&=&l_{\rm pa } \cdot \log_{10}\left(\frac{m_{\rm ini}}{l_{\rm pf}}\right ) + l_{\rm pc},\\ a_{4}&=&\frac{l_{\rm jump } - l_{\rm low}}{\tau_{\rm m } + dt - \tau_{b}},\\ b_{3}&=&l_{\rm low } - ( a_{4 } \cdot ( \tau_{\rm m } + dt)).\end{aligned}\ ] ] the constants are @xmath166 fig .
[ fig : fitlum ] shows the results of the luminosity - fitting formulae . .
[ fig : fitlum],width=302 ]
several different sets of theoretical models for stellar evolution of massive stars exist .
[ fig : compmass ] , [ fig : compteff ] and [ fig : complum ] compare the mass , @xmath154 and luminosity evolution of three different sets of models @xcite .
these models agree qualitatively on the compared stellar properties but show minor differences in the details . .
as the @xmath154 for massive stars are rather similar , the lines in the plot have been shifted ( as indicated in the plot ) in the following way : the upper - most by + 1.0 dex and the second one by + 0.5 dex .
the lowest plot has not been shifted .
[ fig : compteff],width=302 ] | we demonstrate that the mass of the most massive star in a cluster correlates non - trivially with the cluster mass . a simple algorithm according to which
a cluster is filled up with stars that are chosen randomly from the standard imf but sorted with increasing mass yields an excellent description of the observational data .
algorithms based on random sampling from the imf without sorted adding are ruled out with a confidence larger than 0.9999 .
a physical explanation of this would be that a cluster forms by more - massive stars being consecutively added until the resulting feedback energy suffices to revert cloud contraction and stops further star formation .
this has important implications for composite populations .
for example , @xmath0 clusters of mass @xmath1 will not produce the same imf as one cluster with a mass of @xmath2 .
it also supports the notion that the integrated galaxial imf ( igimf ) should be steeper than the stellar imf and that it should vary with the star - formation rate of a galaxy .
[ firstpage ] stars : formation stars : luminosity function , mass function galaxies : star clusters galaxies : evolution galaxies : stellar content galaxy : stellar content |
codimension one foliations on compact manifolds are well - studied . for instance , it s shown that a closed manifold @xmath2 has a smooth codimension one foliation if and only if the euler characteristic of @xmath2 is zero @xcite .
moreover , all leaves of a codimension one foliation without holonomy on a closed manifold are pairwise homeomorphic @xcite .
each codimension one foliation without holonomy on a closed manifold is either compact or minimal , and the base manifold is a fibration over a circle @xmath9 @xcite . since higher
codimension foliations are more flexible than codimension one foliations , most studies of higher codimension foliations are restricted to some classes ( e.g. compact foliations , compact hausdorff foliations ) .
for instance , compact foliations are studied in @xcite and compact hausdorff foliations are studied in @xcite . in particular , it s known that the leaf space of a continuous codimension two compact foliation @xmath10 of a compact manifold @xmath2 is a compact orbifold @xcite .
this implies that each continuous codimension two compact foliation of a compact manifold is @xmath0-closed ( i.e. the leaf class space is hausdorff ) .
on the other hand , there are non-@xmath0-closed compact foliations and non @xmath0-closed flows each of whose orbits is compact for codimension @xmath11 @xcite .
note that the following properties are equivalent for a flow ( resp .
a homeomorphism ) on a compact metric space @xcite : 1 ) @xmath0-closed ; 2 ) @xmath12-stable ; 3 ) of characteristic 0 ; 4 ) weakly almost periodic ; 5 ) lyapunov stable . in this paper , we study codimension two ( resp .
one ) @xmath0-closed foliations , which are generalizations of codimension two ( resp
. one ) compact foliations . to study @xmath0-closed foliations
, we generalize the characterization of hausdorff - ness for a compact foliation in @xcite into one of @xmath0-closedness for a pointwise almost periodic decomposition @xmath13 of a compact hausdorff space @xmath14 .
precisely , the following are equivalent : 1 ) the decomposition @xmath15 of closures of elements of @xmath13 is upper semi - continuous ; 2 ) the class space @xmath16 is hausdorff ; 3 ) the decomposition @xmath13 is @xmath0-closed .
moreover , we study codimension - one - like ( resp .
codimension - two - like ) @xmath0-closed foliations . using the above characterization of @xmath0-closedness , for @xmath17 ( resp .
@xmath18 ) , we show that the leaf class space of a codimension-@xmath19-like @xmath0-closed foliation on a compact manifold is a compact surface with conners ( resp .
@xmath20 $ ] or @xmath9 ) . in @xcite
, it is showed that each action of a finitely generated group @xmath21 on either a compact zero - dimensional space or a graph @xmath14 is pointwise recurrent if and only if it is @xmath0-closed ( i.e. the orbit class space @xmath22 is hausdorff , where the orbit class space @xmath22 is a quotient space @xmath23 defined by @xmath4 if @xmath6 ) . in this paper
, we also study codimension two dynamical systems ( e.g. surface homeomorphisms , flows in @xmath7-manifolds ) .
we show that each minimal set of an @xmath0-closed non - periodic toral homeomorphism @xmath24 is either @xmath25 or a finite disjoint union of essential circloids . moreover , a homeomorphism @xmath26 on a compact surface @xmath8 whose orientable genus is more than one , is @xmath0-closed if and only if it is periodic . for an @xmath0-closed flow @xmath27 on a connected compact @xmath7-manifold @xmath2 ,
one of the following statements holds : 1 ) the flow @xmath27 is minimal or identical ; 2 ) the orbit class space @xmath28 is either @xmath20 $ ] or @xmath29 and each interior point of @xmath28 is two dimensional ; 3 ) the set @xmath30 of periodic points is open dense and the flow is pointwise periodic ( i.e. @xmath31 , where @xmath32 is the disjoint union symbol ) ; 4 ) there is a two dimensional minimal set which is not a suspension of a circloid . in @xcite
, they have shown that if a continuous mapping @xmath33 of a topological space @xmath14 in itself is pointwise recurrent ( resp .
pointwise almost periodic ) then so is @xmath34 for each positive integer @xmath19 .
this result also holds for group - action cases ( see theorem 2.24 , 4.04 , and 7.04 @xcite ) .
we show that an analogous result holds for @xmath0-closed group - action .
indeed , a group acting on a compact hausdorff locally connected space is @xmath0-closed if and only if so is a finite index subgroup .
by a decomposition , we mean a family @xmath10 of pairwise disjoint nonempty subsets of a set @xmath14 such that @xmath35 .
let @xmath10 be a decomposition of a topological space @xmath14 .
for any @xmath36 , denote by @xmath37 or @xmath38 the element of @xmath10 containing @xmath39 . for a subset @xmath40 of @xmath14 ,
write the saturation @xmath41 .
a subset @xmath42 is saturated if @xmath43 . the class @xmath44 of an element @xmath38 is defined by @xmath45 .
then the set @xmath46 of classes is a decomposition , called the class decomposition .
denote by @xmath16 the quotient space , called the class space .
note @xmath47 for a subset @xmath48 of @xmath14 .
recall a point @xmath39 in @xmath14 is said to be of characteristic 0 @xcite if @xmath49 for any @xmath36 , where @xmath50 is its ( bilateral ) prolongation defined as follows :
the decomposition is said to be of characteristic 0 if so is each point of it .
the decomposition @xmath10 is upper semicontinuous ( usc ) if each element of @xmath10 is both closed and compact and for any @xmath52 and for any open neighbourhood @xmath53 of @xmath54 there is a saturated neighbourhood of @xmath54 contained in @xmath53 , is @xmath0-closed if @xmath55 is closed , and is pointwise almost periodic if the set of all closures of elements of @xmath10 also is a decomposition .
an pointwise almost periodic decomposition @xmath1 is weakly almost periodic in the sense of gottschalk w. h. if the saturation @xmath56 of closures of elements for any closed subset @xmath48 of @xmath14 is closed . notice that if @xmath10 is pointwise almost periodic then @xmath15 corresponds to the decomposition of closures of elements of @xmath1 . by the definition of almost periodicity ,
the set of compact foliations on compact manifolds is a proper subset of the set of pointwise almost periodic foliations . in @xcite
, the author has shown that the set of compact hausdorff foliations and of minimal foliations on compact manifolds is a proper subset of the set of @xmath0-closed foliations .
for a non - negative integer @xmath19 , a foliation @xmath1 is said to be codimension-@xmath19-like if all but finitely many leaf classes of @xmath1 are codimension @xmath19 closed submanifolds and the finite exceptions have codimensions each of which is more than @xmath19 .
note that the anosov alternative for transitive anosov flows @xcite implies that the ( un)stable foliation of a non - mixing transitive anosov flow on a closed manifold is a codimension - one - like @xmath0-closed foliation . by a group - action , we mean a continuous action of a topological group @xmath21 on a topological space @xmath14 . for a group - action @xmath21 , denote by @xmath57 the set of orbits of @xmath21 .
recall a group - action @xmath21 is @xmath0-closed if the set @xmath57 of orbits is an @xmath0-closed decomposition . in @xcite
, we have shown that each @xmath0-closed decomposition is pointwise almost periodic .
therefore @xmath21 is @xmath0-closed if and only if @xmath58 is closed . for a non - negative integer @xmath19 , a group - action @xmath21
is said to be codimension-@xmath19-like if all but finitely many orbit closures of @xmath59 are codimension @xmath19 closed connected submanifolds and the finite exceptions is connected subsets each of whose codimension is more than @xmath19 . by a flow , we mean a continuous @xmath60-action on a topological space @xmath14 . for
a flow @xmath27 on a smooth manifold @xmath2 , denote by @xmath61 the decomposition of @xmath2 into orbits of ( the group - action generated by ) @xmath27 .
the flow @xmath27 is said to be @xmath0-closed if @xmath61 is @xmath0-closed .
note @xmath62 , where the orbit class @xmath63 is the subset consisting of elements whose orbit closures correspond to the orbit closure of @xmath39 ( i.e. @xmath64 ) . for an @xmath0-closed flow @xmath27 , write @xmath65 , called the orbit class space of @xmath27 .
first we show the following lemma .
[ lem:001 ] let @xmath10 be a decomposition of a hausdorff space @xmath14 .
if @xmath10 is pointwise almost periodic and @xmath66 is usc , then @xmath10 is @xmath0-closed .
if @xmath14 is compact and @xmath10 is @xmath0-closed , then @xmath10 is pointwise almost periodic and @xmath66 is usc .
suppose that @xmath10 is pointwise almost periodic and that @xmath66 is usc . by proposition
1.2.1 @xcite ( p.13 ) , we have that @xmath67 is hausdorff . by lemma 2.3 @xcite , we obtain that @xmath10 is @xmath0-closed .
conversely , suppose that @xmath14 is compact and @xmath10 is @xmath0-closed .
by lemma 1.6 @xcite , we have that @xmath10 is pointwise almost periodic and that the quotient map @xmath68 is closed . since @xmath14 is compact hausdorff , we obtain that each element of @xmath66 is compact . by proposition
1.1.1 @xcite(p.8 ) , we have that @xmath66 is usc .
notice that the compact condition is necessary .
indeed , the @xmath0-closed decomposition @xmath69 for a non - compact hausdorff space @xmath14 is pointwise almost periodic but @xmath70 is not usc .
we show that each open @xmath13-saturated subset for a pointwise almost periodic decomposition @xmath1 on a hausdorff space is @xmath15-saturated .
[ lem02 ] let @xmath10 be a pointwise almost periodic decomposition of a topological space .
then each open @xmath13-saturated subset is @xmath15-saturated .
let @xmath53 be an open @xmath71-saturated subset of a topological space @xmath14 .
then the complement @xmath72 is a closed @xmath71-saturated subset of @xmath14 and so the closure of each element of @xmath1 contained in @xmath72 is contained in @xmath72 .
the pointwise almost periodicity implies that @xmath72 is closed @xmath73-saturated and so @xmath53 is open @xmath73-saturated . using the previous lemma
, we have the following statement . [ lem:001a ]
let @xmath10 be a pointwise almost periodic decomposition of a normal space @xmath14 .
suppose that , for any open neighbourhood @xmath74 of each element @xmath75 , there is an open saturated neighbourhood of @xmath54 contained in @xmath74 .
then @xmath16 is hausdorff .
suppose that , for any open neighbourhood @xmath74 of each element @xmath75 , there is an open @xmath13-saturated neighbourhood of @xmath54 contained in @xmath74 .
we claim that , for any open neighbourhood @xmath53 of each element @xmath76 , there is an open @xmath15-saturated neighbourhood of @xmath77 contained in @xmath53 .
indeed , fix any open neighbourhood @xmath53 of each element @xmath76 .
the hypothesis implies that , for any @xmath78 contained in @xmath77 , since @xmath53 is also an open neighbourhood of @xmath79 , there is an open @xmath13-saturated neighbourhood @xmath80 of @xmath79 contained in @xmath53 . by lemma [ lem02 ]
, the union @xmath81 is desired .
fix any elements @xmath82 .
since @xmath14 is normal , there are open neighborhoods @xmath83 ( resp .
@xmath84 ) of @xmath77 ( resp .
@xmath85 ) with @xmath86 . by the claim
, there are open @xmath73-saturated neighborhoods @xmath87 ( resp .
@xmath88 ) of @xmath77 ( resp .
@xmath89 ) such that @xmath90 .
notice that the converse is not true .
indeed , consider @xmath91 , @xmath92 , and homeomorphism @xmath33 of @xmath14 by @xmath93 .
then the decomposition @xmath61 on @xmath25 is @xmath0-closed .
then the saturation @xmath94 of @xmath95 , x \in [ 1/2 , 1/2 + y ] \}$ ] contains @xmath96 $ ] and so the open neighborhood @xmath97 of the orbit of @xmath98 contains no saturated open neighborhood . using the characterization of usc decompositions on locally compact hausdorff spaces ( remark after theorem 4.1 @xcite ) , we obtain the following key lemma . [ prop:002 ] let @xmath10 be a pointwise almost periodic decomposition of a compact hausdorff space @xmath14 . the following are equivalent : + 1 ) @xmath13 is @xmath0-closed .
+ 2 ) @xmath15 is usc .
+ 3 ) @xmath16 is hausdorff .
+ 4 ) the quotient map @xmath99 is closed ( i.e. @xmath13 is weakly almost periodic ) .
+ 5 ) for any open neighbourhood @xmath53 of each element @xmath76 , there is an open @xmath100-saturated neighbourhood of @xmath77 contained in @xmath53 .
+ 6 ) @xmath13 is of characteristic @xmath98
. + by corollary 3.5 @xcite , the conditions 1 ) and 6 ) are equivalent . by lemma [ lem:001 ] , the conditions 1 ) and 2 ) are equivalent . by remark after theorem 4.1 @xcite , the conditions 2 ) , 3 ) and 4 ) are equivalent
. apply lemma [ lem:001a ] to @xmath15 , the condition 5 ) implies 3 ) .
finally , it suffices to show that the condition 5 ) can be obtained by other conditions .
fix an open neighbourhood @xmath53 of each element @xmath76 . by the closedness of @xmath72 ,
the condition 4 ) implies that @xmath101 is closed . since @xmath102
, we have @xmath103 .
therefore the complement @xmath104 is a @xmath100-saturated open neighborhood of @xmath77 .
recall that a subgroup @xmath105 of a topological group @xmath21 is syndetic if there is a compact subset @xmath106 of @xmath21 such that @xmath107 .
[ lem31 ] let @xmath21 be a group - action on a topological space @xmath14 , @xmath105 a syndetic subgroup of @xmath21 , and @xmath108 a compact subset with @xmath107 .
if @xmath105 is @xmath0-closed , then so is @xmath21 with @xmath109 for any @xmath36 . for any group - action @xmath110 on a topological space @xmath111
, we claim that @xmath112 is closed for a closed subset @xmath113 of @xmath111 . indeed ,
fix a point @xmath114 .
then @xmath115 is an open neighborhood of @xmath116 and so @xmath117 is an open neighborhood of @xmath118 . for any @xmath119 ,
there are neighborhoods @xmath120 , @xmath121 of @xmath39 , @xmath19 respectively such that @xmath122 .
since @xmath123 is compact , there is a finite subset @xmath124 of @xmath123 such that @xmath125 .
then @xmath126 is an open neighborhood of @xmath39 with @xmath127 . hence @xmath128 .
therefore @xmath129 and so @xmath130 .
this shows that @xmath112 is closed .
let @xmath131 and @xmath132 .
suppose that @xmath105 is @xmath0-closed .
since @xmath133 for any @xmath134 , we have @xmath135 .
since @xmath136 is closed , the above claim implies @xmath137 and so @xmath138 .
consider an action of @xmath21 on @xmath139 by @xmath140 .
then @xmath141 .
since @xmath142 is closed , the above claim implies that @xmath141 is closed .
we generalize lemma 1.1 @xcite to group - actions .
this statement is an analogous result for recurrence and pointwise almost periodicity ( see theorem 2.24 , 4.04 , and 7.04 @xcite ) .
[ prop21 ] let @xmath21 be a group - action on a locally connected compact hausdorff space @xmath14 and @xmath105 a finite index subgroup .
then @xmath21 is @xmath0-closed if and only if so is @xmath105 . by lemma [ lem31 ]
, the @xmath0-closedness of @xmath105 implies one of @xmath21 .
conversely , suppose that @xmath21 is @xmath0-closed .
put @xmath143 . by corollary 1.4 @xcite
, we have that @xmath21 is pointwise almost periodic . by theorem
2.24 @xcite , we have that @xmath105 is also pointwise almost periodic . by lemma [ prop:002 ] ,
the decomposition @xmath73 is usc and it suffices to show that @xmath144 is usc . fix any @xmath36 .
note that @xmath145 and so @xmath146 .
first suppose that @xmath147 .
for any open neighbourhood @xmath148 of @xmath149 , the upper semicontinuity of @xmath73 implies that there is an @xmath73-saturated open neighbourhood @xmath40 of @xmath150 contained in @xmath53 . since @xmath151 , we have that @xmath40 is also an @xmath152-saturated open neighbourhood @xmath40 of @xmath150 .
second assume that @xmath153 .
fix an open neighborhood @xmath83 of @xmath154 .
let @xmath106 be a finite subset of @xmath21 such that @xmath155 .
put @xmath156 and @xmath157 .
let @xmath158 be the saturation of @xmath159 .
since @xmath160 is finite , we obtain @xmath161
. then @xmath77 and @xmath85 are closed . by lemma [ lem31 ]
, the pointwise almost periodicity of @xmath105 implies that @xmath162 .
since @xmath14 is normal , there are disjoint open neighborhoods @xmath163 of @xmath164 respectively such that @xmath165 .
since @xmath73 is usc , there are open subsets @xmath166 , @xmath167 such that @xmath168 is an @xmath73-saturated neighborhood of @xmath169 . since @xmath44 is compact and @xmath14
is locally connected , there are finitely many connected open subsets @xmath170 such that @xmath171 and @xmath172 . put @xmath173 .
for any @xmath174 and @xmath175 , the @xmath144-saturated property of @xmath176 implies that @xmath177 and so @xmath178 .
the @xmath73-saturated property of @xmath168 implies @xmath179 .
since @xmath180 is connected and @xmath178 , we obtain @xmath181 and so @xmath182 .
since @xmath183 , the neighborhood @xmath184 of @xmath154 is desired .
the author would like to know wether the locally connectivity condition is necessary or not , and wether the finite condition can be replaced by the syndetic condition .
let @xmath13 ( resp .
@xmath185 ) be a decomposition of a topological space @xmath14 ( resp .
@xmath111 ) and @xmath186 a continuous mapping . we call that @xmath33 is a semi - conjugacy from @xmath13 to @xmath185 if @xmath33 is surjective and the image of each element of @xmath1 is an element of @xmath185 ( i.e. @xmath187 for any @xmath36 ) .
then @xmath13 is said to be semi - conjugate to @xmath185 .
the semi - conjugacy preserves pointwise almost periodicity .
[ lem040 ] let @xmath13 be a decomposition of a compact space @xmath14 , @xmath185 a decomposition of a hausdorff space @xmath111 , and @xmath188 a semi - conjugacy from @xmath13 to @xmath185 .
if @xmath13 is pointwise almost periodic , then so is @xmath185 . for any element @xmath189
, the surjectivity of @xmath190 implies that there is an element @xmath191 with @xmath192
. then @xmath193 . indeed , we have @xmath194 and so @xmath195 .
since @xmath14 is compact , so is @xmath196 .
the hausdorff property of @xmath111 implies that the compact subset @xmath197 is closed and so @xmath198 .
suppose that there is an element @xmath189 whose closure is not minimal .
then there is an element @xmath199 such that @xmath200 .
fix an element @xmath191 with @xmath192 .
since @xmath193 , there is an element @xmath201 such that @xmath202 and @xmath203 . since @xmath204 , we have @xmath205 .
this contradict to the pointwise almost periodicity of @xmath13 .
thus @xmath185 is pointwise almost periodic .
we show that a semi - conjugacy from @xmath13 to @xmath185 implies the a semi - conjugacy from @xmath15 to @xmath206 if the decompositions @xmath13 and @xmath185 are pointwise almost periodic .
[ lem040 ] let @xmath13 ( resp .
@xmath185 ) be a pointwise almost periodic decomposition of a compact hausdorff space @xmath14 ( resp .
@xmath111 ) and @xmath188 a semi - conjugacy from @xmath13 to @xmath185
. then @xmath207 for any @xmath42 . since @xmath208
, it suffices to show @xmath209 for any @xmath36 .
fix any point @xmath36 .
first we show @xmath210 .
indeed , the compactness of @xmath211 implies that @xmath212 is compact and so closed .
therefore @xmath213 .
conversely , since @xmath214 , we obtain @xmath215 and so @xmath216 .
therefore @xmath217 .
the pointwise almost periodicity implies @xmath218 .
the semi - conjugacy preserves @xmath0-closedness .
[ lem041 ] let @xmath13 ( resp .
@xmath185 ) be a decomposition of a compact hausdorff space @xmath14 ( resp .
@xmath111 ) .
suppose @xmath13 is semi - conjugate to @xmath185 .
if @xmath13 is @xmath0-closed , then so is @xmath185 .
fix any closed subset @xmath106 of @xmath111 . by lemma [ prop:002 ]
, it suffices to show that the saturation @xmath219 of @xmath206 is closed .
let @xmath188 be a semi - conjugacy .
since @xmath220 is closed , the @xmath0-closedness of @xmath1 implies that @xmath221 is closed and so compact , because @xmath14 is compact . by the surjectivity of @xmath190 , we have @xmath222 .
the semi - conjugacy implies that @xmath223 is compact and so closed , because @xmath111 is hausdorff .
consider group - actions to apply lemma [ lem040 ] and proposition [ lem041 ] .
let @xmath21 ( resp .
@xmath105 ) be a group - action on a topological space @xmath14 ( resp .
@xmath111 ) and @xmath188 a continuous mapping . as circle homeomorphism cases , we call that @xmath190 is a semi - conjugacy if @xmath190 is a surjection whose image of an orbit of @xmath21 is an orbit of @xmath105 ( i.e. @xmath224 for any point @xmath36 ) .
then @xmath21 is said to be semi - conjugate to @xmath105 . by lemma [ lem040 ] and proposition [ lem041 ]
, we have the following statement . [ cor042 ]
suppose a group - action @xmath21 on a compact hausdorff space @xmath14 is semi - conjugate to a group - action @xmath105 on a compact hausdorff space @xmath111 .
if @xmath21 is pointwise almost periodic ( resp .
@xmath0-closed ) , then so is @xmath105 .
we also state the inherited property for orientation double covers .
[ lem43 ] let @xmath21 be a group - action on a non - orientable compact manifold @xmath14 and @xmath225 the orientable double cover of @xmath14 .
if @xmath21 is pointwise almost periodic ( resp .
@xmath0-closed ) , then so is the induced action on @xmath225 from @xmath21 .
let @xmath21 be a pointwise almost periodic group - action on a non - orientable compact manifold @xmath14 and @xmath226 the double covering map .
for a point @xmath36 , put @xmath227 . for @xmath42 , we have @xmath228 and so @xmath229 .
we show that the orbit closures @xmath230 and @xmath231 for any point @xmath36 are the orbit classes with either @xmath232 or @xmath233 .
indeed , since @xmath234 , we have @xmath235 . therefore @xmath236 . since @xmath237
, we have @xmath238 . for any @xmath239
we obtain @xmath240 and so @xmath241 .
the symmetry of @xmath242 and @xmath243 implies that @xmath244 .
suppose that there is a point @xmath245 such that @xmath246
. then @xmath247 .
this means @xmath248 .
therefore @xmath249 and so @xmath250 . then @xmath251 .
thus we may assume that @xmath252 for any @xmath239 .
then @xmath248 and so @xmath253 .
this implies @xmath254 .
in particular , we have @xmath255 and @xmath256 . for any @xmath257 , since @xmath258 and @xmath259
, we obtain @xmath260 and so @xmath261 . the symmetry of @xmath243 and @xmath242 implies that @xmath262 and so @xmath263 . thus @xmath21 is pointwise almost periodic .
suppose that @xmath21 is @xmath0-closed .
by lemma [ prop:002 ] , it suffices to show that @xmath264 is hausdorff .
indeed , fix any orbit classes @xmath265 of @xmath21 .
suppose that @xmath266 .
the @xmath0-closedness on @xmath14 implies that there are disjoint open saturated neighborhoods @xmath163 of @xmath267 , @xmath268 respectively .
then the inverse images @xmath269 and @xmath270 are disjoint saturated neighborhoods of @xmath271 and @xmath272 respectively .
suppose that @xmath273 .
put @xmath274 , @xmath275 , and @xmath276 .
fix a riemann metric @xmath277 on @xmath14 .
let @xmath278 be the induced metric of @xmath277 on @xmath225 and @xmath279 .
for a subset @xmath48 , we denote by @xmath280 the open @xmath281-neighborhood of @xmath48 .
fix a small positive number @xmath282 such that the restriction @xmath283 for each point @xmath284 is an embedding . then @xmath285 and @xmath286 .
the @xmath0-closedness on @xmath287 implies that there is a @xmath288-saturated neighborhood @xmath289 of @xmath290 .
we may assume that each connected component of @xmath53 intersects @xmath290 , by replacing @xmath53 with the resulting subset from @xmath53 by removing connected components of @xmath53 each of which does not intersect @xmath290 .
write @xmath291 and @xmath292
. then @xmath293 .
fix any point @xmath294 .
then there is a path @xmath295 from @xmath296 to a point in @xmath297 .
since @xmath293 , the path @xmath298 for each @xmath134 is also a path in @xmath299 from @xmath300 to a point in @xmath297 .
this means that @xmath299 is saturated . by the symmetry , the neighborhoods @xmath299 and @xmath301 of @xmath271 and @xmath272 respectively are disjoint saturated
recall that a minimal set @xmath302 on a surface homeomorphism @xmath303 is an extension of a cantor set ( resp . a periodic orbit ) if there are a homeomorphism @xmath304 and a semi - conjugacy @xmath305 with @xmath306 such that @xmath190 is homotopic to the identity and @xmath307 is a cantor set ( resp . a periodic orbit ) which is a minimal set of @xmath308 . by a continuum we mean a compact connected metrizable space which is not a singleton .
a continuum @xmath113 of a topological space @xmath14 is said to be annular if it has a neighborhood @xmath309 homeomorphic to an open annulus @xmath310 such that @xmath311 has exactly two components each of which is homeomorphic to an open annulus .
we call such @xmath53 an annular neighborhood of @xmath113 .
we show the invariance of cantor sets .
let @xmath312 ( resp .
@xmath313 ) be a homeomorphism on a locally connected compact hausdorff space @xmath14 ( resp .
@xmath111 ) , @xmath13(resp .
@xmath21 ) the decomposition of orbits of @xmath33 ( resp .
@xmath308 ) , @xmath314 a semi - conjugacy from @xmath13 to @xmath185 , and @xmath36 .
suppose @xmath33 is pointwise almost periodic and each inverse image of @xmath315 is finite .
if @xmath316 is a cantor set , then so is @xmath317 .
recall that a compact subset in a metrizable space is a cantor set if and only if it is totally disconnected and perfect . by proposition [ lem43 ] , @xmath308 is also pointwise almost periodic . by lemma [ lem040 ]
, we have @xmath318 .
since @xmath318 is a cantor set , the minimal set @xmath317 has infinitely many points and so is not a periodic orbit .
this implies that the minimal set @xmath317 has no isolated points .
this implies that it is perfect .
assume that there is a non - degenerate connected component @xmath113 of @xmath317 .
since @xmath319 is also connected and so a singleton and @xmath320 is finite , we have that @xmath113 is finite and so degenerate , which contradict to the definition of @xmath113 . thus @xmath317 is a cantor set .
let @xmath312 be a pointwise almost periodic homeomorphism on a locally connected compact hausdorff space @xmath14 and @xmath36 .
if @xmath317 is a cantor set , then so is @xmath321 for any positive number @xmath322 .
fix a positive iteration @xmath323 . since @xmath324 , we have that @xmath325 is totally disconnected .
assume @xmath325 is not perfect .
then there is an isolated point in @xmath325 .
this implies that @xmath325 is periodic and so @xmath317 is periodic , which contradicts that @xmath317 is a cantor set .
we show the connectivity of minimal sets are invariant under iterations .
[ lem63 ] let @xmath21 be a group - action on a locally compact hausdorff space @xmath14 , @xmath105 a finite index subgroup , and @xmath36 .
suppose that @xmath21 is pointwise almost periodic .
if @xmath326 is connected , then @xmath327 . by theorem
2.24 @xcite , the subgroup @xmath105 is also pointwise almost periodic .
assume that @xmath328 .
let @xmath106 be a finite subset of @xmath21 such that @xmath155 .
put @xmath329 . since @xmath105 is of finite index ,
the orbit closure @xmath330 is disconnected , which contradicts to the connectivity of it .
now we show the following tool .
[ lem55 ] let @xmath331 be an @xmath0-closed homeomorphism on a compact surface @xmath8 .
then @xmath33 has no minimal set which is an extension of a cantor set .
suppose that there is a minimal set @xmath332 of @xmath33 which is an extension of a cantor set .
taking the double of @xmath8 , by proposition [ lem43 ] , we may assume that @xmath8 is closed and orientable .
by the definition of an extension of a cantor set , there are a surface homeomorphism @xmath304 , a semi - conjugacy @xmath305 from @xmath33 to @xmath308 , and a cantor minimal set @xmath333 of @xmath308 .
it suffices to show the non - existence of such a minimal set @xmath302 . by corollary [ cor042 ]
, we obtain that @xmath308 is an @xmath0-closed homeomorphism with a cantor minimal set @xmath302 on a closed orientable surface @xmath8 .
since a cantor set is zero - dimensional , there is a basis of @xmath302 consisting of closed and open subsets .
we show that , for a point @xmath334 , there is an open disk @xmath335 which is a neighborhood of @xmath39 and whose boundary @xmath336 does not intersect @xmath302 , where @xmath337 . indeed ,
fix a small open ball @xmath83 centered at @xmath39 . since there is a basis of @xmath302 consisting of closed and open subsets
, there is an open @xmath87 such that @xmath338 .
define the filling @xmath339 of @xmath340 in @xmath83 as follows : @xmath341 if there is a simple closed curve in @xmath340 which bounds an open disk in @xmath83 containing @xmath39 .
then the filling @xmath342 is a desired open disk containing @xmath39 such that @xmath343 .
the compactness of @xmath302 implies that there are finitely many points @xmath344 with @xmath345 .
let @xmath346 be the saturation of the boundary of @xmath347 and @xmath348 . by the @xmath0-closedness of @xmath308 ,
the saturation @xmath12 is closed such that @xmath349 .
since @xmath14 is normal , there are an open neighborhood @xmath350 ( resp .
@xmath351 ) of @xmath302 ( resp .
@xmath12 ) with @xmath352 and @xmath353 such that @xmath354 is an annular neighborhood of @xmath355 and each @xmath356 is a closed disk .
notice that each connected component of @xmath357 is contained in some closed disk @xmath356 and so contractible in @xmath356 . by the @xmath0-closedness of @xmath308 , there is an open saturated neighborhood @xmath358 of @xmath302 .
then each connected component of @xmath53 is contained and contractible in some closed disk @xmath356 .
fix any point @xmath334 .
let @xmath74 be the connected component of @xmath53 containing @xmath39 and @xmath359 the closed disk containing @xmath74 for some @xmath360 .
since @xmath361 , the connected component @xmath74 is the connected component of @xmath362 containing @xmath39 .
define the first return map @xmath363 by @xmath364 for any @xmath365 , where @xmath366 .
the connectivity of @xmath74 implies that @xmath367 is a constant @xmath368 , and that @xmath369 .
this means that @xmath370 and so that @xmath21 is a homeomorphism .
taking an iteration of @xmath21 , we may assume that @xmath21 maps the outer boundary @xmath371 of @xmath74 into itself .
note that @xmath372 .
since @xmath53 is saturated , the image @xmath373 for a simple closed curve @xmath374 in @xmath74 and for @xmath375 is a simple closed curve which is the boundary of an open disk in @xmath376 . by the construction of @xmath377
, we can extend @xmath21 into a mapping @xmath378 .
since the original @xmath308 is @xmath0-closed and homeomorphic , by theorem [ prop21 ] , so is the iteration @xmath21 . by lemma [ prop:002 ]
, @xmath379 is hausdorff .
the @xmath0-closedness of @xmath308 implies that the boundary @xmath380 is a saturated closed subset . by the one point compactification @xmath8 of @xmath377 , the normality of @xmath2
imply that @xmath8 is homeomorphic to @xmath381 and that the orbit class space of the induced homeomorphism on the sphere @xmath8 by @xmath21 is hausdorff , where the equivalent relation @xmath382 is defined as follows : @xmath383 if either @xmath384 or @xmath385 . by lemma [ prop:002 ] ,
the resulting homeomorphism is an @xmath0-closed homeomorphism on a sphere with a cantor minimal set .
this contradicts to corollary 2.6 @xcite .
thus there are no cantor minimal sets . recall that a homeomorphism @xmath312 is periodic if there is a positive integer @xmath386 such that @xmath387 is identical
we say a subset @xmath388 is a circloid if it is an annular continuum and does not contain any strictly smaller annular continuum as a subset .
the combination of lemma [ lem55 ] and theorem 3.2 @xcite implies the following result .
[ cor57 ] each @xmath0-closed toral homeomorphism @xmath33 satisfies one of the following : + 1 .
@xmath33 is minimal .
@xmath33 is periodic .
each minimal set is a finite disjoint union of essential circloids .
we show that an @xmath0-closed homeomorphism on a hyperbolic compact surface is periodic .
[ cor56 ] each homeomorphism on a connected compact surface @xmath8 whose euler number is negative is periodic if and only if it is @xmath0-closed .
let @xmath33 be a homeomorphism on a compact surface @xmath8 whose euler number is negative .
suppose that @xmath33 is periodic .
since the orbit class space of a periodic homeomorphism on a compact manifold is hausdorff , the homeomorphism @xmath33 is @xmath0-closed .
conversely , suppose that @xmath33 is @xmath0-closed .
taking the double of @xmath8 , by proposition [ lem43 ] , we may assume that @xmath8 is closed and orientable and the euler number of @xmath8 is negative .
then the genus of @xmath8 is at least two . by lemma [ lem55 ]
, there is no minimal set which is an extension of a cantor set . by theorem 3.1 @xcite
, we have that @xmath33 is pointwise periodic . by theorem @xcite ,
each pointwise periodic surface homeomorphism is periodic and so is @xmath33 .
recall that a flow is trivial if it is identical or minimal . for a flow @xmath27 on a @xmath7-manifold , a minimal set @xmath332 of @xmath27
is called a suspension of a circloid for @xmath27 if there are an open annulus @xmath48 transverse to @xmath27 and a circloid @xmath113 in @xmath48 whose saturation @xmath389 is @xmath332 with @xmath390 .
we obtain the following statements for flows on @xmath7-manifolds .
[ th:005 ] let @xmath27 be an @xmath0-closed flow on a connected compact @xmath7-manifold @xmath2 .
then the union @xmath53 of suspensions of circloids is open and the quotient space @xmath391 is a @xmath392-manifold .
let @xmath48 be an open annulus transverse to @xmath27 and @xmath393 a circloid whose saturation is a suspension @xmath332 of a circloid with @xmath390 and @xmath394 . then @xmath395 .
since @xmath332 is @xmath396-saturated and connected , there are small connected neighbourhoods @xmath397 of @xmath332 with @xmath398 such that @xmath399 are connected and the first return map @xmath400 is well - defined .
note @xmath401 .
since @xmath27 is @xmath0-closed and so @xmath396 is usc , there is an @xmath396-saturated connected open neighbourhood @xmath402 of @xmath332 .
we show that there is an @xmath396-saturated open connected neighbourhood @xmath403 of @xmath332 such that the first return map @xmath404 is well - defined and homeomorphic , where @xmath405 . indeed , for a point @xmath406
, there is a small open connected neighborhood @xmath407 of @xmath406 such that the first return time map @xmath408 with @xmath409 for any @xmath410 is continuous . since @xmath113 is compact , there are finitely many points @xmath411 such that @xmath412 .
then @xmath413 is an open neighborhood of @xmath113 .
moreover the first return time map @xmath414 is well - defined and continuous .
since @xmath27 is @xmath0-closed , there is an @xmath396-saturated open neighbourhood @xmath415 , y ) \mid y \in w_c \}$ ] of @xmath332 .
put @xmath416 . then @xmath417 .
since @xmath74 is @xmath396-saturated , we have @xmath418 and @xmath419 .
this means @xmath420 and so @xmath421 is a homeomorphism . replacing @xmath422 with the @xmath396-saturation @xmath423 of an open annular neighborhood @xmath424 of
the @xmath425-minimal set @xmath113 , we may assume that @xmath422 is connected .
let @xmath426 be the union of the connected components of @xmath427 which are contractible in @xmath48 .
then it suffices to show that @xmath428 is an open annulus and @xmath429 is an @xmath396-saturated connected open neighbourhood of @xmath332 which consists of suspensions of circloids .
indeed , define an open annulus @xmath430 as follows : @xmath431 if there is a simple closed curve in @xmath422 which bounds a disk in @xmath48 containing @xmath432 .
since each point of @xmath426 is bounded by a simple closed curve in @xmath422 , we have that @xmath433 and @xmath434 is a homeomorphism .
since @xmath113 is a circloid and @xmath48 is the annular neighbourhood , we obtain that @xmath430 is an open annulus . since @xmath27 is @xmath0-closed , we have that @xmath435 is hausdorff and so is @xmath436 .
by the two point compactification of @xmath430 , we obtain the resulting sphere @xmath8 and the resulting homeomorphism @xmath437 with the two periodic points which are the added new points . by the construction
, @xmath438 is the two point compactification of @xmath436 . replacing @xmath439 with @xmath440 if necessary
, we may assume that the added new points are fixed points .
since @xmath441 , we have @xmath442
. then @xmath443 .
indeed , otherwise there is a point @xmath444 . since @xmath445 is compact , there are finitely many points @xmath446 and their @xmath396-saturated neighborhoods @xmath447 of @xmath448 such that @xmath449 and @xmath450 . then @xmath451 is a closed @xmath396-saturated neighborhood of @xmath452 which does not intersect @xmath39 .
this implies that @xmath453 , which contradicts to the choice of @xmath39 .
since @xmath2 is normal , the @xmath0-closedness of @xmath27 implies that @xmath438 is hausdorff and so @xmath454 is @xmath0-closed . by corollary 2.6 @xcite
, we obtain that @xmath8 consists of two fixed points and circloids .
then the open neighbourhood @xmath455 of @xmath332 consists of suspensions of circloids .
therefore the union of suspensions of circloids are open . by dimension , we mean lebesgue covering dimension . by urysohn s theorem ,
the lebesgue covering dimension and the small inductive dimension are corresponded in normal spaces .
a compact metric space @xmath14 whose inductive dimension is @xmath456 is an @xmath386-dimensional cantor manifold if the complement @xmath457 for any closed subset @xmath54 of @xmath14 whose inductive dimension is less than @xmath458 is connected .
it s known that a compact connected manifold is a cantor - manifold @xcite .
[ lem062 ] let @xmath2 be a connected compact manifold , @xmath1 an @xmath0-closed decomposition on @xmath2 , and @xmath459 non - empty open saturated subset . if the codimension of @xmath460 is more than one , then @xmath461 .
the @xmath0-closedness implies that @xmath460 is closed .
fix a metric @xmath277 on @xmath2 .
since a compact connected manifold is a cantor manifold , we have that @xmath462 is connected .
since @xmath463 is closed and open in @xmath462 , the connectivity of @xmath462 implies that @xmath464 . since the codimension of @xmath465 is more than one , we have @xmath466 for any @xmath467 and any @xmath468
. therefore @xmath469 and so @xmath461 .
recall that a minimal set is trivial if it is either a closed orbit or the whole manifold .
[ lem52 ] let @xmath27 be an @xmath0-closed flow on a connected compact @xmath7-manifold @xmath2 .
suppose that each two dimensional minimal set is a suspension of a circloid .
if there is a suspension of a circloid , then the orbit class space @xmath435 of @xmath2 is a closed interval or a circle such that there are at most two orbits each of which is not suspension of a circloid . by lemma [ th:005 ] ,
the union @xmath53 of suspensions of circloids is open and the quotient space of it is a @xmath392-manifold .
the hausdorff property of @xmath435 implies that @xmath470 is either a closed interval or a circle .
the @xmath0-closedness implies that @xmath471 consists of at most two orbit closures each of whose dimension of @xmath471 is at most one . by lemma [ lem062 ] , we have that @xmath461 and so that @xmath472 is a closed interval or a circle .
the following statement is obtained by the similar argument as the proof of lemma [ lem55 ] .
[ lem54 ] let @xmath27 be an @xmath0-closed flow on a connected compact @xmath7-manifold @xmath2 .
each one dimensional minimal set is a periodic orbit .
suppose that there is a one - dimensional minimal set @xmath332 which is not a periodic orbit .
since @xmath395 , there is a neighborhood of @xmath332 . by the @xmath0-closedness of @xmath27 , there is an open saturated neighborhood @xmath53 of @xmath332 without singularities . since a locally compact hausdorff space is zero - dimensional if and only if it is totally disconnected , each nonempty zero - dimensional perfect compact metrizable space is homeomorphic to the cantor set
this implies that the minimal set @xmath332 is transversally a cantor set .
in fact , there is a transverse open disk @xmath473 to @xmath27 such that @xmath474 is a cantor set .
then we may assume that @xmath475 .
indeed , since a cantor set has a basis consisting of open and closed subsets , for any point @xmath334 , there is an open disk @xmath476 containing @xmath39 with @xmath477 .
the compactness of @xmath302 implies that there are finitely many points @xmath344 such that the filling of the union @xmath478 is an open disk whose boundary does not intersect @xmath302 and which contains @xmath302 . replacing @xmath479 with the filling ,
we are done .
taking @xmath480 for any point @xmath334 small , we may assume that the first return time map @xmath481 with @xmath482 for any @xmath483
is continuous . since @xmath484 is closed and disjoint from @xmath332 , there are disjoint open connected saturated neighborhoods @xmath485 , y ) \mid y \in v_{x_i } \}$ ] ( resp .
@xmath486 ) of @xmath332 ( resp .
@xmath12 ) . then @xmath487 and @xmath488 .
let @xmath489 be the connected component of @xmath490 containing @xmath491 .
since @xmath492 , we have @xmath493 and so @xmath494 . since @xmath495 , we obtain @xmath496 .
the inclusion @xmath489 implies @xmath497 . for each @xmath175 and @xmath498
, there is some @xmath499 such that @xmath500 .
then the first return map @xmath501 is well - defined and a continuous injection .
since @xmath502 is saturated and @xmath489 is the connected component of @xmath503 containing @xmath504 , we have that @xmath505 is also the connected component of @xmath503 containing @xmath504 and that @xmath506 .
this implies that @xmath21 is a homeomorphism .
define a disk @xmath507 as follows : @xmath508 if there is a simple closed curve in @xmath40 which bounds a disk in @xmath509 containing @xmath190 .
replacing @xmath21 with @xmath510 if necessary , by the construction of @xmath511 , we can extend @xmath21 into a homeomorphism @xmath512 . by lemma [ prop:002 ] ,
the mapping @xmath21 is a homeomorphism on an open disk @xmath511 such that @xmath513 is hausdorff .
the @xmath0-closedness of @xmath27 implies that @xmath514 is a saturated closed subset of @xmath27 . by the one point compactification of @xmath511 , the resulting space
@xmath515 is a sphere and the orbit class space @xmath516 is hausdorff .
then lemma [ prop:002 ] and the normality of @xmath2 imply that the resulting homeomorphism from @xmath21 is an @xmath0-closed homeomorphism on the sphere @xmath8 with a cantor minimal set .
this contradicts to corollary2.6 @xcite .
thus each one dimensional minimal set is a periodic orbit . by surfaces with conners ,
we mean two dimensional manifolds locally modeled by @xmath517 ^ 2 $ ] .
[ lem:006 ] let @xmath27 be an @xmath0-closed flow on a connected compact @xmath7-manifold @xmath2 .
suppose that each two dimensional minimal set is a suspension of a circloid .
if there are at least three distinct periodic orbits , then @xmath30 is open dense and @xmath31 .
if there is a three dimensional minimal set , then the @xmath0-closedness implies that it has no boundaries and so equals @xmath2 .
thus there are no three dimensional minimal sets .
let @xmath518 be the canonical projection .
by lemma [ lem52 ] , there are no two dimensional orbits closures . since locally connected zero ( resp .
one ) dimensional minimal sets are singular ( resp .
periodic ) orbits , we have @xmath31 . since @xmath519 is closed , we have that @xmath30 is open . by theorem
3.12 @xcite , the quotient space of @xmath30 is a surface with conner .
since @xmath472 is compact hausdorff , the dimension of @xmath520 is at most one .
since @xmath520 consists of singularities , we obtain that the dimension of @xmath521 is at most one . by lemma [ lem062 ]
, we obtain @xmath522 .
now , we state the following trichotomy that an nontrivial @xmath0-closed flow on a @xmath7-manifold is either `` almost two dimensional '' or `` almost one dimensional '' or with `` complicated '' minimal sets .
let @xmath27 be an nontrivial @xmath0-closed flow on a connected compact @xmath7-manifold @xmath2 .
then one of the following holds : + 1 ) the orbit class space @xmath435 of @xmath2 is a closed interval or a circle and each interior point of the orbit class is two dimensional .
+ 2 ) @xmath30 is open dense and @xmath31 .
+ 3 ) there is a two dimensional minimal set which is not a suspension of a circloid .
suppose that each two dimensional minimal set is a suspension of a circloid .
since @xmath27 is nontrivial , there is a nontrivial minimal set @xmath332 .
if the minimal set @xmath332 is two dimensional , then lemma [ lem52 ] implies 1 ) .
thus we may assume that there are no two dimensional minimal sets .
by lemma [ lem54 ] , each one dimensional minimal set is a periodic orbit and so @xmath332 is a periodic orbit . by the flow box theorem ,
this has a neighbourhood without singularities .
since there are no two dimensional minimal sets , there are infinitely many periodic orbits . by lemma [ lem:006 ] , we have that 2 ) holds .
non - trivial @xmath0-closed flows on a connected compact @xmath7-manifolds consisting of closed orbits are of type @xmath523 .
the suspensions of diffeomorphisms which are irrational rotations with respect to an axis on @xmath524 or @xmath25 are of type @xmath525 in the above proposition
. moreover there are homeomorphisms of type @xmath525 with minimal sets which are not circles but circloids .
in fact , recall that there is the remage s @xmath0-closed homeomorphism in example 1 @xcite ( resp .
herman s @xmath526 diffeomorphism
@xcite ) on @xmath527 .
moreover each of them can be modified into a toral @xmath0-closed homeomorphism with pseudo - circles ( resp .
@xmath526 diffeomorphism with non locally connected continua ) as minimal sets @xcite . using @xmath0-closed diffeomorphisms on annuli with minimal sets which are circloids but neither circles nor pseudo - circles @xcite
, one can construct @xmath0-closed diffeomorphisms with such a minimal set on @xmath527 and @xmath25 .
then the suspensions of such diffeomorphisms are also of type @xmath525 .
the author would like to know whether there are @xmath0-closed flows on a connected compact @xmath7-manifolds with nontrivial minimal sets which are not suspensions of circloids .
first , we consider the codimension one case .
[ lem:4.1 ] let @xmath1 be an @xmath0-closed decomposition on a compact connected manifold @xmath2 without codimension zero element classes and @xmath53 a connected component of the union of codimension one elements of @xmath15 .
if @xmath53 is non - empty and open such that each element of @xmath15 in @xmath53 is a closed connected submanifold , then @xmath528 is a closed interval or a circle . by lemma [ prop:002 ] , we have that @xmath528 is hausdorff .
let @xmath53 be the above open connected component and @xmath529 the canonical projection . by theorem 3.3 @xcite
, we have that @xmath530 is a @xmath392-manifold .
suppose @xmath530 is a circle or a closed interval .
then the connectivity of @xmath2 implies @xmath531 .
suppose that @xmath53 is an interval which is not closed . since @xmath532 is hausdorff , each convergence sequence has one limit and so the codimension of @xmath533 is more than one .
lemma [ lem062 ] implies that @xmath461 .
the definition of @xmath53 implies that the closure @xmath534 is a closed interval .
-closed decomposition on a closed orientable surface with genus @xmath535 whose class space is a figure eight curve . ]
the codimension zero condition is necessary , because there are @xmath0-closed decompositions on compact connected surfaces .
indeed , let @xmath12 be a compact three - punctured disk with a decomposition @xmath536 and @xmath537 an open annulus with a decomposition @xmath538 .
gluing @xmath12 and two copies of @xmath539 , we obtain a closed surface @xmath8 of genus @xmath535 and the induced decomposition @xmath1 consists of @xmath12 and essential circles such that the quotient space @xmath540 is neither an interval nor a circle but a graph which is a figure eight curve ( see figure [ fig01 ] ) . recall that each continuous codimension one compact foliation on a compact manifold is @xmath0-closed .
hence the following corollary is a generalization of a fact that the leaf ( class ) space of a continuous codimension one compact foliation of a compact manifold is a closed interval or a circle .
[ cor84 ] let @xmath13 be an @xmath0-closed foliation on a compact manifold @xmath2 . if there is a compact @xmath15-saturated subset @xmath113 whose codimension is more than one such that all leaf closures of @xmath13 in @xmath541 are codimension one closed connected submanifolds , then @xmath528 is either a closed interval or a circle
. applying lemma [ lem:4.1 ] to the group - action cases , we obtain a following statement .
[ prop:003 ] let @xmath21 be an @xmath0-closed group - action on a compact connected manifold @xmath2 , @xmath105 a finite index subgroup of @xmath21 , and @xmath53 a connected component of the union of orbit closures of @xmath105 which are codimension one .
if @xmath53 is non - empty and open such that each orbit closure of @xmath105 in @xmath53 is a closed connected submanifold , then @xmath542 is a closed interval or a circle such that there are at most two elements whose codimensions are more than one . by theorem [ prop21 ]
, we have that @xmath105 is also @xmath0-closed . put @xmath543 . since @xmath21 is a non - minimal group - action ,
there are no codimension zero minimal sets .
applying lemma [ lem:4.1 ] , the class space @xmath544 is either a circle or a closed interval .
since @xmath545 consists of orbit closures , the boundary @xmath471 also consists of orbit closures whose codimension are more than one . by lemma [ lem062 ]
, we obtain that @xmath461 and so that @xmath471 consists of at most two elements of @xmath73 .
hence @xmath528 is either a circle or a closed interval and there are at most two elements whose codimensions are more than one .
since @xmath546 is a finite group acting @xmath532 and since a finite union of closures is a closure of finite union , we have that @xmath547 is either a closed interval or a circle and that there are at most two elements whose codimensions are more than one .
this lemma implies the following application .
[ prop:0073 ] let @xmath21 be an @xmath0-closed group - action on a compact connected manifold @xmath2 and @xmath105 a finite index subgroup of @xmath21 .
suppose that all but finitely many orbit closures of @xmath105 are codimension one closed connected submanifolds . if the codimension of each of finite exceptions is not one , then @xmath542 is a closed interval or a circle .
let @xmath548 be the finite exceptions and @xmath549 .
then @xmath53 is the union of orbit closures of @xmath105 which are codimension one connected submanifolds , and is an open saturated submanifold .
if there is a codimension zero minimal set , then the @xmath0-closedness implies that it has no boundaries and so equals @xmath2 , which contradicts to the non - triviality .
thus there are no codimension zero minimal sets and so the codimension of each @xmath550 is at least two . by lemma [ prop:003 ] , we are done .
second , we consider the codimension two case .
consider the direct system @xmath551 of compact subsets of a topological space @xmath14 and inclusion maps such that the interiors of @xmath552 cover @xmath14 .
there is a corresponding inverse system @xmath553 , where @xmath554 denotes the set of connected components of a space @xmath111 .
then the set of ends of @xmath14 is defined to be the inverse limit of this inverse system .
[ lem:4.2 ] let @xmath13 be an @xmath0-closed decomposition on a compact manifold @xmath2 .
suppose that all but finitely many element closures of @xmath13 are codimension two closed connected submanifolds .
if each of finite exceptions is connected and has codimension at least three , then @xmath528 is a surface with conners .
let @xmath555 be all higher codimension elements of @xmath66 .
removing higher codimensional elements and @xmath556 be the resulting manifold , @xmath557 the resulting decomposition of @xmath556 .
then @xmath557 consists of codimension two closed connected submanifolds and is usc . by theorem
3.12 @xcite , we have that @xmath558 is a surface @xmath559 with conners .
then @xmath560 .
we will show that @xmath559 has @xmath19 ends .
indeed , since the exceptions @xmath561 are finite and connected , there are pairwise disjoint neighbourhoods @xmath447 of them . since @xmath66 is usc , there are pairwise disjoint saturated neighbourhoods @xmath562 of them . since @xmath563 is connected for any connected neighbourhoods @xmath564 of @xmath561 , each end of @xmath559 is corresponded to some @xmath561 .
this shows that @xmath559 has @xmath19 ends corresponding to @xmath565 .
therefore @xmath532 is an end compactification of @xmath558 and so a surface @xmath8 with conners .
recall that each continuous codimension two compact foliation on a compact manifold is @xmath0-closed .
hence the following corollary is a generalization of the fact that the leaf ( class ) space of a continuous codimension two compact foliation of a compact manifold is a compact orbifold .
[ cor86 ] the leaf class space of a codimension - two - like @xmath0-closed foliation on a compact manifold is a surface with conners . applying the previous lemma to the group - action cases
, we have a following statement . [ lem87 ]
let @xmath21 be an @xmath0-closed group - action on a compact manifold @xmath2 and @xmath105 a finite index subgroup of @xmath21 .
if @xmath105 is codimension - two - like , then the orbit class space @xmath542 is a surface with conners . by theorem [ prop21 ]
, we have that @xmath105 is also @xmath0-closed .
applying lemma [ lem:4.2 ] , @xmath566 is a surface with conners .
since @xmath546 implies a finite group - action on @xmath8 and since any finite union of closures is the closure of finite union , we obtain that @xmath567 is a surface with conners .
summarizing corollary [ cor84 ] and [ cor86 ] , we obtain a following statement for foliations .
let @xmath1 be an @xmath0-closed foliation on a compact connected manifold @xmath2 and @xmath17 ( resp .
@xmath18 ) . if @xmath1 is codimension-@xmath19-like , then the leaf class space @xmath528 is a surface with conners ( resp . a closed interval or a circle ) .
similarly , lemma [ lem87 ] and proposition [ prop:0073 ] imply the following statement for group - actions .
let @xmath21 be an @xmath0-closed group - action on a compact connected manifold @xmath2 , @xmath105 a finite index subgroup of @xmath21 , and @xmath17 ( resp .
@xmath18 ) . if @xmath21 is codimension-@xmath19-like
, then @xmath542 is a surface with conners ( resp . a closed interval or a circle ) .
we states several examples as follows .
a denjoy foliation on a torus is codimension - one - like but not @xmath0-closed . moreover
a morse - smale diffeomorphism on a closed @xmath386-manifold is codimension-@xmath386-like but not @xmath0-closed . a spherical flow which is a non - trivial rotation with respect to an axis is codimension - one - like and @xmath0-closed , and so the orbit class space is a closed interval .
moreover , there are codimension - one - like @xmath0-closed spherical homeomorphisms ( resp . diffeomorphisms ) with minimal sets which are not circles but circloids @xcite ( resp .
@xcite ) . on the other hand ,
a toral flow @xmath27 defined by @xmath568 is not @xmath0-closed . for an irrational number @xmath92 , the suspension of a spherical diffeomorphism @xmath33 defined by @xmath569 ( i.e. an irrational rotation with respect to the @xmath570-axis ) is a codimension - one - like @xmath0-closed flow on @xmath571 and so the orbit class space is a closed interval , where @xmath572 .
contrary , the suspension of a periodic spherical diffeomorphism ( e.g. a rational rotation ) is a codimension - two - like @xmath0-closed flow on @xmath571 , and the hopf fibration on @xmath573 is a codimension - two - like @xmath0-closed foliation whose class space is a sphere .
moreover each seifert fibration is a codimension - two - like @xmath0-closed foliation .
note that ( the suspension of ) a spherical diffeomorphism @xmath574 defined by @xmath575 is pointwise almost periodic but not @xmath0-closed .
recall that the topological suspension @xmath576 of a topological space @xmath14 is the quotient space defined by @xmath577 , where @xmath578 if @xmath579 .
since @xmath573 is a topological suspension of @xmath524 , the induced homeomorphism @xmath580 ( i.e. @xmath581 ) : = [ f(x ) , s]$ ] ) on the topological suspension @xmath582 of a codimension - one - like @xmath0-closed spherical homeomorphism @xmath33 is codimension - two - like and @xmath0-closed , and the class space is either a sphere or a closed disk with two corners .
finally , we state two remarks . consider two minimal linear flows on the flat torus @xmath583 generated by vector fields @xmath584 such that @xmath585 , g(v ' , \cdot ) = [ dx_1 + \pi dx_2 ] \in h^1((\r/\z)^2 , \z)$ ] respectively , where @xmath308 is the induced riemann metric .
though the orbit spaces are indiscrete spaces with cardinalities of the continuum and the orbit class spaces are singletons , the minimal flows are not topological equivalent because @xmath586 is transcendental and @xmath587 is algebraic . by the @xmath588 novikov compact leaf theorem @xcite , there are no codimension one foliations on @xmath573 with reeb components and so there are no codimension one @xmath0-closed foliations on @xmath573 .
contrary , consider an irrational flow @xmath589 on a torus @xmath590 , which induces a kronecker foliation .
let @xmath591 be a flow on @xmath592 $ ] defined by @xmath593 .
define an equivalent relation @xmath594 as follows : @xmath595 if either @xmath596 and @xmath597 or @xmath598 and @xmath599 .
then the quotient space @xmath600)/\sim$ ] is a sphere @xmath573 such that the induced flow @xmath27 has exactly two periodic orbits @xmath601 $ ] and @xmath602 $ ] and each orbit class except two periodic orbits is a torus .
this means that the induced foliation @xmath61 is a codimension - one - like @xmath0-closed foliations on @xmath573 .
9 auslander , j. ; glasner , e. ; weiss , b. , _ on recurrence in zero dimensional flows _ forum math .
19 ( 2007 ) , no .
1 , 107114 .
, _ topological properties of solutions of ordinary differential equations _ , _ amer . j. math . _
59(1937 ) , 957982 .
h. colman , s. hurder , _ ls - category of compact - hausdorff foliations _ trans .
( 2004 ) , no .
4 , 14631487 . , _ decompositions of manifolds into codimension one submanifolds _ , _ compositio math . _ 55 ( 1985 ) , no .
2 , 185207 . , _ decompositions of manifolds _ pure and applied mathematics , 124 . academic press , inc . , orlando , fl , 1986 . , _ decompositions into codimension two submanifolds : the nonorientable case _ special volume in honor of r. h. bing ( 19141986 ) .
topology appl . 24
( 1986 ) , no . 1 - 3 , 7181
. k. de cesare and t. nagano , _ on compact foliations , differential geometry _ , ( proc .
pure math .
27 , part 1 ) , amer . math .
soc . , providence , r.i .
, 1975 , pp . 277281 . , _ foliations with all leaves compact _ , _ topology _ , 16 , 1332 , 1977 . s. elaydi , _ on some stability notions in topological dynamics _
j. differential equations , 47 ( 1983 ) , 2434 .
, _ periodic flows on 3-manifolds _ , _ ann . of math .
_ , 95 , 6882 , 1972 , _ foliations with all leaves compact _ , _ ann .
fourier ( grenoble ) .
_ , 26 , 265282,1976 .
d. epstein , h. rosenberg , _ stability of compact foliations _
, springer lecture notes in math . 597 ( 1978 ) , 151160 .
, _ a counter - example to the periodic orbit conjecture in codimension 3 _ , _ ann . of math .
_ ( 2 ) , 108 , 539552 , 1978 . , _ some remarks on almost periodic transformations _ , _ bull .
_ 51 , ( 1945 )
. 126130 .
r. fujisaki , _ stability of singular leaves of compact hausdorff foliations with tori as generic leaves _ , tokyo j. math .
22 ( 1999 ) , no .
1 , 213234 . k. fukui , a. mori , _ codimension two compact hausdorff foliations by hyperbolic surfaces are not stable _ , publ
inst . math .
36 ( 2000 ) , no .
3 , 321336 . , _ topological dynamics _ , amer
36 , american mathematical society , providence , ri , 1955 hattab , h. , _ pointwise recurrent one - dimensional flows _ ,
( 2011 ) , no .
1 , 7783 . , _ construction of some curious diffeomorphisms of the riemann sphere _ , _ j. london math . soc . _
34(1986 ) 375384
. h. holmann , _ seifertsche faserrume _ , math .
157 ( 1964 ) , 138166 .
, _ dimension and zusammenhangsstuffe _ , _ math .
_ , 100 ( 1928 ) pp . 618633 .
ronald a. knight , _ prolongationally stable transformation groups _ , math . zeits .
, 161 ( 1978 ) , 189194 .
kenneth c. millett , _ compact foliations _
, differential topology and geometry ( proc .
colloq . , dijon , 1974 ) springer , berlin , 1975 , 277287 .
lecture notes in math .
, vol . 484 . ,
_ pointwise periodic homeomorphisms _ , _
_ 59 ( 1937 ) , 118120 . , _
surface diffeomorphisms with connected but not path - connected minimal sets containing arcs _ , preprint .
m. nicolau , a. revents , _ compact hausdorff foliations _
, differential geometry ( pescola , 1982 ) , 147153 , lecture notes in math .
, 1045 , springer , berlin , 1984 . j. plante , _ stability of codimension one foliations by compact leaves _
, topology 22 ( 1983 ) , 173177 . , _ on minimal sets in the plane _ , _ proc .
soc . _ 13 1962 4147 .
r. sacksteder , _ foliations and pseudogroups _ , _ amer . j. math .
_ 87 ( 1965 ) , 79102 . v. v. solodov . , _ components of topological foliations _ mat .
, 119(161)(3 ) : 340354 , 447 , 1982 .
, _ a counterexample to the periodic orbit conjecture _ , _ publ .
ihes _ , 46 , 514 , 1976 .
w. p. thurston , _ existence of codimension - one foliations _
, _ ann . of math .
_ ( 2 ) 104 ( 1976 ) , no .
2 , 249268 .
d. tischler , _ on fibering certain foliated manifolds over @xmath9 _ ,
_ topology _ 9 1970 153154 . , _ sur la structure dimensionelle des ensembles ferms _ , _ c.r . acad .
paris _ , 186 ( 1928 )
. 420422 . , _ stable foliations of 4-manifolds by closed surfaces .
i. local structure and free actions of finite cyclic and dihedral groups on surfaces _ , invent .
( 1973 ) , 321348 . , _ foliations of codimension 2 with all leaves compact _ , manuscripta math . ,
18 , 187212 , 1976 e. vogt , _ foliations of codimension 2 with all leaves compact on closed 3- , 4- , and 5-manifolds _ math . z. 157 ( 1977 ) , no .
3 , 201223 . , _ a periodic flow with infinite epstein hierarchy _ , _ manuscripta math .
_ , 22 , 403412 , 1977 . , _ foliations with few non - compact leaves _ , _ algebr .
_ 2 ( 2002 ) , 257284 . , _ periodicity and decomposability of basin boundaries with irrational maps on prime ends _ , trans .
( 1991 ) , 303317 . , _
recurrence , pointwise almost periodicity and orbit closure relation for flows and foliations _ , _ topology appl .
_ 160 ( 2013 ) pp .
21962206 . ,
_ @xmath0-closed homeomorphisms on surfaces _ , _ topology appl . _ 160 ( 2013 ) pp .
18701875 . , _
minimal sets of @xmath0-closed surface homeomorphisms _ , _ c. r. acad .
paris , ser . _
i 350 ( 2012 ) 10511053 .
zhukova , _ local and global stability of compact leaves and foliations _ , zh . mat .
anal . geom . 9
( 2013 ) , no .
3 , 400420 . | in this paper , we study @xmath0-closed foliations which are generalization of compact hausdorff foliations .
we show that the class space of a codimension - two - like @xmath0-closed foliation @xmath1 ( resp .
group - action ) on a compact connected manifold @xmath2 is a surface with conners , which is a generalization of the codimension two compact foliation cases , where the class space is a quotient space @xmath3 defined by @xmath4 if @xmath5 ( resp .
@xmath6 ) .
moreover , a nontrivial @xmath0-closed flow on a connected compact @xmath7-manifold is either `` almost two dimensional '' or `` almost one dimensional '' or with `` complicated '' minimal sets .
in addition , a homeomorphism on a hyperbolic compact surface @xmath8 is periodic if and only if it is @xmath0-closed ( i.e. the orbit class space is hausdorff ) .
each minimal set of an @xmath0-closed non - minimal non - periodic toral homeomorphism is a finite disjoint union of essential circloids . |
diffusion is a fundamental irreversible process intervening in the evolution of many out of equilibrium systems . at sufficiently large scales
it is successfully described by fick s law obtained from non - equilibrium thermodynamics @xcite and , at shorter time scales , by means of kinetic equations of the fokker - planck type derived from mesoscopic non - equilibrium thermodynamics @xcite .
when the particles diffuse through a heterogeneous or complex medium , the structure and the physicochemical properties of the medium may give rise to the appearance of additional kinetic processes that affect the dynamics of the particles@xcite .
an important example of this situation , with applications ranging from biomedicine to engineering , is diffusion in the presence of adsorption in which particles moving through the pores of the material have certain probability to become attached at the walls of the pores @xcite .
two possibilities can be analyzed : total or partial immobilization of the particles at the surface @xcite .
the kinetics is thus dominated by a set of transport equations that may present different relaxation time scales depending on the physicochemical properties of the surface and of the bulk fluid @xcite .
our purpose in this article is to propose a simple model able to analyze the mutual influence of diffusion and adsorption processes in micro and nanopores , a situation observed in zeolites , bentonites or in new engineered highly porous materials @xcite .
we are interested in provide a simple theoretical model able to explain , at least in qualitative form , how this mutual influence is responsible for interesting effects observed in experiments @xcite and recently studied via molecular dynamics and cellular automaton techniques on the diffusion properties of the particles in the pore @xcite .
examples of these effects are the non - monotonic behavior of the diffusivity as a function of particle concentration or particle loading of the pore , or even the strong suppression of self - diffusivity as the number of particles in the pore increases @xcite .
the model describes diffusion of one component in another at rest by means of the fick law and adsorption through a generalized langmuir equation obtained from non - equilibrium thermodynamics @xcite .
using the fact that the concentration of the diffusing particles depends on that of the adsorbed particles , due to mass conservation , we formulate an effective medium theory in which the overall process is described as effective diffusion , with a diffusion coefficient depending on the characteristics of the adsorption process .
our analysis on the effects of adsorption on bulk diffusion is complementary to the usual theoretical approach to the problem which makes emphasis on the surface dynamics of the adsorbed phase @xcite.here , we are considering the case of isothermal adsorption .
the generalization to the case when non - isothermal effects , such as pervaporation associated to the mass transport through a zeolite membrane could also be performed by following a non - equilibrium thermodynamics analysis similar to that of ref .
[ 30 ] .
the effective diffusion coefficient is calculated for a particular adsorption kinetics .
it is shown that it includes a factor accounting for its deviation for the case of absence of adsorption . at high concentrations of adsorbed particles
, a virial expansion is proposed to estimate the effects of interactions .
we conclude that the effective transport properties of the system strongly depend on the adsorption - desorption rate constants as well as on temperature through the corresponding virial coefficients .
the great multiplicity of behaviours found reveals the strong influence that one of these proceses exerts on the other .
the paper is organized as follows . in section 2 ,
we present the model and analyze the case of total immobilization of the adsorbed phase .
section 3 is devoted to study the effect of surface diffusion of the adsorbed phase on the diffusion / adsorption process .
finally , our conclusions are presented in section 4 .
non - equilibrium thermodynamics provides a general framework to describe both adsorption and diffusion process and to analyze their mutual influence @xcite . in many instances , diffusion and adsorption
occur simultaneously .
particles in the bulk phase diffuse and arrive at the surface where they become attached due to the presence of molecular forces .
this last process corresponds to adsorption and has been modeled by considering it as a diffusion through a potential barrier @xcite . .
the light grey circles represent the mobile particles at time @xmath0 that become adsorbed at time @xmath1 .
the lines schematically represent the trajectories they follow .
the black circles represent the particles adsorbed at time @xmath1 ] .
[ fig0 ] when the pore is very narrow , the particles practically move along the pore , in a thin central layer ( see fig .
1 ) . under these conditions transport is quasi - onedimensional and the corresponding descripton can be performed in terms of averaged concentrations depending on the coordinate along the axis of the pore .
it is convenient to stress here that we will only consider the case when the diffusing particles have a coherent structure , such that their geometry can be characterized by an effective radius which is much larger than the molecules of the solvent .
this assumption avoids to consider the rotational dynamics of the particles with more complicated geometries . after averaging the three dimensional diffusion equations over @xmath3-directions
, it can be shown that the evolution equation for the average number concentration @xmath4 in the bulk is given by @xcite @xmath5 where @xmath6 is the diffusion current of particles having dimensions of number of particles per unit area and time along the @xmath7-direction , and @xmath8 has dimensions of number of particles divided by volume and time and represents the net amount of particles leaving the bulk by crossing the dotted lines indicated in figure 1 .
in similar way , the evolution equation for the concentration @xmath9 of the adsorbed phase is @xmath10 here @xmath11 is the corresponding diffusion current density and we used the fact that the net amount of particles arriving at the surface is @xmath12 and that @xmath13 is the reaction term associated to chemisorption . in the following , for simplicity we will only consider the case of physical adsorption , that is , when @xmath14 .
further details on the derivation of eq .
( [ consmasa2 ] ) are given in ref .
the expression for the currents entering eqs .
( [ consmasa1 ] ) and ( [ consmasa2 ] ) follows from the entropy production rate per unit mass @xmath15 of the system @xcite .
this quantity is given by @xmath16 , where the contribution associated to the bulk is @xmath17 and the entropy production at the surface is given by @xcite @xmath18 where @xmath19 ( @xmath20 ) represent the corresponding non - equilibrium chemical potentials .
it is convenient to mention that , by hypothesis , no chemical reactions occur in the bulk and therefore eq . ( [ sigmab ] ) has no contribution due to chemical reactions . from eqs .
( [ sigmab ] ) and ( [ sigmas ] ) and assuming linear relationships between fluxes and forces , one finds : @xmath21 , @xmath22 , @xmath23 . in the dilute case at constant temperature
, the last relation implies that the flow of particles from the bulk to the surface @xmath8 becomes proportional to the pressure difference which in turn is proportional to the concentration ( assuming henry s law in the bulk ) .
this flow has two contributions , @xmath24 , one for the particles leaving the bulk , @xmath25 , and one for the particles arriving to the bulk from the surface , @xmath26 .
the first contribution is proportional to a constant rate @xmath27 giving the probability per unit time that the particles in the bulk is adsorbed by the surface .
however , this probability must depend on the available surface . in langmuir approach
, this factor was assumed as proportional to the concentration of particles at the surface : @xmath28 with @xmath29 the maximum of adsorbing sites at the surface .
however , if interactions exist among these particles , then it depends upon the non - linear function @xmath30 which takes into account excluded volume interactions of the adsorbed phase at the surface @xcite .
that is , @xmath31 with @xmath32 for ideal systems and @xmath30 arbitrary in the general case .
hence , the flow @xmath25 can be written as : @xmath33 ) @xmath34 for langmuir and @xmath35 ) @xmath36 for the non - ideal case . in both expressions
@xmath37 represents the adsorption rate constant .
the second contribution @xmath26 is proportional to the covered surface and is usually expressed as @xmath38 , where @xmath39 is the desorption rate constant . using these relations in eqs .
( [ consmasa1 ] ) and ( [ consmasa2 ] ) one obtains @xmath40 + j,\ ] ] for the mass balance in the bulk , whereas for the adsorbed particles in the nonlinear case we have @xmath41.\ ] ] in writing eqs .
( 5 ) and ( 6 ) we are considering the concentrations @xmath42 and @xmath43 in terms of volume .
this allows us to use the mass conservation condition @xmath44 , in order to considerably simplify the description with out loss of generality , and takes into account the fact that surface diffusion is strictly a three dimensional process that occurs in a small volume surrounding the surface of the solid .
thus , eqs .
( [ r - d1 ] ) and ( [ r - d2 ] ) , together with mass conservation @xmath44 , describe the dynamics of the particles in the pore .
it is convenient to mention that the adsorption rate constant follows an arrhenius law in which this quantity depends on temperature ref .
4 . as a consequence , the interplay between adsorption and diffusion could be controlled by changing the temperature of the bath . instead of proceeding with the numerical solution of eqs .
( [ r - d1 ] ) and ( [ r - d2 ] ) , we will use mass conservation in order to adopt a simpler effective medium description for the interplay between adsorption and diffusion processes in which the boundary term @xmath8 in eq .
( [ r - d1 ] ) does not appears explicitly .
this can be done because this term is coupled to the reaction term in ( [ r - d2 ] ) , which in fact determines a relation between the concentrations @xmath42 and @xmath43 corresponding to the adsorption isotherm .
for instance , in the stationary case with no surface diffusion effects , @xmath8 is strictly zero .
however , even in this case , we now that the diffusion of particles in the bulk is modified by the adsorption process , in such a way that @xmath45 . using the mass conservation relation and the corresponding adsorption isotherm , this effect can be incorporated through an effective diffusion coefficient @xmath46 . @xcite the effective diffusion current for the total number concentration is @xmath47 note that the effective diffusion coefficient @xmath46 can be associated to an effective activity @xmath48 in such a way that we may write @xmath49,\ ] ] where @xmath50 is the effective activity coefficient and @xmath51 is the mobility which in the present case can be well approximated by the stokes friction coefficient , @xmath52 where @xmath53 is the dynamic viscosity of the solvent and @xmath54 is the effective radius of the particle .
using equations ( [ immfluxtotal ] ) and ( [ termoddf ] ) in the mass conservation law for the bulk yields @xmath55 .
\end{aligned}\ ] ] it remains now to determine the expression of @xmath46 which , after using the adsorption kinetics and the mass conservation condition , can be expressed as a function of the total concentration @xmath56 .
this will be performed by analyzing different cases in the next section .
the case in which adsorbed particles are completely immobilized follows from the general model by imposing that the effective diffusion current @xmath57 , referring initially only to the mobile fraction @xmath42 is given by @xmath58 the diffusion process can also be described by means of the effective fick s law ( [ immfluxtotal ] ) .
this fact implies a relationship between the diffusion coefficients @xmath59 and @xmath46 that can be established by considering the constraint @xmath44 , with the subindex @xmath60 corresponding to the immobile fraction . since @xmath43 depends on @xmath42 through the adsorption isotherm : @xmath61 , this gives @xmath62 or equivalently @xmath63 .
thus , using the equality @xmath64 , we finally obtain @xmath65 where the factor @xmath66 accounts for deviations of the diffusion coefficient with respect to its value in the case of absence of adsorption .
the influence of adsorption on diffusion enters through the evolution in time of the concentration of the immobilized particles @xmath43 . an interesting analysis on how distinct kinetic regimes may appear in adsorption - diffusion interplay
is discussed in ref .
this dependence can be determined by means of the generalized langmuir adsorption equation @xcite @xmath67 when adsorption occurs in time scales much shorted than the characteristic time of diffusion , @xmath68 , we may assume that the adsorbed phase is practically in equilibrium . under this condition , one has from ( [ kinetic - langmuir ] ) @xmath69 where @xmath70 . for the case of a non - equilibrium stationary state in which @xmath71 , with @xmath72 a constant current of particles , one obtains @xmath73 [ [ low - coverage ] ] low coverage : when the concentration of the adsorbed particles is small we may assume @xmath74 , with @xmath75 the inverse value of the total concentration of adsorbing sites @xmath76 in the system@xcite . then , eq . (
[ kinetic - langmuir ] ) yields the langmuir isotherm @xcite @xmath77 using eq .
( [ langmuir ] ) in the condition @xmath62 , one obtains @xmath78 where @xmath79 is the total concentration scaled by the total concentration of adsorbing sites . from eq .
( [ c1delangmuir ] ) , the correction factor @xmath80 becomes @xmath81 in fig [ fig2 ] , the correction factor @xmath80 is represented as a function of the total concentration @xmath56 for different values of @xmath82 in the equilibrium case . from now on , we will assume for simplicity that @xmath83 .
values of @xmath84 different from the unity only rescale the results , the qualitative behavior of the functions remains the same .
when adsorption is dominant , as occurs for @xmath85 , the effective diffusion coefficient becomes negligible for concentrations @xmath86 .
increasing the importance of desorption , @xmath87 , makes @xmath80 more significant taking values of about @xmath88 to @xmath89 of @xmath90 .
as expected , this tendency becomes even more manifest when desorption is faster than adsorption , @xmath91 . from ( [ deltaclangmuir ] ) in terms of concentration for different values of the parameter @xmath82 .
solid lines correspond to equilibrium adsorption ( [ langmuir ] ) whereas dashed and dotted lines correspond to the non - equilibrium stationary case ( [ deltaclangmuirstationary ] ) for @xmath92 and @xmath93 , respectively .
the values of @xmath82 used were ( from red to purple ) @xmath94 and @xmath83 . ] in similar form , for the stationary case we have @xmath95 and therefore the correction factor @xmath96 is in this case @xmath97 here we have introduced the parameter @xmath98 , having dimensions of number concentration . dashed and dotted lines in fig [ fig2 ] show the stationary non - equilibrium values of @xmath80 as a function of the total concentration @xmath56 for the same values of @xmath82 in the equilibrium case and for @xmath92 ( dashed ) and @xmath93 ( dotted ) .
the stationary adsorption process is still dominant for @xmath85 , but in this case the effective diffusion coefficient takes values considerably larger than those of the equilibrium case , @xmath99 for @xmath100 and @xmath101 for @xmath102 . as in the equilibrium case , this effect decreases when increasing @xmath82 .
this behavior of the correction factor @xmath80 as a function of particle concentration in the bulk is similar to the one observed for the effective diffusion coefficient in experiments and simulations @xcite .
the relation between both quantities can be established by means of eq .
( [ imm_flux2 ] ) .
unfortunately , a quantitative comparison is not possible due to the fact that reported results in simulations are given in terms of the so - called loading parameter measuring the number of particles in the pore .
[ [ high - coverage ] ] high coverage : when the concentration of adsorbed particles is higher , one has to take into account nonlinear contributions to @xmath103 through the virial expansion @xmath104 where @xmath105 is a constant having dimensions of square volume and may take positive or negative values depending on the nature of the interactions between adsorbed - adsorbed particles and may also depend on temperature .
it is convenient to mention that the expansion ( [ z2interaccion ] ) is consistent with a second order virial expansion in which interactions ( excluded volume and direct ones ) between adsorbed particles are taken into account .
this is the case since the factor @xmath103 plays the role of a fugacity . more important
is that eq .
( [ z2interaccion ] ) allows us to go beyond the classical langmuir model .
using this relation , we obtain the following adsorption isotherm in terms of the total concentration @xmath56 , for @xmath106 and different values of @xmath107 ( from red to purple ) and @xmath83 . _
i _ ) the solid lines represent the behavior of @xmath108 for equilibrium adsorption whereas dashed lines correspond to the ideal case .
interactions between the adsorbed particles introduce a nonmonotonic behavior of the effective diffusion coefficient for @xmath109 , which depends on the value of @xmath105 .
_ ii _ ) the nonideal correction factor for @xmath110 , @xmath83 and @xmath111 ( from red to orange).,title="fig : " ] in terms of the total concentration @xmath56 , for @xmath106 and different values of @xmath107 ( from red to purple ) and @xmath83 .
_ i _ ) the solid lines represent the behavior of @xmath108 for equilibrium adsorption whereas dashed lines correspond to the ideal case .
interactions between the adsorbed particles introduce a nonmonotonic behavior of the effective diffusion coefficient for @xmath109 , which depends on the value of @xmath105 . _
ii _ ) the nonideal correction factor for @xmath110 , @xmath83 and @xmath111 ( from red to orange).,title="fig : " ] @xmath112,\ ] ] where the subindex @xmath113 indicates the non ideal case under consideration .
note that eq .
( [ langmuirnoideal ] ) is valid for @xmath114 .
substitution of this equation into the mass balance leads to the relation @xmath115.\ ] ] this equation is a third degree polynomial equation for @xmath116 that has a real solution from which the correction factor @xmath108 can be calculated .
the resulting relations can be used to estimate the corrections due to the non - ideal nature of the adsorbed phase . the condition @xmath117 , obtained from eqs .
( [ langmuirnoideal ] ) and ( [ c1noideal ] ) , guarantees the existence of real values for the concentration @xmath56 .
figure [ fig3]_i _ ) shows the nonideal correction factor @xmath108 in terms of the total concentration @xmath56 by assuming @xmath118 . for @xmath119 ,
interactions reduce the magnitude of the effective diffusion coefficient although it increases monotically with @xmath56 .
however , for @xmath120 , interactions between particles of the adsorbed phase may cause a non - monotonic behavior of the effective diffusion coefficient for certain values of @xmath105 , making it to decrease for larger values of @xmath56 as it is shown in detail in figure [ fig3]_ii _ ) for which @xmath121 .
this behavior implies that an optimal value for the concentration exists for which the effective diffusion coefficient reaches its maximum value . for @xmath122 , interactions monotonically reduce the effective diffusion coefficient as @xmath56 increases . as in the previous case ,
the non - monotonic behavior manifested by the correction factor agrees well with the non - monotonic behavior of the effective diffusion coefficient observed in simulations and experiments@xcite .
the case in which the adsorbed particles may diffuse on the surface of the sorbent can be analyzed by adding a diffusion term to eq .
( [ kinetic - langmuir ] ) , @xcite .
the evolution equation for @xmath43 takes the form @xmath123,\ ] ] where @xmath124 is the diffusion coefficient at the surface .
an estimation of the contribution of surface diffusion to the adsorption - diffusion interplay can be performed by assuming that @xmath125 where @xmath126 is the characteristic diffusion time and @xmath127 a characteristic length of the sorbent surface . in first approximation , the diffusion term in ( [ kinetic - langmuir2 ] ) can be expressed as @xmath128 \sim \frac{c_2}{\tau_d}.\ ] ] this type of approximation is frequently and succesfully used in transport processes problems to examine and estimate , for instance , viscous or thermal effects near to a surface . at low coverage , when the adsorbed phase is dilute , one can make the appproximation @xmath129 .
( [ kinetic - langmuir2 ] ) reduces to @xmath130 where @xmath131 .
this equation leads to the adsorption isotherm @xmath132 where now @xmath133 . therefore , in the dilute case the influence of surface diffusion on adsorption is to reduce the desorption rate . as a consequence
, the values of the effective diffusion coefficient of the system @xmath46 become shifted to lower concentrations . in the nondilute case ,
both @xmath103 and @xmath134 must include non - linear contributions .
the expression of @xmath103 is given in eq .
( [ z2interaccion ] ) . according to eq .
( [ termoddf ] ) , the activity coefficient can be written in terms of the non - ideal contribution of the chemical potential : @xmath135 .
a virial expansion of @xmath136 at the lower order in concentration leads to the relation @xmath137 , where @xmath138 has dimensions of volume and @xmath139 is the second virial coefficient which must take into account excluded volume effects and direct interactions . in terms of the total concentration @xmath56 for @xmath106 , @xmath140 , @xmath141 and different values of @xmath70 and @xmath142 . in [ fig4]_i
_ ) the solid lines represent the behavior of @xmath143 for equilibrium adsorption whereas dashed lines correspond to @xmath108 for @xmath144 ( low temperatures ) and @xmath145 ( from red to purple ) .
the nonmonotonic behavior of the effective diffusion coefficient is more pronounced for @xmath146 .
_ ii _ ) shows @xmath147 for @xmath148 ( high temperatures ) and @xmath149 ( from red to purple ) and @xmath83 . for the present model , high values of @xmath150 may induce a discontinuity of the correction factor.,title="fig : " ] in terms of the total concentration @xmath56 for @xmath106 , @xmath140 , @xmath141 and different values of @xmath70 and @xmath142 . in [ fig4]_i
_ ) the solid lines represent the behavior of @xmath143 for equilibrium adsorption whereas dashed lines correspond to @xmath108 for @xmath144 ( low temperatures ) and @xmath145 ( from red to purple ) .
the nonmonotonic behavior of the effective diffusion coefficient is more pronounced for @xmath146 .
_ ii _ ) shows @xmath147 for @xmath148 ( high temperatures ) and @xmath149 ( from red to purple ) and @xmath83 . for
the present model , high values of @xmath150 may induce a discontinuity of the correction factor.,title="fig : " ] using the previous expressions for @xmath103 and @xmath151 , one obtains the generalized langmuir equation @xmath152 where now the effective desorption rate is @xmath153 $ ] .
the coefficient @xmath142 can be positive or negative depending on the temperature . in the non - equilibrium stationary case one
obtains from eq .
( [ kinetic - langmuir4 ] ) @xmath154.\ ] ] where @xmath155 . in figure [ fig4 ]
, we show the nonideal correction factor @xmath143 for @xmath156 in the case of equilibrium adsorption with surface diffusion in terms of the total concentration @xmath56 for low , fig .
[ fig4]_i _ ) , and high temperatures , fig .
[ fig4]_ii _ ) . as in fig .
2 , interactions induce a non - monotonic behavior but in the present case , for @xmath157 and low temperatures ( @xmath158 ) , they lead to an enhancement of diffusion for large concentrations due to the attractive nature of direct interactions of the particles in channel . at low concentrations ,
diffusion is significatively lower than in the case of fig .
when the concentration is higher and the temperature is such that @xmath159 , excluded volume effects cause suppression of diffusion , as observed in the figure .
in this article , we have analyzed the combined effect of diffusion and adsorption in porous materials by adopting an effective medium description based on non - equilibrium thermodynamics @xcite .
we have treated the case of very narrow pores , present in some porous materials , microfluidic devices and in biological channels , in which particles move essentially along the direction of the pores and may occasionally become attached to their surface .
the fact that the pores are narrow , makes it possible to adopt a quasi - onedimensional description in terms of an average concentration along the channel and a kinetic mechanism governing the attachment - detachment of the particles to the surface of the channel .
the model proposed considers the combined effect of these two mechanisms and analyzes particle transport by means of effective diffusion in which the diffusion coefficient depends on the adsorption kinetics described in turns by a generalized langmuir equation .
the analysis of the effective diffusion coefficient reveals the existence of a multiplicity of different regimes controlled by different adsorption - desorption kinetics and different coverages of particles at the surface .
we have quantified the change of the diffusion coefficient and found a non - monotonic behavior of this quantity due to excluded volume interactions .
these results are in qualitative agreement with previous ones obtained from numerical models .
the model proposed could be extended to the case of different shapes of the pores by considering entropic barriers emerging from the variation of the cross section of the pores .
this generalization can be performed following the approach presented in refs .
the resulting model may represent a simple and useful tool to analyze particle transport in nanoporous materials and biological channels .
this work was supported by the dgicyt of spanish government under grant no .
fis2008 - 04386 and by unam dgapa grant
in100112 - 2 .
ish appreciates the hospitality of prof .
j. m. rubi and the department of fsica fonamental of the university of barcelona where this work was done during a sabbatical leave .
j. m. rubi acknowledges financial support from generalitat de catalunya under program icrea academia .
z.j.grzywna acknowledges support from mniszw under the project nn508409137 .
40 s. de groot , p. mazur , _ non - equilibrium thermodynamics _
( dover , new york , 1984 ) .
d. reguera , j.m .
rubi , j.m.g .
vilar , the mesoscopic dynamics of thermodynamic systems , j. phys . chem .
b * 109 * ( 2005 ) , 21502 .
j. m. rubi , a. prez - madrid , inertial effects in non - equilibrium thermodynamics , physica a * 264 * ( 1999 ) , 492 .
i. pagonabarraga , j. m. rubi , derivation of the langmuir adsorption equation from non - equilibrium thermodynamics , physica a * 188 * ( 1992 ) , 553 .
d. r. paul and w. j. koros , effect of partially immobilizing sorption on permeability and the diffusion time lag , j. pol .
* 14 * ( 1976 ) , 675 .
b. smit and t. l. maesen , molecular simulations of zeolites : adsorption , diffusion , and shape selectivity , chem
* 108 * ( 2008 ) , 4125 .
hu , c .- s .
chang , r .- c .
ruaan and j .- y .
lai , effect of free volume and sorption on membrane gas transport , j. membr .
* 226 * ( 2003 ) , 51 .
v. saxena and s. a. stern , concentration - dependent transport of gases and vapors in glassy polymers , j. membr .
* 12 * ( 1982 ) , 65 .
j. wloch and j. kornatowski , sorption rate and thermal barriers in a gas - zeolite system : investigation of n - hexane sorption in mfi - type zeolite , langmuir * 20 * ( 2004 ) , 1180 .
g. k. papadopoulos , h. jobic and d. n. theodorou , transport diffusivity of @xmath160 and @xmath161 in silicalite : coherent quasielastic neutron scattering measurements and molecular dynamics simulations , j. phys .
b * 108 * ( 2004 ) , 12748 .
b. baranowski , non - equilibrium thermodynamics as applied to membrane transport , j. membr .
* 57 * ( 1991 ) 119 .
j. krger , d. m. ruthven , _ diffusion in zeolites and other microporous materials _
( wiley , new york , 1992 ) .
roque - malherbe , _ adsorption and diffusion in nanoporous materials _
( crc pres , boca raton , 2007 ) .
k. j. klabunde , r. m. richards , _ nanoscale materials in chemistry _ ( wiley , new york , 2009 ) . j. karger and d. m. ruthven , diffusion in zeolites .
comparison of sorption and nuclear magnetic resonance diffusivities , j. chem .
faraday trans . 1 * 77 * ( 1981 ) , 1485 .
e. a. brinkmann and r. p. king , a kinetic model framework for combined diffusion and adsorption processes , saimm * 2 * ( 1987 ) , 189 .
n. l. filippova , kinetic - diffusion - controlled adsorption and desorption kinetics on planar surfaces j. coll .
* 206 * ( 1998 ) 592 .
a. garca cant ros , j .- s .
mcewen , p. gaspard , effect of ultrafast diffusion on adsorption , desorption , and reaction processes over heterogeneous surfaces , phys .
e * 83 * ( 2011 ) , 021604 .
j. l. vazquez , _ the porous medium equation _ ( oxford university press , 2006 ) .
z. j. grzywna , j. podkowka , effect of immobilizing adsorption on mass transport through polymer films , j. membr .
* 8 * ( 1981 ) , 23 .
s. brandani , j. hufton and d. ruthven , self - diffusion of propane and propylene in 5a and 13x zeolite crystals studied by the tracer zlc method , zeolites * 15 * ( 1995 ) , 624 . f. martin et .
al , tailoring width of microfabricated nanochannels to solute size can be used to control diffusion kinetics , j. cont .
* 102 * ( 2005 ) 123 .
j. f. douglas , h. e. johnson , s. granick , a simple kinetic model of polymer adsorption and desorption , science * 262 * ( 1993 ) 2010 . j .-
simon , j. m. rubi , temperature at small scales : a lower limit for a thermodynamic description , j. phys .
b * 115 * ( 2011 ) , 1422 .
h. jost , _ diffusion in solids , liquids and gases _ ( academic press , new york , 1960 ) .
sheta , a.m. falatah , m.s .
al - sewailem , e.m .
khaled , a.s.h .
sallam , sorption characteristics of zinc and iron by natural zeolite and bentonite , microporous and mesoporous materials * 61 * ( 2003 ) , 127 . o. m. yaghi , m. okeeffe , n. w. ockwig , h. k. chae , m. eddaoudi , j. kim , reticular synthesis and the design of new materials , nature * 423 * ( 2003 ) , 705 .
p. demontis , f. g. pazzona , g. b. suffritti , diffusion in tight connement : a lattice - gas cellular automaton approach .
transport properties , * 126 * ( 2007 ) 194710 .
p. demontis , f. g. pazzona , g. b. suffritti , introducing a cellular automaton as an empirical model to study static and dynamic properties of molecules adsorbed in zeolites , j. phys .
b * 112 * ( 2008 ) 12444 .
a. m. pintus , f. g. pazzona , p. demontis , g. b. suffritti , a parallelizable block cellular automaton for the study of diffusion of binary mixtures containing co2 in microporous materials , j. chem . phys .
* 135 * ( 2011 ) 124110 .
j. kuhn , r. stemmer , f. kapteijn , s. kjelstrup , j. gross , a non - equilibrium thermodynamics approach to model mass and heat transport for water pervaporation through a zeolite membrane , j. membr .
* 330 * ( 2009 ) , 388 .
d. reguera , j. m. rubi , kinetic equations for diffusion in the presence of entropic barriers , phys .
e * 64 * ( 2001 ) , 061106 .
d. reguera , g. schmid , p. s. burada , j. m. rubi , p. reimann , p. hanggi , entropic transport : kinetics , scaling , and control mechanisms , phys .
* 96 * ( 2006 ) , 130603 . | a non - equilibrium thermodynamics model able to analyze the combined effect of diffusion and adsorption in porous materials is proposed .
the model considers the coupled dynamics of the diffusive phase , described by a diffusion type equation , and the adsorbed phase governed by a generalized langmuir equation .
it is shown that the combination of diffusion and adsorption can be treated as diffusion in an effective medium , with an effective diffusion coefficient depending on the characteristics of the adsorption kinetics .
the interplay between these effects leads to the appearance of different regimes in which the effective diffusion coefficient exhibits peculiar behaviours such as a decrease due to the presence of the adsorption process and a non - monotonous behavior resulting from particle interactions .
we outline applications of the model developed to diffusion in porous materials such as zeolites and bentonites , and to engineered highly porous materials . |
the experimental determination of the cabibbo - kobayashi - maskawa ( ckm ) @xcite matrix elements is vital because they control the hadronic sector of the standard model , determining cp violation and flavour mixing .
the magnitudes and phases of these matrix elements must be established to make the standard model predictive , to determine the angles and area of the unitarity triangle , and to look for any inconsistency between the standard model and the experimental data which would point to new physics .
the ckm element @xmath9 is one of the most poorly known , currently uncertain to within a factor of two or three ( at 90% cl ) in magnitude @xcite .
the determination of @xmath5 has traditionally been made from inclusive @xmath10 semileptonic decays @xcite , looking at the lepton energy spectrum beyond the endpoint for decays to charmed final states .
these determinations rely on models incorporating nonperturbative qcd effects to relate the measured spectrum to theoretical predictions .
models used include the quark model @xcite and bound - state models @xcite@xcite together with attempts to combine features from both @xcite .
this results in the inclusion of an additional error reflecting the range of models used @xcite . in this paper
we propose a model - independent method to determine @xmath5 from the exclusive semileptonic charmless @xmath11 decay @xmath0 , which should be measured with improved accuracy in @xmath11 factories and at @xmath12 and hadron colliders in the near future .
currently , only an upper bound exists @xcite for the total decay rate , but new experimental results should be available soon @xcite .
we will show how lattice qcd calculations , which incorporate nonperturbative qcd effects in a systematic way , can be used to extract @xmath5 from experimental measurements of @xmath0 .
exclusive @xmath11 decays have already proved useful for extracting the value of @xmath13 by studying the process @xmath14
@xcite@xcite .
we have also analysed relations between the exclusive processes @xmath15 and @xmath0 which follow from heavy quark symmetry ( hqs ) @xcite .
we discuss the utility of these relations for determining @xmath5 from the experimental measurements of @xmath16 and @xmath17 @xcite@xcite .
the matrix elements we will be considering are of the @xmath18 weak current between @xmath6 and @xmath19 mesons and of the magnetic moment operator between @xmath6 and @xmath20 mesons @xcite .
for @xmath0 the matrix element is , @xmath21 with form factor decomposition , @xmath22 where @xmath23 is the four - momentum transfer and @xmath24 is the @xmath19 polarisation vector .
the form factor @xmath25 can be written as @xmath26 where , @xmath27 with @xmath28 . in the limit of zero lepton masses ,
the term proportional to @xmath25 in equation ( [ eq : btorho2 ] ) does not contribute to the total amplitude and hence to the decay rates .
pole dominance models suggest that @xmath29 , @xmath30 for @xmath31 and @xmath32 correspond to @xmath33 , @xmath34 and @xmath35 exchanges respectively in the @xmath36-channel @xcite .
the main contribution to the @xmath15 decay comes from the matrix element @xmath37 where @xmath38 as above , @xmath24 is now the @xmath20 polarisation vector and @xmath39 for an on - shell photon with @xmath40 , @xmath41 does not contribute to the @xmath15 amplitude and @xmath42 and @xmath43 are related by , @xmath44 hence , for @xmath15 , we need to determine @xmath42 and/or @xmath43 at the on - shell point . neglecting corrections suppressed by inverse powers of the heavy quark mass @xmath45 ,
the following relations hold when @xmath46 is close to the maximum recoil value @xmath47 @xcite @xmath48 where @xmath49 arises from the leading logarithmic corrections and is chosen to be 1 at the @xmath11 mass @xcite , @xmath50 in the calculations reported below , we will use @xmath51 in the quenched approximation and @xmath52 .
the matrix elements of the @xmath18 current and the magnetic moment operator between @xmath6 and identical and @xmath53 quarks . ]
light final - state vector mesons of mass @xmath54 are related in the infinite heavy quark mass limit according to : @xmath55 for values of @xmath46 not too far from @xmath56 , or equivalently , for @xmath57 close to @xmath58 , where @xmath59 here , @xmath60 and @xmath61 are the four - velocities of the heavy - light pseudoscalar meson and the light vector meson respectively .
the equations in ( [ eq : vt1andat2 ] ) relate the physical decay processes @xmath0 and @xmath62 .
if light flavour @xmath63 symmetry is respected then equation ( [ eq : vt1andat2 ] ) relates the processes @xmath0 and @xmath15 @xcite . in the lattice calculations reported below we will test equation ( [ eq : vt1andat2 ] ) directly for different values of the heavy quark mass using identical light meson states for both matrix elements . in the literature ,
versions of equation ( [ eq : vt1andat2 ] ) appear which rely on lowest order hqs but incorporate all @xmath64 corrections from kinematics @xcite corrections it is necessary to consider operators of dimension four in the heavy quark expansion in addition to those of dimension three occurring at leading order . ] .
the modified relations become : @xmath65 2it_2(q^2 ) & = & { \big[(m{+}m)^2 - q^2\big ] \big[(m{-}m)^2 - q^2\big]\over2m(m^2{-}m^2 ) } { v(q^2)\over m+m } \nonumber \\ & & \mbox { } + { q^2 + m^2 - m^2 \over 2 m}{a_1(q^2)\over m - m } \label{eq : at2kinematic}\end{aligned}\ ] ]
lattice calculations with propagating quarks provide matrix elements for heavy quarks around the charm mass over a range of @xmath46 straddling @xmath40 . in extracting form factors from these matrix elements
, we can reach @xmath56 for @xmath66 and @xmath43 only , because the coefficients determining the contribution of the other form factors to the matrix elements of equations ( [ eq : btorho1 ] ) and ( [ eq : btokstargamma ] ) vanish at this kinematic point . to obtain results relevant for @xmath11
decays , we need to extrapolate in the heavy quark mass @xmath45 to the @xmath67 quark mass .
this is simple for fixed @xmath57 , but , at the @xmath11 scale , produces a range of @xmath46 values near @xmath56 and far from @xmath68 .
the results described below come from @xmath69 @xmath63 gauge configurations generated by the ukqcd collaboration on a @xmath70 lattice at @xmath71 in the quenched approximation .
the @xmath72 improved sheikholeslami
wohlert ( sw ) @xcite action was used for fermions , with `` rotated '' fermion fields appearing in all operators used for correlation function calculations @xcite .
the inverse lattice spacing determined from the @xmath19 mass is @xmath73 @xcite .
other physical quantities will lead to slightly different values for the lattice spacing ( @xmath74@xmath75 @xcite ) .
the scale uncertainty should be reflected in the results for dimensionful quantities .
three - point correlators of the heavy - to - light two - fermion operators with a heavy pseudoscalar meson ( the `` @xmath6 '' meson ) and a light vector meson were calculated , as illustrated in figure [ fig : threeptcorrelator ] .
matrix elements were extracted from these correlators by the method detailed in @xcite@xcite .
four heavy - quark hopping parameters , @xmath76 , were used . for the propagator connecting the current operator to the light meson operator ,
two kappa values , @xmath77 , were available .
the subscript @xmath78 is for `` active '' to contrast this propagator with the `` spectator '' propagator joining the heavy - light meson to the light meson .
these @xmath79 values straddle that for the strange quark , @xmath80 @xcite . for @xmath81
, we used three light spectator hopping parameters , @xmath82 , and for @xmath83 we used @xmath84 only .
the critical hopping parameter at this @xmath85 is @xmath86 @xcite . to the lattice calculations were performed with the heavy meson spatial momentum of magnitude @xmath87 or @xmath58 , in lattice units of @xmath88 .
the momentum injected at the operator insertion was varied to allow the modulus of the light meson spatial momentum to take values up to @xmath89 in lattice units ( although some of the momentum choices were too noisy to be used in fits ) .
we refer to each combination of heavy and light meson three - momenta as a channel with the notation @xmath90 in lattice units ( for example @xmath91 or @xmath92 where the subscript @xmath93 indicates that @xmath94 and @xmath95 are perpendicular ) .
the results below have been obtained using uncorrelated fits for the extrapolations in the heavy quark mass .
the extraction of the form factors from the three - point correlation function data used correlated fits @xcite .
statistical errors are 68% confidence limits obtained from @xmath96 bootstrap samples . to make the best use of hqs , we pick momentum combinations which keep @xmath57 constant as the heavy - light meson mass varies .
this allows us to scale in @xmath64 from the charm to the bottom mass scale and gives the form factors for the @xmath6 decays as functions of @xmath57 , and therefore as functions of @xmath46 . when the heavy - light meson is at rest , @xmath57 is independent of the heavy - light meson mass .
as shown in reference @xcite there are some additional channels , where the heavy - light meson is not at rest , but @xmath94 and @xmath95 are perpendicular , for which @xmath57 is very nearly constant as the heavy meson mass varies , and which can be used for heavy quark mass extrapolations .
for those channels where @xmath57 is not strictly constant but which we nevertheless use in our study , the difference between the average @xmath57 value used in our analysis and that for each value of the heavy quark mass is always less than 3% @xcite . for relating lattice results to the continuum , we have used the perturbative values for the renormalisation constants of the vector , axial and magnetic moment operators @xcite : @xmath97 the sw improved action reduces the leading discretisation errors from @xmath98 in the wilson fermion action to @xmath99 , but for quark masses @xmath100 around that of the charm quark , @xmath101 can be of order 10% .
an estimate of the lattice artefacts can be obtained by comparing values of lattice renormalisation constants , computed from different matrix elements , in a non - perturbative way @xcite .
numerical calculations have confirmed that errors of order 10% are present at our value of @xmath85 in the matrix elements of vector and axial vector currents @xcite . we will therefore allow for an extra 10% systematic uncertainty in our results due to possible discretisation errors .
in this section we study the relations between the semileptonic and radiative decay processes of heavy - light mesons to light vector mesons . we will check the hqs relations of equation ( [ eq : vt1andat2 ] ) and the size of @xmath64 corrections to these relations for different values of the heavy - light meson mass . for testing these relations
we use our most accurate light - quark data , @xmath102 : this ensures that the light degrees of freedom are the same for the semileptonic and radiative processes and allows us to check directly , without corrections due to @xmath63 symmetry breaking , the validity of hqs as @xmath45 increases .
specifically , we compare the form factors @xmath66 and @xmath29 for the pseudoscalar to vector semileptonic decay , @xmath103 with the form factors @xmath42 and @xmath43 for the pseudoscalar to vector radiative decay , @xmath104 for different values of the heavy quark mass , @xmath105 , @xmath106 , @xmath107 and @xmath108 , and for different values of @xmath57 close to @xmath109 .
the corresponding heavy - light pseudoscalar masses are in the range @xmath110 to @xmath111 .
we have also extrapolated to the @xmath11 scale and to the infinite heavy quark mass limit .
we have used five momentum channels in our analysis and @xmath112 can be determined directly from the @xmath113 data , but @xmath114 and @xmath115 have to be obtained by an extrapolation in @xmath46 from the measured data . ] : @xmath113 , @xmath91 , @xmath116 , @xmath117 and @xmath92 . in all cases the ratios @xmath118 and @xmath119
are consistent with unity in the heavy quark mass limit , as predicted by hqs , although for certain channels the ratio @xmath118 departs from @xmath58 by up to 75% at the charm scale .
this constitutes a non - trivial test of hqs predictions .
these results are illustrated in figures [ fig : hqs1overm ] and [ fig : hqsomega ] . in figure
[ fig : hqs1overm ] we show the ratios @xmath118 ( for momentum channels @xmath91 and @xmath117 ) and @xmath119 ( for momentum channels @xmath113 and @xmath92 ) for different pseudoscalar meson masses , allowing for linear and quadratic @xmath64 corrections to the infinite mass limit predictions of equation ( [ eq : vt1andat2 ] ) .
the linear and quadratic extrapolations in @xmath64 agree within errors for the extrapolated values , with the exception of three cases of the infinite mass limit points ( one of which is shown in figure [ fig : hqs1overm ] ) .
we trust the linear extrapolations more than the quadratic ones : we are fitting to four points only in a limited region of inverse heavy - light pseudoscalar mass , and quadratic fits can amplify accidental quadratic effects in the fitted points .
all results in the remainder of this section will refer to linear extrapolations .
figure [ fig : hqsomega ] shows the ratios @xmath118 and @xmath119 for five values of @xmath57 at three different heavy - light pseudoscalar masses , around the @xmath120 mass , around the @xmath11 mass and in the infinite mass limit .
the results around the @xmath120 mass are our measured values at @xmath121 ( which corresponds very closely to the charm quark ) .
the @xmath11 mass and infinite mass limit results are extrapolations .
the hqs predictions of equation ( [ eq : vt1andat2 ] ) are well satisfied for both ratios in the infinite mass limit .
the ratio @xmath118 shows large @xmath64 corrections of the order of 75% at the @xmath120 scale and 20% at the @xmath11 scale .
corrections of about 30% at the @xmath120 scale and about 10% at the @xmath11 scale were previously observed for the pseudoscalar meson decay constant @xmath122 @xcite and for the form factor @xmath123 in heavy - to - heavy @xmath124 semileptonic decays @xcite , @xmath125 and @xmath126 . ] .
in contrast , the ratio @xmath119 exhibits small @xmath64 corrections even at the @xmath120 meson scale .
a similar situation occurs for the form factors @xmath125 and @xmath126 in heavy - to - heavy @xmath127 semileptonic decays @xcite and @xmath126 are protected from @xmath64 corrections at zero recoil by luke s theorem @xcite and the leading corrections are of order @xmath128 . ] .
inclusion of kinematic @xmath64 factors , as shown in equations ( [ eq : vt1kinematic ] ) and ( [ eq : at2kinematic ] ) reduces the size of the @xmath64 corrections by about a factor of two for the relation between @xmath42 and @xmath29 , giving results in agreement with those in reference @xcite , but has little effect on the @xmath43 and @xmath66 relation ( note that the @xmath43@xmath66 relation in equation ( [ eq : at2kinematic ] ) reduces to the result in equation ( [ eq : vt1andat2 ] ) at @xmath129 ) .
for this exploratory study we did not have a complete set of light quark kappa values for every heavy - quark kappa value , as discussed in section [ sec : latticedetails ] .
this prevented us , in @xcite , from performing a reliable chiral extrapolation to obtain results for the semileptonic decay @xmath130 .
here we are concerned with semileptonic decays with a light vector meson in the final state : in contrast to the case of a pion , which is a pseudo - goldstone boson , we do not expect a sizable effect from the chiral extrapolation .
indeed , in figure [ fig : sixkappas ] we show the dependence of the form factor @xmath66 on the active and spectator light - quark masses , for two different combinations of momentum channel and heavy quark mass , @xmath105 and @xmath107 , for which six combinations of light quark kappas were computed .
it can be seen that in both cases @xmath66 remains practically constant as the light spectator quark mass decreases ( horizontal movement on the plots ) but there is a dependence on the active quark mass ( vertical movement ) . note that the value of @xmath57 , and hence @xmath46 , depends on @xmath131 as well as @xmath79 , and therefore the results of figure [ fig : sixkappas ] indicate that @xmath132 is nearly independent of @xmath131 , in which case , @xmath133 similar independence of the spectator mass was found in the study of the form factor @xmath43 in the radiative decay @xmath15 @xcite . as we have shown in the previous section , @xmath66 is equal to @xmath134 to good approximation for different heavy quark masses and different momentum channels and therefore the results of @xcite give us further evidence of the spectator mass independence of @xmath66 . to the situation is less clear for the form factors @xmath135 and @xmath29 owing to the larger statistical errors . @xmath135 and
@xmath29 appear to follow the same pattern as @xmath66 , but we can not dismiss a possible mild dependence on the light spectator mass . in figure
[ fig : sixkappas - v - a2 ] we show two examples of the dependence of the form factors @xmath29 and @xmath135 on the active and spectator light quark masses . to
we are going to assume the form factors are independent of the light spectator mass , and therefore we will extrapolate to the chiral limit only for the active light quark , as was suggested for @xmath66 in equation ( [ eq : spec - indepce ] ) .
this will allow us to use results for all four heavy quark masses to guide our extrapolations to the @xmath11 mass .
to quantify the systematic error induced we have compared , for the two heavy kappa values for which we have six combinations of light quark masses , the chiral - extrapolated form factor values for each momentum channel using two procedures : 1 .
using all six data points we can extrapolate to the chiral limit for both the spectator and active quarks ( see reference @xcite ) .
2 . fixing the spectator kappa value to @xmath136 and assuming complete independence of the form factors on the spectator quark mass , we perform the chiral extrapolation for the active quark only , as indicated in equation ( [ eq : spec - indepce ] ) for @xmath66 . for @xmath66
the two procedures give the same results within 5% except for those channels involving a momentum of @xmath89 in lattice units for the light vector meson , where the variation can be 10% .
these errors are comparable with the statistical ones . for procedure 1 above , the chiral extrapolation for the @xmath89 channels is less reliable than for channels with lower light meson momentum , because the statistical fluctuations for three - point functions with the lightest light - quark masses ( @xmath137 ) are larger .
these fluctuations can account for an important part of the 10% variation observed between the two procedures .
we have decided to use 5% to estimate , for all momentum channels , the systematic error on @xmath66 induced by this assumption of independence of the light spectator quark mass .
the statistical fluctuations are much larger for @xmath135 and @xmath29 , making it difficult to check the difference between the results from the two chiral extrapolation procedures above .
we will adopt a conservative position and will admit a 20% systematic error in @xmath135 and a 10% error in @xmath29 , based on the maximum discrepancies observed between the two procedures .
these discrepancies are within the statistical error bars . in figure
[ fig : a1a2v-1overm - extrapolation ] we show extrapolations to the @xmath11 mass scale for the form factors @xmath66 , @xmath135 and @xmath29 , allowing for linear and quadratic @xmath64 corrections to the infinite mass limit predictions of equation ( [ eq : btorho - infmass ] ) .
the extrapolated values agree well for the linear and quadratic fits : we will quote results only from the linear fits in the following , unless stated otherwise .
we have used six momentum channels : the same five channels used in the previous section together with the channel @xmath138 . for @xmath135 and @xmath29 the momentum channels @xmath139 and @xmath117 are extremely noisy and we can not control the extrapolations momentum channel is not measured directly for @xmath135 and @xmath29 . ] .
the measured values of the form factors , extrapolated to the @xmath11 scale for different values of @xmath46 close to @xmath56 , are listed in table [ tab : formfactorvalues ] .
# 1#2@xmath140 to in figure [ fig : a1mbqsq]a
we show @xmath141 at the @xmath11 scale , together with three fits to its @xmath46 dependence : @xmath142 table [ tab : a1mbqsq ] shows the fit parameters for each @xmath46 dependence , with the corresponding @xmath143 per degree of freedom . to to the figure and table indicate that constant - in-@xmath46 behaviour for @xmath66 does not fit the data well .
this feature is even more evident in lattice studies at the @xmath120 scale , where fits to constant behaviour are poor @xcite@xcite , @xcite .
our data favour a pole type behaviour with a pole mass in the expected range for a @xmath34 @xmath144 resonance @xcite .
dipole ( and in general higher powers : tripole , ) and pole fits are hardly distinguishable in the physical range @xmath145 .
in a dipole fit , the mass parameter @xmath146 is roughly given by @xmath147 and hence pole and dipole fits agree in the limited range of values of @xmath46 , well below @xmath148 , explored in figure [ fig : a1mbqsq]a .
an alternative procedure is to use pole fits for extracting @xmath149 from our measured data , at different heavy quark masses around that of the charm and then to extrapolate @xmath150 in @xmath64 to the @xmath11 scale , as shown in figure [ fig : a1mbqsq]b , by using @xcite @xmath151 this gives a result for @xmath150 at the @xmath11 scale ( using linear or quadratic extrapolations in @xmath64 ) of , @xmath152 the quoted errors are purely statistical
. a further 11% systematic error should be added , obtained by combining in quadrature the systematic errors from the spectator quark flavour symmetry breaking and discretisation effects . incorporating the systematic error
makes this method of extracting @xmath150 consistent within errors with the method presented in table [ tab : a1mbqsq ] . in conclusion
, we find that @xmath141 at the @xmath11 scale is fitted by a single pole form , with the parameters given in table [ tab : a1mbqsq ] .
the errors in table [ tab : a1mbqsq ] are statistical only : as mentioned above , a further 11% systematic error should be added . in table
[ tab : a1-zero - results ] we compare our results with other theoretical calculations : the agreement is generally good .
previous lattice results @xcite rely on the assumption of pole behaviour using a lattice determination of the @xmath153 @xmath34 resonance ( @xmath154 ) mass and the value of @xmath141 at a _ single _
@xmath46 value .
in contrast , in this paper , we have tried to determine the @xmath46 dependence of @xmath66 . to statistical and systematic uncertainties in the present study
have prevented us from extracting the @xmath46 dependence of the @xmath135 and @xmath29 form factors .
much effort has recently been devoted by the lattice community to determining the @xmath46 behaviour of @xmath42 and @xmath43 around the @xmath11 scale , without producing a definitive conclusion .
the @xmath46 dependence found here for @xmath66 and the results from the previous section for the ratio @xmath119 may support a pole - type behaviour for @xmath43 at least in a region around @xmath56 .
sum - rules calculations @xcite and a recent calculation by b stech @xcite find that @xmath29 has a more pronounced @xmath46 dependence than @xmath66 , which is consistent with having dipole type behaviour for @xmath29 and pole behaviour for @xmath66 . in the previous section we found that around the @xmath11 scale @xmath42 is roughly equal to @xmath29 .
this supports a dipole behaviour for @xmath42 and consequently further favours a pole behaviour for @xmath43 @xcite .
in this section we employ our lattice determination of the form factors to calculate the differential decay rate @xmath4 for the decay @xmath0 in the region near @xmath56 . experimental data for this region will enable a model - independent determination of @xmath5 .
it has previously been suggested that determinations of the decay rate for @xmath0 near @xmath40 , combined with the experimentally - measured branching fractions @xmath155 and @xmath156 would also allow a determination of @xmath5 .
however , one method relies on three different measurements , @xmath155 , @xmath156 and @xmath157 , together with the theoretical determination of the ratio @xmath158 @xcite .
the evaluation of this ratio of form factors from lattice calculations involves difficult and model - dependent extrapolations from the vicinity of @xmath56 to @xmath40 .
another method uses the ratio of @xmath159 and @xmath160 evaluated on a curve where @xmath161 @xcite , which is independent of any hadronic form factor .
however , it relies on the validity of the leading order hqs relations of equations ( [ eq : vt1kinematic ] ) and ( [ eq : at2kinematic ] ) between the form factors for the radiative and semileptonic decays in the region around @xmath40 , far from the zero recoil point .
furthermore , light flavour @xmath63 symmetry is used to relate the @xmath20 and the @xmath19 mesons .
corrections to these approximations will induce systematic errors in the determination of @xmath5 .
for example , the sum rule calculation of the relevant form factors in reference @xcite finds about 7% corrections to the leading hqs relations and about 25% corrections from @xmath63 breaking .
we propose to look at the region around @xmath56 for the process @xmath0 , beyond the region of charm production which complicates the experimental determination of @xmath10 transitions at low @xmath46 ( note that @xmath162 determined from inclusive decays is expected to be less than @xmath163 ) .
the methods referred to above , based on determinations of the decay rate near @xmath164 , suffer from this experimental difficulty in addition to possible theoretical uncertainties .
lattice techniques , in contrast , allow the form factors to be measured directly near @xmath56 , using only a @xmath64 extrapolation motivated by hqs , and avoiding the problematic model - dependent extrapolation to @xmath68 .
this model - dependence currently plagues the lattice determination of @xmath155 which requires knowledge of the form factors @xmath42 and/or @xmath43 at @xmath40 , where they are equal .
the differential decay rate for @xmath0 is given by @xmath165^{\frac{1}{2 } } \nonumber \\ & & \mbox { } \times \left(|h^{+}(q^2)|^2 + |h^{-}(q^2)|^2 + |h^{0}(q^2)|^2 \right ) \label{eq : diffdecayrate}\end{aligned}\ ] ] where @xmath166 .
@xmath167 comes from the contribution of the longitudinally polarised @xmath19 and is given by @xmath168 where @xmath95 is the momentum of the @xmath19 in the @xmath11-meson rest frame .
@xmath169 correspond to the contribution of the transverse polarisations of the vector meson and are given by @xmath170 looking for semileptonic decays beyond the charm production threshold ( determined by the semileptonic @xmath171 decay ) will make sense only if there are enough events . in figure
[ fig : btorhovsbtopi ] we show the differential decay rates @xmath4 for @xmath0 and @xmath130 .
for @xmath0 we assume that @xmath66 , @xmath135 and @xmath29 are all given by single pole forms with a common pole mass of @xmath172 and a normalisation determined by our measurements for the @xmath91 momentum channel .
for @xmath130 we use the results of reference @xcite with a dipole form for the form factor @xmath173 with @xmath174 and a mass parameter @xmath175 .
these curves are meant to be illustrative , based on reasonable hypotheses : for @xmath130 the results in @xcite were obtained for unphysically large @xmath176 and @xmath53 quark masses ; for @xmath0 we do not necessarily believe that the three form factors involved are all simultaneously single pole types with the same pole mass ( the pole mass of @xmath172 is an estimate of the average of the @xmath33 and @xmath34 @xmath144 resonance masses ) .
the points we wish to make are : 1 . for @xmath0 the events near @xmath40
have already been recommended for study , but the expected number of events at large @xmath46 is at least comparable .
2 . in the region beyond the charm threshold ,
the rates for @xmath0 and @xmath130 are at least comparable ( @xmath130 is not ten times more common than @xmath0 , for example ) , so both should be sizable contributions to the inclusive @xmath10 event rate and either or both can be used for the determination of @xmath5 . to the lattice calculations of the form factors in the previous section provide data points for @xmath4 over a sizable fraction of the region in @xmath46 above the charm endpoint ,
as can be seen in figure [ fig : btorhovsbtopi ] . the uncertainties in determining @xmath135 and @xmath29 in the @xmath113 and @xmath117 momentum channels
do not prejudice our measurement of the differential decay rate .
the rate is forced to vanish for kinematical reasons at @xmath56 so the @xmath113 channel form factor values of @xmath135 and @xmath29 are irrelevant .
moreover , the contribution to the differential decay rate of these two form factors in the @xmath117 channel is highly suppressed ( the @xmath117 channel has a @xmath46 near @xmath56 and the @xmath135 and @xmath29 contributions are suppressed relative to that of @xmath66 by a factor @xmath177 as shown in equation ( [ eq : diffdecayrate ] ) ) . over the range of @xmath46 for which we have measurements , the differential decay rate is dominated by the form factor @xmath66 .
this is shown by the difference between the points marked by squares and crosses in figure [ fig : btorhovsbtopi ] . in table
[ tab : diffdecaypoints ] we give the numerical values we have obtained for the differential decay rate at several @xmath46 values .
we have parametrised the results of the table by fitting our measured points for @xmath4 to the form @xmath178^{\frac{1}{2 } } a^2 \big ( 1 + b(q^2-\qsqmax ) \big),\ ] ] obtained by retaining the phase space dependence and expanding @xmath179 and @xmath167 from equation ( [ eq : diffdecayrate ] ) around @xmath56 .
the results of the fit are @xmath180 where the quoted errors are statistical .
a 24% systematic error should be added as discussed below . with this parametrisation
, we can partially integrate the differential decay rate , giving the results shown in the last column of table [ tab : diffdecaypoints ] .
the table also includes the partially integrated decay rate from the charm threshold to @xmath56 , although higher order terms in the taylor expansion in equation ( [ eq : param ] ) could become sizable at the lower end of this range , where we do not currently have measured points .
the results of table [ tab : diffdecaypoints ] and equation ( [ eq : abvalues ] ) have been obtained by extrapolating the form factors linearly in @xmath64 .
quadratic extrapolations in @xmath64 give results differing in the last significant figure , which is always well within the statistical errors . in figure [ fig : vub ] we show our measured data together with the fit , including 68% confidence level bounds .
to to comparison of any of the values in table [ tab : diffdecaypoints ] with experimental measurements will allow an extraction of @xmath5 with less than 10% statistical and 12% systematic theoretical uncertainties .
another way to determine @xmath5 uses the parametrisation of equation ( [ eq : param ] ) and our results of equation ( [ eq : abvalues ] ) .
it parallels the method used to extract @xmath13 from semileptonic @xmath181 decays ( see references @xcite@xcite ) .
the factor @xmath182 in equation ( [ eq : param ] ) parametrises long distance hadronic dynamics ( it is the analogue of the function @xmath183 in @xcite for the @xmath184 case ) .
the idea is to use experimental measurements of @xmath4 extrapolated to @xmath56 to extract a value for @xmath185 and compare with our theoretical determination of @xmath186 . here
@xmath78 plays the role of @xmath187 in the @xmath181 extraction of @xmath13 .
the lattice simulation provides a systematic determination of the overall normalisation of hadronic dynamical effects , parametrised by @xmath78 . for the extraction of @xmath13 from @xmath181
decays hqs provides such a normalisation ( the isgur - wise function is @xmath58 at zero recoil , up to @xmath128 and short distance perturbative corrections ) . for the heavy - to - light @xmath0 semileptonic decay , such a normalisation
is not provided by hqs . in the fit to equation ( [ eq : param ] )
the @xmath113 channel is not used because @xmath188 is zero and our fitting function automatically incorporates this feature .
our value for @xmath189 from the @xmath113 channel therefore gives an independent measurement of @xmath186 because @xmath29 and @xmath135 do not contribute at this kinematic point .
we find , using the value from table [ tab : formfactorvalues ] , and quoting statistical errors only , @xmath190 in excellent agreement with @xmath186 determined by the fit to the other channels .
this suggests that discretisation and other systematic errors have not conspired to change drastically the shape of the overall @xmath46 dependence .
our final result will use the value of @xmath186 , given in equation ( [ eq : abvalues ] ) , determined from the fit to these other channels , because this does not rely on a single measurement . to estimate systematic uncertainties , we have considered four possible sources of errors : quenching , determination of the lattice spacing in physical units , light spectator mass independence of the form factors and discretisation errors .
the exact effect of ignoring internal quark loops is an unknown systematic error .
however , quenched lattice calculations of form factors for @xmath191 semileptonic decays ( see @xcite and references therein ) give results in agreement with world average experimental values , while quenched calculations of @xmath192 @xcite decays have allowed extractions of @xmath13 in agreement with other determinations of that quantity .
this gives us some confidence that errors due to quenching are most likely within the statistical errors for processes involving a heavy quark .
the error arising from uncertainties in the value of the lattice spacing in physical units should be minimised in our calculation , because we have evaluated only dimensionless quantities : form factors and values of the velocity transfer , @xmath193 .
physical meson masses have been used as input where dimensional results are required . the only dependence on the lattice scale enters in expressing the heavy - light pseudoscalar masses in physical units for the @xmath64 extrapolations of the form factors and in the short distance logarithmic corrections of equation ( [ eq : theta ] ) .
the systematic error induced is much smaller than the other errors we will consider below .
we believe our results show good evidence for light spectator mass independence of the form factors .
we made this independence an assumption in order to use the results for the heaviest light spectator quark only : these results have smaller statistical error and are available for all our heavy quarks .
in the previous section we argued that we could not rule out a 5% error in @xmath66 ( 10% in channels involving @xmath89 momenta for the light meson ) due to this assumption .
the corresponding errors in @xmath135 and @xmath29 were 20% and 10% respectively . in the region
near @xmath56 , @xmath4 is dominated by @xmath66 and the contributions of @xmath29 and @xmath135 are kinematically suppressed , being 5% or less for @xmath194 .
the @xmath116 and @xmath195 channels explore lower @xmath46 values where the @xmath29 and @xmath135 contributions are less suppressed , of order 1520% ( see figure [ fig : btorhovsbtopi ] ) .
if the spectator quark mass independence assumption were violated by as much as 20% for @xmath29 and @xmath135 , the error in the differential decay rate would only be of the order of 2% for the non-@xmath89 channels , and 68% for the @xmath89 channels noted above , smaller in both cases than the likely 10% error induced by the same assumption for @xmath66 ( the differential decay rate is proportional to the square of the form factors ) ) of having a 10% error in the @xmath89 light meson momentum channels .
this changes the parameter @xmath78 by 5% . ] .
systematic errors due to this approximation are within our statistical errors .
a conservative estimate of the systematic error in the differential decay rate arising from this source can be obtained by adding in quadrature the 10% error due to @xmath66 and the worst - case 8% error from @xmath29 and @xmath135 , to get a final error of 13% .
we will see below that discretisation errors are expected to be almost twice as large as this .
we discussed in section [ sec : latticedetails ] that we could not dismiss having 10% errors in the form factors arising from discretisation errors at the value of @xmath85 used . in consequence
, we will admit a 20% error in the decay rate which translates into a 10% error in @xmath5 .
these errors are the largest among the systematic effects considered , and they almost entirely determine our overall systematic error . combining errors from discretisation and our assumption of light quark spectator mass independence of the form factors in quadrature , we finally obtain an estimated 24% systematic error in the decay rate , which should be added to the results of table [ tab : diffdecaypoints ] and equations ( [ eq : param])([eq : a2from00 ] ) .
we have presented a method for extracting the ckm matrix element @xmath5 from experimental measurements of the exclusive decay @xmath0 .
we have shown how lattice simulations provide a model - independent framework in which nonperturbative qcd corrections can be systematically incorporated in the evaluation of the differential decay rate @xmath196 in a region of @xmath46 values from the charm threshold up to the zero recoil point , @xmath56 .
we have measured points spanning roughly two - thirds of this region of @xmath46 .
one of the main features of the approach presented here is that it does not rely upon large , difficult and model - dependent extrapolations in @xmath46 from @xmath56 down to values close to @xmath197 , where the semileptonic decays are dominated by charmed final states , as in other previously suggested methods .
discounting experimental errors , the results of table [ tab : diffdecaypoints ] will allow the determination of the ckm matrix element @xmath5 with a theoretical uncertainty of less than 10% statistical and 12% systematic .
alternatively , determinations of @xmath5 ( with similar theoretical ambiguities ) can be obtained from the overall normalisation of the hadronic effects parametrised by @xmath78 , found to be @xmath198 both statistical and systematic theoretical errors will be reduced in forthcoming lattice simulations with smaller lattice spacing ( to reduce discretisation errors , currently the principal systematic effect ) and higher statistics . currently , only an upper bound for the total decay rate @xmath199 is known experimentally , but new results are forthcoming .
it is important that the differential decay rate near @xmath56 be measured , since , as we have shown , such a measurement can provide a clean determination of @xmath5 in the same way that similar measurements for the @xmath200 semileptonic decay have successfully been used to extract an accurate value for @xmath13 . in this paper
we have also determined the @xmath46 dependence of the form factor @xmath66 ( which dominates the decay rate close to the zero recoil point ) for the @xmath0 decay in the region from @xmath40 to @xmath56 .
the form factors @xmath29 and @xmath135 , for the same semileptonic decay , have been studied in a region of @xmath46 around @xmath56 .
finally , relations between radiative and semileptonic @xmath6 decays into light vector mesons , predicted by hqs , have been examined and found to hold within errors for the ratio @xmath119 and within 20% for the ratio @xmath118 at the @xmath11 scale .
using extra momentum channels where the heavy - light meson is not at rest , but @xmath57 is nevertheless nearly constant as the heavy quark mass changes @xcite , has been essential for studying @xmath46 , or equivalently @xmath57 , dependences of @xmath66 , @xmath119 , @xmath118 and @xmath4 .
knowledge of the @xmath46 dependence of @xmath4 has allowed us to determine the parameter @xmath78 from a set of five measurements as well as from the zero recoil channel , giving us further confidence in our procedure for extracting @xmath5 .
we thank chris sachrajda and hartmut wittig for useful discussions .
this research was supported by the uk science and engineering research council under grants gr / g 32779 and gr / h 49191 , by the particle physics and astronomy research council under grant gr / j 21347 , by the university of edinburgh and by meiko limited .
we are grateful to edinburgh university computing service and , in particular , to mike brown , for maintaining service on the meiko i860 computing surface .
dgr acknowledges the particle physics and astronomy research council for support through an advanced fellowship .
we acknowledge the particle physics and astronomy research council for travel support under grant gr / j 98202 .
99 see a j buras and m k harlander , in _ heavy flavours _ , advanced series on directions in high energy physics vol 10 , eds .
a j buras and m lindner , world scientific ( 1992 ) 58 , and references therein a ali and d london , in proc .
xxvii int .
conf . on high energy physics , glasgow , eds .
p j bussey and i g knowles , institute of physics publishing ( 1994 ) 1133 , and references therein particle data group , phys . rev .
* d50 * ( 1994 ) 1173 cleo collaboration , j bartelt _ et al .
_ , phys . rev . lett .
* 71 * ( 1993 ) 4111 g altarelli _ et al .
_ , nucl . phys .
* b208 * ( 1982 ) 365 n isgur _ et al .
_ , phys . rev .
* d39 * ( 1989 ) 799 ; n isgur and d scora , phys .
rev . * d40 * ( 1989 ) 1491 m wirbel , b stech and m bauer , z. phys .
* c29 * ( 1985 ) 637 ; m bauer and m wirbel , z. phys .
* c42 * ( 1989 ) 671 j g krner and g a schuler , z. phys .
* c38 * ( 1988 ) 511 c ramirez , j f donoghue and g burdman , phys . rev .
* d41 * ( 1990 ) 1496 r j morrison and j d richman , phys . rev . *
50 * ( 1994 ) 1602 cleo collaboration , a bean _ et al .
* 70 * ( 1993 ) 2681 l k gibbons , private communication m neubert , in proc .
xxvii int .
conf . on high energy physics , glasgow , eds .
p j bussey and i g knowles , institute of physics publishing ( 1994 ) 1129 m shifman , in proc .
xxvii int .
conf . on high energy physics , glasgow , eds .
p j bussey and i g knowles , institute of physics publishing ( 1994 ) 1125 ukqcd collaboration , s p booth _
et al . _ ,
* 72 * ( 1994 ) 462 ; ukqcd collaboration , k c bowler _ et al .
_ , southampton preprint shep 9506 , edinburgh preprint 95525 , marseille preprint cpt95p3179 , hep - ph/9504231 l p lellouch in proc .
lattice 94 : 12th international symposium on lattice field theory , bielefeld , germany , eds .
f karsch _ et al .
b ( proc . suppl . )
* 42 * ( 1995 ) 421 ; marseille preprint cpt95/p.3196 , to appear in proc .
xxx rencontres de moriond , les arcs , france , march 1995 cleo collaboration , t e browder , in proc .
xxvii int .
conf . on high energy physics , glasgow , eds .
p j bussey and i g knowles , institute of physics publishing ( 1994 ) 1117 aleph collaboration , i j scott , in proc .
xxvii int .
conf . on high energy physics , glasgow , eds .
p j bussey and i g knowles , institute of physics publishing ( 1994 ) 1121 argus collaboration , h albrecht _ et al .
_ , z. phys . *
c57 * ( 1993 ) 533 n isgur and m b wise , phys . rev . *
d42 * ( 1990 ) 2388 g burdman and j f donoghue , phys . lett .
* b270 * ( 1991 ) 55 p j
odonnell and q p xu , phys . lett . * b325 * ( 1994 ) 219 ; p j odonnell and h k k tung , phys . rev .
* d48 * ( 1993 ) 2145 p santorelli , z. phys . *
c61 * ( 1994 ) 449 p colangelo , f de fazio and p santorelli , bari preprint bari th/94174 , hep - ph/9409438 b grinstein , r springer and m b wise , nucl . phys .
* b339 * ( 1990 ) 269 m neubert , phys . rept .
* 245 * ( 1994 ) 259 p a griffin , m masip and m mcguigan , phys
d50 * ( 1994 ) 5751 b sheikholeslami and r wohlert , nucl . phys . *
b259 * , 572 ( 1985 ) .
g heatlie _ et al .
_ , nucl . phys .
* b352 * , 266 ( 1991 ) .
ukqcd collaboration , c allton _
et al . _ ,
phys . rev .
* d49 * , 474 ( 1994 ) ukqcd collaboration , a k ewing _
et al . _ , southampton preprint shep9520 ukqcd collaboration , k c bowler _
_ , phys . rev .
* d51 * ( 1995 ) 4955 ukqcd collaboration , k c bowler _
_ , phys . rev .
* d51 * ( 1995 ) 4905 v lubicz , g martinelli , m s mccarthy and c t sachrajda , phys .
* b274 * , 415 ( 1992 ) .
c bernard , a el - khadra and a soni , phys .
, 2140 ( 1991 ) ; phys .
* d45 * , 869 ( 1992 ) .
ukqcd collaboration , d r burford _ et al .
_ , southampton preprint shep9509 , hep - lat/9503002 , nucl .
b in press a borrelli _ et al .
_ , nucl .
* b373 * ( 1992 ) 781 g martinelli , c t sachrajda and a vladikas , nucl . phys .
* b358 * ( 1991 ) 212 ukqcd collaboration , r m baxter _ et al .
_ , phys . rev .
* d49 * ( 1994 ) 1594 m e luke , phys .
* b252 * ( 1990 ) 447 a ali , v m braun and h simma , z. phys .
* c63 * ( 1994 ) 437 a abada _
et al . _ ,
phys . * b416 * ( 1994 ) 675 ape collaboration , c r allton _ et al .
_ , phys . lett .
* b345 * ( 1995 ) 513 ape collaboration , a abada _
et al . _ , rome preprint rome 94/1056 , hep - lat/9503020 p ball , v m braun and h g dosch , phys . rev . (
1991 ) 3567 ; p ball , phys . rev .
* d48 * ( 1993 ) 3190 s narison , phys . lett .
* b345 * ( 1995 ) 166 ; phys .
b283 * ( 1992 ) 384 r n faustov , v o galkin and a yu mishurov , cybernetics council , moscow , preprint , hep - ph/9505321 r casalbuoni _ et al .
_ , phys . lett . *
b299 * ( 1993 ) 139 b stech , heidelberg preprint hd thep954 , hep - ph/9502378 | we present results of a lattice computation of the vector and axial - vector current matrix elements relevant for the semileptonic decay @xmath0 .
the computations are performed in the quenched approximation of lattice qcd on a @xmath1 lattice at @xmath2 , using an @xmath3 improved fermionic action .
our principal result is for the differential decay rate , @xmath4 , for the decay @xmath0 in a region beyond the charm threshold , allowing a model - independent extraction of @xmath5 from experimental measurements .
heavy quark symmetry relations between radiative and semileptonic decays of @xmath6 mesons into light vector mesons are also discussed .
# 1 # 2 # 3to0pt-#2 2rho|b^0^+ l^- |_l # 1#1 | # 1| # 1 # 1 # 1 # 1 # 1 makecaption#1#2 tempboxa tempboxa > tempboxa to southampton preprint shep 9518 + edinburgh preprint 95/551 + marseille preprint cpt95/pe.3218 + hep - ph/9506398 lattice study of the decay @xmath7 : + model - independent determination of @xmath8 + _ ukqcd collaboration _ + * j m flynn , j nieves * + department of physics , university of southampton , southampton so17 1bj , uk + * k c bowler , n m hazel , d s henty , h hoeber , r d kenway and d g richards * + department of physics & astronomy , the university of edinburgh , edinburgh eh9 3jz , scotland + * b j gough * + theoretical physics ms106 , fermilab , batavia il 60510 , usa + * h p shanahan *
+ department of physics & astronomy , university of glasgow , glasgow g12 8qq , scotland + * l p lellouch * + centre de physique theorique , cnrs luminy , f13288 marseille cedex 9 , france |
the old concept of multiband / multigap superconductivity @xcite has found its strong experimental evidence only recently in the rich physics of magnesium diboride @xcite .
consequently more superconductors are ( re)considered along this line .
one of the important examples are the iron pnictides , a new class of high-@xmath8 superconductors @xcite for which multigap superconductivity is suggested to lead to an exotic pairing mechanism with a sign reversal of the order parameter between separated fermi - surface sheets .
a revision in dichalcogenides brings more and more signatures of a distribution of superconducting energy gaps , which can be either due to different gaps on different fermi - surface sheets or anisotropic single gap .
transition - metal dichalcogenides 2h - mx@xmath9 ( m = nb , ta , x = s , se ) are materials with layered structure .
nb or ta atoms are trigonally prismatic coordinated by chalcogen atoms and metallic layers are held together by weak van der waals forces .
because of this layered structure , electrical , magnetic , and optical properties show a high degree of anisotropy .
nbs@xmath9 is the only member of the 2h - mx@xmath9 family , which does not undergo a charge - density wave transition @xcite .
this could be a reason for its strong anisotropy , much larger than that of nbse@xmath0 .
nbse@xmath9 had been considered for a long time as being a conventional type ii superconductor @xcite .
later on , effects of the anisotropic and strong coupling interactions were taken into account @xcite .
recent measurements sensitive to the order parameter show evidence that more than one energy scale is necessary to account for establishing superconductivity @xcite .
nbs@xmath9 was also originally considered as just another anisotropic superconductor and its unusual specific heat dependence was not interpreted in detail @xcite .
an important breakthrough came with scanning tunneling microscopy / spectroscopy ( stm ) measurements by guillamn et al .
@xcite , showing strong indications for two superconducting energy gaps instead of a single anisotropic one . since stm is a surface probe , this strong statement certainly needs independent support showing that the two gaps are reflecting the bulk properties of the system . in this paper
we address this issue with bulk thermodynamic measurements of the specific heat at temperatures down to 0.6 k and in magnetic fields up to 14 tesla via ac - calorimetry technique .
we find that the electronic specific heat @xmath2 can not be described by the standard bcs model with a single isotropic energy gap .
first , at the lowest temperatures the data shows that quasiparticles are excited over an energy gap much smaller than the bcs weak coupling limit . the overall temperature dependence of @xmath2 can be described only if two gaps or an anisotropic one gap case is considered .
second , the field dependence of the sommerfeld coefficient @xmath3 shows a strong curvature in striking similarity with that of nbse@xmath0 and mgb@xmath0 .
however , the anisotropy of @xmath3 decreases with magnetic field in an opposite manner compared to the latter system .
finally , the two gap scenario is supported by the absence of in - plane gap anisotropy in recent stm imaging of the vortex lattice in nbs@xmath0 @xcite , and by the fact that the numerical values of the two gaps obtained from fitting our data , @xmath10 and @xmath11 , are also in a very good agreement with the stm data .
details of the synthesis of the single crystalline samples can be found elsewhere @xcite .
the crystals used for the specific heat measurements come from the same batch as those used in the previous stm studies @xcite .
their chemical composition was checked using an energy dispersive x - ray spectroscopy ( eds ) .
single crystals were also confirmed to be of 2h polytype by x - ray diffraction measurements . in our experiment a thin crystal with a well defined hexagonal shape and dimensions 500x500x30 @xmath12 m was chosen .
specific heat measurements have been performed using an ac technique as described elsewhere @xcite .
the high sensitivity of this technique is not only very well adapted to measure the specific heat of very small samples but also to carry continuous measurements during temperature or magnetic field sweeps .
we were thus able to obtain the field dependence of the electronic part of @xmath13 at @xmath14 k which only differs from its zero temperature limit , the sommerfeld coefficient @xmath3 , by about 2@xmath15 .
measurements were performed with the magnetic field aligned along the two main crystallographic orientations , i.e. parallel and perpendicular to the basal @xmath16 plane of the sample .
the temperature oscillations of the sample were recorded by a thermocouple calibrated in magnetic field using measurements on ultrapure silicon .
we performed measurements at temperatures down to 0.6 k and in magnetic fields up to 8 t in the @xmath17he and @xmath18he refrigerators in koice .
supplementary measurements up to 14 tesla and down to 2 k were performed in grenoble .
figure 1 displays the temperature dependence of the specific heat of the sample ( plus addenda ) in selected magnetic fields up to 8 t for @xmath19 and up to 3 t for @xmath4 .
the thermodynamic superconducting transition temperature at zero field was determined from the local entropy balance around the phase transition giving @xmath20 6.05 k. the zero - field anomaly at the transition is sharp ( @xmath21@xmath22@xmath23 @xmath24 0.4 k ) indicating the high quality and homogeneity of the sample .
the position of the specific - heat jumps are gradually shifted toward lower temperatures for increasing magnetic field . despite a significant broadening at high fields ,
the anomaly remains well resolved at all fields .
a field of 3 tesla applied along the c - axis was sufficient to completely suppress superconductivity in all the temperature range .
on the other hand , 8 tesla applied along the ab - planes shifts the superconducting anomaly down to only about 3 - 4k underlying the strong anisotropy of this system .
later we extended the measurements down to 0.6 k in a @xmath17he fridge where the specific heat was measured at zero field and at @xmath25 t. in the case of a very small crystal like ours , it is difficult to evaluate the exact total contribution of the addenda . the electronic part of the total specific heat value can be obtained by extrapolation of @xmath26 for t approaching zero .
this value corresponds to @xmath27 of @xmath28 , with @xmath29 being the electronic heat capacity of the sample in the normal state . to avoid any fitting procedures , the addenda and the phononic contributions
have been eliminated by subtracting the data taken at @xmath25 t from all the other runs .
thus , the electronic specific heat of the sample , @xmath30 , normalized to its normal state value , @xmath29 , can be obtained experimentally by : @xmath31 , where @xmath32t@xmath33 and @xmath28 = @xmath34t@xmath35 .
the result is presented in fig . 2 by the open circles .
the only assumption in this procedure is the absence of magnetic field dependence of the addenda .
this has been previously verified in numerous experiments using the same thermocouple wires and also confirmed here independently by the entropy conservation required for a second order phase transition , proving the thermodynamic consistency of the data and its treatment .
we first compare the electronic specific heat with the isotropic single gap ( isg ) bcs model . the dashed line in fig .
2 presents the isg bcs specific heat ( weak coupling of 2@xmath36 ) .
one can see that while the height of the jump at @xmath8 of the experimental data is quite well reproduced , a significant deviation occurs at lower temperatures .
the discrepancy between the bcs curve and the measured data represents 18 @xmath15 ( 7 @xmath37 ) of the total signal at 1.4k ( 4k ) , which is significantly larger than the error bars of our measurements .
the inset of fig .
2 displays the logarithm of the electronic specific heat versus @xmath38 .
as shown , one obtains an exponential dependence @xmath39 for @xmath40 .
however , the parameter @xmath41 is significantly lower than the value expected for the bcs weak coupling limit in the temperature range @xmath42 @xcite .
this corresponds to coupling ratio 2@xmath21/@xmath43 2.3 that is much smaller than the isg bcs value of 3.52 , indicating that the quasiparticles are activated over a small energy gap .
this fact as well as the overall shape of the specific heat temperature dependence resembles the case of mgb@xmath0 .
a phenomenological @xmath44-model of the specific heat accounting for independent contributions from two bands with two different energy gaps has been successfully applied in the case of mgb@xmath9 @xcite .
the magnitude of the small gap 2@xmath45 and of the large gap 2@xmath46 at @xmath47 are fitting parameters of the model .
the third parameter is the relative fraction of the density of states of the two bands @xmath48 . the full line in fig .
2 represents a fit to the experimental data yielding the following parameters : 2@xmath49 2.1 @xmath500.05 , 2@xmath51 [email protected] , and @xmath48 = 0.67@xmath50 0.15 .
the value of the small gap is close to the one evaluated from the exponential decay shown in the inset of fig .
2 . importantly , both gap values are in striking agreement with those found in the stm experiment @xcite confirming that the latter are characteristic of the bulk . as previously shown by huang et al .
@xcite , the temperature dependence of the specific heat of nbse@xmath0 , another two - gap superconductor , can also be described by an anisotropic @xmath7-wave model , where the gap anisotropy is supposed to be in the form of @xmath52 corresponding to the hexagonal in - plane symmetry .
here , @xmath53 is the average gap value and @xmath44 denotes its anisotropy , yielding @xmath54 and @xmath55 .
this model with parameters @xmath56 and 2@xmath57 fits our data as well as the two gap scenario , @xmath58 @xmath59 @xmath60 @xmath58 @xmath61 @xmath62 @xmath63 @xmath64 .
we remark that the anisotropic gap should leave its footprint in the anisotropic vortex core @xmath65 as it is proportional to the related fermi velocity @xmath66 divided by the gap at zero temperature @xmath67 .
however , in contrast to nbse@xmath0 , for which sts images revealed a sixfold star shape of the vortex cores , the fully isotropic vortices have been imaged in nbs@xmath0 @xcite questioning the applicability of the anisotropic single gap model in our case .
we have measured thoroughly the evolution of the specific heat in the mixed state . at @xmath68
k the electronic specific heat term @xmath69 is very close to the sommerfeld coefficient @xmath3 . its field dependence is displayed in fig . 3a and 3b for both principal field orientations .
our maximum field available at this temperature range ( 8 tesla ) was not sufficient to reach the normal state for @xmath19 , but it was well above the upper critical field value of [email protected] tesla for @xmath4 .
3a emphasizes the strong non - linearity of @xmath70 when @xmath71 is applied perpendicular to the @xmath16 plane .
again , such a non - linearity could be associated with the existence of 2 gaps or a single anisotropic one .
the increase of @xmath3 with magnetic field is mainly due to the quasiparticle contribution inside the vortex cores . in the case of superconductor with a single isotropic gap ,
@xmath3 should increase linearly in small magnetic field and a small nonlinearity in @xmath70 appears above the field where flux lines start overlapping . according to the calculations of nakai et al .
@xcite , much stronger non - linearity of @xmath70 is achieved in case of anisotropic - gap superconductors . the full line in the fig .
3a displays the field dependence of the normalized density of states @xmath72 ( proportional to the sommerfeld coefficient ) calculated by nakai et al . for the anisotropic gap with @xmath73 .
the model qualitatively reproduces the behavior in our data ( open circles ) , but fails to describe the fast increase of @xmath3(h ) at low fields .
this is evident in the inset of fig .
3a where the derivative @xmath74 , i.e. the slope of @xmath75 for @xmath73 ( line ) and the slope of measured sommerfeld coefficient @xmath76 ( open circles ) are compared .
this discrepancy makes the explanation of the specific heat data by the anisotropic gap scenario rather inconsistent .
the observed @xmath70 behavior resembles the two - gap case of mgb@xmath0 ( see fig .
3c , left curve ) @xcite , where the initial rapid increase of @xmath70 due to dominant role of the @xmath77 band with the small gap changes at fields where the @xmath78 band with the large gap comes into play .
next , we compare the behavior of the sommerfeld coefficients in magnetic fields applied parallel and perpendicular to the basal planes of the both materials , nbs@xmath0 ( fig .
3b ) and mgb@xmath0 ( fig .
the latter is taken from ref.@xcite .
following the procedure introduced by bouquet et al . in ref.@xcite
an effective anisotropy @xmath79 can be obtained from these @xmath70 dependences .
it is defined as the ratio of the magnetic fields in the @xmath16 plane and along the @xmath6 axis yielding the same @xmath3 value in fig .
3b and fig .
@xmath79 is plotted in fig 4a for both compounds as a function of the field @xmath4 normalized to its upper critical field value @xmath80 . as discussed already in the work of bouquet et al .
on mgb@xmath0 @xcite the choice of the abscissa is arbitrary .
we chose @xmath4 , but we could plot @xmath79 versus @xmath19 or versus @xmath70 as well . note that this @xmath79 tends towards the usual anisotropy of h@xmath81 , @xmath82 when @xmath83 at large magnetic fields . in mgb@xmath0 at low fields ( @xmath84 ) , the @xmath70 curves for the two principal directions are practically identical which gives @xmath85 as shown in fig .
4a . at larger fields , @xmath79 increases reflecting a reduced contribution from the isotropic @xmath77-band , reaching @xmath86 which is the anisotropy of the dominant @xmath78-band @xcite . in nbs@xmath0
, one observes an opposite field dependence of @xmath79 which starts from a highly anisotropic value @xmath87 at low fields and decreases to @xmath88 at our maximum field .
a field dependent superconducting anisotropy is a typical signature of multigap superconductivity where a role of bands with different gaps can significantly vary with magnetic field @xcite .
in contrast to mgb@xmath0 case , in nbs@xmath0 both bands would be anisotropic , as suggested by analogy with nbse@xmath0 in which band structure calculations @xcite show mostly 4 fermi surface sheets derived from nb @xmath89-bands forming warped cylinders along the @xmath6 axis , centered on the @xmath90 and @xmath91 points in the brillouin zone .
moreover , two sheets derived from the bonding nb @xmath89 band are significantly more warped than the two derived from the antibonding nb @xmath89 band .
different warping of nb sheets can naturally lead to a different level of anisotropy in each band .
thus , a qualitatively different behavior of @xmath92 compared to mgb@xmath0 can be expected . finally , we inspected the upper critical magnetic fields for both principal orientations of magnetic field .
5 summarizes the values of @xmath71@xmath81 derived from the temperature - sweep measurements of the specific heat shown in fig .
1 , as well as from field - sweep measurements . two sets of field - sweeps were performed , one in 14 tesla magnet for @xmath19 in a temperature range down to 2 k , and another one taken in the @xmath93 cryostat down to 0.6 k in the 8 tesla coil .
as stated above , we determined @xmath8 at zero field from the local entropy balance around the anomaly .
however , at finite fields this definition is not very practical for establishing @xmath94 , or @xmath95 . in order to reduce the uncertainty of the @xmath96 value arising from the broadening of the transition particularly at lower temperatures ( higher fields ) , we inspected the temperature shift between two neighboring curves in fig .
1 . a similar procedure was used to determine @xmath96 from field - sweep measurements .
the resulting temperature dependence of @xmath96 is shown in fig . 5 for both @xmath97 and @xmath4 .
importantly , the three independent sets of temperature and field - sweeps measurements yield consistent results with a nice overlap . both temperature dependencies
show a slight positive curvature for temperatures @xmath98 .
the upper critical field in the @xmath16 plane reveals very high values with @xmath99 slope of about 3 tesla / k , close to the pauli paramagnetic limit .
the temperature dependence of the superconducting anisotropy @xmath90 calculated as a ratio @xmath71@xmath100 of the upper critical fields is displayed in fig . 4b together with the results obtained in mgb@xmath0 @xcite . as shown , in contrast to mgb@xmath0 ,
the resulting anisotropy @xmath90 is close to 7 and approximatively constant for @xmath101 .
note that this value might be slightly underestimated in case of a small misalignement of the crystal for @xmath102 .
our results are consistent with those obtained by onabe et al .
@xcite from resistive measurements in a field up to 2 tesla .
the strong decrease of @xmath90 in mgb@xmath0 close to @xmath8 is a direct consequence of the existence of the isotropic @xmath77- band .
this is not a general feature of multiband superconductivity since @xmath103 results from a subtle balance between the fermi velocities and the relative weight in the dos of the different bands @xcite .
these precise calculations are still to be carried out in the case of nbs@xmath0 .
analysis of the zero - field specific heat data has shown that : ( 1 ) the zero - field electronic term of the specific heat can not be described by an isotropic single - gap bcs formula , but it is compatible with the two - gap @xmath44 model ; ( 2 ) the large ( small ) gap is 2@xmath104 4.6 ( 2@xmath105 2.1 ) .
the measurements in the mixed state have supported the two - gap scenario revealing : ( 3 ) a strongly non - linear @xmath106 ; ( 4 ) a field - dependent superconducting anisotropy . even if some of these features of the specific heat could be eventually explained by an extremely anisotropic - gap superconducting model
, this would not be compatible with the observation of two well - resolved gap features with sizes of 2@xmath107 2 and 4 , respectively , and also with the absence of in - plane anisotropy in the vortex lattice images by the scanning - tunneling spectroscopy of guillamn et al . to conclude ,
our bulk thermodynamic measurements are in full agreement with stm spectra , supporting that nbs@xmath0 is another case of well resolved two - gap superconductor .
this work was supported by the ec framework programme mtkd - ct-2005 - 030002 , by the eu erdf ( european regional development fund ) grant no .
itms26220120005 , by the slovak research and development agency , under grants no .
vvce-0058 - 07 , no .
apvv-0346 - 07 , no .
sk - fr-0024 - 09 and no .
lpp-0101 - 06 , and by the u.s .
steel koice , s.r.o .
centre of low temperature physics is operated as the centre of excellence of the slovak academy of sciences .
we thank g. karapetrov for careful reading of the manuscript .
h. suhl , b. t. matthias , and l. r. walker , phys .
* 3 * , 552 ( 1959 ) .
x. x. xi , rep . prog .
phys . * 71 * , 116501 ( 2008 ) . , eds .
p. c. w. chu et al .
, physica c * 469 * , 313 - 674 ( 2009 ) .
d. e. moncton , j. d. axe , and f. j. disalvo , phys .
lett . * 34 * , 734 ( 1975 ) .
b. t. matthias , t. h. geballe , and v. b. compton , rev .
phys . * 35 * , 1 ( 1963 ) .
n. kobayashi , k. noto , and y. muto , j. low .
temp . phys . * 27 * , 217 ( 1977 ) .
d. sanchez , a. junod , j. muller , h. berger and f. lvy , physica b * 204 * , 167 ( 1995 ) .
t. yokoya , t. kiss , a. chainani , s. shin , m. nohara , and h. takagi , science * 294 * , 2518 ( 2001 ) .
j. d. fletcher , a. carrington , p. diener , p. rodire , j. p. brison , r. prozorov , t. olheiser , and r. w. giannetta , phys .
lett . * 98 * , 057003 ( 2007 ) . c. l. huang , j .- y .
lin , y. t. chang , c. p. sun , h. y. shen , c. c. chou , h. berger , t. k. lee , and h. d. yang , phys .
b * 76 * , 212504 ( 2007 ) . t. kiss , t. yokoya , a. chainani , s. shin , t. hanaguri , m. nohara and h. takagi , nature physics * 3 * , 720 ( 2007 ) .
t. valla , a.v .
fedorov , p. d. johnson , p - a .
glans , c. mcguinness , k. e. smith , e.y .
andrei , and h. berger , phys .
* 92 * , 086401 ( 1994 ) .
y. hamaue and r. aoki , j. phys .
. jap . * 55 * , 1327 ( 1986 ) .
dahn , j.f .
carolan , and r.r .
haering , phys .
b * 33 * , 5214 ( 1986 )
. i. guillamn , h. suderow , s. vieira , l. cario , p. diener , and p. rodire , phys .
* 101 * , 166407 ( 2008 ) .
w. fisher and m. sienko , inorg .
* 19 * , 39 ( 1980 ) .
p. f. sullivan and g. seidel , phys
. rev . * 173 * , 679 ( 1968 ) .
superconductivity , edited by r. d. parks , marcel dekker inc .
, new york 1969 .
f. bouquet , y. wang , r. a. fisher , d. g. hinks , j. d. jorgensen , a. junod , and n. e. phillips , europhys .
* 56 * , 856 ( 2001 ) .
n. nakai , p. miranovi , m. ichioka , and k. machida , phys .
b * 70 * , 100503(r ) ( 2004 ) .
z. pribulov , t. klein , j. marcus , c. marcenat , f. levy , m. s. park , h. g. lee , b. w. kang , s. i. lee , s. tajima , and s. lee , phys .
lett . * 98 * , 137001 ( 2007 ) .
f. bouquet , y. wang , i. sheikin , t. plackowski , a. junod , s. lee , s. tajima , phys .
lett . * 89 * , 257001 ( 2002 ) . v. g. kogan and s. i. budko , physica c * 385 * , 131 ( 2003 ) .
t. dahm and n. schopohl , phys .
* 91 * , 017001 ( 2003 ) .
a. a. golubov , a. brinkman , o. v. dolgov , j. kortus , and o. jepsen , phys . rev .
b * 66 * , 054524 ( 2002 ) . m. d. johannes , i. i. mazin , and c. a. howells , phys .
b * 73 * , 205102(2006 ) .
l. lyard , p. samuely , p. szabo , t. klein , c. marcenat , l. paulius , k. h. p. kim , c. u. jung , h .- s .
kang , s. choi , s .-
lee , j. marcus , s. blanchard , a. g. m. jansen , u. welp , g. karapetrov , and w. k. kwok , phys .
b * 66 * , 180502(r ) ( 2002 ) .
k. onabe , m. naito , and s. tanaka , journal of the phys .
soc . of japan
* 45 * , 50 ( 1978 ) . | the heat capacity of a 2h - nbs@xmath0 single crystal has been measured by a highly sensitive ac technique down to 0.6 k and in magnetic fields up to 14 t. at very low temperatures data show excitations over an energy gap ( @xmath1 ) much smaller than the bcs value .
the overall temperature dependence of the electronic specific heat @xmath2 can be explained either by the existence of a strongly anisotropic single - energy gap or within a two - gap scenario with the large gap about twice bigger than the small one .
the field dependence of the sommerfeld coefficient @xmath3 shows a strong curvature for both principal - field orientations , parallel ( @xmath4 ) and perpendicular ( @xmath5 ) to the @xmath6 axis of the crystal , resulting in a magnetic field dependence of the superconducting anisotropy .
these features are discussed in comparison to the case of mgb@xmath0 and to the data obtained by scanning - tunneling spectroscopy .
we conclude that the two - gap scenario better describes the gap structure of nbs@xmath0 than the anisotropic @xmath7-wave model .
pacs : 74.25.bt 74.25.ha 74.70.ad |
electromagnetic fields can be expressed in terms of the scalar and the vector potentials @xmath4 and @xmath5 . while the physical fields are unique , there is a certain degree of freedom in the choice of the potentials .
specifically , electric and magnetic fields in gaussian units are @xmath6 and @xmath7 . a gauge transformation @xmath8 leaves the physical quantities @xmath9 and @xmath10 unchanged , if @xmath11 satisfy the lorentz condition @xcite .
the gauging function @xmath11 allows us to transform between different representations of @xmath5 and @xmath4 in different gauges , although the physical quantities , _ e.g. _ observables , are gauge invariant .
the hamiltonian for an optical electron in external electromagnetic field may be written as @xmath12 where @xmath13 is the core potential seen by the electron . in our current discussion
, we treat the electromagnetic field as a plane wave propagating along the @xmath14-direction , and describe the electric field as @xmath15 . in the velocity gauge ( also known as the transverse or coulomb gauge ) , @xmath16 the length and the velocity gauges are related through the transformation function @xmath17 , leading to potentials in the length gauge @xcite : @xmath18
in the velocity gauge , the hamiltonian becomes , @xmath19 we calculate the energy shifts within the floquet formalism of quasi - energy states @xcite for the third term , and use the first order perturbation theory for the fourth term since it is already second order in the field strength .
the energy shift for a rydberg state @xmath20 therefore becomes @xmath21 the first term can be decomposed into the scalar , vector and tensor contributions .
we focus on the dominant scalar contribution to the stark shift and arrive at @xmath22 where @xmath23 , and we used @xmath24 here @xmath25 is the usual dipole operator .
the summation over the magnetic quantum number can be evaluated explicitly using the wigner - eckart theorem : @xmath26 where @xmath27 are the reduced dipole matrix elements . the first term in eq .
can be immediately identified as the ponderomotive energy shift for a free electron quivering in electromagnetic field .
notice that it came directly from the @xmath28 term in the hamiltonian , and it is the dominant term for a rydberg state .
by contrast , the second term in is the correction to the free electron ponderomotive shift .
all of the resonance structure is included in the second correction term , whereas the free electron shift only provides a smooth background on which the resonance structure sits .
we can express @xmath29 in terms of the conventional ac polarizability through its definition for the plane waves , @xmath30 .
then the ac polarizability in the velocity gauge becomes @xmath31 here @xmath32 is the free electron polarizability .
we now derive eq . in the length gauge , and thereby establish the gauge invariance . using the length gauge vector and scalar potentials , the hamiltonian can be expressed as @xmath33 to calculate the energy shifts resulting from the last three terms in the hamiltonian , we again use the quasi - energy formalism @xcite for the third and the fifth terms , and the first order perturbation theory for the fourth term : @xmath34 upon expanding the square , we encounter three contributions to the first term . out of the three , the dominant one is the @xmath35 term .
the term involving @xmath36 , and the last term involving @xmath37 are suppressed by a factor of @xmath38 and can be safely ignored in the non - relativistic approximation .
the cross term between the @xmath39 and @xmath40 drops out because it is proportional to @xmath41 ^ 2 $ ] which vanishes on the account of @xmath42 being purely imaginary ( due to @xmath43 ) .
we again focus on the dominant scalar contribution to the stark shift , which leaves us with @xmath44 through the definition @xmath30 , the polarizability in the length gauge can be written as : @xmath45 to reveal how this expression relates to @xmath2 and the polarizability in the velocity gauge , we expand the resolvent operator in eq . in powers of @xmath46 :
@xmath47 this results in a series expansion of the polarizability in the length gauge : @xmath48 this series can be resummed such that the first term is separated out , @xmath49 with the help of the oscillator sum rule , @xmath50 we recover the free electron polarizability from the first term in when summed over a complete set of states ( note that @xmath51 ) . on the other hand , with the aid of the expansion , the outer sum in the second term can be collapsed back into the same correction term to the free electron polarizability in the velocity gauge . thus @xmath52 and @xmath53 .
this expression for the length gauge polarizability is identical to the one in eq . in the velocity gauge , confirming the equivalence of the ac stark shifts in both gauges , _ i.e. _ the gauge invariance .
these expressions also show how the free electron term originates , and we will demonstrate below that the correction to @xmath54 is indeed very small for rydberg states . in ref .
@xcite , the emergence of the free electron polarizability in the length gauge has been shown numerically for rydberg states of rb atom .
@xcite has provided an alternative analytical expression using higher rank oscillator sum rules .
however , we find that the formula of ref .
@xcite has poor convergence properties when evaluated numerically .
so far we examined the ac stark effect in the non - relativistic formalism . in the velocity gauge
the dominant contribution came from the expectation value of the vector - potential - squared term @xmath28 .
curiously , the fully - relativistic dirac hamiltonian involves only linear couplings to electromagnetic fields , thereby such term is missing in the lowest - order of perturbation theory .
this naturally raises a question of just how the free - electron polarizability emerges from the relativistic equations .
as we show in this section , it comes from summations over the negative energy ( positron , @xmath55 ) states , i.e. from the dirac sea .
similarly pronounced effects of dirac sea on weak atomic transition amplitudes were found earlier @xcite in relativistic many - body calculations .
we start with the coupling to electromagnetic fields @xmath56 where @xmath57 are the conventional dirac matrices . in the velocity gauge @xmath58 , so that @xmath59 applying the second - order floquet formalism , we arrive at the fully - relativistic ac polarizability @xmath60 the summation over intermediate states in the above equation spans the complete spectrum of the dirac equation , i.e. , both positive and negative energy states .
it can be shown using the pauli approximation that the sum over the conventional positive energy states recovers the second term ( second - order sum ) in the non - relativistic expression for the ac polarizability .
now we demonstrate that the dominant , free - electron polarizability term , emerges from summing over the negative energy states . in this case , the rydberg electron energy @xmath61 is above the dirac sea level and @xmath62 and @xmath63 ( we naturally assume that the photon energy is well below @xmath64 ) . within this approximation ,
the dynamic ac polarizability becomes @xmath65 the dirac matrices @xmath66 mix the large and the small components of the dirac bi - spinor , @xmath67 .
part of the summation on the right hand side of can therefore be written as @xmath68 in the non - relativistic approximation , the negative energy states are complete amongst themselves , i.e. @xmath69 , leading to @xmath70 where we used @xmath71 and @xmath72 a.u .. therefore we recover the free electron polarizability in atomic units .
is not interesting that describing purely classical effect of electron quiver motion requires the notion of anti - mater ?
in all our calculations below , we use the single active electron approximation , and represent the atomic core with the model potential using the correct quantum defects for rb and sr atoms from @xcite : @xmath73 ^ 2 \;,\ ] ] where @xmath74 , @xmath75 and @xmath76 are @xmath77-dependent screening parameters , and @xmath78 is the static polarizability of the core @xcite . for illustrative purposes , we pick rb and sr atoms because they have very different core polarizabilities . for rb , the static core polarizability is @xmath799 a.u . , whereas the polarizability for the sr@xmath80 ion with a 5s valence electron serving as spectator is @xmath7991 a.u .
, an order of magnitude larger .
this results in large differences in quantum defects @xmath81 for the @xmath82 rydberg states : for example in the 100s state of rb @xmath83 and for sr @xmath84 . in our calculations , we assume that the sr atom is in a @xmath85 state , and we ignore contributions from the resonant structure from the 5s valence electron . in order to evaluate polarizabilities in both the length and the velocity gauges , we have to evaluate sums over a complete set of states in and .
we check the completeness of our basis using the well known oscillator sum rule .
we find , @xmath86 while summing over @xmath87 states with @xmath88 through 200 for the 100s states of the rb and sr atoms .
we have shown that the rydberg state polarizability in the length and velocity gauges are identical as a result of the gauge invariance , _
i.e. _ @xmath90 .
furthermore , we have expressed these exact polarizabilities as a sum of the free electron polarizability @xmath2 and an exact correction .
the resonance structure of the polarizabilities is entirely captured by this correction term , whereas @xmath2 only provides a smooth background .
the fractional corrections for various @xmath91s states of rb and sr are shown in fig .
[ fig:01 ] in the ir region of the spectrum .
the resonant structure is evident in the plots , and the widths of the resonances increase with increasing principal quantum number . also , the sizes of the corrections grow smaller as @xmath91 is increased meaning that the near - free electron approximation to the rydberg electron polarizability gradually becomes more accurate away from the resonances . with the exception of the 10s state , the corrections are well below a per cent for both rb and sr , except at the resonances . comparing the upper and lower panels of fig .
[ fig:01 ] , we observe that the errors made by approximating @xmath92 and @xmath93 by @xmath2 are larger for sr than for rb .
this stems from the larger static polarizability of the sr@xmath80 ion with a 5s valence electron compared to the rb@xmath80 ionic core , a property which is contained in the model potential .
although the static core polarizabilities for these atoms differ by an order of magnitude , the corrections to @xmath2 are still much less than a percent in both cases , except in the immediate vicinity of resonances .
another feature seen in fig .
[ fig:01 ] is that the widths of the resonances decrease with increasing principal quantum number .
for example , for rb the width of the 20s-6p resonance is significantly larger than that of the 60s-6p resonance . on the other hand ,
the 10s-6p resonance is so large that the correction to @xmath2 for the 10s state of rb is at least a couple of per cent nearly for all frequencies spanned in fig .
[ fig:01 ] .
similar observations can also be made for sr .
the widening of these resonances with increasing @xmath91 can be qualitatively understood if one realizes that the widths of the resonances in eqs . and are controlled by the square of the dipole matrix elements , which scales as @xmath94 . ref .
@xcite found strong intensity landscape modulations of the effective polarizability for rydberg atoms trapped in ir lattices . because of the low frequency of the trapping field
, the rydberg electron polarizability can potentially deviate quite substantially from the free electron value .
this is because at sufficiently low frequencies the polarizability must approach its static limit , whereas the free electron value diverges . in this paper
, we have shown that the free electron approximation holds , even at ir wavelengths of thousands of nm . | ponderomotive energy shifts experienced by rydberg atoms in optical fields are known to be well approximated by the classical quiver energy of a free electron .
we examine such energy shifts quantum mechanically and elucidate how they relate to the ponderomotive shift of a free electron in off - resonant fields .
we derive and evaluate corrections to the ponderomotive free electron polarizability in the length and velocity ( transverse or coulomb ) gauges , which agree exactly as mandated by the gauge invariance .
we also show how the free electron value emerges from the dirac equation through summation over the dirac sea states .
we find that the free - electron ac stark shift comes as an expectation value of a term proportional to the square of the vector potential in the velocity gauge . on the other hand ,
the same dominant contribution can be obtained to first order via a series expansion of the exact energy shift from the second order perturbation theory in the length gauge .
finally , we numerically examine the validity of the free - electron approximation .
the correction to the free - electron value becomes smaller with increasing principal quantum number , and it is well below a per cent for 60s states of rb and sr away from the resonances .
many of the prominent schemes for realizing quantum logic gates for quantum computing @xcite , and studying few- and many - body physics via quantum simulation experiments using cold rydberg atoms rely on optically trapped rydberg atoms @xcite .
characterizing optical potentials experienced by rydberg atoms plays an essential role in experimental realization of these schemes @xcite .
the trapping ac stark shift seen by the rydberg atom is proportional to its dynamic polarizability , which is essentially that of a free electron quivering in the laser field .
the free electron polarizability @xmath0 , where @xmath1 is the field frequency , can be derived from classical arguments by time averaging the electron kinetic energy in the oscillating electric field @xcite .
it can be also obtained quantum mechanically through the first order perturbation theory , with the perturbation being proportional to the intensity @xcite . in this paper
, we examine the applicability of the free electron approximation for rydberg atoms in length and velocity gauges to gain insight as to how the exact ac stark shift relates to the free electron value in both gauges .
we show that @xmath2 emerges from a term in the hamiltonian proportional to the square of the vector potential @xmath3 in the velocity gauge , whereas it can be recovered to the leading order from a series expansion of the exact energy shift from the second order perturbation theory in the length gauge .
we compute the corrections to the free electron polarizability and conclude that the correction to @xmath2 in rydberg states is relatively small away from resonances .
our discussion elucidates that rather than performing an exact sum over states calculation to evaluate the rydberg state polarizabilities in the length gauge , it is simpler to calculate it as an expectation value of @xmath3 . throughout this paper
, we assume plane wave electromagnetic fields rather than standing wave optical traps .
notice that otherwise the optical trap intensity variation can substantially modulate the rydberg electron polarizability @xcite .
the paper is organized as follows : in the next section , we start by reviewing scalar and vector potentials of traveling waves in the length and velocity gauges . in section
ii , we derive ac stark shifts and polarizabilities in both gauges within the non - relativistic formalism , and examine how they reduce to the free electron value .
we then derive the free electron polarizability from the fully - relativistic dirac hamiltonian in section iii , and find that it comes from summation over the negative energy states in second order perturbation theory .
we conclude in section iv by presenting numerical results illustrating the errors made by approximating the rydberg state polarizabilities by the free electron values in the case of rb and sr atoms .
we use atomic units throughout this report unless specifically stated otherwise . |
the behavior of charged and neutral microparticles in uniform and nonuniform electric fields is a well established issue @xcite . nevertheless , in recent years
, interest has grown in application of electrophoretic ( ep ) and dielectrophoretic ( dep ) forces in a wide variety of micro - systems . in particular , ep and dep forces
are revealing themselves as a useful tool in biology and medicine for precise positioning and manipulation of single cells or bacteria .
in addition , specific field configurations with a high electric field gradient can be designed with the aid of micromachined electrodes allowing one to achieve actual lab - on - a - chip tweezer devices .
micro - objects have been efficiently trapped @xcite , and many biological applications have also been proposed ( see @xcite for a review ) such as , for instance , virus trapping in high - frequency electric field cages @xcite .
in addition , dep forces have been used in microfluidics devices for single- and multi - cell sample preparation @xcite or to achieve lateral deviation of particles in liquid flows @xcite .
dep tweezers have been also realized by lee _
@xcite in a fluid environment by means of a localized 3d movable electric field configuration .
optical tweezers ( ot ) @xcite are formidable and versatile tools for manipulating dielectric particles or cells having size ranging from tens of nanometers up to tens of microns and exerting forces ranging from 10 fn to 100 pn .
therefore , it is not surprising that ep , dep and photonic forces were combined to compare the range of applicability of the two trapping mechanisms , for reciprocal calibrations @xcite , to evaluate trapping efficiencies @xcite , or even to study molecular motors @xcite .
typically , dep forces are estimated by using the method of forced oscillation . in this case ,
an optically trapped bead is driven by a sinusoidal electric field ; the particle displacement is monitored by a position sensor and analyzed with a phase - sensitive detection scheme@xcite .
recently , several experiments demonstrated the possibility to combine photonic and ep forces for estimating the electric charge carried by a single polystyrene microsphere @xcite , or even for mapping the electric field generated in simple electrodes geometries @xcite . in the present work ,
we discuss a new approach to determine simultaneously ep and dep forces in proximity of microelectrodes . with respect to other approaches @xcite ,
our technique is based on the analysis of the deterministic and stochastic motions of a trapped particle in terms of the autocorrelation function , and is able to provide , simultaneously , quantitative and absolute measurements of the particle charge , as well as ep and dep forces .
a charged particle embedded in a fluid and confined in an optical trap , in presence of a nonuniform ( _ i.e. _ , its gradient @xmath0 ) oscillating electric field @xmath1 , undergoes five forces : ( i ) the elastic force of the trap , @xmath2 ; ( ii ) the drag force ( stokes force ) , @xmath3 ; ( iii ) the electrophoretic force , @xmath4 ; ( iv ) the dielectrophoretic force , @xmath5 ; and ( v ) the stochastic thermal force . in the following ,
we will consider a one dimensional geometry for simplicity , however the results can be straightforwardly extended to three dimensional geometries . the particle trajectory @xmath6 of a microsphere of mass @xmath7 , charge @xmath8 and radius @xmath9 , confined into an optical trap of stiffness @xmath10 and embedded in a fluid of viscosity @xmath11 is ruled by the following langevin equation : @xmath12 where @xmath13 is the diffusion coefficient
, @xmath14 is the hydrodynamic factor , @xmath15 is the boltzmann constant , @xmath16 is the absolute temperature , and @xmath17 is a white noise term with zero mean , almost everywhere discontinuous and with infinite variation .
the fourth term in eq . ( [ eq : langevin ] ) represents the dep force , where @xmath18 is the real part of the claussius - mossotti ( cm ) frequency - dependent function and @xmath19 is the relative permittivity of water ( @xmath2080 ) . for micrometric particles in water ( low reynolds number regime ) the inertia is negligible , and by assuming that the particle is driven by a sinusoidal electric field @xmath21 , the solution of eq .
( [ eq : langevin ] ) is given by : x(t ) & = x_th+ sin(2f_m t-_ep ) + & + = + & = x_th + asin(2f_m t-_ep)+b , [ eq : traj ] where @xmath22 is the thermal motion , while the second and third terms correspond , respectively , to the solutions of @xmath23 and @xmath24 forces .
the two phase terms that appear in eq .
( [ eq : traj ] ) are defined as @xmath25 and @xmath26 . the _ corner frequency _ @xmath27
is related to the trap stiffness and to the drag coefficient by the relation @xmath28 .
it is worth noting that the dep solution is composed by two terms : a displacement of the bead from its equilibrium position in the optical trap and an oscillation at a frequency twice the modulation frequency of the forcing electric field .
it is quite easy to calculate the autocorrelation function ( acf ) @xmath29 : @xmath30 .
\label{eq : acf}\end{aligned}\ ] ] the decay time @xmath31 represents the characteristic time of the optical trap and is connected to the trap stiffness and hydrodynamic factor by the relation : @xmath32 .
the experimental setup is based on a custom - built optical microscope described in @xcite . the laser beam (
nd - yag laser , innolight mephisto ne500 , @xmath33= 1064 nm , maximum output power = 500 mw ) was tightly focused with a high - numerical - aperture , water- immersion objective lens ( olympus , uplapo60xw3 , na=1.2 ) .
particle displacements were measured by using a ingaas quadrant photodiode ( qpd , hamamatsu g6849 ) at the back focal plane of the condenser lens @xcite .
calibrations of trap and quadrant photodiode ( qpd ) were carried out using the well established power spectral density method @xcite . in our experiment ,
the laser power was kept at @xmath203 mw , which corresponds to a trap stiffness of about @xmath34 n / m and a characteristic time @xmath35 ms .
negatively charged , sulfate - coated microspheres of polystyrene ( postnova , 1.06 g/@xmath36 density , 1.65 refractive index ) with a diameter of @xmath37 @xmath38 m were diluted in distilled deionized water ( conductivity @xmath39 ) to a final concentration of a few particles per microliter .
a droplet of such solution ( 50 @xmath38l ) was injected inside a sandwiched chamber consisting of a 150 @xmath38m - thick coverslip and a microscope slide .
both glass plates were coated with an indium tin oxide ( ito ) layer , hence forming parallel plate electrodes ( fig.[fig : setup](a ) ) that produced a uniform and predictable electric field @xmath40 at an applied voltage @xmath41 and electrode separation @xmath42 . in the same cell
, we also placed other two electrodes : a sharp tip with radius of curvature @xmath43150 nm and a gold coated tungsten wire ( @xmath44 m ) perpendicular to each other ( see fig.[fig : setup](b ) ) , separated by about 800 @xmath45 m .
they were positioned in the middle of the sample cell ( about 65 @xmath38 m from the bottom coverslip ) to avoid electro - osmotic flow effects@xcite .
we measured the electric charge of a trapped bead applying the external voltage only at the parallel ito electrodes as described in @xcite . in this case , the electric field was uniform ( @xmath46 in eq .
( [ eq : traj ] ) and ( [ eq : acf ] ) , _ i.e. _ , no dep forces were involved ) and a trapped bead moved back and forth along the direction of the electric field ( @xmath47-axis ) . due to the finite drift velocity of free ions , the oscillating amplitude depends on the frequency of the electric field .
we found the modulation frequency @xmath48 = 86.7 hz to optimize the motion amplitude condition ( for more details , see ref .
the bead s effective charge value resulted to be @xmath49 . at the end of the experiment
, we measured again the charge and verified that its value was unchanged .
it is worth noticing that the measured charge resulted about three orders of magnitude lower than the value provided by the manufacturer ( @xmath50 ) , but this discrepancy is ascribable to the screening effects caused by the free ions in water . since the used polystyrene beads showed certain degree of charge and radius heterogeneity , for accurate and reliable simultaneous force measurements , the second dep experiment was carried out on the same trapped bead previously analyzed for charge measurement .
following the electric charge determination , we proceeded to measure the force field near the tip .
in particular , we tracked the trajectories of a trapped microsphere in a raster grid around the tip .
the sample cell was translated in the @xmath51 plane with steps of 2 @xmath38 m in both directions as in scanning probe measurements . at each position of the grid ,
the particle trajectory was acquired for 20 s. all measurements were carried out by using a sinusoidal voltage amplitude @xmath52 v ( between tip and wire ) at a modulation frequency of 86.7 hz .
we chose an applied voltage small enough to avoid nonlinear phenomena in the fluid . at same time , this allowed to acquire bead displacements in a volume much smaller than the length scale of variation of the resulting electric field gradient , therefore enabling a more accurate probing .
we estimated the coefficients @xmath53 and @xmath54 of eq .
( [ eq : traj ] ) , at a given position , from the amplitudes of the acf fuction . at the closest position to the tip where the gradient is larger
, we found @xmath55 .
such values confirm that in our experiment ( i ) the bead s excursion from the center of the optical trap results always small enough to fall in the linear range of our qpd ( @xmath20300 nm in our case ) and that ( ii ) the bead s displacement is also much smaller than its radius .
the first result guarantees that the bead trajectory is correctly sampled , whereas the latter shows that the bead oscillates in a very confined volume where the electric field and its gradient can be reasonably assumed to be constant . )
( solid lines ) measured ( a ) when the bead is trapped near the tip , and ( b ) far away from the tip .
error bars are smaller than the markers size , the fitting produced a value of the reduced @xmath56 of 0.34 ( panel ( a ) ) and 0.47 ( panel ( b ) ) , indicating its goodness ] figure [ fig : acf ] shows the experimental acf curves along the @xmath57-axis , obtained by keeping the trapped bead at the two limit positions : the closest one to the tip ( panel ( a ) ) , and a position far from the tip ( panel ( b ) ) . we can observe that , while the oscillation is symmetric far away from the tip , it is instead clearly asymmetric in its proximity .
this suggests that far from the tip , where the electric field is essentially uniform , the bead motion is governed by the ep force only , _
i.e. _ , the acf is reduced to an oscillating term at frequency @xmath48 with amplitude @xmath58 . on the contrary , near the tip where @xmath59
, the acf contains also a second oscillating term at frequency 2@xmath48 with amplitude @xmath60 .
fitting our experimental data to eq .
( [ eq : acf ] ) as model , we found a very good agreement between theory and experiment as shown in fig .
[ fig : acf ] .
as discussed in ref.@xcite , the same analysis can be performed using the power spectral density .
this last exhibited two sharp peaks , at frequencies @xmath48 and @xmath61 , superimposed to the broad lorentzian shape related to the thermal motion .
the results of the force field patterns measured in a region around the tip are shown in fig .
[ fig : field ] . as it can be noted , both ep and dep forces are below 1 pn , a fact which points out a high sensitivity of the technique .
in particular , ep forces result about four times larger than dep forces .
in addition , at the modulation frequency used , the cm term is positive and the dep force points toward the spatial position where the maximum electric field gradient occurs .
our technique can be easily generalized for the case of higher frequencies , for which the cm term is expected to change sign . indeed , at high frequency ,
although the trapped bead is no more able to follow such a fast oscillation , its center displacement can be detected thanks to the high sensitivity of our position detector .
in conclusion , we have proved that a charged microsphere held in an optical trap can be handled as a probe for measuring simultaneously the absolute values of the electrophoretic and dielectrophoretic forces .
our method is based on the autocorrelation function , which offers high sensitivity and avoids the use of the phase - sensitive detection schemes currently employed . in particular
, our approach allows one to detect very weak dep forces .
we reconstructed the pattern of ep and dep forces in a plane around a tip electrode , a geometry that can be generalized to more complex ones with potential 3d analysis as well .
our experiment points out the necessity to take into account the dielectrophoretic force contribution for determining the actual forces experienced by a dielectric object in presence of a high electric field gradient , and can be of potential interest for studying and mapping ep and dep forces around scanning probe plasmonic tips that are typically characterized by highly localized electric fields@xcite .
p. cheng , m. j. barrett , p. m. oliver , d. cetin , and d. vezenov , `` dielectrophoretic tweezers as a platform for molecular force spectroscopy in a highly parallel format , '' lab chip * 11 * , 42484259 ( 2011 ) .
t. schnelle , t. mller , s. fiedler , s. shirley , k. ludwig , a. herrmann , g. fuhr , b. wagner , and u. zimmermann , `` trapping of viruses in high - frequency electric field cages , '' naturwissenschaften * 83 * , 172176 ( 1996 ) .
n. demierre , t. braschler , p. linderholm , u. seger , h. van lintel , and p. renaud , `` characterization and optimization of liquid electrodes for lateral dielectrophoresis , '' lab chip * 7 * , 355365 ( 2007 ) .
t. schnelle , t. mller , r. hagedorn , a. voigt , and g. fuhr , `` single micro electrode dielectrophoretic tweezers for manipulation of suspended cells and particles , '' biochim .
acta * 1428 * , 99 105 ( 1999 ) .
e. papagiakoumou , d. pietreanu , m. i. makropoulou , e. kovacs , and a. a. serafetinides , `` evaluation of trapping efficiency of optical tweezers by dielectrophoresis , '' j. biomed .
opt . * 11 * , 0140350140358 ( 2006 ) .
y. hong , j .- w .
pyo , s. h. baek , s. w. lee , d. s. yoon , k. no , and b .-
kim , `` quantitative measurements of absolute dielectrophoretic forces using optical tweezers , '' opt .
lett . * 35 * , 24932495 ( 2010 ) .
r. galneder , v. kahl , a. arbuzova , m. rebecchi , j. o. rdler , and s. mclaughlin , `` microelectrophoresis of a bilayer - coated silica bead in an optical trap : application to enzymology , '' biophys . j.
* 80 * , 22982309 ( 2001 ) .
g. s. roberts , t. a. wood , w. j. frith , and p. bartlett , `` direct measurement of the effective charge in nonpolar suspensions by optical tracking of single particles , '' j. chem .
phys . * 126 * , 194503 ( 2007 ) .
g. pesce , g. rusciano , a. sasso , r. isticato , t. sirec , and e. ricca , `` surface charge and hydrodynamic coefficient measurements of bacillus subtilis spore by optical tweezers , ''
colloid surface b * 116 * , 568575 ( 2014 ) .
g. pesce , b. mandracchia , e. orabona , g. rusciano , l. de stefano , and a. sasso , `` mapping electric fields generated by microelectrodes using optically trapped charged microspheres , '' lab chip * 11 * , 41134116 ( 2011 ) .
a. buosciolo , g. pesce , and a. sasso , `` new calibration method for position detector for simultaneous measurements of force constants and local viscosity in optical tweezers , '' opt . commun . * 230 * , 357368 ( 2004 ) .
g. rusciano , g. zito , r. isticato , t. sirec , e. ricca , e. bailo , and a. sasso , `` nanoscale chemical imaging of bacillus subtilis spores by combining tip - enhanced raman scattering and advanced statistical tools , '' acs nano * 8 * , 1230012309 ( 2014 ) . | herein , charged microbeads handled with optical tweezers are used as a sensitive probe for simultaneous measurements of electrophoretic and dielectrophoretic forces .
we first determine the electric charge carried by a single bead by keeping it in a predictable uniform electric field produced by two parallel planar electrodes , then , we examine same bead s response in proximity to a tip electrode . in this case , besides electric forces , the bead simultaneously experiences non - negligible dielectrophoretic forces produced by the strong electric field gradient .
the stochastic and deterministic motions of the trapped bead are theoretically and experimentally analysed in terms of the autocorrelation function . by fitting the experimental data ,
we are able to extract simultaneously the spatial distribution of electrophoretic and dielectrophoretic forces around the tip . our approach can be used for determining actual , total force components in the presence of high - curvature electrodes or metal scanning probe tips . |
x - ray spectra of seyfert 1 galaxies commonly show an iron k@xmath1 emission line at @xmath36kev .
the line is often extremely broad , with a velocity width of @xmath4 .
furthermore , the asymmetries in the line profiles are well explained by relativistic effects . both of these facts suggest that the line is emitted from the surface layers of the accretion disk within a few gravitational radii @xmath5 of the black hole ( bh ) itself .
it is now widely believed that this spectral feature is a relatively clean probe of the immediate environments of supermassive bhs ( see fabian et al . , 2000 , and references therein ) .
one of the best studied broad iron line sources is the nearby bright seyfert 1 galaxy mcg@xmath06 - 30 - 15 ( @xmath6 ) .
this was the first active galactic nucleus ( agn ) for which a relativistic iron line disk profile was measured . in their _ advanced satellite for cosmology and astrophysics _
( asca ) observations , tanaka et al . found that the iron line profile could be explained by emission from an accretion disk around a non - rotating ( schwarzschild ) bh . in a detailed reanalysis ,
however , iwasawa et al . found a much broader line during a period of low continuum x - ray flux , the so - called `` deep minimum state '' .
the line becomes so broad that disk models require emission from within @xmath7 , suggesting either a rotating ( kerr ) bh , with its marginally stable orbit at @xmath8 ( iwasawa et al .
, 1996 , dabrowski et al . ,
1997 ) , or iron fluorescence from material spiraling into the bh at @xmath9 @xcite .
such a broad line has later been confirmed , e.g. , by guainazzi et al .
( 1999 ) and lee et al .
( 1999 ) . using the new generation of x - ray satellites such as _ xmm - newton _ with their improved collecting area and x - ray ccd spectral resolution , the fe line of seyferts can be studied in greater detail than was possible before . in this _
letter _ , we present data from a 100ksec _ xmm - newton _ observation of mcg@xmath06 - 30 - 15 .
our observation covered most of _ xmm - newton _ s orbit 108 , on 2000 june 11/12 , and was simultaneous with the _ rossi x - ray timing explorer _ ( rxte ) . here
, we report on data from the european photon imaging cameras onboard _ xmm - newton _
( epic , strder et al . , 2001 ,
turner et al . , 2001 ) . to prevent photon pile - up ,
the epic - pn camera was operated in its small window mode ( using the medium thick filter to prevent optical light contamination ) , and the epic mos-1 camera was operated in its timing mode .
the mos-2 camera was operated in full - frame mode to study the field surrounding mcg@xmath06 - 30 - 15 .
although we primarily rely on the pn - data here , the pile - up in the mos-2 data can be reduced to @xmath10 when using single events only .
therefore , we can use the mos-2 data to check for the instrument - independency of our results . as of 2001
august , the official version of the _ xmm - newton _ science analysis software ( sas ) only contained a preliminary model for the charge transfer efficiency ( cte ) of the pn - chips .
therefore , we used an internal sas version implementing an improved model for the epic - pn cte ( haberl , priv .
comm . ) . to avoid remaining response matrix uncertainties , we concentrate on the energy range from 0.5 to 11kev . to produce epic - pn source and background spectra , we collected events from two circles of radius 10 pixels on and off the source for those times where the source count rate was below @xmath11 .
this upper limit of the count rate is necessary to avoid phases where the core of the mcg@xmath06 - 30 - 15 point spread function was slightly piled up .
the total resulting epic integration time was 54ksec .
we checked the background lightcurve for periods of severely increased background , but none were found . for spectral modeling ,
response matrices appropriate for the improved cte model and appropriate for the source position on the epic chip were used .
furthermore , we corrected the exposure time in the spectra for the @xmath371% livetime of the pn small window mode @xcite and perform simultaneous fits with single and double events .
we estimate that the overall uncertainties in the spectral calibration and flux calibration due to these procedures are at most a few percent .
figure [ fig : fits]a shows the ratio of the 0.511kev epic - pn data to a spectral model consisting of a pure power - law ( fitted in the 211kev band ) modified by galactic absorption ( @xmath12 , using the xspec phabs model with cross - sections similar to those described by wilms , allen , & mccray , 2000 ) . below @xmath13kev , there is significant spectral complexity which is due to a warm absorber ( nandra & pounds 1994 , reynolds & fabian 1995 , lee et al .
2001 ) , possibly superposed with a complex of relativistically broadened soft x - ray recombination lines @xcite . at 37kev , a broad hump in the spectrum suggests the identification with disk reflection signatures and , especially , the broad iron k@xmath1 line .
the putative broad iron line is more apparent once the soft x - ray spectral complexity has been modeled .
this is not a trivial exercise since the physical nature of this complexity is still a matter of debate . to start with
, we consider a warm absorber scenario ( this will be generalized to include soft x - ray relativistic emission lines below ) .
we assume that the continuum consists of a power - law modified by galactic absorption .
we also allow for the possibility of intrinsic neutral absorption in mcg@xmath06 - 30 - 15 ( although the best fitting column density is always consistent with zero ) . by fitting the high resolution x - ray spectrum from the reflection grating spectrograph ( rgs )
, we construct an empirical warm absorber model ( based on lee et al .
2001 ) with absorption edges of cv , ovii , oviii , and nex ( at 392ev , 740ev , 870ev , and 1362ev , respectively , with best fitting optical depths at threshold of @xmath14 , @xmath15 , @xmath16 , and @xmath17 ) , a gaussian absorption feature below the ovii and oviii edges to crudely model blended resonance absorption lines , and four moderately broad and weak gaussian emission features at observed energies of @xmath18kev , @xmath19kev , @xmath20kev , and @xmath21kev , which are merely included to fit the rgs data ; we do not attempt to make a physical identification of these features . after applying this empirical model to the data ,
the soft x - ray residuals are reduced to less than @xmath22 , and the broad iron line feature is much more obvious ( fig . [
fig : fits]b ) .
the best fit photon index is @xmath23 ( goodness of fit @xmath24 , all uncertainties are quoted at the 90% level for one interesting parameter , @xmath25 ) .
we now consider the nature of the hard spectral complexity .
while it may be tempting to model the hump - like feature as pure broad iron line emission , this is not a physically consistent model .
the equivalent width of this feature ( measured relative to the continuum at the iron k@xmath1 rest - frame energy ) is @xmath26kev , and it would be impossible to obtain such a broad iron line without also observing the x - ray continuum photons that have been backscattered from the accretion disk .
consequently , we will not present pure broad iron line models and jump , instead , to more physically motivated spectral models which include both iron line fluorescence and the backscattered `` reflection '' continuum . within the framework of such models ,
the 37kev hump is a combination of iron line fluorescence and backscattered continuum photons ( with the fe k edge responsible for some fraction of the drop in flux at @xmath27kev ) .
we construct a spectral model appropriate for the case of reflection from the keplerian regions of an accretion disk around a near - extremal kerr black hole ( @xmath28 ) .
the inner edge of the x - ray reprocessing region is taken to be the radius of marginal stability , @xmath29 , and the outer edge is taken to be @xmath30 .
we fix the inclination of the inner disk to be @xmath31 ( tanaka et al . , 1995 ) and note that our results below only slightly depend on the value of @xmath32 chosen .
we assume that the reflected emission has an emissity profile @xmath33 and leave @xmath34 as a free parameter .
the reflection continuum is described using the pexriv model of xspec and normalized to the case where the accretion disk intercepts half of the x - ray power - law continuum photons ( i.e. , @xmath35 ) . both , the iron line profile and the reflection continuum are then relativistically smeared ( using the model of laor , 1991 , as the kernel ) . here , and
throughout the rest of the spectral modeling discussed in this paper , we also include a narrow ( unresolved ) gaussian line at 6.4kev with equivalent width @xmath36 to model the obvious narrow line core .
this component presumably originates from distant material such as the putative molecular torus of seyfert unification schemes , and has recently be seen in several other agn ( e.g. , in ngc 5548 , yaqoob et al . , 2001 , or in mkn 205 , reeves et al . , 2001 ) .
firstly , we consider the case of a fairly cold accretion disk ( rest - frame iron line energy of 6.4kev , disk ionization parameter @xmath37 , and @xmath35 ) .
the resulting best fitting parameters are @xmath38 , @xmath39 , @xmath40ev , and @xmath41ev ( @xmath42 ) .
however , this model fails to explain the data as well as possessing an unphysically large broad line equivalent width .
significant residuals ( up to @xmath43 ) remain in the 311kev band .
the pattern of residuals suggest that the model is failing to produce the red - wing of the 37kev hump , and is underestimating the flux in the blueward edge of this feature .
a hard tail above 8kev suggests that we are also underestimating the strength of the high energy reflection continuum .
if we allow the relative normalization of the reflection continuum , @xmath44 , to be a free parameter , the fit converges to a very large value of @xmath44 ( @xmath45 , also @xmath46 and @xmath47 , @xmath48 ) . despite the slightly better @xmath49 , large residuals around @xmath37kev
remain .
ionization of the accretion disk surface is another way of increasing the amount of reflection continuum . to model this ( with the relative normalization of the reflection continuum fixed to @xmath35 ) , we allow the ionization parameter of the disk @xmath50 , the temperature of the disk @xmath51 and the energy of the iron line @xmath52 to be free parameters ( @xmath52 is constrained to lie in the range of allowable fe-@xmath53 energies ; 6.40kev to 6.97kev ) .
the resulting best fit has @xmath54 , @xmath55 , @xmath56kev , @xmath57 , @xmath58k , @xmath59ev , and @xmath60ev ( @xmath61 ) .
this leads to a significantly better fit with the higher iron edge / line energies explaining the residual at @xmath27kev .
however , significant residuals still exist in the 35kev range ( suggesting that we are still under - predicting the flux in the red - wing of the reflection feature ) and above 8kev .
detailed examination of this last spectral model shows that features in the _ soft _ x - ray band below 1.5kev determine the best fitting ionization parameter and temperature .
consequently , such fits must be viewed with suspicion there are a number of physical processes which are not included in this reflection model but which may well contribute to the soft x - ray spectrum ( most notably , the recombination line and edge emission from c , n , o , and ne )
. it would also not be surprising if our empirical warm absorber model left small soft x - ray residuals which may then be mistakenly fit by ionized reflection models .
firstly , we can consider the ionized reflection fits ignoring all data below 1.5kev . refitting the ionized disk model ( including the empirical warm absorber ) to the 1.511kev band results in a good fit with no clear residuals .
the best fitting parameters are : @xmath62 , @xmath63 , @xmath64kev , @xmath65ev , and @xmath66ev ( @xmath67 ) . @xmath51 and @xmath50 are unconstrained by the data , while the relative normalization of the reflection continuum is constrained to be @xmath68 . during the fitting @xmath44 is anti - correlated with @xmath69 .
these quantities are physically consistent with each other ( i.e. , the iron line has the appropriate strength for that value of @xmath44 and @xmath50 ) when @xmath70@xmath71 ( corresponding to @xmath72@xmath73ev ) .
this spectral modelling suggests that the accretion disk may be ionized .
consequently , the k@xmath1 line need not be intrinsically narrow as we have assumed so far , but might be compton broadened in the hot skin of the accretion disk ( note that compton broadening alone can not explain the large line width ; reynolds & wilms , 2000 ) .
we examine this possibility by relaxing our assumption of a narrow intrinsic line and , instead , model the _ intrinsic _ iron line with a gaussian profile of width @xmath74 .
using the 1.511kev data , the best fitting value is @xmath75 ( with other parameters as above ) .
even if we force the line to be very broad ( e.g. , @xmath761kev , at the expense of reducing the goodness of fit by @xmath77 ) , the data still require @xmath78 . the second approach for handling the soft x - ray spectrum is to explicitly model the recombination k - lines of c , n , o and ne .
qualitatively , this model is the same as the mixed emission - line model of branduardi - raymont et al .
( 2001 ) with the presence of additional warm absorption .
the best fitting parameters are @xmath79 , @xmath80 , @xmath81kev , @xmath82ev , @xmath83 and @xmath84ev ( @xmath85 ) .
again , the relative reflection normalization is poorly constrained by these data , but self - consistency is achieved for @xmath86@xmath71 and @xmath87@xmath73ev .
we conclude that the parameterization of the soft energy spectrum does not qualitatively change the results for the broadened fe k@xmath1 line .
we note that the pexriv reflection model employed here is also incomplete due to the neglect of compton smearing of the absorption edges .
we do not expect this to seriously comprise our results since the observed smearing is much greater than that expected from compton effects .
however , we will examine these effects using improved models ( ballantyne , iwasawa & fabian 2001 ; nayakshin & kallman 2001 ) in a future publication .
preliminary investigations show our current conclusions to be robust . finally , to check the dependence of these result on possible remaining calibration uncertainties we used the data from the mos-2 camera .
the best fit parameters we find are consistent to within the error bars , with slight deviations at the si edge that are a known calibration issue .
we conclude that our results are independent of any remaining calibration issues with the epic instruments .
to summarize , our physically motivated best - fitting model consists of reflection from an ionized accretion disk that emits h - like iron k@xmath1 fluorescence with a relative reflection fraction of @xmath88@xmath71 and a broad iron line equivalent with of @xmath89@xmath73ev .
the 210kev flux of @xmath90 is comparable with the `` deep minimum state '' found by iwasawa et al .
( 1996 , @xmath91 ) . _
the most interesting feature of these spectral models is that a very steep emissivity profile @xmath92@xmath93 for the iron - line / reflection features is required_. we address the implications of this result in the next section .
finally , we note that if @xmath94 is allowed to vary , the data require @xmath95 in order to model the red - wing of the 38kev hump .
we explicitly note that the data can not be adequately described by any reflection model with an inner radius of @xmath96 , the radius of marginal stability around a schwarzschild black hole .
thus , models in which the x - ray reflection occurs in the keplerian part of accretion disks around a non - rotating black hole can not explain these data . in principle , there can be x - ray reflection inside of the radius of marginal stability ( reynolds & begelman 1997 ) .
however , as discussed below , the extremely steep emissivity index @xmath34 required by our data is very hard to understand in the context of schwarzschild geometry .
very steep emission profiles are required in all of our good fits to the fe k@xmath1 line and reflection continuum . here , we explore the implications of this result , assuming that the reflected flux ( including the iron line ) from a local patch of the disk is proportional to the x - ray continuum flux ( at the iron k - shell edge ) irradiating that patch .
we will assume that a fixed fraction @xmath97 of the energy released locally in the body of the accretion disk is transported into an accretion disk corona and then radiated in the x - ray band .
studies of such coronae suggest that @xmath98 in order to produce the observed continua ( haardt & maraschi 1993 , dove et al .
thus , while our assumption may not be true in detail , it must be approximately true across a wide range of disk radii .
can standard bh accretion disks models ( novikov & thorne 1973 , riffert & herold 1995 ) explain the observed emissivity profiles ? with the assumptions described above , this question reduces to an examination of the radial distribution of energy dissipation in such disk models . for a disk around an @xmath28 bh ,
the flux emitted from the disk per unit proper area of the disk , @xmath99 , peaks at @xmath100 and then gradually steepens to approach @xmath101 . at no point
does @xmath99 become as steep as @xmath102 , as required by our iron line observations .
this is true for any assumed bh spin . also , as noted above , the emissivity profile is steeper than accretion disk models even if one accounts for intrinsic broadening of the iron line via comptonization .
we note that there are two complications that may be relevant to real disks , even though they go beyond the realm of standard disk models .
firstly , magnetic fields might couple material within @xmath103 to the rest of the disk , thereby permitting continued energy extraction from this material ( krolik 1999 ; agol & krolik 2000 , hereafter ak00 ) .
non - relativistic accretion disk simulations employing pseudo - newtonian potentials suggest that the stresses and presumably the dissipation remain fairly flat within this region ( hawley & krolik 2001 ; armitage , reynolds & chiang 2001 ) .
hence , it would seem that the required emissivities are difficult to achieve through such effects for disks around schwarzschild holes .
secondly , x - rays produced away from the equatorial plane might be gravitationally focussed into the central regions of the disk .
however , using the method of petrucci & henri ( 1997 ) , we find that such effects can not produce the observed emissivity profile unless the x - ray source is already situated in the very central regions of the disk .
we conclude that the disk has to be irradiated in a more centrally concentrated manner than predicted by current pure - accretion disk models . in other words , we require some additional x - ray source that is both powerful and very centrally concentrated .
there is one obvious candidate : _ x - rays that are associated with the magnetic extraction of bh spin energy . _ here , we explore the suggestion that the central x - ray source is associated with the extraction of bh spin energy , focusing on scenarios in which the bh spin is extracted via magnetic fields that pierce the ( stretched ) event horizon @xcite .
an alternative class of models in which the magnetic fields do not pierce the horizon will be addressed below . in order to produce the required steep irradiation profile ,
the x - ray production must occur close to the disk itself with a production rate that is a steeply declining function of radius .
thus , we are led to consider models in which the magnetic field lines connect the rotating event horizon to the accretion disk and/or disk corona . for definiteness
, we will consider the canonical near - extremal kerr bh . in models
in which the magnetic field lines are strongly coupled to the body of the accretion disk , the magnetic field transmits a retarding couple to the bh if the bh rotates faster than the magnetically connected region of the disk where the extracted rotational energy will be deposited . in order to power the observed x - ray source
, the field strength close to the bh needs to be @xmath104 @xcite .
this is not an impossibly high field ; the corresponding magnetic pressure is still substantially below the ram pressure of the accretion flow within @xmath105 and so can be confined to the bh region .
if a ring within the disk of width @xmath106 is connected to the event horizon with magnetic flux @xmath107 , then the power dumped into the ring is ( setting @xmath108 ) @xmath109 where @xmath110 is the angular velocity of the accretion disk , and the angular velocity of the event horizon is @xmath111 ( li 2000 ) .
the bh resistance , @xmath112 , is a function of disk radius , @xmath113 , defined by a map from the bh horizon to the accretion disk along the magnetic field lines .
viscous forces then transport this energy _
outwards _ by some distance before it is dissipated and radiated .
ak01 and li ( 2000 ) examine this process in detail and compute the radial dependence of the energy dissipation when the magnetic field connects to the accretion disk at one particular radius .
we can use the formulae of ak01 ( also see li 2000 ) as the green s function for computing the more general case .
it is beyond the scope of this letter to address the detailed distribution of the hole - threading magnetic field across the disk surface , @xmath114 , or the nature of @xmath115 .
if all of the extracted spin energy is dumped into a ring at @xmath103 , the emissivity is very steep ( @xmath116 ) within @xmath117 and gradually flattens to @xmath118 at large radius @xcite .
more generally , the spin energy will be deposited into a range of radii , flattening this profile .
we will assume @xmath119 and @xmath120 ( @xmath121 corresponds to a dipole field for @xmath122 ) .
for @xmath123 and @xmath121 , @xmath124 at @xmath125 and is flatter inside that radius . to produce emission profiles as extreme as those required by our data ( i.e. , @xmath126 at @xmath127 ) , field configurations as concentrated as @xmath1284 ( with some dependence on @xmath129 ) . such conditions could be achieved , for example , if the field were `` pinned '' onto the black hole by the ram pressure of the accretion flow . we note that the models addressed by ak00 do not employ the blandford - znajek effect but , instead , magnetically torque the disk via coupling to _ matter _ deep within the plunging region .
for parameters relevant to our discussion , the extra energy source is provided by the black hole spin via the penrose effect occuring within the radius of marginal stability ( but outside of the stretched horizon ) . in this case , all of the extra torque is provided at the inner edge of the disk ( rather than across a range of radii as addressed above ) and so a steep emissivity profile is a natural outcome . finally , we note that the large self - consistent value of @xmath130 may have its origin in general relativity .
some fraction @xmath131 of the upwardly directed disk emission will be bent by the strong gravity and strike the disk again , ( `` returning radiation '' , cunningham 1975 , speith , riffert & ruder 1995 , ak00 ) further enhancing @xmath69 .
this can enhance the relative amount of reflection by up to a factor of 2 .
further computations are required to assess the effect of returning radiation on the emissivity profile in a self - consistent manner .
we have presented _ xmm - newton_-epic observations of mcg@xmath06 - 30 - 15 containing a spectral feature that is best described as an extremely broad and redshifted x - ray reflection feature . both on the basis of the extreme spectrum and the source flux
, it seems likely that we have caught mcg@xmath06 - 30 - 15 in the peculiar `` deep minumum '' state first noted by _ asca _
( iwasawa et al . 1996 ) .
the extreme nature of the line profile leads us to conclude that it originates from the centralmost parts of the accretion disk .
standard accretion disk models can not produce an emissivity profile which is centrally concentrated enough to produce the observed feature .
it also seems unlikely that gravitational focussing of the continuum x - rays or magnetic coupling of the plunging region to the rest of the disk can ameliorate this conclusion .
therefore , we suggest that , during the deep minimum state , x - rays associated with the magnetic extraction of the spin energy of the black hole are dominating the emission , producing a sufficiently compact source to explain our observations . our results thus confirm the suggestion of iwasawa et al .
( 1996 ) and dabrowski et al .
( 1997 ) that mcg@xmath06 - 30 - 15 possesses a rapidly rotating black hole . in a forthcoming paper
, we will present a more detailed analysis of our observation , including simultaneous fits with the rxte pca , a study of the time dependence of the reflection features and the x - ray continuum , and fits using a larger and more detailed set of continuum and reflection models .
we thank a.c .
fabian , f.w .
haberl , j.h .
krolik , and m. kuster for useful conversations , and the referee , chris done , for her very prompt reply and useful comments .
we acknowledge support from dlr grant 50 ox 0002 ( jw , rs , ek ) , hubble fellowship grant hf-01113.01 - 98a ( csr ) , and nsf grant ast 98 - 76887 ( csr , mcb ) . 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 ) . | we present _ xmm - newton _
european photon imaging camera ( epic ) observations of the bright seyfert 1 galaxy mcg@xmath06 - 30 - 15 , focusing on the broad fe k@xmath1 line at @xmath2kev and the associated reflection continuum , which is believed to originate from the inner accretion disk .
we find these reflection features to be _ extremely _ broad and red - shifted , indicating its origin from the very most central regions of the accretion disk .
it seems likely that we have caught this source in the `` deep minimum '' state first observed by iwasawa et al .
( 1996 ) .
the implied central concentration of x - ray illumination is difficult to understand in any pure accretion disk model .
we suggest that we are witnessing the extraction and dissipation of rotational energy from a spinning black hole by magnetic fields connecting the black hole or plunging region to the disk .
[ firstpage ] accretion disks black hole physics
galaxies : individual ( mcg@xmath06 - 30 - 15 ) galaxies : seyferts |
ngc 5128 ( centaurus a ) , the nearby giant elliptical galaxy , hosts the closest powerful active galactic nucleus .
this double - lobed radio source contains a strong jet discovered in x - rays ( @xcite , @xcite ) and well - studied in radio with the vla ( @xcite , @xcite , @xcite ) and with the vlbi ( cf .
@xcite ) . the proximity of ngc 5128 makes it one of the best candidates for studying the inner region around a massive black hole , assumed by the standard model to be at the core of all agns ( e.g. @xcite , @xcite ) .
however , the large warped dust lane , approximately an arcminute wide with associated gas , young stars , and hii regions , dominates the morphology and polarization properties at visible wavelengths , effectively obscuring the nucleus and inner half - kiloparsec of the galaxy and the jet .
ground - based ir observations have provided evidence for a strong , heavily polarized source at the nucleus ( cf .
@xcite , @xcite ) .
previous hst wf / pc-1 imaging polarimetry of the inner region of ngc5128 ( @xcite ) identified the nucleus of the galaxy with the most obscured and reddened feature of the emission , several arcseconds sw of the brightest optical emission near the center of the dust lane and consistent with the ir results . while the discovery of gas disks at the centers of nearby galaxies with hst has opened up new possibilities for determining masses of central black holes by studying kinematics of the disks ( cf .
@xcite , @xcite , @xcite , @xcite ) the dust lane does not permit such studies for ngc 5128 in the optical .
we report here hst near infrared camera and multi - object spectrometer ( nicmos ) observations at 2.2 and which allow us , for the first time , to study the structure of the circumnuclear region at high resolution .
we summarize the observations and data reduction in section 2 , the results in section 3 , and discuss the implications of these data for standard agn and jet models in sections 4 and 5 . throughout
, we assume a distance to centaurus a of 3.5 mpc ( @xcite ) , whence 1@xmath117pc .
the nuclear region of ngc 5128 was observed for three orbits on 11 august 1997 using nicmos camera 2 ( 0075/pixel ) with f222 m , f187n and f190n filters .
all observations were carried out with a multiaccum sequence ( @xcite ) , the detector read out non - destructively several times during each integration to facilitate removal of cosmic rays and saturated pixels .
the integration times were 2304 sec for the f187n and f190n filters ( and adjacent continuum ) and 1280 sec for the f222 m filter ( 2.2medium band continuum ) .
the last observation was also performed off - source to enable telescope thermal background subtraction .
the data were calibrated using the pipeline software calnica v3.0 ( @xcite ) and the best reference files in the hubble data archive to produce flux calibrated images .
a continuum - subtracted image was obtained by direct subtraction of the f187n and f190n images .
we verified the continuum subtraction by rescaling the continuum by up to @xmath4 before subtraction and establishing that this did not significantly affect the observed emission line structure . a small ( few percent ) drift in the nicmos bias level resulted in spatially dependent residuals in the calibrated images ( the `` pedestal '' problem , @xcite ) .
this effect was removed by fitting and subtracting a first degree polynomial surface to each quadrant of the continuum subtracted images .
the calibrated 2.2 @xmath5 m continuum image in fig .
[ fig : continua ] shows the rather smooth and regular structure expected for a typical elliptical galaxy .
this smooth structure on small scales contrasts strongly with the patchy and irregular emission observed on the same spatial scales in r and i with wf / pc-1 ( @xcite ) .
the continuum image is dominated by a strong unresolved source which we interpret as the nucleus of the galaxy .
the intensity of the nucleus is estimated by fitting a model psf derived using the tinytim software ( @xcite ) .
we took particular care in removing the psf artifacts present beyond the first airy ring .
the total non - stellar plus stellar flux of the unresolved nuclear source is ( [email protected] ) erg @xmath7 s@xmath8 @xmath8 ( magnitude [email protected] in the vega reference system ) .
[ fig : radprof ] shows the azimuthally averaged radial light profile obtained by fitting ellipses to the isophotes in the f222 m image and comparing with the model psf fitted to the data .
the nucleus position is ra=13:25:27.46 , dec=-43:01:10.2 ( j2000 ) , with the @xmath61 uncertainty of the hst guide star catalogue ( gsc ) .
this position is consistent with that found in the radio and ground based ir data , and with both the peak of reddening and the peak of polarization found with wf / pc-1 .
in contrast to the relatively smooth light distribution seen in f222 m , the continuum subtracted image ( fig .
[ fig : pa ] ) reveals a number of emission features above the threshold of 6 erg @xmath7 s@xmath8 arcsec@xmath9 ( 3@xmath10 of the noise measured in emission free areas ) .
in addition to the unresolved nuclear source , there is a prominent elongated structure , with some sub - structure , centered on the nucleus , a finger ( n1a ) @xmath109 n - e of the nucleus , two arc - like filaments ( a , b ) located , respectively , at @xmath1 29 n - e and 35 n - n - e of the nucleus , and several other weaker emission knots and filaments distributed over the field of view .
the dominant feature by far is the structure around the nucleus .
isophotal ellipse fitting gives an average ellipticity of 0.5 and a constant position angle of 33 outside a radius of @xmath1101 , within which the nucleus dominates .
the total extension along the major axis is @xmath1 23 ( @xmath1 40pc ) and the total flux from the extended nuclear emission above the level of 1.1 erg @xmath7 s@xmath8 arcsec@xmath9 is ( [email protected] ) erg @xmath7 s@xmath8 . in comparison ,
the observed flux from the unresolved emission is ( [email protected] ) erg @xmath7 s@xmath8 .
the inset in fig .
[ fig : pacont ] overlaying the contours on a gray scale plot of the nuclear emission after subtracting the isophotal fit model clearly shows n1a and another sub - structure s - s - e of the nucleus .
the elongated structure around the nucleus seen in emission has an overall extent of @xmath12 parsecs .
its position angle of 33is approximately perpendicular to that of the dust lane .
it is not oriented along the axis of the x - ray / radio jet , which has a position angle of @xmath11 50 - 55 as seen at x - ray ( @xcite ) , cm ( @xcite ) and vlbi ( @xcite ) wavelengths .
it is also not perpendicular to the jet axis , as reported for an extended structure seen with isocam at much lower resolution in the near ir at p.a .
@xmath13 ( @xcite ) , nor is it consistent with the orientation of a ring with a radius of @xmath1100 pc perpendicular to the radio jet , reported by @xcite . as such , we believe that we are observing a new feature , not directly related to the radio / x - ray jet or to any previously reported optical or ir features .
we have verified that the observed elongation is not the result of varying extinction in front of a smooth circular emission region .
indeed , if the extinction is maximum at the nucleus and decreases outwards in a direction perpendicular to the dust lane , the resulting morphology would be elongated in the direction of this extinction gradient .
we calculated the ir reddening correction map by adopting the procedure presented in @xcite , comparing the colors obtained in recent wfpc2 observations ( @xcite ) with our nicmos continuum images , using the reddening law of @xcite with @xmath14 , and assuming a constant intrinsic color over the field of view of nicmos ; this latter assumption is supported by an i - k vs v - i color diagram where most of the points lie tightly along the reddening line . applying the derived reddening correction to the image does not significantly affect the morphology of the observed structure .
hst observations of other agns suggest three possible explanations for the observed elongated structure : 1 ) gas which is shocked and compressed by interaction with the jet ; 2 ) illumination of gas clouds by an anisotropic nuclear radiation field ( i.e. an ionization cone , cf .
@xcite and references therein ) ; and 3 ) a gaseous disk , such as that of m87 ( @xcite ) .
the structure s position angle differs by @xmath15 from the jet axis ( see fig . [
fig : pacont ] ) .
this significant misalignment suggests the emission is not due to a current jet - cloud interaction , especially when we see other features well correlated with the radio / x - ray jet morphology ( see next section ) .
it has long been hypothesized that the radio jet had a different ( smaller ) position angle in the past , creating the n - s orientation of the outer lobes , and then rotating to form first the inner lobes and then the currently observed x - ray / radio jet . in this
regard we note that the outer edge of the n - e inner lobe does have structure aligned with the position angle of our feature ; both could be gas shocked by the jet ( perhaps two - sided ) at some recent time , and not yet cooled down .
existing radio and x - ray measurements do not provide added information on this small spatial scale . the @xmath15 misalignment with the jet
does allow the elongated structure to lie well within a putative ionization cone .
we note that most of the other detected emission features would also lie within such a cone if it has an opening half - angle of at least @xmath16 , well within the limits postulated by the unified model ( e.g. @xcite ) .
as it stands , we can not rule out clouds densely distributed along this position angle , embedded in a large opening angle radiation field .
however , the relative thickness of the feature in the transverse direction and the symmetry around the nucleus suggest that this is not the case .
we find the simplest interpretation for the elongated emission to be a gaseous disk around the nucleus , as seen on larger scales around other galactic nuclei .
it could be the outer part of an accretion disk , expected around a massive black - hole at the core of this agn .
if the structure is indeed a thin circular disk , the axial ratio suggests an inclination of @xmath160 .
its radius of @xmath17 ( @xmath18pc ) makes it significantly smaller than the hundred parsec scale stellar disks observed by hst at the centers of many galaxies .
this gas disk could be readily ionized by the powerful agn , seen in x- and @xmath19-rays .
the flux of photons required to keep the emitting material ionized can be computed following @xcite : after dereddening ( = 10 mag ) and assuming case b recombination for = k and = @xmath20 , we find a value of 4 photons s@xmath8 .
this is a small fraction of the total expected emission from the agn .
note that the nucleus has an extinction of up to 70mag along the line of sight , estimated from x - ray observations .
however , while being within the ionization cone the disk is well outside the obscuring torus , and subject to a a much lower extinction from the nucleus .
we thus assume = 10 mag as a reasonable upper estimate of the foreground extinction , following schreier et al .
( 1996 ) who found a peak extinction of = 7 mag in the optical .
the mass of the gas responsible for both extended and unresolved nuclear emission is estimated from the standard relation , @xmath214@xmath22 : @xmath23 where is the proton density , v is the volume of the emitting gas , @xmath24 is the line emissivity , @xmath25 is the proton mass , @xmath26 is the observed flux in units of erg @xmath7 s@xmath8 and @xmath27 is the reddening correction if is different from the assumed 10 mag .
this modest mass estimate of m @xmath114 depends mostly on the assumed gas density (= @xmath20 ) and is extremely uncertain , but we believe that a disk of to is quite feasible .
the disk is consistent with being perpendicular to the plane of the dust lane , and lying along the major axis of the large elliptical galaxy ( see fig . [
fig : pacont ] ) .
it is now widely accepted that the large gas / dust disk of cen a ( i.e. the dust lane ) was acquired in a recent merger process , and that differential precession has led to the observed warped structure ( e.g. @xcite ) . in the central region of cen
a , the precession time is only years , and the orthogonal alignment of our small disk may be consistent with numerical studies of the evolution of gas disks in bulge systems ( e.g. @xcite ) . if we thus interpret the observed emission as being from the warped outer portions of an accretion disk around the active nucleus , we conclude that even on the relatively small spatial scale of a few parsecs , the disk is dominated by the gravitational potential of the galaxy as a whole and not by the symmetry of the agn and its jet . for a non - rotating black hole ( bh )
, one would expect the jet to have its direction determined by the angular momentum axis of the inner gaseous disk , while for a kerr black hole , it would be expected to be aligned along the spin axis of the black hole itself ( @xcite ) .
close to a kerr bh , ( @xmath28gm / c@xmath29@xmath11 cm @xmath11 pc ) , the disk itself will warp to become normal to the spin axis . in our cen a data , on a few parsec scale , the normal to the disk and the radio jet are misaligned by @xmath11 70 , in projection .
we conclude that if cen a contains a rotating black hole , then either the gas disk must be outside the sphere of influence of the black hole , or it was formed recently enough that it has not yet become aligned with the spin axis ; for a non - rotating black hole , we find that the disk must become significantly warped away from being normal to the jet inside a radius of @xmath112pc .
@xcite has recently modeled self - induced warping of accretion disks in agns .
he finds that disks are likely to be warped at @xmath30pc around a @xmath11 bh .
our results show a constant position angle for the disk in to @xmath11 2pc , suggesting the somewhat weak upper limit of @xmath11 for the mass of the black hole .
three of the relatively strong features detected appear spatially associated with the x - ray / radio jet of centaurus a. in figure [ fig : radio ] we overlay the image with a 6 cm radio map ( feigelson , private communication ; see also figure 1 of @xcite ) .
we align the two images assuming that the radio and nir nuclei are coincident and note that the extended n - s extent of the radio contours is the result of the @xmath31 radio beam .
the kiloparsec scale x - ray / radio jet has an average pa = 55@xmath67 , the innermost knots have a slightly larger pa@xmath160 and the vlbi milliarcsecond jet and counterjet ( cf .
@xcite ) are at pa@xmath151@xmath63 .
we note that the multiple components of the radio jet within a few arcseconds of the nucleus were not resolved by @xcite and were reported as n1 .
they are clearly visible in the contour map ( fig .
[ fig : radio ] ) and figure 1 of clarke et al .
( 1986 ) , as part of their 6 nuclear jet , although the individual components were not discussed .
we label the compact , innermost knot n1a and see that it corresponds positionally with the `` finger''@xmath109 from the nucleus at pa @xmath32 ( see also the small inset in fig .
[ fig : pacont ] ) .
the two arcs a and b , at distances of @xmath129 and @xmath135 from the nucleus , each extending over @xmath33 pc , are seen to lie @xmath1 15 and 30 pc , respectively , on either side of another knot of the nuclear jet .
these may result from an outgoing shock created by interaction between the radio emitting nuclear ejecta and a gas cloud .
we note that the radio jet has a steep spectrum at this distance from the nucleus ( @xcite ) , suggesting that the gas is still too hot along the jet axis to show line emission .
we can expect , with adequate resolution , to see x - ray emission from this region of the jet .
we summarize the key results of our 2.2 @xmath5 m continuum and observations of the inner region of ngc 5128 as follows : \1 ) we see an extended continuum source with the rather regular and smooth structure expected for an elliptical galaxy .
this smoothness contrasts strongly with the filamentary structure observed in the wf / pc i - band image .
\2 ) we identify a strong unresolved ( @xmath3401 ) central source as the nucleus of the galaxy .
its position is consistent with ground - based ir and radio data and the peak of reddening and polarization found with wf / pc-1 .
the nucleus has an intensity of ( [email protected] ) erg @xmath7 s@xmath8 @xmath8 based on our continuum observations , and a flux of ( [email protected] ) erg @xmath7 s@xmath8 in .
\3 ) we see a prominent elongated structure in centered on the nucleus , with a major to minor axis ratio of @xmath2 , extending @xmath0along position angle @xmath1 33 , perpendicular to the dust lane .
we interpret this as an inclined @xmath3 parsec diameter thin nuclear disk of ionized gas .
the disk , with a radius of @xmath17 , corresponding to @xmath35pc at 3.5 mpc , is among the smallest ever observed at the nucleus of an agn in the optical / near - ir .
it is significantly smaller than the hundred or more parsec stellar disks presumed to exist at the centers of many galaxies .
even at this small scale , the disk is not perpendicular to the jet .
it is , however , consistent with being perpendicular to the dust lane and oriented along the major axis of the bulge .
if it represents the warped outer portion of an accretion disk around the active nucleus , we conclude that even a few parsecs from the nucleus , the disk is dominated by the galaxy gravitational potential and not by the symmetry of the agn and its jet .
further nicmos observations are planned in 2@xmath5 m polarized light and [ feii ] to provide emission mechanism diagnostics and determine the geometry of the nuclear radiation field .
if a warped disk illuminated by the nucleus is the correct model , systematic changes in the polarization angle as a function of disk position should be seen .
high spatial resolution infrared spectroscopy is planned to measure gas kinematics in the disk , and thus determine the mass of the black hole .
a.m. and n.c .
acknowledge support through go grants g005.44800 , g005.76700 , g005.70200 from space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas 526555 .
we thank howard bushouse , alex storrs , john mackenty , chris skinner , eddy bergeron and other stsci nicmos staff for invaluable assistance in understanding the nicmos instrument and its calibration .
we thank nino panagia , martino romaniello , ernesto oliva , george miley , and gary bower for useful comments and discussions .
we thank eric feigelson for providing an unpublished radio map and zolt levay for help in producing plates for this paper .
antonucci r.r.j . ,
1993 , , 31 , 473 bailey j. , sparks w.b . ,
hough j.h .
, axon d.j . , 1986 ,
nature , 322 , 150 bardeen j.m and petterson j.a .
, 1975 , , 195 , 65 blandford r.d . , 1991 , `` physics of agn '' , proceedings of heidelberg conference , springer - verlag , eds .
w.j . duschl and s.j .
wagner , p. 3
burns j.o . , feigelson e.d . ,
schreier e.j . , 1983 , , 273 , 128 bower g.a . , green r.f . ,
the space telescope imaging spectrograph investigation definition team , et al . , 1998 , , in press bushouse h. , skinner c.j . , mackenty j.w , 1997 , nicmos instrument science report , 97 - 28 ( baltimore stsci ) cardelli j.a .
, clayton g.c . and mathis j.s . , 1989 , , 345 , 245 clarke , d.a . , burns , j.o . , and feigelson , e.d . , 1986 , , 300 , l41 dobereiner s. , junkes n. , wagner s.j .
, zinnecker h. , fosbury r. , fabbiano g. , schreier e.j . , 1996 , , 470 , l15 dufour r.j . , van den bergh s. , harvel c.a . , martins d.h . , schiffer f.h . , talbot r.j . , talent d.l .
, wells d.c . , 1979 , , 84 , 284 feigelson e.d . ,
schreier e.j .
, delvaille j.p . , giacconi r. , grindlay j.e . , lightman a.p . , 1981 , , 251 , 31 ferrarese l. , ford h.c . , jaffe w. , 1996 , aj , 470 , 444 harms r.j . , ford h.c . , tsvetanov z.i . , hartig g.f . , dressel l.l . , kriss g.a . , bohlin r.c . , davidsen a.f . , margon b. , kochhar a.k . , 1994 , , 435 , l35 hui x. , ford h.c . , ciardullo r. , jacoby g.h . 1993 , , 414 , 463 krist j.e . and hook r. , 1997 , tinytim user guide , version 4.4 ( baltimore stsci ) jones d.l . , tingay s.j . , murphy d.w . ,
et al . , 1996 , , 466 , l63 mackenty j.w . , et al . , 1997 , nicmos instrument handbook , version 2.0 ( baltimore stsci ) macchetto f.d . , marconi a. , axon d.j . ,
capetti a. , sparks w.b . , crane p. , 1997 , 489 , 579 osterbrock d.e . , 1989 , astrophysics of gaseous nebulae and active galactic nuclei , university science books , p. 146
packham c. , hough j.h . , young s. , chrysostomou a. , bailey j.a .
, axon d.j . ,
ward m.j . , 1996 , , 278,406 pringle j.e . , 1997 , , 292 , 136 quillen a.c . ,
de zeeuw p.t . ,
phinney e.s .
phillips , 1992 , , 391 , 121 robinson a. , 1997 , asp conference series , eds .
peterson , f .- z .
cheng and a.s .
wilson , vol .
113 , p. 280
rydbeck g. , wiklind t. , cameron m. , wild w. , eckart a. , genzel r. , rothermel h. , 1993 , , 270 , l13 schreier e.j .
, feigelson e. , delvaille j. , giacconi r. , grindlay j. , schwartz d.a . , fabian a.c . , 1979 ,
, 234 , l39 schreier e. , burns j.o .
, feigelson e.d . , 1981 , , 251 , 523 schreier e.j .
, capetti a. , macchetto f. , sparks w.b .
, ford h.c . , 1996 , , 459 , 535 schreier e.j . ,
marconi , a. , capetti a. , caon , n. , axon , d. , macchetto f. , in preparation .
skinner c.j . ,
bergeron l.e .
, daou d. , 1997 , hst calibration workshop , eds . s. casertano et al .
( baltimore stsci ) , in press tubbs a.d . , 1980 ,
, 241 , 969 vigroux , iau symposium 186 , 1997 | we report hst nicmos observations of the central region of ngc 5128 at 2.2 and in .
the continuum images show extended emission typical of an elliptical galaxy and a strong unresolved central source we identify as the nucleus of the galaxy .
its position is consistent with ground - based ir and radio data , and with the peak of reddening found with wf / pc-1 . in
, we detect a prominent elongated structure , centered on the nucleus , extended by @xmath0at a position angle of @xmath133 , and with a major to minor axis ratio of @xmath2 .
we interpret this as an inclined , @xmath3 parsec diameter , thin nuclear disk of ionized gas rather than a jet - gas cloud interaction .
we do see several weaker features , some of which may be circumnuclear gas clouds shocked by the x - ray / radio jet .
the disk is one of the smallest ever observed at the nucleus of an agn .
it is not perpendicular to the jet , but consistent with being oriented along the major axis of the bulge .
if it represents the warped outer portion of an accretion disk around a black hole , we conclude that even on the scale of a few parsecs , the disk is dominated by the galaxy gravitational potential and not directly related to the symmetry axis of the agn . |
bulk metallic glasses ( bmgs)@xcite are a new class of advanced materials , which have attractive mechanical properties such as high strengths and large elastic limits arising from their long - range disorder atomic structure .
nevertheless , their applicability as structural materials is largely suppressed by the poor ductility@xcite , which has been regarded as the achill s heel of bmgs . at room temperature ,
plasticity in bmgs is an inhomogeneous process with plastic strain highly localized into shear bands@xcite .
once initiated , shear bands are prone to becoming unstable due to strain softening , causing the catastrophic failure of a bmg sample along the main shear plane@xcite .
consequently , this results in almost zero tensile ductility or very limited compressive plasticity at the macroscopic scale@xcite . in the past decades ,
tremendous efforts@xcitehave been made to overcome the intrinsic weakness of bmg , among which one effective approach is to fabricate bmg matrix composites@xcite .
this often involves ex - situ or in - situ introduction of secondary crystalline phases into a glassy matrix .
based on this approach , significant tensile ductility was achieved in some zr-@xcite and ti - based bmg@xcite composites .
however , in the early development , the bmg composites still displayed macroscopic strain softening under tension@xcite , which led to necking instability upon yielding , a phenomenon that is not desirable in engineering applications .
recently , cuzr - based bmg composites@xcite have attracted a great deal of research interest because of their unique mechanical properties . unlike other types of bmg composites ,
these cuzr - based bmg composites comprise shape - memory b2 cuzr crystalline phases , which could undergo martensitic transformation ( mt ) during deformation and turn into the b19 phase . by a careful control of the size and distribution of the b2 phases ,
it has been shown@xcite that these composites can exhibit significant tensile ductility and also display pronounced work - hardening , an indispensable ingredient for the practical application of structural materials .
apparently , these superior mechanical properties are derived from the interaction between the mt of b2 phases and the plastic flow
carriers ,
i.e. shear bands , in the glassy matrix .
furthermore , it was proposed that the transformation of b2 might exert a compressive stress on the glassy matrix and thus retards the rapid propagation of shear bands and/or cracks@xcite . to validate this proposition , finite element ( fe ) simulations
were utilized to obtain the stress / strain fields in the deformed bmg composites using the commercial packages@xcite ; however , these preliminary analyses are generally insufficient because the transient phase transformation behavior of the b2 phase is not considered in the constitutive model of the commercial fe software@xcite .
therefore , to understand the mechanisms underlying the tensile plasticity in the cuzr - based bmg composites , a precise knowledge of the stress redistribution around the transforming b2 phases is still lacking . in this study , we develop a micromechanical model based on the eshelby theory@xcite to understand the plasticity associated with the mt of b2 in the cuzr bmg composites .
it will be shown that the model predictions , such as the stress state and the resultant distribution of shear bands and strain hardening coefficient , compare very well with the experimental observations .
our current study provides a quantitative understanding of the mechanisms underlying the strain hardened plasticity in the cuzr - based bmg composites , which should be also helpful for a better control of the microstructure in bmg based composites with improved mechanical properties .
master alloy ingot with a nominal composition cu@xmath0zr@xmath0al@xmath1 were prepared by arc melting the mixture of the constitute elements with purity @xmath2 in a ti - getted atmosphere .
each alloys ingots were remelted at least three times for the compositional homogeneity .
plate samples with a thickness of 2 mm , a width of 10 mm and a length about 80 mm were obtained by suction casting into the copper mould . as the morphology of b2 phase is very sensitive to experimental condition in casting process such as the cooling rate and the quench temperature , we control the melting temperature by applying the arc current of 320 a on each alloy ingot for about 20 second before quenching to ensure the microstructure reproducibility of the bmg composite . the specimens with a dimension of about @xmath3 mm were cut from the lowest part of the plate ( the part with the highest cooling rate ) with a diamond saw , and then carefully ground into samples with an aspect ratio of 2:1 for the compression tests .
the four side faces of samples are also carefully polished for the structural characterization after deformation .
the uniaxial compression tests were performed with an instron 5869 electromechanical test system with the maximum load of 50 kn at a constant strain rate of @xmath4 .
the strain were measured by a laser extensometer ( fiedler ) attached to the testing machine .
each compression test was repeated at least three times to ensure the reproducibility of deformation results .
after deformation , the morphology of the shear bands as well as the fracture surface are investigated by the scanning electron microscopy ( sem , gemini 1530).the structure nature of as - cast as well as post - deformed samples is examined by the x - ray diffraction ( xrd , panalytical xpert pro ) with co - ka radiation ( the wave length @xmath5 nm ) and an zeiss axiophot optical microscopy ( om ) . at first , we need to choose a cuzr - based bmg as the base material .
according to the previous study@xcite , the glass - forming ability ( gfa ) and plasticity of the cu - zr - al ternary bmgs strongly depends on the content of al . on one hand ,
it was reported that , when the content of al is around 4@xmath65% ( in atomic percent ) , the cu - zr - al bmgs could possess a relatively large value of poisson s ratio and show a plastic strain of 18% before failure@xcite . on the other hand
, the cu@xmath0zr@xmath0al@xmath1 bmg can be cast into a fully amorphous 2-mm - diameter rod or a 1-mm - thickness plate@xcite , indicative of a very good gfa better than that of the typical binary cu@xmath7zr@xmath7 bmg .
it was noticed that further addition of al into the cu@xmath0zr@xmath0al@xmath1 bmg would enhance the gfa but greatly deteriorate the plasticity .
therefore , we finally chose the cu@xmath0zr@xmath0al@xmath1 bmg as the matrix material and cast it into a 2-mm - thick plate to obtain the bmg composite .
[ fig1 ] shows the typical xrd pattern of the as - cast sample taken from the lowest part of the plate ( the part with highest cooling rate ) , from which one can clearly see the crystalline peaks superimposed on the broad halo of the amorphous matrix .
the majority of the crystalline peaks can be easily identified to be the b2 cuzr phase while the ( 001 ) peak of the martensitic phase b19 could be also found , suggesting a small volume fraction of the b19 phase in the as - cast sample . in ref.34 ,
a similar phenomenon was observed and the small amount of the b19 phase was attributed to the residual - stress induced martensitic transformation of the b2 phase during solidification@xcite . after the deformed to 8% , obvious crystalline peaks of the martensite phase ( b19 ) appeared in the xrd pattern ( see ) , indicating that some b2 phases underwent a martensite phase transition during the deformation . the xrd curves for the as - cast and deformed cu@xmath0zr@xmath0al@xmath1 ( up to @xmath8 ) metallic glass composites .
the b2 phase and martensite phase are identified.,width=377 ] as aforementioned , the size , morphology and distribution of the b2 cuzr phase in the glassy matrix are very sensitive to the alloy composition and experimental parameters , such as the cooling rate and the melting temperature employed for the casting process@xcite .
furthermore , recent studies@xcite also showed that a homogeneous distribution of b2 particles could be achieved by the minor addition of some alloying elements , such as co and ta , into the glassy matrix .
[ fig2 ] displays the morphology and distribution of the b2 cuzr phase on the cross - sections of our as - cast samples . as seen in this figure
, there is a homogeneous distribution of the spherical b2 phases in the glassy matrix , which is different from those typically reported cu - zr - al composites , where the heterogeneous distribution of the b2 phases and even the dendritic structure can be found@xcite . as shown in [ fig2 ] , the diameters of these spherical particles are in the range of 25 - 100 @xmath9 m with an estimated average of @xmath642 @xmath9 m .
here , it is worth noting that the spherical b2 phases are distributed in the 3d glassy matrix and therefore their volume fraction can be not simply estimated from the 2d cross sectional view , as shown in [ fig2 ] .
however , by assuming a thin foil projection@xcite , the volume fraction of the b2 phases can be roughly estimated as : @xmath10 + where @xmath11 is the mean diameter of b2 spheres ( here , @xmath12
@xmath9 m ) , @xmath13 is the thickness of the foil and is roughly equal to@xmath11 , and @xmath14 is the project area fraction from the cross section ( for example , @xmath15 for [ fig2 ] ) .
the volume fraction of b2 cuzr phase calculated by [ eq1 ] is about 5 - 8% for all the samples obtained via the copper - mold casting after an appropriate selection of the casting parameters .
the microstructure of as - cast cu@xmath0zr@xmath0al@xmath1 metallic glass composite plate under optical microscope .
one can see that spherical b2 phases are homogeneously distributed in the glassy matrix.,width=377 ] subsequently we performed uniaxial compression tests on the cu@xmath0zr@xmath0al@xmath1 composite samples . to investigate shear band evolution ,
the compression tests were deliberately stopped at the different strains ( 4% , 8% ) for an sem observation .
[ fig3 ] shows the typical stress - strain curves obtained from these composites . by comparison ,
these curves display the similar yield strength @xmath61.5 gpa or the elastic limit of @xmath62% with those of the corresponding monolithic bmgs ; however , the bmg composite exhibits much better ductility with a plastic strain at least 10% before the final failure occurs .
more importantly , work hardening can be observed during the deformation of the present bmg composite although it may not be as pronounced as that reported in the previous studies[20 , 35 ] , which may be due to the small volume fraction of the b2 phases ( 5 - 8% ) in the current composite . finally , the bmg composite sample failed by the shear fracture along two shear planes initiated from two adjacent sample sides [ ] , rather than alone one single primary fracture plane as seen for most monolithic bmgs .
in addition , a close examination of the stress - strain curves ( see the enlarged view in the rectangular region in fig.3 ) shows that the deviation of the stress - strain curves from linear elasticity occurs at @xmath61 gpa before the nominal yielding point sets out at @xmath61.5 gpa .
this is similar to the first - yielding - point phenomenon as seen in the triple - yielding - behavior of the cuzr - based bmg composites that contain a large volume fraction of b2 phases@xcite .
this makes sense since the activation stress for mt of b2 is known to be lower than the typical yielding strength of a glassy matrix@xcite .
therefore , the initial yielding point of @xmath61 gpa ( it makes sense to see that the yield stress of b2 is much less than bmg materials ) as we herein observed strongly indicates that the mt of b2 occurs prior to the overall yielding of the composite , which causes stress redistribution in the glassy matrix and thus affects the subsequent shear banding , as will be discussed in the later text .
the stress - strain curves of cu@xmath0zr@xmath0al@xmath1 metallic glass composites deformed to different strain levels ( 4% , 8% and final fracture ) in compression .
the inset shows that the stress begins to deviate the linear elasticity at about 1 gpa before the nominal yielding point.,width=453 ] after the deformation , the surface shear - band morphology was systematically investigated under sem . as shown in [ fig4 ] , multiple shear bands
can be clearly observed . especially , many fine and wavy shear bands can be found around the crystalline phases .
these shear bands are sparsely distributed at the relatively low strain@xmath64% and gradually become densely populated across the whole sample with the increasing strain , as shown in ( b ) and ( c ) , respectively .
a careful examination of these sem images reveals that these complex shear bands can be classified into three types .
as indicated by the black arrows in , the type - i shear bands are within the glassy matrix and away from the crystalline particles . like the shear bands in monolithic bmgs@xcite , the type - i shear bands are oriented along the inclined direction of about 45@xmath16 to the loading direction , conforming to the direction of the maximum shear stress . here
it is worth mentioning that the type - i shear bands could be deflected and change their direction of propagation when they approach the crystalline particles , as shown in ( b ) and ( c ) . by comparison , the type - ii shear
bands seemingly emanate from the crystalline phases and are well aligned along the direction of the martensitic transformation
stripes . compared to the type - i shear bands , these type - ii shear bands are short and finely spaced , which disappear within a short distance from the glass / crystalline interface , as indicated by the red arrows in . since the b2 particles embedded in the glassy matrix have different crystal orientations , the direction of the martensitic transformation stripes also varies from one particle to another . as a result
, there is no fixed direction for the type - ii shear bands .
finally , it is worth pointing out that the type - iii shear bands , are also attached with the crystalline phases and are oriented along the fixed direction perpendicular to the loading axis , as indicated by blue arrows . according to our sem observations ,
the three types of shear bands mutually interact and intersect with each other , which contributes to the overall plasticity in the bmg composite and also provides the vivid experimental evidence to decode the underlying deformation mechanisms .
the sem images shows the overall morphology of cu@xmath0zr@xmath0al@xmath1 samples after the deformation : ( a ) 4% ; ( b ) 8% and ( c ) the final fracture.,width=377 ] the shear - band distributions on the surface morphology of the deformed cu@xmath0zr@xmath0al@xmath1 samples .
three types of shear band are identified and labeled by black , red and blue arrows , respectively .
, width=566 ]
theoretically , the martensitic transformation of the embedded b2 phase results in a stress and strain redistribution in the glassy matrix , which may cause the deflection of a propagating shear band , such as the type - i shear band ( ) , and/or even affect the initiation of an embryonic shear band , such as the type - ii and the type - iii shear band ( fig .
5a ) , during the transience of a phase transformation , however , this transient effect is difficult to simulate in the conventional fem@xcite only dealing with quasi - static equilibrium . unlike the previous studies@xcite ,
here we derive the analytical expressions for the stress / strain field arising from the martensitic transformation of the b2 phases by using the eshelby theory@xcite . before proceeding ,
it is worth noting that , in eshelby theory@xcite , a region is termed as an _ inclusion _ if it has the same elastic constants as the matrix material while a region is termed as an _ inhomogeneity _ if it does not . for consistency ,
the same terminologies are also adopted in the following analysis . in principle , a typical eshelby problem , which entails the inelastic deformation of inclusion / inhomogeneity in an elastic matrix , can be solved following three imaginary steps@xcite .
as shown schematically in , at step i , one takes an inclusion out of the matrix and let it undergo an unconstrained transformation as characterized by the transformation strain field @xmath17 ; at step ii , elastic surface tractions @xmath18 are imposed onto the surface of the transformed inclusion such that it can fit back into the original matrix ; at step iii , the surface tractions are removed while the inelastic deformation remains , which leads to the stress / strain redistribution in the inclusion and elastic matrix . since previous experimental studies@xcite already showed that the elastic constants of the b2 phase are very close to those of the glassy matrix , the b2 phase can be here treated as an eshelby inclusion , which is inserted in the homogeneous glassy matrix with the shear modulus @xmath19 and poisson s ratio @xmath20 . following the eshelby s approach as described above , we obtain the following expression for the elastic strain @xmath21 in the matrix which is caused by the inelastic strain , occurring to a spherical inclusion with a radius @xmath22 , during the martensitic transformation ( see the detailed derivation in [ appendixa ] ) , which reads : @xmath23+\\ \\ & \frac{\sigma_{kl}^{t}}{6g(1-\nu)}\left(\frac{a^{3}}{r^{3}}\right ) \left[(1 - 2\nu)+\frac{3}{5 } \left(\frac{a^{2}}{r^{2}}\right)\right ] \left[\frac{2r^{2}\delta_{il}\delta_{jk}-3(x_{k}x_{j}\delta_{il}+x_{k}x_{i}\delta_{jl})}{2r^{2}}\right]+\\ \\ & \frac{\sigma_{kl}^{t}}{4g(1-\nu)}\left(1-\frac{a^{2}}{r^{2}}\right)\left(\frac{a^{3}}{r^{3}}\right ) \left[\frac{r^{2}(x_{k}x_{l}\delta_{ij}+x_{i}x_{k}\delta_{jl}+x_{j}x_{k}\delta_{il})-10x_{i}x_{j}x_{k}x_{l}}{r^{4}}\right ] \end{aligned}\ ] ] + where @xmath24 is the magnitude of the position vector for the centroid of the inclusion ; @xmath25 is in the cartesian coordinate ( @xmath26 ) , and @xmath27 for @xmath28 , and @xmath29 otherwise .
meanwhile , the surface traction @xmath18 : @xmath30 , where @xmath31 and @xmath9 are lam constants of the inclusion . in the inclusion , the elastic strain @xmath32 is homogeneous and related with the inelastic strain @xmath17 through @xmath33:@xmath34 where @xmath35@xcite . as a result , the additional elastic stress ( @xmath36 ) in the matrix as caused by the martensitic transformation can be expressed as @xmath37 . while that ( @xmath38 ) in the inclusion as @xmath39 , since there is already a stress @xmath40 in the inclusion at the end of step ii
. finally , the total stress in the matrix ( or inclusion ) is simply the sum of @xmath36(or @xmath38 ) and the external applied stress @xmath41 . here
, it should be noted that both @xmath36 and @xmath38 are short - lived or instantaneous , which appear during the transience of the martenstic transformation and vanish with the phase transformation being completed . the schematic illustration of the three stages of eshelby approach for solving the stress and strain fields due to the deformation of an inclusion in the matrix.,width=453 ] after the b2 phase is transformed into the martensitic phase , elastic mismatch arises because the newly formed martensitic phase possesses elastic constants different from the glassy matrix according to the previous experiments@xcite .
consequently , this gives rise to an elastic stress field , which may be termed as elastic constraining stress , being superimposed onto the applied stress . according to the eshelby theory@xcite , the effect of the elastic mismatch on the stress / strain distribution within the composite with an inhomogeneity is equivalent to applying a effective transformation strain @xmath17 to an inclusion with the size of the inhomogeneity .
assume that the bulk ( or shear ) modulus of the martensitic phase is @xmath42 ( or @xmath43 ) while that of the glassy matrix is @xmath44 ( or @xmath19 ) , it can be derived that the effective transformation strain[@xmath45@xmath46 with @xmath47 can be expressed as : @xmath48 where @xmath14 and @xmath49 are the two coefficients which are functions of the elastic properties of the imhomogeneity and matrix : @xmath50 + with @xmath51 and @xmath52 . here
the applied far - field strain @xmath41 can be expressed in terms of the volumetric ( @xmath53 ) and deviatoric ( @xmath54 ) component , in which @xmath55 , @xmath56@xmath57 , according to the hook s law . with the knowledge of the @xmath18 from [ eq2 ] and [ eq3 ] , the elastic stress field due to the inhomogeneity
can be readily obtained with the same approach as described for an eshelby inclusion .
likewise , the total stress in the matrix ( or the inhomogeneity ) is the sum of @xmath36 ( or @xmath38 ) and the applied stress @xmath41 .
based on the above analyses , we obtained the analytical expressions for the stress / strain fields in the amorphous matrix composite during the transience of the martensitic transformation and after the transformation .
next , we would study the shear band formation in the bmg composite based on these important results . for simplicity , let us neglect the subtle pressure effect@xcite , although there is no technical difficulty to incorporate the pressure effect into the current analysis , and assume that shear band formation in the glassy matrix simply follows von mises stress criterion . for the stress field @xmath58 ( @xmath59@xmath60 ) , the von mises stress is calculated according to : @xmath61 with @xmath62@xmath60 . the schematic diagram illustrating the coordinate systems for the calculation of stress and strain fields and the principle axis directions of the eigenstrain due to the mt of b2 phase.,width=566 ] first ,
let us calculate the instantaneous stress field during the martensitic transformation of the b2 phases .
since the martensitic transformation is essentially shearing induced , the transformation strain ( or eigenstrain ) @xmath17 is traceless or does not involve any volumetric change ( @xmath63 ) . consequently , the traceless strain tensor @xmath17 can be written as @xmath64 where @xmath65 , @xmath66 , @xmath67 are the eigenvalues of the deformation tensor or the three principal strains with the principal axes aligning along the unit vectors in a cartesian coordinate system@xcite . here , we choose the coordinates so that the principal axis with the unit vector @xmath68 coincides with the @xmath69 axis ( see ) . as such
, @xmath17 can be simplified into the form of plane strain : @xmath70 , where @xmath71 is the magnitude of the transformation strain .
note that all components of the strain tensor @xmath17 are zero except @xmath72 and @xmath73 .
furthermore , we can take @xmath74 and @xmath75 with @xmath76 being the angle between the shear direction and the @xmath77 axis , as shown in . in such a case ,
the direction of the martensitic transformation is symbolized by the angle @xmath76 , which can be also used to track the direction of shear banding that is directly associated with the martensitic transformation .
once @xmath17 is known , we can obtain @xmath18 according to @xmath78 .
afterwards , we can obtain the explicit expression for @xmath32 in the @xmath79 plane by substituting @xmath18 into [ eq2 ] ( see [ b2 ] in the [ appendixb ] ) .
for the calculation of the stress / strain field , typical values are chosen as : @xmath80 , @xmath81 gpa , @xmath82@xcite and the far - field uniaxial compressive stress @xmath83 is oriented along the @xmath84 axis .
since the martensitic transformation takes place before the overall yielding occurs in the glassy matrix , here we take @xmath85 gpa according to our experimental results ( ) . the von mises stress ( @xmath86 ) distribution due to the mt in the vicinity region of spherical phase of @xmath79 plane .
two mappings with @xmath87 ( a ) and@xmath88(b ) are showed here to track the variation of @xmath86 with the mt shear direction .
the arrows indicated the regions with maximum von mises stress.,width=377 ] ( a)-(b ) display the contour plots of the von mises stress @xmath86 in the vicinity of a transforming spherical b2 phase for two shear angles ( a ) @xmath87 and ( b ) @xmath88 . from these figures ,
one can see clearly that @xmath86 is constant inside the crystalline phase but exhibits strong position dependence in the glassy matrix , as consistent with the eshelby theory@xcite . as can be seen
, @xmath86 is small ( generally less than 1 gpa ) in the regions straying away from the direction of the martensitic transformation , such as those colored in blue , while becomes intensified in the regions aligning with the direction of the martensitic transformation , such as those colored in red .
this is particularly so for the high transformation angle @xmath88 , at which the maximum value of @xmath86 could exceed the macroscopic yield stress of the glassy matrix ( @xmath61.5 gpa ) in the directions close to that of phase transformation ( as indicated by the arrows in ( b ) ) .
this indicates that these regions are the preferable sites for shear - band initiation in the glassy matrix during the martensitic transformation .
since the maximum @xmath86 always appear in the direction aligning with the transformation angle @xmath76 , the formation of the type ii shear bands , as observed in our experiments ( ) , can be rationalized as a direct result of the stress redistribution in the glassy matrix during the martensitic transformation of the b2 phases .
next , we would like to calculate the stress / strain distribution in the bmg composite after the martensitic transformation of the b2 phases . for a uniaxial compressive stress ( @xmath89 ) along the y axis , the effective transformation strain or eigenstrain @xmath17 can be written as : @xmath90 where @xmath14 and @xmath49 are the two coefficients associated with the elastic mismatch between the martensitic phase and the glassy matrix .
note that the uniaxial loading along the @xmath84 axis corresponds to @xmath91 and @xmath92 ; the first term in the right hand side of [ eq5 ] is for pure dilation while the second term for pure shear .
likewise , for the pure shear term in [ eq5 ] , the resultant elastic stress / strain field can be derived following a very similar procedure as that used for the transience of the martensitic transformation of the b2 phases ; while for the pure dilation term @xmath93 , where @xmath94 is the magnitude of the dilatational strain , an explicit expression for the corresponding stress / strain field could be also obtained ( see [ c1 ] ) .
afterwards , based on the principle of superposition in linear elasticity , one can obtain the elastic strain / stress field and thus the von mises stress distribution in the presence of a martensitic phase in the glassy matrix . at the present time
, there are no available data reported for the elastic modulus of the cuzr martensitic phase . however , according to the measured hardness data@xcite , the elastic modulus of the martensitic phases should be larger than that of the b2 phases . as a result , this yields the negative values of @xmath14 and @xmath49 , which means that a compressive external strain @xmath95 always corresponds to a tensile effective eigenstrain ( @xmath96 ) , and _ vice versa_.
as will be shown in the later text , this behavior has an important effect on the stability of the particle - glass interface and thus the ductility of the cuzr bmg composite under different loading condition ( compression or tension ) . according to [ eq4 ] the values of @xmath14 and @xmath49
are a function of the elastic constant difference @xmath77 [ @xmath97 , @xmath98 denotes @xmath44 or @xmath19 ] . here
, we take the typical values of @xmath14 and @xmath49 at @xmath100 and 0.50 and calculate the corresponding von mises stress distribution in the @xmath79 plane , as shown in ( a ) and ( b ) respectively . in this case , the maximum values of @xmath86 appear close to the martensitic phase and the direction of the maximum von mises stress is perpendicular to the loading axis , as indicated by the arrows in ( a ) and ( b ) .
in addition , the variation of @xmath77 only changes the intensity of @xmath86 , but the distribution of @xmath86 remains the same .
this behavior indicates that , as the external load continues increasing after the martensitic transformation , the elastic mismatch between the martensitic phase and the glassy matrix will cause shear - band initiation in the direction vertical to the loading axis , which is exactly what we observed in the experiments , that is , the type iii shear bands as seen in .
the von mises stress ( @xmath86 ) distribution due to the constraint of glassy matrix on deformation of martensite phases in the vicinity region of a spherical phase of @xmath79 plane .
two mappings with @xmath101(a ) and @xmath102(b ) are showed here .
the arrows indicated the regions with maximum von mises stress.,width=377 ] compared to monolithic bmgs , one of prominent deformation features of cu - zr based bmg composites is their strain or work hardening ability@xcite . according to the prior works@xcite , the observed work hardening in the cuzr
based bmg composites can not be solely attributed to the intrinsic strain hardening of the b2 phases , but also arises from a constraining effect of the glassy matrix on the martensitic transformation and the subsequent deformation of the transformed phases . in theory , this constraint effect , also called eshelby back stress effect , increases the elastic energy stored in the whole composite system , which leads to an increase in the applied stress and thus manifests as strain hardening in the bmg composite .
now , with the obtained expressions for the various stress / strain fields , the contribution to the strain hardening from this constraint effect can be quantitatively analyzed .
according to eshelby s theory@xcite , the total increase in the elastic energy storage @xmath103 in the bmg composite during the martensitic transformation , i.e. for a spherical inclusion undergoing a pure shear @xmath104(@xmath105 and @xmath106 ) , can be written as : @xmath107 where @xmath108 is the volume of the inclusion .
note that the first term at the right - hand side of [ eq6 ] is the strain energy stored due to the martensitic transformation of the inclusion and the second term is the strain energy stored due to the interaction of the elastic field of the transforming inclusion and the external stress field , i.e. the uniaxial far - field stress @xmath83 along the @xmath84 axis , which can be generally expressed as @xmath109@xcite . since the only non - zero component of @xmath41 is @xmath110(@xmath111 ) , @xmath112 with @xmath113 , where @xmath114 is a parameter that is a function of the orientation of the b2 phase .
it should be noted here that the elastic energy @xmath103 is instantaneous and vanishes after the phase transformation . on the other hand , after the phase transformation , the strain energy stored @xmath115 due to the interaction of the elastic field of the inhomogeneity with the external uniaxial stress field is@xcite : @xmath116 note that the negative sign in the above expression makes @xmath115 positive because the coefficient @xmath14 and @xmath49 are both negative .
assume that the number of the transformed b2 phases is @xmath117 , where @xmath118 is the total number of the b2 phase per volume in the undeformed sample and @xmath119 is the accumulated percentage of transformation .
hence , the number of the b2 phases undergoing the martensitic transformation can be written as @xmath120 , where @xmath121 is the instantaneous probability for a b2 phase to undergo the martensitic transformation . as a result ,
the total elastic energy stored is : @xmath122 where @xmath123 is the elastic energy @xmath103 averaged over all possible orientations of the b2 phases .
the accumulated fraction of the transformed b2 phases , @xmath124 , is assumed to be a single - valued function of the external strain @xmath125 . as a result of the additional elastic energy storage ( [ eq8 ] )
, the applied stress has to increase , resulting in strain hardening .
the strain hardening coefficient @xmath126 is defined as @xmath127 . from [ eq8 ]
, we could obtain @xmath128 . from the simple linear - growth model , one infer that the transformation rate @xmath129 , i.e. , the transformation rate of the b2 phases be proportional to the volume fraction @xmath130 of un - transformed b2 phases and @xmath131 denotes a reference strain .
solving the above equation gives @xmath132 .
combining this equation with the expression for @xmath126 , we finally obtain @xmath133with @xmath134 . to verify our modeling , we collected the strain hardening coefficients as a function of the plastic strain from the stress - strain curves , as reported in the literature@xcite , for the cuzr - based bmg composites with different volume fractions of b2 phases .
as show in , it is evident that our theoretical modeling is in excellent agreement with the experimental data , which indicates that the strain hardening ability of the cu - zr based metallic glass composites can be attributed to the eshelby back stress effect , as proposed in the previous works@xcite .
since the volume fraction of the b2 phases is proportional to @xmath118 , we can thus infer , the strain hardening coefficient should increase with the volume fraction of the b2 phases in a linear way , which is consistent with the previous studies@xcite .
in addition , the dependence of the strain hardening coefficient @xmath135 on the term @xmath134 implies that one has to ensure @xmath136 in order to retain a positive strain hardening coefficient .
in other words , in terms of achieving the strain hardening ability and tensile ductility in the bmg composites , it is beneficial to reduce the extent of the elastic mismatch ( as for a low @xmath115 ) and/or to design the orientation of the b2 phases for a high and efficient instantaneous transformation probability @xmath121 . the variation of the work - hardening coefficient with the applied strain for different volume fraction of b2 phases collected from literatures .
these data points are well fitted by the exponential decaying function @xmath133 .
the values of fitting parameters @xmath137 mpa , @xmath138 for @xmath139% ; @xmath140 mpa , @xmath141 for @xmath142% and @xmath143 mpa , @xmath144 for @xmath145%,width=453 ] based on the experimental results ( ) and theoretical analyses , it is clear that there is a stress - affected zone ( saz ) for every b2 phase embedded in the glassy matrix , within which the martensitic transformation can lead to the formation of type ii and iii shear bands . from an experimental viewpoint , the extension of these shear bands could be used to estimate the saz size .
theoretically , we can assume the saz to be a cubic region for simplicity and the side length b of the saz can be estimated from the decay of the constraining stress with the distance @xmath24 from the center of the b2 phase .
( a ) and ( b ) show the variation of the von - mises stress with @xmath24 along the propagation direction of type ii ( @xmath88 ) and iii shear bands , respectively . as seen in these figures ,
the von mises stress varies significantly in the proximity of the crystalline phase but approaches to the far field stress far away . for type ii shear bands
, @xmath86 reaches the external stress level when the distance @xmath146 , which corresponds to @xmath147 ; while for type iii shear bands , this distance can be set as @xmath148 , which also corresponds to @xmath149 .
therefore , we can safely choose @xmath150 as a theoretical estimate of the saz size , beyond which the back stress effect caused by the martensitic transformation can be neglected . schematic diagram illustrating the stress affected zone ( saz ) due to the stress redistribution of the mt of b2 phases and the subsequent deformation of martensite phases and their deflection on the type i shear band during the deformation of cuzr based metallic glass composites.,width=453 ]
the distribution of the sazs in the glassy matrix could have a significant effect on the deformation stability and ductility of the cu - zr based metallic - glass composites . for a composite with a dilute b2 concentration ,
the final fracture of the composite can take place along the path of the type i shear band as caused by the external stress .
the stability of the type i shear bands determines the overall ductility of the metallic - glass composite , as in monolithic metallic glasses@xcite .
however , a large population of sazs could deflect the propagation of the type i shear band , thus delaying the occurrence of shear - banding instability as illustrated in ( b ) and ( c ) . since the type i shear band propagates through the regions between the sazs , the smaller is the spacing @xmath151 between two adjacent sazs the more effectively is sazs in shear - band deflection . for a homogeneous distribution of the crystalline phases ,
@xmath151 can be estimated as @xmath152{v_{f}/(4\pi a^{3}/3)}-b$ ] by taking the composite sample to be a cubic shape with a unit length , where @xmath153 is the volume fraction of the crystalline phases . from this expression
, one can see that @xmath151 decreases with the increasing @xmath153 , which indicates that a higher volume fraction of crystalline phases can lead to a better ductility of the metallic - glass composite , as consistent with the reported experimental results@xcite .
however , as @xmath153 increases and the sazs begin to overlap , @xmath151 will finally become zero when @xmath153 reaches a critical value . when this happens , the deflection of the type - i shear bands is no longer the controlling mechanism for the deformation stability of the composite .
other mechanisms , such as interface debonding , may come into effect .
taking @xmath150 into the expression for @xmath151 , one obtains the critical volume fraction @xmath154 .
interestingly , this value of @xmath155 is very close to the percolation threshold ( @xmath630% ) as experimentally identified for the crystalline phases in the cu - zr metallic - glass composites@xcite . on the other hand , for a give @xmath153
, @xmath151 decreases with the decreasing radius of the b2 phases .
this implies a size effect of the crystalline phase on the ductility of the metallic - glass composites .
given a fixed volume fraction , the smaller is the size of the individual crystalline phase the denser is its spatial distribution . in turn
, this leads to more effective deflection of the type - i shear bands and thus better ductility .
however , to enhance the ductility of the metallic - glass composite , it can be envisioned that the size of the crystalline phases can not be reduced infinitely because of the finite thickness of a shear band , which ranges from 10 nm to 100 nm[5 ] . when the size of a crystalline phase is comparable to the thickness of a shear band , it becomes less effective to deflect the propagation of the type - i shear band .
consequently , one may foresee an optimum size for the crystalline phase to enhance the overall ductility of the metallic - glass composites , the details of which deserve further theoretical and experimental studies .
finally , we would like to consider the issue of the stability of the interface between the crystalline phase and the glassy matrix .
the debonding of the interface has been observed during the final deformation stage of a cu - zr metallic glass composite under tensile loading@xcite . in principle , the crystalline - amorphous interface is considered stable under a compressive normal stress or a negative hydrostatic pressure ( @xmath156 ) but may debond under a tensile normal stress or a positive hydrostatic pressure ( @xmath157 ) .
now let us calculate the hydrostatic pressure on the crystalline - amorphous interface caused by the martensitic transformation . according to the eshelby theory@xcite
, the mechanical strain across the inclusion - matrix interface satisfies the following relation : @xmath158 where @xmath159 and @xmath160 denote the values of @xmath161 in the matrix and inclusion , respectively , as across the interface ; @xmath162 is the component of the unit vector @xmath163 with @xmath164 being the vector ending at the matrix - inclusion interface . for the martensitic transformation of the b2 phase,@xmath165 , @xmath166 , @xmath167 and thus @xmath168 .
therefore , the interfacial hydrostatic pressure can be expressed as @xmath169 , which varies along the periphery of the crystalline - amorphous interface .
based on the analysis in the [ appendixc ] , the maximum tensile pressure ( @xmath170 ) is attained at the interfacial point with the vector @xmath164 coincident with the principal axis @xmath68 of @xmath171 , which reads @xmath172 , where @xmath173 for @xmath174 . while the maximum compressive pressure ( @xmath175 ) is attained at the interfacial point with the vector @xmath164 perpendicular to the principal axis @xmath68 of @xmath171 , which reads @xmath176 where @xmath177 . with these equations , we can roughly estimate the interfacial bonding strength between the crystalline phases and the amorphous matrix .
this can be done by equating the maximum tensile hydrostatic pressure @xmath178 to the interfacial bonding strength @xmath179 . by taking @xmath180 , @xmath181 , @xmath182 gpa and @xmath183 , the estimated interfacial bonding strength @xmath184 mpa .
this is slightly lower than the yield strength of the metallic - glass composite . in principle ,
based on the theoretical framework , we can extract the interfacial bonding strength if the value of @xmath71 is exactly known or vice versa . following a similar procedure , the hydrostatic pressure acting on the crystalline - amorphous interface after the martensitic transformation can be derived as @xmath185 , where the principal axis of the effective transformation strain coincides with the external @xmath186 axis , such that @xmath91 , @xmath92 .
for the case of uniaxial compression ( @xmath89 ) , the the maximum tensile pressure @xmath187 is attained at the points where the interface intersects the @xmath77-axis or @xmath69-axis ; while the maximum compressive pressure @xmath188 is attained at the points where the interface intersects the @xmath84-axis .
for the case of uniaxial tension , the sign of the above pressures is changed but the magnitude remains the same .
taking @xmath189 @xmath190 gpa , @xmath182 gpa , @xmath191 gpa , @xmath183 , the estimated maximum tensile pressure is 37 mpa , which is much smaller than that due to martensitic transformation of b2 phases .
thus , the hydrostatic pressure due to the elastic mismatch between the martensite phase and the glassy matrix can be reasonably neglected in the estimation of the interface bonding strength .
in summary , salient conclusions can be drawn below based on our current experimental and theoretical studies : 1 .
three types of shear bands with distinct morphological characteristics are identified in the deformed metallic - glass composites .
the type - i shear bands are oriented 45 degrees relative to the loading axis and related to the external stress field ; while the type - ii and type - iii shear bands as seemingly joining or emanating from the crystalline phases have a different orientation .
based on the eshelby theory , it is shown that the type - ii shear bands result from the stress field caused by the martensitic transformation of the b2 phases , while the type - iii shear bands result from the stress field caused by the elastic mismatch induced after the martensitic transformation .
2 . the back stress effect during the martensitic transformation and the elastic misfit effect
afterwards act together and increase the elastic energy storage during the deformation of the metallic - glass composite , giving rise to work hardening .
3 . a stress - affected zone ( saz ) around the crystalline phase is defined based on our stress analysis .
the presence of the sazs rationalizes the phenomenon of the deflection of the type - i shear bands . 4 .
the hydrostatic pressure acting on the crystalline - amorphous interface is derived .
depending on the loading modes , the maximum tensile pressure could be attained at different points along the interface .
interfacial debonding may take place once the tensile hydrostatic pressure exceeds the interfacial bonding strength .
experimental assistance and insightful discussion with prof z. p. lu , dr .
q.wang , , dr . z. y. liu , dr . w. jiao and mr . y. f. ye , mrs y.m .
lu are appreciated .
the work is supported by german science fundation under leibniz program ( grant no .
ec111/26 - 1 ) and the research grant council ( rgc ) of the hong kong government through the general research fund ( grf ) with the account numbers cityu117612 , 102013 , 11209314 , 9042066 and 9054013 .
according to the green s function theory , the displacement at the position @xmath164 due to a point force @xmath192 at @xmath193 in an infinite elastic medium is : @xmath194 where @xmath19 and @xmath20 are the shear modulus and poisson s ratio , respectively . according to [ a1 ] and superposition principle of linear elasticity , the constrained displacement in the stage iii as described in section 3
is : @xmath195 where the integration is performed on the surface of the spherical b2 phase @xmath196 . using the gauss theorem and @xmath197
when acting on the @xmath198 , one obtains : @xmath199 where @xmath200 and @xmath201 are the biharmonic potential and the newtonian potential of attracting matter of unit density filling the inclusion volume @xmath202 . for a spherical b2 phase , we use the spherical coordinate to integrate @xmath203 and @xmath114 , respectively . in the integration
, we choose the @xmath69 axis always aligning along the direction of vector @xmath164 . by doing this ,
the @xmath203 and @xmath114 is integrated as : [ a4 ] @xmath204dr'=\frac{4\pi r}{3}a^{3}+\frac{4\pi}{15r}a^{5 } \\
\nonumber\\ \phi&=&\int_{v}\frac{dv}{|\mathbf{r}-\mathbf{r'}|}=\int_{0}^{a}\int_{0}^{\pi}\int_{0}^{2\pi}\frac{r'^{2}\sin{\theta}dr'd\theta d\varphi}{\sqrt{r^{2}+r'^{2}-2rr'\cos{\theta}}}\nonumber \\ \nonumber\\ & = & 2\pi\int_{0}^{a}\int_{0}^{\pi}\frac{-r'd(r'\cos{\theta})d\theta}{\sqrt{r^{2}+r'^{2}-2rr'\cos{\theta}}}=\frac{4\pi a^{3}}{3r}\end{aligned}\ ] ] + based on [ a4 ] , we have : [ a5 ] @xmath205 taking [ a5 ] into [ a3 ] , we get the constrained displacement due to the spherical inclusion : @xmath206\sigma^{t}_{ij}x_{j}+\frac{1}{4g(1-\nu)}\frac{a^{3}}{r^{5}}\left(1-\frac{a^{2}}{r^{2}}\right)\sigma_{jk}^{t}x_{i}x_{j}x_{k}\ ] ] + and the constrained strain @xmath32 ( as shown in [ eq2 ] in the text ) is readily obtained from @xmath207 .
for ( @xmath208 ) , there are following equivalences@xcite : [ b1 ] @xmath209\\ x_{i}\epsilon_{ij}^{t}x_{j}&=&x_{i}\epsilon^{t}(2n_{i}n_{j}-\delta_{ij})x_{j}=\epsilon^{t}[(\mathbf{n}\centerdot\mathbf{r})^{2}-r^{2}]\end{aligned}\ ] ] taking these equivalences into [ eq2 ] and noticing that @xmath210 , one could get the explicit expression for the constrained strain field due to a pure shear of eigenstrain in @xmath79 plane : @xmath211 - 2r^{2}[(\mathbf{n}\centerdot\mathbf{r})(n_{i}x_{j}+n_{j}x_{i})-x_{i}x_{j}]}{r^{4}}\right\}+\\ \\ & \frac{\epsilon^{t}}{3(1-\nu)}\left(\frac{a^{3}}{r^{3}}\right)\left[(1 - 2\nu)+\frac{3}{5}\left(\frac{a^{2}}{r^{2}}\right)\right]\left\{\frac{2r^{2}(2n_{i}n_{j}-\delta_{ij})-6[(\mathbf{n}\centerdot\mathbf{r})(n_{i}x_{j}+n_{j}x_{i})-x_{i}x_{j}]}{2r^{2}}\right\}+\\ \\ & \frac{\epsilon^{t}}{2(1-\nu)}\left(1-\frac{a^{2}}{r^{2}}\right)\left(\frac{a^{3}}{r^{3}}\right)\left\{\frac{(r^{2}\delta_{ij}-10x_{i}x_{j})[2(\mathbf{n}\centerdot\mathbf{r})^{2}-r^{2}]+2r^{2}[(\mathbf{n}\centerdot\mathbf{r})(n_{i}x_{j}+n_{j}x_{i})-x_{i}x_{j}]}{r^{4}}\right\ } \end{aligned}\ ] ] + for @xmath93 , @xmath212 , taking this into [ eq2 ] and noticing the equivalence : @xmath213 , one could get the constrained stress field caused by a pure dilation of eigenstrain in the matrix : @xmath214\\ \\
& \left(\frac{r^{2}\delta_{ij}-3x_{i}x_{j}}{r^{2}}\right)+\frac{3(1 - 2\nu)\epsilon_{d}}{8(1+\nu)(1-\nu)}\left(1-\frac{a^{2}}{r^{2}}\right)\left(\frac{a^{3}}{r^{3}}\right)\left(\frac{r^{2}\delta_{ij}-8x_{i}x_{j}}{r^{2}}\right ) \end{aligned}\ ] ] + where @xmath215 for the equivalent eigenstrain due to the uniaxial compressive @xmath83 .
in a cartesian coordinate system , the traceless strain tensor @xmath171 is in generally expressed as@xcite : @xmath216 where @xmath65 , @xmath66 , @xmath67 , are the eigenvalues of the strain tensor or the three principal strains with the principal axes aligning along the unit vectors @xmath217 .
thus , @xmath218/r\\ & \epsilon_{ij}^{t}g_{i}g_{j}=\epsilon_{ij}^{t}(x_{i}x_{j}/r^{2})=[(2\lambda_{n}+\lambda_{k})(\mathbf(n)\centerdot\mathbf{r})^{2}+(\lambda_{n}+2\lambda_{k})(\mathbf(k)\centerdot\mathbf{r})^{2}-(\lambda_{n}+\lambda_{k})r^{2}]/r^{2}\end{aligned}\ ] ] when the @xmath68 aligns along the direction of @xmath164 , @xmath219 , @xmath220 , @xmath221 takes the maximum negative value of @xmath222 , thus the pressure takes the maximum positive value @xmath223 . for the @xmath224 , the shear strain size @xmath71
is defined as @xmath225 , we have @xmath226 , where @xmath173 . when the @xmath227 aligns to the vector @xmath164 , @xmath228 , @xmath220 , @xmath221 takes the maximum positive value of @xmath65 .
thus , the pressure takes the maximum negative value @xmath229 , where @xmath177 .
56 natexlab#1#1[2]#2 [ 1]#1 [ 1]http://dx.doi.org/#1 [ ] [ 1]pmid:#1 [ ] [ 2]#2 , , . .
. , . , . . , . , ,
. . , . , , ,
. . , . , , , ,
. . , . , , , , , , ,
. . , . , , , , ,
. . , . , , ,
. . , . , , , , ,
. . , . , , , , , , , ,
. . , . , , , , ,
. . , . , , ,
, , , , , . . ,
, , , . . ,
, , , , , , , a. . , . , , , , , , b. . , . , , , , .
, , , . . ,
, , , , , , . . ,
, , , . . ,
, , , . . ,
, , , . . ,
, , , , , , . . ,
, , , , , , , , , , . . ,
, , , , , , . .
, , , , . . ,
, , , , , a. . , . , , , , , , , b. . , . , , , , , , , , .
, , , , . .
, , , , . . ,
, , , . . ,
, , , , , , , , a. . , .
, , , , , , , , , b. . , .
, , , , , , , . .
, , , , . . ,
, , , , , , . . ,
, , , , , , . . ,
, , , , , , , , , , , . . , . , , , ,
. . , . , , , , , ,
. . , . , , ,
, , , , , a. . , . , , , , , , .
, , , , , , , , , b. . , . , , , , , .
, , , , , , , , , . . ,
, , , , , , , . . ,
, , , . . ,
, , , . . , | in this paper , we present a thorough stress analysis of the cu - zr metallic - glass composite with embedded b2 particles subject to a martensitic transformation . within the framework of the eshelby theory ,
we are able to explain , in a quantitative manner , ( 1 ) the formation of three types of shear bands with distinct morphologies as observed experimentally in the severely deformed cu - zr metallic - glass composite and ( 2 ) the work hardening ability of the cu - zr metallic - glass composite as related to the coupled effects of elastic back stress and elastic mismatch caused by the martensitic transformation .
furthermore , we also discuss the issues about the stress affected zone of the individual b2 phase and the stability of the crystalline - amorphous interface . given the general agreement between the theoretical and experimental findings , we believe that the outcome of our current work can lead to a deeper understanding of the transformation - induced plasticity in the cu - zr based metallic glass composites , which should be very useful to the design of the metallic - glass composites with improved ductility . metallic glass , a.ductility , a.phase transformation , c.analytic functions |
observations of secondary eclipses in exoplanetary systems , starting with hd 209458b @xcite and tres-1 @xcite , have made it possible to measure the integrated day - side flux of hot jupiters ( for a review of transiting exoplanet science see * ? ? ?
* ) . by carefully studying of the shape of the ingress and egress of secondary
eclipses , it should eventually be possible to map the day - side of such planets @xcite . to characterize the planet s longitudinal temperature profile at all longitudes , however , observations must be made at a variety of points in the planet s orbit .
the first successful observations of this sort were reported by @xcite , @xcite and @xcite .
the unprecedented quality of the data in @xcite and @xcite made it possible to model the planet not merely as day and night hemispheres , but to divide the planet into longitudinal slices , hence producing the first ( albeit coarse ) maps of an exoplanet under restricted assumptions . in this paper
we elaborate on the inversion techniques which one can use to obtain a longitudinal brightness map from the light curve of phase variations .
such maps promise to be powerful diagnostic tools for simulations of hot jupiter atmospheric dynamics because they are nearly model - independent .
the detailed study of coupled radiative transfer and dynamics in the atmospheres of hot jupiters is a tremendously complex science ( for a current review of the field , see * ? ? ?
certain models of hot jupiter atmospheres predict significant variability in the integrated brightness of the planets ( * ? ? ?
* and references therein ) .
although it may be possible to glean useful information from the phase function light curves of planets with such variable atmospheres , we elect in this letter to ignore time variability , in line with the detailed modeling of @xcite . we also choose to ignore limb darkening / brightening since most atmospheric models of hot jupiters predict infrared photon absorption lengths much shorter than the scale height of temperature variations .
in any case , the presence of limb darkening would not significantly change the present analysis since the limb does not contribute much to the integrated light of the planet ( we verify this in 3.4 ) .
star spots can plague phase function observations since they change the overall brightness of a planet / star system as they rotate into and out of view .
fortunately , star spots vary on longer timescales ( @xmath1weeks ) than the orbital periods of currently known transiting exoplanets ( @xmath1days ) , and have larger variations in the optical / near - ir which can be used to characterize and subtract the star spot variation ( eg : * ? ? ?
furthermore , some instruments on the spitzer space telescope @xcite exhibit detector ramps which distort phase function observations , especially near the beginning of a time series ( eg : * ? ? ?
* ) . both of these observational challenges can be overcome through longer observations : full orbits or more .
for the purposes of this letter , we assume that any and all variations in the infrared brightness of an extrasolar planetary system are due to changes in the flux from the planet as different portions of the planet rotate in and out of view . stellar variability , which contributes to the photometric scatter in the light curve , will likely be the limiting factor in future phase function analysis .
the mapping technique described here is predicated on the known rotation rate of the planet .
as such , it is only applicable to tidally locked planets , where the rotation and orbital periods are identical .
most hot jupiters are tidally locked so this is not a problematic restriction .
eccentric hot jupiters are thought to be in pseudo - synchronous rotation , in which case their rotational period can be derived from their orbital parameters .
it is unlikely that eccentric planets have steady - state atmospheric dynamics and hence , although the light curve inversion methodology described in this letter may be applicable , it is not clear that the result would be a `` map '' .
the object of this letter is to demonstrate how to transform an observed light curve , @xmath2 , into a longitudinal brightness map of the planet , where the planet s phase , @xmath3 , is the angle in the plane of the planet s orbit between the planet , its host star and the planet s position at superior conjunction ( @xmath4 at superior conjunction ; @xmath5 at inferior conjunction ) . for edge - on systems
, @xmath3 corresponds to the observer
planet star angle ( @xmath6 at secondary eclipse ; @xmath5 at transit ) . in the interest of simplicity , we ignore the dips in light due to transit and secondary eclipse which occur in transiting systems , although these are crucial in pinning down the absolute flux of the planet ( for non - transiting planets @xmath2 is only known to within an additive constant ) . if @xmath7 and @xmath8 are the longitude and latitude on the planetary disk as seen by an observer and @xmath9 is the intensity map of the planet , the total flux from the planet as seen by that observer is : @xmath10 for a tidally locked planet , it is possible to define longitude , @xmath11 , and latitude , @xmath12 , in the planet s rotating frame , such that @xmath13 at the sub - stellar point , @xmath14 at the planet s north pole , and @xmath11 increases in the direction of rotation of the planet .
the rotation that relates @xmath15 to @xmath16 can be expressed in terms of euler angles : @xmath17 , @xmath18 , @xmath19 , where @xmath20 is the inclination of the planet s orbit ( @xmath21 for a face - on orbit ; @xmath22 for an edge - on orbit ) .
the rotation leads to a system of three coupled equations which can be solved for @xmath23 and @xmath24 . by inserting these expressions into equation [ flux_equation ] ,
one obtains an equation for the observed flux from the planet as a function of the specific intensity of the planet at different longitudes and latitudes in effect , a transformation from a brightness map to a light curve : @xmath25 . although the formalism above is sufficient to produce a light curve from a two - dimensional planetary map , it is instructive to further constrain the problem by considering planets with inclinations of _ precisely _ @xmath26 , with the understanding that the resulting solutions should be approximately correct for transiting planets . in the _ worst - case _ scenario of a transiting system with @xmath27 and @xmath28 constant in @xmath12 ,
the edge - on approximation leads to 5% errors in @xmath2 . for edge - on orbits , the rotation relating @xmath15 to @xmath16
simplifies to @xmath29 and @xmath30 .
the expression for @xmath2 can then be written as @xmath31 there are no current observations which can constrain the @xmath12-dependence of @xmath9 , but for edge - on orbits the latitudinal dependence of the intensity is unimportant since one can define @xmath32 , which represents the flux contribution from an infinitesimal slice of the planet when viewed face - on .
the integrated flux from the planet at a given point in its orbit is then @xmath33 it is useful to think of this integral as a convolution @xmath34 , with the piece - wise defined kernel @xmath35 : @xmath36 the kernel , which represents the response of the phase function to a delta function in @xmath37 , is very broad , with a full width at half - maximum of @xmath38 .
in the previous section we developed an analytic expression , equation [ edge_on_integral ] , for the convolution @xmath39 . there is no closed expression for the deconvolution , @xmath40 , and there is no guarantee that solutions are unique .
furthermore , two problems arise in practice : the light curve is only sampled at discrete values of @xmath3 which may not span a full planetary rotation , and the measurements of @xmath2 are not arbitrarily precise but instead have associated uncertainties @xmath41 . given these realities , it is useful to develop model maps which simplify the integral in equation [ edge_on_integral ] and then use numerical methods to solve the problem or
better yet allow for direct inversion .
the planet is divided into @xmath42 longitudinal slices of width @xmath43 .
each slice has a uniform intensity in both longitude and latitude .
this flux distribution is not ruled out by the observations and more importantly smoothing the steps does not significantly change the light curve , provided the total flux from each slice and their brightness - weighted longitude are unchanged .
the @xmath44th slice is centered on @xmath45 , where the phase offset , @xmath46 , is useful to accommodate slight discrepancies between the light curve maximum and superior conjunction .
the intensity map for the planet is given by : @xmath47 since , in practice , one is only ever concerned with comparing the model phase function to data at a finite number of discrete phases , the transformation from n - slice map to light curve can be expressed in matrix form : @xmath48 .
the matrix @xmath49 is defined as @xmath50 , where @xmath51 and @xmath52 represent the leading and trailing edges of the slices : @xmath53\right).\ ] ] sinusoidal basis maps have the advantage of producing sinusoidal light curves _ of the same frequency and phase offset_. if a planet map is composed of sinusoids , @xmath54 , the light curve is simply given by @xmath55 .
the coefficients of @xmath56 are related to those of @xmath37 by : @xmath57 where @xmath44 must be even .
sinusoidal modes with odd @xmath44 ( other than @xmath58 ) do not have a phase function signature . in figure [ sinusoidal_kernel ]
we show the light curve contributions for a handful of sinusoidal modes , assuming that all of the modes have the same amplitude in @xmath37 .
the higher frequency modes are strongly suppressed due to the broad smoothing kernel .
this low - pass filter limits the number of modes which can be meaningfully fit with a given light curve .
the uncertainty in a sinusoidal mode in the light curve is related to the uncertainty in the map sinusoidal modes by equation [ sinusoidal ] ( eg : @xmath59 ) . ) are invisible due to symmetry.,width=317 ] since both the n - slice and sinusoidal models described above provide computationally efficient ways to generate light curves from maps , one can use a @xmath60 fitting routine ( markov chain monte carlo , levenberg - marquardt , etc . ) to produce a map from a given light curve .
it is simply necessary to demand that @xmath61 be strictly positive .
these techniques have the advantage of naturally producing error estimates for the resulting map .
although a unique best - fit map can always be determined in this way , the uncertainty in the fit parameters may be very large if the number of parameters is not commensurate with the signal - to - noise of the light curve . for the n - slice model , the uncertainty balloons for _ all _ the @xmath62 if too many free parameters are used .
we therefore suggest running multiple fits with different numbers of slices .
when the addition of a slice does not improve the @xmath60 , one has achieved the best model that the data can support . instead of repeating the fit with fewer slices
, one can apply smoothing to the map in the form of a bayesian prior ( eg : * ? ? ?
* ) . for well - designed priors
, however , there is no fundamental difference between models with large @xmath42 and long smoothing lengths , versus models with smaller @xmath42 and little or no smoothing . in the interest of simplicity
we recommend using fewer slices rather than smoothing .
an observed light curve can be quickly deconvolved into sinusoidal maps by determining its fourier components , @xmath63 , @xmath64 and @xmath65 , then converting them via the equation [ sinusoidal ] .
it is expedient to assume that there is no power in the odd modes ( other than @xmath58 ) to avoid degenerate solutions , but this may lead to a systematic error in the model map , depending on how much power is present in these modes in the real map .
the uncertainty in the map parameters may be determined using a @xmath60 routine or monte carlo analysis .
since the sinusoidal model has linearly independent modes , only the uncertainty in the highest - frequency modes explodes when too many modes are considered . as a rule of thumb , one should truncate the fourier series once the uncertainty in coefficients becomes greater than the coefficients themselves .
we now turn to an example map and test the ability of the algorithms described above to recover the correct features of this map .
the top panel of figure [ map_cho_03 ] shows the brightness map computed from a snapshot of the atmospheric dynamics model of @xcite .
we performed an analogous test and obtained comparable results using a snapshot from the model of @xcite , which we do not include here in the interest of space .
the brightness map was generated by treating each pixel of the temperature map as a blackbody and computing an associated intensity at @xmath66 m .
the bottom panel of figure [ map_cho_03 ] shows the integrated longitudinal brightness map , @xmath61 .
note that the @xmath67 term in the integral for @xmath37 attenuates the flux contribution from the poles of the planet . also shown in the bottom panel
are the best fit maps for the n - slice and sinusoidal models .
the light curve associated with the map , as well as the best - fit light curves for the models , are shown in figure [ lc_cho_03 ] .
the map @xmath37 was converted to an idealized light curve , and mock observations ( comprised of 100 data points ) were generated by removing the segments of the light curve corresponding to the transit and secondary eclipse , then scaling the planet / star flux ratio and the photometric uncertainties to roughly match those of @xcite .
both models reproduce the features of the map , as well as the light - curve .
the best - fit 5-slice model was determined using a levenberg - marquardt @xmath60-minimization routine .
the sinusoidal map was determined by decomposing the light curve into sinusoidal components to @xmath68 , then converting the coefficients using equation [ sinusoidal ] . the uncertainties estimated by monte carlo analysis in the @xmath69 terms are larger than their amplitudes and are therefore ignored .
the insensitivity of phase functions to the first and most important odd mode , @xmath70 , leads to @xmath71% errors in the resulting map , based on the maps of @xcite and @xcite .
we model the effect of limb darkening by adding @xmath72 $ ] to the expression for @xmath61 and adding @xmath73 $ ] to the integrand of equation [ edge_on_integral ] . we find the resulting light curve to differ by less than @xmath74% , justifying our decision to neglect limb darkening in the formalism above .
k and @xmath75 k. in the lower panel , the solid line is @xmath61 , the histogram and associated error bars represents the best - fit 5-slice model , and the gray band is the @xmath76 confidence interval for the sinusoidal map . note that positive @xmath11 are to the left of the plot , to facilitate comparison with the light curve in figure [ lc_cho_03].,width=317 ] , convolved with photometric scatter comparable to @xcite .
the solid and dashed lines shows the light curves of the best - fit n - slice and sinusoidal models , respectively.,width=317 ]
the best current light curves can be represented by the function @xmath77 which can be directly translated into a longitudinal brightness map using equation [ sinusoidal ] .
the @xmath70 modes should cancel out by symmetry so their presence in the light curve would indicate systematic errors ; @xmath78 modes are generally lost in the noise .
the longitudes and amplitudes of the primary hot - spots and cold - spots can be determined from the @xmath58 terms , while the relative strength of the @xmath58 and @xmath68 modes indicates whether there are secondary local maxima / minima . by the same token ,
a 4-slice model with variable phase offset should be sufficient to model most phase function light curves . for light curves with incomplete phase coverage , the uncertainty in the sinusoidal map is the same at all longitudes whereas the n - slice model naturally has larger uncertainties for slices which were visible for less time , an intuitive and desirable property .
so far only half - orbits have been allocated to phase function studies , but the warm spitzer mission will provide a perfect opportunity to obtain light curves spanning full planetary orbits @xcite . the contribution of sinusoidal modes to the observed light curve decreases precipitously with @xmath44 . not only does the @xmath39 transformation suppress the terms as @xmath79 , but the intrinsic power of these modes in the underlying map might drops as @xmath80 ( as is the case with themap from * ? ? ?
* ) . as a result
, the @xmath69 modes in light curves will be @xmath81 times weaker than the @xmath68 modes and will likely remain undetectable with jwst . even when higher quality light curves are eventually obtained and assuming that stellar variability is not the limiting factor the physical significance of the model map would be questionable due to the insensitivity of the phase function to odd sinusoidal modes .
the power in these modes can be constrained theoretically by dynamical atmospheric models and observationally through secondary eclipse mapping , which promises to be feasible with jwst .
n.b.c . is supported by the natural sciences and engineering research council of canada .
is supported by a national science foundation career grant .
support for this work was provided by nasa through an award issued by jpl / caltech .
the authors wish to thank e. rauscher and a. showman for use of their model temperature maps .
, d. , agol , e. , charbonneau , d. , cowan , n. , knutson , h. , & marengo , m. 2007 , in american institute of physics conference series , vol .
943 , american institute of physics conference series , ed .
l. j. storrie - lombardi & n. a. silbermann , 89100 | we describe how to generate a longitudinal brightness map for a tidally locked exoplanet from its phase function light curve .
we operate under a number of simplifying assumptions , neglecting limb darkening / brightening , star spots , detector ramps , as well as time - variability over a single planetary rotation .
we develop the transformation from a planetary brightness map to a phase function light curve and simplify the expression for the case of an edge - on system .
we introduce two models composed of longitudinal slices of uniform brightness , and sinusoidally varying maps , respectively which greatly simplify the transformation from map to light curve .
we discuss numerical approaches to extracting a longitudinal map from a phase function light curve , explaining how to estimate the uncertainty in a computed map and how to choose an appropriate number of fit parameters .
we demonstrate these techniques on a simulated map and discuss the uses and limitations of longitudinal maps .
the sinusoidal model provides a better fit to the planet s underlying brightness map , although the slice model is more appropriate for light curves which only span a fraction of the planet s orbit .
regardless of which model is used , we find that there is a maximum of @xmath0 free parameters which can be meaningfully fit based on a full phase function light curve , due to the insensitivity of the latter to certain modes of the map .
this is sufficient to determine the longitudes of primary equatorial hot - spots and cold - spots , as well as the presence of secondary maxima / minima . |
we consider a system of hamiltonian differential equations on @xmath0 @xmath1 defined by the hamiltonian function @xmath2 .
we denote the time @xmath3 flow map of these equations by @xmath4 .
the flow of these differential equations has two important features .
the first is that @xmath5 is conserved along trajectories .
that is , @xmath6 for all @xmath7 and all @xmath8 .
the second is that phase space volume is conserved by the flow : if @xmath9 is a bounded open set , then @xmath10 for all @xmath3 .
the latter property is a consequence of the symplecticity of the flow .
( see , for example , @xcite ) . for certain molecular dynamics applications , the ideal numerical integrator would retain these two properties of the flow @xcite .
that is , if the integrator with time step @xmath11 defines a map @xmath12 on @xmath0 , we would like both @xmath13 for all @xmath14 ( energy conservation ) and @xmath15 for all bounded open subsets @xmath16 of @xmath0(volume conservation ) .
symplectic integrators such as the implicit midpoint rule conserve volume exactly for all hamiltonian systems , with any number of degrees of freedom , but do not conserve energy @xcite .
it has already been shown that in certain circumstances it is unreasonable to expect that symplectic integrators which also conserve energy exist @xcite . however , volume - conservation is a weaker property than symplecticity for hamiltonian systems of more than one degree of freedom ( @xmath17 ) .
thus it seems plausible that there is a consistent integration scheme which for any hamiltonian function @xmath5 and time step @xmath11 yields a map @xmath12 which conserves both volume and energy . in this article
we will argue that there is no such integration scheme .
we do this by showing that no numerical integrator is able to integrate _ all _ hamiltonian systems in @xmath18 while simultaneously conserving energy and phase space volume . for special hamiltonian systems in @xmath18
such an integrator is possible .
( see @xcite for examples of such systems of any number of degrees of freedom . ) however , our theorem states that for any energy - conserving numerical integrator from a very broad class , there will be at least one hamiltonian system on @xmath18 ( and in fact very many ) for which it does not conserve phase space volume . before stating and proving our main theorem , in section [ sec : integrator ] we will define what we mean by an integrator . from the class of integrators
we then define the _ computationally reasonable _ integrators .
this will include any explicit or implicit formula for defining a new state @xmath19 as a function of a state @xmath20 and a timestep @xmath11 such that the only information used about @xmath5 is its value and the value of its derivatives at a finite number of points @xmath21 , @xmath22 . in section [ sec : main ] we will state and prove our main result in terms of this definition . we imagine that a numerical analyst has devised a computationally reasonable numerical integrator that is energy - conserving for all hamiltonian systems in @xmath18 .
we apply this integrator to the system with hamiltonian function @xmath23 in @xmath18 . for two different inputs we observe
at which points @xmath24 the integrator depends on the function @xmath5 and its derivatives . using this information
, we construct another hamiltonian function @xmath25 which is arbitrarily close to @xmath5 and has the following property : the numerical integrator can not be volume - conserving for both @xmath5 and @xmath26 .
the main result demonstrates that there is no integration scheme that conserves volume and energy for arbitrary hamiltonian systems of any number of degrees of freedom .
however , we can not conclude that there is no scheme that is energy- and volume - conserving for hamiltonian systems in a particular number of dimensions @xmath27 , @xmath17 .
this question is open . to partially address this issue , in section [ sec : multi ]
we will show that under some reasonable but not essential conditions on an integrator , the problem reduces to that of the @xmath28 case .
thus , there is no energy- and volume - conserving integrator for fixed @xmath29 that satisfies these additional assumptions .
we now explain the relation between the results here and those of the well - known paper of zhong and marsden @xcite . in that paper
the authors consider a hamiltonian system for which there are no invariants except energy .
they show that any integrator that is both symplectic and energy - conserving actually computes exact trajectories of the system up to a time reparametrization . from this result
they conclude that energy - conserving symplectic integration is not possible in general , since presumably the set of hamiltonians for which one could compute trajectories exactly up to a time reparametrization are very small .
though this argument is very plausible , it leaves open two questions : 1 .
how do we make precise the idea of something not being possible for any numerical integrator ? 2 .
is it possible to perform energy - conserving and symplectic integration for general hamiltonians when we restrict ourselves to the @xmath18 case ?
the importance of the latter question is that if it were possible , then volume- and energy - conserving integration would be possible for general hamiltonian systems via @xmath30-splitting @xcite .
we provide an answer to the first question in section [ sec : integrator ] with the definition of a computationally reasonable integrator . as for the second question , since volume - conservation and symplecticity
are identical in @xmath18 , zhong and marsden s result shows that in @xmath18 volume- and energy - conserving integration is equivalent to solving the original system exactly up to a time - reparametrization .
( this result is stated and proved for this case as lemma [ lem : reparam ] in section [ sec : main ] . )
the main result of our paper answers the second question in the negative by showing that it is not possible for general hamiltonian systems in @xmath18 using computationally reasonable integrators .
before we begin we discuss some interesting related work . even though energy and volume conserving integrators may not exist
, the paper @xcite does the next best thing .
there the authors show how to approximate any hamiltonian function arbitrarily well by a special piece - wise smooth function of a form described by @xcite whose trajectories can be integrated while conserving volume and energy .
the original hamiltonian function is not conserved .
however , unlike for standard symplectic methods , a modified hamiltonian function close to the original is conserved exactly @xcite for all time .
an integrator for a hamiltonian system of ordinary differential equations takes a hamiltonian @xmath5 , a step length @xmath11 , and an initial value @xmath31 , and produces a value @xmath32 .
typically , @xmath19 depends on @xmath5 through components of @xmath33 and perhaps @xmath5 itself . if @xmath20 is an approximation to @xmath34 , we take @xmath19 to be an approximation to @xmath35 . [
defn : integrator ] an integrator @xmath36 is a function that takes arguments @xmath37 , @xmath38 , @xmath39 and either returns @xmath32 or is not defined .
we write @xmath40 we have allowed the integrator to not be defined for certain input values .
this is often the case for implicit integrators when the vector field is insufficiently smooth or the time step is too large . in order to get meaningful constraints on what is computationally reasonable
, we can not let arbitrary maps be included in the class of algorithms we study .
after all , the exact flow map has all the qualitative features one could want , but it is not feasible to compute it ( even to machine precision ) for most applications .
we would like our definition to be broad enough to include most existing numerical integrators .
informally , we say an integrator @xmath36 is _ computationally reasonable _ if for each @xmath20 and @xmath11 , @xmath41 depends on @xmath5 only through its value and the value of its derivatives at a finite number of points @xmath21 , @xmath42 . in the following formal definition we use multi - index notation to define higher - order derivatives : for @xmath43 and @xmath44 we let @xmath45 [ defn : reasonable ] an integrator @xmath36 is _ computationally reasonable _ if for each @xmath46 , @xmath39 and @xmath47 there exists 1 .
@xmath48 , 2 .
@xmath49 , 3 .
@xmath50 , such that for any function @xmath51 that satisfies @xmath52 either @xmath53 or both are not defined . in the remainder of this section we discuss examples of the class of computationally reasonable integrators .
first note that all explicit methods fit into this class .
we informally define an integrator to be explicit if it can be implemented by an algorithm that terminates in a finite number of steps using function evaluations of @xmath5 or its derivatives , arithmetic operations , and logical operations .
this includes all of the explicit runge - kutta and partitioned runge - kutta methods , for example .
it also includes integrators that collect information adaptively to perform a step , such as the bulirsch - stoer method @xcite .
taylor series methods are also included .
the class of computationally reasonable integrators also includes implicit methods , such as the implicit runge - kutta methods .
here we define an implicit method to be one where @xmath19 is specified by requiring it to be the solution to a nonlinear system of equations in @xmath5 and its derivatives .
implicit algorithms can not in general be implemented exactly in a finite number of steps , but they still fit into the framework of definition [ defn : reasonable ] . to see this , note that even though solving a system of nonlinear equations exactly typically requires looking at @xmath5 and its derivatives at an infinite number of points ( while performing the newton iteration , for example ) , determining if we have a solution to a nonlinear system of equations only requires examining a finite number of points .
so if @xmath19 solves the equations for @xmath5 , it will still solve the equation for any @xmath26 which is identical to @xmath5 at the points .
similarly , step - and - project methods are included in this class .
these are methods that consist of one step of a simpler method followed by a projection onto a manifold ( * ? ? ?
what integrators do not satisfy definition [ defn : reasonable ] ?
integrators that require the exact computation of integrals of @xmath5 or its derivatives do not . computing the integral of a general function
requires knowing its value at an infinite set of points on the domain of integration . unlike in the case of solving nonlinear equations ,
if we are given a value for the integral , there is no finite set of points at which we can examine the function to verify that the value is correct .
of course , it is possible to define an numerical integrator that uses integrals of functions , but the actual computation of these integrals for general hamiltonian systems would require numerical quadrature .
this in turn would require sampling the function at a finite number of points , and introducing truncation error .
the new method with this additional truncation error does form a computationally reasonable integrator , while the original method with the exact integral does not .
finally , we note that multistep methods are not even integrators according to definition [ defn : integrator ] .
we believe our framework could be extended to multistep method but we do not do so here .
to prove our main result theorem [ thm : main ] we use the following lemma .
it shows that for a hamiltonian system in @xmath54 , the map defined by an energy and area preserving integrator is just a time - reparametrization of the flow map .
as discussed in the introduction , this is essentially zhong and marsden s result @xcite in the two dimensional case .
the only addition is that we show that the time - reparametrization is just a constant rescaling of time where locally the constant does not depend on energy .
[ lem : reparam ] let @xmath55 be a smooth function and @xmath4 its induced hamiltonian flow map .
let @xmath56 be a particular energy .
let @xmath57 be an open set whose intersection with @xmath58 is a simple curve @xmath59 .
suppose that @xmath60 on @xmath59 .
let @xmath12 , @xmath61 be a continuous area - conserving map defined on @xmath57 that conserves @xmath5 . then there is a constant @xmath62 such that @xmath63 for all @xmath64 .
* proof : * on @xmath57 we can define canonical action - angle coordinates @xmath65 in which the hamiltonian function is @xmath66 .
the flow map is then @xmath67 where @xmath68 , since we still have @xmath69^t = \nabla h \neq 0 $ ] in the new coordinates . to study the jacobian of @xmath12 in the coordinates @xmath70
we let @xmath71 for @xmath72 .
then the jacobian of @xmath12 is @xmath73.\ ] ] area conservation implies that the determinant is @xmath74 and so @xmath75 for all @xmath76 .
this yields @xmath77 for all @xmath78 for some @xmath79 which may depend on @xmath80 .
hence , if we choose @xmath62 so that @xmath81 we have @xmath82 for all @xmath83 , as required .
@xmath84 for the main result we will use the following very weak definition of consistency . [ defn : consistent ] let @xmath4 be the flow map for the hamiltonian @xmath46 .
an integrator @xmath36 is consistent at @xmath8
if @xmath85 [ thm : main ] let @xmath36 be an integrator for a hamiltonian system in @xmath18 . let @xmath86 .
let @xmath57 be an open neighbourhood of @xmath87 .
suppose : 1 .
@xmath36 is consistent at @xmath88 ( definition [ defn : consistent ] ) .
@xmath36 is computationally reasonable ( definition [ defn : reasonable ] ) .
3 . @xmath36 conserves energy for any @xmath5 , @xmath14 , and @xmath11 for which it is defined . 4 . for @xmath89
, @xmath90 is defined , depends continuously on @xmath14 , and conserves volume on @xmath57 .
then for all sufficiently small @xmath11 there is a @xmath91 function @xmath92 such that if @xmath93 then @xmath94 is not simultaneously defined , continuous , and volume - conserving on @xmath57 . for each such @xmath11 ,
the @xmath95 constructed can be replaced by @xmath96 for @xmath97 and the same result holds .
* proof : * for any @xmath98 $ ] , since @xmath99 conserves energy and is continuous on @xmath57 , the set @xmath100 is mapped onto itself .
as @xmath99 conserves volume , lemma [ lem : reparam ] shows that it is identical to the flow of the original hamiltonian system on @xmath101 with a rescaling of time : @xmath102 for all @xmath103 , where @xmath62 does not depend @xmath104 .
the consistency condition at @xmath88 implies that for small enough @xmath11 we have that @xmath105 . from now on , we assume @xmath11 is small enough so that @xmath105 .
consider the integrator applied to @xmath5 at the point @xmath106 . since @xmath36 is computationally reasonable ( definition [ defn : reasonable ] )
, there are a finite number of points @xmath107 , @xmath22 , such that @xmath36 only depends on @xmath5 at these points .
choose a @xmath108 big enough so that the interval @xmath109 $ ] contains none of the points @xmath110 and is disjoint from the interval @xmath111 $ ] .
consider the integrator applied to @xmath5 at the point @xmath112 .
there are points @xmath113 such that @xmath114 only depends on @xmath5 at these points .
let @xmath95 be a @xmath115 function such that 1 .
@xmath116 for @xmath104 not in @xmath117 $ ] , and @xmath118 , 2 .
@xmath119 for all @xmath104 , 3 . for some @xmath120 , @xmath121 for @xmath122 .
note that multiplying @xmath95 by any factor in @xmath88 gives a function satisfying the same conditions .
let @xmath123 and let @xmath124 denote the flow map of the system with this hamiltonian .
now if @xmath94 were defined , continuous and volume - conserving on @xmath57 , then by lemma [ lem : reparam ] @xmath125 for all @xmath126 for some constant @xmath127 . we will show that this is impossible .
first note that the integrator gives the same result for both hamiltonians at @xmath88 .
this is because here the result of the integrator only depends on @xmath5 only at the points @xmath128 , at which @xmath5 and @xmath26 agree .
since @xmath129 we must have @xmath130 . on the other hand , the integrators also give the same results for both hamiltonians at @xmath112 .
this implies @xmath131 in the following way .
we have that @xmath132 the inequality follows from lemma [ lem : differ ] , since the hamiltonians differ on the interval @xmath133 $ ] .
so @xmath131 .
this is a contradiction .
therefore , @xmath94 can not be simultaneously defined , continuous , and volume - conserving on @xmath57 .
@xmath84 the following lemma asserts the intuitively clear fact that if @xmath134 is positive for @xmath135 $ ] then the flows starting from @xmath112 corresponding to @xmath5 and @xmath26 are different for positive times . [
lem : differ ] let @xmath136 be @xmath115 with @xmath137 , @xmath138 for all @xmath139 and @xmath121 for all @xmath135 $ ] .
let @xmath140 and @xmath123 .
let @xmath4 and @xmath124 be the respective hamiltonian flow maps of @xmath5 and @xmath26 .
then @xmath141 for @xmath142 . *
the trajectory for the hamiltonian @xmath5 as a function of time is @xmath143 for all @xmath3 .
letting @xmath144 describe the position for the hamiltonian @xmath26 , the usual solution technique gives @xmath145 the function @xmath146 is strictly increasing and so has a well defined inverse .
we can write @xmath147 now @xmath148 for @xmath149 , so @xmath150 for @xmath151 .
hence @xmath152 for @xmath151 and the two flow maps can not be equal for @xmath151 . @xmath84
in the previous section we showed that there can be no general energy- and volume - conserving integration schemes because there are no integrators that conserve energy and volume for all hamiltonians in @xmath18 . however ,
suppose we ask if such integrators exists for hamiltonian systems of dimension @xmath27 , @xmath17 .
we conjecture that a result like theorem [ thm : main ] still holds in this case . however , the method of proof for the @xmath28 case does not extend to this case .
instead we will state two conditions on an integrator , either one of which prevents it from being volume- and energy - conserving for general hamiltonian systems in @xmath0 .
both of these conditions are desirable for an integrator to have , but unlike computational reasonibility , it is not difficult to imagine a practical integrator that did not satisfy them .
the proof of the theorems in this section will work by showing that , if a computational reasonable energy - conserving integrator with either condition exists , a special hamiltonian in @xmath0 can be constructed for which it does not conserve volume . 1 .
@xmath36 is consistent ( definition 3 ) .
@xmath36 is computationally reasonable ( definition 2 ) .
3 . @xmath36 conserves energy for any @xmath5 , @xmath14 and @xmath11 for which it is defined . 4 .
@xmath36 satisfies condition [ cond : untouch ] 5 . for sufficiently small @xmath11 ,
@xmath90 is defined , continuous , and conserves volume for @xmath164 .
then for sufficiently small @xmath11 , there is a @xmath91 function @xmath92 such that if @xmath165 , then @xmath166 is not simultaneously defined , continuous , and volume - conserving on @xmath57 . for each such @xmath11 ,
the @xmath95 constructed can be replaced by @xmath96 for @xmath97 and the same result holds .
* proof : * we will use the hypothesized integrator on @xmath0 to construct an integrator @xmath167 on @xmath18 satisfying the conditions of theorem [ thm : main ] .
let @xmath55 be a given hamiltonian function .
define @xmath168 by @xmath169 .
we define the integrator @xmath167 by @xmath170 let @xmath171 be given by @xmath172 it is straightforward to check that @xmath167 satisfies the conditions of theorem [ thm : main ] on @xmath173 .
thus , by the theorem , we have an arbitrarily small @xmath115 function @xmath95 such that @xmath167 is not simultaneously defined , continuous , and volume - conserving on @xmath173 for @xmath174 .
let @xmath175 for @xmath176 .
now suppose that @xmath177 is defined , continuous , and volume preserving on @xmath57 .
we will derive a contradiction by showing this implies that @xmath178 is , in fact , defined , continuous , and volume - conserving on @xmath173 .
first note that @xmath177 being defined and continuous on @xmath57 implies that @xmath178 is defined and continuous on @xmath173 . to check volume conservation , note that the jacobian of the map @xmath177 has structure @xmath179\ ] ] where we have put the variables in order @xmath180 and @xmath181 is a 2-by-2 matrix .
since the determinant of this matrix is 1 by volume - conservation , the determinant of @xmath181 must be 1 .
but @xmath181 is the jacobian of @xmath178 , so this latter map must be area preserving .
this contradicts our earlier assumption . @xmath182
the second condition states that if the hamiltonian system consists of @xmath29 identical uncoupled one - degree - of - freedom systems , then the integrator itself should consist of @xmath29 identical uncoupled maps on the state - space of each subsystem . though this is certainly a nice property for the integrator to have ( since the flow map has the same property ) there are many integrators for which it does not hold .
for example , step - and - project methods may not satisfy this condition , even if the underlying one - step method does . 1 .
@xmath36 is consistent ( definition 3 ) .
@xmath36 is computationally reasonable ( definition 2 ) .
3 . @xmath36 conserves energy for any @xmath5 , @xmath14 and @xmath11 for which it is defined .
@xmath36 satisfies condition [ cond : prod ] 5 . for sufficiently small @xmath11 ,
@xmath90 is defined , continuous , and conserves volume for @xmath164 .
then for sufficiently small @xmath11 , there is a @xmath91 function @xmath92 such that if @xmath188 then @xmath166 is not simultaneously defined , continuous , and volume - conserving on @xmath57 . for each such @xmath11 ,
the @xmath95 constructed can be replaced by @xmath96 for @xmath97 and the same result holds .
* proof : * this theorem is proven analogously to the previous theorem . for any @xmath189 we define @xmath190 by @xmath191 .
we define the integrator @xmath167 by @xmath192 we define @xmath193 by @xmath194 as in the proof of the previous theorem , @xmath167 and @xmath173 satisfy the conditions of theorem [ thm : main ] .
thus , by the theorem , we have an arbitrarily small @xmath115 function @xmath95 such that @xmath167 is not defined continuous and volume - conserving on @xmath173 for @xmath195 .
now @xmath197 is defined and continuous on @xmath173 . to check volume conservation , note that the jacobian of the map @xmath177 in this case has structure @xmath198\ ] ] where we have put the variables in order @xmath180 and
each @xmath199 is 2-by-2 .
since the determinant of this matrix must be 1 by volume - conservation and the determinants of the @xmath199 are identical , the determinant of @xmath181 must be @xmath200 . as in the proof of the previous theorem
this implies @xmath178 is area conserving on @xmath173 which is a contradiction . @xmath182 | we consider the numerical simulation of hamiltonian systems of ordinary differential equations .
two features of hamiltonian systems are that energy is conserved along trajectories and phase space volume is preserved by the flow .
we want to determine if there are integration schemes that preserve these two properties for all hamiltonian systems , or at least for all systems in a wide class .
this paper provides provides a negative result in the case of two dimensional ( one degree of freedom ) hamiltonian systems , for which phase space volume is identical to area .
our main theorem shows that there are no computationally reasonable numerical integrators for which all hamiltonian systems of one degree of freedom can be integrated while conserving both area and energy . before proving this result
we define what we mean by a computationally reasonable integrator .
we then consider what obstructions this result places on the existence of volume- and energy - conserving integrators for hamiltonian systems with an arbitrary number of degrees of freedom .
ordinary differential equations , numerical integration , hamiltonian systems , geometric integration , no - go theorems , volume - conservation , energy - conservation 65p10 |
the observation of exotic stable massive particles ( smps ) with electric , magnetic or colour charge ( or combinations thereof ) would be of fundamental significance .
searches are therefore routinely made for smps bound in matter , and for smps at cosmic ray and collider experiments @xcite .
furthermore , smps with masses at the tev scale are anticipated in a number of beyond - the - standard - model scenarios , such as supersymmetry and universal extra dimensions @xcite .
smp searches are thus a key component of the early data exploitation program of the lhc experiments @xcite .
to facilitate such searches , models are needed of the interactions of smp as they pass through matter , since such interactions can give rise to a range of different event topologies @xcite .
although robust phenomenological approaches are available to predict the scattering of electromagnetically charged smps , it is not clear how the interactions of hadronic smps ( hereafter termed @xmath0-hadrons - hadron has its origin in the @xmath0-parity quantum number in supersymmetry theories and the work in this paper will be presented in the context of supersymmetric particles .
however , the results are generally applicable to stable heavy exotic hadrons . ] ) should be treated owing to uncertainties associated with the strong scattering processes and the hadronic mass hierarchy .
this paper describes an implementation within geant-4 @xcite and extension of a regge - based model @xcite for hadronic scattering .
the paper is a summary of work contained in ref .
@xcite . to aid the development of search strategies ,
it is important that experiments have access to monte carlo models which span as fully as possible the range of conceivable signatures which could be associated with @xmath0-hadron production . at present ,
one model is available within geant-4 @xcite .
this model , hereafter termed the _ generic model _ , is based on a black disk approximation @xcite and employs pragmatic assumptions regarding possible scattering processes and stable @xmath0-hadron species . this work concerns an approach , hereafter termed the _ regge model _ , which uses triple regge formalism to predict the scattering of @xmath0-hadrons formed from exotic colour triplet squarks ( stop - like ( @xmath2 ) and sbottom - like ( @xmath3 ) ) and electrically neutral gluino - like ( @xmath4 ) colour octet particles .
furthermore , the model uses a different set of well motivated assumptions regarding @xmath0-hadron mass hierarchies than those employed in the generic model .
although there exist no complete calculations of expected @xmath0-hadron mass spectra , estimates have been made using a variety of approaches such as the bag model @xcite and lattice qcd @xcite .
these predictions , together with with measurements of heavy hadron masses , make it possible to approximately determine those features of the mass hierarchies which are most relevant for the modelling of @xmath0-hadron scattering in material .
of particular interest are the masses of the lowest lying states , to which higher mass particles would be expected to decay before being able to interact hadronically .
given the observed spin and flavour independence of interactions of heavy quarks with light quarks @xcite the lightest meson and baryon states can be inferred @xcite from the measured mass spectra of charm and bottom hadrons . thus , charged and neutral squark - based mesons would be approximately mass degenerate and the lightest baryon state to which the other baryons would decay would be the scalar singlet @xmath5 i.e. the exotic equivalent of the @xmath6,@xmath7 baryons .
this picture is supported by calculations of hyperfine splittings of @xmath0-hadron masses @xcite .
the basic properties of gluino @xmath0-hadrons may not be similarly inferred from standard model hadronic mass spectra and a greater reliance on phenomenological approaches is required to estimate the stable states .
possible hadrons include gluino @xmath0-mesons + ( @xmath8 ) and gluino - gluon states ( @xmath9 ) .
bag model @xcite and lattice qcd @xcite calculations estimate that the masses of lowest lying states of each of these hadrons differ by less than a pion mass .
detailed calculations of the lowest lying baryon states @xcite using the bag model predict that the lightest state is the singlet @xmath10 . consequently
, higher mass states would decay weakly into the singlet state over a time scale of @xmath11s @xcite . for the work presented here
, we use the squark and gluino @xmath0-hadron mass hierarchy assumptions given above and use a simplified mass spectra which comprises only those @xmath0-hadrons deemed to be stable .
this leads to @xmath0-hadron event topologies which would be challenging for the one of the most common collider search methods , which looks for a slow muon - like object @xcite .
this is due to the @xmath0-baryon production processes which take place as @xmath0-hadron pass through matter .
such processes , which , as discussed in the next section , imply that a sbottom or gluino @xmath0-hadron , irrespective of its state as it entered a calorimeter , is likely to leave as an uncharged object and therefore escape detection in an outer muon chamber . since the central picture of a low energy light quark system interacting with a stationary nucleon @xcite , @xmath0-hadron scattering can be treated with the phenomenology used to describe low energy hadron - hadron scattering data @xcite , as is done in the regge model of which more details can be found in ref .
the regge model was originally developed to describe squark @xmath0-hadron scattering though has been extended here to also treat gluino @xmath0-hadrons .
this model assumes the stable states described in the previous section . using parameters
fitted to low energy hadron - hadron data , the regge model makes predictions for @xmath0-hadron scattering cross sections , together with energy loss calculations based on the triple regge formalism .
[ fig : fig1 ] ( left ) shows the the model predictions of the scattering cross sections of a squark - based @xmath0-hadron off a stationary nucleon within a nucleus comprising equal numbers of neutrons and protons .
the cross section is shown for different types of squark - based @xmath0-hadrons as a function of the lorentz factor @xmath12 . as can be seen
, there is a large cross section for antibaryon ( @xmath13 ) interactions which is due to a dominant annihilation process with a nucleon in the target .
it is also seen that , at lower values of @xmath12 ( @xmath14 ) , the scattering cross section of squark - based @xmath0-hadrons containing a light valence antiquark ( @xmath15 ) is larger than for antisquark - based @xmath0-mesons , which arises from the presence of reggeon exchange processes which are only permitted for @xmath0-hadrons containing a light anti - quark .
owing to the presence of an additional light quark the scattering cross section for @xmath0-baryons ( @xmath16 ) is twice as large as for the mesons with light quarks . upon an interaction , the probability that a @xmath0-meson to @xmath0-meson process gives
rise to charge exchange is 50% .
similarly , the probability of an interaction in which a @xmath0-meson becomes a baryon is 10% .
once a @xmath0-hadron becomes a baryon it stays in this state .
another process which must be taken into account is the oscillation of a neutral mesonic squark - based @xmath0-hadron into its antiparticle .
since the conversion rate would be model dependent we allow two possibilities here : zero mixing , in which no oscillations take place , and a maximal mixing scenario in which there is a 50% probability that any neutral mesonic squark - based @xmath0-hadron which was produced would automatically be converted to its anti - particle . in fig .
[ fig : fig1 ] ( right ) the predicted cross sections for gluino @xmath0-hadron species are shown .
the baryon cross section is 50% greater than that of stop @xmath0-hadrons owing to the extra light quark in the gluino @xmath0-hadron .
also , since the gluino meson contains a light quark and anti - quark , then both regge and pomeron exchange processes are available .
again , a baryon number transfer probability of 10% is used .
a probability of @xmath17 is assigned to @xmath0-meson to @xmath0-meson processes involving charge exchanges of 0 , @xmath18 or @xmath19 .
the predictions are shown for the regge model ( solid lines ) and the generic model ( dashed lines ) . ] in the dynamical picture of @xmath0-hadron scattering used by geant , the light quark system is decoupled from the heavy spectator parton before interacting with a nucleon according to geant s parameterised model for light hadrons .
in this way secondary particles and so - called black tracks were generated .
following the interaction , the light quark system is recombined with the heavy parton . to implement the regge model in geant-4 , the software architecture used for the generic model
was adapted .
the scattering cross sections and available baryon states were altered and the relative rates of the available processes were adapted to match the prescription in the above section . using single particle simulations
, it was verified that the main features ( energy loss and charge exchange ) of the regge model were well reproduced in the geant-4 implementation .
to estimate the effects of scattering on @xmath0-hadrons at the lhc simulations were made of the passage through iron of @xmath0-hadrons of 300 gev mass .
the kinematic distributions of the @xmath0-hadrons were given by the pythia generator which simulated the direct pair production of @xmath20 , @xmath21 and @xmath22 in proton - proton collisions at 14 tev centre - of - mass energy .
the total energy loss for @xmath0-hadrons passing through 1 m of iron is shown in fig . [
fig : fig2 ] .
stop @xmath0-hadrons which start in the state @xmath23 typically lose more energy than gluino @xmath0-hadrons ( @xmath24 ) which start with the same charge .
this occurs due to the rapid conversion of mesons to baryons .
since the gluino ( stop ) baryon is predicted to be of zero ( positive ) charge , it loses far less energy .
also shown is the number of hadronic interactions .
owing to the additional light quarks in the gluino @xmath0-hadrons more scattering occurs in this case than for the stop @xmath0-hadrons .
a no - mixing scenario is assumed here .
-hadrons of 300 gev mass sent through 1 m of iron .
the top ( bottom ) plots show predictions for stop ( gluino ) @xmath0-hadrons which entered the iron in the state @xmath23 ( @xmath24 ) . ]
an implementation and extension of a regge - based approach to squark and gluino - based @xmath0-hadron scattering has been implemented in geant-4 .
the work allows the modelling of the energy loss and the transformation of @xmath0-hadrons as they propagate through material .
slides containing the presented work can be found in @xcite
david milstead is a royal swedish academy research fellow supported by a grant from the knut and alice wallenberg foundation .
m. s. chanowitz and s. r. sharpe , phys .
b * 126 * , 225 ( 1983 ) .
g. r. farrar , phys .
* 53 * ( 1984 ) 1029 .
f. buccella , g. r. farrar and a. pugliese , phys .
b * 153 * ( 1985 ) 311 .
m. foster and c. michael [ ukqcd collaboration ] , phys .
d * 59 * ( 1999 ) 094509 [ arxiv : hep - lat/9811010 ] .
v. m. abazov _ et al . _
[ d0 collaboration ] , phys .
* 102 * ( 2009 ) 161802 [ arxiv:0809.4472 [ hep - ex ] ] . t. aaltonen _ et al . _
[ cdf collaboration ] , arxiv:0902.1266 [ hep - ex ] .
mermod , arxiv:0909.1911 [ hep - ex ] . | exotic stable massive particles ( smp ) are proposed in a number of scenarios of physics beyond the standard model .
it is important that lhc experiments are able to detect hadronic smp with masses around the tev scale . to do this
, an understanding of the interactions of smps in matter is required . in this paper a regge - based model of @xmath0-hadron scattering is extended and implemented in geant-4 . _
prepared for proceedings of 17@xmath1 international workshop on deep - inelastic scattering and related subjects , madrid , 2009 . _ |
this paper deals with the motion of three bodies in the plane where the only force being considered is the gravitational force .
we will assume that the units have been changed so that the gravitational constant is 1 , the mass of the first body is @xmath8 and the mass of the second body is @xmath9 .
moreover , we will assume that the first body moves according to the formula @xmath10 and the second body moves according to the formula @xmath11 .
the motion of these two bodies is known as the circular solution of the two body problem .
now , let us assume that the third body has a mass @xmath12 so small that it does not affect the motion of the other two bodies .
it follows that if the third body moves according to the formula @xmath13 , then the function @xmath14 and @xmath15 satisfy the following system of differential equations , @xmath16 the problem given by the differential equation above is known as the planar , circular , restricted three body problem , hereafter , pcr3bp .
we will be considering periodic solutions of the pcr3bp .
it is well - known that the period of a periodic solution of the pcr3bp must be a multiple of @xmath17 , @xcite .
when we change the coordinate system so that it rotates along with the motion of the first two bodies , that is , if we consider the functions @xmath18 and @xmath19 instead of @xmath14 and @xmath15 using the relations @xmath20 then , it follows that the functions @xmath18 and @xmath19 satisfy the following system of differential equations @xmath21 where , @xmath22 an important aspect of equation ( [ eq2 ] ) is that it is an autonomous system , and , for this reason , the period of a periodic solution does not have to be a multiple of @xmath17 . regarding the relation between the period of periodic solutions of both systems of differential equations we have that every periodic solution of the pcr3bp defines a periodic solution of equation ( [ eq2 ] ) , and a @xmath0-periodic solutions of equation ( [ eq2 ] ) defines a periodic solution of the pcr3bp only when @xmath23 is a rational number . a detailed discussion on the periodic solutions of equation ( [ eq2 ] ) can be found in @xcite .
a direct computation shows that for every solution of equation ( [ eq2 ] ) we have that @xmath24 where @xmath5 is a constant called the jacobian integral @xcite .
if follows that every solution of equation ( [ eq2 ] ) must be contained in the region @xmath25 we point out that since the function @xmath26 is not defined on the set @xmath27 , then , we are not considering these two points to be part of @xmath28 . in this paper
we prove that if a regular simple closed curve @xmath29 is a @xmath0 periodic solution of equation ( [ eq2 ] ) , and the region @xmath3 enclosed by this closed curve is contained in the region @xmath30 , then , for some integer @xmath2 we have that @xmath31 the choice of @xmath32 depends on the orientation of the periodic solution .
we extend the result to situations when the curve goes several times around the point @xmath33 or around the point @xmath34 .
the proofs of the main results in this paper are very basic since they follow from the stokes theorem . despite the simplicity of these results
, the author considers that they are interesting . the key result that allows to use the stokes theorem is lemma [ le ] .
the author found the formula in this lemma while he was trying to find an explicit solution of the restricted three body problem .
he was trying to find the tread - mill sled of the solution , which is a notion that he has used to explain : the profile curve of surfaces in the euclidean space with constant mean curvature and helicoidal symmetry @xcite , the profile curve of minimal surface in the euclidean space and helicoidal symmetry , @xcite , and the shape of ramp with the property that a block will slide down with constant speed @xcite .
also , along with bennet palmer , they used the treamill sled of a curve to study rotating drops @xcite , @xcite .
https://www.youtube.com/watch?v=5puywqd1xo8 displays a video that explains the notion of treadmill sled of a curve .
let us start this section with the following lemma .
[ le ] if @xmath35 is a regular curve that is a solution of equation ( [ eq2 ] ) , with jacobian integral equal to @xmath5 ( see equation ( [ je ] ) for the definition of @xmath5 ) , then , @xmath36 where @xmath37 satisfies the differential equation @xmath38 where @xmath39 is given by @xmath40 .
see equation ( [ jr ] ) for the definition of @xmath28 . since the curve @xmath41 is regular , @xmath42 never vanishes . by the definition of the jacobian integral we have @xmath43 , that is , @xmath44 .
therefore , for some smooth function @xmath45 , we have that @xmath46 taking the derivative with respect to @xmath47 of the two functions above we get @xmath48 using equation ( [ eq2 ] ) we obtain that @xmath49 multiplying equation ( [ l2 ] ) by @xmath50 and equation ( [ l1 ] ) by @xmath51 and adding them , we obtain the expression for @xmath52 given on equation ( [ dtheta ] ) the following lemma explain how is the behavior of the function @xmath53 near the points @xmath33 and @xmath34 [ l2 ] let @xmath39 be given by @xmath40 . if @xmath54 and @xmath55 , then @xmath56 likewise , if @xmath57 and @xmath55 , then @xmath56 the proof is a direct computation .
[ mtp1 ] let us assume that @xmath35 is a solution of equation ( [ eq2 ] ) , with jacobian integral equal to @xmath5 .
let us further assume that * @xmath58 is periodic , this is , there exists a positive number @xmath0 such that @xmath59 for all @xmath60 .
* @xmath58 is regular , this is , @xmath61 for all @xmath60 .
* @xmath58 is simple , that is , for any @xmath62 is injective . *
the region @xmath3 enclosed by the curve @xmath41 is contained in the set @xmath28 .
we have that * if @xmath41 goes clockwise around the region @xmath3 then @xmath63 * if @xmath41 goes counterclockwise around the region @xmath3 then @xmath64 if @xmath41 goes clockwise around the region @xmath3 , then the normal outer vector of the region @xmath3 is given by @xmath65 and the total change of the angle @xmath45 is @xmath66 . by the stokes theorem we have that ,
@xmath67 if @xmath41 goes counterclockwise around the region @xmath3 , then the normal outer vector of the region @xmath3 is given by @xmath68 and the total change of the angle @xmath45 is @xmath69 . by stokes
theorem we have that , @xmath70 [ mt2 ] let us assume that @xmath35 is a solution of equation ( [ eq2 ] ) , with jacobian integral equal to @xmath5 .
let us further assume that * @xmath58 is periodic , this is , there exists a positive number @xmath0 such that @xmath59 for all @xmath60 .
* @xmath58 is regular , this is , @xmath61 for all @xmath60 .
* @xmath58 is simple , that is , for any @xmath62 is injective . *
the region @xmath3 enclosed by the curve @xmath41 contains the point @xmath33 and is contained in the set @xmath71 .
we have that * if @xmath41 goes clockwise around the region @xmath3 then @xmath72 * if @xmath41 goes counterclockwise around the region @xmath3 then @xmath73 if @xmath41 goes clockwise around the region @xmath3 , then , the normal outer vector of the region @xmath3 is given by @xmath65 and the total change of the angle @xmath45 is @xmath66 . for any small number @xmath74 such
that the disk @xmath75 with center @xmath33 and radius @xmath74 is contained in @xmath3 we define @xmath76 .
if we parametrize the boundary of @xmath75 by @xmath77 , then the outer normal vector of the region @xmath78 along the boundary of @xmath75 is @xmath79 . by the stokes theorem we have that ,
@xmath80 if @xmath41 goes counterclockwise around the region @xmath3 , then , the normal outer vector of the region @xmath3 is given by @xmath68 and the total change of the angle @xmath45 is @xmath69 . by stokes
theorem we have that , @xmath81 a direct computation , see lemma [ l2 ] , shows that @xmath82 therefore we obtain that @xmath83 a similar argument shows the case when the curve @xmath41 goes counterclockwise around the region @xmath3 .
similar arguments show the following theorem .
we will omit the proof due to the similarities with the previous one .
[ mt3 ] let us assume that @xmath35 is a solution of equation ( [ eq2 ] ) , with jacobian integral equal to @xmath5 .
let us further assume that * @xmath58 is periodic , this is , there exists a positive number @xmath0 such that @xmath59 for all @xmath60 .
* @xmath58 is regular , this is , @xmath61 for all @xmath60 .
* @xmath58 is simple , that is , for any @xmath62 is injective .
we have that * if the region @xmath3 enclosed by the curve @xmath41 contains the point @xmath34 and is contained in the set @xmath84 , and @xmath41 goes clockwise around the region @xmath3 , then @xmath72 * if the region @xmath3 enclosed by the curve @xmath41 contains the point @xmath34 and is contained in the set @xmath84 , and @xmath41 goes counterclockwise around the region @xmath3 , then @xmath73 * if the region @xmath3 enclosed by the curve @xmath41 contains the points @xmath33 and @xmath34 and is contained in the set @xmath85 , and @xmath41 goes clockwise around the region @xmath3 , then @xmath86 * if the region @xmath3 enclosed by the curve @xmath41 contains the points @xmath33 and @xmath34 and is contained in the set @xmath87 , and @xmath41 goes counterclockwise around the region @xmath3 , then @xmath88
in this section we consider some special orbits that go around more than once before they closed .
let us start this section with a definition that clarifies the type of orbits that we will be considering .
[ nsimple ] we will say that a closed curve @xmath89\to \mathbb{r}^2 $ ] , with @xmath90 , is @xmath91-simple around the point @xmath92 if there exists a simple regular curve @xmath93\to\mathbb{r}^2 $ ] , with @xmath94 such that @xmath95 we will call the curve @xmath96 the lifting of @xmath41 .
let us assume that @xmath35 is a solution of equation ( [ eq2 ] ) , with jacobian integral equal to @xmath5 .
let us further assume that @xmath41 is @xmath0-periodic and @xmath91 simple around @xmath97 with lifting @xmath96 ( see definition [ nsimple ] ) . if we denote by @xmath98 , @xmath99 the region enclosed by @xmath100 , and @xmath101 where @xmath102 , and @xmath103 we have that * if @xmath96 goes clockwise around @xmath99 and @xmath104 is contained in @xmath71 , then + @xmath105 * if @xmath96 goes counterclockwise around @xmath99 and @xmath104 is contained in @xmath71 , then + @xmath106 let us denote by @xmath107 and @xmath108 the two entries of the curve @xmath100 , that is , @xmath109 .
likewise , we define @xmath110 and @xmath111 and @xmath112 , @xmath113 .
notice that if we define @xmath114 , then , @xmath115 . using the change rule
we obtain that @xmath116 once again , using the chain rule we obtain that @xmath117 since the function @xmath118 is an orientation - preserving conformal map , then @xmath112 and @xmath113 satisfy the cauchy - riemann equations , that is , @xmath119 combining the equations ( [ ep1])-([ep2 ] ) we get that @xmath120 let us denote by @xmath75 a disk with radius epsilon and center at @xmath33 contained in @xmath99 and @xmath121 . notice that if we parametrize the boundary of @xmath75 by @xmath77 , then the outer normal vector of the region @xmath122 along the boundary of @xmath75 is @xmath79 .
if @xmath96 goes clockwise around @xmath99 , by stokes theorem we have that @xmath123 taking the limite when @xmath74 goes to zero we obtain that @xmath124 a similar argument proves the case when @xmath96 goes counterclockwise around @xmath99 a similar result holds when the orbit goes around @xmath34 .
we have [ wi ] let us assume that @xmath35 is a solution of equation ( [ eq2 ] ) , with jacobian integral equal to @xmath5 .
let us further assume that @xmath41 is @xmath0-periodic and @xmath91 simple around @xmath34 with lifting @xmath96 ( see definition [ nsimple ] ) . if we denote by @xmath98 , @xmath99 the region enclosed by @xmath100 and @xmath125 where @xmath102 , @xmath126 we have that * if @xmath96 goes clockwise around @xmath99 and @xmath104 is contained in @xmath84 , then + @xmath127 * if @xmath96 goes counterclockwise around @xmath99 and @xmath104 is contained in @xmath84 , then + @xmath128 the proof follows the same arguments as those in the previous theorem .
once again , the same argument can be applied when the orbit goes around any given point , in particular one of the lagrangian points .
we have , [ wis ] let us assume that @xmath35 is a solution of equation ( [ eq2 ] ) , with jacobian integral equal to @xmath5 .
let us further assume that @xmath41 is @xmath0-periodic and @xmath91 simple around @xmath129 with lifting @xmath96 ( see definition [ nsimple ] ) . if we denote by @xmath98 , @xmath99 the region enclosed by @xmath100 and @xmath125 where @xmath102 , @xmath130 we have that * if @xmath96 goes clockwise around @xmath99 and @xmath104 is contained in @xmath28 , then + @xmath131 * if @xmath96 goes counterclockwise around @xmath99 and @xmath104 is contained in @xmath28 , then + @xmath132
in this section we provide numerically periodic solutions of equation ( [ eq2 ] ) and we use them to verify some of the theorems in section [ mt ] and section [ s3 ] . in order to somehow give these examples a real application
, we took @xmath133 , which is usually the value considered for the system sun - jupiter - asteroid .
@xcite let us consider the initial conditions for equation ( [ eq2 ] ) @xmath134 a direct verification shows that within an error of @xmath135 , the solution @xmath58 is periodic with period @xmath136 .
in this case the jacobian integral is @xmath137 .
figure [ fig1 ] shows the orbit of this periodic motion . ]
a numerical computation shows that @xmath138 and therefore @xmath139 while @xmath140 .
this example verifies theorem [ mtp1 ] when the orbit goes clockwise .
let us consider the initial conditions for equation ( [ eq2 ] ) @xmath141 a direct verification shows that within an error of @xmath135 , the solution @xmath58 is periodic with period @xmath142 .
in this case the jacobian integral is @xmath143 .
figure [ fig2 ] shows the orbit of this periodic motion . ]
numerical approximation shows that @xmath144 and therefore @xmath145 while @xmath146 .
this example verifies theorem [ mt3 ] when the orbit goes counterclockwise and it goes around jupiter .
let us consider the initial conditions for equation ( [ eq2 ] ) @xmath147 a direct verification shows that within an error of @xmath148 , the solution @xmath58 is periodic with period @xmath149 . in this case
the jacobian integral is @xmath150 .
figure [ fig3 ] shows the orbit of this periodic motion . ]
numerical approximation shows that @xmath151 while @xmath152 .
this example verifies theorem [ mt3 ] when the orbit goes clockwise and it goes around the sun and jupiter .
let us consider the initial conditions for equation ( [ eq2 ] ) @xmath153 a direct verification shows that within an error of @xmath148 , the solution @xmath58 is periodic with period @xmath154 . in this case
the jacobian integral is @xmath155 .
figure [ fig4 ] shows the orbit of this periodic motion . .
this orbit @xmath41 goes three times around the sun before closing .
also this orbit is @xmath156-simple around the sun because the curve @xmath96 shown on the right has the property that @xmath157 . in the last expression we had identified @xmath158 with the complex number @xmath159 .
this figure also shows the region @xmath99 in @xmath160 where we integrate on theorem [ wi ] .
notice that @xmath100 is the lifting of @xmath41 .
, title="fig:"].2 cm .
this orbit @xmath41 goes three times around the sun before closing .
also this orbit is @xmath156-simple around the sun because the curve @xmath96 shown on the right has the property that @xmath157 . in the last expression
we had identified @xmath158 with the complex number @xmath159 .
this figure also shows the region @xmath99 in @xmath160 where we integrate on theorem [ wi ] .
notice that @xmath100 is the lifting of @xmath41 .
, title="fig : " ] numerical approximation shows that @xmath161 while @xmath162 .
therefore @xmath163 as predicted by theorem [ wi ] when the orbit goes counterclockwise around the sun more than once .
recall that the function @xmath164 .
we compute the function @xmath165 by first finding the function @xmath166 and @xmath167 sucht that @xmath168 and @xmath169 .
once we have @xmath170 and @xmath166 , we obtain that @xmath171
the main theorems from this paper provide estimates for the period of the periodic orbits and also allow us to provide necessary conditions for the orbit to be either clockwise or counterclockwise .
example [ ex1 ] showes an orbit near the lagrangian point @xmath6 orbiting clockwise .
the following theorem shows that this happens not only for the sun - jupiter system but for any other value of @xmath172 . for any @xmath172
, there exists an @xmath74 such that every periodic solution @xmath35 of the differential equation ( [ eq2 ] ) with jacobian integral different from @xmath173 , contained in the disk @xmath75 with center in @xmath174 must orbit clockwise . the same result holds true for @xmath175 we point out that @xmath177 is a critical value of the function @xmath178 : we can easily check that @xmath179 , @xmath180 and @xmath6 is a local minimum .
if @xmath181 then by the continuity of @xmath26 , there exists a disc @xmath75 centered at @xmath6 such that @xmath182 . since @xmath183 , no point in @xmath58 can be in the disk @xmath75
now , if @xmath184 , then by equation ( [ value of h ] ) we have that @xmath185 and using the continuity of @xmath186 , we can find a disk @xmath75 centered at @xmath6 such that @xmath187 is positive . by theorems [ mtp1 ] and [ wis ]
, we conclude that the periodic solution @xmath41 must move clockwise .
the same argument proves the result for @xmath7 . | we will show that the period @xmath0 of a closed orbit of the planar circular restricted three body problem ( viewed on rotating coordinates ) depends on the region it encloses . roughly speaking ,
we show that , @xmath1 where @xmath2 is an integer , @xmath3 is the region enclosed by the periodic orbit and @xmath4 is a function that only depends on the constant @xmath5 known as the jacobian integral ; it does not depend on @xmath3 .
this theorem has a keplerian flavor in the sense that it relates the period with the space `` swept '' by the orbit . as an application
we prove that there is a neighborhood around @xmath6 such that every periodic solution contained in this neighborhood must move clockwise .
the same result holds true for @xmath7 . |
in our previous papers @xcite , we calibrated the relation between peak energy @xmath0 and peak luminosity @xmath1 of prompt grb emission ( so called yonetoku relation @xcite ) by 33 low - redshift grbs ( @xmath14 ) whose luminosity distances were estimated from sneia @xcite .
then we used the calibrated yonetoku relation as a distance indicator like the period - luminosity relation of cepheid variables , and extended the hubble diagram up to @xmath15 .
@xcite showed that grbs constrain cosmological parameters in a different way from sneia , and grbs could be useful to probe cosmological expansion of high - redshift universe where no snia has been observed .
the amati relation @xcite is another relation for prompt emission property .
it involves peak energy @xmath0 and isotropic energy @xmath16 , and was originally derived under a given set of cosmological parameters .
therefore , the circularity problem arises if one applies naively the amati relation to determine cosmological parameters . to overcome this difficulty , in this paper , we first calibrate the amati relation as we did for yonetoku relation without assuming any cosmological models but using luminosity distance given by sneia for @xmath7 . although the strong correlation between @xmath2 and @xmath1 are generally confirmed , we found little trend between their data residuals of @xmath17 and @xmath18 from the best fit function of each relation .
we suggest statistical independence of amati and yonetoku relations while these two relations have the same parameter @xmath0 ( -2 ) .
next , we will extend the hubble diagram with the obtained relations and make constraints on density parameters , ( @xmath19 ) .
it is shown that the two hubble diagrams differ systematically at high redshifts and , as a result , two different constraints are obtained ( -3 ) .
although the difference is not so significant ( 1-@xmath8 level ) , taking the relatively large systematic errors in the two relations themselves into consideration , it is suggested that there may be a hidden parameter which characterizes the prompt emission and reduces the systematic error of the distance indicator .
we introduce the luminosity time ( @xmath10 ) as a possible hidden parameter to derive the @xmath0-@xmath9-@xmath1 relation , and put constraints on cosmological parameters ( -4 ) .
finally we give some comments on other relations used as distance indicators in the past , and argue the advantage of our new relation ( -5 ) . throughout the paper ,
we fix the current hubble parameter as @xmath20 .
we note that , in whole this paper , tentatively we do not take possible selection effects and evolution effects on relations into accounts although a hot debate @xcite exists .
we think that the final resolution needs the increase of the events with enough well determined parameters such as redshifts , @xmath0 , @xmath1 and @xmath21 since we should divide all data into several sub groups to check possible selection effects and evolution effects .
the typical spectrum of the prompt emission of grbs can be expressed as exponentially connected broken power - law , so called band function @xcite
. then we can determine spectral peak energy @xmath0 , corresponding to the photon energy at maximum in @xmath22 spectra .
there are two empirical relations that relate prompt emission property with @xmath0 .
@xmath0-@xmath2 relation is the first one found by @xcite , which connects @xmath0 with the isotropic equivalent energy @xmath2 .
the second one is @xmath0-@xmath1 relation found by @xcite which was used in our previous papers @xcite .
we first calibrate amati relation in the same way as in our previous papers @xcite and analyze the correlation of the residuals of grb data from the relations and the partial correlation coefficient .
if there are a small degree of correlation between @xmath1 and @xmath3 after removing the effect of @xmath0 , it suggest the possible independence of the distance indicators , and might suggest the existence of a hidden parameter common to each relation .
we found an empirical formula for the luminosity distance as a function of redshift from 192 sneia observations @xcite , @xmath23 the reduced chi - square of the formula is @xmath24 .
note here that the formula is not unique and a different formula is possible .
note also that we do not assume any cosmological models at this stage , but simply assume that the type ia supernovae are the standard candles for @xmath25 .
furthermore , we neglect the errors in eq .
( [ eq : dl_sn ] ) in the following analysis , which leads to the underestimation of errors in cosmological parameters .
our purpose here is to compare distance indicators and find a better indicator so that this neglect would be reasonable .
we apply this formula to 31 low redshifts grbs within @xmath26 .
( for details of our data , see @xcite . ) in fig . [ fig1 ]
we show the peak energy @xmath0 and the isotropic energy @xmath16 of 31 grbs with @xmath14 .
the solid line is the calibrated amati relation given by , @xmath27 where the statistical errors are indicated and @xmath28 where @xmath29 is the bolometric fluence estimated in 1 - 10000 kev energy range in grb rest frame .
the pearson correlation coefficient is 0.943 and the reduced chi - square is @xmath30 with the systematic error @xmath31 .
here we include not only errors in @xmath3 but also errors in @xmath0 , and the systematic error of this relation @xmath32 , so the chi - square function is defined as @xmath33 where the weighting factor @xmath34 .
the value of @xmath32 can be estimated by the value such that a @xmath35 fit to the amati relation calibration produces a reduced @xmath35 of unity .
this kind of systematic error was also found in the yonetoku relation @xcite , @xmath36 where @xmath1 is 1-second peak luminosity .
this yonetoku relation is slightly different from that in the previous work because we include not only @xmath1 error but also @xmath0 error , and the systematic error in this paper .
the pearson correlation coefficient is 0.948 , the reduced chi - square is @xmath37 , and the systematic error is @xmath38 .
the amati relation has slightly larger systematic error than the yonetoku relation and , in both cases , systematic errors are significantly larger than measurement errors . ) and isotropic energy ( @xmath16 ) of 31 grbs with @xmath14 .
the solid line is the calibrated amati relation given by eq ( [ eq : amati ] ) and dashed lines represent the 1-sigma region .
the systematic error of this relation is reduced to @xmath31 which is larger than measurement errors.,height=347 ] fig .
[ fig2 ] shows the correlation between two residuals which are defined as @xmath39 and @xmath40 . here , @xmath41 ( @xmath42 ) is the observed value , and @xmath43 ( @xmath44 ) is the expected quantity from yonetoku and amati relations for observed @xmath0 , respectively .
we can see that there is no correlation between @xmath45 and @xmath46 .
we performed a linear partial correlation test for these 31 samples , and found the partial correlation coefficient is quite small of @xmath4 . here
@xmath47 means the correlation coefficient between the first and the second parameters after fixing the third parameter which is controlling two correlations ( the parameter 13 and 23 relations ) .
this fact indicates that two distance indicators may be independent from each other . in the next section ( -3 )
, we investigate whether the cosmological parameters measured with two independent relations are consistent or not . after that , in
-4 , we explore a hidden parameter which is the cause of the intrinsic dispersion of amati and yonetoku relations . .
here , the residuals are the differences between the observed quantities , @xmath48 and @xmath42 , and the expected quantities , @xmath43 and @xmath44 , from yonetoku and amati relations for the given observed @xmath0 , respectively .
there seems to be no correlation , which implies the two relations are independent distance indicators.,height=317 ]
we apply these amati and yonetoku relations to 29 grbs with high redshifts , @xmath49 , to determine the luminosity distance as a function of @xmath50 .
we include not only measurement errors of @xmath3 ( @xmath1 ) , and @xmath0 but also systematic errors we estimated in -2 and the 1-sigma uncertainties in eq .
( [ eq : amati ] ) ( eq .
( [ eq : yonetoku ] ) ) to estimate luminosity distance @xcite . for example , in case of amati relation , we define @xmath51 and @xmath52 as below , @xmath53.\nonumber\end{aligned}\ ] ] fig . [ fig3 ] shows an extended hubble diagram up to @xmath15 from amati relation ( red ) and yonetoku relation ( blue ) .
a systematic difference between red and green points seems to exist especially in high - redshift region
. then we derive constraints on cosmological parameters . in the @xmath54-cdm model with @xmath55 ,
the luminosity distance is given by , @xmath56 with @xmath57^{-1/2}. \label{f(z)}\ ] ] the likelihood contour is defined by , @xmath58 where @xmath59 and @xmath60 represents the chi - square value for the best - fit parameter set of @xmath5 and @xmath6 . in fig .
[ fig4 ] , we show the likelihood contour from amati ( red ) and yonetoku relations ( blue ) , and the best - fit values with 1-@xmath8 errors are shown in table [ tab1 ] .
interestingly , they are slightly different , although they are consistent in 2-@xmath8 level . )
plane from amati relation ( red ) and yonetoku relation ( blue ) .
the contours correspond to 68.3% confidence regions , and black solid line represents the flat universe .
they are slightly different , although they are consistent in 2-@xmath8 level .
see also table [ tab1].,width=264 ] .constraints on ( @xmath19 ) in non - flat and flat universe with 1-@xmath8 errors from amati , yonetoku and @xmath0-@xmath9-@xmath1 relations and their reduced chi squares .
constraints from amati and yonetoku relations are inconsistent in 1-@xmath8 level . [
cols="^,^,^,^",options="header " , ]
in this section , we seek for a hidden parameter which reduces the systematic error in the relation .
note that the discussion in this section does not concern with independence of the two relations .
if we could find this parameter , we can make strong constraint on cosmological parameters , because the systematic errors in the amati and yonetoku relations are larger than their measurement errors .
@xcite there are some studies using jet break time in afterglow as such a parameter @xcite , but the number of grbs which have observed jet is small . here
we restrict our discussion in a prompt emission property . in the past studies ,
prompt emission was characterized by a time scale , the duration of most intense parts of the grb ( @xmath61 ) @xcite .
here we adopt a time scale called luminosity time @xmath9 introduced by @xcite as , @xmath62 the luminosity time does not depend on detector s energy band because it is defined by bolometric flux and fluence .
we assume the correlation among @xmath0 , @xmath9 and @xmath1 to be of the form , @xmath63 .
then we obtain , @xmath64 from low redshift 30 grbs in @xmath65 . in fig .
[ fig5 ] we show @xmath0-@xmath9-@xmath1 relation .
the pearson correlation coefficient is 0.971 , the reduced chi - square is @xmath66 with the systematic error @xmath67 .
we include not only errors in @xmath68 but also errors in @xmath0 , and @xmath9 so the chi - square function is defined as @xmath69 where the weighting factor @xmath70 , and @xmath71 is estimated by using the error propagation equation without a crossterm between @xmath1 and @xmath3 .
the factor 2c in the front of @xmath72 comes from the fact that the definition of @xmath9 includes @xmath1 .
note that the contribution of the additional error term @xmath73 to chi - square value is little because of its small slope .
thus , we conclude that the additional term really improves relations .
the systematic error is substantially reduced compared to those of amati and yonetoku relations , and now comparable to the measurement error .
thus this relation could be regarded as `` fundamental plane '' @xcite of grb prompt emission .
here we excluded one outlier , grb070521 , from the fitting of eq .
( [ eq : new ] ) .
actually , the host galaxy of grb070521 is detected inside an error circle of xrt by @xcite using subaru telescope after 40 minutes from the trigger , but they could nt detect bright afterglow .
thus , the real redshift may be larger , which is why we exclude grb070521 from our analysis .
note that the probability of the miss identification of the host galaxy from only xrt observation is about 7%@xcite .
futhermore there might be another population of grbs so that the secure classification of grbs is one of the current issue of grb s study . for more detailed discussions about @xmath0-@xmath9-@xmath1 relation ,
see @xcite . and
@xmath1 in 31 grbs with @xmath14 .
the correlation is improved than amati and yonetoku relations .
the solid line shows the best - fit curve without one outlier ( green square : grb070521 ) , and dashed lines represent the 1-sigma region .
the pearson correlation coefficient is 0.971 and reduced chi - square is @xmath66 with the systematic error @xmath74 which is comparable to the measurement error.,height=317 ] we show a possible derivation of the new relation ( eq.([eq : new ] ) under the photospheric model of the prompt emission of grbs ( @xcite and references therein ) .
the luminosity is given by @xmath75 where @xmath76 , @xmath77 and @xmath78 are the photospheric radius , the gamma factor and co - moving temperature of the photosphere , respectively .
since @xcite assume that the energy is supplied by the relativistic collision of the rapid shell of mass @xmath79 and lorenz factor @xmath80 with the slow shell of @xmath81 and @xmath82 .
then under the perfectly inelastic collision model , @xmath77 is given by @xmath83 . since @xmath84 and @xmath85
, we can reduce @xmath86 . if we regard @xmath87 , we can rewrite eq.([eq : photoshere ] ) as @xmath88 now let us assume @xmath76 , @xmath81 and @xmath89 are constants .
then we have @xmath90 the above relation is essentially the same as eq.([eq : new ] ) if we consider 2-@xmath8 error of the power index in eq.([eq : new ] ) .
possible reasons for the above three assumptions are as follows . in @xcite model
@xmath76 is similar to the radius of the progenitor star so that it could be constant .
@xmath9 can be regarded as the effective duration of the burst .
then @xmath91 is the radius that the last rapid shell catches the slow shell and we expect that this is also the order of the radius of the progenitor star , which is constant .
we have no reason why @xmath81 is constant .
however if @xmath81 obeys the log normal distribution like the other observables in grbs we may regard it essentially constant .
if these assumptions are reasonable , the new relation ( eq.([eq : new ] ) ) could be derived in the photospheric model of the prompt emission of grbs like @xcite .
finally we use this new relation to put constraints on cosmological parameters . here
again we include not only measurement errors of @xmath1 , @xmath0 , @xmath9 but also systematic errors and the 1-sigma uncertainties in eq.([eq : new ] ) to estimate luminosity distance @xcite .
the concordance cosmology is still consistent in 1-@xmath8 level .
the constraints on cosmological parameters are @xmath92 and @xmath93 for non - flat universe , @xmath94 and @xmath95 for flat universe .
( see table [ tab1 ] . ) ) plane from @xmath0-@xmath9-@xmath1 relation .
see also table [ tab1].,width=317 ]
recently some authors extended hubble diagram up to @xmath96 using various luminosity indicators @xcite .
@xcite , @xcite , and @xcite are pioneering works for grb cosmology , but they are caught in circularity problem because there are few grbs at low redshift .
@xcite obtained cosmological constraints from lag ( @xmath97)-luminosity relation , variability ( @xmath98)-luminosity relation , @xmath0-jet collimated energy ( @xmath99 ) relation ( so called ghirlanda relation ) , minimum rise time ( @xmath100)-luminosity relation , and yonetoku relation .
@xcite calibrated these relations by luminosity distances from snia observations .
however , @xcite showed redshift dependence of @xmath97-@xmath1 relation analyzing 565 baste grb samples with pseudo - redshifts estimated by yonetoku relation .
this suggests that the relation can not be used as a distance indicator .
the dependence might come from the fact that @xmath101s were evaluated from several fixed energy bands depending on the detectors and they are different energy in grb rest frame , so it suffer from k - correction problem .
this argument applies to @xmath98s and @xmath100s .
although @xmath0-@xmath99 relation and liang & zhang relation have much smaller systematic errors than both yonetoku and amati relations , there are many grbs without a jet break or with multiple jet breaks ( missing or multiple jet break problem ) so that it is not certain whether the jet break can be used to characterize grb emission .
in contrast , our new relation does not suffer from these problems , because it is totally defined by the prompt emission property and it has as a small systematic error as @xmath0-@xmath99 and liang & zhang relations .
thus we could expect that the new relation would put grb cosmology to the next promising stage , as phillips relation @xcite and fundamental plane @xcite did for sneia and elliptical galaxies .
however , we must emphasize that we need detailed studies of the new relation with much larger number of grbs and examination of systematic errors in order for grb to be regarded as a reliable tool for cosmology like snia , cosmic microwave background , baryon acoustic oscillation and gravitational lens . now
ongoing missions like , and , and the collaboration of many observer on ground will promise the progression of grb cosmology .
this work is supported in part by the grant - in - aid from the ministry of education , culture , sports , science and technology ( mext ) of japan , no.19540283 , no.19047004(tn ) , and no.20674002 ( dy ) and by the grant - in - aid for the global coe program _ the next generation of physics , spun from universality and emergence _ from mext of japan .
kt is supported in part by a grant - in - aid from the japan society for the promotion of science ( jsps ) fellows and the global coe programs , `` quest for fundamental principles in the universe : from particles to the solar system and the cosmos '' at nagoya university .
bevington , p. r. , & robinson , d. k. 2003 , data reduction and error analysis for the physical sciences , 3rd ed .
, by philip r. bevington , and keith d. robinson .
boston , ma : mcgraw - hill , isbn 0 - 07 - 247227 - 8 , 2003 . , | we consider two empirical relations using data only from the prompt emission of gamma - ray bursts ( grbs ) , the peak energy ( @xmath0 ) - peak luminosity ( @xmath1 ) relation ( so called yonetoku relation ) and the @xmath0-isotropic energy ( @xmath2 ) relation ( so called amati relation ) .
both relations show high correlation degree , but they also have larger dispersion around the best fit function rather than the statistical uncertainty .
then we first investigated the correlation between the residuals of @xmath1 and @xmath3 from the best function , and found that a partial linear correlation degree is quite small of @xmath4 .
this fact indicates that some kinds of independence may exist between amati and yonetoku relation even if they are characterized by the same physical quantity @xmath0 , and similar quantities @xmath1 and @xmath3 which mean the brightness of the prompt emission
. therefore we may have to recognize two relations as the independent distance indicators . from this point of view , we compare constraints on cosmological parameters , @xmath5 and @xmath6 , using the yonetoku and the amati relation calibrated by low - redshift grbs with @xmath7 .
we found that they are different in 1-@xmath8 level , although they are still consistent in 2-@xmath8 level . in this paper , we introduce a luminosity time @xmath9 defined by @xmath10 as a hidden parameter to correct the large dispersion of the yonetoku relation .
a new relation is described as @xmath11 .
we succeeded in reducing the systematic error about 40% level , and might be regarded as `` fundamental plane '' of grbs .
we show a possible radiation model for this new relation . finally ,
applying the new relation to high - redshift grbs with @xmath12 , we obtain @xmath13 , which is consistent with the concordance cosmological model .
[ firstpage ] gamma rays : bursts gamma rays : observation |
reionisation - the change of state of the intergalactic medium ( igm ) from a neutral state at recombination to the present day ionised state - is an epoch which is currently under intense study .
analyses of the spectra of high redshift qsos indicate a smooth increase in the optical depth of the igm with redshift which accelerates at @xmath3 ( fan et al .
2002 , songaila 2004 , fan et al .
quasars at @xmath3 show almost complete absorption of the flux blue - ward of the lyman@xmath4 resonance [ the gunn - peterson ( 1965 ) effect ] . inferring
the ionisation state of the igm at these high redshifts from qso absorption spectra is difficult and ambiguous ( e.g. oh & furlanetto 2005 , bolton & haehnelt 2006 ) .
if the small amounts of flux transmitted just blue - ward of the lyman@xmath4 centre are attributed to the presence of a locally ionised region embedded within a significantly neutral medium , then lower limits on the neutral fraction of @xmath5 may be inferred ( wyithe & loeb 2004a ; mesinger & haiman 2004 ) .
malhotra & rhoads ( 2005 ) , however , argue that the number density of lyman-@xmath6 emitters can be used to place a lower limit on the ionised fraction of gas at @xmath7 at @xmath8 per cent .
fan et al .
( 2006 ) also come to the conclusion that the igm probed by the highest redshift qsos is predominantly ionised .
the optical depth to thompson scattering inferred from the recent wmap measurements offers an independent constraint on the integrated reionisation history .
interpreted as a sudden reionisation at a redshift @xmath9 the year three wmap results give @xmath10 ( spergel et al .
2006 ) .
the emerging picture is consistent with the reionisation of cosmic hydrogen being complete by @xmath11 and therefore significant contribution to reionisation by sources beyond current detection limits .
experiments to detect the redshifted 21 cm signature of high redshift neutral hydrogen will potentially provide copious amounts of information on the epoch of reionisation .
the european design for a low frequency interferometer , lofar , may be in operation as early as 2007 and other low - frequency experiments ( see [ prospects ] ) are in various stages of planning and development .
however , mapping the high redshift neutral gas distribution in 21 cm will be extremely challenging . the redshifted 21 cm signal due to ionised regions produced by bright
, high redshift qsos surrounded by an otherwise still neutral intergalactic medium may be our best bet for an detection of individual objects due to their 21 cm signature ( see tozzi et al .
2000 for a detailed discussion ) .
kohler et al .
( 2005 ) used numerical simulations to show that ionised regions around high redshift qsos are indeed likely to be the most prominent features in the high redshift neutral gas distribution .
wyithe , barnes & loeb ( 2005 ) use a semi - analytic model for the high redshift qso hii region distribution and predict detectable hii regions at @xmath12 .
however , the growth history of high redshift supermassive bhs is very uncertain . in this paper
we will investigate a range of models for high redshift bh growth which reproduce the bright end of the luminosity function of redshift @xmath0 qsos and sample the range of plausible growth histories at higher redshift .
we begin by reviewing the physics of detecting a 21 cm signal due to neutral hydrogen ( [ 21 cm ] ) . in section
[ qlf_models ] we describe our models for the evolution of the qso luminosity function ( qlf ) .
we discuss the expected size of hii regions due to qsos in [ hii_regions ] . finally , in [ prospects
] , we compare our computed size distributions to the expected sensitivity of lofar to extended sources and discuss observational prospects . throughout this paper
we adopt a cosmological matter density @xmath13 , baryonic matter density @xmath14 , cosmological constant @xmath15 , present day hubble constant @xmath16 km s@xmath17 mpc@xmath18 km s@xmath17 mpc@xmath17 , a mass variance on scales of @xmath19 mpc @xmath20 and a scale invariant primordial power spectrum ( slope @xmath21 ) .
a ground state neutral hydrogen atom can emit or absorb at a wavelength of 21 cm corresponding to the energy of the transition between the singlet and triplet hyperfine levels . at the low frequencies of this transition ,
the specific intensity of the emission is proportional to the brightness temperature of the source @xmath22 as described by the the raleigh - jeans law , @xmath23 the brightness temperature of a typical patch of sky at redshift @xmath24 depends on the optical depth of the gas to 21 cm radiation , the temperature of the cosmic microwave background @xmath25 and the spin temperature of the gas , @xmath26 where @xmath27 depends on the density and ionisation state of the gas .
this results in a differential brightness temperature , @xmath28^{1/2}.\end{aligned}\ ] ] significantly neutral hydrogen gas for which @xmath29 differs from @xmath25 will emit or absorb at 21 cm in contrast to the cmb .
the value of @xmath29 is determined by the kinetic temperature of the igm , @xmath30 , and the strength with which @xmath29 and @xmath30 are coupled via collisional ionisation and/or lyman@xmath4 pumping .
@xmath30 is expected to decouple from @xmath25 as early as @xmath31 as the timescale for compton scattering off residual free electrons becomes longer than the expansion timescale ( e.g. scott & rees 1990 ) . in the absence of a heat source
, the igm cools adiabatically below @xmath25 and therefore the very first ( population - iii ) sources are expected to be born in a cold igm .
the redshift range over which hii regions will appear in absorption is uncertain ; if @xmath29 couples to @xmath30 prior to significant long range x - ray heating there may be a significant absorption epoch .
furlanetto ( 2006 ) argues that if the ( thermal ) radiation from pop - ii stars dominate reionisation then this will be the case , but if pop - iii stars were dominant in reionising the universe then the igm could be heated above @xmath25 before there are sufficient lyman@xmath4 photons to couple @xmath29 to @xmath30 ( see also chen & miralda - escud ' e 2003 and kuhlen , madau & montgomery 2006 ) . for simplicity
we assume here that the temperature of the igm lies well above the temperature of the cmb .
regions of ionised gas then appear as holes in the 21 cm emission due to the igm ( see also kohler et al .
( 2005 ) for a more detailed discussion ) . with the rise of the galaxy and qso populations the ionised fraction of gas increases and the 21 cm signal emission due to
the igm becomes weaker .
the signal will disappear by @xmath32 when we know the neutral hydrogen is predominantly ionised .
there will be a finite period of time when the igm is heated above the temperature of the cmb but most of the gas is still neutral .
this is the ( uncertain ) redshift range over which instruments like lofar can probe .
in addition , the performance of lofar will depend on our ability to model and remove foreground contamination due to the ionosphere , our own galaxy , and both extended and point - like extragalactic sources .
we are interested here in the hii regions around supermassive bhs at high redshift . accretion onto a bh
is an efficient mechanism for producing ionising photons and therefore a supermassive bh that has been built up by gas accretion should generate a very large ionised bubble . if the ionising flux generated by the supermassive bhs at high redshift exceeds that produced by other sources in the same region then the sites of supermassive bhs will show up prominently in 21 cm maps as large ionised regions in an otherwise still neutral igm . in order to estimate the space density and size distribution of these hii regions we require a model for the growth of the supermassive black holes and the production of ionising photons .
our best anchor point for such modelling is the observed luminosity function of @xmath33 qsos .
we therefore consider here models which reproduce this luminosity function but differ in the timescale over which the black holes powering the @xmath33 qsos grow .
the e - folding time for bh growth at the eddington limit is @xmath34 gyr , where @xmath35 is the fraction of the mass energy released as radiation .
estimates for the mass of seed bhs produced by population - iii remnants are @xmath36 m@xmath37 .
such seed masses would require @xmath38 e - foldings to grow to @xmath39 m@xmath40 , the estimated mass of the black holes powering the observed @xmath33 qsos .
the growth of a @xmath41 m@xmath37 bh by a redshift @xmath42 from a population - iii remnant therefore requires more or less continuous accretion if the accretion rate is limited to the eddington accretion rate .
it is a matter of intense debate whether accretion above the eddington rate contributes significantly to the growth of supermassive black holes ( see begelman , volonteri & rees 2006 for a recent discussion ) .
we first construct a simple passive evolution model for the growth of supermassive black holes and the evolution of the qso luminosity function at @xmath43 in which a fixed comoving density of black holes accrete at their eddington limit with a fixed duty cycle @xmath44 .
the bolometric luminosity of a qso powered by a bh of mass @xmath45 radiating at the eddington limit is given by @xmath46 where @xmath47 is the thompson scattering cross section .
the b - band luminosity is assumed to be a fraction @xmath48 of @xmath49 . assuming that a fraction @xmath50 of the mass energy of the accreted mass is released as radiation , the ensemble average luminosity evolution with respect to the redshift @xmath0 population
is then given by , @xmath51 where @xmath52 is the cosmic time at redshift @xmath24 .
we adopt the qlf of [ merger_model ] as our template qlf at @xmath33 . in this model
all luminosities grow exponentially and the shape of the luminosity function is independent of redshift .
we consider two values of the total duty cycle @xmath53 $ ] . to account for a possible contribution to the growth in an optically obscured accretion mode we split the duty cycle of accreting black holes into phases of luminous growth ( @xmath54 ) and growth in an obscured phase ( @xmath55 ) .
we assume the qso spends equal amounts of time in each phase so that @xmath56 $ ] . the observed lifetime of a qso in this model is redshift dependent , and depends on the duty fraction of the luminous population , @xmath57 here @xmath58 is the hubble time at a redshift @xmath24 and @xmath59^{1/2}$ ] is the hubble parameter .
[ plot1 ] the activity powering optically bright qsos is thought to be triggered by galaxy mergers , which drive cold gas onto a central black hole .
various models for the growth of black holes in hierarchically merging galaxies have been built ( e.g. kauffmann & haehnelt 2000 , volonteri et al .
2003 ) and have been successful in reproducing the observed qso luminosity function at @xmath60 .
these models generally assign a qso light curve to each merging event , guided by observationally and/or physically motivated estimates for the central black hole mass and associated accretion rate ( see hopkins et al .
2005 for light curves motivated by numerical simulations of a galaxy merger ) .
below we construct a simple version of such a model for the qlf which uses the ( extended ) press - schechter formalism for dark matter halos to estimate the rate of galaxy mergers ( see erickcek et al .
( 2006 ) for a recent critical assessment of the accuracy of the extended press - schechter formalism ) .
the model is similar to those described by haehnelt , natarajan & rees ( 1998 ) and wyithe & loeb ( 2003 ) .
our model makes a number of simplifying assumptions .
we assume that the merger rate of black holes is equal to the merger rate of the dark matter halos of the galaxies hosting the black holes as estimated using the ( extended ) press - schechter formalism .
we assume a relation of the mass of the bh , @xmath61 , and the mass of the dark matter halo forming in a merger @xmath62 as motivated by the observed correlation between @xmath61 and the velocity dispersion of the host galaxies stellar bulge /dark matter halo , @xmath63 in nearby galaxies ( gebhardt et al .
2000 , ferrarese & merritt 2000 , ferrarese 2002 ) .
this relation appears to hold independently of redshift for qsos out to @xmath64 ( shields et al .
2003 , 2006 ) .
empirical estimates of @xmath6 typically fall in the range @xmath65 . we have chosen @xmath66 which is consistent with a simple self - regulated growth scenario in which the black hole grows until it radiates enough energy to unbind the gas that is feeding it ( e.g. silk & rees 1998 , haehnelt natarajan & rees 1998 , wyithe & loeb 2003 ) .
we further assume that the velocity dispersion of the stellar bulge is proportional to the circular velocity of the dark matter halo at the virial radius , @xmath67 , with ( barkana & loeb 2001 ) , @xmath68^{1/6 } \\ & \times & \left ( \frac{1+z}{7 } \right)^{1/2}{\rm kms}^{-1},\end{aligned}\ ] ] where @xmath69 , @xmath70 and @xmath71 is the overdensity of a virialised halo at redshift @xmath24 . equations ( [ msigma ] ) and ( [ vvir ] ) lead to a dependence of black hole mass on host halo mass @xmath62 and redshift of the form , @xmath72^{\alpha/6}\\\nonumber & & \times ( 1+z)^{\alpha/2 } \left(\frac{m}{10^{12 } m_{\sun}}\right)^{\alpha/3 - 1}m.\\\nonumber\end{aligned}\ ] ] note that @xmath73 sets the ratio of @xmath74 for a given redshift . assuming that all dark matter halos host a central bh with mass given by equation ( [ mbh - mhalo ] ) , the merging rate of halos of mass @xmath75 and @xmath76 is equal to the merger rate of halos hosting bhs of mass @xmath77 with those hosting bhs of mass @xmath78 for @xmath79 and @xmath80 .
@xmath81 is given by the product of the probability per unit time that a halo mass @xmath76 will merge with another halo to form a halo with mass @xmath62 and the space density of halos with the appropriate mass difference , @xmath82 the mass density is derived from the press - schechter ( 1974 ) halo mass function with the modification described by sheth & tormen ( 1999 ) and the halo merger probability is taken from the extended press - schechter formalism ( lacey & cole 1993 ) .
we assume that following a merger , each black hole accretes and radiates at its eddington limit for a period equal to the dynamical time of gas in a disk of radius @xmath83 times the virial radius as in wyithe & loeb 2003 , @xmath84^{-1/2 } \left ( \frac{1+z}{3}\right)^{-3/2},\end{aligned}\ ] ] the light curve is assumed to be a simple top - hat function @xmath85 $ ] ( with @xmath86 the usual heaviside step function ) .
the qlf can then be estimated by integrating over the merger rate of all bh pairs as follows , @xmath87 .
\end{aligned}\ ] ] if we perform the integral over @xmath61 using the @xmath88 function , use equation ( [ mbh - mhalo ] ) to make the change of variable from bh masses to halo masses and assume that the merger rate changes little over the qso lifetime [ i.e. @xmath89 , we get ( see wyithe & loeb 2003 ) , @xmath90 where @xmath91 . note that the integral over @xmath76 is truncated at the lower end to ensure that we are only counting major mergers and at the upper end to count each potential halo merger pair only once . in principle ,
@xmath6 , @xmath73 and @xmath92 may be tuned to fit our model qlf to the observed qlf at @xmath7 .
we have fixed @xmath93 and @xmath94 and varied @xmath73 to fit to the observed luminosity function .
for @xmath95 we obtain an acceptable fit to the observed luminosity function ( see figure 1 ) .
this normalisation is consistent with the corresponding empirical normalisation of equation ( [ msigma ] ) for galaxies in the local universe ( @xmath96 , see ferrarese 2002 ) .
note that with the assumption @xmath97 the qso lifetime at @xmath98 is @xmath99 yrs , which is consistent with other estimates of the duration of the optically bright phase of qsos ( haehnelt , natarajan & rees 1998 , haiman & hui 2001 , martini & weinberg 2001 , hopkins et al .
note that the shorter lifetime in the merger driven model compared to the passive evolution model results in smaller hii regions for the assumed values of @xmath100 .
figure 1 shows the evolution of our model qso luminosity function at @xmath2 ( the passive evolution model with two values of the duty cycle and the merger driven model ) . by construction
all models agree with the observed space density of bright qsos at @xmath7 as determined by fan et al .
the redshift evolution differs dramatically for the three models .
fan et al .
( 2001 ) find that the observed space density of bright qsos decreases with redshift like @xmath101 for @xmath102 .
we compare this to the redshift evolution in our model qlfs by summing over the bright end of each qlf .
we sum from a minimum luminosity @xmath103 l@xmath104 which corresponds to a rest frame @xmath105 absolute magnitude of @xmath106 [ for a conversion to b - band magnitude @xmath107 ( fan et al . 2001 ) ] .
the results are displayed in figure 2 .
for the merger driven model the density of qsos declines with a similar slope to the decline in the observed space density at lower redshift . in the passive evolution model
the rate of decline depends sensitively on the assumed duty fraction . for the lower value of the assumed duty cycle ( @xmath108 )
the black holes grow relatively little at high redshift and the drop in the space density of bright qsos is significantly slower than in the merger driven model . for the high value of the duty cycle ( @xmath109 ) the qsos are growing rapidly and quickly decline in luminosity at higher redshift . as we will discuss in more detail below the three models span a wide range of possible growth histories of supermassive black holes at high redshift .
we should also note that all three models overpredict the space density of bright qsos at lower redshifts .
as discussed in detail by bromley , somerville & fabian ( 2004 ) the rapid growth history of supermassive black holes at high redshift is not easily reconciled with the much slower observed growth at low redshift which is normally modelled with rather severe feedback effects due to energy input by stars and/or the black hole .
as we are only interested in the growth of supermassive black holes before reionisation we make no attempt to model this here .
[ plot2 ] the accretion rate of mass onto black holes during the optically bright phase of accretion can be extracted from the optical qlf as follows , @xmath110 an approximation to the total mass density in bhs can be extracted from the model luminosity function by dividing the luminosity function by the fraction of luminous / active halos at a given time @xmath111 , @xmath112 for the passive evolution model @xmath54 is assumed to be constant .
note , however , that this will somewhat overestimate the bh mass density . for the merger rate model @xmath54
is estimated to be the ratio of the qso lifetime to the average survival time of a dark matter halo as a function of halo mass and redshift [ @xmath113 calculated from the probability distribution in equation ( 2.21 ) of lacey & cole ( 1993 ) . at low redshift @xmath114
is approximately the hubble time , but it decreases to 30 per cent of the hubble time by @xmath115 for the relevant halo masses .
figure 3 shows the evolution of the black hole mass density and the rate of mass accretion during optically bright phases for our three models .
as expected the evolution differs strongly between the three models .
we also show the observed black hole mass density at @xmath116 which is about @xmath117 times larger than our estimates at @xmath33 .
the difference in the integrated observed accretion rate and the mass density in bhs is due to the assumed periods of obscured qso accretion . by construction
the integrated luminous accretion rate will fall short of the mass density in black holes in the passive evolution models by a factor @xmath118 at @xmath98 . in our merger driven model
this difference is much larger .
approximately ten per cent of the inferred bh mass density is accumulated due to luminous accretion . in this model approximately 90 per cent of the growth would have to occur in an obscured phase .
such a growth scenario is consistent with the findings of hopkins et al .
( 2005 ) based on hydrodynamical simulations of qso activity fuelled by merging disk galaxies .
assuming that the qso is embedded in a homogeneous gas distribution with neutral hydrogen number density @xmath119 , the increase of the physical volume of ionised hydrogen @xmath120 generated by an ionising flux @xmath121 is given by ( e.g. madau , haardt & rees 1999 , hereafter mhr99 ) , @xmath122 we assume the cosmic abundance of hydrogen with a neutral fraction @xmath123 and that the temperature of the ionised igm ( @xmath30 ) is @xmath124 k. the average hydrogen recombination timescale @xmath125 for such a gas is ( see abel et al .
1997 ) , @xmath126^{-3}\left ( \frac{c_{\rm eff}}{3}\right)^{-1},\ ] ] where we have assumed an effective clumping factor @xmath127 .
the effective clumping factor relevant for recombinations is expected to be low since the densest self - shielded regions of neutral gas have a small covering factor and should not impede the progression of the ionisation front ( see miralda - escud , haehnelt & rees 2000 for a detailed discussion ) .
equation ( [ ode ] ) is integrated numerically to solve for the physical radius of the hii regions for qsos in each of our qlf models , @xmath128 equation ( [ ode ] ) is approximately equivalent to assuming that the volume of the ionised region is equal to the ratio of the number of emitted ionising photons and the density of neutral hydrogen , @xmath129 for a qso with constant luminosity , this approximation breaks down as @xmath130 and the hii region reaches its str@xmath131mgren radius .
however for a qso with luminosity growing exponentially with the mass of the central bh , most of the hii region growth occurs between @xmath132 and @xmath133 .
since @xmath134 , the result holds for the qsos with @xmath135 in our models . note that since little ionising flux is produced early in the lifetime of a qso powered by a bh continually accreting at the eddington limit , the size of the hii regions at @xmath98 are approximately equal for each of the passive evolution models even though the lifetimes differ by a factor of 10 .
for each of our models the radial extent of the ionised region generated by a qso depends inversely on the surrounding neutral hydrogen fraction @xmath123 and increases with the qso luminosity as @xmath136 .
we calculate @xmath137 at all redshifts by assuming the optical / uv spectral energy distribution [ @xmath138 as given by mhr99 with a high energy cut - off of @xmath139 kev . the spectral energy distribution is normalised by assuming @xmath140 ( with @xmath141 as in [ qlf_models ] ) .
this gives an ionising flux , @xmath142 where @xmath143 is the minimum photoionising frequency .
the corresponding radius of the hii region , for a qso with constant luminosity is , @xmath144^{1/3 } \left(\frac{1+z}{7}\right)^{-1}.\end{aligned}\ ] ]
the next generation of radio telescopes are aiming to explore the low - frequency universe , with detection of an epoch of reionisation signal a major scientific goal .
various low - frequency interferometers are being designed , tested and built with some already taking data .
the largest radio telescope on the horizon is the square kilometre array ( ska ) .
the ska is still in the design phase and not expected to be in operation for more than a decade .
however there are several smaller - scale telescopes including lofar ( to be built in the netherlands ) , the mileura wide - field array ( mwa , sited in western australia ) and the primeval structure formation telescope ( past , currently being built in china ) that are expected to be in operation on much shorter time - scales .
these arrays are typically composed of @xmath145 dipoles ( with approximately 10 per cent of the collecting area and therefore sensitivity of the tentative ska design ) spread over a region with maximum baseline @xmath146 km and designed to operate at subsets of the frequency range @xmath147 mhz .
below we focus on the possibility of detecting hii regions with lofar , but our results are easily transferred to alternate designs for a low - frequency interferometer . [ snr_plot ] [ radius_dist ] we assume that the ionised region has a brightness temperature contrast of @xmath148 mk @xmath149^{1/2}$ ] , corresponding to the limit where the hydrogen in the surrounding igm is completely neutral , and the spin temperature ( @xmath29 ) of the surrounding neutral gas is efficiently coupled to the kinetic temperature of a warm igm .
the compact lofar core , rather than the full array , will offer the highest sensitivity to extended structures . in the following we have adopted array parameters corresponding to the published design study of lofar . assuming a spectral resolution @xmath150 and integration time @xmath151 ,
an array of @xmath152 dual polarisation antennae , each with collecting area @xmath153 and system temperature @xmath154 , will have a brightness temperature sensitivity @xmath155 ^ 2 \left(\frac{n_{\rm stn}}{3200}\right)^{-1 } \left [ \frac{a_{\rm eff}}{18~{\rm m}^2}\right]^{-1}\gamma^{-1},\end{aligned}\ ] ] where @xmath156 is the maximum distance between antennae in the core of the array , @xmath157 and @xmath158 is the solid angle of the synthesised beam .
@xmath159 denotes the boltzmann constant .
the sensitivity to point sources @xmath160 is given by , @xmath161^{-1}\left(\frac{n_{\rm stn}}{3200}\right)^{-1}.\end{aligned}\ ] ] lofar is expected to have a bandwidth of 4 mhz with a spectral resolution of 4 khz which will likely be binned into @xmath162 mhz channels .
the spatial resolution of the lofar core will be inversely proportional to the maximum distance between the antennae ( @xmath163 ) . in comparison , the angular size ( @xmath164 ) and frequency depth ( @xmath165 ) of an hii region with radius @xmath166
are given by @xmath167^{-1},\end{aligned}\ ] ] @xmath168 , \end{aligned}\ ] ] where @xmath169 mhz is the rest - frequency of the signal and @xmath170 is the comoving distance to a source at redshift @xmath24 .
we now calculate the signal to noise ratio ( snr ) with which a spherical bubble of a given radius at a certain redshift can be detected with @xmath171 antennae each with @xmath172 m@xmath173 in a circle of variable diameter @xmath156 .
the brightness temperature sensitivity increases with the compactness of the array .
the surface brightness sensitivity will be maximal when the size of the beam is the same as the angular size of the source @xmath174 , @xmath175 . the snr for a spherical hii region is then given by , @xmath176 the factor @xmath177 accounts for the fraction of the beam that is filled with ionised gas . for a spherical top heat beam @xmath178 .
@xmath177 may be adjusted for different array patterns / synthesised beams and different hii region geometries .
this prescription allows us to estimate the signal to noise ratio for detection of an hii region as a function of its redshift and radius .
figure 4 shows the calculated signal to noise ratio as a function of the physical radius of an hii region at various redshifts .
note that the increase of snr with size of the hii region will become shallower once the size exceeds the beam size for the most compact configuration possible .
this calculation indicates that the smallest bubble surrounded by a neutral igm that may be detected at the @xmath179 level is @xmath180 mpc at @xmath33 .
the minimum radius changes very little with source redshift , and turns over at @xmath181 due to the dependence on the angular diameter distance . unfortunately , instrumental noise is not the only type of noise relevant for the detection of hii regions .
residuals from the subtraction of extragalactic and galactic foregrounds and the correction of the distortion of the signal by the ionosphere will all contribute to the noise .
we currently have little idea how large the noise due to these residuals will be .
the intensity fluctuations in the 21 cm emission due to density fluctuations of the igm will also act as noise ( e.g. kohler et al . 2005 , mellema et al .
2006 ) . for @xmath182
hour integrations with the lofar core the fluctuations due to the igm surrounding the source limit the sensitivity on scales larger than @xmath42 arcminutes ( wyithe , loeb & barnes 2005 ) ; increasing the exposure time will not significantly decrease the minimum size of detectable hii regions .
furthermore , the edge of an hii region surrounding a qso is not expected to be sharp ( see zaroubi & silk 2005 ) and the intrinsic shape will appear distorted due to light travel time effects ( see wyithe & loeb 2004b ; yu 2005 ) . the signal to noise ratios estimated here
should thus be considered as ( probably rather optimistic ) upper limits . in figure 5
we show our predictions for the space density of hii regions as a function of size at @xmath2 in units of @xmath183 for each of the three models .
the vertical and horizontal lines show the typical size of a lofar observing volume ( @xmath184 deg@xmath173 and @xmath185 mhz of bandwidth ) and our estimate for the @xmath179 detection threshold for the size .
the effective recombination timescale for hydrogen gas at @xmath124 k is comparable to or somewhat shorter than the hubble time . as discussed by wyithe ,
loeb & barnes ( 2005 ) hii regions are thus expected to remain as fossils for some time after the qso turns off .
we assume that the hii regions continue to expand with the hubble flow once the quasar turns off and there are no further recombinations , and count the number of fossil regions with physical radius r at @xmath24 by summing over the size distribution function at redshift intervals corresponding to the quasar lifetime .
we assume that the quasars at @xmath33 are on average half way through their optically luminous lifetime .
there are essentially no fossil hii regions in the case of @xmath186 since the qso luminosity grows exponentially on a timescale comparable to the quasar lifetime .
note that for the other two models @xmath187 is @xmath188 times smaller than @xmath189 for the smallest hii regions ( corresponding to the faintest quasars , @xmath190 ) at @xmath98 due to the rapid evolution of the space density of qsos .
@xmath191 asymptotes to one for large hii regions and high redshifts as expected .
we show both the number of hii regions around active qsos ( black lines ) and the total number ( including fossil bubbles ) of hii regions ( grey lines ) . we have assumed that the igm is still neutral all the way down to @xmath33 .
we will come back to this point below . for the merger - driven model we predict @xmath192 detectable sources around active qsos per lofar field at @xmath98 .
this number increases to @xmath193 if we include fossil hii regions around supermassive black holes which are no longer active .
the numbers are similar to those reported by wyithe , loeb & barnes ( 2005 ) for their similar model . in our merger driven model at least one detectable hii region per lofar field is predicted out to @xmath194 . due to the steep slope of the qlf at high luminosities , the numbers would decrease more rapidly with redshift if the effective minimum detectable radius is larger than we have estimated . for our two passive evolution models the predictions are quite different .
the model with the lower duty cycle offers a more optimistic picture than the merger driven model .
the longer qso lifetime results in larger hii regions for the same qso luminosity .
we predict @xmath195 detectable hii regions per lofar field at @xmath196 around active luminous qsos and detectable hii regions are present up to @xmath181 .
the value of the maximum redshift depends sensitively on the duty cycle and could be at least in principle
much larger for smaller values of the duty cycle . this would , however , require the formation of very massive black holes at very high redshift which would be difficult in most formation scenarios .
the passive evolution model with the higher value of the duty cycle depicts a much more pessimistic scenario .
the predicted numbers at @xmath33 are similar to the other models , but the the presence of detectable hii regions does not extend much beyond @xmath33 .
this is easily explained by the very recent growth of the @xmath33 supermassive black holes in this model .
figure 6 shows the redshift evolution of detectable active and fossil hii regions with @xmath197 mpc .
clearly the prospects for detecting hii regions depend sensitively on how bhs grow at high redshift .
so far we have assumed that the igm is still neutral all the way down to @xmath33 .
however the neutral fraction at redshift @xmath0 is uncertain and the universe may already be highly ionised at @xmath198 ( malhotra & rhoads 2005 , fan et al .
a smaller but still significant neutral fraction would result in a larger hii region [ see equation ( [ lum - rad ] ) ] , however the signal to noise ratio for the detection of a resolved hii regions scales with the square root of the neutral fraction ( @xmath199 ) ; sources in a partially neutral igm are more difficult to detect .
for consistency we have calculated the volume filling factor of hydrogen ionised by the qsos in each of our models .
we assume that our model extends to qsos as faint as @xmath200 l@xmath40 , include the contribution due to fossil regions and neglect recombinations .
our filling factor should therefore represent an upper limit on the expected value .
we find a filling factor of hii regions generated by the uv emission of active quasars in our models which is less than @xmath201 per cent at @xmath7 for the model with the low duty fraction and for the merger model and @xmath202 cent for the model with the large duty fraction .
for each model the filling factor drops rapidly at higher redshift .
note again that we have integrated down to rather faint quasars and that taking recombinations into account would also lower these values .
our assumption that the hii regions generated by qsos are isolated may therefore be somewhat questionable at @xmath7 but should be reasonable at @xmath203 .
not surprisingly the quasars in our models which are consistent with the rather low observed space density of qsos at @xmath204 alone are barely able to fully reionise hydrogen by @xmath204 .
the quasars should also contribute to the ionisation of hydrogen with their x - ray emission . due to their long mean free path ,
x - rays will lead to a more uniform ionisation which includes the low - density igm far away from the quasars ( e.g. madau et al . 2004 ,
ricotti & ostriker 2004 ) . the efficiency with which an x - ray photon ionises hydrogen atoms in the igm depends on the average number of ionisations induced by secondary electrons , which is a function of the initial photon energy as well as the existing ionised gas fraction ( see shull & van steenberg 1985 ) .
we follow ricotti & ostriker and assume that each photon with energy @xmath205 kev results in 10 ionisations in a partially ionised igm . assuming again that our model extends to quasars as faint as @xmath200 l@xmath40 the mean ionised fraction at @xmath33 due to the x - ray emission of active qsos in our model is less than 1 percent for all models .
another possibly relevant effect is the clustering of qsos .
bright sources are sufficiently rare that clustering should not be important . for fainter sources overlap of clustered hii region could increase the sizes compared to our estimates in particular if fossil hii regions are taken into account .
[ radius_evo ]
we have investigated the prospects of using the 21 cm signature of the regions around supermassive black holes at @xmath2 to probe the growth history of these supermassive black holes and the ionisation history of their surrounding igm .
we have presented three simple , physically motivated models for the high redshift growth history of supermassive black holes in which the growth is either simple eddington limited accretion or due to short - lived accretion phases triggered by the merger of galaxies .
we have used a simple analytical model for the size of hii regions in order to generate predictions for the size distribution of hii regions for each of our model qlfs .
our results show that models which are all consistent with the luminosity function of bright qsos at @xmath98 but differ in the growth history of the supermassive black holes give very different results for the predicted number of detectable hii regions .
our estimates for the number of detectable hii regions at @xmath7 are promising , up to several hundred detectable sources per lofar field around active qsos and possibly many more fossil regions .
however , the number of detectable hii regions depends sensitively on unknown details of smbh growth and the ionisation state of the igm .
if the qlf evolves to @xmath206 as predicted in our merger - driven model , then isolated hii regions will be prominent features of the high redshift neutral hydrogen distribution providing reionisation was completed late ( @xmath207 ) .
however if the smbh population is growing rapidly due to continuous accretion , or reionisation occurred much earlier than @xmath7 , then the detection of qso hii regions may prove elusive .
the most optimistic picture emerges if supermassive black holes grow more slowly over an extended period of time as in our eddington limited accretion model with a smaller duty cycle ( @xmath208 ) . in this case
detectable sources may be present to very large redshifts and the prospects for detection are good even if reionisation is completed relatively early .
observation of future 21 cm experiments like lofar should not only provide a wealth of information on how the universe was reionised but may also provide important clues for the growth history of supermassive black holes at very high redshift .
the authors wish to thank stuart wyithe for useful comments on a draft and the referee for suggesting calculating the hii region filling factor and the degree of pre - reionisation by x - rays .
kjr is supported by an overseas research scholarship and the cambridge australia trust .
this research was supported in part by the national science foundation under grant no .
phy99 - 07949 . | hii regions surrounding supermassive black holes ( bhs ) in an otherwise still neutral intergalactic medium ( igm ) are likely to be the most easily detectable sources by future 21 cm experiments like lofar . we have made predictions for the size distribution of such hii regions for several physically motivated models for bh growth at high redshift and compared this to the expected lofar sensitivity to these sources . the number of potentially detectable hii regions does not only depend on the ionisation state of the intergalactic medium and the decoupling of the spin temperature of the neutral hydrogen from the cosmic microwave background ( cmb ) temperature , but is also strongly sensitive to the rate of growth of bhs at high redshift . if the supermassive bhs at redshift @xmath0 were built up via continuous eddington - limited accretion from low mass seed bhs at high redshift , then lofar is not expected to detect isolated qso hii regions at redshifts much larger than 6 , and only if the igm is still significantly neutral . if the high - redshift growth of bhs starts with massive seed bhs and
is driven by short - lived accretion events following the merging of bh hosting galaxies then the detection of hii regions surrounding supermassive bhs may extend to redshifts as large as @xmath1 but is still very sensitive to the redshift to which the igm remains significantly neutral .
the most optimistic predictions are for a model where the supermassive bhs at @xmath2 have grown slowly .
hii regions around supermassive bhs may then be detected to significantly larger redshifts .
[ firstpage ] quasars : general - cosmology : theory - intergalactic medium - hii regions |
in the colour - magnitude diagram ( cmd ) of a globular cluster ( gc ) , blue straggler stars ( bsss ) define a sub - population located along the main - sequence ( ms ) in a position brighter and bluer than the current ms - turnoff ( to ) . for this reason
these objects are thought to be hydrogen burning stars more massive than a typical to star .
two physical mechanisms are thought to be responsible for their formation : mass - transfer in a binary system @xcite , and direct stellar collisions @xcite .
since collisions are more frequent in regions of higher density , the relative importance of the different formation channels could depend on the environment ( e.g. , * ? ? ?
* ; * ? ? ?
* ) : bsss in loose gcs might preferentially arise from mass - transfer activity in primordial binaries ( hereafter mt - bsss ; * ? ? ?
* ; * ? ? ?
* ) , while bsss located in high density environments might mainly form from stellar collisions ( hereafter col - bsss ; * ? ? ?
discovered for the first time by @xcite in the external regions of the gc m3 , it was with the advent of the hubble space telescope ( hst ) and the 8-meter class telescopes that it became possible to search for bsss in dense cluster cores ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
bsss have been found in any properly imaged gc @xcite and they are studied also in dwarf spheroidal galaxies @xcite .
since the bss formation mechanisms seems to be tightly connected with the cluster internal dynamics , these stars are commonly recognized as ideal test particles to investigate the impact of dynamics on stellar evolution in different environmental conditions ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
an interesting example comes from the discovery of two distinct sequences of bsss in the core of m30 @xcite .
the authors argued that each of the two sequences is populated by bsss originated by one of the two formation channels , both triggered by the collapse of the cluster core a few gyr ago . on the other side , @xcite
suggest that most bsss , even those found in cluster cores , come from binary systems ( see also * ? ? ?
nevertheless the parent binaries may themselves have been affected by dynamical encounters .
systematic studies of the bss populations in gcs have shown that in most cases their radial distribution is bimodal , with a high peak in the cluster centre , a trough at intermediate radii and a rising branch in the external regions ( see e.g. * ? ? ? * and references therein ) .
dynamical simulations ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) suggest that the observed shape of the bss radial distribution depends on the relative contribution of col - bsss and the efficiency of the dynamical friction , that progressively segregates objects more massive than the average ( as bsss and their progenitors ) towards the cluster centre .
very interesting exceptions are the cases of @xmath1 centauri @xcite , ngc 2419 @xcite and pal 14 @xcite , where the bss radial distribution is completely flat .
this fact indicates that in these clusters the dynamical friction was not effective yet in segregating bsss toward the centre of the potential well .
interestingly enough ngc 2419 and pal 14 are among the most remote gcs in the galaxy , lying at a distance from the galactic centre @xmath0 kpc .
we decided to extend the investigation to other four galactic gcs ( namely am 1 , eridanus , pal 3 , pal 4 ) located in the extreme outer - halo . in section [ sec_data ]
we describe the data - set and reduction procedure .
the cmd and the age estimate for each cluster are discussed in section [ sec_cmd ] .
the determination of the astrometric solutions and the cluster centres of gravity is presented in section [ sec_astro ] .
the radial density profiles from resolved star counts and the estimates of the cluster structural parameters and characteristic time - scales are discussed in section [ sec_dens ] .
section [ sec_bss ] is devoted to the radial distribution of the bss , while a summary is presented in section [ sec_summ ] .
* hereafter s99 ) published deep optical cmds obtained with the wide field planetary camera 2 ( wfpc2 ) on board the hst for eridanus , pal 3 and pal 4 .
the core region of eridanus was roughly centred on the wf3 chip of the wfpc2 mosaic , while pal 3 and pal 4 were centred on the pc chip .
we refer to s99 ( see also * ? ? ?
* ) for a detailed description of data quality and reduction procedure . in brief
, a standard daophotii @xcite point spread function ( psf ) fitting procedure , including allframe @xcite , was adopted to obtain instrumental magnitudes and colours for all measurable stars in the wfpc2 fields .
here we adopted the s99 photometric catalogues of pal 3 , pal 4 and eridanus listing the @xmath2 and @xmath3 magnitudes of 2968 , 4122 and 2340 stars , respectively .
am 1 was observed with the wfpc2 in the @xmath2 and @xmath3 filters during cycle 6 ( go-6512 ; pi hesser ) .
the observation log is described in table 1 of ( * ? ? ?
* hereafter d08 ) .
we retrieved the am 1 images from the stsci archive and performed independent psf fitting photometry .
the psf was modelled on each image using a number ( between 40 and 80 ) of isolated and well sampled stars , adopting the psf / daophotii routine .
a first psf fitting was then performed on every image using daophotii / allstar , while allframe was used to obtain a fine measure of the star magnitudes . for each star
all the magnitudes were normalized to a reference frame and averaged together , and the photometric error was derived as the r.m.s .
of the repeated measurements . for homogeneity with the s99 reduction procedure
, we used the colour terms in table 7 from @xcite to convert the instrumental ( @xmath4 ) magnitudes to the standard johnson v and kron - cousins i ( hereafter v and i , respectively ) . and
@xmath5 , respectively . ]
the cmds obtained for am 1 , eridanus , pal 3 and pal 4 are shown in figure [ fig_cmd ] ( see also figures 13 in s99 ) .
all the detected stars are shown as black dots , while the location of the mean ridge line obtained with a second order polynomial interpolation with a 2.5 sigma clipping rejection criterion is displayed with grey points .
the high quality of the wfpc2 images allows us to properly sample the clusters stars from the tip of the red giant branch ( rgb ) down to @xmath6 mag below the ms - to .
the solid line in figure [ fig_cmd ] shows the location of the best - fit theoretical isochrone from @xcite over - plotted on the cmds of each cluster .
we used the distance modulus and reddening from ( * ? ? ?
* edition ) , while we adopted metallicity [ fe / h]=1.58 and 1.41 from @xcite and @xcite for pal 3 and pal 4 , respectively , [ fe / h]=1.42 from s99 for eridanus , and [ fe / h]=1.5 from d08 for am 1 . with these parameters
we obtain an absolute age of 10.5 gyr for eridanus and pal 4 , and 11 gyr for pal 3 and am 1 , confirming that these clusters are younger than the inner - halo gcs ( s99 and d08 ) .
the location of the to , defined as the hottest point along the ms in the theoretical isochrone of each cluster , is also marked in figure [ fig_cmd ] .
this was used to compute the shifts in magnitude and colour that , after taking into account also the different distance moduli and reddenings , are necessary to co - add the four cmds .
the result is impressive ( see figure [ fig_cmd_all ] ) : the rgb morphologies of the four clusters are in full agreement and consistent with a small metallicity difference among their stellar populations .
indeed , the combined cmd mimics a highly homogeneous single - metallicity population .
these results are in agreement with the study of s99 and d08 , who find that these clusters are younger than m 3 by 1.5 - 2 gyr , and they are homogeneous in terms of both age and metallicity .
l c c cluster & ra@xmath7 & dec@xmath7 + am 1 & @xmath8 & @xmath9 + eridanus & @xmath10 & @xmath11 + pal 3 & @xmath12 & @xmath13 + pal 4 & @xmath14 & @xmath15 + the coordinates of the stars identified in each cluster have been converted from the `` local '' to the absolute astrometric system by using a well tested procedure adopted by our group for more than 10 years ( see e.g. * ? ? ? * and references therein ) . since only a few ( if any ) primary astrometric standards can be found in the central region of gcs , we usually complement hst photometry with ground based wide - field imaging . b- and r - band images of am 1 where taken in december 2000 with the wide field imager ( wfi . ] ) at the mpe / eso 2.2 m telescope ( 66.b-0454(a ) ; pi : gallager ) .
pal 3 , pal 4 and eridanus were observed in the @xmath16 and @xmath17 bands with megacam , the @xmath18 fov ccd camera at the canada - france - hawaii telescope ( cfht ) .
we retrieved the bias - subtracted and flat - field corrected images from the cfht science data archive ( observing run i d : 2010ac06 for pal 3 and pal 4 ; 2009bc02 for eridanus ; pi cote ) .
we analysed the data following the procedure described in section [ sec_data ] .
we analysed only the few chips ( from 2 to 4 , depending on the location of the cluster centre in the fov of each data - set ) which allow us to sample the entire cluster radial extent .
notice that the tidal radius of the four clusters ranges between @xmath19 and @xmath20 @xcite .
unfortunately the ground
based observations reach only the to level and were not deep enough to properly sample the bss region with an appropriate signal - to - noise ratio . for this reason we used this data set only to search for astrometric solutions and to supplement the hst data for the construction of the cluster projected density profiles from resolved star counts ( see sect .
[ sec_dens ] ) .
we used more than one hundred primary astrometric standard stars from the gsc2.3 catalogue in the cluster vicinity to derive an astrometric solution and obtain the absolute equatorial ( ra and dec ) positions of the stars sampled in the ground based catalogues .
cross - correlations of the catalogues and astrometric solutions were calculated with cataxcorr , a code developed and maintained by paolo montegriffo at the inaf - bologna astronomical observatory ( see e.g. * ? ? ?
* ; * ? ? ?
finally , the positions of the stars in the wfpc2 data were transformed into the same celestial coordinate system by adopting @xmath21 stars in common with the ground based catalogues .
the r.m.s .
scatter of the final solution was @xmath22 in both ra and dec for the four clusters .
the centre of gravity ( @xmath23 ) of the four clusters is estimated as the barycenter of the resolved stars ( see e.g. * ? ? ?
* ) . in doing this an iterative procedure was adopted .
the method estimates the centre by averaging the ra and dec positions of all the stars contained within a circular area of a given radius and proceeds until convergence is reached .
we performed this procedure using three different limiting radii and , at any given radius , using only stars with magnitudes brighter than @xmath5 , @xmath24 mag and @xmath25 mag .
the final value of @xmath23 is calculated as the average of the nine measures ( see table [ tab_cent ] ) .
the uncertainties both in ra and dec are smaller than 1 arcsec for eridanus and am 1 , while they increase to @xmath26 for pal 3 and pal 4 because of the extremely low stellar density even in the core region of these clusters . within the errors the new centres are in general good agreement with the ones listed by @xcite .
we used the photometric catalogues and the estimated values of @xmath23 to calculate the projected radial density profile from resolved star counts for each cluster .
we divided the entire sample into concentric annuli , each split into four subsectors ( quadrants ) .
the number of stars within each subsector was counted , and the star density was obtained by dividing these values by the corresponding subsector areas . the stellar density in each annulus was then obtained as the average of the subsector densities .
the star counts in the clusters central regions were performed using the stars with a magnitude @xmath27 from the wfpc2 high - resolution imaging catalogues , while the stars sampled down to the to were taken from the ground based data set for the external regions .
notice that the stars adopted to compute the density profile cover a very limited range of mass in both the hst and the ground based datasets , so that a single mass - model can be used to properly describe the observed profile ( see next section ) .
the final density profiles of the program clusters are shown in figure [ fig_prof ] , where the star counts from ground based catalogues ( open diamonds ) have been normalised to the wfpc2 ones ( solid circles ) using at least two circular area in common between the two data sets .
these are the most accurate and extended radial density profiles ever published for these gcs .
lcccccccccc cluster & age & @xmath28 & @xmath29 & @xmath30 & @xmath31 & @xmath32 & @xmath33 & @xmath34 & @xmath35 + & [ gyr]&@xmath36 $ ] & & & [ pc ] & [ pc ] & @xmath37 & [ yr ] & [ yr ] + am1 & 11.0 & 4.10 & 5.60 & 1.16 & 5.99 & 95.12 & 0.48 & 9.02 & @xmath38 + eridanus & 10.5 & 4.26 & 5.00 & 1.03 & 6.35 & 77.01 & 0.06 & 9.32 & @xmath39 + pal3 & 11.0 & 4.48 & 3.40 & 0.74 & 12.74 & 89.72 & 0.04 & 9.95 & @xmath40 + pal4 & 10.5 & 4.62 & 3.60 & 0.77 & 12.59 & 93.77 & 0.18 & 9.96 & @xmath41 + the derived density profiles allowed us to derive the clusters structural parameters through fitting isotropic king models @xcite . the background density was estimated from the outer data points .
the best fit has been determined by @xmath42 minimization .
as shown in figure [ fig_prof ] , the observed profiles are very well reproduced by isotropic king models characterised by the quoted parameters , with @xmath29 , @xmath30 , @xmath31 and @xmath32 indicating the central dimensionless potential , the concentration , the core and tidal radii , respectively ( see also table [ tab_struc ] ; we adopted the distances quoted by * ? ? ?
these parameters have been used to estimate the core relaxation time ( @xmath43 ) and the half - mass relaxation time ( @xmath44 ) of the four clusters from equation ( 10 ) and ( 11 ) of @xcite .
we assumed the average stellar mass @xmath45 @xcite while the total cluster masses @xmath46 have been derived from the observed total luminosity @xcite and the mass - to - light ratios quoted by ( * ? ? ?
* we adopted an average value of 1.9 for eridanus , which is not included in that work ) .
the resulting values of @xmath47 and @xmath44 are shorter than the age of the clusters ( table [ tab_struc ] ) .
hence , some degree of central mass segregation is expected in the program clusters .
this result is at odds with what found in the two outer halo gcs previously investigated , ngc 2419 and pal 14 , whose @xmath44 ( @xmath48 and @xmath49 gyr , respectively ) are longer than the clusters ages ( @xmath50 and @xmath51 , respectively ; see * ? ? ?
as demonstrated by previous results ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* and references therein ) , bsss in most galactic gcs are more centrally concentrated than normal stars .
since bsss are more massive than the average , and since the half - mass relaxation time in those systems is much smaller than their age , this result is generally ascribed to the effect of dynamical friction and mass segregation ( e.g. * ? ? ?
the clusters studied in this paper contain a significant number of candidate bsss ( see figure [ fig_cmd_all ] ; see also * ? ? ? * hereafter s05 ) . in order to investigate their radial distribution ,
the bss populations have been selected in a homogeneous way in all clusters .
we assumed the to as the reference point and selected as bona - fide bsss all the stars brighter than @xmath5 and @xmath52 bluer than @xmath53 , where @xmath5 and @xmath53 are , respectively , the magnitude and colour of the to as defined in section [ sec_cmd ] , while @xmath54 is the combined photometric uncertainty in the colour . the adopted selection boxes and the resulting bona - fide bss samples are shown in figure [ fig_sel ] . according to @xcite , in order to study the radial distribution of bsss , it is necessary to define a reference population which is expected to follow the cluster light distribution .
since the number of horizontal branch stars in these gcs is very low , we decided to adopt the rgb as the reference population .
we selected as giants all the stars in the range of magnitude @xmath55 , lying at a distance smaller than 3@xmath56 from the mean ridge line ( see asterisks in figure [ fig_sel ] ) , where @xmath56 is the uncertainty associated with the mean ridge line ( horizontal bars in figure [ fig_cmd ] ) .
this choice allows us to obtain a populous sample of reference stars in the same range of v magnitude , i.e. , in the same condition of completeness as for the bsss .
the total number of bsss and rgb stars in each cluster is listed in table [ tab_count ] .
l c c cluster & n@xmath57 & n@xmath58 + am 1 & 31 & 352 + eridanus & 19 & 187 + pal 3 & 15 & 191 + pal 4 & 16 & 354 + in figure [ fig_ks ] we show the cumulative radial distribution of bsss ( solid lines ) and rgb stars ( dashed lines ) as a function of the projected distance from the cluster centre ( @xmath17 ) normalized to the core radius @xmath31 estimated in section [ subsec_struc ] .
a kolmogorov - smirnov test has been performed to check the probability that bsss and rgb stars are extracted from the same parent distribution .
based on the probability values ( @xmath59 ) obtained for each cluster and reported in figure [ fig_ks ] , we conclude that bss are more concentrated than the rgb in am 1 and eridanus , while the result is inconclusive for pal 3 and pal 4 . as second step to investigate the radial distribution of bsss in the four clusters
, we calculated the ratio between the number of bss ( n@xmath57 ) and that of rgb stars ( n@xmath58 ) , counted in concentric annuli centred on the clusters @xmath23 . the radial distribution of n@xmath60 in the four clusters is shown in figure [ fig_rad ] .
consistently with what was found above , bsss appear to be systematically more centrally concentrated than rgb stars .
it is important to emphasise that , because of the insufficient quality of the wide - field data , we are forced to limit our analysis to the region sampled by the hst observations , which do not reach the tidal radius of the program clusters .
thus , unfortunately we can not characterise the full shape of the bss radial distribution , and we are not able to conclude whether the distributions are bimodal or not .
however this analysis demonstrates that dynamical friction was already effective in centrally segregating bsss in the program clusters . ) , for each of the four clusters .
the probability that the two populations are extracted from the same distribution is quoted in each panel . ] ) for each of the four clusters . ]
given the relatively small number of bsss , we applied the method of the minimum spanning tree ( mst ; see * ? ? ? * and references therein ) as an alternative test to evaluate the degree of bss segregation .
the mst is the unique set of straight lines ( edges " ) connecting a given sample of points ( vertices " ; in this case the star coordinates ) without closed loops , such that the sum of the edge lengths is the minimum possible .
hence , the length of the mst is a measure of the compactness of a given sample of vertices .
@xcite showed that the degree of mass segregation in a star cluster can be measured by comparing the length of the mst of two populations with different average masses ( see also * ? ? ?
notice that the mst method is independent of the catalogue astrometric accuracy and of any assumption about the cluster center .
moreover , it gives quantitative measures of both the degree of mass segregation and its associated significance .
here we provide a detailed analysis of the segregation of bsss with respect to rgb stars in the four program clusters , based on a recent modification of the mst method , where , instead of the direct sum of @xmath61 edges of length @xmath62 , their geometric mean is used @xcite : @xmath63.\ ] ] following equation ( 2 ) , we have computed @xmath64 for the most massive population ( i.e. , the observed samples of @xmath65 bsss in each cluster ) .
then , we randomly extracted @xmath66 sets of @xmath65 objects from the reference population ( the combination of the bss and the rgb samples ) , and we computed their mean and standard deviation , @xmath67 and @xmath68 respectively .
finally , the level of bss segregation with respect to giant stars and its associated uncertainty have been estimated as : @xmath69 the geometrical mean has the very useful property of effectively damping contributions from extreme edge lengths , so that @xmath70 of a compact configuration of even a few stars will not be much affected by an `` outlier '' .
note that @xmath71 has the dimension of a length , while @xmath72 is dimensionless .
a value of @xmath73 is found if the two populations have the same radial distribution , while @xmath74 indicates the presence of mass segregation meaning a more concentrated distribution of massive stars compared to the reference population . in the latter case ,
the significance of mass segregation is provided by the value of @xmath75 for which @xmath76 .
table [ tab_mst ] lists the values of @xmath72 and the related @xmath77 uncertainties computed for the bss populations in the four target clusters , and , for comparison , for ngc 2419 and pal 14 . in agreement with the results of @xcite and @xcite
, we find that bss and rgb stars share the same radial distribution in ngc 2419 and pal 14 ( @xmath78 ) .
conversely , bsss appear to be significantly more segregated than giant stars in the other four gcs .
in particular , am 1 and eridanus show the highest level of mass segregation ( with @xmath79 ) , with high statistical significance ( @xmath80 ) .
pal 3 and pal 4 show a slightly lower degree of bss segregation ( @xmath81 ) , at the @xmath82 level .
.values of the mst tree ( @xmath72 ) and its @xmath77 uncertainty ( @xmath83 ) for the bss in the four target clusters and for ngc 2419 and pal 14 . [ cols="<,^,^",options="header " , ]
using hst - wfpc2 data we studied the bss population in the central regions of four galactic halo gcs , namely am 1 , eridanus , pal 3 and pal 4 .
these clusters , together with ngc 2419 and pal 14 , represent the entire group of gcs at @xmath0 kpc .
a proper comparison of the derived cmds with theoretical isochrones confirms the younger ages ( 10.5 - 11 gyr ) of these systems with respect to the inner - halo gcs ( see figure [ fig_cmd ] ) .
the impressive similarity of their rgb morphology ( fig .
[ fig_cmd_all ] ) also indicates a small metallicity difference ( @xmath84[fe / h]@xmath85 , depending on the cluster ) , in agreement with previous findings by s99 and d08 . by complementing hst data with wide - field catalogues from ground based imaging ,
we have derived the most extended radial density profiles from resolved star counts ever published for these clusters ( fig .
[ fig_prof ] ) .
all profiles are well fit by isotropic king models with the structural parameters listed in table [ tab_struc ] .
unluckily , the insufficient quality of the wide field catalogues did not allow us to study the bss populations in the cluster outskirts and look for possible signatures of an external rising branch , similar to that found in most of the previously surveyed gcs ( see * ? ? ?
* and references therein ) .
we therefore limited the analysis to the area covered by the hst data .
bsss have been selected in a homogenous way in the four program clusters and their radial distribution has been compared to that of rgb stars , taken as a reference .
we found that bsss are significantly more centrally concentrated than giants in all four systems ( see figs .
[ fig_ks ] , [ fig_rad ] and table [ tab_mst ] ) .
since bsss are assumed to be more massive than normal cluster stars , their higher central concentration is interpreted in terms of an effect of mass segregation .
indeed , the half - mass relaxation time of the four program gcs is found to be smaller than their age ( see table [ tab_struc ] ) , thus indicating that dynamical friction has been already effective in segregating the most massive stars in the cores . by measuring the degree of mass segregation with the @xmath86 test ,
we conclude that am 1 and eridanus are the dynamically oldest systems in the group , while pal 3 and pal 4 are slightly less dynamically evolved , and ngc 2419 and pal 14 show no signature of mass segregation yet ( see also * ? ? ?
* ; * ? ? ?
the results shown in this paper once more indicate that , indeed , bsss represent the ideal population to investigate the dynamical state of dense stellar systems .
the authors thank the anonymous referee for her / his useful comments on the first version of the paper .
this research is part of the project cosmic - lab funded by the european research council ( under contract erc-2010-adg-267675 ) .
the research leading to these results has received funding from the european community s seventh framework programme ( /fp7/2007 - 2013/ ) under grant agreement no 229517 . based on observations made with the nasa / esa hubble space telescope , obtained from the data archive at the space telescope institute .
stsci is operated by the association of universities for research in astronomy , inc . under the nasa contract
nas 5 - 26555 .
this research used the facilities of the canadian astronomy data centre operated by the national research council of canada with the support of the canadian space agency .
appreciates funding by the german research foundation ( dfg ) , grant ol 350/1 - 1 .
n.l thanks holger baumgardt for his great help to understand and apply the king models used in this paper .
_ facilities : _ , .
lanzoni , b. , dalessandro , e. , ferraro , f. r. , mancini , c. , beccari , g. , rood , r. t. , mapelli , m. , & sigurdsson , s. 2007a , , 663 , 267 lanzoni , b. , et al . 2007b , , 663 , 1040 lanzoni , b. , ferraro , f. r. , dalessandro , e. , et al .
2010 , , 717 , 653 | using hst / wfpc2 data , we have performed a comparative study of the blue straggler star ( bss ) populations in the central regions of the globular clusters am 1 , eridanus , palomar 3 , and palomar 4 .
located at distances @xmath0 kpc from the galactic centre , these are ( together with palomar 14 and ngc 2419 ) the most distant clusters in the halo .
we determine their colour - magnitude diagrams and centres of gravity .
the four clusters turn out to have similar ages ( 10.5 - 11 gyr ) , significantly smaller than those of the inner - halo globulars , and similar metallicities . by exploiting wide field ground based data
, we build the most extended radial density profiles from resolved star counts ever published for these systems .
these are well reproduced by isotropic king models of relatively low concentration .
bsss appear to be significantly more centrally segregated than red giants in all globular clusters , in agreement with the estimated * core * and half - mass relaxation times which are smaller than the cluster ages . assuming that this is a signature of mass segregation , we conclude that am 1 and eridanus are
slightly dynamically more evolved than pal 3 and pal 4 . |
in recent years the `` shashlik '' technology has been extensively studied to assess its performance at @xmath1 , @xmath2 and @xmath3 accelerator experiments @xcite .
shashlik calorimeters are sampling calorimeters in which scintillation light is read - out via wavelength shifting ( wls ) fibers running perpendicularly to the converter / absorber plates @xcite .
this technique offers the combination of an easy assembly , good hermeticity and fast time response . in many applications it also represents a cheap solution compared to crystals or cryogenic liquid calorimeters .
shashlik calorimeters are , in particular , considered to be good candidates for barrel electromagnetic calorimetry at future linear @xmath1 colliders @xcite . in this context , the physics requirements impose @xmath4 , at least three longitudinal samplings , transversal segmentation of the order of @xmath5 ( @xmath6 @xmath7 ) and the possibility of performing the read - out in a 3 t magnetic field . the present shashlik technology can satisfy these requirements , except for the optimization of longitudinal segmentation which still needs development . the solution proposed in this paper
consists of thin vacuum photodiodes inserted between adjacent towers in the front part of the calorimeter .
they measure the energy deposited in the initial shower development that allows for longitudinal sampling and @xmath0 separation .
a prototype detector was exposed to a beam with the aim of measuring the sampling capability and demonstrating that the insertion of diodes neither deteriorates critically the energy response nor produces significant cracks in the tower structure .
the tested prototype had 25 pb / scintillator towers , assembled in a @xmath8 matrix .
each tower consisted of 140 layers of 1 mm thick lead and 1 mm thick scintillator tiles , resulting in a total depth of @xmath9 .
the sampling was the finest ever used with the shashlik technique .
the transversal dimension of each tower was @xmath10 @xmath7 .
in the first @xmath11 the tiles had a smaller transverse dimension to provide room for the housing of the diodes . plastic scintillator consisting of polystyrene
doped with 1.5% paraterphenyl and 0.05% popop was used .
optical insulation between the towers was provided by white tyvek paper .
as it is custumary in shashlik technique , the blue light produced in the scintillator was carried to the photodetector at the back of the calorimeter by means of plastic optical fibers doped with green wls .
the 1 mm diameter fibers crossed the tiles in holes drilled in the lead and scintillator plates and they were uniformly distributed with a density of 1 fiber/@xmath7 . in the scintillator tiles the holes were 2 mm larger ( 4 mm in the lead ) than the fiber diameter .
the light transmission between the plastic scintillator and the fibers was in air .
all the fibers from the same tower were bundled together at the back and connected to photodetectors .
two types of fibers were tested : bicron bcf20 fibers and kuraray y11 . in both cases ,
the emission peak was at about 500 nm .
light collection was increased by aluminizing the fiber end opposite to the photodetector by sputtering .
the light from the fibers was viewed after a 5 mm air gap by 1
hamamatsu r2149 - 03 phototetrodes .
each tetrode was placed inside an aluminium housing , containing a charge sensitive jfet preamplifier and a high voltage divider .
the differential output signals were shaped with a shaping time of 1.500 @xmath12s and digitized .
four towers were read - out with hamamatsu avalanche photodiodes instead of tetrodes .
a plexiglass light guide was used to match the smaller apd sensitive area to the fiber bundle .
preamplifiers and voltage dividers were housed in the same mechanical structure as the tetrodes .
two types of vacuum photodiodes , viewed with a bialcali photocathode , were produced by emi ( hamamatsu ) with a rectangular ( squared ) front surface of @xmath13 @xmath7 ( @xmath10 @xmath7 ) and a thickness of 5 mm .
the diodes were installed in the first part of the towers in order to sample the energy deposited in the first 8 @xmath14 .
they were in optical contact with the lateral side of the scintillator tiles and the light emitted in first part of the detector was therefore read - out twice since the photons crossing the lateral scintillator surface were collected by the diode while those reaching the fibers , either directly or after reflections , were seen by the tetrodes . due to the direct coupling ,
the light collection efficiency of the diodes was much larger than that of the tetrodes / apd s and this compensated for the absence of gain in the diodes .
most of the cells were equipped with emi vacuum photodiodes .
one diode prototype from hamamatsu , sampling only 4 @xmath14 , was successfully tested during the last part of the data taking .
technical characteristics of these devices are listed in table [ diodechar ] .
the hamamatsu prototype dimensions are such that it is possible to house two diodes in the same tower in order to obtain three longitudinal samplings .
for all diodes , the same front - end electronics and read - out chain as for the tetrodes were used .
the read - out electronics was positioned above the tower stacks ( cfr.fig.[install ] ) . .
technical characteristics of vacuum photodiodes . [ cols="<,<,<",options="header " , ] fig.[ediode ] shows @xmath15 versus @xmath16 for pions and electrons at 20 gev .
the discriminating power of the different variables in terms of pion contamination for 90% electron efficiency , at energies ranging from 20 to 50 gev is shown in fig.[episep ] . in most of the cases ,
purely calorimetric variables improve the overall separation capability with a factor @xmath17 compared with @xmath18 by itself . at 50 gev
the pion contamination for 90% electron efficiency is @xmath19 .
the present test has demonstrated the technical feasibility of longitudinally segmented shashlik calorimeters in which lateral sampling is performed by vacuum photodiodes . due to the small dimension of the diodes and to the tilt of fibers and diodes with respect to the incoming particles ,
no significant cracks or dead zones are introduced .
performance in terms of energy resolution , impact point reconstruction and @xmath0 separation seem to be adequate for applications at future @xmath1 collider experiments .
the ihep workshop staff has been essential for the construction of the prototype : we are greatly indebted with a. kleschov , p. korobchuk and a. tukhtarov .
we wish to thank also c. fanin for the mechanical project , v. giordano and r. cereseto for the careful work in the realization of the prototype , g. rampazzo for the invaluable effort in the construction of the read - out chain and all the staff and technical support of the sl - ea group for the smooth operation of the accelerator during the testbeam .
a special thank to c. luci for providing the simulation code at the early stage of this work and to m. pegoraro and g. zumerle for usefull suggestions and for granting the use of emi diodes .
j. badier et al . , nucl .
instr . and meth .
* a348 * ( 1994 ) 74 hera - b design report , desy / prc 95 - 01 .
lhcb technical proposal , cern / lhcc 98 - 4 .
alsvaag et al . ,
cern / ep 98 - 132 . h. fessler et al . , nucl .
instr . and meth . *
a240 * ( 1985 ) 284 .
atoyan et al . , nucl .
instr . and meth .
* a320 * ( 1992 ) 144 .
r. brinkmann , g. materlik , j. rossbach , a. wagner ( eds . ) , desy 1997 - 048 .
cms technical proposal , cern / lhcc 94 - 38 g.a .
akopdjanov et al . , nucl .
instr . and meth .
* 140 * ( 1977 ) 441 . | a novel technique for longitudinal segmentation of shashlik calorimeters has been tested in the cern west area beam facility .
a 25 tower very fine samplings e.m .
calorimeter has been built with vacuum photodiodes inserted in the first 8 radiation lengths to sample the initial development of the shower .
results concerning energy resolution , impact point reconstruction and @xmath0 separation are reported . , , , , , , , , , , , , , , . |
recent years have seen an uprising interest in thermodynamics at atomic scale @xcite due to the latest - generation manipulation of few , if not single , elementary quantum systems @xcite . in particular ,
out - of - equilibrium thermodynamics of quantum systems represents one of the most active research areas in the field @xcite . in this context , triggered by vast technological outcomes @xcite , the concept of quantum absorption thermal machine
@xcite has been reintroduced @xcite .
these machines are particularly convenient since they function without external work , extracting heat from thermal reservoirs through single atomic transitions to provide thermodynamic tasks ( e.g. , refrigeration ) .
nonetheless , fundamental issues remain unsolved .
a first one is the connection a single atomic transition to given thermal reservoirs , posing serious obstacles to practical realizations of such machines .
a second , more theoretical issue concerns the role of quantumness .
indeed , in typical models quantum features are not required @xcite , and only recently the advantages of quantum properties in thermal reservoirs have been pointed out @xcite .
the role of quantum features in the machines itself is debated @xcite , so that the advantages of quantum machines over standard ones remains partially unclear . in this paper
we address both of these open problems by introducing a new quantum thermal machine setting , based on an out - of - thermal - equilibrium ( ote ) electromagnetic bath naturally ( i ) coupling to each single atomic transitions , and ( ii ) creating quantum features in the machine .
the field is produced by macroscopic objects , and acts on each atomic transition as a different thermal bath at an effective temperature , hence providing all the elements needed for quantum absorption tasks .
this paper is structured as follows : the physical system is introduced in section [ system ] , along with the master equation governing its dynamics , while section [ thermodynamics ] is devoted to the introduction of thermodynamic quantities characterising the heat exchanges happening between atoms and field . in section [ tasks ] the first part of the results
is given , concerning the action of the machines , the different tasks it can produce and their intrinsic quantum origin .
the second part of the results of this work about the machine efficiency and its carnot limit is given in section [ efficiency ] .
finally , remarks and conclusions are drawn in section [ conclusions ] .
the setup of this paper is schematically depicted in fig . [ system ] , where a slab of thickness @xmath0 at temperature @xmath1 is placed in the blackbody radiation emitted by some walls at temperature @xmath2 .
the total electromagnetic field embedding the space between the slab and the walls is therefore given by the sum of four contribution : the direct blackbody radiation of the walls , the radiation emitted naturally by the slab and the walls radiation after being either reflected or transmitted by the slab .
such an ote field has been studied in the context of casimir - lifshitz force and heat transfer @xcite , where its properties have been characterised in terms of the field correlators through a scattering matrix approach .
the slab and the walls , macroscopic objects , are here the only ones directly connected to thermal baths .
in addition , a three - level atom @xmath3 ( machine ) and a two - level atom @xmath4 ( target body ) are placed at the same distance @xmath5 from the surface of the slab and spatially separated by a distance @xmath6 .
the atomic open system involves then four transitions : the body transition labeled as @xmath4 and the three machine transitions labeled as @xmath7 .
transition @xmath8 connects the two lowest - lying energy eigenstates ( red transition in fig . [ scheme ] ) and transition @xmath9 the two highest ones ( green transition in fig . [ scheme ] ) .
the ote field interacts with them through the hamiltonian @xmath10 , where @xmath11 is the dipole moment of the @xmath12-th transition of the atomic system and @xmath13 is the electromagnetic field at its position @xmath14 .
the total hamiltonian of the system is @xmath15 where @xmath16 is the hamiltonian of the ote field . in the following
we will not need the explicit expression of @xmath16 since only the field correlations will enter the master equation describing the dynamics of the atoms .
the free atomic hamiltonians @xmath17 and @xmath18 have expressions @xmath19 being @xmath20 ( @xmath21 ) the lowering ( raising ) operator of the body @xmath4 and @xmath22 , @xmath23 ( @xmath24 ) the lowering and raising operators of @xmath3 .
@xmath25 here represents a shift of the energy of each level in the @xmath26th transition due to the local interaction with the field and @xmath27 . in the third equality the renormalised transition frequencies @xmath28 for @xmath3 and @xmath4
have been introduced to account for the effects of the shifts @xmath25 . throughout this work
we will always assume the physical consequences of such frequency renormalisation to be negligible , such that @xmath29 .
this assumption has been fully confirmed by extended numerical simulations , having always detected the relative error introduced by neglecting these shifts to be less than @xmath30 .
it is worth stressing here that , differently from previous works on atomic - scale thermal machines @xcite , each atomic transition interacts here with _ the same _ electromagnetic field , which embeds all the space where the atoms are placed . as we will show in what follows
, there is then no need to conceive different environments , each interacting with a single atomic transition : a single non - equilibrium electromagnetic field is here able to produce all the physics needed for quantum thermodynamic tasks . in @xcite
the master equation ( me ) for two emitters in such a field has been derived under the markovian limit as @xmath31+d_{b}(\rho)+\sum_{n=1}^3d_n(\rho)+d_d(\rho),\ ] ] where @xmath32 .
@xmath33 is an effective _ field - mediated _ dipole interaction coupling resonant atomic transitions .
here we assume @xmath4 and @xmath3 to be resonant through transitions at frequency @xmath34 , and all their dipoles to have the same magnitude and to lie along the line joining the two atoms , and oriented from @xmath4 to @xmath3 .
@xmath35 is the effective interaction strength and @xmath20 ( @xmath36 ) is the lowering operator of the body ( of the resonant transition of the machine ) .
@xmath37 originates from the correlations of the fluctuations of atomic dipoles due to the common field .
the derivation of the master equation has been performed under the markovian and rotating wave approximations .
it involves the average photon number @xmath38 at frequency @xmath39 and temperature @xmath40 and the two functions @xmath41 which encompass all the properties of the environment , such as the dielectric properties of the slab and the correlation functions of the field .
for their explicit expressions we refer the interested reader to @xcite . the dissipative effects due to the atom - field coupling
are accounted for by the dissipators @xmath42 with expressions @xmath43
@xmath44 @xmath45 where @xmath46 .
one recognises standard local dissipation terms ( @xmath47 and @xmath48 ) , each associated to the degrees of freedom of a well - identified atom , and non - local dissipation ( @xmath49 ) which describes energy exchanges at frequency @xmath34 of the atomic system _ as a whole _ with its ote environment , not separable in machine or body contributions , its action involving degrees of freedom of both atoms in a symmetric way .
the parameters @xmath50 ( the rates of the dissipative processes of absorption and emission of photons through local or non - local interactions ) depend on local or non - local correlations of the field in the atomic positions , which in turn are functions of the temperatures @xmath1 and @xmath51 and the dielectric properties of the slab @xmath52 as @xmath53\alpha_w^i(\omega)\nonumber \\ & + & \big[1+n(\omega , t_s)\big]\alpha_s^i(\omega),\\ \frac{\gamma^-_i(\omega)}{\gamma^0_i(\omega)}&=&n(\omega , t_w)\alpha_w^i(\omega)^*+n(\omega , t_s)\alpha_s^i(\omega)^*,\end{aligned}\ ] ] where @xmath54 , for @xmath55 , is the vacuum spontaneous emission rate of the @xmath12-th atomic transition having a dipole moment @xmath11 , and @xmath56 .
thanks to the functional dependence of these parameters on the frequency and on the position of the atom , and to the critical behaviour shown in correspondence to the resonance frequency @xmath57 of the slab material , thermodynamic tasks become achievable . to simplify the notation , in the rest of this work the explicit @xmath39-dependence in all the @xmath58s
will be omitted .
after having introduced all the dynamic effects characterizing the atomic system , we want in this section to introduce some quantities which will characterize the machine tasks and functioning . in order to describe the machine thermodynamics , it is convenient to introduce two kind of temperatures . a first one characterizes the action of the field on the atoms : it has been shown @xcite that the atom - field interaction can be effectively rewritten as if each atomic transition felt a local _ equilibrium _ environment whose temperature depends on the transition frequency , on the properties of the slab and on the slab - atom distance @xmath5 .
these effective _ environmental temperatures _ depend on the rates @xmath59 as @xmath60 with @xmath61 .
it is important to stress here that , despite these effective environments can be characterised by a temperature , their spectra are not simply blackbody spectra as they have their own transition - dependent purcell factor @xcite.in this framework we study thermodynamic effects of _ stationary _ heat fluxes between @xmath3 and @xmath4 , mediated and sustained by the ote environment . to characterize the effects of these fluxes a second kind of temperature has to be introduced
indeed , as much as the environmental temperatures characterise the thermodynamics of the ote field , we need a second parameter to describe the energetics of atoms . in particular , atoms exchange energy under the form of heat with their surroundings by emitting photons through one of their transitions .
this means that the possibility of such heat exchanges is related to the distribution of population in each atomic level .
note that , from the very definition of @xmath62 , the environmental temperature depends on how the field tends to distribute atomic population in each pair of levels , due to the presence of the ratio @xmath63 .
a transition is therefore in equilibrium with its effective local environment if and only if its two levels @xmath64 and @xmath65 are populated such that @xmath66 . if not , the field and the atom will exchange heat along such a transition until such a ratio is reached .
this suggests to introduce a second temperature , hereby referred to as _ population temperature _ , which for a transition of frequency @xmath67 ( @xmath68 ) is defined as @xmath69 @xmath70 ( @xmath71 ) being the stationary population of the ground ( excited ) state of the @xmath72-th transition .
the result of a stationary thermodynamic task on the body , be it refrigeration , heating or population inversion , is then to modify its population temperature @xmath73 .
the condition @xmath74 is satisfied only if detailed balance ( @xmath75 ) holds .
it can be proven that detailed balance can be broken in a three level atom in ote fields . as a consequence the machine @xmath3 produces non - zero stationary heat fluxes with @xmath4 and the field environment , one for each dissipative process @xmath48 in the me .
these fluxes , following the standard approach in the framework of markovian open quantum systems @xcite , are given as @xmath76 $ ] , where @xmath77 is here the stationary atomic state and @xmath78 is a suitable atomic hamiltonian which can be @xmath17 , @xmath18 or @xmath79 depending on which part of the atomic system the heat flows into .
note that this definition implies an outgoing heat flux to be negative .
following their definition , these heat fluxes depend both on the field properties ( through the structure of the dissipators @xmath48 ) and on the properties of the atoms through their stationary state .
this dependence , for the local dissipators , can be put under the very clear thermodynamic form @xmath80 where @xmath81 , @xmath82 is a positive function of @xmath83 ( and of other parameters such as the frequency of the transition ) and the second approximated equality holds in the limit @xmath84 .
equation shows that the direction of heat flowing is uniquely determined by the sign of the difference @xmath85 , matching the thermodynamic expectation that heat flows naturally from the hotter to the colder body and strengthening the physical meaning of @xmath83 .
there being no time - dependence in the hamiltonian of the model , the first law of thermodynamics at stationarity for the total atomic system comprises only heat terms and assumes the form @xmath86 in addition , energy is exchanged between the machine and the body thanks to their field - induced interaction @xmath37 . in appendix
a , following the general scheme developed in @xcite , we show such an exchange to be under the form of heat .
seen by @xmath3 such a flux is @xmath87\big)/\hbar$ ] while as expected @xmath4 sees the flux @xmath88 . by introducing the explicit expressions for @xmath17 and @xmath37 , one can obtain a particularly simple form for @xmath89 as @xmath90 where @xmath91 . in an analogous way , by employing eq .
, one can evaluate the change in internal energy of @xmath3 due to the non - local heat flux exchanged by the atomic system with the ote environment , given by @xmath92 $ ] .
it is @xmath93\big\}.\ ] ] finally , the change in the internal energy of @xmath4 due to the same effect , @xmath94 $ ] , is given by same expression . fig .
[ scheme ] shows the full scheme of such heat fluxes for a particular configuration of the system .
the two levels of @xmath4 will be labeled here as @xmath95 and @xmath96 . despite the two - level assumption
might seem specific , it has been shown in various contexts @xcite that quantum thermal machines only couple to some effective two - level subspaces in the hilbert space of the body they are working on .
a two - level system is therefore the fundamental building block of the functioning of quantum thermodynamic tasks .
the main result of this paper is the possibility to drive the temperature @xmath73 of the body outside of the range defined by the external reservoirs at @xmath51 and @xmath1 .
the body , without the effect of the machine , would thermalise at the local environmental temperature ( @xmath97 ) , corresponding to @xmath98 .
this temperature is necessarily constrained within the range @xmath99 $ ] @xcite.due to the particular form of the master equation , in which all collective atomic terms involve only resonant atomic transitions , in the non - resonant subspace the collective atomic state will be diagonal in the eigenbasis of @xmath100 .
this is due to the fact that local dissipation in eq .
of section [ system ] induces a thermalisation with respect to the free atomic hamiltonians . on the other hand , in the resonant atomic subspace of the eigenbasis of @xmath100 spanned by the states @xmath101 of @xmath4 and the two states @xmath102 of the transition of @xmath3 at frequency @xmath34 ,
the most general form of the atomic stationary state is @xmath103 a coherence @xmath104 is present in the decoupled basis between the two atomic states @xmath105 and @xmath106 having the same energy .
note that the temperature @xmath73 of the body increases monotonically with the ratio @xmath107 . by tracing out the machine degrees of freedom from the master equation ,
one obtains a diagonal state with stationary populations @xmath108 and @xmath109 of the body @xmath4 .
be now @xmath110 the flux @xmath111 seen by @xmath4 .
then the expressions for heat fluxes exchanged by @xmath4 with its surroundings are @xmath112\big\rangle,\label{dotqrq}\\ \dot{q}_d&=&\gamma_{d}^+\big[\big\langle\sigma^{\dag}h_b\kappa_{r}\big\rangle-\frac{1}{2}\big\langle\{h_b,\sigma^{\dag}\kappa_{r}\}\big\rangle\big]\nonumber\\ & + & \gamma_{d}^-\big[\big\langle\sigma h_b\kappa_{r}^{\dag}\big\rangle-\frac{1}{2}\big\langle\{h_b,\sigma\kappa_{r}^{\dag}\}\big\rangle\big]+c.c.\label{qqtq},\\ { \dot{q}}_b&=&\gamma_b^+\big[\big\langle \sigma^{\dag } h_b \sigma \big\rangle-\frac{1}{2}\big\langle
\ { h_b,\sigma^{\dag}\sigma \ } \big\rangle\big]\nonumber \\ & + & \gamma_b^-\big[\big\langle \sigma h_b \sigma^{\dag } \big\rangle-\frac{1}{2}\big\langle \ { h_b,\sigma\sigma^{\dag } \ } \big\rangle\big],\label{qb}\end{aligned}\ ] ] where the mean values are evaluated over the stationary state of the total system . exploiting its general form , it is just a matter of straightforward calculations to evaluate all the mean values above . imposing the sum of , and to vanish ( first law for @xmath4 , analogous to eq . ) , one obtains @xmath113 where @xmath114\big\}.\ ] ] note now that , thanks to eq . , @xmath115 and @xmath116 , such that the first term in @xmath117 stems from the resonant heat @xmath118 exchanged with the machine , while the second is due to the non - local heat flux @xmath119 .
show that the thermal machine works _ only if a stationary quantum coherence @xmath104 is present_. remarkably , it can be shown @xcite that quantum discord @xcite ( a key measure of purely quantum correlations ) is a monotonic function of the absolute value of the coherence @xmath104 in our system . differently from previous studies @xcite , here
discord between @xmath3 and @xmath4 is a necessary condition for any thermodynamic task , and represents a resource the machine can use through the two different processes @xmath89 and @xmath119 .
. means that a quantum coherence between machine and body modifies the stationary temperature of the body with respect to @xmath120 .
this modification is reported in fig .
[ temp1 ] , where the behaviour of @xmath73 as a function of the slab - atoms distance @xmath5 is shown for two different slab thicknesses @xmath0 .
four possible regimes can be singled out : both during refrigeration ( @xmath121 ) and heating ( @xmath122 ) , @xmath73 can be either driven outside of the range @xmath99 $ ] ( strong tasks ) or kept within it ( light tasks ) . as a limiting case of strong heating
, the body can be brought to infinite temperature ( @xmath123 ) and , further on , to negative ones , producing population inversion . of the body ( solid black line ) , machine resonant temperature @xmath124 in the absence of @xmath4 ( dotted pink line ) and local - environment temperature @xmath120 felt by the body ( dot - dashed green line ) versus @xmath5 .
the slab is made of sapphire and kept at @xmath125 while @xmath126 .
the machine transition frequencies are @xmath127 , @xmath128 and @xmath129 , @xmath130 being the first resonance frequency of sapphire ( the optical data for the dielectric permittivity of the slab material are taken from @xcite ) .
the two atoms are placed at a distance @xmath131 from each other .
panel ( a ) : numerical data for a semi - infinite slab .
light and strong refrigeration are achieved in this configuration .
panel ( b ) : same quantities for a slab of finite thickness .
the plotted functions are in this case @xmath132 , @xmath133 and @xmath134 ( left vertical scale ) , with the same color code as before .
the population inversion corresponds to divergent @xmath73 and @xmath124 .
the corresponding value of temperature can be read on the right vertical scale .
all tasks are in this case obtained . ] as one can easily see from fig .
[ temp1 ] , the physics behind the absorption tasks is enclosed in the strong sensitivity of the population temperature @xmath73 of the body to the population temperature @xmath124 of the machine along the resonant transition when the body is not present . it is shown in fig . [ temp1 ] that the machine has a very high thermal inertia , such that the body , when put into thermal contact with the machine having a certain temperature @xmath124 , thermalizes with it and @xmath135 .
[ temprate ] shows the mechanism the machine uses to modify its population temperature @xmath124 in absence of the body , thanks to the different environmental temperatures each of its transition feels .
this drives @xmath3 out of detailed balance condition and allows @xmath3 to keep its resonant transition temperature almost constant .
we label here the three transitions of the machine as high frequency ( @xmath136 ) , average frequency ( @xmath137 ) and low frequency ( @xmath138 ) , one of which ( suppose here @xmath138 , connecting states @xmath139 and @xmath140 ) is resonant with @xmath34
. for simplicity , let us focus on refrigeration only , which we suppose to happen either through transition 2 ( connecting first and the second excited states ) , since in this configuration the high - frequency transition @xmath141 is always used by an absorption refrigerator to dissipate heat into the environment @xcite .
as shown in fig .
[ temp1 ] , to obtain a low @xmath73 the resonant machine transition must be made cold .
this is achieved by reducing the ratio @xmath142 , which in turn happens when : + ( a ) the effective environmental temperature @xmath143 felt by the high frequency transition is very cold . in this way
the environment contributes in increasing the population of the ground state of @xmath3 at the expenses of the population of its most energetic state .
the resonant transition involves necessarily one of these two levels , and in both cases the effect of the high frequency transition helps reducing @xmath142 ; + ( b ) the effective environmental temperature felt by the average transition is very hot .
this , following the same idea , would either mean reducing the population of @xmath144 or increasing the one of @xmath145 , thus reducing @xmath142 .
+ , @xmath146 and @xmath147 of the local environments felt by the three machine transitions , for @xmath148 , @xmath149 and @xmath150 ( @xmath151rad@xmath152 ) versus the machine - slab distance @xmath5 in the absence of the body . the slab and walls
temperatures are @xmath153k and @xmath154k , and the slab thickness is @xmath155 m .
as the plot shows , the transition having the same frequency as the slab resonance is much more strongly affected by the field emitted by the slab , such that its rate temperature is kept much lower than @xmath146 .
this produces a mechanism , shown in the inset , according to which excitations ( yellow dots ) are transferred to the intermediate level of the machine and removed from its upper one .
this in turn drives the population temperature @xmath156 of the transition at @xmath138 ( transition 2 in the example ) to values lower than the one @xmath147 of its local environment , allowing the machine to refrigerate objects . in this configuration , introducing a body @xmath4 at @xmath157 m from the slab and @xmath158 m from @xmath3 , one obtains @xmath159k @xmath160 . ]
when these two conditions are met , the machine can always redistribute its populations such that the ratio @xmath142 can be kept low and almost unaffected by the presence of another atom .
the advantage of the ote field configuration is that the effective field temperatures can be manipulated through a wide set of parameters involving @xmath5 , @xmath0 , @xmath51 and @xmath1 . in particular
, the role of the resonance of the slab material is crucial @xcite , as explained in the caption of fig .
[ temprate ] . in the case
@xmath161 , transitions strongly affected by the field emitted by the slab feel a cold local environment .
moreover , provided @xmath137 is far enough from @xmath57 , one can at the same time have @xmath162 . by this mechanism
, @xmath3 can change the temperature @xmath73 , bringing it to values far outside the range @xmath163 $ ] .
we stress here that the difference between light and strong tasks is a fundamental one : better than light tasks could in principle be done by direct connection of the body to one of the two real reservoirs at @xmath1 or @xmath51 , while strong tasks can not be achieved by a simple thermal contact with anything in the system.@xmath117 strongly depends on the slab - matter system distance @xmath5 and on the external temperatures through @xmath104 , @xmath164 and @xmath165 .
one can thus engineer one or many of these regimes at will as shown in the functioning - phase diagram of the machine in fig .
[ phases ] for a fixed thickness @xmath166 .
all the strong and light functioning phases of the machine are found as a function of both @xmath167 and @xmath5 . and @xmath168 .
the sapphire slab has a thickness @xmath166 and its temperature is continuously changed in the range @xmath169 $ ] .
the walls are at @xmath126 .
strong refrigeration ( blue areas ) , strong heating ( red areas ) , light refrigeration ( cyan ) , light heating ( orange ) and population inversion ( green ) can all be obtained . ]
consider now the refrigerating regime in which the machine extracts heat from the body through the transition @xmath9 .
the scheme of heat fluxes is then exactly the one depicted in fig .
[ scheme ] .
the efficiency of this process is @xmath170 due to the fact that @xmath89 is the power produced by the machine , which absorbs energy from its surroundings through transitions @xmath8 and @xmath9 ( the equivalent of a work input ) while uses transition @xmath141 to dissipate part of the absorbed energy after use ( the equivalent of the spiral in a normal fridge ) .
the corresponding carnot limit @xmath171 can be obtained by analysing the machine functioning in its reversible limit ( zero entropy production ) .
the instantaneous entropy production rate @xmath172 for quantum systems is defined as @xcite @xmath173 where @xmath174 is the so - called relative entropy @xcite , never increasing in time under a markovian dynamics .
following @xcite , one can apply equation term by term to each dissipator in the master equation thanks to the fact that they all are under a markovian form .
one thus obtains @xmath175\geq 0,\ ] ] where @xmath176 , @xmath177 is the kernel ( stationary state ) of the single dissipator @xmath42 .
the @xmath141 local dissipators @xmath48 of the machine and the local dissipator @xmath47 of the body induce stationarity under the standard gibbs form at the effective environmental temperature , diagonal in the free atomic hamiltonian basis .
the nonlocal dissipator @xmath49 , in the case studied here where the dipoles of @xmath4 and @xmath3 lie along the line connecting the two atoms ( and more in general when @xmath178 ) , has the same kernel at environmental temperature @xmath179 , local in the degrees of freedom of @xmath3 and the @xmath4 . introducing these single - dissipator stationary states into equation one obtains @xmath180 which is a form of the second law at stationarity for our system . with the help of the first law in eq . of section [ thermodynamics ] , the known property of three - level atomic heat fluxes @xcite @xmath181 @xmath182 ( where @xmath183 is the total flux along the @xmath72-th transition ) and the fact that in refrigeration @xmath184 and @xmath185 ( as commented in section [ optimal ] ) and under the condition @xmath186 ( other cases can be treated analogously ) , one obtains from another first degree inequality
this has a non - trivial solution only if @xmath187 , from which a bound on the efficiency in eq .
can be obtained as shown in appendix b. such a bound depends only on the three frequencies of the machine and the temperatures of the effective local and non - local environments . in the case of refrigeration along transition @xmath9 the carnot efficiency assumes the form @xmath188 an important figure of merit for the realistic functioning of any thermal machine is how close to its carnot limit it works when delivering maximum power ( i.e. , when @xmath89 is maximised ) .
many bounds are known for different setups , limiting the efficiency at maximum power @xmath189 to some fractions of @xmath190 @xcite .
remarkably our structured ote environment allows for refrigeration tasks with @xmath189 much closer to @xmath190 than the bound known for quantum absorption machines @xcite based on ideal blackbody reservoirs , reading for our system @xmath191 .
this is exemplified in fig .
[ effmax1 ] for a particular configuration of the model .
the blue triangles ( left vertical scale ) represent the ratio @xmath192 , plotted versus @xmath193 while keeping fixed @xmath194 .
the red dots ( right vertical scale ) are the power @xmath89 plotted versus the same quantity , while the red dashed line is the machine - body discord ( right vertical scale ) .
it is clear that the power is maximised at @xmath195 , corresponding to @xmath196 .
@xmath89 starts decreasing , as classically expected when the efficiency approaches @xmath190 , around @xmath197 , but suddenly increases again when @xmath193 approaches @xmath57 .
this behaviour is due to the fact that , when one atomic transition is resonant with the characteristic frequency of the slab material , the atomic populations are strongly affected by the field emitted by the slab . hence the not - black - body nature of the total field become crucial ( e.g. , the atomic decay rate is no longer proportional to @xmath198 ) , allowing to overcome bounds set by the blackbody physics .
the role of discord as machine resource is clearly shown here , where discord at resonance has a sharp peak leading to the high - power performance of @xmath3 .
( blue triangles , left vertical axis ) , the power of the machine ( @xmath89 , red dots , right vertical axis in units of @xmath199 ) and machine - body discord ( dashed red line , right vertical scale in units of @xmath200 ) versus the scaled machine transition frequency @xmath201 .
the sapphire slab is semi - infinite and at @xmath202 , while @xmath203 and @xmath204 .
the transition frequency of the body is fixed as @xmath205 ( @xmath130 ) , resonant with transition @xmath9 of the machine .
the maximum power is reached for @xmath195 , corresponding to @xmath206 .
remarkably , discord shows a sharp resonance peak , similarly to @xmath89 . ]
one could wonder whether such an exceptionally high efficiency at maximum power is only seldomly attained for the kind of machines described here .
to answer such a question on quantitative bases , we performed a random sampling of over @xmath207 thermal machines , all delivering thermodynamic tasks on the same fixed body . in the simulations performed and reported in fig .
[ effmax2 ] , the machines work as a quantum refrigerator delivering strong refrigeration using a semi - infinite slab . in this sampling ,
the machine - slab distance @xmath5 has been , for each machine , randomly drawn in the range @xmath208 $ ] , the walls temperature has been selected randomly in @xmath209 $ ] and , for each value of @xmath51 , the slab temperature has been chosen at random in @xmath210 $ ] .
the internal structure of the body is kept fixed during the simulations , with a frequency @xmath211 resonant with the transition 2 of @xmath3 . for each machine thus generated , we have then maximised the delivered power by modifying the two other machine frequencies over every possible value of @xmath212 and @xmath193 compatible with the condition @xmath213 . finally , once obtained the configuration corresponding to the maximum power , we have computed the efficiency of the process . fig . [ effmax2 ]
shows the histogram of the distribution of the ratio @xmath214 of efficiency at maximum power to the corresponding carnot efficiency in the interval @xmath215 $ ] within these @xmath207 random refrigerators .
it is remarkable that around @xmath216 of these machines work at maximum power with efficiencies higher than the bound @xmath217 in @xcite and that none of them have been found to work at maximum power with efficiencies lower than @xmath218 .
moreover , as can be clearly seen in fig .
[ effmax2 ] , a small but non - negligible fraction of them can reach @xmath219 . for a random sampling of @xmath207 thermal machines , always in resonance with the same body @xmath4 . for each machine ,
@xmath5 has been randomly generated in the range @xmath208 $ ] , @xmath209 $ ] and , for each value of @xmath51 , @xmath210 $ ] .
the internal structure of the body is kept fixed during the simulations , with a frequency @xmath211 resonant with the transition 2 of @xmath3 .
the maximisation is performed over every possible value of @xmath212 and @xmath193 compatible with the condition @xmath213 . around @xmath216 of machines thus generated
have @xmath220 . ]
this work introduces a new realization of a quantum thermal machine using atoms interacting with single non - equilibrium electromagnetic fields . by simply connecting two thermal reservoirs to _ macroscopic objects _ , their radiated field allows the atomic machine to achieve all quantum thermodynamic effects ( heating , cooling , population inversion ) , without any direct external manipulation of atomic interactions .
this overcomes the usual difficulty of connecting single transitions to thermal reservoirs , in a realistic and simple configuration where the field - mediated atomic interaction modifies at will stationary inter - atomic energy fluxes . despite the environmental dissipative effects , atoms share steady quantum correlations @xcite which we showed to be necessary for one atom to deliver a thermodynamic task on the other , uncovering genuinely non - classical machine functioning .
these particular features affect the tasks efficiency , which can be remarkably high also at maximum power , defying the known bounds for quantum machines based on ideal and independent blackbody reservoirs thanks to the fundamental effect of the resonance with the real material of which the slab is made .
moreover , such a remarkably high efficiency at maximum power is strongly connected to the presence of a peak in quantum correlations between the machine and the body , which represent the resource the machine uses for its tasks .
these results tackle major open problems on quantum thermal machines , paving the way for an efficient quantum energy management based on the potentialities of non - equilibrium and quantum features in atomic - scale thermodynamics .
the authors acknowledge fruitful discussions with n. bartolo and r. messina , and financial support from the julian schwinger foundation .
in this appendix we demonstrate that the resonant energy exchange between @xmath3 and @xmath4 due to the field - mediated coherent interaction @xmath37 consists only of heat . following the approach of @xcite ,
the dynamics of the sole @xmath3 induced by the hamiltonian interaction @xmath37 comprises in general an hamiltonian and a dissipative part and can be written as @xmath221+d_{mb}(\rho),\ ] ] where @xmath222 is a non - unitary dissipative term for @xmath3 due to the interaction with @xmath4 , which depends however on the total state @xmath223 because , in general , the two subparts are correlated .
@xmath224 is a renormalised free hamiltonian of subsystem @xmath3 due to the interaction with @xmath4 . defining the two marginals @xmath225 and the correlation operator @xmath226 , it is shown in @xcite that @xmath227\big).\label{dab}\end{aligned}\ ] ] introducing @xmath228 as the part of @xmath224 which commutes with @xmath17 and @xmath229 which does
not , directly from equation one has , for the internal energy of @xmath3 @xmath230 , @xmath231\rho_m\big ) .
\end{split}\ ] ] it is custom to identify heat terms as the ones producing a change in the entropy of a subsystem : all the rest is identified as work @xmath232 .
eq . can then be split in @xmath233\rho_m\big).\label{dotwa}\end{aligned}\ ] ] introducing the symbols @xmath234 ( @xmath235 ) for the coherences _ of the marginal _ @xmath236 ( different then from the coherence @xmath104 introduced in eq . of section [ tasks ] which is a two - atom coherence )
, equation becomes @xmath237 by tracing out the machine or the body degrees of freedom from equation , one can prove that the two stationary marginals @xmath236 and @xmath238 are always diagonal in the eigenbases of their respective free hamiltonians , so that @xmath239 .
no renormalisation to the machine hamiltonian comes therefore from the interaction with @xmath4 , which means that equation vanishes , proving that no work is involved in machine - body energy exchanges .
as for the heat , considering that @xmath240=0 $ ] , eq . reduces to @xmath241\big)={\dot{q}}_r,\ ] ] with the same @xmath89 given in eq . .
in this appendix we deduce eq . of section [ efficiency ] for the carnot efficiency in refrigeration along transition 2 , and under the condition @xmath186 .
in addition to eqs . and , the condition @xmath181 gives for @xmath242 and @xmath243 the following @xmath244 solving eqs . and for @xmath245 and @xmath119 and using these solutions into one obtains for @xmath89 @xmath246-{\dot{q}}_2\ ] ] which , used in eq .
of section [ efficiency ] , gives a bound on @xmath247 as a function of @xmath248 and @xmath249 .
finally , using the fact that such a bound is a decreasing function of @xmath249 , one obtains the carnot efficiency as the limit for @xmath250 , which turns out to be independent on @xmath248 and gives ultimately the first line of eq . .
on the other hand , in the case @xmath251 , one can not obtain anything like eq . and
the only possibility for the machine to work without producing entropy is therefore to have vanishing heat flux from / to the body .
this means @xmath252 which , inserted in the expression for the efficiency and using again @xmath181 leads to the second line of eq . .
32 j. gemmer , m. michel , and g. mahler , _ quantum thermodynamics - emergence of thermodynamic behavior within composite quantum systems _ ( springer , 2010 ) .
v. blicke and c. bechinger , nature physics * 8 * , 143 ( 2012 ) . n. brunner , n. linden , s. popescu , and p. skrzypczyk , phys . rev .
e * 85 * , 051117 ( 2012 ) .
m. horodecki and j. oppenheim , nature comm .
* 4 * , 2059 ( 2013 ) .
s. haroche , rev .
* 85 * , 1083 ( 2013 ) .
m. esposito , u. harbola , and s. mukamel , rev .
81 * , 1665 ( 2009 ) .
s. deffner and e. lutz , phys .
lett . * 107 * , 140404 ( 2011 ) .
b. leggio , a. napoli , h .-
breuer , and a. messina , phys .
e * 87 * , 032113 ( 2013 ) .
b. leggio , a. napoli , a. messina , and h .-
breuer , phys .
a * 88 * , 042111 ( 2013 ) .
o. abah and e. lutz , europhys . lett . * 106 * ,
20001 ( 2014 ) . m. o. scully , phys .
lett . * 104 * , 207701 ( 2010 ) .
p. hnggi and f. marchesoni , rev .
* 81 * , 387 ( 2009 ) . h. e. d. scovil and e. o. schulz - dubios , phys .
* 2 * , 262 ( 1959 ) .
n. linden , s. popescu , and p. skrzypczyk , phys .
lett . * 105 * , 130401 ( 2010 ) . a. levy and r. kosloff , phys .
* 108 * , 070604 ( 2012 ) .
l. a. correa , j. p. palao , d. alonso , and g. adesso , sci .
rep . * 4 * , 3949 ( 2014 ) .
d. venturelli , r. fazio , and v. giovannetti , phys .
lett . * 110 * , 256801 ( 2013 ) .
l. a. correa , j. p. palao , g. adesso , and d. alonso , phys .
e * 87 * , 042131 ( 2013 ) . n. brunner , m. huber , n. linden , s. popescu , r. silva , and p. skrzypczyk , phys . rev .
e * 89 * , 032115 ( 2014 ) .
m. antezza , l. p. pitaevskii , and s. stringari , phys .
a * 70 * , 053619 ( 2004 ) .
m. antezza , l. p. pitaevskii , and s. stringari , phys .
lett . * 95 * , 113202 ( 2005 ) ; j. m. obrecht , r. j. wild , m. antezza , l. p. pitaevskii , s. stringari and e. a. cornell , phys . rev . lett . * 98 * , 063201 ( 2007 ) .
m. antezza , j. phys .
a : math . gen . * 39 * , 6117 ( 2006 ) .
r. messina and m. antezza , phys .
a * 84 * , 042102 ( 2011 ) ; europhys . lett . * 95 * , 61002 ( 2011 ) . b. bellomo and m. antezza , europhys . lett .
* 104 * , 10006 ( 2013 ) ; new j. phys . *
15 * , 113052 ( 2013 ) . supplementaty material at _
supplementary material_. b. bellomo , r. messina , and m. antezza , europhys . lett . * 100 * , 20006 ( 2012 ) ; b. bellomo , r. messina , d. felbacq , and m. antezza , phys . rev .
a * 87 * , 012101 ( 2013 ) . s. de liberato and m. ueda , phys .
e * 84 * , 051122 ( 2011 ) .
breuer and petruccione , _ the theory of open quantum systems _
( oxford university , berlin , 2002 ) .
h. weimer , m. j. henrich , f. rempp , h. schrder , and g. mahler , europhys . lett .
* 83 * , 30008 ( 2008 ) .
d. spehner and m. orszag , j. phys .
a * 47 * , 035302 ( 2014 ) .
h. ollivier and w. h. zurek , phys .
. lett . * 88 * , 017901 ( 2001 )
. _ handbook of optical constants of solids _ , edited by e. palik ( academic press , new york , 1998 ) .
v. vedral , rev .
mod . phys . * 74 * , 197 ( 2002 ) .
f. curzon and b. ahlborn , am . j. phys .
* 43 * , 22 ( 1975 ) .
b. bellomo and m. antezza , arxiv:1409.7178 . | we propose a scheme for a quantum thermal machine made by atoms interacting with a single non - equilibrium electromagnetic field .
the field is produced by a simple configuration of macroscopic objects held at thermal equilibrium at different temperatures .
we show that these machines can deliver all thermodynamic tasks ( cooling , heating and population inversion ) , and this by establishing quantum coherence with the body on which they act .
remarkably , this system allows to reach efficiencies at maximum power very close to the carnot limit , much more than in existing models .
our findings offer a new paradigm for efficient quantum energy flux management , and can be relevant for both experimental and technological purposes . |
ultracold atomic gases are novel systems with a high degree of controllability , making them very useful in the studies on quantum phenomena . the possibility to control
the parameters during experiments is essential , for example , in quantum information processing ( e.g. , refs .
@xcite ) and matter - wave interferometry ( e.g. , refs . @xcite ) . in this paper
, we discuss the dynamics of an ultracold bosonic gas trapped in a time - periodically modulated potential .
the dynamics of periodically modulated quantum systems has attracted both theoretical ( e.g. , refs .
@xcite ) and experimental ( e.g. , refs .
@xcite ) interest during recent years .
it is known that a modulated system has resonances at which the tunneling is either suppressed or enhanced . in the neighborhood of a resonance ,
the behavior of the system is very sensitive to the modulation frequency .
the suppression of tunneling by modulating the energy difference between the modes is known as the coherent destruction of tunneling ( cdt ) @xcite .
cdt was discovered in ref .
@xcite , where the motion of a charged particle in a lattice under the influence of an oscillating electric field was studied .
it was shown that an initially localized particle remains localized in a one - dimensional lattice if the amplitude and frequency of the electric field are chosen suitably . in ref .
@xcite , cdt was found to occur in systems consisting of a particle subjected to a periodic force and trapped in a double - well potential .
recently , the coherent destruction of tunneling has been actively studied in the context of ultracold bosonic atoms ( e.g. , refs .
@xcite ) .
unlike the cdt , which is typically observed under the condition that the tunneling coupling is larger than or comparable to the interaction energy , the enhancement of tunneling by modulation can take place in a system where the interaction energy dominates over the tunneling coupling . in the absence of modulation , the large interaction energy suppresses tunneling for energetic reasons @xcite .
this leads to a very long tunneling period ( the time needed for @xmath0 particles to tunnel from one mode to another and back ) .
however , by modulating the tunneling matrix element , it is possible to enhance the many - particle tunneling and thereby reduce the tunneling period @xcite . in this paper
, we analyze the reasons behind the enhanced tunneling with the help of a detailed floquet analysis . in order to make the analysis more complete , we consider also systems where either the interaction strength or the energy difference between the modes is modulated .
we find that these two methods provide an alternative way to enhance tunneling .
it is shown that the width of the resonance , that is , the range of modulation frequencies corresponding to the enhanced tunneling , depends strongly on whether the tunneling matrix element , the interaction strength , or the energy difference between the two modes is modulated ; the resonance is wider in the former two cases .
we explain this difference with the help of floquet theory and the eigenvalues of the non - modulated hamiltonian .
we analyze also the coherent destruction of tunneling using floquet theory .
it is known that cdt can be caused by modulating either the energy difference between the modes or the interaction strength .
we study only the floquet spectrum of the former system because it has not received much attention in the literature , unlike the floquet spectrum of the interaction - modulated system @xcite .
in addition to this , we present two ways to generate noon - like ( schrdingers - cat like ) states .
the first is based on the cdt induced by a large - amplitude modulation of interaction strength , whereas the second makes use of a small - amplitude modulation of the tunneling coupling .
this paper is organized as follows . in sec .
[ sec_bh ] , we define the modulated bose hubbard hamiltonian . in sec .
[ sec_floquet ] , we give a short summary of the floquet theory used in this article .
section [ sec_selftrapping ] discusses in depth the results of the floquet analysis for systems in the self - trapping regime subjected to a small - amplitude modulation . in sec .
[ sec_josephson ] , the coherent destruction of tunneling is examined using floquet theory .
it is also shown that noon states can be created with the help of partial cdt . in sec .
[ sec_noon ] , a way to create noon states using adiabatic sweep across an avoided crossing is presented . finally , the conclusions are in sec .
[ sec_conclusions ] .
we consider a system described by the two - mode bose - hubbard hamiltonian . for definiteness , we assume that this model is realized physically by bosons trapped in a double - well potential .
we consider situations where either the tunneling amplitude , the interaction strength , or the energy difference between the wells is modulated periodically in time .
this system is described by the hamiltonian @xmath1 here @xmath2 is the tunneling matrix element , @xmath3 is the on - site interaction , and @xmath4 is the energy difference between the wells ( tilt ) .
we have introduced the su(2 ) generators defined as @xmath5 where @xmath6 annihilates ( creates ) an atom in mode @xmath7 .
we define the time - dependent tunneling matrix element as @xmath8 , \label{eq_jt}\ ] ] where @xmath9 is the amplitude of the time - independent part and @xmath10 $ ] gives the relative amplitude of the time - dependent tunneling matrix element . the modulated tilt and interaction strength
are defined as @xmath11 where @xmath12 ( @xmath13 ) are the amplitudes of the static ( time - dependent ) part of the interaction and the tilt , respectively . in the above equations ,
@xmath14 is the modulation frequency and @xmath15 , and @xmath16 are the phase offsets . in this paper , time is measured in units of @xmath17 which is the tunneling period in the absence of the interaction ( @xmath18 ) and the tilt ( @xmath19 ) . here and in what follows , we set @xmath20 .
if the hamiltonian @xmath21 is periodic in time , floquet theory provides a powerful tool to analyze the dynamics of the system . in the following ,
we denote the modulation period of @xmath21 by @xmath22 . in our case ,
the modulation is sinusoidal and hence @xmath23 . according to the floquet theorem ( see , e.g. , ref .
@xcite ) , the time - evolution operator @xmath24 determined by the hamiltonian of eq .
( [ eq_h ] ) can be written as @xmath25 where @xmath26 is a periodic matrix with minimum period @xmath22 and @xmath27 and @xmath28 is a time - independent operator .
we define the floquet operator @xmath29 as @xmath30}\right\ } , \end{aligned}\ ] ] where @xmath31 is the time - ordering operator . at times
@xmath32 , where @xmath33 is an integer , we get @xmath34 .
the floquet operator is a mapping between the state at @xmath35 and the state after one modulation period @xmath23 : @xmath36 .
the columns of the floquet operator @xmath29 can be obtained by following the time evolution of the basis states for one modulation period .
each time - evolved basis state forms a column of the matrix @xmath29 .
the hilbert space of a two - mode system containing @xmath0 bosons is @xmath37 .
the basis of this hilbert space can be chosen to be @xmath38 , where @xmath39 is a state with @xmath40 particles in mode @xmath41 and @xmath42 particles in mode @xmath43 .
any pure state of the system can be written as @xmath44 where the amplitudes @xmath45 are complex numbers .
if @xmath0 is even ( odd ) , @xmath46 takes only even ( odd ) values . in order to characterize the eigenstates of the floquet operator
, we define the parity operator @xmath47 as @xmath48 it can alternatively be written as @xmath49 .
the eigenvalues of @xmath47 are @xmath41 and @xmath50 , corresponding to even and odd parity , respectively . because @xmath51 , the hamiltonian , and consequently the time - evolution operator , commutes with @xmath47 if the tilt vanishes .
then the eigenstates of @xmath29 are also eigenstates of @xmath47 and either @xmath52 or @xmath53 holds for the components of the eigenvectors of @xmath29 . in the former case ,
the eigenstate has even parity and is said to be symmetric , while in the latter case the parity is odd and the eigenstate is called antisymmetric . furthermore , the absolute values of the coefficients @xmath54 of an eigenstate have maxima at @xmath55 , where @xmath56 is an integer .
we denote such an eigenstate by @xmath57 , where @xmath58 @xmath59 indicates that the eigenvector is symmetric ( antisymmetric ) . if @xmath60 and the signs of the eigenvectors are defined appropriately , we get @xmath61 which is valid to zeroth order in @xmath62 . for non - zero @xmath63 or @xmath64 ,
the floquet eigenstates are neither exactly symmetric nor antisymmetric because @xmath65 is not invariant under the parity operator .
however , since the time average of the @xmath65 term is zero ( we assume that @xmath66 ) , the floquet eigenstates are almost symmetric or antisymmetric , provided @xmath64 is small .
we thus use the notation @xmath67 also in this case . note that , in the case of large - amplitude modulation of the tilt , the floquet eigenstates can not be classified in this way .
the floquet operator is a unitary operator , and therefore the eigenvalue corresponding to the eigenvector @xmath68 can be written as @xmath69 .
the eigenvalue equation becomes @xmath70 in this paper , we call @xmath71 the phase of a floquet eigenvalue
. assume that the initial state is @xmath72 .
at @xmath32 ( @xmath73 ) , the state reads @xmath74}\psi_n^{(-)}\right).\end{aligned}\ ] ] if @xmath75 , we get @xmath76 ; that is , the system has tunneled from @xmath77 to @xmath78 . in this paper , we define the tunneling period as the time needed for the system to tunnel from @xmath77 to @xmath78 and back . in terms of the phases of the floquet eigenvalues , the tunneling period reads @xmath79 increasing @xmath80 reduces the tunneling period and vice versa .
when @xmath81 , the tunneling period diverges .
in this section , we consider the tunneling of bosons in the self - trapping regime characterized by @xmath82 .
assume that in the initial state all @xmath0 particles are either in site 1 or site 2 .
the reduction of the interaction energy by single - particle tunneling is of order @xmath83 .
this reduction can not be compensated by the increase of the kinetic energy , which is approximately given by @xmath84 . as a consequence , single - particle tunneling is suppressed ( self - trapping ) , and all @xmath0 particles stay in the same well for a long time . in this situation , oscillations between the states @xmath77 and @xmath78 occur through higher - order co - tunneling @xcite . in ref . @xcite it was found that the tunneling period of the higher - order co - tunneling can be drastically reduced by modulating the tunneling matrix element @xmath2 . as we show here
, a similar phenomenon can be seen when the tilt is modulated ( we set @xmath66 ) . in figs .
[ fig_modj ] and [ fig_modv ] , we show the tunneling period @xmath85 for the modulated tunneling matrix element and tilt , respectively . as an example
, we have set @xmath86 and @xmath87 in the both cases .
we see that the behavior of @xmath85 as a function of the modulation frequency @xmath14 depends strongly on whether @xmath2 or @xmath4 is modulated .
this difference can be explained using floquet analysis . before analyzing the system in detail , we first summarize two key points .
one is the parity of the operator whose coefficient is modulated , and the other is the shift in the phases of the floquet eigenvalues due to an avoided crossing .
the parity of @xmath88 is even and that of @xmath65 is odd .
therefore , @xmath88 couples floquet eigenstates of the same parity , and @xmath65 couples those of the opposite parity .
this means that in the case of modulated @xmath2 [ @xmath4 ] , there is an avoided crossing between floquet eigenstates of the same [ opposite ] parity .
the differences in the behavior of the tunneling period can be attributed to the parities of the states undergoing an avoided crossing .
below we show that usually @xmath90 ( @xmath91 ) holds for odd ( even ) @xmath0 . however , this is not necessarily the case near avoided crossings where the values of @xmath92 are shifted .
these shifts lead to either suppression or enhancement of the tunneling .
we have chosen @xmath86 and @xmath87 in this section .
the results can , however , be straightforwardly generalized to any value of @xmath0 and @xmath93 , as long as @xmath94 and @xmath95 .
we remark that figs .
[ fig_modj ] , [ fig_floq](b ) , and the top panel of fig .
[ fig_amp ] are taken from ref .
@xcite , but the detailed floquet analysis of the @xmath2 modulation , as well as the entire analysis of the effects of the tilt and interaction modulation , has not been presented elsewhere .
if the modulation amplitude is small and the system is not near an avoided crossing , the floquet eigenstates and eigenvalues turn out to be close to the ones determined by the time - independent part of the hamiltonian , given by @xmath96 as a consequence , some important properties of the modulated system , such as the positions of the resonances , can be explained by analyzing the spectrum of @xmath97 .
we assume that @xmath98 and @xmath66 . in order to compare the time - evolution operator of the original time - dependent modulated system with that determined by hamiltonian ( [ h0 ] )
, we define the floquet operator @xmath99 corresponding to @xmath97 as @xmath100 we denote the eigenvalues of the time - independent hamiltonian by @xmath101 , where we use the same indexing as in the case of the eigenvectors of the floquet operator @xmath29 .
the phases of the floquet eigenvalues are given by @xmath102 we find that @xmath103 for odd @xmath0 and @xmath104 for even @xmath0 . because of the minus sign in eq .
@xmath105 , the opposite holds for the phases of the floquet eigenvalues @xmath106 .
the situation is similar in the presence of a small - amplitude modulation , and thus , normally , @xmath90 ( @xmath91 ) for odd ( even ) @xmath0 . now
@xmath107 if @xmath108 . using this and the equation @xmath109 , we see that @xmath106 , and therefore also @xmath110 , increases faster as a function of @xmath14 the larger @xmath111 is .
this means that if @xmath112 , @xmath113 approaches @xmath114 from below as @xmath14 increases [ see fig .
[ fig_floq](b ) for an example ] .
the phases @xmath115 cross repeatedly as @xmath14 increases .
a crossing occurs when @xmath14 satisfies @xmath116 with @xmath117 .
the last crossing between @xmath118 and @xmath119 is at @xmath120 . in the limit of @xmath121 , the phases of all the floquet eigenvalues approach zero from the negative side . in the specific case @xmath86 and @xmath87 , corresponding to figs .
[ fig_modj ] and [ fig_modv ] , the crossing points between @xmath122 and the other phases are at @xmath123 ( crossing with @xmath124 ) , @xmath125 ( @xmath126 ) , @xmath127 ( @xmath128 ) , and @xmath129 ( @xmath130 ) .
these are obtained from eq .
( [ eq_res ] ) with @xmath131 .
note that @xmath132 and @xmath133 , and thus @xmath134 and @xmath135 , are almost identical .
the crossing points corresponding to @xmath131 and @xmath43 in the region @xmath136 are shown by the vertical red dotted lines and arrows in figs .
[ fig_modj ] and [ fig_modv ] .
we see that the positions of all the resonances shown in figs .
[ fig_modj ] and [ fig_modv ] are well explained by the energy eigenvalues of the time - independent hamiltonian .
based on this fact , we can say that the positions of the crossings are the same irrespective of the modulated variable .
next we consider the modulation of the tunneling matrix element ; see figs .
[ fig_modj ] and [ fig_floq ] .
the value @xmath137 can be non - zero only if the floquet eigenstates @xmath138 and @xmath139 have the same parity .
consequently , there is an avoided crossing between eigenstates with the same parity . in fig .
[ fig_floq](b ) , we show the phases of the floquet eigenvalues as a function of @xmath140 for the parameters used in fig .
[ fig_modj ] [ and fig .
[ fig_floq](a ) ] . from fig .
[ fig_floq ] we see that a large change of @xmath85 occurs when @xmath141 crosses the other @xmath142 s [ circles and ellipses in figs .
[ fig_floq](a ) and [ fig_floq](b ) ] .
the behavior of the phases of the floquet eigenvalues near the crossings is schematically shown in fig .
[ fig_floq](c ) . in an @xmath0-particle system
, there are @xmath143 different types of crossings between @xmath144 and other @xmath145 s @xcite .
because now @xmath86 , we have three types of crossings ; each resonance corresponds to one of these . in the following ,
we analyze in detail each of these three crossing types .
a crossing between @xmath141 and @xmath146 leads to a reduction of @xmath85 in a wide range of @xmath14 around @xmath147 .
this crossing is indicated in figs .
[ fig_floq](a ) and [ fig_floq](b ) by the solid magenta circle .
the detailed structure of the crossing is shown schematically in the top figure in fig .
[ fig_floq](c ) . since @xmath128 and @xmath130 are almost equal , the avoided crossings between @xmath130 and @xmath135 and between @xmath128 and @xmath134 occur almost simultaneously [ the red solid circles in fig .
[ fig_floq](c ) ] . because @xmath90 for odd @xmath0 outside the crossing region ( see sec . [ sec_timeindep ] ) and @xmath88 couples floquet eigenstates with the same parity , the first avoided crossing occurs between @xmath130 and @xmath135 [ the left red solid circle ] as the modulation frequency increases . due to the repulsion between these two levels , the splitting between @xmath141 is increased near the avoided crossing and thus the tunneling period is reduced .
the second avoided crossing takes place between @xmath128 and @xmath134 [ the right red solid circle ] .
note that , after the first avoided crossing , the states @xmath148 and @xmath149 have been interchanged [ between the red solid circles ] and the energy splitting between @xmath141 remains large until the second avoided crossing at which @xmath150 and @xmath151 are interchanged .
these successive avoided crossings lead to a reduction of the tunneling period in a wide range of @xmath152 . because of the large quasienergy splitting between @xmath124 and @xmath126 , the points where @xmath141 crosses @xmath124 and @xmath126 are far apart .
we call a crossing between @xmath141 and @xmath126 a type 2 crossing and that between @xmath141 and @xmath124 a type 3 crossing . with increasing @xmath14 , a type 2 crossing first yields a reduction and then an enhancement of the tunneling period .
we show the schematic structure of a type 2 crossing in the middle figure in fig .
[ fig_floq](c ) .
the resonances around @xmath153 and @xmath154 , indicated by the green dashed curves in figs .
[ fig_floq](a ) and [ fig_floq](b ) , correspond to type 2 crossings .
suppose that the crossing is approached from the small @xmath152 side .
far from the avoided crossing @xmath155 as explained in sec .
[ sec_timeindep ] .
since @xmath88 couples floquet eigenstates with the same parity , @xmath126 undergoes an avoided crossing with @xmath135 ( the large red solid circle ) .
near the avoided crossing , the energy splitting between @xmath141 is enhanced , which leads to the reduction of the tunneling period .
just after the avoided crossing ( around the vertical dashed line ) , the states @xmath156 and @xmath149 are interchanged and , unlike in the usual situation , @xmath157 . since @xmath134 is larger than @xmath135 far from the avoided crossing also on the large @xmath152 side , @xmath134 and @xmath135 cross each other ( the small blue solid circle ) , which yields a divergence of the tunneling period .
as opposed to a type 2 crossing , a type 3 crossing ( crossings between @xmath141 and @xmath124 ) gives first an enhancement and then a reduction of the tunneling period .
the resonances at @xmath158 and @xmath159 , which are indicated by the blue dotted ellipses and circles in figs .
[ fig_floq](a ) and [ fig_floq](b ) , correspond to type 3 crossings .
a detailed schematic structure of a type 3 crossing is shown in the bottom figure in fig .
[ fig_floq](c ) .
suppose again that we approach the crossing from the small @xmath152 side . in this case , we have an avoided crossing between @xmath124 and @xmath134 .
the phase @xmath134 , which is located above @xmath135 far from the crossing , is pushed downward due to the avoided crossing with @xmath124 , and thus @xmath134 and @xmath135 cross each other ( the small blue solid circle ) .
this leads to the divergence of the tunneling period .
after this , there is an avoided crossing between @xmath124 and @xmath134 ( the large red solid circle ) , leading to an enhancement of the quasienergy splitting between @xmath141 .
this yields a reduction in the tunneling period .
after the avoided crossing , the states @xmath160 and @xmath161 are interchanged . in the top panel of fig .
[ fig_amp ] , we show the tunneling period @xmath85 for various values of the modulation amplitude @xmath163
. with decreasing @xmath163 , the resonance around @xmath164 becomes narrower and finally separates into two resonances ( see the case @xmath165 shown by the blue dashed - dotted line ) .
the schematic behavior of the phases of the floquet eigenvalues near the crossing is shown in the bottom panel of fig .
[ fig_amp ] .
we call this a type 1@xmath162 crossing .
the major difference between type 1@xmath162 and type 1 crossings is the existence of two points where @xmath141 cross each other .
these are indicated by the small blue solid circles , and they are located between the two avoided crossings ( the large red solid circles ) .
one can also view a type 1@xmath162 crossing as a combination of type 2 and type 3 crossings .
when @xmath163 is small , the coupling between the two states that undergo an avoided crossing is small .
thus the difference @xmath166 remains very small even near the avoided crossing .
therefore , unlike in a type 1 crossing , the inverted configuration of @xmath141 ( i.e. , the situation @xmath167 ) can not be sustained throughout the region between the two avoided crossings .
this leads to the appearance of two crossing points , indicating a diverging tunneling period . in ref .
@xcite it has been shown that the divergences are present if @xmath168 . as can be seen from figs .
[ fig_modj ] and [ fig_modv ] , the tunneling period behaves differently when @xmath4 is modulated . in fig .
[ fig_modv ] , a noticeable change in the tunneling period @xmath85 can be seen around @xmath164 .
there are also small , narrow resonances at @xmath169 and @xmath170 @xcite .
unlike in the case of modulated @xmath2 , the resonance at @xmath164 is not a wide and smooth reduction of @xmath85 for any value of the modulation amplitude @xmath64 . as in the case of modulated @xmath2 ,
the resonance around @xmath164 is caused by a crossing between @xmath141 and @xmath146 .
however , unlike @xmath88 , the operator @xmath65 has odd parity , and it thus couples floquet eigenstates of opposite parity . in fig .
[ fig_floq_tilt ] , we show the schematic behavior of the floquet eigenstates near @xmath164 .
suppose that the crossing is approached from the small @xmath152 side .
as @xmath152 increases , the states @xmath148 and @xmath151 undergo an avoided crossing , and @xmath134 is pushed downward .
far from the avoided crossing the relation @xmath90 holds . because of this and the fact that @xmath134 is pushed downward , the phases @xmath134 and @xmath135 cross ( the left small blue circle ) before the avoided crossing ( the left large red circle ) . after the first avoided crossing , the states @xmath148 and @xmath151 are interchanged .
next , @xmath128 and @xmath135 undergo an avoided crossing ( the right large red circle ) , and the corresponding states are interchanged . because now @xmath167 , these phases cross after the second avoided crossing ( the right small blue circle ) , so that @xmath171 far away from the crossing .
the two crossing points and the two avoided crossings correspond to the two divergences and the two reductions of the tunneling period , respectively .
these are shown in fig .
[ fig_modv ] near @xmath164 .
because in the present case the avoided crossings occur between floquet eigenstates of opposite parity , the phases @xmath141 necessarily cross each other outside the region of the successive avoided crossings .
for this reason , a smooth reduction of @xmath85 in a wide region of the modulation frequency @xmath14 can not be achieved by modulating the tilt .
this is one of the major findings of this paper .
we note that all the other resonances are also much narrower than in the case of modulated @xmath2 .
this is because the operator @xmath65 , which is related to the tilt , does not contribute to the single - particle tunneling , unlike @xmath88 .
the range of @xmath14 characterizing the width of the resonance is comparable to the quasienergy separation at the avoided crossing .
this is approximately proportional to @xmath172 in the case of @xmath2 modulation and to @xmath173 in the case of @xmath4 modulation .
the latter is smaller than the former by a factor @xmath174 .
this will be discussed in more detail in sec .
[ sec_noon ] .
finally , we consider the case in which the on - site interaction strength @xmath3 is modulated weakly @xmath175 .
the hamiltonian in this case is @xmath176 , with @xmath177 given by eq .
( [ eq_ut ] ) . since this hamiltonian can be rewritten as @xmath178,\ ] ] with @xmath179 and @xmath180,\ ] ] we can expect that the dynamics can be reproduced by modulating @xmath2 with the amplitude @xmath181 and phase @xmath182 instead of modulating @xmath3 .
this observation is confirmed by the result shown in fig .
[ fig_modu ] , where we compare the tunneling period @xmath85 as a function of @xmath14 in the cases of modulated @xmath2 and modulated @xmath3 . in this example
@xmath183 and @xmath184 .
the result for the modulated @xmath2 is taken from fig .
[ fig_modj ] ( @xmath185 ) . by setting @xmath186 , i.e. , @xmath187 in the present case
, these two results almost coincide with each other .
note that since @xmath85 does not noticeably depend on the phase of the modulation , @xmath186 is a sufficient condition for the behaviors of the tunneling periods to coincide .
an analysis of the phases of the floquet eigenvalues in the case of the @xmath3 modulation shows that the schematic behavior of these phases around each resonance is the same as in the case of the @xmath2 modulation shown in fig .
[ fig_floq](c ) .
this can be understood by noting that @xmath88 and @xmath188 have the same parity .
in this section , we first study a system characterized by a weak interaction and a large - amplitude tilt modulation , concentrating on the properties of the floquet spectrum .
after this we examine the effects of a large - amplitude modulation of the interaction . the floquet spectrum of this system has been analyzed elsewhere ( see refs .
@xcite ) and will not be discussed here .
instead , we propose a way to create noon states using selective tunneling originating from the modulation of the interaction strength .
next , we consider a case where the interaction is weak , @xmath189 , and the amplitude of the modulation of the tilt is large , @xmath190 .
we assume that the tunneling matrix element @xmath2 and the interaction strength @xmath3 are time - independent , that is , @xmath191 and @xmath192 .
furthermore , we set @xmath66 . in this case , it is well - known that the effect of the modulation of the tilt can be approximately described by a renormalized tunneling term . in more detail ,
the original tunneling term @xmath193 is replaced by an effective one @xcite : @xmath194 } \hat{s}_x\\ & - \sin{\left[\frac{v_1}{\omega}\cos{\phi_v}\right ] } \hat{s}_y \big\},\end{aligned}\ ] ] where @xmath195 is the zeroth - order bessel function ( see appendix [ sect_jeff ] for the derivation ) .
coherent destruction of tunneling takes place when @xmath196 is equal to one of the zeros of @xmath195 . in the rest of this section ,
we discuss cdt in terms of the floquet eigenvalues .
this discussion holds for any @xmath197 . in fig .
[ fig_cdt ] , we show the tunneling period @xmath85 as a function of the modulation frequency @xmath14 in the regime of weak interaction and large - amplitude modulation . in this calculation , we have set @xmath86 , @xmath198 , @xmath66 , @xmath199 , and @xmath200 , and in the initial state all particles are in site 1 .
the first five zeros of @xmath201 are at @xmath202 , @xmath203 , @xmath204 , @xmath205 , and @xmath206 : they correspond to @xmath207 , @xmath208 , @xmath209 , @xmath210 , and @xmath211 , respectively .
these frequencies are shown by the vertical red dotted lines in fig .
[ fig_cdt ] .
there is good agreement between these dotted lines and the actual positions of the peaks of @xmath85 . in fig .
[ fig_floq_cdt ] , we plot the phases of the floquet eigenvalues for the parameters used in fig .
[ fig_cdt ] .
when the cdt occurs , the phases gather in pairs , the phases in each pair being almost equal , and all the pairs gather in a narrow region ( red arrows in fig . [ fig_floq_cdt ] ) .
this behavior can be understood by noting that the hamiltonian is effectively @xmath212 at the point where cdt occurs , and thus @xmath46 becomes a good quantum number , with a twofold degeneracy with respect to @xmath213 . finally , we discuss the difference between even and odd @xmath0 cases . for even @xmath0 ,
the number of the floquet eigenvalues is @xmath214 , which is odd .
therefore , when cdt occurs , the floquet eigenvalues are grouped into one trio and @xmath215 pairs [ cf .
@xmath216 pairs for odd @xmath0 ] .
a key point is that , for even @xmath0 , there is a fock state @xmath218 , which does not have a degenerate pair , unlike the other fock states . in this case ,
the floquet eigenstates near the value of @xmath14 at which cdt occurs can be classified into three types : 1 ) one floquet eigenstate that has maximum amplitude at @xmath219 component , 2 ) @xmath220 floquet eigenstates that do not have maximum amplitude at the @xmath219 component but that always have a non - zero @xmath219 component , and 3 ) @xmath220 floquet eigenstates that do not have maximum amplitude at the @xmath219 component and where this component becomes zero when cdt occurs .
the trio consists of all the three types , and the @xmath215 pairs consist of the second and third types .
we note that , for even @xmath0 , the degeneracies of the trio and of all the pairs are incomplete provided @xmath221 @xcite , while all the pairwise degeneracies are complete for odd @xmath0 .
consequently , cdt is more complete for odd @xmath0 than even @xmath0 .
due to the non - linear dependence of the interaction on @xmath46 , the cdt caused by a large - amplitude modulation of the interaction strength ( @xmath222 , @xmath223 ) is state dependent @xcite . here
we assume @xmath224 for simplicity .
in this case , a condition for partial cdt between the states @xmath225 and @xmath226 ( @xmath227 is a positive integer ) is @xmath228 = 0;\ ] ] see appendix [ sect_umod ] for the derivation . unlike in the case of modulated @xmath4 shown in fig .
[ fig_floq_cdt ] , only the floquet eigenstates relevant to partial cdt show the degeneracy in the phases of the floquet eigenvalues ( see , e.g. , fig . 1 of ref
. @xcite ) . for odd
@xmath0 , each partial cdt is associated with a perfect degeneracy of the phases of the floquet eigenvalues while , for even @xmath0 , some degeneracies ( but not all ) are incomplete provided @xmath221 . as in the case of modulated @xmath4 ,
these incomplete degeneracies are caused by the existence of the fock state @xmath229 .
consequently , partial cdt is generally more complete for odd @xmath0 than for even @xmath0 .
the floquet spectrum in the case of large - amplitude modulation of @xmath3 and weak interaction has been studied in depth in refs .
we refer to these references for further discussion . finally , we point out that it is possible to create mesoscopic schrdingers - cat like states [ noon - like states @xcite , i.e. , states proportional to @xmath230 , where @xmath231 is a phase ] using the state - dependent cdt . in this scheme , we assume that @xmath232 and choose @xmath77 as the initial state .
we modulate @xmath3 at a frequency @xmath14 that corresponds to a partial cdt between @xmath77 and @xmath233 , that is , @xmath234 = 0 $ ] . at this frequency
the phases of the floquet eigenstates @xmath235 , which are very close to noon states , become degenerate @xcite . by detuning from this partial cdt
, we have a coherent oscillation ( with period @xmath85 ) between @xmath236 and @xmath237 . as a result
, the initial state @xmath77 evolves into a noon - like state at @xmath238 , with @xmath239 . with increasing the absolute value of the detuning ,
the period @xmath85 decreases but the amplitudes of the components other than @xmath240 increase , so that the oscillation between the noon states is disturbed .
therefore , @xmath14 ( more precisely , @xmath241 ) should be optimized . in fig .
[ fig_noonmodu ] , we show the time evolution of the normalized population imbalance @xmath242 and its variance @xmath243 for @xmath244 and @xmath245 as examples . here
@xmath246 and @xmath247 with @xmath248 .
these are optimized cases with the amplitude of the wiggles in the oscillation of @xmath242 being @xmath249 .
when @xmath250 , @xmath251 is almost equal to @xmath41 , which is the largest possible value ; this is a unique property of noon states . note that the oscillation periods are comparable in the two cases : @xmath252 and @xmath253 for @xmath244 and @xmath245 , respectively .
this shows that an advantage of the present scheme is that the optimized @xmath85 does not increase exponentially with @xmath0 unlike the tunneling period of the higher - order co - tunneling in the self - trapping regime .
this may be understood by the fact that the static part of the interaction strength @xmath223 is very small ( @xmath254 ) .
a disadvantage is that we need to know the number of particles exactly and to fine - tune @xmath241 .
this scheme can be used regardless of the value of @xmath0 if @xmath255 .
in this section we propose another scheme to create noon - like states .
this scheme uses an adiabatic sweep of the modulation frequency .
it enables us to obtain noon - like states with @xmath256 particles starting from the ground state @xmath257 of the time - independent hamiltonian @xmath97 .
the basic idea is to create an avoided crossing between the floquet eigenstate corresponding to @xmath257 and the one corresponding to the noon - like eigenstate @xmath258 by time - periodic modulation , which changes the geometry of the ( quasi)energy space to be periodic . here
, we modulate the hopping parameter @xmath2 and set the tilt @xmath259 .
since the phase @xmath260 of the modulation does not affect the result , we choose @xmath261 for definiteness .
the time - independent part @xmath97 of the hamiltonian @xmath262 is given by eq .
( [ h0 ] ) , while the time - dependent part @xmath263 is @xmath264 for even @xmath0 , the crossing used in the creation of the noon state is the one between @xmath236 and @xmath265 . for odd @xmath0
, it is the one between @xmath236 and @xmath266 .
we consider the regime @xmath267 , where @xmath236 is a noon - like state .
the ground state @xmath257 of @xmath97 corresponds to @xmath265 ( even @xmath0 ) or @xmath266 ( odd @xmath0 ) , and the eigenvalue of @xmath97 corresponding to @xmath257 is denoted by @xmath268 .
similarly , the noon - like eigenstate @xmath258 of @xmath97 corresponds to @xmath269 .
state @xmath258 has the highest energy among symmetric eigenstates of @xmath97 , and its eigenenergy is denoted by @xmath270 . because @xmath271 , the eigenstates of @xmath97 are almost equal to the floquet eigenstates except near the crossing points .
therefore , @xmath272 for even @xmath0 , @xmath273 for odd @xmath0 , and @xmath274 . as discussed in sec .
[ sec_timeindep ] , when @xmath14 is decreased from a sufficiently large value , the first crossing occurs between @xmath275 and @xmath276 for even @xmath0 and between @xmath275 and @xmath277 for odd @xmath0 @xcite .
therefore , in principle , this scheme can be used without knowing precisely the total number of particles . the avoided crossing between the phases
@xmath275 and @xmath278 or @xmath124 is approximately at @xmath279 . in the @xmath86 case discussed earlier
, this crossing corresponds to the rightmost circle in fig .
[ fig_floq](b ) .
let us take @xmath257 as the initial state .
if we sweep @xmath14 adiabatically across the avoided crossing , @xmath257 undergoes an almost perfect transition to @xmath258 .
we consider a linear sweep of the form @xmath280 where @xmath281 is the location of the crossing and @xmath282 is the sweep rate .
the initial and final times of the sweep are denoted by @xmath283 and @xmath284 , respectively . in the following calculations , we set @xmath86 and @xmath87 .
the avoided crossing is at @xmath159 . in fig .
[ fig_lz ] , we show the time evolution of the probability @xmath285 [ blue ( dark gray ) lines ] at which the system stays in the initial state @xmath257 and the probability @xmath286 [ red ( medium gray ) lines ] at which the system undergoes a transition to the target state @xmath258 . note that @xmath287 shown by the green ( light gray ) lines in fig .
[ fig_lz ] is very close to unity throughout the calculations ( the deviation is within @xmath288% ) , and the system is , to a very good approximation , restricted to the subspace spanned by the two states .
thus the crossing can be described by the landau - zener ( lz ) model @xcite .
we denote the modulation period at the crossing point by @xmath289 .
the difference between the phases of the floquet eigenvalues at @xmath290 is @xmath291 . here
, we shift the phase difference so that the crossing at @xmath292 is passed at @xmath35 , in accordance with the standard expression of the lz hamiltonian . for the linear sweep of eq .
( [ eq_sweep ] ) , we get @xmath293 . here
we assume that @xmath294 .
we obtain the quasienergy separation @xmath295 corresponding to @xmath296 near the crossing as @xmath297 where we have approximated @xmath298 .
the diagonal matrix elements @xmath299 and @xmath300 of the lz hamiltonian are thus @xmath301 , where the upper sign corresponds to @xmath299 and the lower one corresponds to @xmath300 .
we found that the off - diagonal elements @xmath302 and @xmath303 of the effective hamiltonian are to a good approximation given by @xmath304 .
consequently , the asymptotic value @xmath305 of the transition probability @xmath306 , @xmath307 , is @xcite @xmath308}\nonumber\\ = & \exp{\left[-\frac{\pi j_0 ^ 2 a_j^2|\langle \psi_h|\hat{s}_x|\psi_g\rangle|^2}{\alpha } \right]}. \label{eq_p0}\end{aligned}\ ] ] in fig .
[ fig_p0 ] , we show the probability @xmath305 as a function of the inverse sweep rate @xmath309 [ fig .
[ fig_p0](a ) ] and the modulation amplitude @xmath163 [ fig .
[ fig_p0](b ) ] . since @xmath306 and @xmath310 continue to oscillate around the asymptotic value until far after the crossing ( see fig .
[ fig_lz ] ) , we calculate @xmath305 by taking the time average of @xmath306 after its oscillation amplitude becomes small and almost time independent .
these results are shown by circles in fig .
[ fig_p0 ] .
semianalytical results obtained from eq .
( [ eq_p0 ] ) are shown by the red solid lines .
for the parameters used here ( @xmath86 and @xmath87 ) , we have @xmath311 , @xmath312 , @xmath313 , and @xmath314 .
the agreement between the semianalytical and numerical results is very good .
finally , we examine the experimental feasibility of this scheme . according to eq .
( [ eq_p0 ] ) , to obtain a noon - like state , we should satisfy the adiabaticity condition : @xmath315 in addition , the initial and the final frequencies should be outside the crossing region . since the range of @xmath14 characterizing the crossing region is comparable to the level separation @xmath316 at the avoided crossing , the initial time @xmath283 and the final time @xmath284 of the sweep have to satisfy @xmath317 .
also the landau - zener formula is valid under this condition .
taking into account the adiabaticity condition ( [ eq_ad ] ) , this leads to the requirement @xmath318 as an example , let us estimate the time scale given by this equation by using the parameters used in the experiment of ref . @xcite . in this experiment ,
the frequency of the pair tunneling is @xmath319 hz for @xmath320 ; thus @xmath321 ms . if @xmath322 , the right - hand side of eq .
( [ eq_tsweep ] ) is @xmath323 ms for @xmath324 , @xmath325 ms for @xmath326 , and @xmath327 ms for @xmath328 .
therefore , a noon state with @xmath329 could be created within an experimentally accessible time , provided the value of @xmath14 can be controlled with a sufficiently high accuracy .
we note that , more generally , an upper limit for @xmath0 for this scheme to work is @xmath330 .
since the width of the peaks in the probability distribution ( in the fock space ) of @xmath257 and @xmath258 scales as @xmath331 , a few times @xmath332 should be larger than @xmath0 in order to have an overlap between @xmath257 and @xmath258 and to have a significant nonzero value of @xmath333 . in the present scheme , the modulation of the hopping parameter works much better than the modulation of the tilt .
this can be seen using perturbation theory .
a straightforward calculation shows that for odd number of particles @xmath334 and for even number of particles latexmath:[$\langle\psi_h perturbation theory shows that latexmath:[$\langle\psi_h ' @xmath0 and @xmath336 for even @xmath0 . here
@xmath337 is the antisymmetric eigenstate of @xmath97 with the highest energy .
we see that @xmath338 conclusions ----------- in this paper , we have considered a time - periodically modulated two - mode bose - hubbard model . we have discussed three types of modulations , one where the tunneling amplitude is modulated , another where the interaction strength is modulated , and a third where the energy difference between the modes ( tilt ) is modulated .
we focused mainly on the self - trapping regime , characterized by @xmath339 , and assumed that the amplitude of the modulation is small .
it is known that a modulation of the tunneling amplitude can lead to a drastic reduction of the tunneling period @xcite .
we found that a similar effect can be induced by modulating the interaction strength or the energy difference between the modes .
we have analyzed this phenomenon using floquet theory as the main tool .
we found that regardless of the modulated variable , the system has resonances at some modulation frequencies , corresponding to either greatly reduced or enhanced tunneling periods .
to a good approximation , the locations of the resonances can be obtained with the help of the energy eigenvalues of the time - independent part of the hamiltonian .
consequently , the locations of the resonances are almost independent of whether the tunneling , interaction , or tilt is modulated .
we found numerically that if the tunneling amplitude or interaction strength is modulated , the system has a wide resonance ; that is , the tunneling period is greatly reduced in a wide range of modulation frequencies .
this resonance is present also in the case of modulated tilt , but it is much narrower .
furthermore , the behavior of the tunneling period as a function of the modulation frequency is not smooth in this case ; see fig .
[ fig_modv ] .
these differences can be explained using floquet theory .
the presence of resonances is related to avoided crossings of the phases of the floquet eigenvalues . in the case of a modulated tunneling matrix element or interaction strength ,
the avoided crossings correspond to floquet eigenstates with the same parity . in the case of modulated tilt
, these avoided crossings correspond to eigenstates with opposite parity . in sec .
[ sec_modv ] , it is shown that due to this difference , a wide smooth resonance can not be obtained in the case of modulated tilt .
we have also analyzed cases where the interaction energy is weak in comparison with the tunneling energy , @xmath340 , and the modulation amplitude of either the interaction strength or the tilt is large . under these conditions , tunneling can be suppressed at some specific modulation frequencies .
this phenomenon , the coherent destruction of tunneling , has been extensively studied in the literature .
we concentrated on a property of cdt that has received less attention in the previous studies , namely , the floquet spectrum of a system where the tilt is modulated .
as expected , we found that the suppression of tunneling takes place when the phases of the floquet eigenvalues become degenerate .
for an even number of particles the suppression is more complete than that for an odd number of particles .
finally , we have proposed two ways to create a noon state .
one is based on coherent oscillation resulting from a detuning from a partial cdt caused by modulated interaction strength .
an advantage of this method is that the tunneling period does not increase exponentially with the total number of particles @xmath0 .
the other method is based on sweeping the modulation frequency of the tunneling term adiabatically .
this scheme requires neither precise knowledge of the number of particles nor fine - tuning of the modulation frequency .
we have shown that by using the latter method and the parameters of a recent experiment @xcite , it is possible to obtain noon states of @xmath329 particles .
it is known that the mean - field theory of the time - periodically modulated two - mode bose - hubbard model shows chaotic dynamics ( e.g. , refs .
@xcite ) . in the future
, it would be interesting to study the connection between the floquet spectrum of the original quantum system and the chaotic mean - field dynamics . another interesting problem to study would be the quantum dynamics determined by a time - periodically modulated hamiltonian in the presence of dissipation . in particular , the engineered dissipation leading to squeezed states proposed in ref .
@xcite is of interest .
we acknowledge ippei danshita , chris pethick , and sukjin yoon for helpful discussions .
gw acknowledges the max planck society , the korea ministry of education , science and technology , gyeongsangbuk - do , and pohang city for the support of the jrg at apctp .
gw is also supported by basic science research program through the national research foundation of korea ( nrf ) funded by the ministry of education , science and technology ( no .
2012008028 ) .
here we derive the effective tunneling amplitude in the limit of large - amplitude tilt modulation .
the system follows the schrdinger equation @xmath341 with @xmath342 and @xmath343 given by eq .
( [ eq_vt ] ) .
we go to a rotating system by defining @xmath344 where @xmath345\\ & = v_0 t+\frac{v_1}{\omega}[\cos\phi_v -\cos(\omega t+\phi_v)].\end{aligned}\ ] ] using this , the schrdinger equation becomes @xmath346 where @xmath347\,\hat{s}_x-\sin[\alpha(t ) ] \,\hat{s}_y\right)+u_0 \hat{s}_z^2.\end{aligned}\ ] ] assuming that the modulation period @xmath23 is the shortest time scale in the system , it is possible to obtain an effective hamiltonian by averaging over @xmath22 as @xmath348 the effective tunneling amplitudes are defined as @xmath349 \ , d\tau\\ { j_y^{\textrm{eff}}}(t)&= -\frac{j_0}{t_{\omega}}\int_t^{t+t_{\omega } } \sin[\alpha(\tau ) ] \ , d\tau.\end{aligned}\ ] ] instead of calculating @xmath350 and @xmath351 separately , we write @xmath352}.\end{aligned}\ ] ] this integral can be calculated easily using the equation @xmath353 where @xmath354 are bessel functions of the first kind .
we thus obtain @xmath355}\vspace{1mm}\\ \displaystyle \quad\times \sum_{n=-\infty}^{\infty}\ , { \cal j}_n\left(\frac{v_1}{\omega}\right ) \frac{e^{in(\omega t+\phi_v-\frac{\pi}{2})}}{v_0+n\omega},&\displaystyle \frac{v_0}{\omega}\not\in \mathbb{z}\vspace{3mm}\\ \displaystyle { j_0 \cal j}_{k}\left(\frac{v_1}{\omega}\right)e^{i\frac{v_1}{\omega}\cos\phi_v } e^{-ik ( \phi_v+\frac{\pi}{2})},&\displaystyle \frac{v_0}{\omega}=k \in \mathbb{z}. \end{cases } \label{eq_jeff}\end{aligned}\ ] ] in the special case @xmath356 , the original tunneling amplitudes @xmath357 and @xmath358 are replaced by effective ones , @xmath359,\\ { j_y^{\textrm{eff}}}(t ) & = j_0 { \cal j}_{k}\left(\frac{v_1}{\omega}\right)\sin\left[k\left(\phi_v+\frac{\pi}{2}\right)-\frac{v_1}{\omega}\cos\phi_v\right],\end{aligned}\ ] ] where @xmath64 is non - zero .
in the case of large - amplitude modulation of the interaction strength , the coherent destruction of tunneling is state - dependent @xcite . here , we derive the effective hamiltonian for this case .
we start from the time - dependent schrdinger equation ( [ eq_tdsch ] ) with the hamiltonian @xmath360 where @xmath177 is given by eq .
( [ eq_ut ] ) . for simplicity
, we set @xmath259 . as in appendix
[ sect_jeff ] , we go to the rotating frame by defining @xmath361 where @xmath362\nonumber\\ = & \ u_0 t + \frac{u_1}{\omega } \left[\cos\phi_u - \cos{(\omega t+\phi_u)}\right].\end{aligned}\ ] ] thus the schrdinger equation becomes @xmath363 , with @xmath364,\ ] ] where @xmath365 . we have used the equations @xmath366=\hat{s}_+ ( 2\hat{s}_z+1)$ ] , @xmath367=-(2\hat{s}_z+1)\hat{s}_-$ ] , and @xmath368 to obtain @xmath369.\end{aligned}\ ] ] by time averaging over one modulation period @xmath22 , the effective hamiltonian reads @xmath370.\end{aligned}\ ] ] here @xmath371 is defined as @xmath372 } e^{i\left[u_0(t+\frac{\pi}{\omega } ) + \frac{u_1}{\omega}\cos{\phi_u } \right](\delta n+1)}\vspace{1mm}\\ \displaystyle \qquad\times\sum_{n=-\infty}^{\infty } { \cal j}_{n}\left[\frac{u_1}{\omega}(\delta n+1)\right ] \frac{e^{-in(\omega t+\phi_u+\frac{\pi}{2})}}{u_0 ( \delta n+1)-n\omega}|\delta n\rangle,\quad & \displaystyle\frac{u_0}{\omega}(\delta n+1)\not\in \mathbb{z},\vspace{3 mm } \\
\displaystyle { \cal j}_{k}\left[\frac{u_1}{\omega}(\delta n + 1)\right ] e^{i\frac{u_1}{\omega } ( \delta n+1 ) \cos{\phi_u } } e^{-ik(\phi_u+\frac{\pi}{2})}|\delta n\rangle,\quad & \displaystyle\frac{u_0}{\omega}(\delta n + 1)=k \in \mathbb{z } , \end{cases}\label{eq_a}\ ] ] where we have used the equation @xmath373 and @xmath38 is the basis of the system . in this basis @xmath374
is a tridiagonal matrix .
note that @xmath371 , unlike eq .
( [ eq_jeff ] ) , depends on @xmath46 . if @xmath375 ( here we assume @xmath376 without loss of generality ) , we get @xmath377 . in the special case @xmath378=k\in \mathbb{z}$ ] ,
the condition for partial cdt between states @xmath379 and @xmath380 , @xmath375 , can be written as @xmath381=0.\ ] ] if this equation holds , the hilbert space can be written as a direct sum of three uncoupled subspaces , spanned by @xmath382 , @xmath383 , and @xmath384 . as we shall see later , for odd @xmath0 these @xmath143 crossing types consist of one type where @xmath385 crosses @xmath144 successively and @xmath386 types where @xmath387 crosses @xmath144 once .
for even @xmath0 , there is one crossing type with @xmath385 and @xmath388 types with each @xmath389 and one type with @xmath390 . in the case of modulated @xmath2 ( @xmath4 ) ,
the phase of the floquet eigenvalue @xmath390 undergoes avoided crossings with the phases of the eigenstates of even ( odd ) parity . for @xmath393 , the effective hamiltonian [ see eq .
( [ eq_heff ] ) ] vanishes when the condition for cdt is satisfied : @xmath394 .
when this equation holds , the phases of all the floquet eigenvalues vanish , and the degeneracy is complete irrespective of whether @xmath0 is even or odd . for even @xmath0
this degeneracy involves three states if @xmath393 : @xmath235 and another floquet eigenstate whose quasienergy is identically zero .
the third state disturbs the desired oscillation between @xmath235 .
the third state can be easily lifted to change the threefold degeneracy to a twofold one by introducing nonzero , but small , @xmath223 such that @xmath254 .
this can be understood by the fact that , with decreasing @xmath14 , @xmath276 for even @xmath0 and @xmath277 for odd @xmath0 deviate from zero at the lowest rate and @xmath275 deviates at the highest rate among symmetric states . | we study the tunneling dynamics in a time - periodically modulated two - mode bose - hubbard model using floquet theory .
we consider situations where the system is in the self - trapping regime and either the tunneling amplitude , the interaction strength , or the energy difference between the modes is modulated . in the former two cases ,
the tunneling is enhanced in a wide range of modulation frequencies , while in the latter case the resonance is narrow .
we explain this difference with the help of floquet analysis .
if the modulation amplitude is weak , the locations of the resonances can be found using the spectrum of the non - modulated hamiltonian .
furthermore , we use floquet analysis to explain the coherent destruction of tunneling ( cdt ) occurring in a large - amplitude modulated system .
finally , we present two ways to create a noon state ( a superposition of @xmath0 particles in mode 1 with zero particles in mode 2 and vice versa ) .
one is based on a coherent oscillation caused by detuning from a partial cdt .
the other makes use of an adiabatic variation of the modulation frequency .
this results in a landau - zener type of transition between the ground state and a noon - like state . |
in recent years , various ideas have been proposed to realize intelligent highway systems to reduce traffic congestions and improve safety levels .
it is envisioned that navigation , communication and automatic driver assistance systems are critical components @xcite . a great deal of monitoring and controlling capabilities
have been implemented through roadside infrastructures , such as cameras , sensors , and control and communication stations .
such systems can work together to monitor in real time the situations on highways and at the same time guide vehicles to move in a coordinated fashion , e.g. to keep appropriate distances from the vehicles in front of and behind each individual vehicle .
in intelligent highway systems , the guiding commands are expected to be simple and formatted as short digital messages to scale with the number of vehicles and also to avoid conflict with the automatic driver assistance systems installed within the vehicles .
similar guided formation control problems also arise when navigating mobile robots or docking autonomous vehicles @xcite . motivated by this problem of guiding platoons of vehicles on highways
, we study in this paper the problem of controlling a one - dimensional multi - agent formation using only _ coarsely _ quantized information .
the formation to be considered are rigid under inter - agent distance constraints and thus its shape is uniquely determined locally .
most of the existing work on controlling rigid formations of mobile agents , e.g. @xcite , assumes that there is no communication bandwidth constraints and thus real - valued control signals are utilized .
the idea of quantized control through digital communication channels has been applied to consensus problems , e.g. @xcite and references therein , and more recently to formation control problems @xcite .
the uniform quantizer and logarithmic quantizer @xcite are among the most popular choices for designing such controllers with quantized information . moreover , the paper @xcite has discussed krasowskii solutions and hysteretic quantizers in connection with continuous - time average consensus algorithms under quantized measurements .
the problem studied in this paper distinguishes itself from the existing work in that it explores the limit of the least bandwidth for controlling a one - dimensional rigid formation by using a quantizer in its simplest form with only two quantization levels . as a result , for each agent in the rigid formation , at most four bits of bandwidth is needed for the communication with the navigation controller .
the corresponding continuous - time model describing the behavior of the overall multi - agent formation is , however , non - smooth and thus an appropriate notion of solution @xcite has to be defined first .
we use both the lyapunov approach and trajectory - based approach to prove convergence since the former provides a succinct view about the dynamic behavior while the latter leads to insight into the set of initial positions for which the proposed controller may fail .
we also discuss some situations when different assumptions about the quantization scheme are made and indicate those scenarios in which the formation control problem with quantized information can be challenging to solve .
the rest of the paper is organized as follows .
we first formulate the one - dimensional guided formation control problem with coarsely quantized information in section [ se : formulation ] . then in section
[ se : analysis ] , we provide the convergence analysis results first using the lyapunov method and then the trajectory - based method . simulation results are presented in section [ se : simulation ] to validate the theoretical analysis .
we make concluding remarks in section [ se : conclusion ] .
the one - dimensional guided formation that we are interested in consists of @xmath0 mobile agents .
we consider the case when the formation is rigid @xcite ; to be more specific , if we align the given one - dimensional space with the @xmath1-axis in the plane and label the agents along the positive direction of the @xmath1-axis by @xmath2 , then the geometric shape of the formation is specified by the given pairwise distance constraints @xmath3 , @xmath4 , where @xmath5 are desired distances .
although the guidance system can monitor the motion of the agents in real time , we require that it can only broadcast to the mobile agents quantized guidance information through digital channels .
in fact , we explore the limit for the bit constraint by utilizing the quantizer that only has two quantization levels and consequently its output only takes up one bit of bandwidth .
the quantizer that is under consideration takes the form of the following sign function : for any @xmath6 , @xmath7 each agent , modeled by a kinematic point , then moves according to the following rules utilizing the coarsely quantized information : [ quantized.n.agent.system ] x_1 & = & -k_1 ( x_1-x_2 ) ( |x_1-x_2|-d_1 ) + x_i & = & ( x_i-1-x_i ) ( |x_i-1-x_i|-d_i-1)- + & & k_i ( x_i - x_i+1 ) ( |x_i - x_i+1|-d_i ) , + & & i=2, ,n-1 + x_n & = & ( x_n-1-x_n ) ( |x_n-1-x_n|-d_n-1 ) where @xmath8 is the position of agent @xmath9 in the one - dimensional space aligned with the @xmath1-axis , and @xmath10 are gains to be designed . note
that since each agent is governed by at most two distance constraints , as is clear from ( [ quantized.n.agent.system ] ) , a bandwidth of four bits is sufficient for the communication between the guidance system and the agents @xmath11 and the required bandwidths for the guidance signals for agents @xmath12 and @xmath0 are both 2 bits .
hence , in total only @xmath13 bits of bandwidth is used .
the main goal of this paper is to demonstrate under this extreme situation of using coarsely quantized information , the formation still exhibits satisfying convergence properties under the proposed maneuvering rules . towards this end , we introduce the variables of relative positions among the agents [ z ] z_ix_i - x_i+1,i=1,2, ,n-1 .
let us express the system in the @xmath14-coordinates to obtain [ quantized.n.agent.system.z ] z_1 & = & -(k_1 + 1 ) ( z_1 ) ( |z_1|-d_1 ) + & & + k_2 ( z_2 ) ( |z_2|-d_2 ) + z_i & = & ( z_i-1 ) ( |z_i-1|-d_i-1 ) + & & -(k_i+1 ) ( z_i ) ( |z_i|-d_i ) + & & + k_i+1 ( z_i+1 ) ( |z_i+1|-d_i+1 ) , + & & i=2,
,n-2 + z_n-1 & = & ( z_n-2 ) ( |z_n-2|-d_n-2 ) + [ 2 mm ] & & -(k_n-1 + 1 ) ( z_n-1 ) ( |z_n-1|-d_n-1 ) . to study the dynamics of the system above , we need to first specify what we mean by the solutions of the system .
since the vector field @xmath15 on the right - hand side is discontinuous , we consider krasowskii solutions , namely solutions to the differential inclusion @xmath16 , where @xmath17 @xmath18 denotes the involutive closure of a set , and @xmath19 is the ball centered at @xmath14 and of the radius @xmath20 .
the need to consider these solutions becomes evident in the analysis in the next section . since the right - hand side of ( [ quantized.n.agent.system ] ) is also discontinuous ,
its solutions are to be intended in the krasowskii sense as well .
then we can infer conclusions on the behavior of ( [ quantized.n.agent.system ] ) provided that each solution @xmath1 of ( [ quantized.n.agent.system ] ) is such that @xmath14 defined in ( [ z ] ) is a krasowskii solution of ( [ quantized.n.agent.system.z ] ) .
this is actually the case by @xcite , theorem 1 , point 5 ) , and it is the condition under which we consider ( [ quantized.n.agent.system.z ] ) .
it turns out that the @xmath14-system ( [ quantized.n.agent.system.z ] ) is easier to work with for the convergence analysis that we present in detail in the next section .
in this section , after identifying the equilibria of the system , we present two different approaches for convergence analysis . the first is based on a lyapunov - like function and the second examines the vector field in the neighborhood of the system s trajectories .
we start the analysis of system ( [ quantized.n.agent.system.z ] ) by looking at the discontinuity points of the system .
a discontinuity point is a point at which the vector field on the right - hand side of the equations above is discontinuous .
hence , the set @xmath21 of all the discontinuity points is : @xmath22 it is of interest to characterize the set of equilibria : [ lemma.equilibria ] let @xmath23 , @xmath24 for @xmath25 , and @xmath26 .
the set of equilibria , i.e. the set of points for which @xmath27 with @xmath15 being the vector field on the right - hand side of ( [ quantized.n.agent.system.z ] ) , is given by @xmath28 for @xmath29 , if @xmath30 for @xmath31 , and @xmath27 , then @xmath32 .
_ proof : _
suppose by contradiction that @xmath33 .
this implies that in a neighborhood of this point , the state space is partitioned into different regions where @xmath15 is equal to constant vectors . in view of ( [ quantized.n.agent.system.z ] ) , the component @xmath9 of these vectors is equal to one of the following values : @xmath34 , @xmath35 , @xmath36 , @xmath37 , if @xmath38 , or @xmath39 , @xmath40 , @xmath41 , @xmath42 , if @xmath43 .
any @xmath44 is such that its component @xmath9 belongs to ( a subinterval of ) the interval @xmath45 $ ] if @xmath38 ( respectively , to the interval @xmath46 $ ] if @xmath47 ) . in both cases ,
if @xmath24 , then the interval does not contain @xmath48 and this is a contradiction .
this ends the proof of the lemma .
@xmath49 _ proof of proposition [ lemma.equilibria ] :
_ first we show that if @xmath27 , then @xmath50 . as a first step , we observe that @xmath27 implies @xmath51 .
in fact , suppose by contradiction that the latter is not true .
this implies that at the point @xmath14 for which @xmath27 , any @xmath52 is such that the first component takes values in the interval @xmath53 $ ] , or in the interval @xmath54 $ ] . in both cases ,
if @xmath23 , then @xmath48 does not belong to the interval and this contradicts that @xmath27 .
hence , @xmath51 .
this and lemma [ lemma.claim ] show that @xmath55 , consider the last equation of ( [ quantized.n.agent.system.z ] ) , and again suppose by contradiction that @xmath56
. then the last component of @xmath57 belongs to a subinterval of @xmath58 $ ] or to a subinterval of @xmath59 $ ] .
if @xmath26 , then neither of these intervals contain @xmath48 and this is again a contradiction .
this concludes the first part of the proof , namely that @xmath27 implies @xmath50 .
now we let @xmath50 and prove that @xmath60 . by definition , if @xmath50 , then @xmath14 lies at the intersection of @xmath61 planes , which partition @xmath62 into @xmath63 regions , on each one of which @xmath15 is equal to a different constant vector .
any @xmath64 is the convex combination of these @xmath65 vectors , which we call @xmath66 .
we construct @xmath64 such that @xmath67 .
we observe first that , the component @xmath12 of the vectors @xmath68 s can take on four possible values , namely @xmath69 , @xmath70 , @xmath71 , @xmath72 , and that there are exactly @xmath73 ( we are assuming that @xmath74 , as the case @xmath75 is simpler and we omit the details ) vectors among @xmath76 whose first component is equal to @xmath69 , @xmath73 whose first component is equal to @xmath70 and so on . as a consequence , if @xmath77 for all @xmath78 , then @xmath79 .
+ similarly , the component @xmath9 , with @xmath80 , can take on eight possible values ( @xmath39 , @xmath40,@xmath81 see the expression of @xmath82 in ( [ quantized.n.agent.system.z ] ) ) and as before , the set @xmath76 can be partitioned into @xmath83 sets , and each vector in a set has the component @xmath9 equal to one and only one of the eight possible values .
moreover , these values are such that @xmath84 . + finally , if @xmath85 , the set @xmath76 can be partitioned into four sets , and each vector in a set has the last component equal to one and only one of the four possible values @xmath86 , @xmath87 , @xmath88 , @xmath89 .
hence , @xmath90 .
let now @xmath57 be such that @xmath91 , with @xmath77 for all @xmath9 . since @xmath84 for all @xmath92 , then @xmath67 and this proves that for all @xmath50 , we have @xmath27 .
this completes the proof .
@xmath49 next , we show that the equilibrium set @xmath93 is attractive .
now we are in a position to present the main convergence result .
if [ gains ] k_1k_2 , k_ik_i+1 + 1 , i=2, ,n-2 , k_n-11 , then all the krasowskii solutions to ( [ quantized.n.agent.system.z ] ) converge to ( a subset of ) the equilibria set @xmath94 .
_ proof : _ let @xmath95 be a smooth non - negative function .
we want to study the expression taken by @xmath96 , where @xmath15 is the vector field on the right - hand side of ( [ quantized.n.agent.system.z ] ) .
we obtain : @xmath97 \\
\qquad i=1\\ \qquad \\ z_i(z_i^2-d_i^2)[-(k_i+1 ) \textrm{sgn}(z_i ) \textrm{sgn}(|z_i|-d_{i } ) \\ + \textrm{sgn}(z_{i-1 } ) \textrm{sgn}(|z_{i-1}|-d_{i-1 } ) \\ + k_{i+1 } \textrm{sgn}(z_{i+1 } ) \textrm{sgn}(|z_{i+1}|-d_{i+1 } ) ] \\ \qquad i=2,\ldots , n-2\\ \qquad \\ z_{n-1}(z_{n-1}^2-d_{n-1}^2)[\textrm{sgn}(z_{n-2 } ) \textrm{sgn}(|z_{n-2}|-d_{n-2 } ) & \\
-(k_{n-1}+1 ) \textrm{sgn}(z_{n-1 } ) \textrm{sgn}(|z_{n-1}|-d_{n-1 } ) ] \\ \qquad
i = n-1 \ea \right.\end{aligned}\ ] ] if @xmath98 , i.e. if @xmath14 is not a point of discontinuity for @xmath15 , then : @xmath99 -(k_i - k_{i+1 } ) |z_i|\,|z_i^2-d_i^2| & i=2,\ldots , n-2\\[2 mm ] -k_{n-1 }
|z_{n-1}|\,|z_{n-1}^2-d_{n-1}^2| & i = n-1 \ea \right.\end{aligned}\ ] ] where we have exploited the fact that @xmath100 . hence , if ( [ gains ] ) holds , then @xmath101 if @xmath102 , we look at the set @xmath103 we distinguish two cases , namely ( i ) @xmath104 and ( ii ) @xmath105 . in case ( i ) ,
@xmath106 , and therefore , @xmath107 . in case ( ii )
, there must exist at least one agent such that @xmath108 and at least one agent such that @xmath109 .
let @xmath110 ( respectively , @xmath111 ) be the set of indices corresponding to agents for which @xmath108 ( @xmath109 ) . clearly , @xmath112 .
+ since @xmath113 if @xmath114 , then @xmath115 let @xmath116 and @xmath57 . in view of ( [ quantized.n.agent.system ] ) , for @xmath92
, it holds : @xmath117\}\;,\end{aligned}\ ] ] with @xmath118 then @xmath119 by ( [ gains ] ) , @xmath120 for all @xmath9 , and therefore , if @xmath105 , then latexmath:[\[\ba{rcl } \nabla
v(z ) \cdot v & \le & -{\displaystyle}\sum_{i\in { \cal i}_2(z ) } @xmath57 .
this shows that for all @xmath122 , either @xmath123 or @xmath124 . in summary , for all @xmath125 , either @xmath126 or @xmath124 , and @xmath127 if and only if @xmath50 .
it is known ( lemma 1 in @xcite ) that if @xmath128 is a solution of the differential inclusion @xmath16 , then @xmath129 exists almost everywhere and @xmath130 .
we conclude that @xmath131 is non - increasing .
let @xmath132 , with @xmath133 a compact and strongly invariant set for ( [ quantized.n.agent.system.z ] ) .
for any @xmath134 , such a set exists and includes the point @xmath135 ( hence @xmath136 ) , by definition of @xmath137 and because @xmath137 is non - increasing along the solutions of ( [ quantized.n.agent.system.z ] ) .
since @xmath138 or @xmath124 for all @xmath125 , then by the lasalle invariance principle for differential inclusions @xcite , any solution @xmath128 to the differential inclusion converges to the largest weakly invariant set in @xmath139 ( @xmath94 is closed ) . since the choice ( [ gains ] ) yields that the gains @xmath140 s satisfy the condition in lemma [ lemma.equilibria ]
, @xmath94 is the set of equilibria of ( [ quantized.n.agent.system.z ] ) ( and therefore it is weakly invariant ) and since @xmath136 , we conclude that any solution @xmath128 converges to the set of points @xmath141 .
@xmath49 since the equilibrium set @xmath93 contains those points for which two agents coincide with each other , it is of interest to characterize those initial conditions under which the asymptotic positions of some of the agents become coincident . in the next subsection
, we use a three - agent formation as an example to show how such analysis can be carried out .
we specialize the rigid formation examined before to the case @xmath142 . letting @xmath143 ,
the one - dimensional rigid formation becomes : [ quantized.3.agent.system ] x_1 & = & -(x_1-x_2 ) ( |x_1-x_2|-d_1 ) + x_2 & = & ( x_1-x_2 ) ( |x_1-x_2|-d_1)- + & & ( x_2-x_3 ) ( |x_2-x_3|-d_2 ) + x_3 & = & ( x_2-x_3 ) ( |x_2-x_3|-d_2 ) .
let us express the system in the coordinates @xmath144 , @xmath145 , so as to obtain : [ quantized.3.agent.system.z ] z_1 & = & -2 ( z_1 ) ( |z_1|-d_1)+(z_2 ) ( |z_2|-d_2 ) + z_2 & = & ( z_1 ) ( |z_1|-d_1)- 2(z_2 ) ( |z_2|-d_2 ) .
we study the solutions of the system above . in what follows ,
it is useful to distinguish between two sets of points : @xmath146 clearly , @xmath147 .
we now prove that all the solutions converge to the desired set @xmath148 except for solutions which originates on the @xmath144- or the @xmath145-axis : [ p1 ] all krasowskii solutions of ( [ quantized.3.agent.system.z ] ) converge in finite time to the set @xmath149 .
in particular , the solutions which converge to the points @xmath150 must originate from the set of points @xmath151 .
moreover , the only solution which converges to @xmath152 is the trivial solution which originates from @xmath152 . _
proof : _ because of the symmetry of @xmath15 , it suffices to study the solutions which originate in the first quadrant only .
in the first quadrant we distinguish four regions : ( i ) @xmath153 , ( ii ) @xmath154 , ( iii ) @xmath155 , ( iv ) @xmath156 .
now we examine the solutions originating in these regions .
+ ( i ) @xmath157 . if both @xmath158 and @xmath159 , then the system equations become @xmath160 and the solution satisfies @xmath161 . in other words
, the solution evolves along the line of slop @xmath162 and intercept @xmath163 .
if @xmath164 , then the solution @xmath165 converges to the point @xmath166 in finite time .
in particular @xmath167 with @xmath168 .
if @xmath169 , then @xmath165 converges in finite time to the semi - axis @xmath170 .
this is a set of points at which @xmath15 is discontinuous , since for @xmath171 , @xmath172 , and for @xmath173 , @xmath174 .
since at these points @xmath175 , denotes the smallest closed convex set which contains @xmath176 . ] and vectors in @xmath177 intersect the tangent space at the semi - axis in those points , a sliding mode along the semi - axis must occur . since @xmath178
, we conclude that the sliding mode must satisfy the equations @xmath179 and therefore , after a finite time , the solution converges to the point @xmath180 . on the other hand ,
if @xmath181 , then the solution reaches the ray @xmath182 .
similar considerations as before can show that a sliding mode occurs along the ray and that it satisfies the equations @xmath183 and again convergence in finite time to @xmath180 is inferred .
finally we examine the case @xmath184 . at the point @xmath180 , @xmath185 i.e. @xmath186 and @xmath180 is an equilibrium point .
similarly as before , one shows that the solution which originates from @xmath180 must stay in @xmath180 .
+ ( ii ) @xmath187 . if @xmath188 and @xmath189 , then the map @xmath15 is equal to the vector @xmath190 and the solution @xmath165 satisfies @xmath191 . if @xmath192 , then @xmath165 converges to @xmath180 , while if @xmath192 , it first converges to the semi - axis @xmath193 , and then it slides towards @xmath180 . when @xmath192 , the solution reaches the segment @xmath194 . on this segment , @xmath195 , and since this intersects the tangent space at the segment , a sliding mode occurs .
the sliding mode obeys the equations @xmath196 which show that the state reaches @xmath180 .
+ if @xmath197 and @xmath189 , then the initial condition lies on another discontinuity surface of @xmath15 .
observe that , for those points such that @xmath198 and @xmath199 , @xmath172 .
hence , @xmath200 intersects the tangent space at the semi - axis in those points , and the solutions can slide along the semi - axis until they reach the point @xmath201 and stop , or can enter the region @xmath202 , and then converge to @xmath180 , or they can enter the region @xmath203 and converge to the point @xmath204 .
+ the point @xmath201 is an equilibrium , and if @xmath205 , solutions stay at the equilibrium . +
we review the remaining cases succinctly , as they are qualitatively similar to the cases examined above .
+ ( iii ) @xmath206 . if @xmath207 for @xmath208 , then the solutions converge to @xmath180 possibly sliding along the segments @xmath209 or @xmath210 .
if @xmath197 and @xmath211 , then the solution can converge to the points @xmath212 , @xmath213 or @xmath180 .
if @xmath214 and @xmath215 , then the solutions can converge to @xmath180 , @xmath216 or @xmath217 . finally ,
if @xmath218 for @xmath208 , the solutions can converge to any of the points in @xmath219 .
in particular , a possible solution is the one which remains in @xmath152 .
+ ( iv ) @xmath220 .
solutions which start from initial conditions such that @xmath221 and @xmath211 converge to @xmath222 . if @xmath223 and @xmath211 , then the solution converge to @xmath222 possibly sliding on the segment @xmath224 .
if @xmath221 and @xmath215 , the solutions can converge to one of the three possible points : @xmath225 , @xmath226 , @xmath222 .
@xmath49 a few comments are in order : * sliding modes arise naturally for those situations in which , for instance , the state reaches the semi - axis @xmath227 .
this forces us to consider krasowskii solutions rather than carathodory solutions . on the other hand
, the set of krasowskii solutions may be too large in some cases , as it is evident for instance for those solutions which start on the @xmath144- or @xmath145-axis .
* the occurrence of sliding modes are not acceptable in practice as they would require fast information transmission . a mechanism to prevent sliding modes in the system ( [ quantized.3.agent.system ] )
can be introduced following @xcite .
in this section , we present simulation results for the guided formation control with coarsely quantized information .
we consider a formation consisting of 6 agents , labeled by @xmath228 .
the distance constraints are @xmath229 , @xmath230 .
the initial positions of agents 1 to 6 are 0 , 0.5 , 1 , 2 , 4 and 5 respectively .
then the shape of the initial formation is shown in figure [ fig1 ] .
we choose @xmath231 , @xmath232 , @xmath233 , @xmath234 and @xmath235 and simulate the agents motion under the control laws ( [ quantized.n.agent.system ] ) . in figure [ fig2 ] , we show the shape of the final formation . to see how the shape evolves with time
, we present the curve of the lyapunov function @xmath236 in figure [ fig3 ] .
since our analysis has been carried out using krasowskii solutions , when we further look into the dynamics of @xmath14 , it is clear that the sliding mode may still happen when the krasowskii solution converges . but
this effect due to the system s non - smoothness is within an acceptable level as shown in figure [ fig4 ] which presents the curve of @xmath144 .
in this paper , we have studied the problem of controlling a one - dimensional guided formation using coarsely quantized information .
it has been shown that even when the guidance system adopts quantizers that return only the one - bit sign information about the quantized signal , the formation can still converge to the desired equilibrium under the proposed control law .
the point model we have used throughout the analysis is a simplified description of vehicle dynamics .
when more detailed models are taken into consideration , we need to deal with collision avoidance and other practical issues as well .
so it is of great interest to continue to study the same problem with more sophisticated vehicle models and more physical constraints from the applications . | motivated by applications in intelligent highway systems , the paper studies the problem of guiding mobile agents in a one - dimensional formation to their desired relative positions .
only coarse information is used which is communicated from a guidance system that monitors in real time the agents motions .
the desired relative positions are defined by the given distance constraints between the agents under which the overall formation is rigid in shape and thus admits locally a unique realization .
it is shown that even when the guidance system can only transmit at most four bits of information to each agent , it is still possible to design control laws to guide the agents to their desired positions .
we further delineate the thin set of initial conditions for which the proposed control law may fail using the example of a three - agent formation .
tools from non - smooth analysis are utilized for the convergence analysis . |
photonic entanglement is an essential resource for many quantum information processing applications , such as linear - optics quantum computing ( loqc ) @xcite , quantum teleportation @xcite , and quantum key distribution ( qkd ) @xcite .
the most widely used method for generating polarization - entangled photon pairs is spontaneous parametric down - conversion ( spdc ) , which can be realized in a number of configurations and by using different crystals @xcite . a well optimized spdc source with high flux and high entanglement quality can be utilized to develop advanced enabling quantum technology such as a random number generator @xcite .
certain applications such as quantum communications and loqc are often designed to operate with a system clock to allow for synchronization and , for that purpose , a pulsed source of entangled photons is needed .
in other applications such as free - space entanglement - based qkd in the daylight , pulsed operation may be preferred to simplify signaling of the arrival times of the entangled photons and to provide temporal discrimination against undesirable background light .
many spdc sources , however , are continuous - wave ( cw ) pumped which can not yield any timing information about when an entangled photon pair is generated .
it is therefore of interest to develop an efficient pulsed source of entangled photons that can offer unique operational and application - specific advantages over a cw spdc source .
we have previously developed a cw spdc source of polarization - entangled photons using a polarization sagnac interferometer ( psi ) configuration that is highly efficient and yields a high visibility in two - photon quantum interference @xcite .
the sagnac source configuration is a phase - stable , single - crystal implementation of interferometrically combined outputs of two identical , coherently - driven down - converters @xcite .
the psi configuration eliminates the need for spatial ( aperture ) , spectral ( interference filter ) , and temporal ( timing compensator ) filtering because the two down - converter outputs are completely indistinguishable @xcite .
all the output photons are strongly polarization - entangled without the need for filtering , thus leading to a much higher generation efficiency than that of other approaches . in principle
, the sagnac configuration can be used for both cw and pulsed operation .
shi and tomita have previously utilized a non - polarizing sagnac interferometer to realize a highly stable pulsed down - conversion entanglement source @xcite .
the use of a non - polarizing beam splitter incurs a 50% reduction in the entanglement generation rate because when both photons exit the same port of the beam splitter ( half of the time ) it does not produce polarization entanglement . in this work
, we employ a polarizing beam splitter ( pbs ) to eliminate this 50% loss , and we demonstrate the first pulsed operation of a sagnac down - conversion source pumped by a narrowband , mode - locked ultraviolet ( uv ) source . unlike most pulsed spdc sources that are pumped by femtosecond lasers @xcite , the pulsed sagnac source is pumped with @xmath550-ps pulses with a narrow spectral bandwidth . while the cw and pulsed sagnac sources share a common design , there can be major differences between the two types of sources . in cw
down - conversion the output flux is often limited by available cw pump power , especially in the uv region .
therefore cw sources are often characterized by a normalized generation efficiency in terms of the number of entangled pairs generated per second per nanometer of detection bandwidth for 1 mw of pump power . in general
, the situation is different for pulsed pumping because high peak power is easily obtainable at many wavelengths , either directly from a laser or by efficient harmonic generation of a high power pulsed laser @xcite . the high peak pump power for pulsed down - conversion benefits applications that can take advantage of a substantial pair generation probability per pulse such as those requiring multiphoton coincidences for the generation of novel multipartite states in loqc @xcite . on the other hand , with high generation probabilities one must be careful with the impact of multiple - pair events on the system performance of some quantum information processing applications
for instance , in free - space entanglement - based qkd the entanglement source may be strongly driven to achieve a desirable key generation rate but it should not exceed a level that may compromise the security of the generated secret keys @xcite . we have developed a pulsed sagnac source based on type - ii phase - matched periodically poled ktiopo@xmath0 ( ppktp ) and pumped with a custom - built high - power mode - locked uv source at 390.35 nm @xcite .
the sagnac source is capable of producing polarization - entangled photon pairs at 780.7 nm with a mean pair generation of as much as one per pulse , even in a small bandwidth of 0.15 nm .
the narrowband pulsed sagnac source is particularly suitable for free - space entanglement - based qkd .
pulsed operation affords synchronized detection of the entangled photons and provides temporal discrimination against temporally random background photons or detector dark counts .
the narrowband outputs can be used with narrowband spectral filtering that is essential for blocking out ambient light during daylight operation . in the next section ,
we describe our experimental setup in the construction of the pulsed sagnac source . in section iii ,
we characterize the entanglement quality of the psi output with three different types of measurements : two - photon quantum interference , violation of clauser - horne - shimony - holt ( chsh ) form of bell s inequality @xcite , and quantum state tomography . in section iv
, we focus on the strongly pumped sagnac source with its high - flux outputs and the accompanying visibility degradation due to occurrence of multiple - pair events , before concluding in section v.
we develop the narrowband pulsed sagnac source for potential use in free - space entanglement - based qkd that must satisfy a number of operational constraints . for line - of - sight free - space qkd
it has been suggested that the optimal operating wavelength is around 780 nm that takes into account atmospheric transmission and the characteristics of commercially available si avalanche photodiodes ( apds ) as single - photon counters @xcite .
commercially available si apds typically have a detection quantum efficiency of @xmath550% and a rise time of @xmath5300ps .
the pump for the sagnac source should have a pulse width that is small compared with the detector rise time so that the pump does not add to the timing uncertainty . at the same time
, we prefer to operate the down - converter in a quasi - cw manner with a pump bandwidth that is small compared with the down - conversion phase - matching bandwidth or the spectral filter bandwidth . otherwise , some of the photon pairs generated by the broadband pump would be outside of the measurement bandwidth and the system throughput efficiency is reduced .
we choose the pump requirements to have a pulse width of 100ps or less and a bandwidth of less than 0.1 nm .
a narrowband down - conversion output allows the use of narrowband spectral filters inserted before the si apd detectors to screen out background light from the sky during daylight qkd operation .
the pulse width and bandwidth requirements can accommodate a range of pump operating conditions and we note that transform limited pulses are not required .
figure 1 shows the experimental setup for the pulsed sagnac source together with the measurement apparatus .
we will briefly describe the psi configuration and more details can be found in previous cw implementations @xcite .
the psi , as shown in the dashed box of fig . 1 ,
is composed of two flat mirrors and a pbs that serves as the input and output optical element . within the interferometer is a nonlinear crystal for generating entangled photons by spdc .
bidirectional pumping of the crystal in counter - propagating directions creates coherent superposition of the counter - propagating down - conversion outputs at the pbs , whose output photons are entangled in polarization .
interferometric combination of two outputs is usually sensitive to path - length perturbations , but the psi configuration eliminates the need for path - length stabilization through its common path arrangement @xcite .
moreover , the phase of the output state can be chosen by simply adjusting the relative phase between the horizontally ( @xmath6 ) and vertically ( @xmath7 ) polarized components of the pump . for our pulsed sagnac source
, we used a 10-mm long flux - grown ppktp crystal from raicol with a grating period of 7.85@xmath8 m for type - ii phase matching and degenerate wavelength output at 780.7 nm .
the crystal was temperature stabilized at a phase - matching temperature of 28.6@xmath9c using a thermoelectric heater with a temperature stability of better than 0.1@xmath9c . using cw second - harmonic generation , we measured an effective nonlinear coefficient , @xmath10 of @xmath52.4pm / v . to ensure that the counter - propagating pump components reach the crystal at the same time , thus eliminating any temporal distinguishability between the two outputs , we positioned the crystal at the center of the interferometer within @xmath51 mm , which is much less than that spanned by the 50-ps pump pulse .
the pulsed uv pump centered at 390.35 nm was a compact , home - built narrowband picosecond uv source based on frequency quadrupling of an amplified mode - locked fiber laser @xcite .
it had a bandwidth of less than 0.1 nm , full - width at half maximum ( fwhm ) .
the passively mode - locked erbium - doped fiber laser with a narrowband intracavity filter was operated at 1561.4 nm at a repetition rate of 31.1mhz .
the fiber laser output was amplified with a 5-w polarization - maintaining erbium - doped fiber amplifier and followed by two stages of single - pass second - harmonic generation to yield the pulsed uv output with a maximum average power of 400mw and a pulse width of @xmath550ps .
because of the frequency quadrupling configuration , there was a significant amount of residual 780.7 nm light in the uv source output .
as a uv pump for the sagnac source , it was necessary to filter this undesirable near - infrared ( near - ir ) light before reaching the psi .
we used a pair of dichroic mirrors ( dms ) that were highly transmissive at 780.7 nm and highly reflective at 390.35 nm , and constructed a 7-bounce passage for the pump beam to filter the residual 780.7 nm light .
we estimate that a suppression of 140db was achieved using this multiple - bounce dichroic mirror pair arrangement with negligible loss of uv power .
the linearly polarized , collimated uv beam after the dm filter was passed through a half - wave plate ( hwp ) and a quarter - wave plate ( qwp ) to adjust the relative phase and amplitude between the @xmath6- and @xmath7-polarized components .
we focused the pump into the psi using a plano - convex lens with a 15-cm focal length and with broadband anti - reflection ( ar ) coating that covered both uv and near - ir wavelengths .
the ppktp crystal was centered at the pump focal spot that was measured to have a beam diameter of @xmath590@xmath8 m .
the sagnac configuration calls for an input / output pbs that works at both the pump and down - conversion wavelengths .
instead , we used a commercially available standard pbs cube designed and ar - coated for 780 nm but not for 390 nm .
the pbs provided acceptable extinction ratio ( @xmath520:1 ) to separate the @xmath6 and @xmath7 components of the uv pump , with absorption and reflection losses of 35% at 390 nm .
the pbs served to produce counter - propagating pumps for bidirectional pumping of the down - conversion crystal .
an ar - coated dual - wavelength half - wave plate ( dw - hwp ) designed for 390.35 nm and 780.7 nm rotated the @xmath7-polarized pump to @xmath6 polarization that was required ( along the crystal s @xmath11-axis ) for type - ii phase matching .
we note that the low extinction ratio pbs introduced a small amount of @xmath7-polarized pump at the crystal but it was not phase matched to generate any down - converted photon pairs .
before the signal and idler outputs from the two bidirectionally pumped down - converters were combined at the pbs , the clockwise circulating output was rotated by 90@xmath9 with the dw - hwp such that the signal and idler beams after the pbs were spatially separated , as indicated in fig . 1 .
each of the signal and idler outputs was collimated with a 15-cm focal length plano - convex lens and passed through a matching circular aperture for spatial - mode filtering .
we sent the output beams to polarization analyzers each comprising a hwp and a high - extinction polarizer to measure the polarization states .
a qwp could also be added for quantum state tomography measurements .
we filtered both beams spectrally using interference - filters ( ifs ) with a fwhm bandwidth of @xmath50.15 nm centered at 780.7 nm and with peak transmission of 60% .
the outputs were subsequently focused onto two perkinelmer si apds with @xmath550% detection efficiency .
we recorded individual apd counts as well as coincidence counts and a 1.8-ns coincidence window was used @xcite .
[ figure1 ] we aligned the sagnac interferometer such that the two down - converter outputs for the signal ( and for the idler ) were spatially and temporally overlapped at the pbs output to erase any `` which - path '' information .
this required symmetric positioning of the ppktp crystal inside the sagnac interferometer . at this point ,
the pair generation rates from both paths were balanced and the necessary input pump polarization adjustments were made to generate a singlet - state output @xmath12 where the subscript @xmath3 ( @xmath13 ) refers to the signal ( idler ) photon . in the next two sections , we show the results of our source characterization in the low and high flux regimes .
unlike most cw spdc sources , the combination of a high - power pulsed uv pump and the efficient psi source can lead to a substantial pair generation probability per pulse . in order to assess the entanglement quality of the pulsed sagnac source with minimal degradation due to multiple - pair events , we measured the characteristics of the psi output at low pumping powers .
in particular , we made measurements of the two - photon quantum interference , the chsh form of bell s inequality , and quantum state tomography of the psi output state .
we show that the three types of characterization methods are consistent with each other .
the signal - idler quantum interference measurement in two incompatible polarization bases , @xmath6-@xmath7 and @xmath1445@xmath9 antidiagonal - diagonal ( @xmath15-@xmath16 ) , is a common and relatively simple method to assess the entanglement quality of the source . for each of the polarization measurement bases ,
the signal polarization analyzer angle is set at one of the basis polarization axes and the coincidence counts are monitored as a function of the idler analyzer angle .
the quantum - interference visibility is given by @xmath17 where @xmath18 and @xmath19 are the maximum and minimum coincidence counts , respectively , and an ideal entanglement source yields @xmath20 .
[ figure2 ] -@xmath7 ( squares ) and @xmath15-@xmath16 ( circles ) bases .
@xmath6-@xmath7 ( @xmath15-@xmath16 ) visibility is 99.79% ( 98.11%).,width=264 ] we observed the quantum interference fringes at different pump power levels and different aperture sizes .
figure 2 shows the coincidence counts in the @xmath6-@xmath7 and @xmath15-@xmath16 bases and the corresponding sinusoidal fits to data at a pump power of 0.1mw , as measured at the entrance to the psi .
each data point represents an average of 30 1-s measurements , and we apply no background - count subtraction to the coincidence count data in fig .
2 . using the measured maximum and minimum coincidence counts we obtain from eq .
( 2 ) an interference visibility of @xmath21% in the @xmath6-@xmath7 basis and @xmath22% in the @xmath15-@xmath16 basis . the data in fig . 2 were taken using a circular aperture with a full collection divergence angle of @xmath513mrad . from the singles count rate of @xmath51,600/s
we infer a @xmath59.5% conditional detection probability , limited in large part by the si apd detector efficiency of @xmath550% , the if transmission of @xmath536% ( 60% peak transmission and 60% double - lorentzian filter shape ) , and the measured 95% transmission efficiency through the other optical components .
the remaining reduction in the conditional detection probability can be attributed to spatial filtering that was provided by the 2-mm diameter aperture , which is a typical problem due to the multimode nature of spdc with an unfocused or weakly focused pump @xcite .
we measured a si apd dark - count rate of @xmath550/s , and the non - parametric fluorescence counts were less than 5% of the measured singles which , at this pump level , did not contribute to any accidental coincidences .
figure 3 shows the quantum interference measurements at a pump power of 1.1mw and with all other parameters being the same as in fig . 2 . the singles rate increased to 16,000s@xmath23 at this power level , and
the quantum - interference visibilities are found to be @xmath24% in the @xmath6-@xmath7 basis and @xmath25% in the @xmath15-@xmath16 basis . from the singles count rate and the 9.5% conditional detection probability
, one can estimate the pair generation rate to be 1.1% per pulse .
this high pump power caused the slight reduction in visibilities due to multiple - pair emission events , which are analyzed in detail in the next section .
[ figure3 ] -@xmath7 ( squares ) and @xmath15-@xmath16 ( circles ) bases . with slightly higher photon pair generation rates ,
we obtain 98.04% @xmath6-@xmath7 visibility and 96.64% @xmath15-@xmath16 visibility.,width=264 ] we have also measured the variation in the quantum - interference visibility at different divergence angles , as set by the aperture diameters , and at a constant input power of 1.1mw . for each data point in fig . 4 , we average 30 1-s measurements of @xmath18 and @xmath19 , and obtain the visibility calculated from eq .
( [ v ] ) . at a given pump power
a larger aperture size allowed more light to be collected which increased the effective pair generation rate . as a result , multiple - pair events increased ( see next section for details ) and caused the visibility to degrade slightly , as shown in figure 4 for the measurements in the @xmath6-@xmath7 ( squares ) basis .
the @xmath15-@xmath16 data shows a more pronounced deterioration in the visibility at larger collection angles .
the @xmath6-@xmath7 basis was the natural basis aligned with the ppktp s crystal principal axes .
therefore , measurements in the @xmath15-@xmath16 basis required the coherent superposition of the two counter - propagating down - converter outputs , and the entanglement quality was sensitive to their spatial mode distinguishability and to the phase variation of the output state across the spatial extent . in our case , the @xmath15-@xmath16 data in fig .
4 clearly show that the singlet - state entangled output of the psi degraded at larger divergence angles . in designing the sagnac source for a specific application
, one must therefore consider the trade - off between generation efficiencies and entanglement quality as a function of the collection angles .
[ figure4 ] -@xmath7 ( squares ) and @xmath15-@xmath16 ( circles ) bases for various aperture sizes at a constant input pump power of 1.1mw.,width=264 ] violation of the chsh form of bell s inequality @xcite is another common method of entanglement characterization by measuring its @xmath3 parameter . for an ideal entangled state a maximum value of @xmath26
is predicted by quantum mechanics , whereas @xmath3 can not be greater than 2 classically .
therefore , the state is considered nonclassical if @xmath27 and the closer @xmath3 is to @xmath28 , the higher is the entanglement quality .
we follow the measurement procedure of ref .
the @xmath3-parameter is given by four measurement expectation values @xcite @xmath29 where @xmath30 and @xmath31 are the polarization analyzer angles for signal and idler , respectively .
we chose to maximize @xmath3 by using @xmath32 , @xmath33 , @xmath34 , and @xmath35 .
each expectation value @xmath36 is obtained from coincidence measurements in 4 different polarization analyzer combinations : the chosen signal and idler set and their orthogonal sets : @xmath37 , @xmath38 , @xmath39 and @xmath40 .
one calculates the expectation value according to @xmath41 where the subscript @xmath42 represents the angle set @xmath37 and the @xmath43 subscript substitutes it with the orthogonal angle .
we made the @xmath3-parameter measurements for the low - flux case with an input pump power of @xmath570@xmath8w at a full divergence angle of 13mrad .
a total of 16 coincidence measurements was made , each consisting of an average of 30 1-s data sets .
we obtain [email protected] , indicating a violation of bell s inequality of more than 6 standard deviations . the entanglement quality given by @xmath3
is comparable to that indicated by the quantum - interference visibility measurements .
a more detailed characterization of the output state can be obtained by quantum state tomography .
one can reconstruct the density matrix , @xmath44 , for a two - photon output state through a series of projective measurements .
the coincidence counts recorded in these measurements can be used to obtain individual density matrix elements by a maximum likelihood algorithm .
we utilize the protocol outlined in @xcite , to estimate @xmath44 from 16 projective measurements for an input pump power of 0.1mw at a full divergence angle of 13mrad . for each projective measurement
, we averaged 30 1-s coincidence counting data points .
coincidence data from these measurements are used in a nonlinear least squares algorithm to estimate the bipartite density matrix from which fidelity measures can be calculated .
the real and imaginary parts of @xmath44 are plotted in fig .
this density matrix estimation yields 98.85% fidelity for the psi singlet state output , where fidelity is calculated as tr@xmath45 , with @xmath46 being the singlet state density operator .
another entanglement measure , tangle , is calculated from the measurements to be 0.9589 , where a factorizable state yields zero tangle and a maximally - entangled bell state yields unity .
we observe that all three types of measurements are equally effective and useful in characterizing the psi output state as highly entangled .
[ figure5 ]
in the previous section the sagnac spdc source output has been characterized in the low - flux limit . from the fig .
3 measurements we have determined that the pair generation rate was 1% per pulse per mw of average pump power for a full divergence angle of 13 mrad .
availability of over 400mw pump power gives us the capability to drive the spdc process at a significant generation rate at which multiple pairs could be produced in large numbers .
this generation capability allows one to tailor the spdc source for a specific application .
one may , for instance , improve the entanglement source performance by increasing the pump power and reducing the divergence angle to achieve the desired generation rate with a higher entanglement quality .
however , a higher flux output comes with its disadvantages .
the main concern is an increase of accidental coincidences that reduces quantum - interference visibility , leading to potential errors in quantum information processing tasks .
take entanglement - based qkd as an example , while a higher secret key rate can be obtained for a mean pair generation number @xmath47 , the maximum @xmath1 that can be used without compromising its security depends on a number of operational system parameters @xcite . as a result , it is highly useful to have a simple and accurate estimate of the pair generation rate under all circumstances . in this section
, we will consider the occurrence of accidental coincidences and how a pulsed source can lead to more errors than a cw source .
we will also concentrate on the effect of multiple pair generation on entanglement quality and how that effect can be utilized to measure the pair generation probability accurately .
there are three main contributions to the observation of accidental coincidences in the typical measurement setup of fig . 1 .
the first type of contribution is due to the generation of independent multiple entangled pairs from a multimode spdc source under strong pumping .
if the detected coincident photon pair originates from two different signal - idler pairs , there is no polarization correlation and therefore an error may occur .
considering the case of two - pair events , the production rate is proportional to the square of the pump power .
the second type of contribution is caused by the detection of one photon from a down - converted photon pair and a uv - induced fluorescence photon emitted by the ppktp crystal which we measured to have a generation rate of less than 5% of the singles count rate .
since the fluorescence varies linearly with the pump power , the accidental coincidence rate is proportional to the square of the pump power .
there is also the possibility of an accidental coincidence due to the detection of two fluorescence photons , but the probability is much lower than the case when one of the detected photons belongs to a down - converted pair .
the third contribution comes from background photons from stray light and from detector dark counts .
both events are independent of the pump power and generally the background and dark counts are low enough that accidental coincidences caused by them are negligible .
the above discussion on accidental coincidences applies to both cw and pulsed spdc .
however , more accidental coincidences are observed in the pulsed case than in the cw case for the same average pumping power .
if we assume the same spdc setup , then the same average input power yields the same number of down - converted pairs per second for the cw and pulsed cases . in cw operation , the spdc pair or fluorescence photon generation probability within a coincidence window of duration @xmath48 is proportional to @xmath48 . the cw accidental coincidence rate resulting from two - pair events
is then given by @xmath49 , where @xmath50 is a proportionality constant .
on the other hand , for pulsed spdc with a repetition rate of @xmath51 , the pair or fluorescence emission is localized within each pump pulse duration , whose width is typically smaller than @xmath48 , as in our setup . therefore , the spdc pair ( or fluorescence photon ) generation probability per pulse ( alternatively , per coincidence window ) is proportional to @xmath52 .
the accidental coincidence probability per pulse is then proportional to @xmath53 , and the accidental coincidence rate is @xmath54 .
note that the cw and pulsed cases have the same proportionality as long as the same type of accidental coincidences is considered . for the same input power ,
the ratio of their accidental rates is @xmath55 using a typical value for the coincidence window duration @xmath56ns , the repetition rate of @xmath57mhz for our pulsed pump implies that @xmath58 .
that is , our pulsed system is much more susceptible to visibility degradation due to accidental coincidences compared with a cw source with the same entangled pair generation rate .
equivalently , the cw source can tolerate a much higher average pump power than a pulsed source ( @xmath518@xmath59 in our case ) for the same amount of accidental coincidences .
this problem can be minimized by increasing the repetition rate until it is comparable to @xmath60 . increasing the repetition rate , however , has its own problems .
for instance , typical si apds are not able to handle detection rates much higher than a few mhz .
also , it is not desirable to have @xmath61 because of the need for temporal separation of the pulses . for a pulsed source designed for a specific application such as qkd , one must be aware of the trade - offs between the need for a high pair generation rate and the desire for high entanglement quality , and strike an application - specific balance for the appropriate combination of pump powers , repetition rates , detector speeds , and error budget .
the uv - induced fluorescence rate was measured by detuning the phase matching temperature so that only fluorescence photons were detectable and we found that it was only a small fraction ( @xmath55% ) of the detected singles rate . therefore multiple - pair generation is the main contributor to accidental coincidences in our setup .
the limitations due to multi - pair events have previously been observed @xcite .
measured the multi - pair visibility degradation in their study of stimulated parametric emission in a multi - pass down - conversion configuration @xcite . in this section
we use a simple model to quantify the effect of multiple - pair generation on two - photon quantum interference measurements in spontaneous parametric emission in a single - pass down - conversion setup .
we compare the theoretical predictions to our experimental observations at high pump powers .
consider an ideal pulsed spdc source that generates singlet - state polarization - entangled photon pairs , and there are no background or fluorescence photons .
we assume that the free - space output is spatially multimode and that the output is also temporally multimode because the coincidence measurement time is large compared with the output pulse width .
the number of entangled pairs in the output pulse is poisson distributed with a mean pair generation number @xmath1 : @xmath62 where @xmath63 is the probability weight for obtaining exactly @xmath64 pairs per pulse .
we further assume that the @xmath64 entangled pairs are independent because of the multimode nature of the output .
that is , each mode has an occupation number of either 0 or 1 .
we note that the assumption of the poisson distribution is not valid for single - mode outputs , but we have found that the results below do not change significantly when we replace it with a thermal distribution .
we focus on the effect of multiple pairs on the two - photon quantum - interference visibility . in calculating the visibility of eq .
( [ v ] ) , we only need to obtain the maximum and minimum coincidence probabilities , @xmath18 and @xmath19 , respectively . for singlet - state
entangled output , we consider a coincidence measurement along @xmath6 polarization for the signal photon and along @xmath7 ( @xmath6 ) polarization for the idler photon to measure @xmath18 ( @xmath65 ) , and we assume ideal polarization analyzers .
other polarization settings ( such as @xmath15-@xmath16 ) work just as well , as long as the polarizers are orthogonal for measuring @xmath18 and parallel for measuring @xmath19 . to complete the model for the calculations , we further assume ideal photon - number non - resolving detectors and both the signal and idler paths have a system detection efficiency @xmath66 .
the minimum and maximum coincidence probabilities are given by @xmath67 where @xmath68 and @xmath69 are the minimum and maximum coincidence probabilities for an output with exactly @xmath64 pairs of entangled photons .
note that the summation starts from @xmath70 because @xmath71 .
for @xmath64 independent photon pairs , there are @xmath72 different ways to arrange their polarization orientations relative to the analyzer settings , and they follow the binomial distribution . for example , the @xmath64 signal photons may be @xmath6-polarized for @xmath73 of them ( with corresponding @xmath7-polarized idler photons ) , and @xmath7-polarized for the rest , or @xmath74 of them , and it has a probability of @xmath75 with the binomial coefficient @xmath76 for @xmath77 . for a given polarization arrangement ,
one requires at least one detected @xmath6-polarized event at the signal detector and at least one detected @xmath7-polarized event at the idler detector to yield a coincidence count for @xmath69 .
similarly , to obtain @xmath68 it is required to detect an @xmath6-polarized event at each detector . combining the above requirements we obtain @xmath78 \left[1-(1-\eta)^{n - k}\right]\ , , \nonumber \\
max}(n)&= & \frac{1}{2^n}\sum_{k=1}^{n } { n\choose k } \left[1-(1- \eta)^k\right]^2\ , .
\label{coeff}\end{aligned}\ ] ] the above coefficients allow us to compute the minimum and maximum coincidence probabilities from eqs .
( [ cmin]-[cmax ] ) and obtain the visibility of eq .
( [ v ] ) as a function of the mean pair generation probability per pulse @xmath1 .
it is instructive to evaluate the multiple - pair effect for small @xmath1 in which case we ignore terms with more than two pairs of entangled photons and simplify the poisson distribution coefficients @xmath79 and @xmath80 . in this case , we obtain @xmath81\,,\end{aligned}\ ] ] and one obtains the visibility to the first order of @xmath1 as @xmath82 the linear dependence of the quantum - interference visibility on the mean pair generation probability shows clearly how the entanglement quality degrades with increasing pump powers .
it is a little surprising to see in eq .
( [ vdependence ] ) that there is no dependence on the system detection efficiency @xmath66 .
a physical explanation is that for a pair of photons that survive the system loss and are detected , one does not know if they belong to the same entangled pair or not and therefore it should only depend on the generation statistics and not on the system loss . for a system with high loss that results in a low detection rate the entanglement quality
may still be compromised by multi - pair events .
the above results can be directly compared with measurements using our setup in fig . 1
. the free - space output of our pulsed sagnac source with a weakly focused pump had many spatial modes that can be described by a poisson distribution .
moreover , the pumping rate was low enough that even at the highest measured @xmath1 value in the experiment , stimulated emission can be neglected .
we measured quantum - interference visibilities in both @xmath6-@xmath7 and @xmath15-@xmath16 bases at different input pump powers and at a fixed full divergence angle of 13mrad .
the pair generation rate per pulse @xmath1 is inferred from singles and coincidence measurements , which have yielded a conditional detection efficiency of @xmath83% under these operating conditions .
the maximum pump power was @xmath570mw which corresponds to @xmath84 .
we measured @xmath5110,000 pairs / s with @xmath51mhz singles rate at this power level .
we did not use higher @xmath1 values to avoid any possible crystal damage at elevated pump power levels .
[ figure6 ] -@xmath7 ( squares ) and @xmath15-@xmath16 ( circles ) bases as a function of mean pair generation number @xmath1 .
the solid curve is computed from our theoretical model .
visibilities in the @xmath15-@xmath16 basis are lower than those in the @xmath6-@xmath7 basis due to other factors that are unrelated to multi - pair occurrences .
inset shows the linear behavior of the quantum interference visibility in the low - flux regime ( @xmath850.1).,width=302 ] figure 6 plots the measured visibilities and compares them with our calculated values .
the @xmath6-@xmath7 results match the theoretical values very well in both the low and high @xmath1 regimes .
in particular , at low @xmath1 , the visibility degradation is linear with respect to the mean pair generation number .
the @xmath15-@xmath16 results also show the expected linear dependence on @xmath1 except there is a fixed amount of visibility loss , which is related to other factors .
we speculate that the discrepancy between the @xmath6-@xmath7 and @xmath15-@xmath16 visibilities is due to wavefront distortion in the optical components of the sagnac interferometer that created partial spatial distinguishability .
we have verified that this problem did not originate from the input pump profile because we observed no visibility improvement after we cleaned up the pump spatial mode by coupling it into a single mode fiber . at this point
, we suspect that the surface quality of the pbs was responsible for the visibility loss .
further testing with a higher quality dual - wavelength pbs is needed to support this claim .
since the @xmath6-@xmath7 measurements were made in the down - converters natural polarization basis , their results were not affected by these distinguishability issues .
it is useful to note that one can take advantage of the @xmath1 dependence of @xmath6-@xmath7 visibility measurements to deduce accurately the mean pair generation probability @xmath1 .
we have presented a compact , pulsed , narrowband source of polarization - entangled photons that is of interest to entanglement - based free - space qkd . pumped by a home - built , narrowband ,
pulsed uv source the sagnac geometry ensures phase - stable generation of any desired bell state at the degenerate wavelength of 780.7 nm . in principle , the bidirectionally pumped single - crystal sagnac configuration eliminates spatial , spectral , and temporal distinguishability in the interferometric combination of the two counter - propagating down - converter outputs .
experimentally , we observed high quality entanglement except for a slight quantum interference visibility loss caused probably by the spatial distinguishability problem .
pulsed operation allows a qkd system to easily manage the arrival times of the photons by synchronization of the transmitter and receiver clocks .
it also provides temporal discrimination against background light that arrives outside of the clocked coincidence window .
the narrowband spdc output allows a matching narrowband spectral filter to be used to screen out ambient light especially during daylight qkd operation .
our pulsed sagnac source is highly efficient and it can be pumped to have a significant pair generation per pulse , producing 0.01 pair per pulse at an average pump power of @xmath51mw at a pump repetition rate of 31.1mhz . we show that at high pair generation the presence of multiple pairs degrades the two - photon entanglement quality that may not be desirable in some quantum information processing tasks . if the pair generation probability per pulse is limited by multi - pair events , then the only way to increase the flux without reducing the entanglement quality is to increase the repetition rate , up to the point where the detector speed may impose a practical limit .
it is common to evaluate the performance of a cw spdc source by its quantum - interference visibility and its spectral brightness , which is the number of detected pairs per second per mw of pump in a 1-nm bandwidth .
however , the spectral brightness is not an appropriate performance metric for a pulsed source . instead of spectral brightness , it is more useful to specify the pair generation probability per pulse and the repetition rate of the source . with the high power pumping capability ,
we have observed the quantum interference visibility as a function of pair generation rate and reached an @xmath1 value as high as 0.7 .
the pulsed sagnac source is capable of achieving a high enough pair generation rate for optimal secret key generation in an entanglement - based qkd system in the presence of channel losses and non - ideal detection efficiencies @xcite .
high pair generation rates have been reported for fiber - based @xmath86 downconversion sources @xcite .
our system performance compares well in terms of achievable generation efficiency with negligible amount of fluorescence and scattering effects . in conclusion ,
we have developed a highly efficient spdc source utilizing the sagnac configuration to generate pulsed , narrowband , polarization - entangled photons .
we have demonstrated high entanglement quality and high pair generation rates with the sagnac source . under strong
pumping , the pulsed source shows entanglement quality degradation due to multi - pair events , in accordance with our theoretical analysis .
the source can be utilized in many quantum information processing applications including quantum key distribution .
in addition , because of its high efficiency and compactness , the sagnac source is also useful for more advanced quantum enabling technologies such as an on - demand source of single photons @xcite . | we demonstrate pulsed operation of a bidirectionally pumped polarization sagnac interferometric down - conversion source and its generation of narrowband , high - visibility polarization - entangled photons .
driven by a narrowband , mode - locked pump at 390.35 nm , the phase - stable sagnac source with a type - ii phase - matched periodically poled ktiopo@xmath0 crystal is capable of producing 0.01 entangled pair per pulse in a 0.15-nm bandwidth centered at 780.7 nm with 1mw of average pump power at a repetition rate of 31.1 mhz .
we have achieved a mean photon - pair generation rate of as high as 0.7 pair per pulse , at which multi - pair events dominate and significantly reduce the two - photon quantum - interference visibility . for low generation probability @xmath1 , the reduced visibility
@xmath2 is independent of the throughput efficiency and of the polarization analysis basis , which can be utilized to yield an accurate estimate of the generation rate @xmath1 . at low @xmath1
we have characterized the source entanglement quality in three different ways : average quantum - interference visibility of 99% , the clauser - horne - shimony - holt @xmath3 parameter of @xmath4 , and quantum state tomography with 98.85% singlet - state fidelity .
the narrowband pulsed sagnac source of entangled photons is suitable for use in quantum information processing applications such as free - space quantum key distribution . |
graphene , a newly realized two - dimensional crystal of carbon atoms ordered on a honeycomb lattice,@xcite is being extensively studied due to its unusual electronic and structural properties .
undoped graphene is a zero - gap semiconductor with a linear low - energy dispersion relation and a vanishing density of states at the fermi level.@xcite although the coulomb interaction is unscreened in this regime , the system behaves mostly as a non - interacting electron liquid , where the minor effects of interactions are encoded in a renormalization of the fermi velocity and the quasiparticle lifetime.@xcite the opposite case is that of a divergent density of states , the so called van hove ( vh ) singularity , which is associated to a saddle point in the band dispersion.@xcite if the fermi level lies near such a singularity , screening is perfect for wave - vectors connecting vh singularities and correlation effects may be enhanced .
recently , several experimental groups have succeeded on making this regime accessible for graphene.@xcite the techniques used for this aim are different : in ref .
they have used twisted graphene , obtained from a rotation between stacked graphene layers , which allows one to tune the position of the vh singularity .
other methods involve chemical doping of graphene@xcite or intercalation of gold clusters between the graphene layers.@xcite this opens up the fascinating possibility to study correlated electronic phases in this material , such as superconductivity , charge - density wave ( cdw ) or spin - density wave ( sdw ) .
the proximity of the fermi level to a vh singularity in most of the cuprate superconductors has triggered large efforts to understand the role of a peaked density of states ( dos ) on the electronic properties of an electron liquid.@xcite superconductivity , itinerant ferromagnetism ( fm ) , cdw , and sdw are examples of competing instabilities associated with the vh scenario . in graphene ,
the existence of some of these instabilities has already been evidenced by recent experiments , such as cdw@xcite or superconducting pairing due to electron - electron interactions,@xcite following the kohn - luttinger mechanism.@xcite cdw and sdw phases usually occur in systems with fermi surface nesting , i.e. when the fermi surface can be mapped onto itself by a ( nonzero ) @xmath2-vector ( nesting vector ) . at low doping , graphene shows some nesting , since the fermi surface is composed of two circles around the @xmath3 and @xmath4 points .
however , since the dos is very low , these peaks are small and a very high ( hubbard ) interaction is required to enter the dw regime.@xcite the situation changes if we strongly dope graphene around the value @xmath5 , where @xmath6 is the chemical potential and @xmath7ev is the hopping parameter .
the fermi surface then acquires a hexagonal shape and nesting occurs in three directions ( see fig .
[ fig1 ] ) .
furthermore , the vh singularities lie at the fermi surface for this doping . at these points ,
the dos diverges and we expect some non - trivial peaks in the susceptibility . in this paper , we investigate the hubbard model in a honeycomb lattice using a field theoretical approach . by performing a hubbard - stratonovich transformation , we determine the effective action in terms of an 8-component order parameter , which accounts for charge- and spin - degrees of freedom in the a and b graphene sublattices .
we find that the charge susceptibility never diverges for repulsive interactions , thus excluding the possibility of a cdw formation , whereas the spin susceptibility exhibits several peaks .
the peaks at the nesting wavevectors @xmath2 are stronger than the one at @xmath8 ( see fig .
[ fig2 ] ) , thus signaling that sdw is the leading instability , that wins against fm for repulsive interactions . in the following ,
we introduce the model and outline the main steps of the calculations in sec .
, we present the temperature vs interaction vs doping phase diagram and show that the metallic state becomes unstable towards a sdw phase for realistic values of the interaction parameter @xmath9 in sec .
finally , we discuss possible experimental techniques which could be used to observe our results and draw our conclusions in sec .
due to the strong screening of interactions by the electron gas at the vh filling , we consider only the local coulomb repulsion ( the usual hubbard u term ) , which is a good approximation around this doping level.@xcite this term leads to a deformation of the band dispersion toward the saddle point,@xcite an effect that has been observed experimentally by arpes.@xcite notice that longer - range interaction terms , such as the nearest - neighbor repulsion , would lead to a richer phase diagram , with the possibility of a pomeranchuk instability.@xcite this scenario is in sharp contrast with low doped graphene , for which the dos vanishes and the coulomb interaction will be only slightly screened and therefore will be long ranged .
hence , we use a tight - binding model with a hubbard interaction @xmath9 , @xmath10 where the operator @xmath11 can be either @xmath12 or @xmath13 , depending whether @xmath14 is a label of the @xmath15 or @xmath16 sublattice , respectively . defining @xmath17
, the hubbard term can be rewritten using the relation , @xmath18 where @xmath19 is the on site number operator and @xmath20 , describes the spin on the lattice site @xmath14 .
note that the inner product @xmath21 , where @xmath22 is some vectorial field , is defined by @xmath23 , and that @xmath24 is the vector of pauli matrices .
the grand - canonical partition function describing the system reads @xmath25 \textrm{d } [ c ] e^{-s[c^\dagger , c]/\hbar},\ ] ] where the action is given by @xmath26.\ ] ] in this expression , @xmath27 is the imaginary time variable , @xmath28 and @xmath29 is defined in eq .
( [ eq1h ] )
. we will investigate the possible appearance of cdw and sdw instabilities by using a path integral formalism recently developed by some of the authors.@xcite in the following , we briefly outline the procedure .
we start by performing a hubbard - stratonovich transformation that eliminates the quartic term in the action , but introduces eight auxiliary bosonic fields @xmath30 ,
@xmath31 , @xmath32 , and @xmath33 related , respectively , to the electronic and the magnetic degrees of freedom of the fermionic fields of each sublattice . in fourier space , the action then reads @xmath34,\end{aligned}\ ] ] where @xmath35 .
the inverse green s function is defined by @xmath36 , where the bare green s function reads @xmath37 \delta_{{{\bf k}},{{\bf k ' } } } \delta_{n , n'},\end{aligned}\ ] ] with @xmath38 a @xmath39 identity matrix , @xmath40 the fermionic matsubara frequency , and @xmath41 . in what follows we set the lattice constant @xmath42 .
the self energy is given by @xmath43,\end{aligned}\ ] ] where @xmath44 is the number of sites of each sublattice and both , @xmath45 and @xmath46 carry a subindex @xmath47 . by introducing the eight component vector @xmath48 and integrating out the fermionic fields
, we obtain the effective action @xmath49,\ ] ] where @xmath50 and the partition function becomes @xmath51 \exp(-s_\textrm{eff}/\hbar)$ ] .
finding the susceptibility from here is a two steps process .
first , we introduce external source fields @xmath52 and then we expand the fields in the action around their mean field value @xmath53 .
this results in a self - consistent equation for the mean field values of the fields . using this equation
, we may determine how the bosonic fields react on a distortion of the source fields .
for @xmath54 .
brighter regions correspond to higher peaks in the spin susceptibility .
the peaks at the three nesting wavevectors ( see fig . [ fig1 ] ) are higher than the one at @xmath55 , indicating that the sdw instability wins over the fm one . ]
the mean field values are those at which the action has an extremum .
thus , the linear terms in @xmath56 are imposed to vanish , yielding the self - consistent equation @xmath57,\ ] ] where @xmath58^{\rm t}$ ] and the trace is taken over the matrices only , because the sum takes care of the trace in @xmath2- and frequency - space .
after differentiating this equation with respect to the source @xmath52 , one finds the susceptibility through the relation @xmath59 where @xmath60 moreover , we can also determine the free energy by performing a legendre transformation @xmath61= u^{-1}\langle\mathbf{m}\rangle^{\dag}_{j } \cdot \mathbf{j}-\ln(z[\mathbf{j}])$ ] on the partition function which , up to quadratic order in the deviation @xmath62 and without an additive constant , is @xmath63=\frac{1}{2\hbar u^2}\left ( \delta\langle\mathbf{m}\rangle_{j}\right)^{\dag } \cdot ( \mathbf{\chi}_{0}^{\rm rpa})^{-1 } \cdot\delta\langle\mathbf{m}\rangle_{j}.\ ] ] the susceptibility @xmath64 is evaluated in the absence of the source field , @xmath65 . from here on , we assume that the system is in a homogeneous state , for which the mean field values of the boson fields are given by @xmath66 as a consequence , the self - energy , the green s function and the susceptibility matrix @xmath67 are all diagonal in momentum and frequency space . due to a nonzero self - energy , the hamiltonian gets renormalized , such that @xmath68 where @xmath69 is given by the first term in eq .
( 1 ) . assuming a ( renormalized ) hamiltonian , which may be diagonalized by using the unitary operators @xmath70 , such that @xmath71 where @xmath72 and @xmath73 , the susceptibility can be written as @xcite @xmath74,\end{aligned}\ ] ] where @xmath75 are bosonic matsubara frequencies and @xmath76 is the fermi distribution function .
notice that the overlap between the electron and hole wave - functions is taken into account through the term @xmath77 . for @xmath78 for @xmath79 .
notice that the peak corresponding to sdw at @xmath80 is larger than the fm peaks at @xmath81 and @xmath82 . ]
.numerical values of the critical coupling @xmath83 ( in units of @xmath84 ) for a fm and a sdw phases calculated on a 10000@xmath8510000 mesh .
[ cols="<,<,<,<",options="header " , ]
next , we investigate the possible instabilities by considering the static susceptibility @xmath86 . to evaluate the latter , one needs to perform a sum over all momenta in the brillouin zone .
we proceed by dividing the brillouin zone in an @xmath87 mesh , which implies that the sum in the expression for the susceptibility has @xmath88 terms .
the instability condition for repulsive interactions requires the interaction strength @xmath9 to exceed the critical value @xmath83 defined by @xmath89 this relation links the critical interaction strength to the largest eigenvalue @xmath90 of the matrix @xmath91 by @xmath92 in fig .
[ fig2 ] a density plot of the largest eigenvalue of the matrix @xmath91 is shown for a temperature of @xmath93 k. one can distinguish four inequivalent peaks ( bright regions in fig .
[ fig2 ] ) .
the one at @xmath94 corresponds to a fm instability , while the other three , at nonzero @xmath95 , correspond to the three independent nesting vectors ( see fig .
[ fig1 ] ) and hence give rise to a sdw .
although not expected , for all temperatures the sdw peaks turn out to be higher than the peak corresponding to the fm phase ( see fig . [ fig3 ] ) .
however , this result depends heavily on the size of the mesh : for low temperatures , the sum converges slowly , and a wrong result can be inferred if @xmath44 is not large enough .
indeed , by considering a @xmath96 mesh , we find that a fm phase would set in at low temperatures [ see fig .
[ fig4c ] ] , whereas the results for a finer mesh ( @xmath97 ) indicate that the true ground state is a sdw [ see fig .
[ fig4a ] and fig . [ fig4b ] ] .
scales as @xmath98 and will therefore increase if the doping level is tuned away from @xmath99 . ] for zero temperature and a doping value exactly at the vh singularity , the critical coupling for both fm and sdw is zero . as explained below ,
our formalism is not suited to determine which state is more favorable . for finite temperatures ,
the sdw phase transition has a lower critical @xmath9 than the fm one .
therefore , between the red and blue lines in fig . [ fig4 ]
we find a sdw phase . above the blue line ,
the system allows for a fm phase transition .
the starting point of our formalism is a homogeneous ground state , after which we can determine the critical @xmath9 that is needed to enter a more ordered phase , like the fm , sdw or cdw phase . to reach the regime above the blue line in fig .
[ fig4 ] , one has to start from a sdw ground state and subsequently determine if a phase transition to the fm phase will occur .
since the sdw ground state is inhomogeneous , our formalism is thus unable to determine the leading instability when @xmath9 is in this regime .
if the system is tuned away from optimal doping , the vh singularities will no longer lie on the fermi surface and nesting is reduced in general .
this will result in a lowering of the peaks in the maximal eigenvalue of the susceptibility and hence the critical couplings will increase .
this behavior is shown in fig .
the sdw peak is always higher than the fm one , but the height decreases quite rapidly as function of doping . the interaction strength in graphene can not be tuned externally . although at low doping the onsite hubbard interaction is estimated to be around @xmath100,@xcite values of @xmath9 for highly doped graphene are , to the best of our knowledge , not yet known , but should not be different from the low - doping values .
we therefore expect that a sdw phase should be observed experimentally , in the regime of @xmath1 and @xmath9 parameters shown in fig .
[ fig5 ] ( see also table [ tab1 ] ) .
recently , a cdw has been experimentally observed in twisted graphene bilayers doped at the vh singularity , but it has disappeared when the hopping parameter between the two layers has been put to zero.@xcite a cdw phase seems also to be present for graphene grown on top of a superlattice of gold intercalated clusters,@xcite and a superconducting phase has been conjectured to occur for graphene heavily doped with ca and k on both sides.@xcite however , it is unclear which is the appropriate theoretical model to describe graphene under these circumstances , and for the conventional hubbard model in a single layer , a cdw can only occur for attractive on - site interactions . in conclusion ,
we have evaluated possible sdw and cdw instabilities from a metallic phase for the honeycomb lattice of a single layer of graphene doped at the vh filling .
we found that charge and spin degrees of freedom are decoupled , and that cdws occur for attractive , whereas sdws occur for repulsive interactions .
a peak in the spin - susceptibility at zero wave - vector has also been found , although precise calculations indicate that it is always sub - dominant with respect to the ones appearing at the nesting wavevectors , thus determining that the sdw instability must win over fm in the neighborhood of the metallic phase .
this finding contradicts a previous result in the literature,@xcite and could only be reached after a very careful and time consuming computation in a very large mesh .
an interesting point is that the sdw phase is expected to occur at rather high temperatures .
this is not very surprising , given that phenomena such as cdw or the integer quantum hall effect , that usually occur at very low temperatures , have already been detected at high temperatures in graphene ( @xmath101k for the first and room temperature for the latter).@xcite we hope that our results may stimulate further experiments to unveil the existence of a sdw phase in heavily doped graphene .
we acknowledge financial support from the netherlands organization for scientific research ( nwo ) .
r. roldn thanks the eu - india fp-7 collaboration under monami .
we are indebted to beln valenzuela for a careful reading of this manuscript and for pointing out some ambiguities in the first draft of this paper .
novoselov , a.k .
geim , s.v .
morozov , d. jiang , y. zhang , s.v .
dubonos , i.v .
grigoriev , and a.a .
firsov , science * 306 * , 666 ( 2004 ) . for a review on graphene ,
see : a.h.c neto , f. guinea , n.m.r .
peres , k. novoselov , and a.k .
geim , rev .
phys . * 81 * , 109 ( 2009 ) .
j. gonzlez , f. guinea , and m.a.h .
vozmediano , nucl .
b * 424 * , 595 ( 1994 ) .
l. van hove , phys . rev . * 89 * , 1189 ( 1953 ) .
g. li , a. luician , j.m.b.l .
dos santos , a.h.c .
neto , a. reina , j. kong , and e.y .
andrei , nature physics * 6 * , 109 ( 2010 ) .
mcchesney , a. bostwick , t. ohta , t. seyller , k. horn , j. gonzlez , and e. rotenberg , phys .
* 104 * , 136803 ( 2010 ) .
m. cranney , f. vonau , p.b .
pillai , e. denys , d. aubel , m.m.d .
souza , c. bena , and l. simon , epl . *
91 * , 66004 ( 2010 ) . for a comprehensive review of the van hove scenario in the context of high-@xmath102 superconductors ,
see : r.s .
markiewicz , j. phys .
solids * 58 * , 1179 ( 1997 ) . w. kohn and j.m .
luttinger , phys . rev
. lett . * 15 * , 524 ( 1965 ) .
n. m. r. peres , m. a. n. araujo , and daniel bozi , phys .
b * 70 * , 195122 ( 2004 ) .
j. gonzlez , f. guinea , and m.a.h .
vozmediano , nucl .
b * 485 * , 694 ( 1997 ) .
r. roldn , m.p .
lpez - sancho , and f. guinea , phys .
b * 77 * , 115410 ( 2008 ) . t. stauber , phys
b * 82 * , 201404(r ) ( 2010 ) .
b. valenzuela , and m. a. h. vozmediano , new journal of physics * 10 * , 113009 ( 2008 ) .
lamas , d.c .
cabra , and n. grandi , phys .
b * 80 * , 075108 ( 2009 ) .
d. makogon , i. spielman , and c.m .
smith , arxiv:1007.0782 ( 2010 ) .
t. wehling , e. sasioglu , c. friedrich , a. lichtenstein , m. katsnelson , and s. blgel , phys .
lett . * 106 * , 236805 ( 2011 ) .
k. s. novoselov , z. jiang , y. zhang , s. v. morozov , h. l. stormer , u. zeitler , j. c. maan , g. s. boebinger , p. kim , a. k. geim , science * 315 * , 1379 ( 2007 ) . | we study the instability of the metallic state towards the formation of a new ground state in graphene doped near the van hove singularity .
the system is described by the hubbard model and a field theoretical approach is used to calculate the charge and spin susceptibility .
we find that for repulsive interactions , within the random phase approximation , there is a competition between ferromagnetism and spin - density wave ( sdw ) .
it turns out that a sdw with a triangular geometry is more favorable when the hubbard parameter is above the critical value @xmath0 , which depends on the temperature @xmath1 , even if there are small variations in the doping .
our results can be verified by arpes or neutron scattering experiments in highly doped graphene . |
nonlinear physics of dna has experienced an enormous development in the previous years .
there are many experimental data and theoretical results published about the nonlinear properties of dna ( for a review see , e.g. , @xcite .
the dna molecule is a discrete system consisting of many atoms having a quasi - one - dimensional structure .
it can be considered as a complex dynamical system , and , in order to investigate some aspects of the dynamics and the thermodynamics of dna , several mathematical models have been proposed . among them , it is worth remarking the peyrard bishop model @xcite introduced for the study of dna thermal denaturation .
this model , and some variations of it , has also been used extensively for the study of the dynamical properties of dna . in dna there
exist different kinds of interactions between the main atomic groups .
one of them is the stacking interaction between neighbouring bases along the dna axis , these are short range forces which stabilize the dna structure and hold one base over the next one forming a stack of bases .
there exist also long range forces , due to the finite dipole moments of the hydrogen bonds within the nucleotides .
the existence of these forces is corroborated both by theoretical and experimental studies .
recently , quantum chemical calculations , using the second moler - plesset perturbation method , have determined bonds lengths and bonds angles involved in hydrogen bonds between the bases of dna base pairs , as well as dipole moments based at the mp2/6 - 31g(d ) and mp2/6 - 31g(d , p ) basis sets @xcite .
the dipole moments calculated for the adenine - thymine base pair at equilibrium are , respectively , 1.44 d and 1.29 d. the values calculated for the guanine - cytosine base pair at equilibrium are , respectively , 5.88 d and 5.79 d. all these values are within the range of the parameters values for the dipole moments considered in this paper .
some experiments carried out on the dynamics of the b - a transition of dna double helices , through the analysis of the transients observed under electric field pulses , show that dipolar strectching is the main driving force for the b - a reaction @xcite .
the dipole
dipole interactions between different base pairs should never be ignored and they may play a crucial role for the dynamical properties of dna , for example when the geometry of the double strand of dna is taken into account , as the distance between base pairs and , therefore , the intensity of the coupling between them , depends on the shape of the molecule @xcite . most of the studies have been done considering dna homogeneous models .
nevertheless , the dna molecule is essentially an inhomogeneous system , and this inhomogeneity is characterized by the existence of different local ranges , or functional regions , with very specific sequences of base pairs and very specific functions .
numerical simulations show that discrete breathers ( dbs ) , which are spatially localized nonlinear oscillations , can appear in models of crystals , biomolecules , and many others nonlinear discrete systems @xcite .
it is well known that dbs can be static , but they can also move under certain physical conditions @xcite , constituting a mechanism for the transport of energy and information along discrete systems .
they are called _ moving breathers _ ( mbs ) and have been obtained in many different systems @xcite . in the peyrard bishop model ,
the existence of dbs has been demonstrated @xcite , and dbs are thought to be the precursors of the bubbles that appear prior to the transcription processes in which large fluctuations of energy have been experimentally observed @xcite .
the study of mbs in klein - gordon models of dna chains with the inclusion of long range interactions was initiated in @xcite using an augmented peyrard bishop model .
the system considered is homogeneous and without bending , that is , all the base pairs are identical , aligned in a straight line , and all the dipole moments are parallel along the same direction .
the results show that mbs exist for a wide range of the parameter values , although the mobility is hindered when the intensity of the long range interaction increases .
essentially , the effective mass of a breather increases as the intensity of the dipole
dipole interaction increases , and there exists a threshold of the dipole
dipole coupling constant above which the breather is not movable .
the interaction of mbs with different kind of inhomogeneities in the dna molecule may be also of great importance for the localization of energy @xcite . as a consequence of this interaction , mbs
may be trapped in small regions , that is , vibrations remain affecting only to a small number of consecutive base pairs , producing an accumulation of energy , which can initiate denaturation or transcription bubbles .
also , trapping phenomena may be on the basis for the appearance of secondary breaks when dna is irradiated with ionizing radiation @xcite .
structural studies carried out with non coding sequences of dna have shown that poly(da.dt ) sequences ( called t - tracts ) are abundant genomic dna elements .
suter et al .
@xcitehave shown that they exist as rigid dna structures in nucleosome - free yeast promoters in vivo , that is , in vivo they are not folded in nucleosomes .
their data support that transcription activation depends on the length of the t - tracts and the same occurs when they are replaced by poly(dg.dc ) sequences , another rigid structures .
this fact suggests that the t - tracts operate not by recruitment of specific transcription factors , but rather by their intrinsic dna structure , and that transcription activation must be intimately related with the dynamical properties of t - tracts .
these considerations motivated us to study mbs in some types of poly(da.dt ) or poly(dg.dc ) sequences . in this paper
we study the effects of the collisions of mbs with local inhomogeneities in two different but related systems .
the first system consists of an inhomogeneous dna molecule formed by two consecutive homogeneous regions . all the base pairs are identical , but the dipoles of a region are oriented in opposite direction to the dipoles of the other one , as sketched in fig .
[ fig1](a ) .
one possibility is for a sequence being as @xmath0 , where @xmath1 and @xmath2 represent , respectively , the adenine - thymine base pair and the thymine - adenine base pair .
the other possibility is for a sequence being as @xmath3 , where @xmath4 and @xmath5 represent , respectively , the cytosine - guanine base pair and the guanine - cytosine base pair .
the inhomogeneity is determined by the existence of an interface that separates these regions , that is , there exists a local inhomogeneity and the interface marks a discontinuity in the system .
these kind of chains are called _ interfaced piecewise chains _ , they exist in real dna and can also be synthesized in the laboratory @xcite .
the second system consists in a dna chain of identical base pairs with all the dipoles oriented in the same direction , except one of them which has an opposite orientation , that is , the local inhomogeneity is at the site of the inverted dipole , as sketched in fig .
[ fig1](b ) .
in both systems , the simulations of mbs launched towards the inhomogeneity show that if the translational kinetic energy is below a critical value , which is determined by the values of the coupling parameters , the breather is trapped . for greater values , the breather
is transmitted to the chain , losing part of its energy as phonon radiation . moreover ,
reflection of the colliding breather never occurs .
these results emphasizes the enormous importance of a change of the orientation of a single base pair in a homogeneous track of dna chain , as it demonstrates that a very simple structural change creates a mechanisms for localization and accumulation of energy .
moving breathers can also collide with a trapped breather .
our study of these collision processes shows that multiple scenarios are possible , one of them is the possibility of getting an accumulation of energy at the interface up to a saturation threshold value .
all the accumulated energy can be carried away by a successful incoming breather .
these phenomena suggest a way for the dna molecule to accumulate enough energy to break the hydrogen bonds between the base pairs on opposing strands .
another important result concerning these systems , with a local inhomogeneity and with competing short and long range interactions , is that there exists a relevant parameter determining the range of translational kinetic values for which mbs are trapped , this is the quotient between the dipole
dipole and the stacking coupling constants .
the critical kinetic energy increases , or the size of the `` trapping window '' , if this quotient increases . this paper is organized as follows : in section [ model ]
, we introduce the interfaced piecewise dna model and the single inverted dipole dna model , which includes the long range interactions and has into account their respective chain conformations . in section [ linmod ]
, we investigate the characteristics of the linear modes of the first system , as they are important for determining possible behaviors of mbs when they collide with the interface .
it is shown that the linear modes spectrum of this system is similar to the spectrum of a homogeneous dna molecule with a single inverted dipole . in section [ numres ] , we generate numerically mbs which are launched towards the interface and we study the collision phenomena , exploring all the possible scenarios that appear when the translational kinetic energies of the mbs and the coupling parameters are varied .
thereafter , we show that the collisions of mbs in the case of a single inverted dipole dna molecule , bring about similar effects to the interfaced case .
section [ collisions ] presents some interesting phenomena that can appear when moving breathers collide with a previous trapped breather .
finally , section [ conclu ] summarizes our results and contains some conclusions .
we consider a modification of the peyrard - bishop dna model @xcite , with the addition of a dipole - dipole energy term .
thus , the hamiltonian of the system can be written as @xcite @xmath6 the term @xmath7 represents the kinetic energy of the nucleotide of mass @xmath8 at the @xmath9th site of the chain , and @xmath10 is the variable representing the transverse stretching of the hydrogen bond connecting the base at the @xmath9th site .
the morse potential , i.e. , @xmath11 , represents the interaction energy due to the hydrogen bonds within the base pairs , being @xmath12 the well depth , which corresponds to the dissociation energy of a base pair , and @xmath13 is related to the width of the well .
the stacking energy is @xmath14 , where @xmath15 is the stacking coupling constant .
the last term of the hamiltonian , i.e. , @xmath16 is the long - range dipole
dipole interaction term , where @xmath17 is the dipole dipole coupling constant .
the expression for this constant is @xmath18 @xcite , @xmath19 being the charge transfer within the dipole and @xmath20 the distance between neighbouring base pairs , which is supposed to be constant .
finally , @xmath21 is the dipole - dipole coupling factor given by @xmath22 where @xmath23 is the normalized distance between base pairs , and @xmath24 takes the value @xmath25 if the dipoles at @xmath9 and @xmath8 are parallel , and @xmath26 if they are antiparallel . in order to keep the spatial homogeneity in a finite system with periodic boundary conditions @xcite , the number of particles affected by the long range interaction must be limited . in consequence , if n is the number of base pairs , we suppose that the long range interaction affects to @xmath27 neighboring base pairs , if n is odd , or @xmath28 , if n is even , to each direction of a given site of the chain .
in fact , we use periodic boundary conditions in a way that both the last dipole and the first one are pointing along the same direction , this implies that the interfaced piecewise chain must has the geometry of a mbius band , and in this way there exists only one interface in the chain . as an example , we can consider a small system with @xmath29 , with the dipoles located at @xmath30 pointing in one direction and the other ones located at @xmath31 pointing in opposite direction . then , the dipole dipole coupling factors @xmath32 can be written in a matrix form as follows : + @xmath33 we perform the following changes of variables : @xmath34 where @xmath35 is the frequency of an isolated oscillator in the harmonic limit . it becomes unity in the scaled system . with these changes ,
the dynamical equations become : @xmath36
the study of the linear modes ( phonons ) of a nonlinear discrete system gives a necessary information for predicting some properties of the mbs with a given frequency .
for example , the frequencies of mbs must be not too close to the phonon band .
otherwise , they emit a large amount of phonon radiation simultaneously to their movement . furthermore
, as it was shown in @xcite , the analysis of linear modes allows to predict the scenario of the collisions of mbs with inhomogeneities without the necessity of performing numerical simulations .
the dynamical equations of the interfaced piecewise system can be linearized if we assume that the amplitudes of the oscillations are small enough .
then , the linearized dynamical equations are : @xmath37 the inhomogeneity of these equations determines the existence of two localized modes , one of them is above ( top mode ) and the other one is below ( bottom mode ) the extended modes ( phonon band ) .
we have calculated numerically the frequencies of the linear modes for different values of the parameters @xmath15 and @xmath17 .
[ fig2](left ) shows the dependence of the frequency spectrum with respect to the parameter @xmath17 for the fixed value of @xmath38 , which is an appropriate value to obtain mbs with low dispersion @xcite . in this figure
, it can be observed the existence of localized modes whose origin relies in the fact that the interface acts as an inhomogeneity in the lattice .
the localized modes are two , one of them is above ( top mode ) and the other one is below ( bottom mode ) the extended modes ( phonon band ) . for @xmath38 .
the upper line represents the top mode and the lower line represents the bottom mode .
right : top and bottom mode frequencies with respect to @xmath17 .
crosses and dots represent , respectively , analytical and numerical values.,scaledwidth=70.0% ] the generic profiles of the localized modes are shown in fig .
their vibration patterns depend on the parameter @xmath17 .
in particular , we have observed that there exists a critical value @xmath39 such that for @xmath40 , the top mode has a zig - zag vibration pattern , while the bottom mode vibrates in phase . for @xmath41 ,
the vibration patterns are interchanged .
as it is shown below , @xmath42 ( e.g. for @xmath38 , @xmath43 ) .
this fact is related to the existence of a competition between the attractive ( stacking ) and the repulsive ( dipole dipole ) interaction .
the first one is dominant for @xmath44 and vice versa . and
@xmath38.,scaledwidth=70.0% ] an approximate explicit expression that gives the dependence of the localized modes frequencies with respect to the coupling constant @xmath17 can be obtained considering that the long - range interaction is limited to nearest neighbour base pairs .
a study of infinity - range interaction could be performed , but it would be limited to frequencies close to the phonon band @xcite .
this approximation is justified as the interaction decays rapidly .
we have checked that the results are practically coincident except for values of @xmath17 far from the ones considered in our study .
the localized modes can be found using green s lattice function methods @xcite .
however , the same results can be obtained in a more straightforward way using the following ansatz : @xmath45 where r is a spatial decay factor .
the sign of r indicates the vibration pattern of the localized mode . if @xmath46 , the particles vibrate in phase , and if @xmath47 , the mode has a zig - zag vibration pattern .
we have supposed that the dipoles at @xmath48 are all antiparallel to the dipoles at @xmath49 , then , there exists only two neighboring dipoles which have a different orientation , and the interface is located between @xmath50 and @xmath51 . as the localized modes are centered between both ( see fig .
[ fig3 ] ) , we assume that : @xmath52 with the application of these anstze we obtain that the frequencies of the top and bottom modes are , respectively : @xmath53 the decay factors for the top and bottom modes are , respectively : @xmath54 > from these expressions , it is easy to obtain that @xmath55 .
the discrepancies from this value and the numerical one rely on our assumption that long - range interactions are limited to nearest neighbours .
the values of the parameters for which the breather frequency @xmath56 are related to the scattering properties , see ref .
however , these values are excluded in our study , because as we require the existence of moving breathers , we need that @xmath57 and @xmath58 , with @xmath59 @xcite . in other words
, @xmath60 is a necessary condition for the existence of moving breathers .
the equality @xmath56 implies that @xmath61/2 $ ] , with @xmath62 .
thus , it is straightforward to demonstrate that @xmath17 must be greater than @xmath63 in order to get breather resonance with the bottom mode frequency .
this region is inaccessible for the moving breathers , as it is shown above .
the analytical and numerical results that give the dependence of the frequencies of the localized modes with respect to @xmath17 are represented in fig .
[ fig2](right ) .
as can be appreciated , there is an excellent agreement between the two approaches .
the single inverted dipole system also has two linear localized modes , but now they are localized at the inverted dipole site .
the frequencies of these modes are : @xmath64
the dynamics of mbs can be strongly affected by the existence of the interface .
it is expected that mbs differing only in their translational kinetic energies can produce different effects when they interact with the interface . in order to study these effects , we have simulated many collisions processes launching mbs towards the interface with different translational kinetic energies .
mbs can be generated numerically by adding to the velocities of a static breather a perturbation of magnitude @xmath65 collinear to the pinning mode @xcite .
the pinning mode is an anti
symmetric linear localized mode , which may appear in the set of linear perturbations of the system when the coupling is strong enough .
this perturbation breaks down the translational symmetry of the system and the breather moves with a translational kinetic energy given by @xmath66 .
the time evolution of the breathers have been studied using a calvo s 5th order symplectic algorithm @xcite .
recent quantum mechanical calculations ref .
@xcite establishes that in dna the order of magnitude of the parameter @xmath17 is @xmath67 , therefore , as we are interested in applications to dna , we have limited our study to this order .
mbs are not exact solutions of the dynamical equations and some phonon radiation is emitted simultaneously to the movement of breathers .
this radiation is minimized if the breather frequency is not close to the phonon band . in all the simulations ,
the breather frequency has been fixed to the value of @xmath68 , which is an appropriate value . for smaller values of @xmath69 ,
the breather gets more and more sharply localized and highly pinned to the lattice , and for @xmath70 its movement is not possible . also , phonon radiation increases when the value of the translational kinetic energy increases and some effects due to the size of the system can appear . for that reason
the calculations have been tested using different systems sizes , to make sure that the results are not modified by boundary effects ( most of the simulations have been done with @xmath71 ) .
the values of the parameter @xmath65 have been limited to the interval @xmath72 , as for this interval the phonon radiation is small enough .
finally , the first group of simulations have been done with the value of the stacking coupling constant @xmath38 .
we have performed a large number of simulations and , essentially , we have found that there exist only two different dynamical regimes , which can be characterized as follows : trapping regime : : : + if the translational kinetic energy of a mb is smaller than a critical value , i.e. @xmath73 , the breather is trapped at the discontinuity region , that is , only some particles around the interface remain oscillating .
the critical translational kinetic energy depends on the values of the parameters @xmath74 and @xmath75 .
an example of the time evolution of a mb with @xmath65 smaller than the critical value , is observed in fig .
[ fig4](a ) . however , not all the moving breather energy is trapped after the collision because some energy is transferred to the chain as phonon radiation . to see this ,
we have studied the evolution of the `` central energy '' , i.e. the energy of some few particles around the interface , which is represented in fig .
[ fig5](a ) . before
the collision the central energy is zero , then it increases quickly taking the value of the incident breather energy , and a small decay of this energy occurs which is due to phonon radiation .
this can be appreciated in the central zoom figure of fig .
[ fig5](a ) . + and ( b ) @xmath76 , with @xmath77 , @xmath38 and @xmath68 . for these parameters
values , @xmath78 .
the interface is located between @xmath50 and @xmath51.,scaledwidth=70.0% ] + ( a ) ; ( b ) in fig .
[ fig4](b ) .
the central figure is a zoom of the left one , scaledwidth=90.0% ] + the analysis of the fourier spectrum of this trapped breather , carried out after the initial decay of the central energy and at an early stage of the evolution , shows a frequency of value 0.8 , which is the internal frequency @xmath75 of the launched breather , and a frequency of value 0.02 , which corresponds to an oscillating movement of the breather around the interface .
this type of movement can be appreciated in fig .
[ fig4](a ) .
the absolute values of the fourier components calculated soon after the collision of the breather with the interface are represented in fig .
[ fig6 ] , and the central energy of this breather has the value @xmath79 .
the same calculations carried out after a time of @xmath80 breather periods , shows a decrease of the fourier component corresponding to @xmath81 , and a central energy of value @xmath82 .
a very slow decaying process occurs and the oscillating trapped breather approaches to a static breather centered at the interface with frequency @xmath83 . the energy of this existing static breather is @xmath84 .
then , a necessary condition for the trapping effect is the existence of an _ inhomogeneity breather solution _ ( ib ) with the frequency of the mb and with a smaller energy .
it consists of a static breather solution centered at the inhomogeneity to which the trapped breather decays .
+ ( a ) calculated soon after the collision , scaledwidth=60.0% ] + the trapping phenomena for small enough translational kinetic energies , and the oscillatory behavior as a whole of the trapped breather can be partially explained with the help of the collective coordinate method .
+ in fact , the long range interaction term of the hamiltonian may be represented as follows : @xmath85 where the first summation term in the rhs represents an effective on - site energy @xcite , with the on - site potential + @xmath86 + and the second summation term represents an effective dispersion energy . in the case of a spatially homogeneous system
the on - site potential does not depend on the index n , but for the interfaced piecewise system the potential @xmath87 , with @xmath77 , has the well profile shown in fig .
[ fig7 ] ( dots ) .
it has a symmetric profile , and the depth is close to @xmath17 . the dna chain with a point inhomogeneity given by a single inverted
dipole has an effective potential with the well profile shown in fig .
[ fig7 ] ( open circles ) .
it is also a symmetric profile , and the depth is close to @xmath88 .
+ the set of dynamical equations derived from the hamiltonian ( [ ham ] ) can be transformed , in the continuum limit , into the corresponding continuous partial differential equation . within this context
the collective coordinate method provides a good way to analyze the influence of a perturbation on a soliton ( in our case we can consider the long range interaction term as a perturbation ) .
the idea is the same as using the center of mass to analyze the behavior of a system of particles .
the method introduces the center of the breather `` x '' as a description variable , and the on - site potential affects as a potential energy of the form @xmath89 if the breather is well localized , as occurs in our cases , @xmath90 . in other words , the translational movement of the breather center `` x '' is affected by a symmetric potential well with the same form as the shown , for each case , in fig .
if the translational kinetic energy of the breather is small enough , the phonon radiation emitted as it passes through the potential well prevents the transmission of the breather and it is trapped with an oscillating movement like a particle in a potential well .
the remaining terms in eq .
( [ l ] ) also contribute to the center of breather motion but they are not so significant . in fact , in the continuum approximation they lead to a space - dependent effective mass of excitation
. + , versus site number , n , for the interfaced piecewise system ( dots ) and for the single inverted dipole system ( open circles).,scaledwidth=40.0% ] transmission regime : : : + if the translational kinetic energy of a mb is bigger than the critical value , i.e. @xmath76 , the breather crosses the interface and continues its movement along the chain with smaller translational kinetic energy .
in other words , the breather is transmitted as a consequence of the interaction with the interface , and a small part of its energy is transferred to the chain as phonon radiation .
an example of a transmitted mb can be seen in fig .
[ fig4](b ) , and the evolution of the corresponding central energy is shown in fig .
[ fig5](b ) .
the same qualitative results have been found in the case of the parallel - oriented dipole chain with a single inverted dipole . in this case , the trapping site is centered at the inverted dipole .
both this system and the interfaced piecewise system have a symmetric effective on site potential well due to the dipole coupling . also , both of them have two linear localized modes . the nonexistence of an effective on site potential barrier makes impossible the appearance of a reflection regime . as mentioned before
, the critical value @xmath91 depends on the values of the parameters @xmath17 , @xmath15 and the breather frequency @xmath92 .
we have numerically determined the dependence of @xmath91 with respect to @xmath17 for some different values of @xmath15 and the fixed value @xmath83 .
the results are shown in fig .
[ fig8 ] , where it can be observed that , for a fixed value of @xmath15 , @xmath91 increases monotonously with @xmath17 , and , for a fixed value of @xmath17 , @xmath91 decreases monotonously with @xmath15 . with respect to @xmath17 for some different values of @xmath15.,scaledwidth=40.0% ] equation ( [ eq : rtopbottom ]
) shows that the extent of localization of the linear localized modes depends on the values of the parameters @xmath17 and @xmath15 .
in fact , for the interfaced system the decay factor for the bottom mode is given approximately by @xmath93 then , the extent of localization depends basically on the quotient @xmath94 .
we have observed that varying @xmath17 and @xmath15 with constant @xmath95 , the values of @xmath91 are practically coincident .
also , when @xmath95 increases , or @xmath96 decreases , @xmath91 increases .
the ib solution is even more localized than the bottom mode because @xmath97 .
the values of the critical translational kinetic energies that determine the transitions from the trapping to the transmission regimes are related to the extent of localization of the ib solution .
an important result is that for a given system the `` trapping window '' , defined as the range of translational kinetic energy for which trapping occurs , is related to the quotient @xmath95 .
higher localization hinders the capability of breathers for being transmitted to the chain .
a similar conclusion can be obtained for the single inverted dipole chain . in this case
the values of @xmath91 are higher than the corresponding to the interfaced chain .
for example , with @xmath77 and @xmath98 we obtain @xmath99 . with the same value of @xmath63 and with @xmath100 , @xmath101 , although this value is out of the interval of @xmath65 values that we have considered in this paper .
the problem of the interaction of mbs with an impurity in a homogeneous klein - gordon chain was considered in @xcite . in that case
mbs can be reflected , trapped or transmitted by the impurity , and a necessary condition for the appearance of trapping is the existence of an impurity breather solution , that is , a static breather solution centered at the impurity .
similarly , we have seen that for our systems a necessary condition for trapping is the existence of an _
inhomogeneity breather solution _ ( ib ) with the frequency of the mb and an internal energy smaller than the internal energy of the mb . fig .
[ fig9 ] shows that , for @xmath98 and @xmath83 , the difference between the inhomogeneity breather energy ( @xmath102 ) and the internal energy of the moving breather ( @xmath103 ) is negative .
this confirms the existence of inhomogeneity breathers and that its energy is smaller than the moving breather energy , which is the necessary condition for trapping .
it can also be observed in this figure that the magnitude the difference between the inhomogeneity breather energy ( @xmath102 ) and the internal energy of the moving breather ( @xmath103 ) increases monotonously with @xmath17 .
thus , the phonon radiation emitted by a trapped breather increases with @xmath17 . ) and the internal energy of the moving breather ( @xmath103 ) with respect to @xmath17 , for @xmath98 and @xmath83 . ( a ) the interface case , ( b ) the single inverted
dipole case , scaledwidth=70.0% ]
the study of the trapping of several breathers at the interface region is of great interest . when a mb is trapped , other incoming breathers
can also be trapped after colliding with the interface .
this process could increase the energy density locally , thus collecting or accumulating energy in the interface region .
another point of interest is whether there is a saturation threshold for this accumulated energy . in general
, the study of energy exchange in collisions of mbs is a difficult problem .
some results are known for a special type of fpu lattice @xcite , and in a one dimensional dna model @xcite , but at present no general results are known concerning to collisions of mbs in klein gordon lattices .
the simulations show that when a mb collides with a trapped breather , the effects of the collision are strongly dependent on their exact dynamical details when the effective collision processes begin .
we have performed a large number of simulations in a chain of n=200 base pairs , with the coupling parameters @xmath77 and @xmath98 .
all the breathers have the frequency @xmath83 .
we have found that under favorable conditions the accumulation of energy is possible as a consequence of the trapping of successive breathers .
we show in fig .
[ fig10](a ) contour plots for the evolution of the energy density of five breathers launched towards the interface at different instant of time .
all the translational kinetic energies are different . ) .
( b ) time evolution of the central energy.,scaledwidth=75.0% ] ) .
after the last collision the accumulated energy is transmitted to the right side of the chain .
( b ) time evolution of the central energy.,scaledwidth=75.0% ] the first breather is launched towards the interface at @xmath104 with a translational kinetic energy smaller than the critical value , @xmath105 , so that the breather is trapped at the discontinuity region .
the phonon radiation must be removed because it distorts the numerical simulations when it is reflected at the boundaries . for that purpose , just before each breather is launched , we set to zero the displacements and velocities outside the interface region , @xmath106 . the second breather is launched with @xmath107 , which is greater than @xmath91 , but it is also trapped at the interface . we have observed that the results of the collision between these two breathers depends strongly on their relative phase when the collision processes begin .
we have chosen the translational kinetic energy of this second breather in order to obtain an efficient trapping . using the same procedure as before ,
three consecutive breathers are launched with @xmath108 , respectively .
all of them are trapped , and to illustrate the evolution of the total collected energy at the interface , we have represented in fig .
[ fig10](b ) the central energy versus time .
we observe that after each collision the total trapped energy increases , but only part of the energy of each incident breather is accumulated at the interface region .
the numerical simulations show that the breather is partly transmitted and partly reflected , and also some little phonon radiation appears .
it is interesting to observe that the phonon radiation increases as the number of trapped breathers or the energy accumulated increases .
the possibility for trapping more energy at the inhomogeneity region by additional incoming breathers seems to have a limit due to a saturation effect
. the accumulation of energy at the interface region could start local base pair openings in the dna molecule .
nevertheless , the simulations show that the accumulation of energy can , occasionally , disappears after an appropriate collision .
an example of this is shown in fig .
[ fig11 ] , where the last collision induces the transmission of the accumulated energy to the right side of the chain . in this simulation
five breathers are launched successively towards the interface with @xmath109 , @xmath107 , @xmath110 , @xmath111 and @xmath112 , respectively . to gain deeper insight into the possible reasons for these striking features
, the authors are currently investigating this issue and will be object of a future work .
we have studied the evolution , and some collision processes , of mbs in two different dna molecules .
the first dna chain consists of a sequence of identical base pairs , where there exists an interface such that the dipole moments at each side are oriented in opposite directions .
the second chain consists also of a sequence of identical base pairs but there exists a single dipole moment which is oriented in opposite direction to that of the others ones .
i both cases there exists a local inhomogeneity defined by the interface or the single inverted dipole , respectively .
the hamiltonian of the peyrard bishop model is augmented with an energy term that takes into account the long range forces due to the dipole dipole coupling between the base pairs .
we have generated mbs with different translational kinetic energies and studied the effects of the collisions with the interface or with the single inverted dipole , respectively . in each case , the outcome depends on the values of the stacking coupling constant , the dipole dipole coupling constant , the frequency and the translational kinetic energy of the mbs .
taking fixed values for the first three quantities , we have found that if the kinetic energy of the mb is smaller than a critical value , the breather is trapped after colliding with the interface , whereas if the kinetic energy is larger , the breather is transmitted to the other side of the interface loosing energy .
therefore , there exist two different dynamical regime : a ) the trapping regime , when mbs are trapped at the interface ; b ) the transmission regime , when mbs are transmitted to the other side of the interface .
the same effects appear when mbs collide with a single inverted dipole in the chain . in both cases ,
a necessary condition for the trapping of a mb is the existence of a stationary breather solution with the same frequency , centered at the inhomogeneity site , which we call inhomogeneity breather ( ib ) , and with a smaller energy than the internal energy of the mb .
when collision occurs part of the energy is transferred to the chain as phonon radiation .
the existence , for each case , of a symmetric effective on site potential well due to the dipole - dipole coupling , explains qualitatively the found behavior .
the value of the critical translational kinetic energy that determines the transition from the trapping regime to the transmission regime is directly related to the extent of the localization of the ib .
this extent of localization is related to the quotient between the long range and the short range coupling parameters . trapping or accumulation of energy
is facilitated when this quotient increases .
it is important to state that the most simple structural modification in a homogeneous dna molecule , which is to reverse the orientation of a single dipole , provides a mechanism for the trapping of moving breathers .
the simulations show that some incoming mbs can also be trapped after colliding with a trapped breather , producing an accumulation of energy at the inhomogeneity region .
the interface or the inverted dipole can play the role of chargers of vibrational energy .
occasionally , a successful incoming breather can collide with this excited region and all the accumulated energy can be carried away to the rest of the dna chain . at this point , if these processes occur really in nature , one is tempted to imagine that succsesive pulses of accumulated energy can arrive to a coding sequences producing the initiation of the transcription processes .
the design of experiments in real or synthetic dna are encouraged in order to verify the results presented in this paper .
we would like to thank professor yu b gaididei and professor jm romero for helpful discussions . | we study the dynamics of moving discrete breathers in an interfaced piecewise dna molecule .
this is a dna chain in which all the base pairs are identical and there exists an interface such that the base pairs dipole moments at each side are oriented in opposite directions .
the hamiltonian of the peyrard bishop model is augmented with a term that includes the dipole dipole coupling between base pairs .
numerical simulations show the existence of two dynamical regimes .
if the translational kinetic energy of a moving breather launched towards the interface is below a critical value , it is trapped in a region around the interface collecting vibrational energy . for an energy larger than the critical value , the breather
is transmitted and continues travelling along the double strand with lower velocity .
reflection phenomena never occur .
the same study has been carried out when a single dipole is oriented in opposite direction to the other ones .
when moving breathers collide with the single inverted dipole , the same effects appear .
these results emphasize the importance of this simple type of local inhomogeneity as it creates a mechanism for the trapping of energy .
finally , the simulations show that , under favorable conditions , several launched moving breathers can be trapped successively at the interface region producing an accumulation of vibrational energy .
moreover , an additional colliding moving breather can produce a saturation of energy and a moving breather with all the accumulated energy is transmitted to the chain . |
anderson localization@xcite as well as strong correlation effects due to interactions@xcite have been studied intensely for decades in solid state physics .
it is well established that both phenomena modify the transport properties of solids in a crucial way .
in particular , disorder as well as correlations are capable of inducing metal - insulator transitions known as the anderson transition and the mott - hubbard transition , respectively .
although each fundamental phenomenon - correlations and disorder - on its own gives rise to challenges to both experimental and theoretical research and is still a subject of current investigations , the realistic modeling of materials requires the simultaneous consideration of both effects .
this interplay is of central interest within present day s solid state physics.@xcite theoretical investigations , e.g. of the anderson - hubbard model combining the hubbard model and the anderson model are notoriously difficult .
the most intriguing effects , such as the mott transition and the anderson transition , take place at intermediate interaction and disorder strength requiring a non - perturbative approach .
extensions of the dynamical mean - field theory ( dmft)@xcite to disordered systems are especially promising in this respect , since local correlations are incorporated non - perturbatively .
such extensions were performed on a level of the coherent potential approximation@xcite and on a level of typical medium theory ( tmt ) incorporating localization effects.@xcite a shortcoming of both approaches is given by the non - consideration of disorder - induced fluctuations of local observables .
dobrosavljevi and kotliar extended the dmft to include disorder fluctuations in space on a fully stochastic level,@xcite which is referred to as statistical dmft .
@xcite this theory is capable of describing mott and anderson - mott transitions non - perturbatively and allows for the self - consistent determination of the probability distribution function ( pdf ) of local observables . among the investigations of the box disordered anderson - hubbard model in @xmath0 dimensions which are non - perturbative and incorporate localization effects
the tmt was used in the most comprehensive way.@xcite within these works the ground state phase diagram has been determined for a half - filled band @xcite and the physical two fluid picture of the anderson - mott transition has been developed.@xcite the statistical dmft has also been used to study the box disordered anderson - hubbard model.@xcite the anderson - mott transition has been characterized in detail by an accurate treatment of the low energy excitations.@xcite remarkably , two critical disorder strengths have been found , one corresponding to the entering of a non - fermi liquid phase and a second one corresponding to the anderson - mott transition .
a special focus was laid on the zero - bias anomaly @xcite in two dimensions.@xcite however , partially due to the immense numerical effort and partially due to the restrictions of the used impurity solvers the slave - boson approach@xcite and the hubbard - i approximation@xcite the ground state phase diagram was not determined . in this work
we revisit the anderson - hubbard model with box disorder by means of the statistical dmft . by using the modified perturbation theory ( mpt)@xcite as an impurity solver we determine the pdf of the local density of states ( ldos ) for finite interaction strengths which was not done before .
our mpt investigation shows that the pdf of the ldos is not well fitted by a log - normal distribution function within the metallic phase in contrast to the non - interacting case.@xcite finally , the paramagnetic ground state phase diagram consisting of disordered correlated metal , mott insulator , and anderson - mott insulator is determined and critically discussed with regard to established results obtained by means of tmt earlier.@xcite
a tight - binding model that describes strong correlations and randomness is given by the anderson - hubbard hamiltonian @xmath1 where @xmath2 ( @xmath3 ) denotes the fermionic creation ( annihilation ) operator at a lattice site @xmath4 with spin @xmath5 .
the number operator is given by @xmath6 .
the hopping amplitude @xmath7 for nearest neighbor ( n.n . )
sites @xmath4 and @xmath8 only and zero otherwise , the local interaction strength is denoted by @xmath9 , and @xmath10 represents the chemical potential . in the following , we consider electrons on the bethe lattice),@xcite which is characterized by the connectivity @xmath11
it is related to the lattice coordination number @xmath12 via @xmath13 . within the tight - binding model the disorder manifests itself by random on - site energies @xmath14 , which are drawn from a pdf @xmath15 .
here , we consider the box distribution @xmath16 which is commonly used model in theoretical investigations .
therein , the disorder strength is denoted by @xmath17 and we assume that the on - site energies of all lattice sites are independently and identically distributed .
the statistical dmft@xcite is an extension of the well - known dmft to cope with disordered systems , which enables the self - consistent determination of the pdf of the local single - particle green s functions @xmath18 $ ] . here , @xmath19 is the fourier transform of the retarded green s function @xmath20_+ \rangle$ ] , where @xmath21_+$ ] represents the anticommutator bracket . in the discussion that follows ,
the spin index @xmath22 is omitted , since we consider paramagnetic solutions of the anderson - hubbard model . in the non - interacting system
, the renormalized perturbation theory@xcite shows that the local green s function is given by @xmath23 where the hybridization function @xmath24 describes the coupling of site @xmath4 to the nearest neighbor lattice sites .
the introduction of a finite broadening factor @xmath25 is essential for the numerical distinction between localized and extended states as discussed in detail later .
the hybridization function can be expressed by an infinite , renormalized series @xmath26 where @xmath27 is the diagonal cavity green s function of the system when the sites @xmath28 are removed . the corresponding diagonal cavity green s functions are determined by similar series on cavity lattices leading to a hierarchy of equations . on the bethe lattice with connectivity @xmath11 this series is exactly truncated after the first term and the hybridization functions are exactly given by@xcite @xmath29 additionally , the cavity hybridization functions of lattices with several sites removed are also simplified due to the absence of loops on the bethe lattice .
the corresponding equations reproduce their structure from the second equation and therefore there are only two classes of hybridization functions ( and corresponding green s functions ) determined by the equations @xmath30 here , both sums extend over the nearest neighbors and the second one is defined on the cavity lattice .
next , we employ the further approximation that was introduced and successfully used by abou - chacra _ _ et al.__@xcite therein , the structural differences in the equations ( [ ld_bethe_hyb_eq1 ] ) and ( [ ld_bethe_hyb_eq2 ] ) are neglected and replaced by a single equation @xmath31 this approximation is better the higher the connectivity is , since the difference between @xmath12 and @xmath11 becomes smaller .
furthermore , we note that all equations in the hierarchy incorporate the sum over @xmath11 diagonal green s functions except the first.@xcite thus , @xmath11 can be considered as a typical number of summands in the equations of the hierarchy.@xcite following dobrosavljevi and kotliar@xcite for finite interactions , an additional local self - energy @xmath32 , taking into account the interaction effects , is incorporated via @xmath33 $ ] , where @xmath34 corresponds to the green s function for the same disorder realization in the non - interacting case .
the local single - particle green s function is consequently given by @xmath35 the approximation of a local self - energy is exact for infinite connectivity , as was shown by metzner and vollhardt in a seminal paper@xcite and is known to be a good approximation in three spatial dimensions .
since the translational invariance is broken in a disordered system , the local single - particle green s function becomes site - dependent , giving rise to a pdf @xmath36 $ ] . within statistical dmft
, this pdf is determined by an ensemble with a large number @xmath37 of green s functions .
each of these green s functions corresponds to an impurity problem defined by a site - dependent hybridization function @xmath38 .
the solution of each impurity problem results in local self - energies @xmath39.@xcite hence , the statistical dmft maps the original lattice model onto an ensemble of impurities , whose pdf has to be determined self - consistently .
the statistical dmft naturally adopts anderson s point of view@xcite that an investigation of disordered systems should focus on the pdfs of local observables . the self - consistent calculation scheme is implemented by the following algorithm : the starting point is an initial pdf @xmath36 $ ] and the calculation is performed by ( i )
drawing a random on - site energy @xmath14 from the pdf @xmath40 given in eq .
( [ pd_onsite_e ] ) for each sample of the ensemble ; ( ii ) the hybridization function @xmath24 , with the local single - particle green s function @xmath41 of the nearest neighbors randomly sampled from the pdf @xmath36 $ ] , is determined via eq .
( [ eq_hybrid ] ) for each sample ; ( iii ) the local self - energy @xmath39 is calculated from the solution of the local impurity problem by using an impurity solver ; ( iv ) the local single - particle green s function @xmath42 is calculated using eq .
( [ eq_m4 ] ) ; ( v ) having calculated a completely new ensemble @xmath43 , a new pdf @xmath36 $ ] is obtained and we return to step ( i ) . the steps ( i)-(v )
are to be repeated until self - consistency for the pdf @xmath44 $ ] is achieved .
we note that this method incorporates spatial fluctuations , i.e. quantum interference effects via ( ii ) .
statistical fluctuations can be minimized by an artificial increase of the size of the ensemble after having reached self - consistency . in this work ,
typically @xmath45 successive dmft iterations are used .
one physical observable that is capable of describing the mott transition and the anderson - mott transition on equal footing is given by the ldos.@xcite it is defined by @xmath46 and is a random quantity in disordered systems .
the corresponding pdf @xmath47 $ ] is obtained by counting all values of the ldos for each frequency and by constructing a histogram . from the calculated pdf
we can extract all moments of the pdf of the ldos @xmath48 \rho'^k \label{eq_p_moments}.\ ] ] in particular , the first moment ( @xmath49 ) is the arithmetic average of the ldos @xmath50 \ , \rho'(\omega ) \label{eq_arith_av}\;,\ ] ] where the dependence on @xmath51 is to be understood parametrically .
the arithmetic average of the ldos corresponds to the density of states ( dos ) of the system .
furthermore , the typical value , i.e. the most probable value , can be extracted . in tmt ,
the typical value is approximated by the geometric disorder average average@xcite @xmath52 since statistical dmft gives access to the full pdf of the ldos , we will consider the geometric average of the pdf @xmath53 \
, \ln \rho'(\omega ) \label{eq_geo_av}\ ] ] only for comparison purposes . in the following
, we will also use the cumulative pdfs @xmath54 = \int\limits_0^{\rho(\omega ) } d\rho ' ( \omega)\ ; p[\rho'(\omega ) ] \;.
\label{eq_cumulative}\ ] ] to the end of this section we shortly introduce the impurity solver that is used to solve the impurity part of the dmft equations .
since we are dealing with a large number of single sites , we have to be restricted to impurity solvers with a feasible amount of computational time . in this work
we employ the mpt,@xcite which properly reproduces the non - interacting limit and the atomic limit .
mpt is the generalization of the iterated perturbation theory ( ipt)@xcite to systems with arbitrary fillings .
ipt qualitatively describes the mott - hubbard metal - insulator transition at a finite critical interaction strength @xmath55@xcite by calculating the self - energy to the second order in @xmath9 in the non - renormalized perturbation expansion .
using this impurity solver , green s function ensemble sizes of the order @xmath56 are computationally feasible within a parallelized code run on a cluster . within mpt
the local self - energy is given by@xcite @xmath57 where @xmath58 is the second order perturbation contribution to the self - energy.@xcite the perturbation expansion uses the non - renormalized hartree - fock green s functions@xcite @xmath59 as propagators .
the parameter @xmath60 is fixed by requiring @xmath61.@xcite the parameters @xmath62 and @xmath63 are determined by the requirement that the spectral moments @xmath64 reproduce an equation of motion analysis up to @xmath65 and that the atomic limit is recovered correctly.@xcite the corresponding self - consistency equations that accomplished the requirement were worked out by m. potthoff _ _
in our calculations , we consider the half - filled case , i.e. band filling @xmath66 , which is accomplished by setting @xmath67 for box disorder .
we work in energy units of the non - interacting bandwidth @xmath68 . since we used the approximate equation ( 10 ) , the dos of non - disorder system is always semielliptic for each @xmath11 . in order to investigate the physics of strongly correlated electrons in disordered lattices we need to define all relevant phases .
the correlation - induced _
mott insulator _ is incompressible and exhibits an energy gap in the spectrum , which is proportional to the interaction strength @xmath9 .
a gapped spectrum in turn corresponds to a vanishing arithmetic average of the ldos at the fermi level , i.e. @xmath69 .
an insulating phase that originates from the randomness is the well - known _ _
anderson insulator__@xcite for zero interaction .
this phase is compressible and its spectrum is dense point - like , which can be attributed to an absence of diffusion.@xcite in the interacting case , we define a state to be localized or extended , if the spectrum of the single - particle green s function is a dense point - like or continuous respectively , which is a natural generalization of the localization concept to interacting systems . consequently , the _ anderson - mott insulator _ is defined by localized single - particle excitations at the fermi level .
so far it is not established on a mathematical level of rigor , that localized single - particle excitations correspond to a vanishing dc conductivity , but strong evidence was recently given in case of weak interactions.@xcite how we numerically detect localized states within the statistical dmft will be explained in the following subsection [ sec_localization ] . finally , the _ paramagnetic metal _ is compressible and therefore has a non - vanishing dos @xmath70 at the fermi level and the corresponding states are extended .
of one random sample out of the ensemble for two different box disorder strengths @xmath17 : ( a ) @xmath71 and ( b ) @xmath72 . parameters
are @xmath73 , @xmath74 , @xmath75 , and @xmath76.,scaledwidth=48.0% ] localized states are defined by the absence of diffusion:@xcite the state of a propagating particle initially localized at one single lattice site exhibits a finite return probability in the thermodynamic limit .
it was shown that the finite return probability implies a dense distribution of poles on the real axis of the single - particle green s function in contrast to extended states , which are given by a branch cut on the real axis of the single - particle green s function.@xcite this difference in the analytic properties of the single - particle green s function can be utilized for a numerical distinction between localized and extended states within the statistical dmft and its non - interacting limit , i.e. the local distribution approach.@xcite extended and localized states are characterized by different behavior of the pdf @xmath77 $ ] in the limit of vanishing broadening @xmath78,@xcite as is explained subsequently in more detail . in a finite lattice
the local one - particle green s function at site @xmath79 is given by @xcite @xmath80 here , @xmath81 denotes the overlap of the eigenstate @xmath82 with the wannier function on site @xmath79 , @xmath83 denotes the complete set of single particle energy eigenstates with eigenenergies @xmath84 .
it was shown that the return probability is determined by @xmath85.@xcite in case of an extended eigenstate , the residue @xmath86 is proportional to the inverse number of occupied lattice sites @xmath87 .
consequently , the return probability approaches zero in the thermodynamic limit . in case of an exponentially localized state ,
the residues are finite even in the thermodynamic limit .
if these values were ordered , they would decrease exponentially and only some of them would contribute significantly to the sum ( [ eq_ef_representation_gf ] ) .
in particular , if we introduce a small coupling @xmath88 to a dissipative bath , which regularizes the singularities , and consider a contour which encloses a small energy interval @xmath89 , then the most probable value of the sum of the residues of poles enclosed by this contour will decrease exponentially proportional to the ratio @xmath90 .
from here it follows , that the most probable value of the imaginary part of the green s function is proportional to @xmath88.@xcite hence , the ldos will be highly fragmented in case of localized eigenstates , characterized by dominating well - separated resonances . in contrast , the arithmetically averaged spectral function , i.e. the dos of the system , does not exhibit this high fragmentation , since the spectral weight must be located somewhere in the lattice for every energy .
this theoretical reasoning can be verified within our numerical investigation .
as explained in the method section we determine the pdf of the local green s function by calculating the local green s function for each sample of the ensemble .
each of them is given by a random on - site energy , a local self - energy and a hybridization function which is randomly sampled by @xmath11 local green s functions out of the ensemble according to equation [ eq_hybrid ] .
having calculated all local green s functions , we can investigate the local spectrum of one single sample out of the ensemble .
two example spectra are displayed in fig .
[ fig_spectra_extended_localised ] .
panel ( a ) shows a typical spectrum for small disorder strengths , namely @xmath71 .
the spectrum is smooth and furthermore broadened in comparison to @xmath91 due to the disorder .
in contrast , the spectrum at higher disorder strength @xmath92 , shown in panel ( b ) , is highly fragmented and consists of various delta peaks , which are broadened by the artificial value @xmath88 .
this can be seen more clearly in the inset .
in fact , this is exactly what we expect : extended states of the system are characterized by a branch cut on the real axis of the local green s function , whereas localized states are given by poles . .
parameters are @xmath73 , @xmath74 , @xmath75 , and @xmath76.,scaledwidth=48.0% ] so far , the spectrum of a single sample was considered , but it is also interesting to study the distribution of spectra of the whole ensemble .
the disorder causes a broadening of the spectrum as can be clearly seen from fig .
[ fig_spectra_no_ww ] , where the geometrically averaged ( [ eq_geo_av ] ) and arithmetically averaged spectral functions ( [ eq_arith_av ] ) are plotted for selected values of the disorder strength and zero interaction .
the broadening is naturally accompanied by a drop - off of the geometric average as well as the arithmetic average .
this is because the spectral weight is distributed over a bigger range of energies with increasing disorder strength . as an important feature we note the much stronger decline of the geometric average to zero , as found
previously within tmt.@xcite since the arithmetically and geometrically averaged spectral function are plotted as a function of the energy , we are also able to note that the decline is not uniform , but rather strongly depends on the energy . for instance , the geometric average of the ldos remains clearly finite in the band center for @xmath93 , whereas it is close to zero for the states in the outer parts of the spectrum .
$ ] for increasing disorder strength @xmath17 .
the dashed lines correspond to least square fits with a log - normal function .
parameters are @xmath75 , @xmath94 and connectivity @xmath76.,scaledwidth=48.0% ] next we discuss how the overall pdf of the ldos is affected by disorder . in fig .
[ fig_histo_development_with_d ] the pdf of the ldos is displayed for several disorder strengths .
the initial delta function for the homogeneous system located at @xmath95 ( not displayed ) is broadened by a small amount of disorder ( @xmath96 ) due to the disorder - induced fluctuations . by increasing the disorder strength further to @xmath97 and @xmath98 the pdf spreads more and more , long tails develop and more weight is shifted to smaller values .
this becomes evident , if we focus on the most probable value of the pdf , also called the _ typical value _ , which approaches zero with increasing disorder strength .
the dashed lines in fig .
[ fig_histo_development_with_d ] correspond to least square fits with theoretically predicted log - normal distributions@xcite @xmath99 with fitting parameters @xmath10 and @xmath22 . obviously , the approximation becomes better for stronger disorder , in agreement with recent high accuracy kernel polynomial method calculations for various bravais lattices in the non - interacting system.@xcite $ ] for decreasing broadening @xmath88 at two different disorder strengths @xmath17 : ( a ) @xmath71 and ( b ) @xmath72
. parameters are @xmath76 , @xmath74 , and @xmath75.,scaledwidth=48.0% ] now we discuss how the broadening @xmath88 is used for a distinction between localized and extended states . in the localized phase the green s function
is given by a distribution of poles .
hence , an arbitrarily chosen frequency @xmath51 lies with probability one between the poles , resulting in a value of the ldos equal to zero .
if it is exactly a @xmath100-peak , the resulting value would be infinity .
however , the probability for this event is zero .
the artificially introduced coupling to a dissipative bath via the small factor @xmath88 broadens the @xmath100-peaks to lorentzians with a width proportional to @xmath88 and generates a finite probability to obtain a finite value of the ldos .
moreover , let us consider an energy in the vicinity of a pole present in the considered sample .
the corresponding value of the ldos is proportional to @xmath88 , since the ldos is given by the imaginary part of the single - particle green s function .
now the observed features of the pdf of the ldos in fig .
[ fig_localized_extended ] can be understood : in a system where the local spectra are highly fragmented due to the disorder , most ensemble samples will exhibit a nearly zero ( @xmath101 ) value of the ldos for a given frequency @xmath51 . on the other hand , a small fraction will contribute high values of the ldos stemming from the broadened @xmath100-peaks , constituting the long tails of the pdf .
consequently , the maximum of the pdf of the ldos of an extended state saturates at a finite value for @xmath78 , but an increasing amount of the pdf s weight shifts to zero in the case of localized states .
$ ] for fixed disorder strength @xmath102 and different values of @xmath88 : ( a ) @xmath73 , ( b ) @xmath103 , ( c ) @xmath104 and ( d ) @xmath94 .
parameters are @xmath76 , @xmath74 , and @xmath75.,scaledwidth=48.0% ] this allows for a numerical distinction between localized and extended states .
clearly , the pdf of the ldos of an extended state will not be affected upon lowering of @xmath88 below a certain value .
in contrast , in the localized phase , the pdf will be affected dramatically : with decreasing broadening @xmath88 an increasing amount of the pdf s weight is shifted to smaller values , and the most probable value will shift to zero .
[ fig_localized_extended ] displays the evolution of the pdf of the ldos at frequency @xmath105 on a log - log scale when the broadening is decreased . in panel
( a ) the disorder strength @xmath17 is equal to one and the pdf is found to saturate .
we conclude that the state is extended . for @xmath72 in panel ( b )
the pdf exhibits the above described behavior .
we conclude that the state is localized .
it should be shortly mentioned here that the pdfs in the localized phase are not well described by log - normal functions .
in fact , the shape of the pdfs results from a collection of narrow lorentzians .
the width of these lorentzians is determined by the broadening @xmath88 and so is the shape of the pdf .
therefore , the log - normal distributions can only be used as suitable approximations for extended states due to the finite broadening . when the analysis is performed with a complete frequency resolution
, the mobility edges of the system can be identified .
for example this is done in fig .
[ fig_mobility_edge ] , where the natural logarithm of the pdf @xmath77 $ ] is shown for fixed disorder strength @xmath102 in the @xmath51-@xmath106-plane .
the broadening factor @xmath88 is lowered from @xmath107 in panel ( a ) to @xmath108 in panel ( d ) .
when the broadening is decreased , we notice clearly a low - energy core of the band , where the pdfs are unaffected upon further lowering of the broadening . on the other hand , outside of this core the pdf shows the behavior for localized states .
the separating energies represent the mobility edges we are looking for .
it is important to note that the resolution of this procedure is determined by the lowest possible value of @xmath88 , which is basically determined by the finite bath size . in any case
, the procedure will give the lower bound of the critical disorder strength for the transition from extended to localized states , since an extended state will never be mistaken as a localized state .
this question of resolution was examined in more detail in ref . .
when the mobility edges cross the fermi level in the band center all states become localized , corresponding to the anderson insulator .
since localized states do not contribute to the dc conductivity , a metal - insulator transition is obtained , caused by the disorder .
the existence of this transition on the bethe lattice was analytically shown @xcite and recently checked numerically.@xcite $ ] of the non - interacting system with decreasing @xmath88 in ( a ) for @xmath109 and @xmath96 , ( b ) for @xmath110 and @xmath102 , and ( c ) for @xmath76 and @xmath102 .
parameters are @xmath74 and @xmath75.,scaledwidth=48.0% ] compared to cpa and tmt , the connectivity @xmath11 represents an additional free parameter within the statistical dmft . it is well established that no localization occurs in the limit of infinite spatial dimensions , or for infinite connectivity , respectively . on the other hand ,
it is known from scaling theory that in one dimension and in two dimensions
i.e. for small connectivities already arbitrarily small disorder suffices to localize all states in the absence of interactions.@xcite hence , this strong dependence of localization on the connectivity should be captured by statistical dmft .
we expect that for higher connectivity the tendency towards delocalization is enhanced .
let us , for instance , consider the evolution of the cumulative pdf of the ldos at the fermi level defined in equation ( [ eq_cumulative ] ) when the broadening @xmath88 approaches zero for several parameter sets , as given in fig .
[ fig_psum_vgl_k ] . in panel
( a ) we show a small disorder strength @xmath96 for connectivity @xmath109 , i.e. a linear chain .
we note that the @xmath105 state is localized . for @xmath110 (
not displayed ) our analysis shows that the state is extended and higher disorder strengths are needed to localize the state at the fermi level , as can be seen for example for @xmath102 in panel ( b ) .
finally , in panel ( c ) we investigate the connectivity @xmath76 for the same disorder strength . in the former case
the state at @xmath105 is localized , whereas in the latter case with a larger connectivity the state is extended . in this subsection
we will address the question of disorder - induced localization in the presence of interactions .
the evolution of the cumulative pdf of the ldos with decreasing broadening is plotted in fig . [ fig_loc_with_u ] for the weakly interacting case @xmath111 . in panel
( a ) the disorder strength is given by @xmath102 and in panel ( b ) @xmath112 , corresponding to the strongly disordered regime .
we note that the pdf saturates for @xmath102 when the broadening approaches zero , corresponding to an extended state .
in contrast , for very strong disorder strength @xmath112 the pdf does not saturate . here , the state is localized within the resolution given by the size of our stochastic green s function bath . by comparison to fig .
[ fig_psum_vgl_k ] , we note that the shape of the pdfs of the interacting system differs from those in the non - interacting case . in particular
, the slope of the cumulative pdfs varies with the broadening , which indicates that significant weight of the pdf is distributed over several orders of magnitude of the ldos .
$ ] with decreasing @xmath88 ( a ) for @xmath102 and ( b ) for @xmath112 at fixed interaction strength @xmath111 .
parameters are @xmath113 , @xmath75 , and @xmath67.,scaledwidth=48.0% ] the statistical dmft allows for studying how the pdf of the ldos evolves from the non - interacting case and weakly interacting case to the regime of strong correlations . in fig .
[ fig_pdf_vary_u ] the pdfs are plotted for several selected values of the interaction and fixed disorder strength @xmath102 and connectivity @xmath113 . in panel
( a ) weak interactions ( @xmath111 ) are introduced and stepwise increased to the intermediate regime ( @xmath114 ) . in comparison to the non - interacting case ( corresponding to a localized state as we will see clearly in the next section ) , we note a striking redistribution already for weak interactions . the pdf is still extended over many decades of values of the ldos , but a big fraction of the pdf is shifted to much higher values of the ldos compared to the non - interacting case
. we will understand this severe modification later .
further increase of the interaction strength to @xmath115 systematically fortifies this redistribution .
more and more sites exhibit comparable values of the ldos , as can be seen from the larger slope of the cumulative pdf .
this behavior contradicts a strong fragmentation of the single - particle excitation spectrum .
moreover , we note that the cumulative pdf gains most of its weight in two regions of a big slope for @xmath115 .
this feature erodes for bigger interaction strengths as can be seen from the cumulative pdf for @xmath114 . in panel
( b ) the interaction strength is stepwise increased further to @xmath116 .
in contrast to the behavior found in panel ( a ) it can be clearly seen that the opposite behavior is obtained in that case .
the pdfs shifts again to smaller and smaller values of the ldos .
obviously $ ] with increasing interaction @xmath9 for fixed disorder strength @xmath102 . in panel
( a ) the pdf of the non - interacting system is compared to the pdfs of the interacting system from weak interactions ( @xmath111 ) to intermediate interactions ( @xmath117 ) . in panel ( b ) the interaction strength @xmath9 is further increased stepwise to the strongly interacting regime ( @xmath118 ) .
parameters are @xmath113 , @xmath104 , @xmath75 , and @xmath67.,scaledwidth=48.0% ] correlations do not monotonically influence the pdf of the ldos , but rather lead to two different regimes .
we expect this behavior to have a strong influence on the conduction properties of the system , and that it should also be mirrored in other observables . to understand this behavior the pdf is compared to least squares fits of log - normal distributions for @xmath93 in fig .
[ fig_ln_fit_finiteu ] . at zero interaction
the log - normal distribution represents a good approximation of the pdf , but for increasing interaction strengths @xmath111 and @xmath119 strong modifications are observed .
the weight of the pdf is shifted to higher values of the ldos and an additional interaction - induced peak is observed .
the latter becomes particularly apparent on a linear scale as shown in the inset . from these results
it is obvious that the log - normal distribution is not sufficient to approximate the pdf of the ldos in the interacting case .
it is interesting to note that the formation of a second peak is reversed upon further increase of the interaction strength as we learned from the discussion of fig .
[ fig_pdf_vary_u ] .
$ ] for fixed disorder strength @xmath93 and increasing interaction : @xmath120 .
the dashed lines correspond to least square fits with a log - normal function .
parameters are @xmath113 , @xmath73 , @xmath75 and @xmath67.,scaledwidth=48.0% ] $ ] for fixed interaction strength @xmath121 with increasing disorder strength : @xmath122 .
the dashed lines correspond to least square fits with a log - normal function .
parameters are @xmath113 , @xmath73 , @xmath75 , and @xmath67.,scaledwidth=48.0% ] next we discuss the evolution of the pdf with increasing disorder and for fixed interaction strength @xmath121 as plotted in fig .
[ fig_ln_fit_finiteu_vgld ] . in the homogeneous system the pdf of the ldos
is given by a delta - function at the value @xmath95 , which corresponds to the value of the non - interacting system due to the luttinger theorem .
finite disorder strength ( @xmath71 ) broadens the pdf similar to the behavior found in the non - interacting case as shown in fig .
[ fig_histo_development_with_d ] . however , we clearly recognize the two - peak structure of the pdf for @xmath93 and @xmath123 .
this structure is also observable for @xmath124 , but a suitable approximation via the log - normal distribution is clearly restored in the strongly disordered regime .
an important feature of the system considered here is given by the mott transition in the homogeneous system . by construction
the statistical dmft reduces to dmft in the homogeneous case .
since box disorder exhibits a bounded range of accessible on - site energies , it is reasonable to expect the mott transition to be also present in the disordered case at least for moderate disorder strengths .
an example of such a transition is given in fig .
[ fig_box_mott_d3 ] , in which the geometrically and arithmetically averaged spectral functions are displayed for fixed disorder strength @xmath123 and increasing interaction .
for @xmath125 we recognize a spectral function corresponding to a disordered metal , as at the same time it was checked that the states at the fermi level are extended .
this means that the system exhibits no excitation gap .
when the interaction strength is increased ( @xmath126 ) , the spectrum becomes more and more dominated by incoherent excitations .
finally , the spectra belonging to @xmath116 and @xmath127 evidently correspond to mott insulating states , since @xmath128 . obviously , a mott transition takes place between @xmath126 and @xmath116 . moreover , it is interesting to note that we find no evidence for a zero - bias anomaly as can be seen from the spectra .
in contrast , recent exact diagonalization investigations of the anderson - hubbard hamiltonian within the hartree - fock approximation by shinaoka and imada@xcite established the existence of a soft hubbard gap .
the physical origin of the zero - bias anomaly was analyzed in more detail by chen _
et al._@xcite pointing out the importance of non - local correlations which are not included in our theory .
please note that in our recent statistical dmft investigation of fermions in speckle disordered optical lattices@xcite we compared results for the bethe lattice as presented here with results for a finite - size square lattice .
we found evidence for the existence of a zero - bias anomaly on the square lattice but not on the bethe lattice . ) for fixed disorder strength @xmath123 .
parameters are @xmath73 , @xmath67 , @xmath75 , and @xmath113.,scaledwidth=48.0% ] it is of fundamental interest how the anderson transition of the non - interacting system is modified in the presence of interactions and how the mott transition of the homogeneous system is affected by disorder .
the nature of the anderson - mott transition has been investigated in detail within tmt@xcite and within statistical dmft.@xcite in the latter , also a griffiths phase was revealed as precursor for the anderson - mott transition .
however , the extent of the metallic phase has not been systematically discussed within statistical dmft so far , but only within tmt , which neglects spatial fluctuations.@xcite we will discuss how the homogeneous system in the mott insulating state is affected by disorder . in fig .
[ fig_box_delocal_fixu ] the evolution of the geometrically and arithmetically averaged spectral function is presented at fixed interaction strength @xmath126 .
( @xmath129 ) for fixed interaction strength @xmath126 .
parameters are @xmath73 , @xmath67 , @xmath75 , and @xmath113.,scaledwidth=48.0% ] in the homogeneous system the corresponding state is mott insulating .
the mott excitation gap persists for small disorder strengths . from the spectrum at @xmath93
we clearly see that the mott gap is reduced by disorder .
this can be understood by a redistribution of states into the gap . at @xmath123
the gap is completely filled .
the geometric average is finite in the core of the band , but zero in the outer parts of the spectrum .
this suggests that a metallic phase is obtained .
this is confirmed by a proper investigation ( which is not shown here ) of the evolution of the pdf of the ldos in the limit @xmath78 to rule out localized states at the fermi level .
hence , the increase of the disorder strength causes a delocalization of the local magnetic moments of the mott insulator and the system undergoes a transition from the mott insulator to a disordered metal .
this result is in agreement with findings obtained by tmt.@xcite moreover , we note that the kondo resonance corresponding to coherent quasi - particle excitations is present in the disordered metal .
the question how the correlation - induced metal - insulator transition is affected by randomness arises naturally . in previous tmt studies@xcite
it was observed that the luttinger theorem , i.e. the pinning of the ldos at its non - interacting value@xcite , is not fulfilled in the presence of disorder .
moreover , it was shown that the metallicity , given by the dos at the fermi level , grows with increasing interaction strength so that the luttinger theorem is nearly fulfilled again.@xcite sufficiently strong interactions hinder a decay of the quasi - particle excitations .
further increase of the interaction strength leads to a rather abrupt transition to the mott insulating phase . within statistical dmft
we observe a similar behavior as given in the upper panel of fig .
[ fig_box_metallicity ] , where the arithmetic average of the ldos at the fermi level is plotted as a function of the interaction for three different disorder strengths . for three different values of the disorder strength @xmath17 : @xmath96 ( solid squares ) , @xmath93 ( open circles ) and @xmath123 ( solid circles ) in the upper panel . in the lower panel
the arithmetically averaged ldos is plotted as a function of the disorder strength for three different values of the interaction strength : @xmath121 ( solid squares ) , @xmath125 ( open circles ) and @xmath126 ( solid circles ) in the upper panel .
parameters are @xmath67 , @xmath73 , @xmath75 , and @xmath113.,scaledwidth=48.0% ] for @xmath74 the arithmetic average ldos decreases for higher disorder strengths as expected . as a consequence of increasing interaction we observe an increase of the metallicity for each value of the considered disorder strengths in agreement with the tmt results . for @xmath96
the metallicity increases gradually until it suddenly drops to zero for @xmath130 .
this behavior indicates a first order transition in agreement with the observed hysteresis region within tmt.@xcite however , no sudden transition to the mott insulator occurs for higher disorder strengths as for example displayed for @xmath93 and @xmath102 but a rather smooth decline .
our data does not allow for answering the question whether the decline corresponds to a second order transition or a crossover , whereas the previous tmt analysis suggests a crossover.@xcite the increase of metallicity is conform with the observed influence of the interaction strength on the pdf of the ldos as shown in fig .
[ fig_pdf_vary_u ] for @xmath102 .
the two regimes of distinct behaviors of the pdf upon increase of the interaction discussed previously are not reflected in the metallicity .
the metallic phase obviously features non - trivial properties , which might be associated to the non - fermi liquid phase representing the precursor of the anderson - mott transition.@xcite future research is needed to substantiate this possibility and the additional determination of the pdf of further local observables such as the quasi - particle decay rate might be necessary to properly characterize this phase . the lower panel of fig . [ fig_box_metallicity ]
displays the metallicity as a function of disorder strength for three different values of the interaction .
the finite metallicity of the homogeneous system found for small interaction remains for increasing disorder strength , but reduces gradually .
starting within the mott insulating phase of the homogeneous system , an increase of the interaction strength results in a finite metallicity from some critical value on , which confirms the disorder - induced delocalization .
$ ] with decreasing @xmath88 for the ( a ) non - interacting system and ( b ) in presence of interactions , @xmath126 , at fixed disorder strength @xmath102 .
parameters are @xmath67 , @xmath75 , and @xmath113.,scaledwidth=48.0% ] next we discuss how anderson localized states are influenced by interactions .
panel ( a ) of fig .
[ fig_deloc_by_u ] shows the evolution of the pdf of the ldos at the fermi level upon @xmath78 for the non - interacting system with @xmath102 and @xmath113 .
this state is anderson insulating . in panel
( b ) the corresponding evolution is shown for finite interactions , namely @xmath126 . clearly , the pdf saturates corresponding to an extended state .
the interaction obviously causes a transition from the anderson insulator to the disordered metal .
thus , a second delocalization transition caused by the increase of the interaction strength is identified , which is in agreement with the corresponding tmt analysis.@xcite now we are also able to understand the dramatic effect of the introduced correlations on the pdf of the ldos as displayed in panel ( a ) of fig .
[ fig_pdf_vary_u ] , since the system is in the metallic state already for @xmath111 . , @xmath131 , and @xmath113.,scaledwidth=48.0% ] our systematic analysis results in the paramagnetic phase diagram depicted in fig .
[ fig_pd_box ] . both above described delocalization processes
are clearly visible , as can be judged from the extent of the metallic phase . for strong interactions
a mott insulator occurs , which is continuously connected to the anderson - mott insulator for strong disorder .
that means , the mott insulating state can be continuously tuned into an anderson - mott insulating state by changing the disorder and the interaction strength .
the dashed line corresponds to a vanishing arithmetic average of the ldos at the fermi level , as a reasonable distinction between the mott insulator and the anderson - mott insulator . within tmt
studies a distinction was achieved by the criterion of vanishing hubbard subbands@xcite or whether the quasi - particle weight drops to zero for all on - site energies or not.@xcite here , we see that the anderson - mott insulator and the mott insulator are separated both by the metallic phase and for strong interactions by a crossover line @xmath132 .
future research is needed to quantify the critical behavior of the transition from metall to mott insulator for higher disorder strengths .
statistical dmft investigations in combination with more powerful and accurate impurity solvers such as the numerical renormalization group@xcite could give valuable insight in this respect .
it is interesting to compare our statistical dmft phase diagram with the phase diagram obtained by means of the tmt.@xcite the overall structure is reproduced . within both methods
one finds a metallic core for intermediate strengths of both the disorder and the interaction .
also the shape of the metallic core is in agreement .
the two delocalization processes are obtained within both methods . however , on a quantitative level the methods differ substantially . in the non - interacting system ,
the tmt predicts the anderson localization to occur at a critical disorder strength @xmath133.@xcite within statistical dmft the transition is found for significantly higher disorder strength . for the connectivity @xmath113 the critical disorder strength @xmath134 results in a discrepancy of a factor @xmath135 .
we remark that within tmt no fluctuations due to disorder are incorporated .
furthermore , the choice of the mean , i.e. the geometric average , is to some degree arbitrary and has a strong influence on the extent of the metallic phase.@xcite this discrepancy was already discussed by alvermann and fehske for the non - interacting system.@xcite therein it was also shown that tmt misses the re - entrant behavior of the mobility edge that was found by means of the local distribution method . moreover
, our discussion above shows that within statistical dmft the extent of the metallic phase depends on the connectivity @xmath11 as it is expected from a physical reasoning .
for higher connectivities , within statistical dmft , this discrepancy to tmt findings will even be consolidated , since the critical disorder strength for anderson or anderson - mott localization will be higher . as a consequence , the crossover line between the metallic phase and the mott insulating phase exists for much larger disorder strengths than found within tmt .
on the other hand , it should be mentioned that within the tmt investigation the numerical renormalization group was employed as impurity solver .
therewith , the mott transition is determined essentially numerically exactly .
in contrast , within the here presented statistical dmft approach the approximate mpt was used , which is known to differ from the most accurate values known.@xcite to sum up the comparison , our findings indicate that criticism regarding the detection of localization within tmt@xcite is justified on a quantitative level . however
, our investigation supports that statistical dmft and tmt lead to qualitatively agreeing results in general . in this sense ,
the tmt represents a valuable investigation tool for strongly correlated , disordered systems .
an important advantage of the tmt is the considerable smaller computational effort , which enables the use of more powerful impurity solvers than it is feasible to date within statistical dmft .
the statistical dmft was applied to the box disordered anderson - hubbard hamiltonian .
we have confirmed that statistical dmft is able to describe localization transitions in the presence of interactions and disorder .
localization is judged by the evolution of the pdf of the ldos in the limit @xmath136 , which enables the investigation of the anderson - mott transition .
it was found that the presence of interactions in the anderson insulating phase causes delocalization for intermediate disorder strengths . on the other hand , also disorder delocalizes the mott insulating phase by the redistribution of states into the mott gap
. however , if the disorder or respectively the interaction is strong enough , no metallic phase occurs anymore .
consequently , the mott insulator and the anderson - mott insulator are continuously connected .
the resulting paramagnetic ground state phase diagram was found to be in qualitative agreement with tmt results .
however , considerable quantitative differences between statistical dmft and tmt were found regarding the extent of the metallic phase .
our analysis of the probability distribution functions of the local density of states in the interacting case revealed a non monotonic dependence on the interaction strength .
the most striking feature is an emerging second peak in the probability distribution function which was not observed before . in particular
, the log - normal distribution does not serve as a suitable approximation in the interacting case .
our results may be not only relevant for the solid state community but also for future experiments with ultracold fermions in disordered optical lattices,@xcite in which the disorder strength and the interaction can be controlled and tuned independently .
we acknowledge useful discussions with u. bissbort and j. wernsdorfer .
this work was supported by the deutsche forschungsgemeinschaft ( dfg ) via forschergruppe for 801 .
computations were performed at the center for scientific computing ( csc ) at the goethe university frankfurt . k.b . acknowledges support by the grant trr80 of the dfg and the grant no .
n n202 103138 by the polish ministry of science and education . 99 p. w. anderson , phys . rev . * 109 * , 1492 ( 1958 ) . b. kramer and a. mackinnon , rep .
. phys . * 56 * , 1469 ( 1993 ) .
n. f. mott , proc .
a * 62 * , 416 ( 1949 ) ; _ metal insulator transitions _ , 2nd edn .
( taylor and francis , london 1990 ) . m. imada , a. fujimori , and y. tokura , rev . mod .
phys . * 70 * , 1039 ( 1998 ) .
p. a. lee and t.v .
ramakrishnan , rev .
phys . * 57 * , 287 ( 1985 ) .
b. l. altshuler and a. g. aronov , in _ electron - electron interactions in disordered systems _ , edited by m. pollak and a. l. efros ( north - holland , amsterdam , 1985 ) .
d. belitz and t. r. kirkpatrick , rev .
phys . * 66 * , 261 ( 1994 ) .
s. kravchenko and m. sarachik , rep .
. phys . * 67 * , 1 ( 2004 ) .
e. miranda and v. dobrosavljevi , rep .
. phys . * 68 * , 2337 ( 2005 ) .
a. georges , g. kotliar , w. krauth , and m. j. rozenberg , rev .
* 68 * , 13 ( 1996 ) .
g. kotliar , s. y. savrasov , k. haule , v. s. oudovenko , o. parcollet , and c. a. marianetti , rev .
mod . phys . * 78 * , 865 ( 2006 ) .
w. metzner and d. vollhardt , phys .
lett . * 62 * , 324 ( 1989 ) .
r. vlaming and d. vollhardt , phys .
b * 45 * , 4637 ( 1992 ) .
m. ulmke , v. jani and d. vollhardt , phys .
b * 51 * , 10411 ( 1995 ) . v. dobrosavljevi , a. a. pastor , and b. k. nikoli , europhys . lett . * 62 * , 76 ( 2003 ) .
k. byczuk , w. hofstetter , and d. vollhardt , phys .
. lett . * 94 * , 056404 ( 2005 ) .
m. c. o. aguiar , v. dobrosavljevi , e. abrahams , and g. kotliar , phys .
b * 73 * , 115117 ( 2006 ) . m. c. o. aguiar , v. dobrosavljevi , e. abrahams , and g. kotliar , phys .
lett . * 102 * , 156402 ( 2009 ) .
k. byczuk , w. hofstetter , and d. vollhardt , phys .
lett . * 102 * , 146403 ( 2009 ) .
k. byczuk , w. hofstetter , and d. vollhardt , int . j. mod .
* 24 * , 1727 ( 2010 ) .
v. dobrosavljevi , int . j. mod
b * 24 * , 1680 ( 2010 ) .
v. dobrosavljevi and g. kotliar , phys .
lett . * 78 * , 3943 ( 1997 ) ; phil .
lond . a * 356 * , 57 ( 1998 ) .
e. miranda and v. dobrosavljevi , phys .
lett . * 86 * , 264 ( 2001 ) .
m. c. o. aguiar , e. miranda and v. dobrosavljevi , phys .
b * 68 * , 125104 ( 2003 ) . f. x. bronold , a. alvermann , and h. fehske , phil . mag
. b * 84 * , 673 ( 2004 ) .
tran , phys .
b * 76 * , 245122 ( 2007 ) . y. song , r. wortis , and w. a. atkinson , phys . rev .
b * 77 * , 054202 ( 2008 )
. y. song , s. bulut , r. wortis , and w. a. atkinson , j. phys .
matt * 21 * , 385601 ( 2009 ) .
d. semmler , k. byczuk , and w. hofstetter , phys .
b * 81 * , 115111 ( 2010 ) .
d. semmler , j. wernsdorfer , u. bissbort , k. byczuk , and w. hofstetter , phys .
b * 82 * , 235115 ( 2010 ) . a. l. efros and b. i. shklovskii , j. phys . * c8 * , l49 ( 1975 ) .
b. l. altshuler and a. g. aronov , solid state commun .
* 30 * , 115 ( 1979 ) . s. chiesa , p. b. chakraborty , w. e. pickett , and r. t. scalettar , phys . rev
* 101 * , 086401 ( 2008 ) .
g. kotliar and a. e. ruckenstein , phys .
lett . * 57 * , 1362 ( 1986 ) .
j. hubbard , proc .
royal soc .
london a , * 276 * , 238 ( 1963 ) .
h. kajueter and g. kotliar , phys .
lett * 77 * , 131 ( 1996 ) .
m. potthoff , t. wegner , and w. nolting , phys . rev .
b * 55 * , 16132 ( 1997 ) .
a. d. mirlin , phys .
b * 53 * , 1186 ( 1996 ) .
a. d. mirlin , physics reports * 326 * , 259 ( 2000 ) .
g. schubert , j. schleede , k. byczuk , h. fehske , and d. vollhardt , phys .
b * 81 * , 155106 ( 2010 ) .
m. eckstein , m. kollar , k. byczuk , and d. vollhardt , phys .
b * 71 * , 235119 ( 2005 ) .
m. kollar , m. eckstein , k. byczuk , n. blmer , p. van dongen , m. h. radke de cuba , w. metzner , d. tanaskovic , v. dobrosavljevi , g. kotliar , and d. vollhardt , ann .
( leipzig ) * 14 * , 642 ( 2005 ) .
e. n. economou , _
green s functions in quantum physics _
( springer - verlag , berlin , 2006 ) .
r. abou - chacra , p. w. anderson , and d. j. thouless , j. phys .
c * 6 * , 1734 ( 1973 ) .
r. abou - chacra and d.j .
thouless , j. phys .
c * 7 * , 65 ( 1974 ) . even on a real bravais lattice restricted to the leading contribution
, the first equation would extend over @xmath12 neighbors , but all subsequent equations would incorporate @xmath137 or less summands . e.
mller - hartmann , z. phys .
b * 76 * , 211 ( 1989 ) .
x. y. zhang , m. j. rozenberg , and g. kotliar , phys .
rev . lett . * 70 * , 1666 ( 1993 ) .
e. n. economou and m. h. cohen , phys .
b * 5 * , 2931 ( 1972 ) .
d. basko , i. aleiner , and b. altshuler , ann . phys . * 321 * , 1126 ( 2006 ) .
a. alvermann and h. fehske , j. phys .
* 35 * , 145 ( 2006 ) .
a. alvermann and h. fehske , eur .
b * 48 * , 295 ( 2005 ) .
d. c. licciardello and e. n. economou , phys .
b * 11 * , 3697 ( 1975 ) .
d. j. thouless , j. phys .
c * 3 * , 1559 ( 1970 ) .
e. abrahams , p. w. anderson , d. c. licciardello , and t. v. ramakrishnan , phys .
lett . * 42 * , 673 ( 1979 ) .
h. shinaoka and m. imada , phys .
lett . * 102 * , 016404 ( 2009 ) .
h. shinaoka and m. imada , j. phys .
* 78 * , 094708 ( 2009 ) .
chen , r. wortis , and w. a. atkinson , phys .
b * 84 * , 045113 ( 2011 ) .
h. r. krishna - murthy , j. w. wilkins , and k. g. wilson , phys .
b * 21 * , 1003 ( 1980 ) .
w. hofstetter , phys .
lett . * 85 * , 1508 ( 2000 ) .
r. bulla , t. costi , t. pruschke , rev .
phys . * 80 * , 395 ( 2008 ) .
a. m. c. souza , d.o .
maionchi , and h. j. herrmann , phys .
b * 76 * , 35111 ( 2007 )
. y. song , w. a. atkinson , and r. wortis , phys .
b * 76 * , 045105 ( 2007 ) .
r. bulla , t. a. costi , and d. vollhardt , phys .
b * 64 * , 045103 ( 2001 ) . a. aspect and m. inguscio , physics today * 62 * , 30 ( 2009 ) .
l. sanchez - palencia and m. lewenstein , nature phys .
* 6 * , 87 ( 2010 ) . | strongly correlated electrons with box disorder in high - dimensional lattices are investigated .
we apply the statistical dynamical mean - field theory , which treats local correlations non - perturbatively .
the incorporation of a finite lattice connectivity allows for the detection of disorder - induced localization via the probability distribution function of the local density of states .
we obtain a complete paramagnetic ground state phase diagram and find correlation - induced as well as disorder - induced metal - insulator transitions .
our results qualitatively confirm predictions obtained by typical medium theory .
moreover , we find that the probability distribution function of the local density of states in the metallic phase strongly deviates from a log - normal distribution as found for the non - interacting case . |
modern low cost devices are increasingly based on organic semiconductors . this important class of materials allows to achieve well - priced thin film transistors and optical components such as light emitting diodes .
an important intermediate step in this technology is the possibility to control the growth behaviour of active organic materials in terms of orientation and structure on short length scales .
to do so patterned thin surface layers or self - assembled monolayers can be used .
the patterning of such films can be achieved in various ways including but not limited to soft lithography,@xcite scanning probe techniques ( dip pen lithography,@xcite nanografting,@xcite etc . ) , energetic beams ( uv - light,@xcite electrons,@xcite etc . ) and many more.@xcite here , we present a friction force microscopy ( ffm ) study of thin surface layers of a photosensitive thiocyanate - functionalized trialkoxysilane on silicon oxide ( sio@xmath0 ) . these films can be reliably prepared on this technological important surface with a high degree of control over the final film thickness . however , more important is the fact that these films can easily be modified by uv light and subsequently functionalized.@xcite as different end groups of the molecule will have different interaction with the probe of the atomic force microscope , ffm@xcite allows to differentiate between them . using this approach ,
four different terminations could be distinguished and hierarchically ordered on a sample that has been subsequently irradiated twice using line masks with different feature spacings . as a result
the different terminations could be hierarchically ordered by their interaction strength with the afm probe .
for the preparative work of the organic thin surface layers , hazardous chemicals and solvents are used ( ammonium thiocyanate , methanol , propylamine , 2,2,2-trifluoroethylamine , and piranha solution ) .
in addition , piranha solution is explosive , and its preparation is highly exothermic ( up to 120@xmath1c ) .
therefore , reactions must be carried out in a fume hood , and protective clothes and goggles must be used !
uv irradiation causes severe eye and skin burns .
precautions ( uv protective goggles and gloves ) must be taken !
the photoreactive surface layers were prepared by immersion of pretreated ( piranha solution ) boron doped silicon wafers into a solution of trimethoxy[4-(thiocyanatomethyl)phenyl]silane ( si - scn ) in toluene .
x - ray reflectivity measurements revealed a film thickness of 6 nm for these films .
it has to be emphasized here , that obviously this is not a monolayer but an oligolayer with a thickness corresponding to five or six individual layers ( assuming upright standing molecules ) .
the formation of oligolayers is attributed to cross linking of the trimethoxy - silane groups in the presence of water.@xcite in a subsequent step , the samples were illuminated by uv light under inert gas to avoid photo - oxidation ( 254 nm , 80@xmath2 ) .
the illumination leads to an isomerization of the benzyl thiocyanate ( si - scn ) group to the corresponding benzyl isothiocyanate ( si - ncs ) .
this illumination step was done by utilizing a contact mask with equidistant lines and spaces to create a pattern on the surface which consists of alternating stripes of si - scn and si - ncs . for selected samples
an additional post - isomerization modification was performed by exposing the surface to vapours of propylamine . in this process
the isothiocyanate group reacts to the corresponding thiourea group ( si - pa ) . the sequence of the reaction steps together with the respective molecular structure is shown in fig .
[ fig : reactions ] .
a detailed description of the film preparation has already been given elsewhere.@xcite the afm results were obtained with a digital instruments multimode iiia atomic force microscope . to reduce damage to the film , the topographic images were recorded in intermittant contact mode , eliminating effectively lateral forces between the tip and the sample surface . for intermittant mode
normal si probes with a typical resonance frequency of 300khz were used . for roughness characterization ,
the root mean square roughness @xmath3 , the lateral correlation length @xmath4 , and the hurst parameter @xmath5 were calculated from the images using the height - height correlation function ( hhcf).@xcite all roughness parameters have been obtained by analyzing at least three independent 5@xmath6 m images .
ffm ( also called lateral force microscopy ( lfm ) or chemical force microscopy ) @xcite is a special type of contact mode atomic force microscopy .
lateral forces acting on the tip will twist the cantilever , when scanned perpendicular to its long axis , leading to a deflection of the laser on the four quadrant photodiode in lateral direction .
the twist of the cantilever depends on the friction between the tip and the sample surface . as cantilevers ,
specially designed ffm silicon rectangular beam type cantilevers are used .
the cantilevers have a nominal length of 225@xmath6 m and a force constant of typical 0.2n / m .
the lateral force acting on the tip is influenced by the friction coefficient between tip and sample surface .
this coefficient depends on the interaction between the tip and the terminating group of the molecules forming the thin film.@xcite for a clearer contrast , ffm images are calculated from trace and retrace images obtained simultaneously with the topographic image .
this effectively reduces false ffm contrast originating from the surface morphology .
the presented images are therefore always calculated from @xmath7 . for the presented ffm images no scale is given as no force calibration
was performed prior to the measurement .
thus information that can be obtained is purely qualitative , however , with sufficient accuracy allowing to establish a hierarchy of adhesion forces.=1
figure [ fig : unpatt ] presents afm topography images demonstrating the effect of film preparation on surface roughness . homogeneous films of si - scn ( b ) , and propylamine modified si - ncs ( si - pa ) ( c ) films were prepared and compared to the surface of the bare substrate ( a ) . the surface of the substrate ( fig .
[ fig : unpatt](a ) ) shows a uniform featureless topography as expected for a silicon dioxide surface .
the root mean square ( rms ) roughness of @xmath3=0.2 nm , the lateral correlation length @xmath4=30 nm , and the hurst parameter @xmath5=0.5 confirm the qualitative observation .
other investigations report much larger correlation length for sio@xmath8 measured by x - ray and optical techniques.@xcite the smaller values given here are related to the shorter length scales that can be evaluated with afm .
deposition of a thin layer of si - scn and subsequent illumination with 254 nm uv light for 20min ( resulting in a si - ncs film shown in fig .
[ fig : unpatt](b ) ) does not lead to a strong change in the roughness parameters : @xmath3=0.3 nm , @xmath4=30 nm and @xmath5=0.5 . as mentioned above
, x - ray reflectivity measurements revealed a film thickness of 6 nm for this layer.@xcite modifying the surface with vapors of propylamine ( fig .
[ fig : unpatt](c ) ) leads to a further slight increase in surface roughness ( @xmath9 nm @xmath10 nm , and @xmath11 ) .
the changes in the roughness parameters can be mainly attributed to contamination of the sample surface due to handling and residue from the chemical compounds used during modifications and cleaning . in the following ,
ffm data of the patterned films are presented .
figure [ fig : scn - ncs ] compares the topographic image ( a ) obtained on a si - scn layer that was illuminated through a stripe mask with 10@xmath6 m lines and 10@xmath6 m spaces to the simultaneously recorded ffm image ( b ) .
whereas in the topography image no stripe pattern is visible , a clear stripe pattern with a 10@xmath6 m pitch appears in the ffm image . in all ffm images presented ,
bright areas mean higher friction and dark ones correspond to lower friction .
it will be demonstrated below that the high friction areas correspond to the illuminated si - ncs stripes and the low friction areas are the pristine si - scn stripes covered by the mask during illumination . mask . in the ffm image ( b )
bright areas correspond to si - ncs terminated areas showing high friction .
the dark stripes are the non - illuminated si - scn areas . ]
figure [ fig : scn - pa ] shows afm images obtained after the patterned sample described above has been exposed to vapours of propylamine . while the topographic image presented in fig .
[ fig : scn - pa](a ) shows a weak stripe pattern with the expected spacing , the corresponding ffm image in fig .
[ fig : scn - pa](b ) allows a clear identification of the line pattern created by contact lithography and of errors in the masking process ( lower right corner ) . from the combined cross section in fig .
[ fig : scn - pa](c ) the height difference of 0.5 nm between the modified si - pa and the pristine si - scn stripes is clearly discernible .
the addition of an alkyl group to the molecule will result in an increase of the film thickness .
an increased height is therefore only expected for the modified parts of the surface , which are the si - pa areas .
these areas show a lower friction signal .
areas of lower height correspond to non - illuminated ( and unmodified ) zones of the layer containing si - scn structures , which give a higher friction than the si - pa structure .
the increase in layer thickness after reaction with propylamine was 0.6 nm which equals 10% of the initial layer thickness .
this corresponds to results obtained with photoreactive polymers containing benzyl thiocyanate side groups.@xcite in a further experiment , the surface has been exposed twice to illumination using crossed masks with an intermediate propylamine reaction step .
the whole sample preparation process is sketched in fig .
[ fig : cross](a ) . in a first step ,
a mask with 5@xmath6 m lines and spaces was employed during uv illumination .
after the si - scn film has been exposed to vapors of propylamine , a surface layer made up of alternating 5@xmath6 m stripes of si - scn and si - pa is created .
the resulting pattern is similar to the one presented in fig .
[ fig : scn - pa ] .
this modified film was now illuminated for a second time through a mask with 10@xmath6 m lines and spaces oriented perpendicular to the first mask pattern . during this step both stripes , si - pa as well as si - scn , are illuminated partly .
it can be expected that the si - pa surface will not change significantly during this process .
however , the remaining si - scn stripes will be converted into alternating 10@xmath6 m patches of si - scn and si - ncs .
the resulting surface morphology and the friction image are presented in fig .
[ fig : cross](b , c ) . as before , the main features in the topographic image ( fig .
[ fig : cross](b ) ) are not related to the mask process but result from contamination and small long range undulations in the sio@xmath0 surface of the wafer .
however , the ffm image shows a clear pattern of regular 5 by 10@xmath6 m patches of four different shadings , i.e. friction levels .
we can identify the four areas using the information on the friction contrast obtained from the previous samples .
the dark areas in fig .
[ fig : cross](b ) are the si - pa areas ( labeled pa ) created in the first illumination step .
the neighboring brighter patches ( pi ) in the 5@xmath6 m stripe are the si - pa areas that were subsequently illuminated a second time . in the neighboring 5@xmath6 m stripes ( shadowed during the first illumination )
one patch has been protected by the mask in both illumination steps ( da ) while the other one has been exposed to uv - light a single time during the second illumination ( il ) .
the last two areas are comparable to those shown in fig .
[ fig : scn - ncs ] and therefore allow the identification of the stripes in fig .
[ fig : scn - ncs](b ) . the areas marked pa and da are comparable to the combination shown in fig .
[ fig : scn - pa ] . the observed friction contrast can therefore be ordered in the following way : the highest tip - film interaction and therefore the largest friction is observed for si - ncs ( il ) followed by si - scn ( da ) and the two propylamine modified surfaces ( pi and pa in fig .
[ fig : cross](b ) ) .
a possible explanation fort his sequence can be given by the different polarity of the individual endgroups at the surface ( thiocyanate ( si - scn ) , isothiocyanate ( si - ncs ) , and propyl ( ch@xmath8-ch@xmath8-ch@xmath12 , si - pa ) ) and by the stiffness of the molecular endgroups .
comparing data on the dipole moment of ethyl isothiocyanate ( 3.67 debye ) and ethyl thiocyanate ( 3.33 debye ) as well as on the surface tension @xmath13 ( at 20@xmath1c ) of these compounds ( ethyl isothiocyanate : @xmath13 = 36.0 mn / m , and ethyl thiocyanate : @xmath13 = 34.8 mn / m ) it is found that isothiocyanates are of higher polarity than the corresponding thiocyanates.@xcite assuming that the friction force between the surface and the silicon tip ( which is covered with a native oxide layer ) increases with the polarity of the surface , for the uv illuminated regions ( containing ncs units at the surface ) a higher friction force will be recorded than for the non - illuminated regions bearing scn units . after reaction of the photogenerated ncs groups with propylamine ( pa ) , the surface
is terminated with non - polar alkyl groups .
it is therefore not surprising that the post - exposure derivatization with propylamine will result in lower friction force .
also the stiffness of the molecule is reported to influence the resulting friction coefficient .
@xcite the flexible alkyl group that terminates si - pa will therefore also reduce the observed friction by bending under the applied normal load .
in contrast , the shorter and stiffer thiocyanate and iso - thiocyanate groups ( si - ncs has two double bonds between sulfur , carbon , and nitrogen ; si - scn has a triple bond between carbon and nitrogen while the sulfur is linked by two single bonds ) can not bend under the applied load .
the intermediate friction contrast observed for areas that were illuminated and modified with propylamine , and then illuminated for a second time , can also be explained that way . during the second illumination , residual scn units which have remained unreacted in the first illumination step are converted into ncs groups leading to an increased interaction between the surface and the afm tip .
ncs groups which have reacted with propylamine ( to yield propyl thiourea units ) are expected to remain unchanged during the second illumination step since n , n-dialkyl substituted thiourea groups are stable under uv light . especially from the last sample ,
the hierarchy in the interaction forces ( si - ncs @xmath14 si - scn @xmath14 si - pa ) between the individual terminating groups becomes evident .
for lithographically patterned photoreactive surface layers we demonstrated that friction force microscopy is not only able to distinguish between different head groups , but also that this can be done simultaneously for at least three different terminations .
the method is able to detect minute changes in the molecular structure .
the case of the thiocyanate ( si - scn ) and isothiocyanate ( si - ncs ) head groups is especially relevant since these are educts and products of an isomerization reaction .
ffm is able to discern the two terminations , although , only the sequence of the terminating three atoms is altered .
however , the resulting change in polarity of the terminating molecular groups is large enough to be detected by ffm as different friction levels .
in addition , the post isomerization modification with propylamine could be clearly visualized with ffm . in this case , a further reduction of polarity together with a change in the stiffness of the molecule leads to the lowest friction observed in the investigated system .
currently , contact angle measurements are underway to obtain an independent confirmation of the observed hierarchy .
the next step in future work will be to quantify the adhesive and frictional forces responsible for the qualitative results presented here .
for this task a well known approach suggested in literature will be used to calibrate the afm probes with sufficient accuracy.@xcite these studies shall include the use of functionalized tips to fine tune the probe sample interaction.@xcite | precise control of surface properties including electrical characteristics , wettability , and friction is a prerequisite for manufacturing modern organic electronic devices .
the successful combination of bottom up approaches for aligning and orienting the molecules and top down techniques to structure the substrate on the nano and micrometer scale allows the cost efficient fabrication and integration of future organic light emitting diodes and organic thin film transistors .
one possibility for the top down patterning of a surface is to utilize different surface free energies or wetting properties of a functional group . here , we used friction force microscopy ( ffm ) to reveal chemical patterns inscribed by a photolithographic process into a photosensitive surface layer .
ffm allowed the simultaneous visualization of at least three different chemical surface terminations .
the underlying mechanism is related to changes in the chemical interaction between probe and film surface . |
pinterest is an online catalog used to discover and save ideas .
hundreds of millions of users organize pins around particular topics by saving them to boards .
each of the more than 50 billion pins saved on pinterest has an image , resulting a large - scale , hand - curated collection , with a rich set of metadata .
most pins images are well annotated : when a person bookmarks an image on to a board , a _
pin _ is created around an image and a brief text description supplied by the user . when the same pin is subsequently saved to a new board by a different user , the original pin gains additional metadata that the new user provides .
therefore , the data structure surrounding each pin continues to get richer each time the pin is re - saved .
furthermore , boards ( i.e. collections of pins ) reveal relations _ between _ pins : if many users save these two pins together , there is a high likelihood that another user may find them to be related as well
. such aggregated _ image co - occurrence _ statistics are found to be useful for related content recommendation .
this work explores how user curation signals can be used in conjunction with content - based features to improve recommendation systems .
specifically we introduce related pins , an _ item - to - item _ content recommendation service triggered when a pin closeup is shown to the user , and describe in detail our experiments using visual features ( such as those obtained from convolutional neural networks ) , which are of particular interest since this system ultimately recommends visual content .
as one of the most popular features on pinterest , related pins is a recommendation system that combines collaborative filtering @xcite with content - based @xcite retrieval . since may 2015 , the user engagement metric on pin recommendations has improved by more than 50% .
note that the improvement is the result of using both visual features and other metadata signals in the learning - to - rank framework the scope of this paper is limited to the understanding of user curation and visual features in the context of recommendation systems .
this work makes two contributions : first , we demonstrate that `` pinning , '' a form of user curation , provides valuable user signals for content recommendation .
specifically we present our use of _ image / board co - occurrences _ , including the _ pinjoin _ data structure used to derive this signal .
second , we demonstrate that combining collaborative filtering with content - based retrieval methods , such as applying a learning - to - rank framework @xcite to a set of semantic and visual features associated with the candidate pins , can significantly improve user engagement .
in particular , our a / b experiments demonstrate that the use of recently developed visual features ( when used in conjunction with other text and graph signals ) , such as those obtained from vgg @xcite and faster r - cnn @xcite yield significant gains in recommendation quality .
collaborative filtering @xcite using user - generated signals ( e.g. co - views , co - clicks ) is widely used in commercially deployed recommendation systems such as youtube related videos @xcite and amazon related items @xcite .
this work investigates the use of _ user curation _ signals derived from pins image / board co - occurrences , which are unique to pinterest .
visual features are widely used in both content - based recommendation systems and image search systems @xcite , and the the learning - to - rank framework used in this paper from @xcite has been widely used in industry @xcite . to our best knowledge
this work contains the first published empirical results on how the latest convolutional neural network ( cnn ) based visual features ( e.g. vgg @xcite ) and large - scale object detection using faster r - cnn@xcite can improve commercial recommendation systems .
visual features are computed using a distributed process described in our previous work @xcite .
content curation is the process of organizing and collecting content relevant to a particular topic of interest .
pinterest is a user - powered content curation service as content is collected and grouped into topic boards , creating a rich set of metadata associated with pins images .
for example , during pin creation , users typically provide a text description of the images as shown in figure 3 .
although any single instance of text description can be noisy , an aggregated collection reveals important annotations relevant to the pin s image .
furthermore , when a pin is saved to a board , one can infer the categorical information of the pin s image from the category the user selected for the board .
formally , we denote the data structures associated with pins and boards in the following way : each _ pinjoin _ is a 3-tuple @xmath0 , where @xmath1 is the image url , @xmath2 is the collection of pins generated for that image , and @xmath3 is the aggregation of text annotations or keywords ( extracted from board titles and descriptions ) . each _
boardjoin _ is represented as a 2-tuple @xmath4 , where t is the board title and p is a list of pins .
pinjoin is conceptually similar to visual synsets @xcite except in this case , all the images within a visual synset are exact duplicates of each other and the collection is curated manually . in practice , both of these structures contain additional metadata , such as category and topic information , which is used for deriving other features during re - ranking .
user curation reveals relations _ among _ images : we observed that images of pins on the same board are semantically ( and to some extent visually ) related to each other .
therefore , if enough users save these two pins together , there is a high likelihood that a new user may also find them to be related .
this is helped by the fact that users on pinterest actively curate content our engaged users have an average of 24 boards .
an example of image / board co - occurrences is shown in figure [ fig : occur_examples ] .
this section presents the architecture that powers the pinterest related pins recommendation system .
the first step described relies on collaborative filtering over user curation signals ( image co - occurrences on boards ) to generate a candidate sets .
the second step uses content - based ranking approach to rank the candidates based on content signals such as visual features , textual signals , and categories signals derived from _
pinjoins_. the first step of the pipeline is to generate a set of image candidates for each query image , which will serve as candidate sets for content - based re - ranking .
we adopt a classic collaborative filtering approach to exploit the pin / board co - occurrences as described in section 3 . for each pin , we select up to 10,000 pins with the highest number of shared boards . in practice
, the candidate generation process is accomplished through a mapreduce job , which takes _ boardjoin _ @xmath5 as input .
the mapping stage outputs image pairs @xmath6 for all image pairs in each board @xmath7 , and in the reduce stage all the related images are grouped by the same query image . for computational efficiency , we sample images based on the quality / popularity of the images . for pins that do not generate enough candidates through board co - occurrence
, we rely on a content - based retrieval service described in our previous paper @xcite . after generating a set of candidates , we re
- rank with a set of both content pair - features ( defined between a query and a candidate ) and query - independent features .
in addition to standard features such as text annotation match and topic vector similarity , we were particularly interested in the effectiveness of visual similarity features for our re - ranking step .
examples of visual features include the _ fc6 _ and _ fc8 _ activations of intermediate layers of deep convolutional neural networks ( cnns ) @xcite based on alexnet @xcite and vgg @xcite .
these features are binarized ( _ fc6 _ ) and sparsified ( _ fc8 _ ) for representation efficiency and compared using hamming distance ( _ fc6 _ ) and cosine similarity ( _ fc8 _ ) , respectively .
we use the open - source caffe @xcite framework to perform training and inference of our cnns on multi - gpu machines . in this work ,
we also trained an object detection module using faster r - cnn @xcite , initially fine - tuned on a dataset containing the most common objects found on pinterest , including home decor and fashion categories such as various furniture types , shoes , dresses , glasses , bags and more .
to learn the weight vector for our linear model , we adopted the learning - to - rank approach from joachims @xcite .
given training data in the form of relative ranking triplets @xmath8 , where document @xmath9 is considered to be more relevant to query @xmath10 than document @xmath11 , the ranksvm algorithm described in @xcite approximates a weight vector @xmath12 which maximizes the number of training examples satisfying @xmath13 , where @xmath14 gives @xmath15 features of the document @xmath16 in the context of query @xmath17 .
the relevance triplets we use in training are generated through user clicks and impression logs , normalized by position and device to account for position bias , using a clicks over expected clicks @xcite ( coec ) model .
for each query @xmath10 in our training set , given the set of observed results @xmath18 , we generate the training triplet : @xmath19 corresponding to the query , best engaged document , and worst engaged document .
we also generate random negative examples : @xmath20 on the intuition that even poorly engaged candidates generated through our board co - occurrence signal should still be more relevant than a random document from our corpus of pins .
in this subsection we present a qualitative analysis of using user curation signals to generate candidates for related pins .
as we described previously , board co - occurrences for pins is a strong signal of relevance and has been the foundation of candidate generation for our system .
figure [ fig : occur_examples ] illustrates that the relevance of the candidates grows gradually when the number of co - occurrences with the query pin increases .
we also show the percentage of image pairs having different number of board co - occurrences in figure [ fig : cooccurrence ] .
note that the majority of the image pairs ( around 80% ) only co - occur once on the same board , which suggests that ranking based on content features as the next step is important for finding high - quality recommendations . on the other hand , there are only a handful of image pairs which co - occur many times on the same boards .
we found that one of the most important features for re - ranking the pin candidates generated through board co - occurrence is visual similarity .
we validated this by setting up a series of a / b experiments , where we selected five million popular pins on pinterest as queries , and re - ranked their recommendations using different sets of features .
the control group re - ranked related pins using a linear model with a set of standard features : text annotation similarity , topic vector similarity , category vector similarity , as well as query - independent features .
the treatment group re - ranked using fine - tuned vgg _
fc6 _ and _ fc8 _ visual similarity features along with indicator variables ( in addition to the features used in control ) . across the 5 m query pins ,
the treatment saw a 3.2% increase in save / clickthrough rate . after expanding the treatment to 100 m query pins
, we observed a net gain of 4.0% in propensity to engage with related pins , and subsequently launched this model into production .
similar experiments with a fine - tuned alexnet model yielded worse results ( only 0.8% engagement gain ) .
when broken down by category , we noted that the engagement gain was stronger in predominantly visual categories , such as art ( 8.8% ) , tattoos ( 8.0% ) , illustrations ( 7.9% ) , and design ( 7.7% ) , and lower in categories which primarily rely on text , such as quotes ( 2.0% ) and fitness planning ( 0.2% ) . given the difference in performance among categories , we performed a follow - up experiment where we introduced a cross feature between the category vector of the query and the scalar _ fc6
_ visual similarity feature ( between the query and candidate ) .
this introduces 32 new features to the model , one for each of our site - wide categories ( these features are sparse , since the pinterest category vector thresholds most values to zero ) .
the result from this was a further 1.2% engagement increase in addition to the gains from the initial visual re - ranking model .
further work into component and cross product features is of interest to us , as they are essentially free to compute at rank - time , since the raw feature data is already stored .
users are sometimes interested in the _ objects _ in the pin s image , instead of the full image ( as shown in figure [ fig : visualobject ] ) .
we therefore speculate that object detection , when feasible , should improve relevance targeting . after applying non - maximum suppression ( nms ) to the proposals generated by our fine - tuned faster r - cnn module mentioned in section , we considered query pins where the largest proposal occupies at least 25% of the pin s image , or if the proposal is smaller , it passes a confidence threshold of 0.9 in faster r - cnn .
we categorize these images as containing a dominant visual object , and using the best - performing fine - tuned vgg re - ranking variant from the previous section as our control , we experimented with the following treatments : * _ variant a _ :
if a dominant visual object is detected in the query pin , we compute visual features ( vgg ) on just that object . * _ variant b _ : same as _ variant a _ , but we also hand - tune the ranking model by increasing the weight given to visual similarity by a factor of 5 .
the intuition behind this variant is that when a dominant visual object is present , visual similarity becomes more important for recommendation quality . * _ variant c _ :
if a dominant visual object is detected in the query pin , we still use the features from the entire image ( as the control does ) , but increase the weight given to visual similarity by a factor of 5 , as in _
variant b_. in this variant , we assume that the presence of detected visual objects such as bags or shoes indicates that visual similarity is more important for this query .
.results when using cross features and object detection , measured over a 7 day period in oct .
2015 [ cols="<,<,^",options="header " , ] results for these variants are listed in table [ tbl : visualsearchrel ] .
variants a and b of the object detection experiments suggest that the tight bounding boxes from our object detection module do not provide enough context for our cnn models , but variant c , which results in an additional 4.9% engagement gain over the vgg similarity feature control , demonstrates that the presence of visual objects indicates that visual similarity should be weighed more heavily .
based on these results , our future focus is scaling up the number of object categories we can detect , and tuning the weight given to visual similarity in variant b and c.
the related pins system described in this work has improved user engagement metric and traffic on pin recommendations by more than 50% from may 2015 to november 2015 .
this demonstrates that signals derived from user curation and the activity of users organizing content contain rich information about the images and are very effective when used in conjunction with collaborative filtering .
we also demonstrate that visual features such as representations learned from cnns or presence of detected visual objects can be used in the learning - to - rank framework to improve item - to - item recommendation systems .
one important component not discussed in this work is our use of user signal in the form of _ navboost _ , which also uses a model based on coec ( extended to actions beyond clicks ) to re - rank content based on user engagement .
our future work includes exploring a richer set of features ( e.g. sparse features , dense features , cross - product features , more object categories ) and real - time recommendations ( enabling re - ranking based on locale , current search query , and other forms of personalization ) .
we would like to thank our colleagues on the visual discovery and recommendations teams at pinterest , in particular dmitry chechik , yunsong guo , and many others .
we d also like to acknowledge jeff donahue and trevor darrell from berkeley vision and learning center ( bvlc ) for their collaboration with pinterest and their work on caffe .
s. baluja , r. seth , d. sivakumar , y. jing , j. yagnik , s. kumar , d. ravichandran , and m. aly .
video suggestion and discovery for youtube : taking random walks through the view graph . in _ proceedings of the 17th international conference on world wide web _ , www 08 , pages 895904 , new york , ny , usa , 2008 .
m. bendersky , l. garcia - pueyo , j. harmsen , v. josifovski , and d. lepikhin .
up next : retrieval methods for large scale related video suggestion . in _ proceedings of the 20th acm sigkdd international conference on knowledge discovery and data mining _ , kdd 14 , pages 17691778 , new york , ny , usa , 2014 .
o. chapelle and y. zhang .
a dynamic bayesian network click model for web search ranking . in _ proceedings of the 18th international conference on world wide web _ , www 09 , pages 110 , new york , ny , usa , 2009 .
acm .
r. datta , j. li , and j. z. wang .
content - based image retrieval : approaches and trends of the new age . in _ proceedings of the 7th acm sigmm international workshop on multimedia information retrieval _ , mir 05 , pages 253262 , new york , ny , usa , 2005 .
d. a. ferrucci , e. w. brown , j. chu - carroll , j. fan , d. gondek , a. kalyanpur , a. lally , j. w. murdock , e. nyberg , j. m. prager , n. schlaefer , and c. a. welty . building watson : an overview of the deepqa project .
, 31(3):5979 , 2010 .
y. jing , d. liu , d. kislyuk , a. zhai , j. xu , j. donahue , and s. tavel . visual search at pinterest . in _ proceedings of the 21th acm sigkdd international conference on knowledge discovery and data mining _ , kdd 15 , pages 18891898 , new york , ny , usa , 2015 .
t. joachims . optimizing search engines using clickthrough data . in _ proceedings of the eighth acm
sigkdd international conference on knowledge discovery and data mining _ , kdd 02 , pages 133142 , new york , ny , usa , 2002 .
acm .
m. richardson , a. prakash , and e. brill . beyond pagerank : machine learning for static ranking . in _ proceedings of the 15th international conference on world wide web _ , www 06 , pages 707715 , new york , ny , usa , 2006 .
b. sarwar , g. karypis , j. konstan , and j. riedl .
item - based collaborative filtering recommendation algorithms . in _ proceedings of the 10th international conference on world wide web _ , www 01 , pages 285295 , new york , ny , usa , 2001 . | this paper presents pinterest related pins , an item - to - item recommendation system that combines collaborative filtering with content - based ranking .
we demonstrate that signals derived from _ user curation _ , the activity of users organizing content , are highly effective when used in conjunction with content - based ranking .
this paper also demonstrates the effectiveness of visual features , such as image or object representations learned from convnets , in improving the user engagement rate of our item - to - item recommendation system . |
at future colliders , the lhc and the ilc , experiments can be performed in the supersymmetric particle sector with very high precision
experimental accuracies are expected at the per - cent down to the per - mille level @xcite .
this should be matched from the theoretical side @xcite .
therefore a well defined theoretical framework is necessary for the calculational schemes in perturbation theory as well as for the input parameters .
motivated by the experience in analyzing data at the former @xmath0 colliders lep and slc , the spa convention and project @xcite has been proposed .
it provides + @xmath1 a convention for calculating masses , mixings , decay widths and production cross sections , + @xmath1 a program repository of codes ( rge , spectrum calculators , fitting routines , event generators etc . )
that will be expanded continuously in the future , + @xmath1 a list of future tasks on both the theoretical and the experimental sides needed before data from future experiments could be evaluated at the desired level of accuracy , + @xmath1 a susy reference point sps1a@xmath2 as a general setup for testing these tools in practice .
+ combining the experimental information from lhc and ilc will provide a high - precision picture of supersymmetry at the tev scale @xcite which subsequently may lead to the reconstruction of the fundamental supersymmetric theory at a high scale and the mechanism of supersymmetry breaking @xcite .
the spa convention and project is a joint inter - regional experimental and theoretical effort .
the current status of the project is documented on the routinely updated web - page +
building on vast experience in susy calculations and data simulations and analyses , the spa convention consists of the following propositions : * the masses of the susy particles and higgs bosons are defined as pole masses . *
all susy lagrangian parameters , mass parameters and couplings , including @xmath3 , are given in the @xmath4 scheme and defined at the scale @xmath5 1 tev .
+ * gaugino / higgsino and scalar mass matrices , rotation matrices and the corresponding angles are defined in the @xmath4 scheme at @xmath6 , except for the higgs system in which the mixing matrix is defined in the on - shell scheme , the scale parameter chosen as the light higgs mass . *
the standard model input parameters of the gauge sector are chosen as @xmath7 , @xmath8 , @xmath9 and @xmath10 .
all lepton masses are defined on - shell .
the @xmath11 quark mass is defined on - shell ; the @xmath12 quark masses are introduced in @xmath13 at the scale of the masses themselves while taken at a renormalization scale of 2 gev for the light @xmath14 quarks .
* decay widths , branching ratios and production cross sections are calculated for the set of parameters specified above .
the @xmath4 scheme , based on dimensional reduction and modified minimal subtraction , is designed to preserve supersymmetry and it is technically very convenient .
recently it has been shown that it can be formulated in a mathematically consistent way @xcite and to comply with the factorization theorem , see @xcite .
the physical on - shell masses are introduced in the decay widths and production cross sections such that the phase space is treated in the observables closest to experimental on - shell kinematics .
to use the highly developed @xmath13 infrastructure for proton colliders a repository contains the translation tools between the @xmath4 and @xmath13 schemes , as well as the on - shell renormalization schemes .
the responsibility for developing codes and maintaining them up to the current theoretical state - of - the - art precision rests with the authors .
the slha @xcite convention is recommended for communication between the codes .
the repository contains links to codes grouped in several categories : 1 .
scheme translation tools for definitions and relations between on - shell , @xmath4 and @xmath13 parameters
spectrum calculators for transition from the lagrangian parameters to a basis of physical particle masses and the related mixing matrices .
3 . calculation of other observables : decay tables , cross sections , low - energy observables etc .
cosmological and astrophysical aspects : cold dark matter relics , cross sections for cdm particle searches , astrophysical cross - sections in the susy context etc .
event generators that generate event samples for susy and background processes in realistic collider environments .
analysis programs to extract the lagrangian parameters from experimental data by means of global analyses . 7 .
rge programs to connect the low - energy effective lagrangian parameters to the high - scale where the model is supposed to be matched to a more fundamental theory .
auxiliary programs and libraries : structure functions , beamstrahlung , numerical methods , sm backgrounds etc .
the spa convention and project is a very ambitious and extended experimental and theoretical effort .
it is set up to cover general susy scenarios .
however , to perform first checks of its internal consistency and to explore the potential of such coherent data analyses a mssm reference point sps1a@xmath2 has been proposed as a testing ground .
of course , in future the spa has to be tested in more extended mssm as well as more complicated scenarios .
the roots defining the point sps1a@xmath2 are the msugra parameters [ in the conventional notation for cmssm ] in the set @xmath15 where the universal gaugino mass @xmath16 , the scalar mass @xmath17 and the trilinear coupling @xmath18 [ yukawa couplings factored out ] , are defined at the gut scale @xmath19 .
the point is close to the original snowmass point sps1a
@xcite and to point b@xmath2 of @xcite . with the sm input parameters given explicitly in the spa document @xcite , extrapolation of the above parameters down to the @xmath20 1 tev scale generates the mssm lagrangian parameters as shown in the left part of table [ tab : par ] . here
the rge part of the program spheno @xcite has been used ( for internal or external comparison , other codes can equally be used ) .
@xmath21 @xmath22 the physical [ pole ] masses of the supersymmetric particles are collected in the right part of table [ tab : par ] .
the connection between the lagrangian parameters and the physical pole masses can presently be controlled at the 1-loop level for the masses of the susy particles , and at the 2-loop level for the higgs masses .
the qcd effects on the heavy quark masses are accounted for to 2-loop accuracy . for illustration ,
the left panel of figure [ fig : xsec ] displays cross sections for the production of squarks and gluinos at the lhc as functions of the squark mass crossing the point sps1a@xmath2 , while the right panel shows the production cross section of pairs of charginos for the point sps1a@xmath2 at the ilc as a function of the collider energy . to perform experimental simulations ,
the branching ratios of the decay modes are crucial .
the spa document and the spa web page provide results of calculations using feynhiggs @xcite and sdecay @xcite .
the born cross sections ( broken lines ) are shown for a few channels.__,title="fig:",width=226,height=188 ] annihilation .
the born cross sections ( broken lines ) are shown for a few channels.__,title="fig:",width=226,height=188 ] if sps1a@xmath2 , or a susy scenario with mass scales similar to this point , is realized in nature , a plethora of interesting channels can be exploited to extract the basic supersymmetry parameters when combining experimental information from mass distributions at lhc with measurements of decay spectra and threshold excitation curves at an @xmath0 collider with energy up to 1 tev @xcite . from the simulated experimental errors
the data analysis performed coherently for the two machines gives rise to a very precise picture of the supersymmetric particle spectrum as demonstrated in the left part of table [ tab : massesa ] .
@xmath23 + . for the point sps1a@xmath2
_ _ accuracies for representative mass measurements of susy particles in individual lhc , ilc and coherent lhc+ilc analyses . @xmath24 and @xmath25 represent @xmath26 . excerpt of extracted susy lagrangian mass and higgs parameters at the supersymmetry scale @xmath27 1 tev .
masses in gev . _
_ [ cols="^,^",options="header " , ]
the results for sps1a@xmath2 presented here are based on preliminary experimental simulations and extrapolations from earlier analyses for sps1a and other reference points as a substitute of missing information necessary for a first comprehensive test of all aspects of the spa project .
althought current spa studies are very encouraging , much additional work both on the theoretical as well as on the experimental side will be needed to achieve the spa goals .
in particular * the present level of theoretical calculations still does not match the expected experimental precision , particularly in coherent lhc+ilc analyses .
new efforts in higher - order susy calculations are necessary for the interpretation of experimental analyses .
* there is no complete proof that @xmath4 scheme preserves supersymmetry and gauge invariance in all cases . as the precision of susy calculations is pushed to higher orders , the spa project also requires an improved understanding of the @xmath4 scheme .
recently it has been shown that for massive final state particles spurious density functions for the @xmath28 gluon components have to be introduced to comply with the factorization theorem @xcite which opens a way to formulating an efficient combination of the most attractive elements of @xmath4 and @xmath13 schemes in describing hadronic processes . *
a limited set of observables included in experimental analyses by no means exhausts the opportunities which data at lhc and at ilc are expected to provide in the future .
likewise , in most analyses errors are purely experimental and do not include the theoretical counterpart which must be improved considerably before matching the experimental standards .
this is particularly important for the coherent combination of future data obtained at lhc and ilc .
feedback and coherently combined analyses , which will greatly benefit from a concurrent running of both colliders , are indispensable for a meaningful reconstruction of the underlying theory @xcite . * astrophysical data
play an increasingly important role in confronting supersymmetry with experiments .
models with r - parity conservation predict a stable weakly interacting , massive particle . in this case on the one hand their relic abundance imposes crucial limits on supersymmetric scenarios and specific requirements on the accuracies must be achieved when the cdm particle is studied in high - energy physics laboratory experiments @xcite . on the other , predictions based on the comprehensive parameter analysis of high - energy experiments determine the cross sections for astrophysical scattering experiments by which the nature of the cold dark matter particles can be established .
for example , a study of the sps1a point at lhc , based on very large statistics @xcite , indicates that the relic density can be determined to @xmath29% for the sps1a@xmath2 scenario . at the ilc
a precision of @xmath30% should be achievable . *
the parameter set sps1a@xmath2 chosen for a first study provides a benchmark for developing and testing the tools needed for a successful analysis of future susy data .
however , neither this specific point nor the mssm itself may be the correct model for low - scale susy . while versions of msugra and of gaugino mediation have also been analyzed in some detail , the analyses have to be extended systematically to other possibilities .
in particular , cp violation , r - parity violation , flavor violation , nmssm and extended gauge groups are among scenarios which might be realised in the susy sector .
the spa conventions are general enough to cover all these scenarios .
the spa project , a joint theoretical and experimental effort , aims at providing + @xmath1 a well defined framework for susy calculations and data analyses , + @xmath1 all necessary theoretical and computational tools , + @xmath1 a testground scenario sps1a@xmath2 , + @xmath1 a platform for future extensions and developments .
+ first results for the reference point spa1a@xmath2 are very encouraging , however it is clear that much work is still needed on the experimental and theoretical sides to achieve the desired level of accuracy .
the spa project is a dynamical system expected to evolve continuously and to encompass more general supersymmetry scanarios .
+ * acknowledgments : * + i would like to thank all the spa members for their efforts to make the spa projects a success . special thanks go to peter zerwas and uli martyn .
+ i am also deeply indebted to prof .
wojciech marczyski , md , dr . tomasz skrok , md , and the whole personnel of the department of traumatology and orthopedics of the central clinical hospital of wam , warsaw , for their hospitality and care while this write - up has been prepared .
aguilar - saavedra _ et al .
_ , tesla technical design report , desy 01 - 011 and arxiv : hep - ph/0106315 ; t. abe _ et al . _ [ american lc wg ] , in _ proceedings of the aps / dpf / dpb summer study on the future of particle physics ( snowmass 2001 ) _ , slac - r-570 and arxiv : hep - ex/0106055 - 58 ; t. abe _ et al . _
[ asian lc wg ] , kek - report2001 - 011 and arxiv : hep - ph/0109166 .
j. kalinowski , supersymmetry parameter analysis , _ proceedings international conference on supersymmetry and unification of findamental interactions susy02 ( hamburg 2002 ) _ , arxiv : hep - ph/0212388 .
j. a. aguilar - saavedra _ et al .
_ , supersymmetry parameter analysis : spa convention and project , arxiv : hep - ph/0511344 . b. c. allanach _ et al .
lhc+lc report _
@xcite , [ and arxiv : hep - ph/0403133 , arxiv : hep - ph/0407067 ] .
g. a. blair , w. porod and p. m. zerwas , phys .
d * 63 * ( 2001 ) 017703 and eur .
j. c * 27 * ( 2003 ) 263 .
b. c. allanach , d. grellscheid and f. quevedo , jhep * 0205 * ( 2002 ) 048 .
g. l. kane , j. lykken , s. mrenna , b. d. nelson .
l. t. wang and t. t. wang , phys .
d * 67 * ( 2003 ) 045008 .
d. stckinger , jhep * 0503 * ( 2005 ) 076 .
a. signer and d. stockinger , phys .
b * 626 * ( 2005 ) 127 .
skands _ et al .
_ , jhep * 0407 * ( 2004 ) 036 . b. c. allanach _ et al .
_ , eur .
j. c * 25 * ( 2002 ) 113 . | reconstruction of the fundamental supersymmetric theory and its breaking mechanism will require high - precision tools . here a brief introduction to spa , the supersymmetry parameter analysis ( spa ) convention and project , is presented which is based on a consistent set of conventions and input parameters .
= by -1 |
in network communications , the communication could fail if some nodes or some edges are broken .
though the failure of a modem could be considered the failure of a node , we can model this scenario also as the failure of the communication link ( the edge ) attached to this modem .
thus it is sufficient to consider edge failures in communication networks .
it is also important to note that several nodes ( or edges ) in a network could fail at the same time .
for example , all brand x routers in a network could fail at the same time due to a platform dependent computer worm ( virus ) attack . in order to design survivable communication networks , it is essential to consider this kind of homogeneous faults for networks .
existing works on network quality of services have not addressed this issue in detail and there is no existing model to study network reliability in this aspect . in this paper , we use the colored edge graphs which could be used to model homogeneous faults in networks .
the model is then used to optimize the design of survivable networks and to study the minimum connectivity ( and design ) requirements of networks for being robust against homogeneous faults within certain thresholds .
a colored edge graph is a tuple @xmath0 , with @xmath1 the node set , @xmath2 the edge set , @xmath3 the color set , and @xmath4 a map from @xmath2 onto @xmath3 .
the structure @xmath5 be distinct nodes of @xmath6 .
@xmath7 are called @xmath8-_color connected _ for @xmath9 if for any color set @xmath10 of size @xmath11 , there is a path @xmath12 from @xmath13 to @xmath14 in @xmath6 such that the edges on @xmath12 do not contain any color in @xmath15 . a colored edge graph @xmath6 is @xmath8-_color connected _ if and only if for any two nodes @xmath13 and @xmath14 in @xmath6 , they are @xmath8-color connected .
the interpretation of the above definition is as follows . in a network ,
if two edges have the same color , then they could fail at the same time .
this may happen when the two edges are designed with same technologies ( e.g. , with same operating systems , with same application software , with same hardware , or with same hardware and software ) .
if a colored edge network is @xmath8-color connected , then the network communication is robust againt the failure of edges of any @xmath11 colors ( that is , the adversary may tear down any @xmath11 types of devices ) . in practice
, one communication link may be attached to different brands of network devices ( e.g. , routers , modems ) on both sides . for this case ,
the edge can have two different colors .
if any of these colors is broken , the edge is broken .
thus from a reliability viewpoint , if one designs networks with two colors on the same edge , the same reliability / security can be obtained by having only one color on each edge . in the following discussion
, we will only consider the case with one color on each edge .
meanwhile , multiple edges between two nodes are not allowed either .
we are interested in the following practical questions . for a given number @xmath16 of nodes in @xmath1 ( i.e. , the number of network nodes ) , a given number @xmath17 of the colors ( e.g. , the number of network device types ) , and a given number @xmath11 , how can we design a @xmath8-color connected colored edge graphs @xmath18 with minimum number @xmath19 of edges ? in another words , how can we use minimum resources ( e.g. , communication links ) to design a network that will keep working even if @xmath11 types of devices in the network fail ? for practical network designs
, one needs first to have an estimate on the number of homogeneous faults .
for example , the number @xmath11 of brands of routers that could fail at the same time .
then it is sufficient to design a @xmath8-color connected network with @xmath20 colors ( e.g. , with @xmath21 different brands of routers ) .
necessary and sufficient conditions for this kind of network design will be obtained in this paper .
another important issue that should be taken into consideration in practical network designs is that the number @xmath17 of colors ( e.g. , the number of brands for routers ) is quite small .
for example , @xmath17 is normally less than five .
necessary and sufficient conditions for network designs with @xmath22 and with optimized resources will be obtained in this paper .
note that for cases with small @xmath17 , we may have @xmath23 .
the outline of the paper is as follows .
section [ meqtp1sec ] describes the necessary and sufficient conditions for the case of @xmath20 without optimizing the number of edges in the networks .
section [ generalnecsec ] gives a necessary condition for colored edge networks in terms of optimized number of edges .
section [ practiffsec ] shows that the necessary conditions in section [ generalnecsec ] are also sufficient for the most important three cases : ( 1 ) @xmath20 ; ( 2 ) @xmath24 ; and ( 3 ) @xmath22 .
section [ hardsec ] shows that it is * np*-hard to determine whether a given colored edge graph is @xmath8-connected .
though colored - edge graph is a new concept which we used to model network survivability issues , there are related research topics in this field . for example , edge - disjoint ( colorful ) spanning trees have been extensively studied in the literature ( see , e.g. , @xcite ) .
these results are mainly related to our discussion in the next section for the case of @xmath20 .
a colored edge graph @xmath6 is _ proper _ if whenever two edges share an end point they carry different colors . a spanning tree for a colored edge graph is called colorful if no two of its edges have the same color .
two spanning trees of a graph are edge disjoint if they do not share common edges . for a non - negative integer @xmath25 , let @xmath26 denote the complete graph on @xmath25 vertices .
a classical result of euler states that the edges of @xmath27 can be partitioned into @xmath16 isomorphic spanning trees ( paths , for example ) and each of these spanning trees can easily be made colorful , but the resulting edge colored graph usually fails to be proper .
though it is important to design colored edge graphs with required security parameters , for several scenarios it is also important to calculate the robustness of a given colored edge graphs .
roskind and tarjan @xcite designed a greedy algorithm to find @xmath8-edge disjoint spanning trees in a given graph .
this is related to the questions @xmath8-color connectivity for the case of @xmath20 .
we are not aware of any approximate algorithms for deciding @xmath8-color connectivity of a given colored edge graph .
indeed , we will show that this problem is * np*-hard .
in this section , we show necessary and sufficient conditions for some special cases .
[ lemma1s ] a colored edge graphs @xmath0 is @xmath8-color connected if and only if , for all @xmath28 , @xmath29 , @xmath30 , @xmath31 , @xmath32 is a connected graph , where @xmath33 is a partition of @xmath2 under the @xmath17 different colors . as we have mentioned in the previous section , the classical result by euler states that @xmath27 can be partitioned into @xmath16 spanning trees .
thus , by lemma [ lemma1s ] , we have the following theorem .
[ eulerresult ] ( euler ) for @xmath34 , there is a coloration @xmath0 of @xmath35 such that @xmath6 is @xmath36-color connected . in the following ,
we extend theorem [ eulerresult ] to the general case of @xmath37 .
[ lemma2s ] for @xmath37 and @xmath38 , there exists a graph @xmath18 with @xmath39 , and @xmath40 such that the following conditions are satisfied : 1 .
@xmath41 is a connected graph for all @xmath42 ; 2 .
@xmath43 for all @xmath44 .
* we prove the lemma by induction on @xmath16 and @xmath17 . for @xmath45 and @xmath46 ,
the lemma holds obviously .
assume that the lemma holds for @xmath47 . in the following ,
we show that the lemma holds for @xmath48 and for @xmath49 .
let @xmath50 be the graph with @xmath51 , and @xmath52 such that the conditions in the lemma are satisfied : for the case of @xmath53 and @xmath54 , let @xmath55 where @xmath56 is a new node that is not in @xmath57 , and let @xmath58 , @xmath59 , @xmath30 , @xmath60 where @xmath61 are distinct nodes from @xmath57 .
it is straightforward to show that @xmath39 , @xmath41 is a connected graph , and @xmath43 for all @xmath44 .
thus the lemma holds for this case . for the case of @xmath62 and @xmath63 ,
let @xmath64 where @xmath65 are new nodes that are not in @xmath57 , and define @xmath66 as follows .
1 . set @xmath67 and @xmath68 , where @xmath69 is a temporary variable .
2 . define @xmath70 : 1 .
select an edge @xmath71 .
2 . let @xmath72 .
3 . let @xmath73 and @xmath74 .
3 . define @xmath75 for @xmath76 : 1 .
select @xmath77 .
2 . let @xmath78 .
3 . let @xmath79 and @xmath80 .
it is straightforward to show that @xmath81 ( thus @xmath82 ) , @xmath41 is a connected graph , and @xmath43 for all @xmath44 .
this completes the proof of the lemma .
[ iffthmt1 ] given @xmath83 with @xmath20 , there exists a @xmath8-color connected colored edge graphs @xmath0 with @xmath84 and @xmath85 if and only if @xmath37 .
* by lemma [ lemma1s ] , a @xmath8-color connected colored edge graphs @xmath0 with @xmath84 and @xmath86 contains at least @xmath87 edges .
meanwhile , @xmath0 contains at most @xmath88 edges .
thus for @xmath89 , we have @xmath90 . in nother words , for @xmath89
, there is no @xmath8-color connected colored edge graphs @xmath0 with @xmath84 and @xmath86 .
now the theorem follows from lemmas [ lemma1s ] and [ lemma2s ] .
first we note that for a colored edge graph @xmath6 to be @xmath8-color connected , each node must have a degree of at least @xmath21 .
thus the total degree of an @xmath16-node graph should be at least @xmath91 .
this implies the following lemma .
[ nesslemma1 ] for @xmath92 , and a @xmath8-color connected colored edge graph @xmath0 with @xmath84 , @xmath93 , and @xmath85 , we have @xmath94 . in the following , we use
cover free family concepts to study the necessary conditions for colored edge graphs connectivity .
let @xmath95 be a finite set with @xmath96 and @xmath97 be a set of mutually disjoint subsets of @xmath95 with @xmath98 .
then @xmath99 is called a @xmath100-partition of @xmath95 if @xmath101 .
let @xmath102 be positive integers .
an @xmath100-partition @xmath99 is called a @xmath103-cover free family ( or @xmath103-cff(@xmath104 ) ) if , for any @xmath11 elements @xmath105 , we have that @xmath106 it should be noted that our above definition of cover - free family is different from the generalized cover - free family definition for set systems in the literature ( see , e.g. , @xcite ) . in @xcite , a set system @xmath99
is called a @xmath107-cover free family if for any @xmath108 blocks @xmath109 and any @xmath11 blocks @xmath105 , one has @xmath110 .
specifically , there are two major differences between our @xmath111-partition system and the set systems in the literature . 1 . for a set system @xmath99 , @xmath97 may contain repeated elements . 2 .
for a set system @xmath99 , the elements in @xmath97 are not necessarily mutually disjoint .
it is straightforward to show that a colored edge graph @xmath6 is @xmath8-color connected if and only if for any color set @xmath10 of size @xmath11 , after the removal of edges in @xmath6 with colors in @xmath112 , @xmath6 remains connected .
assume that @xmath6 contains @xmath16 nodes .
then a necessary condition for connectivity is that @xmath6 contains at least @xmath113 edges . from this discussion , we get the following lemma . for a colored edge graph @xmath0 , with @xmath84 , @xmath93 , @xmath85 , a necessary condition for @xmath0 to be @xmath8-color connected
is that the @xmath100-partition @xmath99 is a @xmath103-cff(@xmath114 ) with @xmath115 and @xmath116 where @xmath117 . in the following ,
we analyze lower bounds for the number @xmath19 of edges for the existence of a @xmath103-cff(@xmath114 ) . for a set partition @xmath99 and a positive integer @xmath11 , let @xmath118 it is straightforward to see that a @xmath100-partition @xmath99 is a @xmath103-cff(@xmath114 ) if and
only if @xmath119 . given positive integers @xmath120 , let @xmath121 from the above discussion and lemma [ nesslemma1 ] , we have the following theorem .
[ dogthm ] let @xmath120 be given positive integers .
@xmath122 and @xmath94 are necessary conditions for the existence of a @xmath8-color connected colored edge graph @xmath0 , with @xmath84 , @xmath93 , @xmath85 .
[ sufbound ] let @xmath120 be given positive integers .
then we have @xmath123 * proof . * for a given @xmath100-partition @xmath99 , let @xmath124 be an enumeration of elements in @xmath97 such that @xmath125 for all @xmath126 .
it is straightforward to show that @xmath127 .
thus @xmath128 takes the maximum value if @xmath129 is maximized .
it is straightforward to show that this value is maximized when the @xmath100-partition @xmath99 satisfies the following conditions : 1 .
@xmath130 for @xmath131 , and 2 .
@xmath132 for @xmath133 .
the theorem follows from the above discussion . for @xmath134 , and @xmath135 , we have @xmath136
. however , @xmath137 .
this shows that the condition @xmath94 in theorem [ dogthm ] is not redundant .
there are no @xmath8-color connected colored edge graphs @xmath0 for the following special cases : 1 .
2 . @xmath139 .
3 . @xmath140
. * proof . *
before we consider the specific cases , we observe that , when @xmath17 and @xmath11 are fixed , the function @xmath141 is nondecreasing when @xmath19 increases .
in this case , the maximum value that @xmath19 could take is @xmath142 .
thus @xmath143 . that is
, there is no @xmath144-cff(@xmath145 ) , which implies the claim .
note that this result also follows from theorem [ iffthmt1 ] .
\2 . in this case , the maximum value that @xmath19 could take is @xmath146 .
thus @xmath147 .
we only show this for the case @xmath148 . in this case , the maximum value that @xmath19 could take is @xmath149 .
thus @xmath150 .
note that this result also follows from theorem [ iffthmt1 ] .
the following theorem is a variant of theorem [ dogthm ] .
[ necessarybound ] for @xmath151 , a necessary condition for the existence of a @xmath8-color connected colored edge graph @xmath0 with @xmath84 , @xmath93 , and @xmath85 is that @xmath94 and the following conditions are satisfied : * if @xmath152 for some integer @xmath153 , then @xmath154 . * if @xmath155 for some integer @xmath153 , then @xmath156 . *
if @xmath157 for some integer @xmath153 , then @xmath158 .
* @xmath159 * if @xmath160 for some integer @xmath153 , then @xmath161 .
* proof . * for @xmath23 , by theorem [ sufbound ] , we have @xmath162 thus the necessary condition @xmath122 in theorem [ dogthm ] can be interpreted as the following conditions : @xmath163 in other words , for a @xmath8-color connected colored edge graphs @xmath0 , the following @xmath164 conditions ( the disjunction not conjunction ) are satisfied : * @xmath165 , and @xmath85 . * @xmath166 , and @xmath85 . * @xmath159 * @xmath167 , and @xmath85 . by distinguishing the cases for @xmath152 , @xmath155 , @xmath168 , and @xmath160 , and by reorganizing above lines , these necessary conditions can be interpreted as the following @xmath164 conditions : * @xmath152 and @xmath154 for some @xmath153 .
note that this follows from the last line of the above conditions ( one can surely take other lines , but then the value of @xmath19 would be larger ) .
this comment applies to following cases also .
* @xmath155 and @xmath156 for some @xmath153 .
* @xmath157 and @xmath158 for some @xmath153 .
* @xmath159 * @xmath160 and @xmath161 for some @xmath153 .
generally we are interested in the question whether the necessary condition in theorems [ dogthm ] and [ necessarybound ] are also sufficient . in the following ,
we show that this is true for several important practical cases .
[ mistp1 ] the necessary condition in theorem [ dogthm ] is sufficient for the case of @xmath20 .
* proof . *
since @xmath169 is the remainder of @xmath19 divided by @xmath17 , we trivially have @xmath170 .
now assume that @xmath171 . by theorem [ sufbound ] , we have @xmath172 .
the rest follows from theorem [ iffthmt1 ] . before we show that the necessary conditions in theorems [ dogthm ] and [ necessarybound ] are sufficient for the case of @xmath24 , we first present two lemmas whose proofs are straightforward . for @xmath173 and @xmath24 ,
the following @xmath17-node circle graph is @xmath174-color connected : @xmath175 with @xmath176 for @xmath126 and @xmath177 .
[ lemmacircle ] for @xmath24 , @xmath178 , and @xmath179 , the graph in figure [ figcircle ] that is defined in the following is @xmath174-color connected @xmath180 with @xmath181 [ tp1case ] the necessary conditions in theorems [ dogthm ] and [ necessarybound ] are sufficient for the case of @xmath24 .
* for the case of @xmath182 and @xmath24 , it follows from theorem [ mistp1 ] .
now assume that @xmath183 and @xmath24 . in this special case ,
the necessary conditions in theorem [ necessarybound ] is as follows : * @xmath184 and @xmath154 for some @xmath153 .
* @xmath185 and @xmath156 for some @xmath153 .
* @xmath186 and @xmath187 for some @xmath153 .
* @xmath159 * @xmath188 and @xmath161 for some @xmath153 . in the following we first show that the condition `` @xmath185 and @xmath189 '' is sufficient .
let the graph in figure [ figt1basic ] be defined as follows : @xmath190 for each @xmath191 with @xmath192 , let @xmath193 .
then it is straightforward to check that the colored edge graphs @xmath0 is @xmath174-color connected , @xmath194 , and @xmath195 .
now we show that the condition `` @xmath196 and @xmath197 for @xmath198 '' is sufficient .
let @xmath0 be the colored edge graph that we have just constructed with @xmath194 , and @xmath195 .
let @xmath199 .
define a new colored edge graph @xmath200 ( see figure [ figt1case1 ] ) by attaching the following edges to the @xmath17-node circle @xmath201 : @xmath202 the colors for the new edges are defined by letting @xmath203 for @xmath204 and @xmath205 .
it is straightforward to check that @xmath200 is @xmath174-color connected , @xmath206 , and @xmath207 . for @xmath24 and @xmath17 , @xmath208
, there exists an @xmath174-color connected colored edge graph @xmath0 with @xmath84 and @xmath93 if and only if @xmath209 * proof . *
it follows from the proof of theorem [ tp1case ] .
[ m4t2 ] the conditions in theorems [ dogthm ] and [ necessarybound ] are sufficient for the case of @xmath210 .
* proof . *
it is sufficient to show that both of the conditions `` @xmath155 and @xmath189 '' and `` @xmath157 and @xmath158 '' are sufficient ( note that @xmath211 and @xmath135 ) . in the following
we first show that the condition `` @xmath155 and @xmath189 '' is sufficient by induction on @xmath212 . for the case of @xmath213 , we have @xmath214 , and @xmath135 .
let the graph @xmath215 in figure [ figm4t2 ] be defined as @xmath216 where @xmath217 means that the edge @xmath218 takes color @xmath219 .
it is straightforward to check that @xmath215 is @xmath220-color connected . for the case of @xmath221 , we have @xmath222 , and @xmath135 .
let the graph @xmath223 in figure [ fign7m4t2 ] be defined as @xmath224 where @xmath217 means that the edge @xmath218 takes color @xmath219 .
it is straightforward to check that @xmath223 is @xmath220-color connected .
now for @xmath225 ( @xmath226 ) , we have @xmath227 and @xmath228 .
if we glue the @xmath229 node of @xmath230 copies of @xmath215 , we get a @xmath8-color connected colored graph @xmath6 with @xmath231 and @xmath232 .
thus the condition for the case of @xmath225 holds . for @xmath233 ( @xmath226 )
, we have @xmath234 and @xmath235 .
if we glue glue the @xmath229 node of @xmath236 copies of @xmath215 and one copy of @xmath223 , we get a @xmath8-color connected colored graph @xmath6 with @xmath237 and @xmath238 .
thus the condition for the case of @xmath233 holds .
this completes the induction . for the condition `` @xmath157 and @xmath158 ''
, one can add one node to the graph for the case `` @xmath155 and @xmath189 '' with @xmath142 edges ( with distinct colors ) to any three nodes .
the resulting graph meets the requirements .
theorem [ m4t2 ] could be extended to the case of @xmath239 and @xmath240 .
[ met2 ] the conditions in theorems [ dogthm ] and [ necessarybound ] are sufficient for the case of @xmath239 and @xmath240 . * proof .
* it is sufficient to show that both of the conditions `` @xmath155 and @xmath189 '' and `` @xmath157 and @xmath158 '' are sufficient ( note that @xmath241 ) . in the following we first show that the condition `` @xmath242 and @xmath189 '' is sufficient by induction on @xmath212 and @xmath17 .
for @xmath239 and @xmath213 , we have @xmath243 .
the graph in figure [ fign5m5t3 ] shows that the condition is sufficient also . for the case of @xmath221
, we have @xmath244 .
the graph in figure [ fign7m5t3 ] shows that the condition is sufficient also . for @xmath225 ( @xmath226 ) ,
the condition becomes @xmath227 and @xmath245 .
if we glue the @xmath229 node of @xmath230 copies of @xmath246 , we get a @xmath8-color connected colored graph @xmath6 with @xmath231 and @xmath247 .
thus the condition for the case of @xmath225 holds . for @xmath233 ( @xmath226 ) ,
the condition becomes @xmath234 and @xmath248 .
if we glue glue the @xmath229 node of @xmath236 copies of @xmath246 and one copy of @xmath249 , we get a @xmath8-color connected colored graph @xmath6 with @xmath237 and @xmath250 .
thus the condition for the case of @xmath233 holds .
this completes the induction . for the condition `` @xmath157 and @xmath158 ''
, we have @xmath251 and @xmath252 .
we can add one node to the graph for the case `` @xmath155 and @xmath189 '' with @xmath253 edges ( with distinct colors ) to any four nodes .
the resulting graph meets the requirements .
* open questions : * we showed in this section that the conditions in theorems [ dogthm ] and [ necessarybound ] are sufficient for practical cases
. it would be interesting to show that these conditions are also sufficient for general cases .
we leave this as an open question .
we have given necessary and sufficient conditions for @xmath8-color connected colored edge graphs .
sometimes , it is also important to determine whether a given graph is @xmath8-color connected .
unfortunately , the following theorem shows that the problem ceconnect is * conp*-complete .
the ceconnect problem is defined as follows .
before we prove the hardness result , we first introduce the concept of color separator . for a colored edge graph @xmath254 , a color separator for two nodes @xmath13 and @xmath14 of the graph
@xmath6 is a color set @xmath257 such that the removal of all edges with colors in @xmath258 from the graph @xmath6 will disconnect @xmath13 and @xmath14 .
it is straightforward to observe that @xmath13 and @xmath14 are @xmath8-color connected if and only there is no @xmath11-size color separator for @xmath13 and @xmath14
. * proof . *
it is straightforward to show that the problem is in * conp*. thus it is sufficient to show that it is * np*-hard .
the reduction is from the vertex cover problem .
the vc problem is as follows ( definition taken from @xcite ) : question : is there a vertex cover of size @xmath11 or less for @xmath6 , that is , a subset @xmath261 such that @xmath262 and , for each edge @xmath263 , at least one of @xmath56 and @xmath264 belongs to @xmath265 ? for a given instance @xmath259 of vc , we construct a colored edge graph @xmath266 as follows .
first assume that the vertex set @xmath1 is ordered as in @xmath267 .
let @xmath268 in the following , we show that there is a vertex cover of size @xmath11 in @xmath6 if and only if there is a @xmath11-color edge separator for @xmath269
. without loss of generality , assume that @xmath270 is a vertex cover for @xmath6 .
then it is straightforward to show that @xmath271 is a color separator for @xmath269 since each incoming path for @xmath14 in @xmath269 contains two colors corresponding to one edge @xmath272 in @xmath6 .
for the other direction , assume that @xmath273 is a @xmath11-color separator for @xmath269 .
let @xmath274 . by the fact that @xmath258 is a color separator for @xmath269 , for each edge
@xmath275 in @xmath6 , the path @xmath276 in @xmath269 contains at least one color from @xmath258 .
since this path contains only two colors @xmath277 and @xmath278 , we know that @xmath279 or @xmath280 or both belong to @xmath265 . in another word , @xmath265 is a @xmath11-size vertex cover for @xmath6 .
this completes the proof of the theorem .
y. desmedt , y. wang , and m. burmester . a complete characterization of tolerable adversary structures for secure point - to - point transmissions without feedback . in _ proc .
isaac 2005 _ , pages 277287 .
lecture notes in computer science 3827 .
springer verlag 2005 .
h. wang and d. m. blough .
construction of edge - disjoint spanning trees in the torus and application to multicast in wormhole - routed networks . in _ proc .
1999 intl conf . on parallel and
distributed computing systems _
pages 178184 , 1999 . | in this paper , we use the concept of colored edge graphs to model homogeneous faults in networks .
we then use this model to study the minimum connectivity ( and design ) requirements of networks for being robust against homogeneous faults within certain thresholds .
in particular , necessary and sufficient conditions for most interesting cases are obtained .
for example , we will study the following cases : ( 1 ) the number of colors ( or the number of non - homogeneous network device types ) is one more than the homogeneous fault threshold ; ( 2 ) there is only one homogeneous fault ( i.e. , only one color could fail ) ; and ( 3 ) the number of non - homogeneous network device types is less than five . |
grand unified theories ( guts ) @xcite based on the gauge group @xmath0 @xcite are very attractive candidates to unify the strong , weak and electromagnetic forces into a single force , as well as to shed light on some of the open questions of the standard model ( sm ) .
quarks and leptons of each family are unified into a single irreducible representations of @xmath0 group , the 16-dimensional spinor , which also contains the right - handed neutrino .
the presence of the right - handed neutrino makes the seesaw mechanism @xcite for generating small neutrino masses very compelling in these theories .
since @xmath0 gauge symmetry is automatically anomaly - free @xcite , it provides a nice explanation for the miraculous cancelation of anomalies that occurs within each fermion family .
the observed quantization of electric charges is also understood in these theories owing to their non - abelian nature .
unifying all fermions into a single multiplet gives us the hope of understanding some aspects of the flavor puzzle in these theories .
unification of gauge couplings occurs naturally at an energy scale of @xmath10 gev @xcite , as @xmath0 admits an intermediate symmetry group unlike theories based on @xmath11 which must break directly to the sm .
it is of course well known that if supersymmetry is assumed to be present in its minimal version at the tev scale , one - step breaking of @xmath0 directly down to the sm can be realized at an energy scale of @xmath12 gev @xcite .
the focus of this paper is , however , @xmath0 theories without the assumption of supersymmetry .
we wish to inquire what an economic yukawa sector would look like in renormalizable @xmath0 theories .
this may appear to be a well understood issue , but as we suggest here , this question has not been properly resolved .
economy may be viewed as having the least number of higgs fields as well as yukawa parameters while being realistic .
assuming that there are no new fermions beyond the three families of chiral 16s ( or other real representations ) with gut scale masses exist and mix with the chiral 16s , new possibilities are available , see for e.g. , ref .
] the answer to this question may be found in the group theory of fermion bilinears : @xmath13 here the subscripts @xmath14 and @xmath15 stand for symmetric and antisymmetric components ( in family space ) .
the 10 and the 120 are real representations in @xmath0 , while the 126 is complex .
the most general renormalizable yukawa couplings in @xmath0 theories then would take the form @xmath16 here the index @xmath17 takes values @xmath18 where @xmath19 is the number of @xmath1 fields employed , and similarly the index @xmath20 and @xmath21 with @xmath22 and @xmath23 being the number of @xmath2 and @xmath24 present in the theory .
the yukawa coupling matrices @xmath25 and @xmath26 are @xmath27 complex symmetric matrices in family space , while @xmath28 are complex antisymmetric matrices .
we wish to identify the smallest possible set of \{@xmath29 } that would lead to a realistic spectrum of quark and lepton masses as well as mixing angles .
this set will turn out to be the choice @xmath30 , as we shall see .
this result is satisfying , as it suggests that nature has utilized each possible higgs field for fermion mass generation exactly once , without any replication . before establishing this assertion , which will be done in the next section ,
let us note that _ a complex _
@xmath31 can be constructed from _ two real _ 10s in @xmath32 ) : @xmath33 . similarly , _ a complex _
@xmath34 may be constructed from _ two real _ 120s . in these cases ,
the yukawa couplings will involve terms of the type @xmath35 as well as @xmath36 with completely independent yukawa coupling matrices , and similarly for the @xmath34 field .
it is possible to assign a charge exterior to @xmath0 to these fields such as the peccei
quinn @xmath37 motivated on other grounds so that the yukawa couplings contain only the @xmath35 term , and not the @xmath36 term . these restricted class of yukawa couplings in @xmath0 have been studied extensively @xcite . while interesting , the predictions of such models are those of @xmath38 where @xmath39 is a symmetry exterior to @xmath0 , and not of the true grand unified symmetry @xmath0 itself .
our inquiry relates to the minimal yukawa sector in theories where only the @xmath0 gauge symmetry plays a role .
it should be noted that in theories which assume supersymmetry ( susy ) , which is not the focus of the present work , chiral superfields are necessarily complex , thus requiring the complexification of 10 and 120 higgs fields .
holomorphy of the superpotential would imply that the coupling @xmath36 is not present simultaneously with the superpotential term @xmath40 .
these models share some of the features of non - susy models based on @xmath41 , although the renormalization group evolution of the fermion mass parameters between the weak scale and the gut scale would be different in the two classes of theories .
supersymmetric @xmath0 models have been studied extensively , and it has been shown that economic models where only a ( complex ) @xmath1 and a @xmath24 couple to fermions can be predictive and consistent with all fermion masses and mixings @xcite @xcite . if the additional higgs fields needed for symmetry breaking are restricted to a @xmath3 and a @xmath42 , split supersymmetry may be required for consistency @xcite . alternatively , a ( complex ) @xmath2 may be introduced to relax some of the restrictions imposed by the symmetry breaking sector @xcite .
have also been applied in the context of renormalizable susy @xmath0 with some success in explaining the fermion spectrum .
see for example ref .
] our goal in this paper is to identify the analog of the minimal susy @xmath0 yukawa sector , but for @xmath0 theories without supersymmetry .
the rest of the paper is organized as follows . in sec .
2 we present our proof that the economic higgs sector will have @xmath30 . in sec .
3 we analyze the predictions of this model for quark and lepton masses and mixings . here
we present our numerical study which shows full consistency with experimental data . in sec .
4 we present the constraints on these models from the unification of gauge couplings ; in sec .
5 we calculate the proton decay branching ratios . in sec .
6 we conclude .
four appendices contain a discussion on how the fine - tuning is achieved , technical details on the proof of economic yukawa sector as well as the best fit parameters for the minimal model .
in this section we establish the assertion that @xmath30 is the economic choice of yukawa sector in non - supersymmetric @xmath0 theories .
this corresponds to choosing one real @xmath1 , one real @xmath2 and a complex @xmath24 of higgs bosons that have yukawa couplings with the three chiral families of @xmath43 .
an additional higgs filed belonging to @xmath44 , @xmath45 or @xmath42 would be needed for completing the symmetry breaking .
these fields , however , do not have yukawa couplings with the @xmath43 , and the precise choice is not so important for now .
a proof of our assertion would require that the choice @xmath30 leads to a realistic fermion spectrum , and no other simpler choice exists consistent with realism .
the former part of the proof is delegated to sec .
3 where we perform a numerical analysis of this economic yukawa sector ; here we address the latter part .
if only one higgs field among @xmath1 , @xmath2 and @xmath24 is present in a theory , there would be no flavor mixing as the yukawa coupling matrix of this single higgs field can be diagonalized using an @xmath0 rotation .
thus at least two higgs fields are needed for realistic fermion spectrum .
one of the fields used must be a @xmath3 , since it gives large majorana masses to the righ - handed neutrinos directly .
this field also plays a role in the symmetry breaking sector , as it breaks @xmath0 down to @xmath11 , reducing the rank .
one could consider replacing the @xmath3 with a @xmath46 which can play a similar role in rank reduction .
in such a case the right - handed neutrino can acquire a large majorana mass via the two - loop witten diagram @xcite involving gauge boson and scalar loops .
the induced majorana mass can be estimated @xcite to be of order @xmath47 where @xmath48 is the @xmath0 gauge coupling , @xmath49 is the yukawa coupling of @xmath1 , and @xmath50 is the @xmath51 breaking vacuum expectation value ( vev ) of the @xmath46 . in a nonsupersymmetric @xmath0 theory @xmath50 is well below the gut scale for consistency with gauge coupling unification , with its range being @xmath52 gev depending on the surviving intermediate symmetry .
@xmath53 is then of order @xmath54 gev or less , which is too small to reproduce the correct order of magnitude for the light neutrino masses .
the susy particle masses should be of order the gut scale to prevent additional suppression factor of @xmath55 in eq .
( [ two - loop ] ) , which may cause a problem with generating a higgs boson mass of 125 gev @xcite . ] keeping one @xmath3 field in the theory , we seek if a realistic fermion spectrum can be generated with the addition of a second higgs field .
this turns out to be not possible . if the second higgs field is a @xmath3 , the mass relations @xmath56 and @xmath57 will result at the gut scale , which are inconsistent with observations .
the ratio @xmath58 is found to be about 1.7 at the gut scale ( with small input errors ) when the low energy mass parameters are evolved up to the gut scale using sm renormalization group equations .
we found that this ratio is more realistically in the range @xmath59 , when intermediate scale threshold effects arising from the right - handed neutrino sector and the gauge bosons of @xmath60 are included .
each of the two threshold effects causes a decrease in the ratio @xmath58 at the gut scale .
we conclude that the relation @xmath61 is clearly excluded .
the relation @xmath62 is not too far off ( our rge evolution shows the ratio @xmath63 to be about 4 at the gut scale ) , while @xmath64 is off by an order of magnitude . thus a minimal yukawa sector consisting of two copies of @xmath24 is not realistic .
if the second higgs field is a real @xmath1 , two complex symmetric yukawa matrices can be written down , one with the @xmath1 , and one with the @xmath24
. however , the higgs doublet in the @xmath1 is self - conjugate , and is contained in the @xmath65 representation of the pati - salam subgroup @xmath66 .
this field can be written as @xmath67 in general , if the ( 1,1 ) element of @xmath68 is independent from the ( 2,2 ) element , we can denote their respective vacuum expectation values to be @xmath69 and @xmath70 with @xmath69 giving mass to the up - quarks and dirac neutrinos , while @xmath70 generates down - quark and charged lepton masses .
the reality of @xmath1 implies that @xmath71 , and thus the ratio @xmath72 . with @xmath73 ,
the needed splitting between the top and bottom quark masses can not be achieved .
note that @xmath73 is a special case of the general @xmath41 models with @xmath74 .
such models have been studied , which find the phenomenological requirement @xmath75 . a three generation analysis of fermion masses and mixings with a complex @xmath1 in ref .
@xcite shows that a realistic fit requires @xmath76 , which is well outside of the prediction of @xmath73 in the case of real @xmath1 .
thus we conclude that one @xmath24 and one real @xmath1 is not realistic @xcite . what about using one @xmath24 and one @xmath2 ? as shown in ref .
@xcite , this case also can not reproduce fermion masses correctly .
the ratio @xmath77 comes out to be of order one , rather than the phenomenological value of @xmath78 .
in addition , as we shall show , this model predicts the gut scale mass ratio @xmath79 , with any deviation of order @xmath80 .
as already noted , the ratio @xmath81 at the gut scale in @xmath0 models under discussion .
thus we conclude that only two higgs fields being responsible for yukawa couplings can not be realistic .
when three higgs fields are introduced , the choice of one @xmath1 , one @xmath2 and one @xmath24 appears attractive , as there in no replication here .
this choice can indeed lead to a realistic fermion mass spectrum , as we elaborate in the next section .
there would be two complex symmetric yukawa coupling matrices in this case , along with one complex antisymmetric matrix .
if an alternative choice of one @xmath24 and two copies of @xmath2 can lead to a realistic spectrum , that would have less parameters with one symmetric and two antisymmetric yukawa matrices .
however , as we show in appendix b , this choice would lead to the relation @xmath61 with corrections of order 5% , even when one allows for large off - diagonal contributions to the mass matrices from the @xmath2 .
models with one @xmath24 and two copies of @xmath1 would be realistic ; however , these models would have three complex symmetric yukawa matrices which have more parameters compared to the case of one @xmath1 , one @xmath2 and one @xmath24 .
this completes the first part of the proof that @xmath30 is the economic choice for the yukawa sector . to complete the proof we establish in the next section that this choice is indeed realistic .
[ [ section ] ] as argued in the previous section , the minimal yukawa sector of @xmath0 makes use of one real @xmath1 , one real @xmath2 and one complex @xmath24 of higgs bosons that couple to the three families of fermions in the @xmath43 representation . here
we proceed to establish the consistency of such a theory with observed fermion masses and mixings . with no symmetry other than the gauge symmetry of @xmath0 imposed ,
the most general yukawa interactions of the model can be written down as @xmath82 here @xmath49 and @xmath83 are complex symmetric yukawa matrices , while @xmath84 is a complex antisymmetric matrix . under the pati - salam subgroup @xmath85 , these fields decompose as @xmath86 the @xmath1 has one sm doublet higgs field contained in the bidoublet ( 2,2,1 ) , while the @xmath2 has two sm higgs doublets , one each belonging to ( 2,2,1 ) and ( 2,2,15 ) .
the reality condition for the ( 2,2,1 ) from @xmath1 is listed in eq .
( [ real ] ) , while those from the @xmath2 would imply @xmath87 and @xmath88 with the superscripts @xmath89 and @xmath90 denoting the ( 2,2,1 ) and the ( 2,2,15 ) fragments .
the @xmath3 contains two sm higgs fields contained in the complex bidoublet ( 2,2,15 ) fragment , which is not subject to the reality condition .
we denote the up - type and down - type electroweak vevs of the @xmath24 as @xmath91 and @xmath92 respectively .
note also that the @xmath93 fragment of @xmath24 contains a sm singlet field which generates large majorana masses for the right - handed neutrinos once it acquires a vev .
the up - quark , down - quark , charged leptons , dirac neutrino and majorana neutrino mass matrices derived from eq .
( [ yukawa])can be now written down : @xmath94 now defining @xmath95 and going into a phase convention where @xmath96 is real ( this can be done by an @xmath97 rotation ) , we get @xmath98 these matrices are written in a basis @xmath99 .
the light neutrino mass matrix , obtained from the see - saw formula , is given by @xmath100 without loss of generality one can choose a basis where @xmath101 is real , positive and diagonal .
in this basis , @xmath101 would have 3 real parameters while @xmath102 has 6 complex parameters . since the matrix @xmath103 is antisymmetric
, it has 3 complex parameters .
there are 4 additional complex parameters in @xmath104 and one phase @xmath105 .
an overall phase either from @xmath106 or @xmath107 will be irrelevant in the matrix @xmath108 .
altogether there are then 16 real parameters and 13 phases . with these parameters
one should fit 18 observables : 6 quark masses , 3 quark mixing angles , 1 ckm phase , 3 charged lepton masses , 2 neutrino mass squared differences , and 3 mixing angles in the neutrino sector .
if we assume dominance of either type - i or type - ii seesaw , then the parameter set is reduced by 1 magnitude and 1 phase .
although the number of model parameters is larger than the number of observables , it is nontrivial to find an acceptable fit owing to the fact that 12 or 13 parameters are phases which can not be manipulated much . the type - ii contribution to the light neutrino mass matrix originates in the model from terms such as @xmath109 in the scalar potential .
when decomposed into the the pati - salam symmetry group , this term would contain terms of the type @xmath110 .
when the singlet vev of @xmath111 and the doublet vev of @xmath65 are inserted in this term , a linear term in @xmath112 would result , which leads to an induced vev for its neutral component : @xmath113 .
we note that with the right - handed neutrino mass given as in , the mass of the @xmath114 gauge bosons which are outside of @xmath11 but mediate proton decay is given as @xmath115 , where @xmath116 is the @xmath0 gauge coupling . in this section
we discuss the procedure we follow for the numerical analysis to the fermion masses and mixings and present our fit results . for optimization purpose
we do a @xmath117-analysis .
the pull and @xmath117-function are defined as : @xmath118 where @xmath119 represent experimental 1@xmath120 uncertainty and @xmath121 , @xmath122 and @xmath123 represent the theoretical prediction , experimental central value and pull of observable @xmath17 .
we fit the values of the observables at the gut scale , @xmath124 gev . to get the gut scale values of the observables we take the central values at the @xmath125 scale from table-1 of ref . @xcite . with this input
we do the renormalization group equation ( rge ) running of the yukawa couplings @xcite and the ckm parameters @xcite within the sm up to the gut scale .
for the associated one sigma uncertainties of the observables at the gut scale , we keep the same percentage uncertainty with respect to the central value of each quantity as that at the @xmath125 scale . for the charged lepton masses , a relative uncertainty of 0.1@xmath126
is assumed in order to take into account the theoretical uncertainties arising for example from threshold effects .
all these inputs are presented in table [ tab:04 ] .
the rge running factors for the yukawa couplings @xmath127 and for the ckm mixing angles @xmath128 are taken to be : @xmath129 the low scale inputs as shown in table [ nu ] in the neutrino sector are taken from ref .
@xcite . for neutrino observables , we run the rge for the dimension five operator from low scale to the @xmath50 scale @xcite and use these new values during the fitting produce . for this running purpose
, we have assumed hierarchical structure of the neutrinos and used the approximations @xmath130 and @xmath131 .
the running values of the observables at the high scale depend on the scale @xmath50 , this is why we present the neutrino mass squared differences resulting from running in table [ tab:01 ] at the relevant scale @xmath50 corresponding to two different fits ( type - i dominance and type - i+ii case ) , while all the other inputs are at @xmath132 gev . in @xmath0 gut models such as the one we are considering ,
the ( 3,3 ) entry of the dirac neutrino yukawa coupling matrix @xmath133 is expected to be of the order of unity , and thus rge corrections proportional to @xmath133 can be important in the momentum range @xmath134 . this effect could have a sizeable contribution to the tau lepton mass only , since for the first and second generation dirac yukawa couplings turn out to be small . including this effect of the heavy right - handed neutrinos thresholds ,
the dirac neutrino mass matrix gets modified at the gut scale as @xmath135 m_{\nu_d}.\end{aligned}\ ] ] while the modified charged lepton mass matrix becomes @xmath136 m_{e } .\end{aligned}\ ] ] to be clear , the tau lepton mass decreases in going from the @xmath137 mass scale to the gut scale due to the dirac neutrino yukawa correction . in the fitting procedure it was thus @xmath138 from eq . to be compared to the experimental values at @xmath139 , while in @xmath140 has been replaced by @xmath141 .
notice that @xmath142 and @xmath143 in eqs . and
are defined in eqs . and .
we investigate three different scenarios , type - i dominance , type - ii dominance and the general scenario where both contributions are present , type - i+ii .
the fit results corresponding to our numerical analysis is presented in table [ tab:01 ] .
we found good solutions for both type - i and type - i+ii with total @xmath144 0.45 and 0.004 respectively , but not for type - ii scenario ( the total @xmath145 in this case ) .
for the type - ii case , our numerical analysis shows that , for the best fit , the worst fitted quantity corresponds to @xmath146 that comes out to be @xmath147 times smaller ( with pull @xmath148 -32 ) than the experimental data .
the other discrepancy is of the quantity @xmath149 that is @xmath1481.5 times smaller compared to the experimental central value . with these fit results
the predictions of the model for these two scenarios are listed in table [ tab:02 ] .
the parameter set corresponding to these best fit results are presented in appendices c and d for type - i and type - i+ii cases respectively .
we conclude that the model gives an excellent fit to all observables in the fermion sector .
this completes our proof of the minimality of the yukawa sector in @xmath0 models . .
observables in the neutrino sector used in our fits taken from ref .
[ cols="^,^",options="header " , ]
as is well known , the three gauge couplings of the sm do not unify at a common scale .
@xmath0 models provide a way to achieve coupling unification by virtue of an intermediate scale . in our proposed framework ,
the first stage of symmetry breaking can be achieved by employing a real @xmath44 , or a real @xmath45 or a real @xmath42 , along with a complex @xmath3 .
employing @xmath44 higgs would require relying on the quantum corrections in the higgs potential @xcite , while there is no such problem with the use of @xmath42 . in both cases the discrete d parity symmetry
would be broken at the gut scale @xcite .
the intermediate gauge symmetry may be @xmath66 when a @xmath42 is used , while it is @xmath150 if the @xmath44 is used .
alternatively , a @xmath45 can break @xmath0 down to pati - salam symmetry preserving d parity . in this case
the unification scale tends to be lower , of order @xmath151 gev , if threshold effects arising from the scalar multiplets are ignored .
this can potentially be in conflict with proton decay limits .
it has been recently shown in ref .
@xcite that symmetry breaking with a @xmath45 and a @xmath3 can lead to higher values of @xmath152 consistent with proton lifetime , when threshold effects are properly included . here
we present for completeness our results on the unification of gauge couplings assuming the intermediate symmetry to be @xmath66 with or without d parity . since breaking @xmath0 gauge group by @xmath45 higgs preserves the discrete parity , that demands the equality of the @xmath97 and @xmath153 gauge couplings ( @xmath154 ) at the ps intermediate scale .
the low energy data completely determines the value of this scale as well as the gut scale with the assumption of survival hypothesis @xcite .
the one - loop beta function coefficients for the evolution of the @xmath37 , @xmath97 and @xmath155 gauge couplings are @xmath156 @xcite . to determine the intermediate scale , we use the low energy values from ref.@xcite : @xmath157 , @xmath158 and @xmath159 ( only the central values are quoted here ) .
then from the intermediate scale to the gut scale we run the rges with one - loop coefficients @xmath160 for the group @xmath161 that determines the gut scale .
the existence of the multiplets @xmath162 , @xmath163 and @xmath164 is assumed at the intermediate scale while the rest of the multiplets are assumed to have gut scale mass following the survival hypothesis .
one - loop running of the rges of the gauge couplings are shown in fig .
[ uni0 ] ( left plot ) . model .
the left plot corresponds to the case when the gut symmetry is broken by @xmath45 higgs that leaves the discrete symmetry @xmath154 unbroken .
the right plot is for the case when @xmath45 is replaced by @xmath42 higgs that does not preserve the discrete symmetry .
, title="fig : " ] model .
the left plot corresponds to the case when the gut symmetry is broken by @xmath45 higgs that leaves the discrete symmetry @xmath154 unbroken .
the right plot is for the case when @xmath45 is replaced by @xmath42 higgs that does not preserve the discrete symmetry .
, title="fig : " ] from this fig .
[ uni0 ] one sees that the gut scale is @xmath165 gev , which is about a factor of 2.5 smaller compared to what is needed to save the theory from the experimental proton decay limit @xmath166 yrs @xcite .
certainly the assumption made that all scalar particles have a common mass at the assumed scale is too restrictive : the higgs multiplets are likely to have non - degenerate mass spectrum with masses scattered around each scale under consideration .
if one includes this threshold correction arising from the higgses , the unification scale can be raised as shown in an explicit calculation in ref .
there is no strict guideline , however , on how much the mass spectrum may be scattered ; this would lead to significant uncertainty in proton lifetime estimate .
as we show in the next section , the branching ratios for proton decay are much more stable and can be used to test these theories .
if instead of a @xmath45 a @xmath42 is used to break the gut symmetry , then the unification scale is naturally raised to about @xmath167 gev .
this is because d parity is broken by the vev of @xmath42 , and as a result , with the assumption of survival hypothesis , the intermediate scale scalar spectrum is left - right asymmetric .
although @xmath42 allows for other intermediate symmetries , here we focus on the pati - salam symmetry . the gauge coupling evolution with the ps intermediate symmetry
is presented in fig .
[ uni0 ] ( right plot ) . following survival hypothesis
, we consider the multiplets @xmath162 and @xmath168 at the intermediate scale with the rest of the multiplets lying at the gut scale . with these multiplets ,
the one - loop rge coefficients are @xmath169 for the group @xmath161 .
this plot clearly shows that the gut scale can be raised by about an order of magnitude compared to the @xmath45 scenario and one does not need to rely on the threshold correction to save the theory from rapid proton decay .
it should be noted that the scenario with @xmath42 has a drawback that the intermediate scale is relatively low @xmath170 gev , which does not fit the right - handed neutrino mass spectrum as well as the @xmath45 model where this scale is around @xmath171 gev .
a look at the heaviest right - handed neutrino mass from table [ tab:02 ] suggests that the case of type - i seesaw prefers symmetry breaking by a @xmath45 , while the type i + type ii scenario can accommodate breaking by a @xmath42 .
in non - susy @xmath0 models , proton decay mediated by the gauge bosons are the most important . the lifetime of the proton is extremely sensitive to the superheavy gauge bosons masses ( @xmath172 ) since the lifetime goes as @xmath173 , where @xmath174 is the proton mass and @xmath116 is the unified gauge coupling . as noted in the previous section , there is a large uncertainty in the determination of @xmath175 from low energy data , owing to unknown high scale threshold effects . on the other hand ,
proton decay branching ratios are less sensitive to these threshold effects , and so we focus on the predictions of the model for branching ratios .
the gauge bosons of @xmath0 belong to the adjoint @xmath176 .
the decomposition of this field under the sm gauge symmetry is given by : @xmath177 the gauge bosons responsible for proton decay are the @xmath178 and @xmath179 .
the gauge interaction lagrangian of these bosons with the fermions in the current eigenstate basis is given by @xcite : @xmath180 + [ \overline{\nu}_l\gamma^{\mu}\overline{y}^{i}_{\mu}d^c_{li } ] + [ \overline{d}_{l i}\gamma^{\mu}\overline{x}^{i}_{\mu}e^c_{l } + \epsilon_{ijk}\overline{u}^{ck}_l\gamma^{\mu}\overline{x}^{i}_{\mu } u^{j}_l ] \nonumber \\ & + [ -\overline{u}_{l i}\gamma^{\mu}\overline{y}^{i}_{\mu}e^c_{l } + \epsilon_{ijk}\overline{u}^{ck}_l\gamma^{\mu}\overline{y}^{i}_{\mu } d^{j}_l ] + [ -\epsilon_{ijk}\overline{d}^{ck}_l\gamma^{\mu}\overline{x}^{\prime i}_{\mu } d^{j}_l ] + [ -\overline{u}_{r i}\gamma^{\mu}\overline{x}^{\prime i}_{\mu } \nu^c_r ] + [ -\overline{u}_{l i}\gamma^{\mu}\overline{x}^{\prime i}_{\mu } \nu^c_l ] \nonumber \\ & + [ \epsilon_{ijk}\overline{d}^{c k}_l\gamma^{\mu}\overline{y}^{\prime i}_{\mu } u^{j}_l ] + [ -\overline{u}_{r i}\gamma^{\mu}\overline{y}^{\prime i}_{\mu } e^+_r ] + [ -\overline{d}_{l i}\gamma^{\mu}\overline{y}^{\prime i}_{\mu } \nu^c_l ] + h.c . \},\end{aligned}\ ] ] where @xmath181 are color indices and we have suppressed the family indices and @xmath97 indices .
the resulting @xmath182 effective operators of the form @xmath183 responsible for proton decay can be constructed from this lagrangian @xcite : @xmath184 here , @xmath185 and @xmath186 , @xmath187 and @xmath188 .
the indices @xmath181 are color indices , @xmath189 are family indices and @xmath190 are @xmath97 indices . in the physical basis
these operators will be modified as : @xmath191 where @xmath192 ; \label{cia } \\
c(e_\alpha , d^c_\beta ) & = k_1 ^ 2 v_1^{11 } v_3^{\beta \alpha}+k_2 ^ 2 \left(v_4v^\dagger_{ud}\right)^{\beta 1 } \left ( v_1 v_{ud } v_4^\dagger v_3\right)^{1 \alpha } ; \\
c(\nu_l , d_\alpha , d_\beta^c)&=k_1 ^ 2 \left ( v_1 v_{ud}\right)^{1\alpha}\left ( v_3 v_{en}\right)^{\beta l } + k_2 ^ 2 v_4^{\beta \alpha } \left ( v_1 v_{ud } v_4^\dagger v_3 v_{en}\right)^{1l};\\ c(\nu_l^c , d_\alpha , d_\beta^c)&= k_2 ^ 2 \left[\left(v_4 v_{ud}^\dagger \right)^{\beta 1 } \left(u_{en}^\dagger v_2\right)^{l \alpha } + v_4^{\beta \alpha } \left(u_{en}^\dagger v_2 v_{ud}^\dagger\right)^{l 1}\right ] ; \alpha=\beta \neq 2 . \label{cib}\end{aligned}\ ] ] in the above @xmath193 etc are mixing matrices defined so that @xmath194 , @xmath195 , @xmath196 , @xmath197 , @xmath198 , @xmath199 and @xmath200 , where @xmath201 define the diagonalizing matrices given by @xmath202 then the partial decay width of the decay @xmath203 ( @xmath204 , @xmath205 and @xmath206 is anti - lepton ) is given by : @xmath207 ^ 2 \left| \sum_i c^i w^i_0(n\to p ) \right|^2 , \end{aligned}\ ] ] where the coefficients @xmath208 are given in eqs . - and the relevant form factors @xmath209 are obtained by using lattice qcd computations @xcite : @xmath210 in table [ tab:03 ] we present the @xmath182 proton decay branching ratios calculated for our best fit parameter sets .
we find that the two dominant modes are @xmath211 and @xmath212 .
a comparison of these modes with those of more general @xmath182 proton decay studies @xcite shows similarity .
the near dominance of the @xmath213 mode may be taken as a test of the yukawa sector presented here .
in this work , we have presented an economic yukawa sector for @xmath0 models .
the main feature of this construction is that only the @xmath0 symmetry is used to constrain the yukawa parameters .
the higgs system consists of a _ real _ @xmath1 , a _ real _ @xmath2 and a complex @xmath24 that have yukawa couplings . in most nonsupersymmetric @xmath0 models in the literature symmetries outside of @xmath0 such as a peccei - quinn @xmath37
are used to constrain the yukawa sector .
that would require the complexification of the _ real _ @xmath1 and _ real _ @xmath2 .
the model presented here deviates from this , and yet is quite constraining .
we showed that , with a limited number of yukawa parameters , a good fit to all fermion masses and mixings , including the neutrino sector is possible .
once the flavor sector is fixed , we are able to calculate the proton decay branching ratios .
the dominant decays of the proton are found to be @xmath214 and @xmath215 , which may provide partial tests of the model .
this work is supported in part by the u.s .
department of energy grant no .
de - sc0010108 ( k.s.b and s.s ) and by the slovenian research agency ( b.b . ) .
the authors wish to thank barbara szczerbinska and the cetup * 2016 workshop for hospitality and for providing a stimulating environment for discussions .
numerical calculations are performed with cowboy supercomputer of the high performance computing center at oklahoma state university ( nsf grant no .
oci-1126330 ) .
as is well known , any grand unified theory has to address the question of making one higgs doublet light , while its color triplet gut partner remains superheavy so as to not cause rapid proton decay .
this doublet - triplet splitting problem is present in both susy and non - susy minimal guts .
if the higgs doublet mass is not split from the color triplet mass , either the electroweak symmetry would break at the gut scale , or not break at all , or the light color - triplet would lead to far too fast proton decay .
a fine - tuning is necessary to bring the higgs doublet mass down to the weak scale . in supersymmetric versions , this fine - tuning
is done at the tree level , susy would guarantee its stability against quantum corrections . in non - supersymmetric @xmath0 theories ,
the tuning must be done after taking account of loop corrections to a very high order .
the induced higgs mass from quantum loops would be at the @xmath216-loop level of order @xmath217 . for @xmath218 gev
, the tuning must be done after @xmath219 loop corrections are included .
we note that this is nevertheless only one fine - tuning , albeit not easily enforceable by actual calculations .
the hermitian higgs doublet mass matrix is tuned to have near - zero determinant .
in contrast , in minimal susy guts , the needed tree level tuning requires the determinant of the complex doublet higgsino mass matrix to be near zero .
recall that in susy all mass parameters in the superpotential are complex in general .
such a tuning amounts to two conditions , unlike the non - supersymmetric tuning , which requires only one such condition .
although the higgs mass can be ensured only after including very high order corrections in non - susy @xmath0 , we find it intriguing that the fine - tuning condition is more minimal here compared to minimal susy @xmath0 .
without loss of generality , we can diagonalize the yukawa coupling of @xmath3 .
we focus on the second and third generation down quarks and charged leptons .
their mass matrices can be written down as @xmath220 this form persists even when many of @xmath2 fields are used , with their mass contribution going into the off - diagonal entries differently in @xmath221 and @xmath142 .
the exact invariants of these matrices are then @xmath222 from these relations it follows that @xmath223~.\ ] ] the undetermined parameter @xmath224 is bounded by eq .
( [ app ] ) , and no matter how we vary @xmath224 , the deviation from 3 in the ratio @xmath58 is of order @xmath80 . the inclusion of the first family is not expected to change considerably this result .
this proves that a higgs sector consisting of one @xmath3 and two or any number of copies of real @xmath2 can not lead to realistic fermion masses .
in this appendix we present the parameter set corresponding to the @xmath225 best fit for the type - i scenario .. for both the type - i and type - i+ii scenarios , one must keep all the significant digits of the parameters presented in these appendices . this high level of accuracy
is needed to reproduce the neutrino observables ; it is due to the fact that the right - handed neutrino mass spectrum in both cases shows extreme hierarchy among the generations , see table .
[ tab:02 ] . since this hierarchy between the first and the second generations is extreme , chopping - off digits effects mainly the quantity @xmath146 . ]
@xmath226 @xmath227 @xmath228 @xmath229
in this appendix we present the parameter set corresponding to the @xmath225 best fit for the type - i+ii scenario .
j. c. pati and a. salam , `` lepton number as the fourth color , '' phys . rev .
d * 10 * , 275 ( 1974 ) erratum : [ phys . rev .
d * 11 * , 703 ( 1975 ) ] .
h. georgi and s. l. glashow , `` unity of all elementary particle forces , '' phys .
* 32 * , 438 ( 1974 ) .
h. georgi , h. r. quinn and s. weinberg , `` hierarchy of interactions in unified gauge theories , '' phys .
lett . * 33 * , 451 ( 1974 ) .
h. georgi , in particles and fields ( edited by c.e .
carlson ) , a.i.p .
( 1975 ) ; h. fritzsch and p. minkowski , `` unified interactions of leptons and hadrons , '' annals phys .
* 93 * ( 1975 ) 193 .
p. minkowski , phys .
b * 67 * ( 1977 ) 421 ; t. yanagida , proceedings of the _ workshop on unified theories and baryon number in the universe _ , tsukuba , 1979 , eds .
a. sawada , a. sugamoto ; s. glashow , in _ cargese 1979 , proceedings , quarks and leptons _ ( 1979 ) ; m. gell - mann , p. ramond , r. slansky , proceedings of the _ supergravity stony brook workshop _ , new york , 1979 , eds .
p. van niewenhuizen , d. freeman ; r. mohapatra , g. senjanovi ' c , phys.rev.lett .
* 44 * ( 1980 ) 912 .
j. schechter and j. w. f. valle , `` neutrino masses in su(2 ) x u(1 ) theories , '' phys .
d * 22 * , 2227 ( 1980 ) ; g. lazarides , q. shafi and c. wetterich , `` proton lifetime and fermion masses in an so(10 ) model , '' nucl .
b * 181 * , 287 ( 1981 ) ; r. n. mohapatra and g. senjanovi ' c , `` neutrino masses and mixings in gauge models with spontaneous parity violation , '' phys .
d * 23 * , 165 ( 1981 ) ; c. wetterich , `` neutrino masses and the scale of b - l violation , '' nucl .
b * 187 * , 343 ( 1981 ) ; j. schechter and j. w. f. valle , `` neutrino decay and spontaneous violation of lepton number , '' phys .
d * 25 * , 774 ( 1982 ) .
t. g. rizzo and g. senjanovi ' c , `` can there be low intermediate mass scales in grand unified theories ?
, '' phys .
lett . *
46 * , 1315 ( 1981 ) ; t. g. rizzo and g. senjanovi ' c , `` grand unification and parity restoration at low - energies .
phenomenology , '' phys .
d * 24 * , 704 ( 1981 ) [ phys .
d * 25 * , 1447 ( 1982 ) ] ; t. g. rizzo and g. senjanovi ' c , `` grand unification and parity restoration at low - energies .
unification constraints , '' phys .
d * 25 * , 235 ( 1982 ) .
w. e. caswell , j. milutinovi ' c , and g. senjanovi ' c , `` predictions of left - right symmetric grand unified theories , '' phys .
d * 26 * , 161 ( 1982 ) .
d. chang , r. n. mohapatra and m. k. parida , `` decoupling parity and su(2)-r breaking scales : a new approach to left - right symmetric models , '' phys .
lett . * 52 * , 1072 ( 1984 ) .
j. m. gipson and r. e. marshak , `` intermediate mass scales in the new so(10 ) grand unification in the one loop approximation , '' phys .
d * 31 * , 1705 ( 1985 ) .
d. chang , r. n. mohapatra , j. gipson , r. e. marshak and m. k. parida , `` experimental tests of new so(10 ) grand unification , '' phys .
d * 31 * , 1718 ( 1985 ) .
n. g. deshpande , e. keith and p. b. pal , `` implications of lep results for so(10 ) grand unification , '' phys .
d * 46 * , 2261 ( 1993 ) .
n. g. deshpande , e. keith and p. b. pal , `` implications of lep results for so(10 ) grand unification with two intermediate stages , '' phys . rev .
d * 47 * , 2892 ( 1993 ) [ hep - ph/9211232 ] .
s. bertolini , l. di luzio and m. malinsk , `` intermediate mass scales in the non - supersymmetric so(10 ) grand unification : a reappraisal , '' phys .
d * 80 * , 015013 ( 2009 ) [ arxiv:0903.4049 [ hep - ph ] ] .
s. bertolini , l. di luzio and m. malinsk , `` on the vacuum of the minimal nonsupersymmetric so(10 ) unification , '' phys .
d * 81 * , 035015 ( 2010 ) [ arxiv:0912.1796 [ hep - ph ] ] .
s. bertolini , l. di luzio and m. malinsk , `` the quantum vacuum of the minimal so(10 ) gut , '' j. phys .
ser . * 259 * , 012098 ( 2010 ) [ arxiv:1010.0338 [ hep - ph ] ] .
k. s. babu and s. khan , `` minimal nonsupersymmetric @xmath0 model : gauge coupling unification , proton decay , and fermion masses , '' phys .
d * 92 * , no . 7 , 075018 ( 2015 ) [ arxiv:1507.06712 [ hep - ph ] ] .
l. grf , m. malinsk , t. mede and v. susi , `` one - loop pseudo - goldstone masses in the minimal @xmath0 higgs model , '' arxiv:1611.01021 [ hep - ph ] .
s. dimopoulos , s. raby and f. wilczek , `` supersymmetry and the scale of unification , '' phys .
d * 24 * , 1681 ( 1981 ) ; l. e. ibez and g. g. ross , `` low - energy predictions in supersymmetric grand unified theories , '' phys .
b * 105 * , 439 ( 1981 ) ; m. b. einhorn and d. r. t. jones , `` the weak mixing angle and unification mass in supersymmetric su(5 ) , '' nucl .
b * 196 * , 475 ( 1982 ) ; w. j. marciano and g. senjanovi ' c , `` predictions of supersymmetric grand unified theories , '' phys . rev .
d * 25 * , 3092 ( 1982 )
. k. s. babu and s. m. barr , `` an so(10 ) solution to the puzzle of quark and lepton masses , '' phys .
rev . lett .
* 75 * , 2088 ( 1995 ) [ hep - ph/9503215 ] .
k. s. babu , b. bajc and s. saad , `` new class of so(10 ) models for flavor , '' phys .
d * 94 * , no . 1 , 015030 ( 2016 ) [ arxiv:1605.05116 [ hep - ph ] ] .
g. lazarides , q. shafi and c. wetterich , `` proton lifetime and fermion masses in an so(10 ) model , '' nucl .
b * 181 * , 287 ( 1981 ) .
a. davidson , v. p. nair and k. c. wali , `` mixing angles and cp violation in the @xmath234 model , '' phys .
d * 29 * , 1513 ( 1984 ) .
g. lazarides and q. shafi , `` fermion masses and mixings in so(10 ) , '' nucl .
b * 350 * , 179 ( 1991 ) . k. s. babu and r. n. mohapatra , `` predictive neutrino spectrum in minimal so(10 ) grand unification , '' phys .
lett . * 70 * , 2845 ( 1993 ) [ hep - ph/9209215 ] .
d. g. lee and r. n. mohapatra , `` an so(10 ) x s(4 ) scenario for naturally degenerate neutrinos , '' phys .
b * 329 * , 463 ( 1994 ) [ hep - ph/9403201 ] .
b. bajc , a. melfo , g. senjanovi ' c and f. vissani , `` yukawa sector in non - supersymmetric renormalizable so(10 ) , '' phys .
d * 73 * , 055001 ( 2006 ) [ hep - ph/0510139 ] .
a. s. joshipura and k. m. patel , `` fermion masses in so(10 ) models , '' phys .
d * 83 * , 095002 ( 2011 ) [ arxiv:1102.5148 [ hep - ph ] ] .
t. fukuyama and n. okada , `` neutrino oscillation data versus minimal supersymmetric so(10 ) model , '' jhep * 0211 * , 011 ( 2002 ) [ hep - ph/0205066 ] .
b. bajc , g. senjanovi ' c and f. vissani , `` b - tau unification and large atmospheric mixing : a case for noncanonical seesaw , '' phys .
* 90 * , 051802 ( 2003 ) [ hep - ph/0210207 ] .
h. s. goh , r. n. mohapatra and s. p. ng , `` minimal susy so(10 ) , b tau unification and large neutrino mixings , '' phys .
b * 570 * , 215 ( 2003 ) [ hep - ph/0303055 ] .
h. s. goh , r. n. mohapatra and s. p. ng , `` minimal susy so(10 ) model and predictions for neutrino mixings and leptonic cp violation , '' phys .
d * 68 * , 115008 ( 2003 ) [ hep - ph/0308197 ] .
k. s. babu and c. macesanu , `` neutrino masses and mixings in a minimal so(10 ) model , '' phys .
d * 72 * , 115003 ( 2005 ) [ hep - ph/0505200 ] .
s. bertolini , t. schwetz and m. malinsk , `` fermion masses and mixings in so(10 ) models and the neutrino challenge to susy guts , '' phys .
d * 73 * , 115012 ( 2006 ) [ hep - ph/0605006 ] .
b. bajc , i. dorner and m. nemevek , `` minimal so(10 ) splits supersymmetry , '' jhep * 0811 * ( 2008 ) 007 [ arxiv:0809.1069 [ hep - ph ] ] .
t. fukuyama , k. ichikawa and y. mimura , `` revisiting fermion mass and mixing fits in the minimal susy @xmath0 gut , '' phys .
d * 94 * , no . 7 , 075018 ( 2016 ) [ arxiv:1508.07078 [ hep - ph ] ] .
s. bertolini , m. frigerio and m. malinsk , `` fermion masses in susy so(10 ) with type ii seesaw : a non - minimal predictive scenario , '' phys .
d * 70 * , 095002 ( 2004 ) [ hep - ph/0406117 ] .
w. m. yang and z. g. wang , `` fermion masses and flavor mixing in a supersymmetric so(10 ) model , '' nucl .
b * 707 * , 87 ( 2005 ) [ hep - ph/0406221 ] .
b. dutta , y. mimura and r. n. mohapatra , `` suppressing proton decay in the minimal so(10 ) model , '' phys .
* 94 * , 091804 ( 2005 ) [ hep - ph/0412105 ] . c. s. aulakh and s. k. garg , `` the new minimal supersymmetric gut : spectra , rg analysis and fermion fits , '' nucl .
b * 857 * , 101 ( 2012 ) [ arxiv:0807.0917 [ hep - ph ] ] .
m. c. chen and k. t. mahanthappa , `` fermion masses and mixing and cp violation in so(10 ) models with family symmetries , '' int .
j. mod .
phys . a * 18 * , 5819 ( 2003 ) [ hep - ph/0305088 ] .
w. grimus and h. kuhbock , `` fermion masses and mixings in a renormalizable @xmath235 gut , '' phys .
b * 643 * , 182 ( 2006 ) [ hep - ph/0607197 ] .
y. cai and h. b. yu , `` a so(10 ) gut model with s4 flavor symmetry , '' phys . rev .
d * 74 * , 115005 ( 2006 ) [ hep - ph/0608022 ] .
a. albaid , `` flavor violation in a minimal @xmath236 susy gut , '' int .
j. mod .
a * 27 * , 1250005 ( 2012 ) [ arxiv:1106.4070 [ hep - ph ] ] .
p. m. ferreira , w. grimus , d. juriukonis and l. lavoura , `` flavour symmetries in a renormalizable so(10 ) model , '' nucl .
b * 906 * , 289 ( 2016 ) [ arxiv:1510.02641 [ hep - ph ] ] .
e. witten , `` neutrino masses in the minimal o(10 ) theory , '' phys .
* 91b * , 81 ( 1980 ) .
g. f. giudice and a. strumia , `` probing high - scale and split supersymmetry with higgs mass measurements , '' nucl .
b * 858 * , 63 ( 2012 ) [ arxiv:1108.6077 [ hep - ph ] ] .
s. antusch and v. maurer , `` running quark and lepton parameters at various scales , '' jhep * 1311 * , 115 ( 2013 ) [ arxiv:1306.6879 [ hep - ph ] ] .
m. e. machacek and m. t. vaughn , `` two loop renormalization group equations in a general quantum field theory .
yukawa couplings , '' nucl .
b * 236 * , 221 ( 1984 ) ; h. arason , d. j. castao , b. keszthelyi , s. mikaelian , e. j. piard , p. ramond and b. d. wright , `` renormalization group study of the standard model and its extensions .
1 . the standard model , '' phys . rev .
d * 46 * , 3945 ( 1992 ) .
k. s. babu , `` renormalization group analysis of the kobayashi - maskawa matrix , '' z. phys .
c * 35 * , 69 ( 1987 ) .
g. l. fogli , e. lisi , a. marrone , d. montanino , a. palazzo and a. m. rotunno , `` global analysis of neutrino masses , mixings and phases : entering the era of leptonic cp violation searches , '' phys .
d * 86 * , 013012 ( 2012 ) [ arxiv:1205.5254 [ hep - ph ] ] .
k. s. babu , c. n. leung and j. t. pantaleone , `` renormalization of the neutrino mass operator , '' phys .
b * 319 * , 191 ( 1993 ) [ hep - ph/9309223 ] ; s. antusch , m. drees , j. kersten , m. lindner and m. ratz , `` neutrino mass operator renormalization revisited , '' phys .
b * 519 * , 238 ( 2001 ) [ hep - ph/0108005 ] .
h. georgi , `` towards a grand unified theory of flavor , '' nucl .
b * 156 * , 126 ( 1979 ) .
f. del aguila and l. e. ibez , `` higgs bosons in so(10 ) and partial unification , '' nucl .
b * 177 * , 60 ( 1981 ) .
r. n. mohapatra and g. senjanovi ' c , `` higgs boson effects in grand unified theories , '' phys .
d * 27 * , 1601 ( 1983 ) .
d. r. t. jones , `` the two loop beta function for a g(1 ) x g(2 ) gauge theory , '' phys .
d * 25 * , 581 ( 1982 ) .
h. nishino _ et al . _
[ super - kamiokande collaboration ] , `` search for nucleon decay into charged anti - lepton plus meson in super - kamiokande i and ii , '' phys .
d * 85 * , 112001 ( 2012 ) [ arxiv:1203.4030 [ hep - ex ] ] . | we show that in @xmath0 models , a yukawa sector consisting of a real @xmath1 , a real @xmath2 and a complex @xmath3 of higgs fields can provide a realistic fit to all fermion masses and mixings , including the neutrino sector .
although the group theory of @xmath0 demands that the @xmath1 and @xmath2 be real , most constructions complexify these fields and impose symmetries exterior to @xmath0 to achieve predictivity . the proposed new framework with _ real _ @xmath1 and _ real _ @xmath2 relies only on @xmath0 gauge symmetry , and yet has a limited number of yukawa parameters .
our analysis shows that while there are restrictions on the observables , a good fit to the entire fermion spectrum can be realized .
unification of gauge couplings is achieved with an intermediate scale pati - salam gauge symmetry .
proton decay branching ratios are calculable , with the leading decay modes being @xmath4 and @xmath5 .
osu - hep-16 - 08 + * yukawa sector of minimal unification * + k.s .
babu@xmath6 , borut bajc@xmath7 and shaikh saad@xmath6 @xmath8_department of physics , oklahoma state university , stillwater , ok , 74078 , usa _
@xmath9_joef stefan institute , 1000 ljubljana , slovenia _ |
nanocomposite films made of nano - sized metallic particles embedded in an insulating matrix display peculiar electron transport regimes depending on the relative amount of the metallic phase compared to the insulating one . above a critical concentration @xmath0 of the metallic phase , macroscopic clusters of connected metallic particles span the entire sample , giving rise to an electrical conductivity @xmath1 characterized in the vicinity of @xmath0 by a power - law behavior of the form : @xmath2where @xmath3 is the transport exponent which the percolation theory predicts to take the values @xmath4 and @xmath5 for strictly two - dimensional ( 2d ) and three - dimensional ( 3d ) systems , respectively.@xcite equation describes well the conductivity behavior in the @xmath6 region of several nanogranular metal films grown by different methods,@xcite with observed values of the exponent that are consistent with the effective dimensionality of the films . below @xmath0 the metallic phase is broken up into disconnected metallic regions , so that for concentrations sufficiently smaller than @xmath0 electrons have to tunnel across the insulating barrier separating the conducting fillers . in this regime
, the conductivity of nanocomposites with embedded , isolated metallic particles follows approximately:@xcite @xmath7 where @xmath8 is the tunneling decay length which , depending on the nature of the composite constituents , ranges from a fraction to a few nanometers , and @xmath9 is a typical distance between the surfaces of two conducting particles . in the dilute limit @xmath10 ,
dimensional considerations@xcite imply that @xmath11 for particles of mean size @xmath12 randomly dispersed in a @xmath13-dimensional volume , as also inferred from transport measurements on 2d and 3d nanogranular metal composites.@xcite recently , a survey of the conductivity data of several thick nanogranular composite films has evidenced that the percolation behavior of eq
. evolves into the tunneling one of eq . as the result of the competition between percolation and tunneling transport mechanisms as
@xmath14 decreases.@xcite in this article we study the conductivity behavior as a function of gold content in au - implanted nanocomposite films above and below @xmath0 .
we consider two distinct series of systems : au - implanted alumina and au - implanted polymethylmethacrylate ( pmma ) films .
au - alumina films have thicknesses @xmath15 a few times larger than the mean au particle size @xmath12 , while the conducting layer of au - pmma is basically formed by a single layer of gold particles ( @xmath16 ) , which makes this system strictly two dimensional .
we show that for @xmath6 the conductivity of both systems follow eq . with transport exponents very close to the 2d value @xmath4 , in accord with the prediction of the percolation theory of nanocomposites with nanometric values of @xmath15 .
to study the tunneling behavior at au concentrations below @xmath0 , we analyze the conductivities of au - pmma and au - alumina using an effective medium theory formulated for film thicknesses ranging from @xmath17 to @xmath18 , as to describe the evolution from 2d to 3d of the conductivity .
we find that , in contrast to the 2d nature of the au - pmma films , the au - alumina films are large enough to sustain tunneling conductivity with 3d character .
the effective dimensionality of our ion - implanted au - alumina nanocomposites thus increases from 2d to 3d as the system crosses over from the percolation to the tunneling regimes .
this conclusion is supported by the value of @xmath8 extracted from the conductivity data , which coincides with that observed@xcite in thick ( i.e. , 3d ) co - sputtered au - alumina nanocomposite films.@xcite
the experimental data used in the present work are basically the same ones presented in refs . , but with a completely different approach and exploring a different range of the data .
the mentioned works addressed specifically implantation doses above the percolation thresholds , while the present work focuses in doses below the percolation thresholds , where the particles are isolated from each other and the tunneling effect is dominant . a summary of the experimental procedure will be done here and details can be obtained in the mentioned references .
pmma was deposited on glass microscope slides using a spin coater , generating a film thickness about @xmath19 nm .
electrical contacts were formed at both ends of the substrates by plasma deposition of thick ( @xmath20 nm ) platinum films .
very low energy ( @xmath21 ev ) was used for ion implantation using a streaming ( unidirectionally drifting ) charge - neutral plasma formed by a vacuum arc plasma gun .
gold ion implantation in alumina using @xmath22 kev was proceeded in an implanter.@xcite electrical contacts were formed at both ends of the alumina sample using silver paint . for both au - pmma and au - alumina systems , in situ resistance measurements
were performed as the ion implantation proceeded:@xcite after a known dose of au ions is implanted , the implantation process is temporary halted and the resistance is measured .
this process is repeated to determine the sample conductivity as a function of ion implantation dose . in the present work
, computer simulation using the tridyn computer code@xcite was used to estimate the depth profiles of the ion implanted gold in the alumina substrate .
tridyn is a monte carlo simulation program based on the trim ( transport and range of ions in matter ) code.@xcite the program takes into account compositional changes in the substrate due to : previously implanted dopant atoms , and sputtering of the substrate surface .
note that , the parameters @xmath3 ( transport exponent ) and @xmath0 ( critical concentration of the metallic phase ) were recalculated in the present work , using a different approach , presenting some deviation from those obtained in the previous works.@xcite
solid lines are fits with eq . .
] results of our conductivity measurements on au - pmma and au - alumina nanocomposites are shown in fig .
[ fig1 ] ( open squares ) , where the conductivity ratio @xmath23 is plotted as a function of the normalized au concentration @xmath24 . the saturation dose @xmath25 is defined as the implantation dose @xmath26 above which a continuous metal film starts to be deposited on the insulator surface , while the saturation conductivity @xmath27 is the conductivity measured at @xmath28 ( i.e. , at @xmath29 ) .
we have determined @xmath30 atoms @xmath31 and @xmath32 s / m for au - pmma,@xcite and @xmath33 atoms @xmath31 and @xmath34 s / m for au - alumina.@xcite as is apparent from the semi - log plot of fig .
[ fig1 ] , the @xmath14-dependence of @xmath23 for both au - pmma and au - alumina films presents a double hump shape , which is a feature commonly observed also in several other nanogranular metals composites.@xcite the hump at values of @xmath14 larger than @xmath35 stems from the presence of clusters of coalesced au particles that extent across the entire composite layer . in this region
, @xmath23 is reasonably well fitted by eq .
( shown by solid lines in fig .
[ fig1 ] ) with @xmath36 and @xmath37 for au - pmma , and @xmath38 and @xmath39 for au - alumina composite films .
the fitted values of the transport exponents are both consistent with the 2d value @xmath4 , indicating that ion - implanted au - pmma and au - alumina composites behave as 2d percolating composites for @xmath6 .
this result is consistent with the values of the thickness of the conducting layers of au - pmma and au - alumina extracted from the trydin analysis and from cross sectional tem images of au - pmma . for the case of au - pmma films
we observed a conducting layer of thickness @xmath40 nm ( see fig . [ fig2](a ) ) formed by au particles of size @xmath41 nm,@xcite which indicates that au - pmma films are composed basically of one monolayer of au particles embedded into the pmma substrate .
au - implanted pmma films are thus strictly 2d systems . on the contrary ,
the conducting layer of au - alumina films is not strictly two - dimensional : the mean au particle size extracted from tem is @xmath42 nm,@xcite while trydin analyses indicate that @xmath43 nm ( see fig .
[ fig2](b ) ) , that is , about @xmath44 times larger than @xmath12 .
the 2d character of the percolation conductivity of au - alumina films is explained by observing that the relevant length scale that governs the power - law behavior of eq .
( 1 ) is the correlation length @xmath45 ( not to be confused with the tunneling decay length @xmath8 ) , which measures the typical size of finite clusters of connected particles.@xcite when @xmath14 approaches @xmath0 from either above or below the percolation threshold @xmath0 , the correlation length increases as @xmath46 , where @xmath47 is of the order of the particle size @xmath12 , and @xmath48 is the correlation length exponent . for concentrations such that @xmath45 becomes larger than the film thickness @xmath15
, the composite behaves effectively as a 2d system,@xcite so that in the vicinity of @xmath0 the conductivity is expected to follow eq .
with @xmath4 , as observed in our au - alumina samples.@xcite ev and dose @xmath49 atoms @xmath31 , and into ( b ) alumina with ion energy @xmath22 kev and dose @xmath50 atoms @xmath31 . with these doses ,
the systems are below their respective percolation thresholds , presenting isolated gold nanoparticles . the arrows in ( b ) delimit @xmath51 nm that is the conducting layer thickness considered .
] dispersed within the space delimited by two parallel hard planes of area @xmath52 and separated by a distance @xmath53 .
the spheres can not penetrate the confining walls .
( b ) cross - sectional view illustrating the volume @xmath54 available for the sphere centers ( grey region ) . ] concerning the percolation thresholds extracted from the fits , we note that the values of @xmath0 of au - pmma and au - alumina are basically the same , although the percolation threshold is generally expected to depend on the particular combination of metal and insulator components constituting the films and on the details of the fabrication process.@xcite we have no arguments ruling out that the equivalence of the two percolation thresholds is just a mere coincidence , as very similar thresholds have already been observed in granular films with different components.@xcite interestingly , however , the value @xmath55 of our au - alumina films is larger than the percolation threshold of three - dimensional au - alumina composites ( @xmath56 ) , which is consistent with the observation that the percolation threshold generally diminishes as the dimensionality increases.@xcite for values of @xmath14 lower than @xmath0 clusters of coalesced particles no longer span the entire composite layer and the conductivity is dominated by tunneling processes , which give rise to the hump of @xmath1 at @xmath57 in fig .
[ fig1 ] . in this region
the conductivities of au - pmma and of au - alumina films are expected to follow eq . with typical inter - particle distances @xmath9
whose @xmath14-dependence is influenced by the effective dimensionality of the system .
although for the case of au - pmma films , which are strictly two - dimensional , @xmath9 can be determined by considering dispersions of conducing particles in a plane , for au - alumina films one must consider the effect of the finite width @xmath15 of the conducting layer , as its observed value ( @xmath58 ) is such that tunneling processes between au particles at different depths within the conducting layer are also possible . to tackle this problem , we consider an effective medium approximation ( ema ) for the tunneling conductivity applied to a dispersion of @xmath59 spherical particles of diameter @xmath12 that are confined by two parallel hard walls separated by a distance @xmath60 and of macroscopic area @xmath52 , as shown schematically in fig .
[ fig3](a ) .
the ema conductivity @xmath61 is the solution of the following equation:@xcite @xmath62 where @xmath63 $ ] , with @xmath64 , is the tunneling conductance between two spherical particles @xmath65 and @xmath66 , @xmath67 is the distance between their centers located at @xmath68 and @xmath69 , and @xmath70 is a tunneling prefactor . the angular brackets in eq .
denote a statistical average over the positions of the spheres occupying the volume delimited by the two parallel planes located at @xmath71 , as shown in fig .
[ fig3](b ) .
since the spherical particles can not penetrate the hard walls , the available volume for the sphere centers is @xmath54 .
assuming that the particles are uncorrelated , the average reduces to : @xmath72 so that eq . can be rewritten as : @xmath73 + 1}=2,\ ] ] where @xmath74 , @xmath75 for @xmath76 and @xmath77 for @xmath78 , and @xmath79 . using @xmath80 where @xmath81
, we express eq . in terms of an integral over @xmath82 which , due to the exponential decay of the tunneling conductance , can be extended over the whole space : @xmath83 + 1}=2,\ ] ] where @xmath84 is the particle number density . finally , we pass to spherical coordinates and integrate over the angles to find that the ema conductivity @xmath85 satisfies : @xmath86 + 1}=2\ ] ] for @xmath87 , and @xmath88 + 1}\nonumber\\ & + \frac{24 h\eta}{(h - d)d^3}\int_d^{h - d}\!dr\,\frac{r^2}{g^*\exp[2(r - d)/\xi]+1}\nonumber\\ & -\frac{12 h\eta}{(h - d)^2d^3}\int_d^{h - d}\!dr\,\frac{r^3}{g^*\exp[2(r - d)/\xi]+1}=2\end{aligned}\ ] ] for @xmath89 .
in eqs .
and we have introduced the dimensionless density @xmath90 which , for the case of uncorrelated ( i.e. , penetrable ) spheres , is related to the volume fraction @xmath14 of the metallic phase through @xmath91.@xcite the 2d and 3d limits are obtained by setting , respectively , @xmath17 in eq .
( 8) and @xmath18 in eq .
it is instructive to express @xmath85 as a tunneling conductance between the surfaces of two spheres separated by a characteristic distance @xmath92 : @xmath93 in this way , the term @xmath94 + 1\}$ ] that appears in the integrands of eqs . and reduces for @xmath95 to : @xmath96 + 1}&=\frac{1}{\exp[2(r - d-\delta^*)/\xi]+1}\nonumber\\ & \approx\theta(\delta^*+d - r).\end{aligned}\ ] ] using eq . in eq .
, we find thus that the characteristic distance @xmath92 for a strictly 2d system becomes : @xmath97^{1/2}-1,\ ] ] while , since the first and third integrals in eq
. vanish for @xmath18 , for a 3d system we obtain : @xmath98^{1/3}-1.\ ] ] for @xmath99 the above expressions correctly reproduce the dimensional scaling @xmath100 expected to hold true in the dilute limit . furthermore , eqs . and evidence that , for moderately small values of @xmath14 , the relevant tunneling distance for 2d systems is about twice that for 3d systems , as illustrated in fig .
[ fig4](a ) where eqs . and
are shown by solid lines . between the surfaces of two spheres as a function of the metallic volume fraction @xmath14 and for different thicknesses @xmath15 of the film .
@xmath92 is calculated from @xmath101 , where @xmath85 is the numerical solution of the ema equations and with tunneling decay length fixed at @xmath102 .
solid lines are results for the limits 2d and 3d given in eqs . and .
( b ) @xmath92 as a function of @xmath15 for selected values of the volume fraction of the metallic spheres . for @xmath103
the films behave practically as 3d systems . ] to assess how the ema conductivity evolves from the 2d to the 3d limits , we solve numerically eqs . and for films thicknesses ranging from @xmath17 to @xmath18 . expressing the resulting @xmath85 in terms of the distance
@xmath92 as specified in eq . , we find that @xmath92 rapidly decreases as the film thickness increases from @xmath17 , and essentially matches the 3d limit already for @xmath104 , as shown in fig .
[ fig4](b ) where @xmath92 is presented as a function of @xmath105 for selected @xmath14 values . the rapid evolution from 2d to 3d is also illustrated in fig .
[ fig4](a ) , where the numerical values of @xmath92 as function of @xmath14 and for several film thicknesses are compared with eqs . and .
( open squares ) as a function of @xmath106 for ( a ) au - pmma and ( b ) au - alumina nanocomposite films .
the characteristic ema distance @xmath92 is assumed to be given by the 2d limit of eq . for au - pmma and by the 3d limit of eq . for au - alumina .
solid lines are the best fits of the data with eq . .
] the above analysis suggests that it suffices to have film thicknesses only a few times larger than the particle size to induce a 3d character to the tunneling conductivity .
this result is consistent with the observation that the relevant length scale for tunneling is given by the mean inter - particle distance , and has interesting consequences with respect to the nanocomposite films considered here . indeed ,
if on the one hand au - pmma films are expected to display 2d conductivity in the whole range of @xmath14 because @xmath16 , on the other hand the conductivity of au - alumina nanocomposites should be understood as having 2d character for @xmath6 but 3d character for @xmath57 .
this is so because the measured thickness of the conducting layer of au - alumina ( @xmath58 ) is sufficiently larger than the ema 2d-3d crossover value @xmath107 ( see fig .
[ fig4 ] ) to induce 3d tunneling .
assuming that the above considerations capture the essential physics of the problem , we interpret the measured conductivity data of au - pmma and au - alumina films in terms of eq . , where @xmath9 is identified with @xmath92 as given by eqs . and ,
respectively . consistently with the exponential form of eq . , the @xmath108 data of both systems follow approximately a linear dependence as a function of @xmath92 , as shown in fig .
[ fig5 ] . from linear fits of @xmath108 with @xmath109
we extract the values of @xmath110 that best fit the data to find @xmath111 for au - pmma and @xmath112 for au - alumina .
remarkably , the value of @xmath113 found for au - alumina coincides with the value extracted from analyses of co - sputtered au - alumina granular thick films,@xcite which are three - dimensional systems .
using the estimated mean size of au particles in our films ( @xmath114 nm ) we find @xmath115 nm , which compares fairly well with @xmath116 nm , obtained from @xmath117 , where @xmath118 is the electron mass and @xmath119 ev is the estimated barrier height for tunneling between au and alumina.@xcite it is worth noting that fitting the au - alumina conductivity data using eq . with @xmath92 as given by the 2d limit of eq
. gives @xmath120 nm , which is about twice the value found assuming 3d tunneling . from @xmath121 found for au - pmma films and from the corresponding mean au particle size ( @xmath122 nm ) we find @xmath123 nm , which indicates that the tunneling decay length of au - pmma is substantially larger than that of au - alumina nanocomposites .
calculated for dispersions of impenetrable disks in 2d,@xcite and impenetrable spheres in 3d.@xcite solid lines are @xmath124 for the 2d case and @xmath125 for the 3d case , where @xmath126 and @xmath127 are the expressions of the corresponding mean distances between the centers of nearest neighboring particles given in eqs . and . ] to assess the robustness of the results based on the ema approach , we repeat the above analysis by identifying @xmath128 of eq . with the critical distance @xmath129 , as prescribed by the critical path approximation ( cpa).@xcite according to cpa , @xmath129 is defined as the shortest among the inter - particle distances @xmath130 such that the set of bonds satisfying @xmath131 forms a percolating cluster . for dispersions of metallic particles
in which the distances @xmath132 span several multiples of @xmath8 , cpa ensures that eq . with @xmath133
gives a good estimate of the composite conductivity.@xcite ( open squares ) as a function of @xmath129 for ( a ) au - pmma and ( b ) au - alumina nanocomposite films .
the critical distance @xmath129 is assumed to be given by the 2d limit for au - pmma and by the 3d limit for au - alumina .
solid lines are the best fits of the data with eq . .
] monte carlo results of @xmath129 for impenetrable spheres dispersed in a 3d volume@xcite and impenetrable disks dispersed in a 2d area@xcite are shown in fig .
[ fig6 ] by open symbols . for the 2d case
, we set @xmath134 to convert the area fraction @xmath135 covered by the disks to the volume fraction @xmath14 of spheres of equal diameter with centers lying on a plane . in analogy with the ema results of fig .
[ fig4 ] , the critical distance for the 2d case is systematically larger than the critical distance in 3d . for @xmath14 larger than about @xmath136 the monte carlo data are well reproduced by setting @xmath124 for the 3d case and @xmath125 for the 3d case ( solid lines in fig .
[ fig6 ] ) , where @xmath137 and @xmath138 are the mean distances between the centers of nearest neighboring spheres in 2d and 3d , respectively.@xcite we fit the @xmath108 data of au - pmma and au - alumina with eq .
using respectively the 2d and 3d functional dependences of @xmath129 , as shown in fig .
[ fig7 ] . from the slopes of the straight lines we extract @xmath139 for au - pmma and @xmath140 for au - alumina .
we immediately see that for the au - alumina films the estimates of @xmath113 from cpa and ema coincide within errors , while those for au - pmma films differ by almost a factor @xmath141 .
this discrepancy could be attributed to possible effects of local particle correlations , neglected within our ema approach but fully accounted for in cpa , which are expected to be more prominent in 2d than in 3d . combining the estimates from ema and cpa , and using the measured mean size of au particles , we infer that in 2d au - pmma films the tunneling decay length is @xmath142 nm .
we have presented a study of the dependence of the electrical conductivity on the gold concentration in au - implanted composite thin films with different insulating matrices .
we have evidenced that the film thickness may influence substantially the behavior of the conductivity below the percolation threshold @xmath0 , where tunneling between isolated gold particles dominates .
specifically , we have shown that an effective medium theory predicts a crossover from two - dimensional to three - dimensional tunneling behavior when the film thickness @xmath15 is larger than only about 3 times the mean au particle size @xmath12 .
au - implanted pmma films , that have thickness@xmath16 , are thus strictly 2d systems in the tunneling regime , while au - implanted alumina films are expected to show 3d tunneling behavior , as for this system @xmath58 .
interestingly , the dimensionless tunneling decay length @xmath113 extracted from the tunneling conductivity data of au - implanted alumina films coincides with previous estimates of @xmath113 in co - sputtered au - alumina thick film composites .
we have also shown that above the percolation threshold the measured conductivity of both au - pmma and au - alumina follows a percolation power - law behavior with 2d transport exponent , in accord with the theory of percolation in thin disordered films .
au - pmma films have thus 2d character in the whole range of au concentrations , while the effective dimensionality of au - implanted alumina films increases from 2d to 3d as the system crosses over from the percolation regime to the tunneling regime . these results and interpretations
could find a firmer confirmation by measuring the conductivity behavior in films with thicknesses ranging continuously from @xmath16 to @xmath18 . from our model
we expect that the tunneling conductivity crosses over from two - dimensional to three - dimensional behaviors when the film thickness is about @xmath143-@xmath144 , as inferred from the behavior shown in fig .
[ fig4 ] .
this work was supported by the fundao de amparo pesquisa do estado de so paulo ( fapesp ) and the conselho nacional de desenvolvimento cientfico e tecnolgico ( cnpq ) , brazil .
we are grateful to the institute of ion beam physics and materials research at the forschungszentrum dresden - rossendorf , germany , for the tridyn - fzr computer simulation code .
v. ambegaokar , b. i. halperin , and j. s. langer , phys .
b * 4 * , 2612 ( 1971 ) ; m. pollak , j. non - cryst .
solids * 11 * , 1 ( 1972 ) ; b. i. shklovskii and a. l. efros , sov .
jetp * 33 * , 468 ( 1971 ) ; * 34 * , 435 ( 1972 ) .
we approximate the typical inter - particle distance @xmath128 by the mean distance between nearest neighboring particles . for a system of randomly distributed particles of number density @xmath145 ,
there is in average one particle contained within a volume @xmath146 centered about a given particle : @xmath147 , so that @xmath148 .
ziegler , j.p .
biersack , u. littmark , _ the stopping and range of ions in solids _ , ( pergamon , new york , 1985 ) ( new edition in 1996 ) .
the computer code is downloadable from the srim website http://www.srim.org/. it is possible to estimate the values of @xmath14 below which au - alumina films behave as 2d percolating systems by using @xmath149 , where @xmath150 is the 3d correlation length exponent .
using @xmath55 and @xmath58 , we obtain that @xmath45 becomes greater than the film thickness when the au concentration is lower than about @xmath151 . | we study the dependence of the electrical conductivity on the gold concentration of au - implanted polymethylmethacrylate ( pmma ) and alumina nanocomposite thin films . for au contents larger than a critical concentration ,
the conductivity of au - pmma and au - alumina is well described by percolation in two dimensions , indicating that the critical correlation length for percolation is larger than the thickness of the films . below the critical loading ,
the conductivity is dominated by tunneling processes between isolated au particles dispersed in pmma or alumina continuous matrices . using an effective medium analysis of the tunneling conductivity ,
we show that au - pmma behaves as a tunneling system in two dimensions , as the film thickness is comparable to the mean au particle size . on the contrary ,
the conductivity of au - alumina films is best described by tunneling in three dimensions , although the film thickness is only a few times larger than the particle size .
we interpret the enhancement of the effective dimensionality of au - alumina films in the tunneling regime as due to the larger film thickness as compared to the mean interparticle distances . |
matrix multiplication is one of the most fundamental tasks in mathematics and computer science .
while the product of two @xmath1 matrices over a field can naturally be computed in @xmath2 arithmetic operations , strassen showed in 1969 that @xmath3 arithmetic operations are enough @xcite .
the discovery of this algorithm for matrix multiplication with subcubic complexity gave rise to a new area of research , where the central question is to determine the value of the exponent of square matrix multiplication , denoted @xmath4 , and defined as the minimal value such that two @xmath1 matrices over a field can be multiplied using @xmath5 arithmetic operations for any @xmath6 .
it has been widely conjectured that @xmath7 and several conjectures in combinatorics and group theory , if true , would lead to this result @xcite .
however , the best upper bound obtained so far is @xmath8 , as we explain below .
coppersmith and winograd @xcite showed in 1987 that @xmath9 .
their approach can be described as follows .
a trilinear form is , informally speaking , a three - dimensional array with coefficients in a field @xmath10 . for any trilinear form @xmath11
one can define its border rank , denoted @xmath12 , which is a positive integer characterizing the number of arithmetic operations needed to compute the form . for any trilinear form @xmath11 and any real number @xmath13 $ ]
, one can define a real number @xmath14 , called the _ value _ of the trilinear form . the theory developed by schnhage @xcite shows that , for any @xmath15 and any @xmath13 $ ] , the following statement hold : @xmath16 here the notation @xmath17 represents the trilinear form obtained by taking the @xmath18-th tensor power of @xmath11 .
coppersmith and winograd presented a specific trilinear form @xmath19 , obtained by modifying a construction given earlier by strassen @xcite , computed its border rank @xmath20 , and introduced deep techniques to estimate the value @xmath21 .
in particular , they showed how a lower bound @xmath22 on @xmath21 can be obtained for any @xmath13 $ ] by solving an optimization problem . solving this optimization problem
, they obtained the upper bound @xmath23 , via statement ( [ statement ] ) with @xmath24 and @xmath25 , by finding the smallest @xmath26 such that @xmath27 .
they then proceeded to study the tensor power @xmath28 and showed that , despite several new technical difficulties , a similar approach can be used to reduce the computation of a lower bound @xmath29 on @xmath30 to solving another optimization problem of several variables .
they discovered that @xmath31 ^ 2 $ ] , due to the fact that the analysis of @xmath28 was finer , thus giving a better upper bound on @xmath4 via statement ( [ statement ] ) with @xmath24 and @xmath32 .
solving numerically the new optimization problem , they obtained the upper bound @xmath9 . in view of the improvement obtained by taking the second tensor power ,
a natural question was to investigate higher powers of the construction @xmath19 by coppersmith and winograph . investigating the third power
was explicitly mentioned as an open problem in @xcite .
more that twenty years later , stothers showed that , while the third power does not seem to lead to any improvement , the fourth power does give an improvement @xcite ( see also @xcite ) .
the improvement was obtained again via statement ( [ statement ] ) , by showing how to reduce the computation of @xmath33 to solving a non - convex optimization problem .
the upper bound @xmath34 was obtained in @xcite by finding numerically a solution of this optimization problem .
it was later discovered that that solution was not optimal , and the improved upper bound @xmath35 was given in @xcite by exhibiting a better solution of the same optimization problem .
independently , vassilevska williams @xcite constructed a powerful and general framework to analyze recursively powers of a class of trilinear forms , including the trilinear form @xmath19 by coppersmith and winograd , and showed how to automatically reduce , for any form @xmath11 in this class and any integer @xmath36 , the problem of obtaining lower bounds on @xmath37 to solving ( in general non - convex ) optimization problems .
the upper bound @xmath38 was obtained @xcite by applying this framework with @xmath24 and @xmath39 , and numerically solving this optimization problem .
obtained for the eighth power is stated as @xmath40 in the conference version @xcite , the statement has been corrected to @xmath38 in the most recent version @xcite , since the previous bound omitted some necessary constraints in the optimization problem .
our results confirm the value of the latter bound , and increase its precision . ]
a natural question is to determine what bounds on @xmath4 can be obtained by studying @xmath41 for @xmath42 .
one may even hope that , when @xmath43 goes to infinity , the upper bound on @xmath4 goes to two .
unfortunately , this question can hardly be answered by this approach since the optimization problems are highly non - convex and become intractable even for modest values of @xmath43 . in this paper
we show how to modify the framework developed in @xcite in such a way that the computation of @xmath44 reduces to solving @xmath45 instances of _ convex _ optimization problems , each having @xmath45 variables . from a theoretical point a view , since a solution of such convex problems can be found in polynomial time , via statement ( [ statement ] ) we obtain an algorithm to derive an upper bound on @xmath4 from @xmath41 in time polynomial in @xmath43 . from a practical point of view
, the convex problems we obtain can also be solved efficiently , and have several desirable properties ( in particular , the optimality of a solution can be guaranteed by using the dual problem ) .
we use this method to analyze @xmath46 and @xmath47 , and obtain the new upper bounds on @xmath4 described in table [ table : chart ] . besides leading to an improvement for @xmath4
, these results strongly suggest that studying powers higher than 32 will give only negligible improvements .
our method is actually more general and can be used to efficiently obtain lower bounds on @xmath48 for any trilinear forms @xmath11 and @xmath49 that have a structure similar " to @xmath19 .
indeed , considering possible future applications of our approach , we have been attentive of stating our techniques as generally as possible .
to illustrate this point , we work out in the appendix the application of our method to an asymmetric trilinear form , originally proposed in @xcite . | this paper presents a method to analyze the powers of a given trilinear form ( a special kind of algebraic constructions also called a tensor ) and obtain upper bounds on the asymptotic complexity of matrix multiplication .
compared with existing approaches , this method is based on convex optimization , and thus has polynomial - time complexity . as an application , we use this method to study powers of the construction given by coppersmith and winograd [ journal of symbolic computation , 1990 ] and obtain the upper bound @xmath0 on the exponent of square matrix multiplication , which slightly improves the best known upper bound . |
with the rhic now in its second year of running , it is important to pin down lattice predictions for the high temperature phase of qcd . in this context ,
fully non - perturbative measurements of quark number susceptibilities @xcite are important for four reasons .
firstly , they are directly related to experimental measurements of event - to - event fluctuations in particle production @xcite .
secondly , earlier results @xcite showed a strong jump in the susceptibility across the phase transition , but indicated a significant departure from weak - coupling behaviour , and it is important to check whether this persists into the continuum .
third , resummed perturbative computations of this quantity have now become available @xcite , thus making it possible to accurately test the importance of non - perturbative contributions to this quantity .
finally , with continuum extrapolated results in hand , one can address the question of whether the strangeness production seen in heavy - ion collisions can be quantitatively explained as a signal of the quark - gluon plasma .
we have recently presented systematic results for quark number susceptibilities in quenched qcd @xcite as well as for qcd with two flavours of light dynamical quarks @xcite . in these studies a large range of temperatures , @xmath0 , was covered at a series of different quark masses , @xmath3 .
these computations were done at a fixed cutoff with the lattice spacing , @xmath4 while keeping the finite volume effects under control so that the thermodynamic limit could be taken reliably .
there was a 35% difference between the quenched and dynamical computations . in this paper
we examine the cutoff dependence of the quenched results and report their continuum ( zero lattice spacing ) limit .
the partition function of qcd with three quark flavours is z(t,_u,_d,_s ) = u _ f = u , d , sm(t , m_f,_f ) , [ part]where the temperature @xmath0 determines the size of the euclidean time direction , @xmath5 is the gluonic part of the action and the determinants of the dirac matrices , @xmath6 , contain as parameters the quark masses , @xmath7 , and the chemical potentials @xmath8 for each of the flavours @xmath9 .
we also define the chemical potentials @xmath10 , @xmath11 and @xmath12 , which correspond to the diagonal flavour @xmath13 generators .
note that @xmath14 is the usual baryon chemical potential and @xmath15 is an isovector chemical potential .
quark number densities are defined as n_i(t,_u,_d,_s ) = t v , [ numb]and the susceptibilities as _
ij(t,_u,_d,_s ) = t v , [ susc]where @xmath16 denotes the spatial volume .
the subscripts @xmath17 and @xmath18 are either of the index sets @xmath19 or @xmath20 , 3 and 8 .
we lighten the notation by writing the diagonal susceptibilities @xmath21 as @xmath22 .
we determine the susceptibilities at zero chemical potential , @xmath23 . in this limit , all @xmath24 .
since we work with @xmath25 , we also have @xmath26 .
flavour off - diagonal susceptibilities such as _ ud = ( ) m_u^-1 m_u m_d^-1 m_d[offd]are given entirely in terms of the expectation values of disconnected loops .
since @xmath27 , we obtain @xmath28 with each defined by an obvious generalization of the formula above . of the flavour diagonal susceptibilities
we shall use _ s = ( ) .
[ chis]@xmath29 are given by a generalization of this formula .
numerically , the simplest quantity to evaluate is the diagonal iso - vector susceptibility _ 3 = 12 ( ) ( m_u^-1 m_u - m_u^-1m_um_u^-1m_u ) .
[ chi3]two more susceptibilities are of interest .
these are the baryon number and electric charge susceptibilities , _ 0 = 19(4_3+_s+4_ud+4_us ) _ q = 19(10_3+_s+_ud-2_us ) .
[ chi0q]note that @xmath30 is the baryon number susceptibility for three flavours of quarks . as a result , this expression differs from the iso - singlet quark number susceptibility for two flavours , defined in @xcite , both in overall normalization and by terms containing strangeness . in our numerical work
we have chosen to use staggered quarks .
hence , to normalise to one flavour of continuum quarks and compensate for fermion doubling on the lattice , we have to multiply each of the traces in eqs.([offd][chi3 ] ) by a factor of @xmath31 .
the quark mass appears in two logically distinct places first in the operators which define the susceptibilities in eqs.([offd][chi3 ] ) , and secondly in the determinant in the partition function of eq .
( [ part ] ) which defines the weight for the averaging of these operators .
the first defines a valence quark mass and the second the sea quark mass . in this paper
we adopt the quenched approximation , whereby the determinants in eq .
( [ part ] ) are set equal to unity .
however , we shall vary the valence quark masses . as shown earlier , unquenching the sea quarks changes results by 35% @xcite .
note that all the flavour off - diagonal susceptibilities are exactly zero in an ideal gas .
we shall denote by @xmath32 the ideal gas value for @xmath33 . on an @xmath34 lattice with @xmath35 colours and lattice spacing @xmath36
, we find a^2 ^ 3_fft = _ p \
{ d^-2 ^2p_0 ^ 2p_0 + d^-1(^2p_0 - 12 ) } , [ fft]where the spectrum of momenta is @xmath37 with @xmath38 and @xmath39 with @xmath40 , and @xmath41 . for a given @xmath42 and @xmath43 ,
the value of the quark mass in lattice units is ma=(1n_t ) ( mt_c)(t ) .
[ mass]the formula in eq .
( [ fft ] ) is normalised such that there is exactly one copy of each flavour of quarks .
the values of @xmath44 , @xmath30 and @xmath45 for an ideal gas can then be simply obtained as linear combinations of @xmath32 for different valence quark masses .
for a given coupling @xmath46 .
crosses denote data for @xmath47 lattices and pluses for @xmath48 lattices .
the scale has been set by a corrected qcd 2-loop formula @xcite .
the difference between the two sets of data is a measure of scale breaking by residual power - law corrections and is negligible within statistical errors . ]
we have previously reported measurements of @xmath33 in quenched qcd with 4 time slices ( @xmath47 ) at temperatures of @xmath49 , @xmath50 , @xmath51 , @xmath52 and @xmath53 @xcite .
we have extended these measurements to higher temperatures through two simulations on @xmath54 lattices at @xmath55 and @xmath56 .
these correspond to temperatures of @xmath57 and @xmath58 .
all these computations are performed at fixed bare quark masses of @xmath59 , @xmath60 , @xmath61 , @xmath62 and 1 ( the data for @xmath63 are new ) . the continuum limit has been taken by going to several smaller lattice spacings .
one set of computations is performed with a lattice spacing @xmath64 which is 33% smaller than the lattice spacing used earlier .
we have taken data on @xmath65 lattices at @xmath66 , @xmath67 and @xmath56 corresponding to @xmath68 , 2 and @xmath69 , and on a @xmath70 lattice at @xmath71 corresponding to @xmath72 .
one further computation was made on an @xmath73 lattice at @xmath74 .
this corresponds to @xmath75 with lattice spacing @xmath76 .
the quark mass @xmath42 is kept independent of @xmath36 and @xmath0 , and hence @xmath77 decreases with increasing @xmath0 at fixed @xmath78 or with increasing @xmath78 at fixed @xmath0 according to eq .
( [ mass ] ) . the lattice scale has been set using the plaquette measurements of @xcite and the analysis performed in @xcite .
the conversion of @xmath46 to @xmath0 using data for @xmath47 and 8 give the two curves in figure [ fg.temp ] .
the difference between the two curves is within the statistical uncertainty of 5% on the scale assignment .
there is the same degree of statistical uncertainty in the bare quark masses and in all other scales .
we shall not display this inherent scale uncertainty in the measurements , but it should be kept in mind .
the simulations have been performed with a cabbibo - marinari pseudo - heatbath technique with 3 @xmath79 subgroups updated on each hit .
an initial 1000 sweeps have been discarded for thermalisation . on the @xmath80 lattices
we have used 80 configurations separated by 1000 sweeps for the measurements . on the @xmath48 lattice
we have used 55 configurations separated by 500 sweeps . at @xmath81 shown as a function of @xmath82 for @xmath59 .
the ratio reaches its continuum limit when evaluated on lattices with @xmath83 . ] in quenched qcd , shown as a function of @xmath0 for the values of @xmath42 indicated .
pluses denote data for lattice spacing @xmath84 and circles for @xmath64 .
the lines are cubic spline fits to the data .
] we have previously checked that the spatial lattice size , @xmath85 , is quite irrelevant to the values of @xmath86 measured as long as it is sufficiently large compared to the inverse pion mass @xcite .
in fact , at all the couplings and quark masses we have used , @xmath87 .
the most important other constraint on @xmath88 comes from the requirement that it should be large enough to prevent spatial deconfinement @xcite .
this is ensured by taking @xmath89 in all our simulations .
the traces are evaluated by the usual stochastic technique , a = 12n_i = i^n r_i^ar_i , [ stoch]where @xmath90 are a set of @xmath91 uncorrelated vectors with components drawn independently from a gaussian ensemble with unit variance .
each vector has three colour components at each site of the lattice .
since we use a half lattice version of the dirac operator for staggered fermions , the number of components of each vector @xmath90 is one and a half times the number of lattice points .
@xmath92 is obtained by dividing the set of @xmath90 into disjoint blocks , constructing @xmath93 in each block , taking all possible distinct pairs of such estimates , multiplying them and then averaging over the pairs .
for @xmath47 it is possible to get accurate results with @xmath94 , although we have chosen to use @xmath95 for each gauge configuration @xcite .
it is easy to check that the variance of the estimator above is proportional to @xmath96 . since the diagonal element of the fermion matrix is proportional to the quark mass @xmath77 , whereas the off - diagonal elements are bounded by unity , with increasing @xmath0 or decreasing @xmath36 , @xmath93 decreases linearly , while its variance remains constant .
hence , for sufficiently small @xmath77 the number of vectors has to increase quadratically with @xmath77 .
we found that @xmath97 was sufficient for @xmath48 at @xmath81 , although at larger @xmath78 or @xmath0 significantly larger values of @xmath98 are needed .
a further numerical problem arises from the cancellation of the two matrix elements which give @xmath33 ( see eq .
[ chi3 ] ) . while increasing @xmath78 and keeping the physical size of the lattice fixed , each of these terms increases quadratically but mutually cancel to give a number which decreases quadratically with @xmath78 .
at fixed word length , this implies a reduction in accuracy by a factor of @xmath99 . , and the quark number susceptibility , @xmath33 at @xmath52 .
the measurements have been made on a @xmath73 lattice at a series of quark masses .
the line is an exponential fit to the data . ]
.results for the continuum limit of quark number susceptibility in quenched qcd .
we have taken @xmath100 as appropriate to full qcd .
our @xmath47 results indicate that at higher @xmath0 the ratios @xmath101 remain flat .
[ cols="<,>,>,>,>,>",options="header " , ] in figure [ fg.extrap ] we show how the continuum limit of @xmath33 is reached .
there is an 8% decrease in @xmath102 in going from @xmath47 to @xmath83 lattices .
however , within measurement errors , the continuum limit is reached already at @xmath83 .
this implies a cancellation of the @xmath78-dependence of @xmath33 and @xmath32 .
the reason for the atypical behaviour for @xmath47 is not in the numerator , which is non - perturbative , but in the denominator , which is the ideal gas . for @xmath47 ,
the spectrum of @xmath103 consists of @xmath104 . as a result , @xmath105 and the second term in the sum on the right of eq .
( [ fft ] ) vanishes identically .
this accident occurs only for @xmath47 , and for all other values of @xmath78 this term contributes a non - vanishing value .
it is easy to see that in the limit @xmath106 this term contributes to the leading part of @xmath107 .
this is the reason the values for @xmath108 for @xmath47 lie well away from the continuum limit .
our complete results for the dependence of @xmath102 on @xmath43 at several different quark masses for @xmath47 and 6 are shown in figure [ fg.tdep ] . in view of the data shown in figure [ fg.extrap ] ,
our results for @xmath83 are the continuum limit for the quenched theory : these are summarised in table [ tb.res ] .
note that even at temperature as high as @xmath109 the result is far below the ideal gas limit .
closer to @xmath110 there is an even more dramatic fall in the value of @xmath102 .
we have also measured the off - diagonal susceptibility @xmath111 . as already seen for @xmath47
this is consistent with zero . in table
[ tb.res ] we show that the continuum limit of @xmath111 also vanishes within errors of @xmath112 .
it is therefore much smaller than the 3-loop perturbative result given in @xcite .
note that the continuum limit of our measurements lie 2030% below the ideal gas result .
they also differ significantly from htl as well as the skeleton graph resummed results of @xcite . at @xmath113 , htl predicts @xmath114@xmath115 on varying the scale of @xmath116 between @xmath117 and @xmath118 , and the resummed computation gives @xmath119@xmath120 .
our measurement shows @xmath121 at @xmath53 ( see table [ tb.res ] ) . a second computation of the htl result is available for @xmath122 @xcite .
since this agrees with the htl result of @xcite for @xmath122 , we believe the htl result for @xmath102 is under control .
there is thus a genuine discrepancy between these lattice results and existing perturbative computations .
the quark number susceptibilities are closely related to the screening correlator of a one - link separated quark bilinear operator which corresponds to a @xmath123 meson at zero temperature .
the @xmath124 transfer matrix of the problem mixes this with a quark bilinear that corresponds to the @xmath125 @xcite .
we have earlier shown evidence for @xmath47 that @xmath33 is closely related to the only known non - perturbative quantity among the screening correlators , that is the screening mass @xmath126 coming from the two degenerate correlators that descend from the @xmath125 and @xmath127 operators of the zero temperature theory @xcite . in view of the strongly non - perturbative character of @xmath33 , as revealed by the comparison with htl and resummation explained above , examining this correlation in the continuum becomes more significant . in figure [ fg.mpi ]
we show data obtained on @xmath73 lattices which indicate that this correlation survives into the continuum .
, obtained from various experimental measurements , compared with the value obtained from the lattice measurements reported here . ]
measurements of hadron yields in relativistic heavy - ion collisions have been analysed extensively @xcite and the observed enhancement of strangeness has been claimed to be evidence for quark - gluon plasma formation .
the chemical composition is observed at the hadron freezeout temperature , which has been found to be very close to the value of @xmath110 in full qcd .
it has been argued that this chemical composition can not arise due to hadronic rescattering @xcite . it would be interesting if it could be directly determined whether or not the composition is characteristic of the plasma for @xmath128 .
we argue below that this question can be answered by a lattice qcd measurement such as we have presented above .
the quantity which seems to be directly connected to lattice measurements is the wroblewski parameter @xcite , @xmath129 , which measures the ratio of newly created primary strange to light quarks _ s = .
[ wrob]a fluctuation - dissipation theorem relates the rate of production of quark pairs by the plasma in equilibrium to the imaginary part of the full complex quark number susceptibilities @xcite .
if the inverse of the characteristic time scales of the qcd plasma are not close to the typical energy scales for the production of strange and lighter quarks , then the ratio of their production rates is just the ratio of the static susceptibilities that we have measured .
then , if the observed chemical composition is created in equilibrium , we should have _
s = , [ relwrob]where the last equality holds in the limit of equal @xmath130 and @xmath131 quark masses .
these susceptibilities have to be evaluated on the lattice at the temperature and chemical potential , @xmath14 , relevant to the collision .
however , it has been shown recently @xcite that @xmath129 is insensitive to @xmath14 for sps and rhic energies .
furthermore a lattice measurement shows that the screening mass , @xmath126 , does not change rapidly with increasing chemical potential @xcite .
then , in view of the correlation between @xmath126 and @xmath86 shown in figure [ fg.mpi ] , we do not expect rapid changes with chemical potential in the ratio above .
both these arguments show that as a first approximation one can take the ratio in eq .
( [ relwrob ] ) at zero chemical potential . from our measurements of @xmath86 reported here
, we can form the lattice `` prediction '' of @xmath129 created in the plasma at @xmath110 .
this has been done by taking our measurements at @xmath132 and 2 , and extrapolating them linearly to @xmath110 .
for @xmath133 we use data taken for @xmath59 , which is light enough for this purpose .
since scaled quantities feel the smallest correction for unquenching @xcite , we take @xmath44 to correspond to @xmath134 , which is appropriate to full qcd . in figure [ fg.wrob ] the resultant prediction is compared with the data collected in @xcite .
it can be seen that there is a fair agreement with the data .
however , this is subject to several assumptions 1 .
the foremost assumption is that the characteristic time scales of the plasma are not close to the inverse energy scale of the production processes .
it is not possible to test this assumption in an euclidean computation .
however , it has been suggested that these characteristic time scales could be observed in real or virtual photon production as spikes in the spectrum @xcite . if these are seen and they lie in the energy range for the production , the assumption would be falsified .
another important assumption is that chemical equilibration takes place in the plasma .
if chemical equilibrium is not achieved but the light degrees of freedom achieve energy equipartition , then the effective temperature could be higher .
we have assumed slow variation of the ratio @xmath135 with chemical potential . while this seems to be a reasonable estimate in view of the results of @xcite and our own observations about the relation between the lowest screening mass and the susceptibilities
, one should keep in mind possible changes in the ratio due to violation of this assumption .
our continuum results are obtained in quenched qcd , whereas the data corresponds to full qcd . as mentioned already ,
we have taken @xmath134 , as appropriate to full qcd , in order to correct for this .
however , an additional 510% shift in our results could arise when unquenching @xcite .
this is not shown in figure [ fg.wrob ] .
our measurements are made away from @xmath110 and extrapolated down to this temperature . in principle
this can be corrected by a computation made directly at @xmath110 . however , due to the enormous cost of extracting physical quantities at @xmath110 , we postpone this to the time when we attempt a full qcd simulation .
it might seem surprising that @xmath129 corresponds to the chemical composition at @xmath110 .
however , a moment s thought shows that this is entirely in accord with the assumption of chemical equilibrium .
even if the system is thermalized at a much higher temperature , it remains in equilibrium until it reaches @xmath110 .
departure from equilibrium as the system cools through the transition is unlikely , and if it occurs , should have a visible signature such as the formation of a disoriented chiral condensate @xcite . in the absence of such phenomena , if @xmath129 does not correspond to the chemistry at @xmath110 , then the second assumption above would be demonstrated to be false . also , with increasing ion energy the system moves closer to zero baryon number , as seen at rhic .
this makes our third assumption better at the rhic and lhc .
the fourth assumption can be removed by a computation with full qcd .
such a computation is planned , where also the final assumption can be removed , and will be reported elsewhere .
in summary , we have presented new and precise results on quark number susceptibilities over a wide range of temperatures and quark masses in the high temperature phase of qcd .
the main aim of this study was to obtain the continuum extrapolation of these susceptibilities .
as shown in figure [ fg.extrap ] , while there is a significant change in going from @xmath47 to @xmath83 lattices , there is no statistically significant change in the susceptibilities beyond this .
the results obtained for @xmath83 lattices can then be taken as the continuum limit .
this limit is shown in figure [ fg.tdep ] and table [ tb.res ] .
there is a strong discrepancy between the continuum extrapolated lattice results and htl computations for these quantities varying between about 12% at the highest @xmath0 to even larger at smaller @xmath0 .
the off - diagonal susceptibility , @xmath111 , is zero , in contrast to the htl results .
there is a somewhat larger discrepancy between the continuum extrapolated lattice results and the skeleton graph resummed results .
the conjecture that there is a strong non - perturbative component to the quark number susceptibilities is supported by an observed strong correlation between the smallest quark bilinear screening mass , @xmath126 , and the susceptibility @xmath33 ( figure [ fg.mpi ] ) .
it is interesting to note that the continuum extrapolated results for the strange and light quark susceptibilities can be used to give a surprisingly good description of the chemical composition of hadrons at freezeout in sps and rhic experiments ( figure [ fg.wrob ] ) .
this has to be treated as a preliminary estimate due to the many caveats which we have listed , and some of which we plan to check in future lattice measurements .
99 s. gottlieb _ et al .
_ , _ phys .
_ 59 ( 1987 ) 2247 .
r. v. gavai _ et al .
_ , _ phys .
_ d 40 ( 1989 ) 2743 ; + c. bernard _ et al .
_ , _ phys .
_ d 54 ( 1996 ) 4585 ; + s. gottlieb _ et al .
_ , _ phys . rev .
_ , d 55 ( 1997 ) 6852 .
r. v. gavai and s. gupta , _ phys .
_ d 64 ( 2001 ) 074506 .
r. v. gavai , s. gupta and p. majumdar , hep - lat/0110032 , _ phys .
_ d , in press .
m. asakawa , u. heinz and b. mller , _ phys .
_ 85 ( 2000 ) 2072 ; + s. jeon and v. koch , _ ibid . _ 85 ( 2000 ) 2076 ; + d. bower and s. gavin , _ phys . rev .
_ c 64 ( 2001 ) 051902 ; + s. jeon , v. koch , k. redlich and x. n. wang , nucl - th/0105035 .
blaizot , e. iancu and a. rebhan , hep - ph/0110369 .
s. datta and s. gupta , _ phys .
_ , b 471 ( 2000 ) 382
. g. boyd _ et al .
_ , _ nucl .
_ , b 469 ( 1996 ) 419 . s. gupta , _ phys .
_ , d 64 ( 2001 ) 034507 .
p. chakraborty , m. g. mustafa and m. h. thoma , hep - ph/0111022 , to appear in _ euro .
j. _ c. s. gupta , _ phys .
_ , d 60 ( 1999 ) 094505 .
braun - munzinger _
_ , _ phys .
_ , b 465 ( 1999 ) 43 ; + j. letessier and j. rafelski , _ nucl . phys .
_ , a 661 ( 1999 ) 97c
. f. becattini _ et al .
_ , _ phys .
_ c 64 ( 2001 ) 024901 .
u. heinz _
j. phys .
_ g 25 ( 1999 ) 263 ; + r. stock , _ phys .
_ b 456 ( 1999 ) 277 .
a. wroblewski , _ acta phys .
_ , b 16 ( 1985 ) 379 .
p. c. martin , in _ many body physics _ , proceedings of the 1967 les houches school , eds . c. dewitt and r. balian , gordon and breach , new york , 1968 ;
+ w. marshall and s. w. lovesey , _ theory of thermal neutron scattering _ , oxford university press , london , 1971 ; + a. l. fetter and j. d. walecka , _ quantum theory of many - particle systems _ , mcgraw - hill , new york , 1971 . j. cleymans , e - print hep - ph/0201142 . s. choe _ et al .
_ , e - print hep - lat/0110223 .
h. a. weldon , _ nucl .
phys . _ a 525 ( 1991 ) 405c ; + s. gupta , _ nucl .
phys . _ a 566 ( 1994 ) 69c .
k. rajagopal and f. wilczek , _ nucl .
_ , b 404 ( 1993 ) 577 . | we report the continuum limit of quark number susceptibilities in quenched qcd .
deviations from ideal gas behaviour at temperature @xmath0 increase as the lattice spacing is decreased from @xmath1 to @xmath2 , but a further decrease seems to have very little effect .
the measured susceptibilities are 20% lower than the ideal gas values , and also 10% below the hard thermal loop ( htl ) results .
the off - diagonal susceptibility is several orders of magnitude smaller than the htl results .
we verify a strong correlation between the lowest screening mass and the susceptibility .
we also show that the quark number susceptibilities give a reasonable account of the wroblewski parameter , which measures the strangeness yield in a heavy - ion collision . |
let @xmath1 be a symplectic manifold without boundary , and @xmath0 a compact connected lie group with lie algebra @xmath4 .
we fix a hamiltonian action of @xmath0 on @xmath5 and an ( equivariant ) moment map @xmath6 .
assume that the following hypothesis is satisfied : + ( * h ) * _ @xmath0 acts freely on @xmath7 and the moment map @xmath8 is proper . _
+ then the symplectic quotient @xmath9 is well - defined , smooth and closed .
we denote by @xmath10 ( @xmath11 ) the equivariant ( co-)homology of @xmath5 with integer coefficients ( equipped with the cup product @xmath12 ) , @xmath13 etc . , by @xmath14
the kirwan map , and by @xmath15^g\in h_g^2(m)$ ] the class of @xmath16 in the cartan model .
we define @xmath17 to be the set of maps @xmath18 such that @xmath19^g , b\ran\leq c\big\}\big|<\infty,\quad\forall c\in{\mathbb{r}},\ ] ] we equip @xmath17 with the convolution product @xmath20 , and call the triple @xmath21 the _ equivariant novikov ring_. we denote by @xmath22 the ( small ) quantum cohomology of @xmath23 with coefficients in @xmath17 , and by @xmath24 the space of @xmath0-invariant and @xmath25-compatible almost complex structures on @xmath5 . for @xmath26
we denote by @xmath27 the infinitesimal action .
we call @xmath1 _ equivariantly convex at @xmath28 _ iff there exists a proper @xmath0-invariant @xmath29 , @xmath30 and @xmath31 such that @xmath32 for every @xmath33 and @xmath34 . here
@xmath35 denotes the levi - civita connection of the metric @xmath36 .
a symplectic manifold @xmath37 is called semi - positive iff for every @xmath38 the conditions @xmath39,b\ran>0 $ ] and @xmath40 imply that @xmath41 ( see @xcite ) .
note that this holds for example if @xmath37 is weakly monotone or @xmath42 .
we call @xmath37 aspherical iff @xmath43 for every @xmath44 .
[ conj : qk g ] assume that ( h ) holds , @xmath1 is equivariantly convex at @xmath28 and aspherical , and @xmath23 is semi - positive ( see @xcite ) .
then there exists a @xmath17-algebra homomorphism @xmath45 of the form @xmath46 .
this conjecture ( without specification of the quantum coefficient ring involved ) was formulated by d. a. salamon .
the idea of proof outlined below is also due to him .
the present article is part of a project whose goal is to make this idea rigorous and hence prove the conjecture .
as an example , consider @xmath47 with the standard structure @xmath48 , and a linear action of @xmath0 .
then @xmath1 is equivariantly convex at @xmath28 ( see example 2.8 in @xcite ) and aspherical . assume that @xmath0 is the torus @xmath49 , and let @xmath50 be the weights of the action and @xmath51 be given by @xmath52 , for some @xmath53 . then @xmath8 is proper if and only if there exists @xmath54 such that @xmath55 , for @xmath56 ( see proposition 4.14 in @xcite ) .
the action of @xmath0 on @xmath7 is free if and only if for every subset @xmath57 the following holds . if there exist @xmath58 , for @xmath59 , such that @xmath60 then @xmath61 , for @xmath59 , generate the integral lattice in @xmath62 ( see lemma 5.20 in @xcite ) .
the quotient @xmath23 is semi - positive if for every @xmath54 the conditions @xmath63 , @xmath64 imply @xmath65 .
assume that the hypotheses of conjecture [ conj : qk g ] are satisfied , the action of @xmath0 on @xmath5 is monotone ( and hence @xmath23 is monotone ) , and @xmath66 is generated by classes of degree less than twice the minimal maslov number of this action . with these hypotheses
r. gaio and d. a. salamon @xcite proved a version of the conjecture involving the novikov ring of @xmath23 .
this was used by k. cieliebak and d. a. salamon @xcite to compute @xmath67 for monotone torus actions on @xmath68 with maslov number at least 2 . a potential application of conjecture [ conj : qk g ] is to extend these computations to the more general setting of this conjecture .
the idea of proof of conjecture [ conj : qk g ] is to define @xmath69 by counting symplectic vortices on the plane . to explain this
, we fix a @xmath0-invariant inner product @xmath70 on @xmath4 and identify @xmath62 with @xmath4 via @xmath70 .
let @xmath71 be a real surface equipped with an area form and a compatible complex structure , and @xmath72 a principal @xmath0-bundle .
we denote by @xmath73 the space of smooth equivariant maps from @xmath74 to @xmath5 , and by @xmath75 the space of smooth connections on @xmath74 .
a ( symplectic ) vortex is a solution @xmath76 of @xmath77 here @xmath78 denotes the complex anti - linear part of @xmath79 , which we think of as a one - form on @xmath80 with values in the complex vector bundle @xmath81 .
similarly , we view the curvature @xmath82 of @xmath83 as a two - form on @xmath80 with values in the adjoint bundle @xmath84 .
finally , we view @xmath85 as a section of @xmath86 . the vortex equations ( [ eq : vort p ] )
were discovered by k. cieliebak , a. r. gaio and d. a. salamon @xcite , and independently by i. mundet i riera @xcite .
the energy density and energy of a @xmath87 are given by @xmath88 and @xmath89 .
consider @xmath90 , equipped with the standard area form @xmath91 and complex structure @xmath92 , and @xmath93 .
let @xmath94 . the group @xmath95 of smooth gauge transformations on @xmath74 acts on the set of solutions of ( [ eq : vort p ] ) .
we denote by @xmath96 the set of gauge ( equivalence ) classes of vortices @xmath97 on @xmath74 for which @xmath98 , @xmath99 is compact and @xmath100 represent the equivariant homology class @xmath101 ( see @xcite ) .
we denote by @xmath102 a contractible topological space on which @xmath0 acts continuously and freely .
there are natural evaluation maps @xmath103 ( at @xmath104 ) and @xmath105 ( at @xmath106 ) ( see again @xcite ) .
heuristically , for @xmath107 and @xmath108 , we define @xmath109 we fix dual bases @xmath110 and @xmath111 of @xmath112 , in the sense that @xmath113 .
the idea is now to define a map @xmath114 as in ( [ eq : phi h g m ] ) by @xmath115 the goal of the present and two subsequent papers @xcite is to make this definition rigorous : in the present article i introduce weighted sobolev spaces @xmath116 and @xmath117 that will serve as model spaces for a banach manifold @xmath118 of gauge classes of pairs @xmath119 and the fiber of a banach bundle @xmath120 . furthermore , i show that the vertical differential of the equations ( [ eq : vort p ] ) ( viewed as a section @xmath121 of @xmath122 ) is a fredholm map between @xmath116 and @xmath117 . in @xcite
the notion of a stable map of vortices on @xmath2 and ( pseudo-)holomorphic spheres in @xmath123 is defined , and a bubbling result for vortices over @xmath2 is proven .
the hypotheses of equivariant convexity and asphericity are needed here .
the former forces vortices on @xmath2 to stay in a fixed compact subset of @xmath5 , and the latter rules out bubbling of holomorphic spheres in @xmath5 .
the present paper and @xcite are based on my ph.d .-
thesis @xcite . in @xcite ,
i define an atlas for @xmath118 consisting of charts with targets open subsets of the spaces @xmath116 ( and a similar atlas for @xmath122 ) .
the charts are based on the choice of a @xmath0-invariant riemannian metric on @xmath5 whose exponential map along @xmath7 is compatible with the moment map @xmath8 .
( this is needed to make the transition maps well - defined . )
furthermore , i show how to perturb @xmath121 in order to make it transverse to the zero section . combining all these results
, it follows that @xmath124 is a smooth finite dimensional submanifold , and the evaluation map @xmath125 is a pseudo - cycles . here
the semi - positivity of @xmath23 is needed .
the formula ( [ eq : qk g b ] ) is then made rigorous as an intersection number of pseudo - cycles . in order to show that @xmath3 intertwines @xmath12 with @xmath126
, it suffices to prove that @xmath127 for every @xmath128 , @xmath129 and @xmath130 .
the idea for proving this is the following .
we choose `` oriented submanifolds '' @xmath131 that are `` poincar dual '' to @xmath132 and @xmath133 , and an oriented submanifold @xmath134 representing @xmath135 .
( to make this rigorous one has to pass to some finite dimensional approximation of @xmath102 , a compact submanifold of @xmath5 with boundary and rational multiples of @xmath136 and @xmath135 . )
consider the marked points @xmath137 , @xmath138 , and a sequence of gauge classes of vortices @xmath139 , such that @xmath140 , @xmath141 , and @xmath142 . by the main result of @xcite a subsequence of @xmath143 converges to a stable map of ( gauge classes of ) vortices on @xmath2 and holomorphic spheres in @xmath123 . in the transverse case ,
this map consists of two classes @xmath144 and @xmath145 of vortices on @xmath2 , each equipped with a marked point @xmath146 , and a holomorphic map @xmath147 , equipped with a marked point @xmath148 .
@xmath149 is attached to @xmath150 at the nodal point @xmath151 , for @xmath152 .
the total homology class of the stable map equals @xmath101 , @xmath153 , for @xmath152 , and @xmath154 .
the number of such stable maps equals the right hand side of ( [ eq : ring ] ) .
consider now the marked points @xmath155 , @xmath138 , and a sequence @xmath139 , such that @xmath140 , @xmath141 and @xmath142 .
using @xcite again , a subsequence of @xmath143 converges to a stable map of vortices on @xmath2 and spheres in @xmath123 . in the transverse case ,
this stable map consists of a single class @xmath156 , satisfying @xmath157 , and a ghost bubble @xmath158 that is attached to @xmath159 at some point in @xmath2 and contains two marked points .
the number of such stable maps equals the left hand side of ( [ eq : ring ] ) .
equality ( [ eq : ring ] ) follows by combining this with the argument of the last paragraph and a gluing argument for vortices on @xmath2 and spheres in @xmath123 .
the gluing result is part of my future research .
a similar argument involving an adiabatic limit in the vortex equations over @xmath160 will show that @xmath3 intertwines the genus 0 symplectic vortex invariants with the gromov - witten invariants of @xmath23 .
the idea of proof of conjecture [ conj : qk g ] presented here is different from the construction used by gaio and salamon in @xcite .
they use an adiabatic limit argument , which fails in the more general situation considered here , because of bubbling off of vortices on @xmath2 .
assume now just that ( h ) holds and @xmath1 is equivariantly convex at @xmath28 .
we denote by @xmath161 the @xmath0-equivariant gromov - witten theory of @xmath1 . in joint work with christopher woodward @xcite
we interpret @xmath3 as a morphism of cohomological field theories between @xmath161 and @xmath162 .
we formulate `` functoriality '' for @xmath163 under reduction in stages .
we also conjecture quantum generalizations involving @xmath3 of non - abelian localization and abelianization . for a vector bundle @xmath164 and @xmath165
we denote by @xmath166 the space of its smooth sections and by @xmath167 the bundle of @xmath168-forms on @xmath169 with values in @xmath170 . for an almost complex manifold @xmath169 and a complex vector bundle @xmath164
is we denote by @xmath171 the bundle of anti - linear one - forms on @xmath169 with values in @xmath170 .
we equip the bundle @xmath172 with the complex structure induced by @xmath173 .
we define @xmath174 the group @xmath175 acts naturally on @xmath176 and @xmath177 .
we denote by @xmath178 and @xmath179 the quotients .
the map @xmath180 is @xmath175-equivariant , and hence induces a map @xmath181 .
note that @xmath182 is the set of gauge classes of vortices .
assume that the action of @xmath175 on @xmath176 is free .
( this is satisfied for @xmath90 , @xmath183 and @xmath92 under hypothesis ( h ) , see lemma [ le : free ] below . )
then heuristically , @xmath184 is an infinite dimensional manifold , @xmath185 is an infinite dimensional vector bundle over @xmath184 , and @xmath121 is a smooth section of @xmath185 .
assume that @xmath186 .
then formally , there is a canonical map @xmath187 , where @xmath188 denotes the fiber over @xmath159 .
the vertical differential of @xmath121 at @xmath159 is given by @xmath189 .
let @xmath190 be a representative .
we denote by @xmath191 the infinitesimal action of @xmath175 on @xmath176 .
formally , @xmath192 , @xmath193 , and @xmath194 .
assume that @xmath176 and @xmath195 are equipped with a @xmath175-invariant riemannian metric and @xmath175-invariant inner product respectively .
for @xmath196 we denote by @xmath197 the adjoint of @xmath198 .
then formally , the tangent space @xmath199 is the quotient of the disjoint union of @xmath200 , with @xmath100 ranging over all representatives of @xmath159 , by the linearized action of @xmath175 .
we now equip @xmath5 with the riemannian metric @xmath36 , and @xmath201 with the induced bundle metric .
the bundle @xmath86 inherits a bundle metric from @xmath70 .
we equip @xmath195 and @xmath202 with the corresponding @xmath203 inner products
. then @xmath204 is given by @xmath205 here we view @xmath206 as a map from @xmath86 to @xmath201 , and we denote by @xmath207 the adjoint of @xmath206 , and @xmath208 , where @xmath209 denotes the hodge - star operator on sections of @xmath86 with respect to the metric @xmath210 .
the levi - civita connection @xmath35 of @xmath36 and @xmath83 induce a connection @xmath211 on @xmath201 ( see section [ sec : back ] ) . for @xmath212
consider the operator @xmath213 , @xmath214 for @xmath186 the map @xmath215 is given by @xmath216 .
( this follows for example from @xcite , formula ( 23 ) , p. 27 . )
consider now the case @xmath217 and @xmath92 .
the purpose of this article is to find a banach space setup in which the vertical differential @xmath218 is fredholm . to this end , let @xmath219 , @xmath220 $ ] , @xmath221 , and @xmath222 be a measurable function .
we denote @xmath223 $ ] , and define @xmath224 .
assume now that @xmath225 .
we denote by @xmath226 and @xmath227 the spaces of @xmath0-equivariant maps from @xmath228 and connections on @xmath74 , of class locally @xmath229 .
we abbreviate @xmath230 , and we define @xmath231 furthermore , we denote by @xmath232 the group of locally @xmath233 gauge transformations on @xmath74 , and we define @xmath234 .
let @xmath235 be a smooth pair .
@xmath236 and the levi - civita connection of the standard metric @xmath237 on @xmath2 induce a linear connection @xmath211 on @xmath238 .
for @xmath239 we define @xmath240.\ ] ] here the pointwise norms are taken with respect to the metrics @xmath36 and @xmath237 , and @xmath70 . we define @xmath241 here @xmath204 is as in ( [ eq : l w * ] ) . in @xcite ,
the set @xmath118 will be equipped with a banach manifold structure , such that for every @xmath242 admitting a smooth representative @xmath243 , the tangent space @xmath244 can be identified with @xmath116 .
furthermore , the spaces @xmath117 will be identified with the fibers of a banach bundle @xmath120 . + * from now on throughout this article , we assume that hypothesis ( h ) is satisfied .
* + we denote by @xmath245 the maslov index of @xmath100 , see section [ sec : back ] .
the main result of this article is the following .
[ thm : fredholm ] assume that @xmath246 .
let @xmath225 , @xmath247 and @xmath235 be a smooth pair . then the following statements hold .
[ thm : fredholm : x y ] the normed vector spaces @xmath248 and @xmath249 are complete .
[ thm : fredholm : ddd ] if @xmath250 then the operator @xmath251 given by the formula ( [ eq : d w ] ) is well - defined and fredholm of real index @xmath252 .
the condition @xmath250 in this theorem captures the geometry of finite energy vortices .
more precisely , let @xmath253 be a finite energy vortex such that @xmath99 is compact . then for every @xmath254 there exists a constant @xmath255 such that @xmath256 , for every @xmath257 ( see @xcite , corollary 4 ) .
it follows that @xmath243 if @xmath258 . this bound is sharp . to see this ,
let @xmath259 , let @xmath260 and @xmath261 , and consider the inclusion @xmath262
. on the other hand , every @xmath263 satisfies @xmath98 if @xmath225 and @xmath247 . the latter condition is sharp .
namely , let @xmath264 , and consider @xmath265 with the standard symplectic form @xmath266 , complex structure @xmath261 and the action of the trivial group @xmath267 .
we choose a number @xmath268 and a smooth map @xmath269 such that @xmath270 , for @xmath257
. then @xmath271 and @xmath272 .
the condition @xmath258 is also needed for @xmath273 to have the right fredholm index . namely , let @xmath247 be such that @xmath274 , and @xmath243 . then the proof of theorem [ thm : fredholm ] shows that @xmath273 is fredholm with index equal to @xmath275 , where @xmath168 is the largest integer less than @xmath276 .
in particular , the index changes when @xmath277 passes the value @xmath278 .
the definition of the space @xmath248 is natural , since it parallels the definition of @xmath279 .
namely , by linearizing with respect to @xmath280 and @xmath83 the terms @xmath281 and @xmath85 occurring in the energy density @xmath282 , we obtain the terms @xmath283 and @xmath284 .
these expressions occur in @xmath285 , except for the factor @xmath206 in @xmath284 .
( it follows from ( h ) and lemma [ le : si ] ( appendix ) that this factor is irrelevant . ) the expression @xmath286 is needed in order to make @xmath287 non - degenerate .
* remark . *
_ naively , one could define the domain of @xmath288 to be the kernel of @xmath204 defined on the space of usual @xmath229-sections , and its target to consist of @xmath289-sections . however , then in general @xmath288 would not have closed image , and hence is not fredholm . note also that the 0-th order terms @xmath290 and @xmath291 in ( [ eq : d w ] ) are not compact ( neither with @xmath116 and @xmath117 defined as in ( [ eq : xx w],[eq : yy w ] ) nor the naive choices ) .
the reason is that the kondrachov compactness theorem fails on @xmath2 .
observe also that because of these terms , @xmath288 is not well - defined if we choose spaces that look like the standard ( weighted ) sobolev spaces in `` logarithmic '' coordinates @xmath292 ( with @xmath293 ) .
_ the proof of theorem [ thm : fredholm ] is based on a fredholm result for the augmented vertical differential ( theorem [ thm : fredholm aug ] ) and surjectivity of @xmath204 ( theorem [ thm : l w * r ] ) .
the proof of theorem [ thm : fredholm aug ] has two main ingredients .
the first one is a suitable complex trivialization of the bundle @xmath238 .
for @xmath294 large , @xmath295 and @xmath296 this trivialization respects the splitting @xmath297 , where @xmath298 denotes the complexified infinitesimal action , for @xmath26 . the second ingredient are two propositions stating that the standard cauchy - riemann operator @xmath299 and a related matrix differential operator are fredholm maps between suitable weighted sobolev spaces .
these results are based on the analysis of weighted sobolev spaces carried out by r. b. lockhart and r. c. mcowen .
( see @xcite and references therein . )
note that for a _ compact _ riemann surface @xmath80 without boundary , in @xcite k. cieliebak et al . proved that the augmented vertical differential of the vortex equations is fredholm .
section [ sec : back ] contains some background about the connection @xmath211 and the definition of the maslov index @xmath245 . in section [ subsec : reform ] a fredholm theorem for the augmented vertical differential ( theorem [ thm : fredholm aug ] ) , and an existence result for a right inverse for @xmath204 ( theorem [ thm : l w * r ] ) are stated .
furthermore , the main result is deduced from these results .
section [ subsec : proof : fredholm aug ] contains the core of the proof of theorem [ thm : fredholm aug ] . here the notion of a good complex trivialization is introduced and an existence result for such a trivialization is stated ( proposition [ prop : triv ] ) . furthermore , a result is formulated saying that every good trivialization transforms @xmath288 into a compact perturbation of the direct sum of @xmath299 and a matrix operator ( proposition [ prop : x x w ] ) .
the results of section [ subsec : proof : fredholm aug ] are proved in section [ subsec : proofs ] . in section [ subsec :
proof : thm : l w * r ] theorem [ thm : l w * r ] is proved , using the existence of a right inverse for @xmath300 ( proposition [ prop : right ] ) .
appendix [ sec : weighted ] contains a hardy - type inequality , which is used in the proof of proposition [ prop : x x w ] , some standard embedding and compactness results for weighted sobolev spaces , and fredholm results for the cauchy - riemann operator and a matrix valued operator on @xmath2 ( propositions [ prop : d l l ] and [ prop : d a b d ] and corollary [ cor : d l l ] ) . in appendix [ sec : proof : prop : right ] proposition [ prop : right ] is proved .
appendix [ sec : aux ] contains some other auxiliary results .
this article arose from my ph.d .- thesis .
i would like to thank my adviser , dietmar a. salamon , for the inspiring topic .
i highly profited from his mathematical insight .
i am very much indebted to chris woodward for his interest in my work , for sharing his ideas with me , and for his continuous encouragement .
it was he who coined the term `` quantum kirwan map '' .
i would also like to thank urs frauenfelder and kai cieliebak for stimulating discussions . finally , financial support of the dissertation from the swiss national science foundation is gratefully acknowledged .
let @xmath301 be a real ( smooth ) vector bundle .
we denote by @xmath302 the affine space of ( smooth linear ) connections on @xmath170 .
let @xmath303 .
let @xmath304 be a smooth manifold , and @xmath305 be a smooth map .
we denote by @xmath306 the pullback bundle .
the pullback connection @xmath307 is uniquely determined by @xmath308 , for every @xmath309 and every @xmath310 .
let @xmath0 be a lie group , @xmath311 a ( right-)principal @xmath0-bundle , and @xmath312 a @xmath0-equivariant vector bundle .
then the quotient @xmath313 has a natural structure of a vector bundle over @xmath169 .
let now @xmath301 be a @xmath0-equivariant vector bundle .
we denote by @xmath314 the space of @xmath0-invariant connections on @xmath170 .
we fix @xmath315 , @xmath316 , and @xmath317 .
we define @xmath318 by @xmath319 , for @xmath320 , @xmath321 , and @xmath322 .
we denote @xmath323 .
the connection @xmath324 is basic ( @xmath0-invariant and horizontal ) , hence there exists a unique @xmath325 with the following property .
let @xmath326 and @xmath327 .
we define @xmath328 , where @xmath329 is such that @xmath330 , and @xmath320 is a @xmath0-invariant section such that @xmath331 , for every @xmath321 .
assume now that @xmath169 is an open subset of @xmath332 , and let @xmath333 be a bundle map ( fixing the base ) .
we define @xmath334 by @xmath335 , for every @xmath336 , @xmath337 and @xmath338 .
( here we think of @xmath100 as a constant section of @xmath339 . )
let @xmath340 and @xmath173 be as in section [ sec : main ] , @xmath341 , @xmath342 a principal @xmath0-bundle , @xmath343 and @xmath344 .
the definition of the maslov index of @xmath100 is based on the following .
[ prop : p ] there exists an extension of @xmath74 to some smooth @xmath0-bundle @xmath345 , such that @xmath280 extends to a continuous map from @xmath346 to @xmath5 .
furthermore , if @xmath347 are two such extensions of @xmath74 and @xmath348 the corresponding extensions of @xmath280 , then there exists an isomorphism of continuous @xmath0-bundles @xmath349 such that @xmath350 .
the proof of proposition [ prop : p ] is postponed to the appendix ( page ) .
we choose @xmath346 and @xmath351 as in proposition [ prop : p ] .
then @xmath25 induces a fiberwise symplectic form @xmath352 on the continuous bundle @xmath353 .
[ defi : m w ] we define the maslov index @xmath245 to be the first chern number of @xmath354 .
it follows from proposition [ prop : p ] that @xmath245 does not depend on the choice of the extension @xmath346 .
note that it only depends on the gauge equivalence class of @xmath100 .
the condition @xmath247 is needed for @xmath245 to be well - defined for @xmath243 .
consider for example the case @xmath355 and @xmath267 .
let @xmath225 and @xmath356 .
we choose @xmath357 , and a smooth map @xmath358 such that @xmath359 , for @xmath257
. then @xmath360 , and @xmath361 diverges , as @xmath362 , for every @xmath363 .
therefore , we can not associate any maslov index with @xmath364 .
let @xmath340 and @xmath173 be as in section [ sec : main ] ( assuming hypothesis ( h ) ) , @xmath365 and @xmath342 a principal @xmath0-bundle . in the present section we always * assume that @xmath366*. let @xmath225 , @xmath367 , @xmath279 be defined as in ( [ eq : bb p lam ] ) , and @xmath235 be a smooth map .
we denote @xmath368 , and by @xmath369 the orthogonal projection onto @xmath370 .
@xmath371 induces an orthogonal projection @xmath372 onto @xmath373 .
for @xmath374 we write @xmath375 .
note that @xmath370 is in general not a subbundle of @xmath376 , since the dimension of @xmath377 may vary with @xmath26 .
for @xmath378 we define @xmath379 , where @xmath285 is as in ( [ eq : ze w p lam ] ) .
recall the definition ( [ eq : yy w ] ) of @xmath117 .
we define @xmath380 here @xmath381 is defined as in ( [ eq : d w ] ) .
note that the map @xmath382 given by @xmath383 is well - defined and bounded .
this follows from the fact @xmath384 ( for every @xmath26 ) and compactness of @xmath385 .
[ thm : fredholm aug ] let @xmath225 and @xmath386 be real numbers , and @xmath387 a smooth pair . then the following statements hold .
[ thm : fredholm x w ] the normed spaces @xmath388 and @xmath389 are complete . [
thm : fredholm dd w ] assume that @xmath390 .
then the operator @xmath391 is fredholm of real index @xmath392 this theorem is proved in section [ subsec : proof : fredholm aug ] .
the proof relies on the existence of a suitable trivialization of @xmath238 in which the operator @xmath393 becomes standard .
[ thm : l w * r ] let @xmath225 , @xmath247 , and @xmath394 .
then the map @xmath395 admits a bounded ( linear ) right inverse .
the proof of theorem [ thm : l w * r ] is postponed to section [ subsec : proof : thm : l w * r ] ( see page ) .
it is based on the existence of a bounded right inverse for the operator @xmath300 over a compact subset of @xmath332 diffeomorphic to @xmath396 ( proposition [ prop : right ] ) and the existence of a neighborhood @xmath397 of @xmath7 such that @xmath398 .
we define @xmath399 let @xmath225 , @xmath247 , and @xmath235 be a smooth pair .
we * prove ( [ thm : fredholm x w])*. [ claim : xx x ] we have @xmath400 , and the restriction of the norm @xmath401 to @xmath402 is equivalent to @xmath287 .
it suffices to prove that @xmath403 and this inclusion is bounded .
it follows from hypothesis ( h ) that there exists @xmath404 such that @xmath405 .
we have @xmath406 .
lemma [ le : si ] below implies that there exists @xmath407 such that @xmath408 .
let @xmath409 .
then @xmath410 , and thus , using the last assertion of remark [ rmk : c ] below , @xmath411 hence @xmath403 , and this inclusion is bounded .
this proves claim [ claim : xx x ] .
part ( [ thm : fredholm x w ] ) follows from part ( [ thm : fredholm : x y ] ) of theorem [ thm : fredholm aug ] and claim [ claim : xx x ] . *
part ( [ thm : fredholm : ddd ] ) * follows from part ( [ thm : fredholm dd w ] ) of theorem [ thm : fredholm aug ] , theorem [ thm : l w * r ] and lemma [ le : x y z ] ( appendix ) with @xmath412 , @xmath413 given by ( [ eq : d w ] ) and @xmath414 .
this proves theorem [ thm : fredholm ] .
we denote by @xmath415 and @xmath416 the standard coordinates in @xmath2 . for @xmath417
we denote @xmath418 .
for @xmath419 we define @xmath420 , @xmath421 .
we equip the bundle @xmath422 with the ( fiberwise ) complex structure @xmath423 defined by @xmath424 .
furthermore , we denote @xmath425 , @xmath426 , and for @xmath427 we denote by @xmath428 the map @xmath429 . for @xmath26 we write @xmath430 for the complex linear extension of @xmath431 .
we define @xmath432 note that in general , the union @xmath433 of all the @xmath434 s is not a smooth subbundle of @xmath376 , since the dimension of @xmath434 may depend on @xmath435 .
however , there exists an open neighborhood @xmath397 of @xmath7 such that @xmath436 is a subbundle of @xmath437 .
let @xmath225 , @xmath386 and @xmath235 be a smooth pair .
for @xmath438 we define @xmath439 .
consider a complex trivialization ( i.e. bundle isomorphism fixing the base @xmath2 ) @xmath440 [ defi : triv ] we call @xmath441 _ good _ , if the following properties are satisfied . *
( splitting)*[defi : triv split ] for every @xmath438 we have @xmath442 furthermore , there exists a number @xmath407 , a smooth section @xmath443 of @xmath444 , and a point @xmath445 , such that the following conditions are satisfied . for every @xmath295 we have @xmath446 @xmath447 converges to @xmath448 , uniformly in @xmath363 , as @xmath362 , @xmath449 , and for every @xmath295 and @xmath450 , we have @xmath451 [ defi : triv c ] there exists a number @xmath452 such that for every @xmath453 @xmath454 [ defi : triv na ] we have @xmath455 .
[ prop : triv ] if @xmath225 , @xmath386 and @xmath235 is smooth then there exists a good complex trivialization of @xmath238 .
the proof of this proposition is postponed to subsection [ subsec : proofs ] ( page ) .
the next result shows that a good trivialization transforms @xmath456 into some standard operator .
for @xmath457 we denote @xmath458 and @xmath459 .
let @xmath460 , @xmath219 , @xmath165 , @xmath367 , @xmath461 an open subset , @xmath159 a real or complex vector space , and @xmath462 a @xmath168-times weakly differentiable map .
we define @xmath463,\\ \nn\vert u\vert_{w^{k , p}_\lam(\om , w)}&:=&\sum_{|\al|\leq k}\vert\langle\cdot\rangle^\lam\d^\al u\vert_{l^p(\om , w)}\in[0,\infty],\\ \label{eq : l k p}l^{k , p}_\lam(\om , w)&:=&\big\{u\in w^{k , p}_\loc(\om , w)\,|\,\vert u\vert_{l^{k , p}_\lam(\om , w)}<\infty\big\}\\ \label{eq : w k p } w^{k , p}_\lam(\om , w)&:=&\big\{u\in w^{k , p}_\loc(\om , w)\,|\,\vert u\vert_{w^{k , p}_\lam(\om , w)}<\infty\big\}.\end{aligned}\ ] ] if @xmath464 , @xmath465 , are normed vector spaces then we endow @xmath466 with the norm @xmath467 .
let @xmath468 . if @xmath469 then we choose @xmath470)$ ] such that @xmath471 for @xmath472 and @xmath473 for @xmath474 .
in the case @xmath475 we set @xmath476 .
the isomorphism of lemma [ le : x d iso ] ( appendix ) induces norms on @xmath477 and @xmath478 .
we define @xmath479 @xmath480 for a complex vector space @xmath159 we denote by @xmath481 ( @xmath482 ) the operator @xmath483 ( @xmath484 ) acting on functions from @xmath485 to @xmath159 .
we denote by @xmath486 the hermitian inner product on @xmath487 ( complex anti - linear in its first argument ) extending @xmath70 .
we define @xmath488 by @xmath489 and @xmath490 , and @xmath491 .
[ prop : x x w ] let @xmath492 , @xmath386 , @xmath493 be smooth and @xmath494 a good trivialization . then the following statements hold . [
prop : x x w iso ] the following maps are well - defined isomorphisms of normed spaces : @xmath495 [ prop : x x w fredholm ] there exists a positive @xmath485-linear map @xmath496 ( i.e. @xmath497 for every @xmath498 ) such that the following operator is compact : @xmath499 the proof of proposition [ prop : x x w ] is postponed to subsection [ subsec : proofs ] ( page ) . it is based on some inequalities and compactness properties for weighted sobolev spaces ( proposition [ prop : lam d morrey ] ) and a hardy - type inequality ( proposition [ prop : hardy ] ) . [
proof : fredholm aug ] let @xmath225 , @xmath386 , and let @xmath500 be a smooth pair .
the space ( [ eq : l k p ] ) is complete , see @xcite . by proposition
[ prop : lam d morrey]([prop : lam w ] ) ( appendix ) the same holds for the space ( [ eq : w k p ] ) . combining this with propositions [ prop : triv ] and [ prop :
x x w]([prop : x x w iso ] ) , * part ( [ thm : fredholm x w ] ) * follows . *
part ( [ thm : fredholm dd w ] ) * follows from propositions [ prop : triv ] and [ prop : x x w]([prop : x x w fredholm ] ) , corollary [ cor : d l l ] and proposition [ prop : d a b d ] ( appendix ) .
this proves theorem [ thm : fredholm aug ] .
[ rmk : fredholm aug ] an alternative approach to prove theorem [ thm : fredholm aug ] is to switch to `` logarithmic '' coordinates @xmath292 ( defined by @xmath501 ) . in these coordinates and a suitable trivialization
the operator @xmath393 is of the form @xmath502 .
hence one can try to apply the results of @xcite . however , this is not possible , since @xmath503 contains the operator @xmath504 , which diverges for @xmath505 .
[ proof : triv ] let @xmath506 and @xmath100 be as in the hypothesis .
we choose a section @xmath443 of @xmath507 and a point @xmath445 as in lemma [ le : si ] .
[ claim : u triv ] there exists an open @xmath0-invariant neighborhood @xmath397 of @xmath448 such that @xmath436 is a smooth subbundle of @xmath376 with the following property .
there exists a smooth complex trivialization @xmath508 satisfying @xmath509 , for every @xmath510 , @xmath511 and @xmath512 . by hypothesis ( h )
we have @xmath513 , where @xmath514 is defined as in ( [ eq : m * ] ) .
we choose a @xmath0-invariant neighborhood @xmath515 of @xmath448 so small that @xmath516 and @xmath517 intersect transversely , for every @xmath518 .
then @xmath519 is a smooth subbundle of @xmath520 .
furthermore , by the local slice theorem there exists a pair @xmath521 , where @xmath522 is a @xmath0-invariant neighborhood of @xmath448 and @xmath523 is a submanifold of dimension @xmath524 that intersects @xmath525 transversely in exactly one point , for every @xmath511 .
we choose a complex trivialization of @xmath526 and extend it in a @xmath0-equivariant way , to obtain a trivialization @xmath527 of @xmath436 .
this proves claim [ claim : u triv ] .
we choose @xmath528 and @xmath527 as in claim [ claim : u triv ] .
it follows from lemma [ le : si ] that there exists @xmath529 such that @xmath530 , for @xmath296 , if @xmath295 .
we define @xmath531 by @xmath532 this is a smooth complex trivialization of @xmath533 . [
claim : wt psi 1 ] @xmath534 extends to a smooth complex trivialization of @xmath201 .
we define @xmath535 by @xmath536 .
we choose a complex trivialization @xmath537 .
we define @xmath538 by @xmath539 for a continuous map @xmath540 we denote by @xmath541 its degree . [
claim : d m(w ) ] the index @xmath245 equals @xmath542 .
we define @xmath346 to be the quotient of @xmath543 under the equivalence relation generated by @xmath544 , where @xmath510 is determined by @xmath545 , for @xmath546 , @xmath547 .
furthermore , we define @xmath548 by @xmath549):=u(p)$ ] , for @xmath321 , and @xmath550):=g^{-1}x_\infty$ ] , for @xmath510 .
the statement of lemma [ le : si ] implies that this map is continuous and extends @xmath280 .
the ( fiberwise linear ) complex structure @xmath551 on @xmath552 descends to a complex structure @xmath553 on @xmath353 . by definition
, we have @xmath554 .
we define the local trivialization @xmath555 by @xmath556,\psi^u_{u\circ\si(z)}v_0+l^\c_{u\circ\si(z)}\al\big),&\textrm{if } z\in { \mathbb{r}}^2\wo b_r,\\ g\cdot\big([\infty,\one],\psi^u_{x_\infty}v_0+l^\c_{x_\infty}\al\big),&\textrm{if } z=\infty .
\end{array}\right.\ ] ] then @xmath557 , for @xmath558 , and therefore @xmath559 is the transition map between @xmath560 and @xmath561 .
claim [ claim : d m(w ) ] follows from this . by claim [ claim : d m(w ) ] and lemma [ le : aut ] in the appendix the maps @xmath559 and @xmath562 are homotopic .
hence there exists a continuous map @xmath563 such that @xmath564 , if @xmath565 , and @xmath566 , if @xmath558 .
we define @xmath567 by @xmath568 on @xmath569 , and @xmath570 , for @xmath571 , @xmath572 . smoothing @xmath573 out on the ball @xmath574
, we obtain the required extension of @xmath534 .
this proves claim [ claim : wt psi 1 ] .
we define @xmath575 by @xmath576 [ claim : hhat psi 1 ] @xmath577 extends to a smooth complex trivialization of the bundle @xmath578 .
we denote by @xmath579 and @xmath580 the adjoint representations of @xmath0 on @xmath4 and @xmath487 respectively . for every @xmath510 we have @xmath581 .
we choose a continuous section @xmath582 of the restriction @xmath583 .
we define @xmath584 to be the unique map such that @xmath585 , for every @xmath558 .
it follows that @xmath586
. therefore , @xmath587 .
hence lemma [ le : aut ] ( appendix ) implies that there exists a continuous map @xmath588 satisfying @xmath589 , for every @xmath558 .
we define @xmath590 by @xmath591 on @xmath569 and by @xmath592 , where @xmath593 , for @xmath571 , @xmath594 .
smoothing @xmath595 out on the ball @xmath574 , we obtain the required extension of @xmath577 .
this proves claim [ claim : hhat psi 1 ] .
we choose extensions @xmath573 and @xmath595 of @xmath596 and @xmath597 as in claims [ claim : wt psi 1 ] and [ claim : hhat psi 1 ] .
[ claim : psi good ] @xmath598 is a good complex trivialization of @xmath238 .
condition * ( [ defi : triv split ] ) * of definition [ defi : triv ] follows from the construction of @xmath441 . to prove * ( [ defi : triv c ] ) * , note that for @xmath599 and @xmath600 , we have @xmath601 here we used the fact @xmath602 , for every @xmath26 . by our choice of @xmath528
, @xmath603 is a smooth subbundle of rank @xmath604 .
it follows that @xmath605 is a smooth subbundle of @xmath437 of rank @xmath606 .
hence @xmath430 is injective , for every @xmath511 .
since by assumption @xmath99 is compact , the same holds for the set @xmath607 .
it follows that there exists a constant @xmath452 such that @xmath608 for every @xmath609 , @xmath610 and @xmath611 . combining this with equality ( [ eq : psi z z ] ) , condition ( [ defi : triv c ] ) follows .
we check condition * ( [ defi : triv na])*. let @xmath612 , @xmath609 and @xmath613 .
we choose a point @xmath546 and a vector @xmath322 such that @xmath330 .
then @xmath614 furthermore , for every smooth vector field @xmath169 on @xmath528 we have @xmath615 we define @xmath255 to be the maximum of @xmath616 over all @xmath617 , @xmath618 and @xmath619 such that @xmath620 , @xmath621 . furthermore , we define @xmath622 . by ( [ eq : na a v wt psi ] ) and ( [ eq : wt na a wt v x ] ) with @xmath623 , we have @xmath624 we now define @xmath625 to be the unique equivariant maps such that @xmath626 , @xmath627 .
we have @xmath628 $ ] , and similarly for @xmath629 .
since @xmath630 , it follows that @xmath631\big|.\ ] ] combining this with ( [ eq : vert na a v wt psi ] ) and the facts @xmath632 and @xmath633 , condition ( [ defi : triv na ] ) follows .
this proves claim [ claim : psi good ] and concludes the proof of proposition [ prop : triv ] .
[ proof : x x w ] let @xmath634 and @xmath441 be as in the hypothesis .
we choose @xmath470)$ ] such that @xmath471 for @xmath635 and @xmath473 for @xmath257 .
we fix @xmath636 , @xmath443 and @xmath448 as in definition [ defi : triv]([defi : triv split ] ) .
we abbreviate @xmath637 . *
we prove ( [ prop : x x w iso ] ) .
* for every @xmath638 leibnitz rule implies that @xmath639 [ claim : f x bdd ] the first map in ( [ eq : x y psi ] ) is well - defined and bounded .
proposition [ prop : lam d morrey]([prop : lam morrey ] ) below and the fact @xmath386 imply that there exists a constant @xmath640 such that @xmath641 we choose a constant @xmath642 as in part ( [ defi : triv c ] ) of definition [ defi : triv ] . then by ( [ eq : c ] ) and ( [ eq : lan ran - d i d ze ] ) , we have @xmath643 it follows from ( [ eq : psi infty x h ] ) and ( [ eq : psi 0 al be ] ) , the definition @xmath644 and the compactness of @xmath385 that there exists @xmath645 such that , for every @xmath646 , @xmath647 where @xmath648 . for @xmath649 we denote @xmath650 and @xmath651 .
we define @xmath652 . by condition ( [ defi : triv na ] ) of definition [ defi : triv ]
we have @xmath653 .
let @xmath646 .
then by ( [ eq : na phi psi f ze ] ) we have @xmath654 we define @xmath655
. then @xmath656 by proposition [ prop : lam d morrey]([prop : lam d iso ] ) . combining this with ( [ eq : na a psi ze p lam ] ) and
( [ eq : lan ran - d i d ze ] ) , we get @xmath657 by a direct calculation there exists a constant @xmath658 such that @xmath659 , for every @xmath646 .
claim [ claim : f x bdd ] follows from this and ( [ eq : psi ze infty],[eq : d mu u v],[eq : na a psi ze ] ) .
[ claim : f -1 x bdd ] the map @xmath660 is well - defined and bounded .
we choose a neighborhood @xmath397 of @xmath7 as in lemma [ le : u ] ( appendix ) , and define @xmath661 as in ( [ eq : c inf ] ) , and @xmath662 .
since @xmath447 converges to @xmath448 , uniformly in @xmath363 , as @xmath362 , there exists @xmath663 such that @xmath530 , for every @xmath664
. then ( [ eq : psi infty x h],[eq : psi 0 al be ] ) and ( [ eq : c inf ] ) imply that @xmath665 where @xmath666 , for every @xmath667 .
[ claim : d rho 0 p ] there exists a constant @xmath668 such that for every @xmath669 , we have @xmath670 it follows from equality ( [ eq : na phi psi f ze ] ) and conditions ( [ defi : triv c ] ) and ( [ defi : triv na ] ) of definition [ defi : triv ] that there exist constants @xmath255 and @xmath671 such that @xmath672 for every @xmath669 . on the other hand ,
leibnitz rule implies @xmath673 hence by a short calculation , using leibnitz rule again , it follows that there exists a constant @xmath674 such that @xmath675 for every @xmath676 . combining this with ( [ eq : p - d i d f ] ) , claim [ claim : d rho 0 p ] follows .
let @xmath669 .
we denote @xmath677 . by inequality ( [ eq : d rho 0 p ] ) the hypotheses of proposition [ prop : hardy ] with @xmath678 and @xmath277 replaced by @xmath679 are satisfied .
it follows that there exists @xmath680 , such that @xmath681 , uniformly in @xmath363 , as @xmath362 , and @xmath682 since @xmath386 , we have @xmath683 . hence the convergence @xmath684 and the estimate ( [ eq : vert al be vert ] ) imply that @xmath685 .
we choose a constant @xmath452 as in part ( [ defi : triv c ] ) of definition [ defi : triv ] .
the convergence @xmath686 and the first inequality in ( [ eq : c ] ) imply that @xmath687 we define @xmath688 .
proposition [ prop : lam d morrey]([prop : lam d iso ] ) in the appendix and inequalities ( [ eq : wt ze ze infty ] ) and ( [ eq : d rho 0 p ] ) imply that there exists a constant @xmath658 ( depending on @xmath689 and @xmath441 , but not on @xmath690 ) such that @xmath691 finally , by a straight - forward argument , there exists a constant @xmath692 ( independent of @xmath690 ) such that @xmath693 .
combining this with ( [ eq : vert al be vert],[eq : v infty 1],[eq : v 1 v 2 v bar n ] ) , claim [ claim : f -1 x bdd ] follows .
claims [ claim : f x bdd ] and [ claim : f -1 x bdd ] imply that the first map in ( [ eq : x y psi ] ) is an isomorphism ( of normed vector spaces ) .
it follows from condition ( [ defi : triv c ] ) of definition [ defi : triv ] that the second map in ( [ eq : x y psi ] ) is an isomorphism .
this completes the proof of * ( [ prop : x x w iso])*. * we prove statement ( [ prop : x x w fredholm ] ) . *
recall that we have chosen @xmath694 and @xmath448 as in definition [ defi : triv]([defi : triv split ] ) .
we define @xmath695 to be the complex linear extension of @xmath696 . by our hypothesis ( h )
the lie group @xmath0 acts freely on @xmath7 .
it follows that @xmath697 is injective .
therefore @xmath698 is positive with respect to @xmath486 . by ( [ eq : psi z c bar n ] ) and ( [ eq :
psi z 0 ] ) there exist complex trivializations @xmath699 such that @xmath700 .
we denote by @xmath701 ( @xmath702 ) the inclusion as ( the projection onto ) the second factor .
we define @xmath703 @xmath704 @xmath705 note that @xmath706 and @xmath707 .
we define @xmath708 as in ( [ eq : f psi dd ] ) .
since @xmath709 is finite dimensional , @xmath710 is compact .
hence it suffices to prove that @xmath711 is compact .
to see this , we denote @xmath712 and we define @xmath713 ( for @xmath714 ) and @xmath715 by @xmath716 @xmath717 @xmath718 here @xmath719 and @xmath720 are as in ( [ eq : f 1 f 2 ] ) and @xmath721 , for @xmath611 ( viewed as a constant section of @xmath722 ) .
a direct calculation shows that @xmath723 . for a subset @xmath724 we denote by @xmath725 its characteristic function .
it follows that @xmath726 is of 0-th order .
since it vanishes outside @xmath727 , it follows that this map is compact .
[ claim : s i j compact ] the operators @xmath728 , @xmath714 , and @xmath729 are compact . to see that the map @xmath730 is compact ,
note that leibnitz rule and holomorphicity of @xmath731 imply that @xmath732 since @xmath247 , assertions ( [ prop : lam d iso ] ) and ( [ prop : lam morrey ] ) of proposition [ prop : lam d morrey ] imply that the map @xmath733 is well - defined and compact . by condition ( [ defi : triv na ] ) of definition [ defi : triv ] , the map @xmath734 is bounded .
condition ( [ defi : triv c ] ) of definition [ defi : triv ] implies boundedness of the map @xmath735 .
compactness of @xmath730 follows . by the definition of @xmath279
, we have @xmath736 .
this together with proposition [ prop : lam d morrey]([prop : lam d iso ] ) and ( [ prop : lam morrey ] ) and definition [ defi : triv]([defi : triv c ] ) implies that the map @xmath737 is compact .
furthermore , it follows from definition [ defi : triv]([defi : triv split ] ) that @xmath738 . to see that @xmath739 is compact , we define @xmath740 by setting @xmath741 to be the complex linear extension of the map @xmath742 . since @xmath447 converges to @xmath448 , uniformly in @xmath69 , as @xmath362 , the map @xmath743 converges to 0 , uniformly in @xmath69 , as @xmath744 . hence by proposition [ prop : lam d morrey]([prop : lam w cpt ] ) ,
the map @xmath745 is compact .
definition [ defi : triv]([defi : triv split ] ) implies that @xmath746 .
it follows this map is compact .
finally , proposition [ prop : lam d morrey]([prop : lam morrey ] ) and parts ( [ defi : triv na ] ) and ( [ defi : triv c ] ) of definition [ defi : triv ] imply that the map @xmath747 is compact .
claim [ claim : s i j compact ] follows .
this completes the proofs of statement * ( [ prop : x x w fredholm ] ) * and proposition [ prop : x x w ] .
let @xmath219 , @xmath748 , @xmath749 , @xmath0 a compact lie group , @xmath70 an invariant inner product on @xmath750 , @xmath751 an open subset , @xmath752 a principal @xmath0-bundle and @xmath753 .
then @xmath83 and the standard metric on @xmath754 induce a connection @xmath211 on @xmath755 .
for @xmath756 we define @xmath757 where the pointwise norms are taken with respect to @xmath70 .
we denote @xmath758 for a subset @xmath759 we denote by @xmath760 its interior . for the proof of theorem [ thm : l w * r ]
we need the following .
[ prop : right ] let @xmath761 and @xmath70 be as above and @xmath762 a compact subset diffeomorphic to @xmath396 .
then for every @xmath0-bundle @xmath763 and @xmath753 there exists a bounded right inverse @xmath294 of the operator @xmath764 furthermore , there exist @xmath254 and @xmath452 such that for every principal @xmath0-bundle @xmath763 and every @xmath753 satisfying @xmath765 the map @xmath294 can be chosen such that @xmath766 .
the proof of proposition [ prop : right ] is postponed to the appendix ( page ) .
[ proof : thm : l w * r ] let @xmath506 and @xmath767 be as in the hypothesis .
it follows from hypothesis ( h ) that there exists @xmath404 such that @xmath405 ( defined as in ( [ eq : m * ] ) ) .
we define @xmath768 .
it follows from lemma [ le : si ] that there exists a number @xmath769 such that @xmath770 , for every @xmath771 .
we choose constants @xmath772 and @xmath640 as in the second assertion of proposition [ prop : right ] ( corresponding to @xmath773 and @xmath255 , for @xmath774 ) .
furthermore , we choose constants @xmath775 and @xmath668 as in lemma [ le : infty 1 p a ] ( corresponding to @xmath773 and @xmath255 ) .
we define @xmath776 . by assumption
we have @xmath777 .
hence there exists an integer @xmath778 such that @xmath779 .
we choose a smooth map @xmath780\to[0,1]$ ] such that @xmath781 on @xmath782\cup[3/4,1]$ ] , @xmath783 on @xmath784 $ ] , and @xmath785 and @xmath786 , for all @xmath787 $ ] .
we choose a bijection @xmath788 .
we define @xmath789 $ ] by @xmath790 and @xmath791 $ ] by @xmath792 .
furthermore , for @xmath793 we define @xmath794 $ ] by @xmath795 .
we choose a compact subset @xmath796 ^ 2 $ ] diffeomorphic to @xmath396 , such that @xmath797 ^ 2\sub \int k_0 $ ] , and we denote @xmath798 .
furthermore , we choose a compact subset @xmath799 ^ 2 $ ] diffeomorphic to @xmath396 , such that @xmath800\sub \int k$ ] .
for @xmath793 we define @xmath801 . for @xmath802
we define @xmath803 . by the first assertion of proposition [ prop :
right ] there exists a bounded right inverse @xmath804 of @xmath805 .
we fix @xmath793 .
since @xmath806 , by our choice of @xmath304 we have @xmath807 .
hence it follows from the statement of proposition [ prop : right ] that there exists a right inverse @xmath808 of @xmath809 , satisfying @xmath810 we define @xmath811 each section @xmath812 induces a section @xmath813 . for @xmath814
we have @xmath815 , and therefore the map @xmath816 is invertible .
hence we may define @xmath817 for @xmath818 , and @xmath819 , for @xmath820 .
furthermore , we define @xmath821 by @xmath822 .
it follows that @xmath823 .
theorem [ thm : l w * r ] is now a consequence of the following : [ claim : ze x w ] @xmath294 restricts to a bounded map from @xmath389 to @xmath824 .
we choose a constant @xmath825 so big that @xmath826 , for every @xmath802 . for a weakly differentiable section @xmath827
we denote @xmath828 .
[ claim : d a * wt al ] there exists a constant @xmath829 such that @xmath830 , for every @xmath831 .
let @xmath831 .
we denote @xmath832 and @xmath833 .
since @xmath834 , a straight - forward calculation shows that @xmath835 fix @xmath438 .
then @xmath836 .
hence ( [ eq : d a * wt al ] ) implies that @xmath837 inequalities ( [ eq : d a * p ] ) and ( [ eq : r i 1 p ] ) imply that @xmath838 equality ( [ eq : d a * wt al ] ) implies that @xmath839 combining this with ( [ eq : r i 1 p ] ) , claim [ claim : d a * wt al ] follows .
we choose @xmath829 as in claim [ claim : d a * wt al ] .
let @xmath831 .
we abbreviate @xmath840 . by the fact @xmath841 , lemma [ le : infty 1 p a ] , the fact @xmath806 and claim [ claim : d a * wt al ] , we have @xmath842 recall that @xmath369 and @xmath372 denote the orthogonal projections onto @xmath370 and @xmath373 .
claim [ claim : ze x w ] is now a consequence of the following three claims .
[ claim : hat r wt r ] we have @xmath843 let @xmath831 be such that @xmath844 . for @xmath802
we denote @xmath832 .
since @xmath845 , for every @xmath438 , we have @xmath846 . using ( [ eq : r i 1 p ] ) , it follows that @xmath847 .
we define @xmath848 . by remark [ rmk : c ] below
, the statement of claim [ claim : d a * wt al ] and the fact @xmath849 , we have @xmath850 inequality ( [ eq : wt v infty ] ) and remark [ rmk : c ] imply that @xmath851 .
we fix @xmath793 .
lemma [ le : infty 1 p a ] , ( [ eq : r i 1 p ] ) and the fact @xmath806 imply that @xmath852 .
furthermore , @xmath853 it follows that @xmath854 .
claim [ claim : hat r wt r ] follows .
[ claim : na a p lam ] we have @xmath855 .
let @xmath831 .
we define @xmath840 , @xmath856 and @xmath857 as in ( [ eq : xi rho v v ] ) . by lemma [ le : na a l u ] below , we have @xmath858 using the second part of lemma [ le : na a l u ] ( with @xmath859 ) , it follows that @xmath860 we choose a constant @xmath255 so big that @xmath861 and @xmath862 , for every @xmath863 , @xmath864 and @xmath865 .
we define @xmath866 .
since @xmath867 , equalities ( [ eq : na a l u eta],[eq : d a wt xi ] ) imply that @xmath868 here we used remark [ rmk : c ] . since @xmath869 , claim [ claim : d a * wt al ] and ( [ eq : wt v infty ] ) now imply claim [ claim : na a p lam ] .
[ claim : na a al p lam ] we have @xmath870 .
let @xmath831 be such that @xmath871 .
we write @xmath832
. then @xmath872 . setting @xmath873
, it follows that @xmath874 here in the second inequality we used the fact @xmath875 , and ( [ eq : r i 1 p ] ) .
this proves claim [ claim : na a al p lam ] , and completes the proofs of claim [ claim : ze x w ] and theorem [ thm : l w * r ] .
let @xmath419 . the following lemma is used in section [ subsec : proof : fredholm aug ] in order to define a norm on @xmath876 . if @xmath469 then let @xmath470)$ ] be such that @xmath471 for @xmath472 and @xmath473 for @xmath474 . in the case
@xmath475 we set @xmath476 .
recall that @xmath420 , @xmath421 .
[ le : x d iso ] for every @xmath877 and @xmath878 the map @xmath879 is an isomorphism of vector spaces .
this follows from a straight - forward argument .
the following proposition was used in the proof of theorem [ thm : fredholm aug ] and proposition [ prop : x x w ] . for every normed vector space @xmath880 we denote by @xmath881 the space of bounded continuous maps from @xmath332 to @xmath880 .
we denote @xmath882 .
[ prop : lam d morrey ] let @xmath219 .
then the following statements hold . [
prop : lam morrey ] let @xmath883 .
then for every @xmath367 there exists @xmath452 such that @xmath884 if @xmath885 then @xmath886 is compactly contained in @xmath887 .
[ prop : lam w ] for every @xmath165 , @xmath877 and @xmath367 the map @xmath888 is a well - defined isomorphism ( of normed spaces ) .
[ prop : lam w cpt ] let @xmath889 , @xmath367 , and @xmath890 be such that @xmath891 , for @xmath892
. then the map @xmath893 is compact .
[ prop : lam d iso ] for every @xmath877 , @xmath367 , @xmath468 and @xmath894 we have @xmath895 [ proof of proposition [ prop : lam d morrey ] ] inequality ( [ eq : lam morrey ] ) follows from inequality ( 1.11 ) in theorem 1.2 in the paper by r. bartnik @xcite .
assume now that @xmath885 .
then it follows from morrey s embedding theorem that there exists a canonical bounded inclusion @xmath896 . in order to show that this inclusion is compact ,
let @xmath897 be a sequence such that @xmath898 .
by kondrachov s compactness theorem on @xmath899 ( for @xmath900 ) , and a diagonal subsequence argument there exists a subsequence @xmath901 of @xmath902 that converges to some map @xmath903 , weakly in @xmath904 , and strongly in @xmath905 , for every @xmath900 .
[ claim : u u nu j ] we have @xmath906 and @xmath901 converges to @xmath280 in @xmath887 .
we choose a constant @xmath671 as in the first part of ( [ prop : lam morrey ] ) . for every @xmath407 we have @xmath907 .
hence @xmath908 .
since @xmath885 , by inequality ( [ eq : lam morrey ] ) , this implies @xmath906 . to see the second statement , we choose a smooth function @xmath909 $ ] such that @xmath910 for @xmath911 , @xmath912 for @xmath913 , and @xmath914 .
let @xmath636 .
we define @xmath915 $ ] .
let @xmath900 .
abbreviating @xmath916 , we have @xmath917 inequality ( [ eq : lam morrey ] ) implies that @xmath918 furthermore , @xmath919 . combining this with ( [ eq : vert v j ] ) and ( [ eq : vert v j rho r ] ) , and the fact @xmath920
, it follows that @xmath921 .
since @xmath885 and @xmath636 is arbitrary , it follows that @xmath901 converges to @xmath280 in @xmath887 .
this proves claim [ claim : u u nu j ] and completes the proof of statement * ( [ prop : lam morrey])*. * statement ( [ prop : lam w ] ) * follows from a straight - forward calculation .
* proof of ( [ prop : lam w cpt ] ) : * let @xmath890 be as in the hypothesis .
let @xmath922 be a sequence such that @xmath923 . by kondrachov
s theorem on @xmath899 ( for @xmath900 ) and a diagonal subsequence argument there exists a subsequence @xmath924 and a map @xmath925 , such that @xmath926 converges to @xmath927 , strongly in @xmath928 , as @xmath929 , for every compact subset @xmath762 .
standard arguments show that @xmath930 and @xmath926 converges to @xmath927 in @xmath931 .
this proves * ( [ prop : lam w cpt])*. * statement ( [ prop : lam d iso ] ) * follows from a straight - forward calculation .
this completes the proof of proposition [ prop : lam d morrey ] .
the next result was used in the proof of proposition [ prop : x x w ] and will be used in the proof of lemma [ le : si ] .
[ prop : hardy ] let @xmath219 , @xmath749 , @xmath885 and @xmath932 be such that @xmath933
. then @xmath934 converges to some @xmath935 , uniformly in @xmath936 , as @xmath362 , and @xmath937 for the proof of proposition [ prop : hardy ] , we need the following . [
le : hardy ] let @xmath219 , @xmath877 , @xmath885 and @xmath932 .
if there exists @xmath407 such that @xmath938 then @xmath939)$ ] .
if @xmath280 is smooth then the stated inequality follows from exercise 21 , chapter 6 , in the book by o. kavian @xcite .
the general case can be reduced to this case by mollifying the function @xmath280 .
this proves the lemma .
let @xmath940 as in the hypothesis .
we set @xmath941 .
[ claim : u x u y ] there exists a constant @xmath640 such that for every weakly differentiable map @xmath942 and
@xmath943 satisfying @xmath944 , we have @xmath945 . by morrey s theorem
there is a constant @xmath255 such that @xmath946 , for every @xmath649 , weakly differentiable @xmath947 and @xmath948 .
let @xmath949 and @xmath950 be as in the hypothesis of the claim .
let @xmath951 be such that @xmath952 .
for @xmath953 we define @xmath954 . furthermore , we set @xmath955 and @xmath956 , and we choose points @xmath957 , @xmath958 , such that @xmath959 , for @xmath960 . for @xmath961
we have @xmath962 .
hence it follows from the statement of morrey s theorem that @xmath963 .
moreover , for @xmath964 we have @xmath965 , and hence analogously , @xmath966 . using the inequality @xmath967 , claim [ claim : u x u y ] follows .
let @xmath932 be such that @xmath933 .
it follows from claim [ claim : u x u y ] that there exists @xmath968 such that @xmath934 converges to @xmath969 , as @xmath362 , uniformly in @xmath936 . to prove inequality ( [ eq : hardy y ] ) , we choose a smooth map @xmath970 $ ] such that @xmath971 for @xmath972 , @xmath973 for @xmath974 and @xmath975 .
we fix a number @xmath407 and define @xmath976 $ ] by @xmath977 .
we abbreviate @xmath978 . using lemma [ le : hardy ] with @xmath280
replaced by @xmath979 , we have @xmath980 combining this with a calculation using leibnitz rule , it follows that @xmath981 claim [ claim : u x u y ] implies that @xmath982 , for @xmath983 . using the equalities @xmath984 and @xmath985
, it follows that @xmath986 inequality ( [ eq : hardy y ] ) follows by inserting this into ( [ eq : vert v lam ] ) and sending @xmath294 to @xmath28 .
this proves proposition [ prop : hardy ] .
the next result will be used to prove corollary [ cor : d l l ] below , which in turn is used in the proof of theorem [ thm : fredholm aug ] . for every @xmath468 we define @xmath987 and @xmath988 to be the spaces of polynomials in @xmath989 and @xmath990 of degree _ less than _ @xmath991 .
( note that if @xmath992 then @xmath993 . )
we abbreviate @xmath994 , @xmath995 , @xmath996 and @xmath997 .
let @xmath169 be a normed vector space and @xmath998 be a closed subspace .
we denote by @xmath999 the dual space of @xmath169 and equip @xmath1000 with the quotient norm .
[ prop : d l l ] for every @xmath468 , @xmath877 and @xmath390 the following conditions hold .
[ prop : d l l : fredholm ] the operator @xmath1001 is fredholm .
l l : kernel ] we have @xmath1002 . [ prop : d l l : coker ] the map @xmath1003 , @xmath1004 is well - defined and an isometric isomorphism .
here we equip @xmath1005 with the quotient norm .
[ rmk : x y ] for every normed vector space @xmath169 and every closed subspace @xmath998 the map @xmath1006 , @xmath1007 , is well - defined and an isometric isomorphism .
this follows from a straight - forward argument .
we denote by @xmath121 the space of schwartz functions on @xmath485 and by @xmath1008 the space of temperate distributions . by @xmath1009
we denote the fourier transform , and by @xmath1010 the inverse transformation .
[ proof : d l l ] let @xmath1011 and @xmath1012 be as in the hypothesis .
* we start by proving ( [ prop : d
l l : kernel ] ) . *
a calculation in polar coordinates shows that for every polynomial @xmath280 in @xmath1013 we have @xmath1014 hence our assumption @xmath1015 implies that @xmath1016 .
* we prove that @xmath1017 .
* let @xmath1018 .
then @xmath1019 ( as temperate distributions ) .
it follows that the support of @xmath1020 is either empty or consists of the point @xmath1021 .
hence the paley - wiener theorem implies that @xmath280 is real analytic in the variables @xmath415 and @xmath416 , where @xmath1022 , and there exists @xmath1023 such that @xmath1024 .
( see for example theorem ix.12 in vol .
i of the book @xcite . )
therefore , by liouville s theorem @xmath280 is a polynomial in the variable @xmath1013 .
since by our assumption @xmath386 , it follows from ( [ eq :
u l ] ) that @xmath1025 .
this proves that @xmath1017 and completes the proof of ( [ prop : d l l : kernel ] ) .
* we prove ( [ prop : d l l : fredholm ] ) and ( [ prop : d l l : coker ] ) . * we define @xmath1026 .
we identify the spaces @xmath1027 and @xmath1028 via the isometric isomorphism @xmath1029 .
then the adjoint operator @xmath1030 is given by @xmath1031 , where the derivatives are taken in the sense of distributions .
[ claim : t p ] @xmath1032 . for every polynomial @xmath280 in @xmath990 we have @xmath1033 our assumption @xmath386 and ( [ eq : u
l * ] ) imply that @xmath1034 .
furthermore , the inclusion @xmath1035 is proved analogously to the inclusion @xmath1017 , using @xmath1015 and ( [ eq : u l * ] ) .
this proves claim [ claim : t p ] .
we apply now theorem 4.3 in the paper by r. b. lockhart @xcite with @xmath1012 ( case @xmath1036 ) or @xmath1030 ( case @xmath1037 ) .
the hypotheses of that theorem are satisfied , since by our assumption @xmath390 , and since the operator @xmath1038 ( @xmath1030 ) has constant coefficients and is elliptic , in the sense that its principal symbol @xmath1039 ( @xmath1040 ) does not vanish on @xmath1041 .
hence that theorem implies that in the case @xmath1036 the map @xmath1012 is fredholm , and in the case @xmath1037 the operator @xmath1030 is fredholm .
it follows that @xmath1042 is closed if @xmath1036 . on the other hand ,
if @xmath1037 then @xmath1043 is closed , hence the same holds for @xmath1042 .
statements ( [ prop : d l l : coker ] ) and ( [ prop : d l l : fredholm ] ) follow now from statement ( [ prop : d l l : kernel ] ) , remark [ rmk : x y ] applied with @xmath1044 and @xmath1045 , and claim [ claim : t p ] .
this proves proposition [ prop : d l l ] .
let @xmath468 , @xmath877 , @xmath390 , and @xmath1046 $ ] be a smooth function that vanishes on @xmath1047 and equals 1 on @xmath1048 .
we equip @xmath1049 with the norm induced by the isomorphism of lemma [ le : x d iso ] .
this norm is complete .
( see e.g. @xcite . )
[ cor : d l l ] the following map is fredholm , with real index @xmath1050 : @xmath1051 the composition of the isomorphism of lemma [ le : x d iso ] with ( [ eq : d bar z rho ] ) is given by @xmath1052 , where @xmath1053 and @xmath1054 .
the map @xmath1012 is the composition of the canonical projection @xmath1055 with the operator @xmath1056 . using proposition [ prop : d l l ]
, it follows that @xmath1012 is fredholm of real index @xmath1050 .
furthermore , @xmath1057 is compact , since it equals the composition of the canonical projection @xmath1058 ( which is compact ) with a bounded operator .
corollary [ cor : d l l ] follows .
the next result is used in the proof of theorem [ thm : fredholm aug ] .
let @xmath1059 be a finite dimensional hermitian vector space , @xmath1060 positive linear maps , @xmath367 and @xmath877 .
we define @xmath1061 [ prop : d a b d ] the operator @xmath1062 is fredholm of index 0 . for the proof of proposition [ prop :
d a b d ] we need the following result .
[ prop : calderon ] let @xmath1063 and @xmath83 be as above , and @xmath219 . then the map @xmath1064 is an isomorphism ( of banach spaces ) .
consider first the case @xmath1065 and @xmath1066 .
we define @xmath1067 .
the map @xmath1068 is well - defined . by caldern
s theorem this map extends uniquely to an isomorphism @xmath1069 ( see theorem 1.2.3 . in the book @xcite . )
note that @xmath1070 , for every @xmath1071 .
it follows that the inverse of ( [ eq : l g u ] ) is given by @xmath1072 .
hence this is an isomorphism .
the general case can be reduced to the above case by diagonalizing the map @xmath83 .
this proves proposition [ prop : calderon ] .
[ proof : d a b d ] we abbreviate @xmath1073 , etc . * assume first that @xmath1074 .
* we denote by @xmath1075 the unique positive linear maps satisfying @xmath1076 , @xmath1077 .
we define @xmath1078 a short calculation shows that @xmath1079 [ claim : l ] the operator @xmath1080 is an isomorphism .
we define @xmath1081 by a short calculation we have @xmath1082 .
since the linear maps @xmath1083 are positive , proposition [ prop : calderon ] implies that @xmath1084 is an isomorphism .
we denote by @xmath1085 its inverse and define @xmath1086 .
then @xmath294 is bounded and @xmath1087 . by a short calculation
, we have @xmath1088 , for every schwartz function @xmath1089 .
this implies that @xmath1090 , and therefore @xmath1091 .
since @xmath1092 is continuous and @xmath1093 is dense , it follows that @xmath1094 .
claim [ claim : l ] follows .
the maps @xmath1095 and @xmath1096 are automorphisms . therefore ,
( [ eq : a b t ] ) and claim [ claim : l ] imply that @xmath1012 is an isomorphism .
consider now the * general case @xmath367*. the map @xmath1097 is an isometric isomorphism .
furthermore , by proposition [ prop : lam d morrey]([prop : lam w ] ) the map @xmath1098 is well - defined and an isomorphism .
we define @xmath1099 .
direct calculations show that @xmath1100 , @xmath1101 and @xmath1102 .
therefore , proposition [ prop : lam d morrey]([prop : lam w cpt ] ) implies that the operator @xmath1057 is compact .
we proved that @xmath805 is an isomorphism .
it follows that @xmath1062 is a fredholm map of index 0 .
this proves proposition [ prop : d a b d ] in the general case .
for the proof of proposition [ prop : right ] we need the following three results .
let @xmath219 , @xmath1103 , @xmath0 a compact lie group with lie algebra @xmath4 and @xmath70 an invariant inner product on @xmath4 .
for a riemannian manifold @xmath169 and a @xmath0-bundle @xmath1104 we denote by @xmath1105 the space of @xmath229-connections on @xmath74 .
[ prop : uhlenbeck ] let @xmath762 be a compact subset diffeomorphic to @xmath396 . if @xmath1106 then there exist constants @xmath254 and @xmath255 such that for every principal @xmath0-bundle @xmath763 , @xmath1107 , and @xmath1108 the following holds . if @xmath1109 is flat and @xmath765 then there exists a gauge transformation @xmath1110 such that @xmath1111 .
let @xmath1112 and @xmath1113 be as in the hypothesis .
we denote by @xmath1114 the trivial connection on @xmath1115 . by uhlenbeck s
gauge theorem there exist constants @xmath254 and @xmath255 such that for every connection @xmath1116 satisfying @xmath1117 there exists @xmath1118 such that @xmath1119 .
( this follows for example from theorem 6.3 in @xcite . )
let @xmath763 be a @xmath0-bundle , @xmath1107 be flat , and @xmath1108 be such that @xmath765 .
since @xmath1109 and @xmath1114 are flat , there exists a smooth isomorphism of @xmath0-bundles @xmath1120 ( with fixed base ) such that @xmath1121 .
we choose @xmath1118 as in the conclusion of uhlenbeck s theorem with @xmath1122 .
we define @xmath1123 .
a straight - forward calculation shows that @xmath1124 .
the statement of proposition [ prop : uhlenbeck ] follows from this .
[ prop : right d a * d a ] let @xmath219 , @xmath877 , @xmath0 a compact lie group with lie algebra @xmath4 , @xmath70 an invariant inner product on @xmath4 , and @xmath762 a compact subset diffeomorphic to @xmath396 .
then there exists a constant @xmath255 such that for every principal @xmath0-bundle @xmath763 and every smooth flat connection @xmath83 on @xmath74 there exists right inverse @xmath294 of @xmath1125 satisfying @xmath1126 .
let @xmath219 and @xmath877 .
for an open subset @xmath1127 we denote by @xmath1128 the compactly supported smooth functions on @xmath528 .
we define the map @xmath1129 as follows .
we denote by @xmath1130 the fundamental solution of the laplace equation ( see e.g. @xcite , p. 22 ) .
let @xmath1131 .
we define @xmath1132 to be the extension of @xmath1133 by 0 outside @xmath1134 .
we denote by @xmath209 convolution in @xmath332 and define @xmath1135 .
note that @xmath559 is locally integrable , hence the convolution is well - defined .
furthermore , @xmath1136 is smooth , and @xmath1137 .
( the first assertion follows from differentiation under the integral , and for the second see for example theorem 1 on p. 23 in the book @xcite . )
[ claim : c tf ] there exists a constant @xmath255 such that @xmath1138 , for every @xmath1131 .
young s inequality states that @xmath1139 , for every @xmath1131 .
furthermore , the caldern - zygmund inequality states that there exists a constant @xmath255 such that for every @xmath1140 we have @xmath1141 .
( see for example theorem b.2.7 in @xcite .
note that @xmath1142 . )
claim [ claim : c tf ] follows from this .
we fix a constant @xmath255 as in claim [ claim : c tf ] . by this claim
the map @xmath1143 uniquely extends to a bounded linear map @xmath1144 . since @xmath1145 , for every @xmath1146
, a density argument shows that @xmath1147 , for every @xmath1148 , @xmath1012 is a right inverse for @xmath1149 .
let now @xmath1150 and @xmath1113 be as in the hypothesis . without loss of generality
we may assume that @xmath1151 .
let @xmath1152 be a @xmath0-bundle , and @xmath753 be flat .
we fix a point @xmath1153 , and denote by @xmath1154 the @xmath83-horizontal section through @xmath1155 .
this is the unique smooth section of @xmath74 satisfying @xmath1156 and @xmath1157 .
for @xmath1158 we define the map @xmath1159 by @xmath1160 .
this is an isometric isomorphism .
we define @xmath1161 .
it follows that @xmath1162 , where we use various operator norms .
a straight - forward calculation shows that @xmath1163 .
it follows that @xmath294 is a right inverse for @xmath1164 .
this proves proposition [ prop : right d a * d a ] .
[ le : infty 1 p a ] let @xmath219 , @xmath749 , @xmath0 a compact lie group with lie algebra @xmath4 , @xmath70 an invariant inner product on @xmath4 and @xmath1165 a compact subset diffeomorphic to @xmath396 .
then there exist constants @xmath255 and @xmath254 such that for every principal @xmath0-bundle @xmath1166 , @xmath1108 , @xmath1167 and @xmath1168 the following holds . if @xmath1169 then @xmath1170 .
let @xmath1171 and @xmath1113 be as in the hypothesis .
we choose constants @xmath254 and @xmath1172 as in proposition [ prop : uhlenbeck ] .
since by assumption @xmath749 , it follows from morrey s theorem that there exists a constant @xmath668 with the following property . if @xmath1173 is a @xmath0-bundle , @xmath1107 is flat , @xmath1167 and @xmath1174 then @xmath1175 . let @xmath1166 be a @xmath0-bundle and @xmath1108 be such that @xmath1169 .
we choose @xmath1176 as in proposition [ prop : uhlenbeck ] .
let @xmath1168 .
we set @xmath1177 , @xmath1178 and @xmath1179\,|\,\xi,\eta\in \g:\,|\xi|\leq1,\,|\eta|\leq1\big\}$ ] .
a direct calculation shows that @xmath1180 $ ] .
it follows that @xmath1181 by the statement of proposition [ prop : uhlenbeck ] , we have @xmath1182 . combining this with morrey s theorem and ( [ eq : al infty ] ) , lemma [ le :
infty 1 p a ] follows .
let @xmath1183 and @xmath83 be as in the hypothesis .
we show that the operator ( [ eq : d a w ] ) admits a bounded right inverse .
[ claim : l ] there exists a bounded linear map @xmath1184 such that @xmath1185 .
we may assume without loss of generality that @xmath1151 .
we define @xmath1186 to be the subset consisting of all @xmath1187 such that @xmath1188 , where @xmath1189 .
furthermore , we denote by @xmath1190 the @xmath83-parallel transport in @xmath1191-direction .
let @xmath1192 .
we denote @xmath1193 .
then @xmath1194 , where @xmath1195 is the unique path satisfying @xmath1196 , for every @xmath1197 , @xmath1198 and @xmath1199 .
let @xmath1200 .
we define @xmath1201 by the condition @xmath1202=\xi\circ \pi(p)$ ] , for @xmath321 .
let @xmath1203 and @xmath1204 .
we define @xmath1205 .
furthermore , we define the section @xmath1206 by the condition @xmath1207 $ ] , for every @xmath321 , and @xmath1208 . then @xmath1080 has the required properties .
this proves claim [ claim : l ] .
we choose a map @xmath1080 as in claim [ claim : l ] .
furthermore , we choose a smooth flat connection @xmath1109 on @xmath74 . by proposition [ prop : right d
a * d a ] there exists a bounded right inverse @xmath804 of @xmath1209 .
we define the map @xmath1210 the first assertion of proposition [ prop : right ] is now a consequence of the following .
[ claim : r ] the map ( [ eq : r d a r 0 ] ) is well - defined and bounded , and @xmath1211 .
a short calculation shows that @xmath1212 is of first or zeroth order .
hence it defines a bounded linear map from @xmath1213 to @xmath1214 .
furthermore @xmath1215 .
this implies that @xmath294 is well - defined and bounded .
a short calculation shows that @xmath1211 .
this proves claim [ claim : r ] . to prove the second statement of proposition [ prop : right ] , we choose @xmath1216 as in lemma [ le : infty 1 p a ] ( corresponding to @xmath1217 ) and @xmath1218 as in proposition [ prop : uhlenbeck ] ( corresponding to @xmath1217 ) .
we also fix a constant @xmath1219 as in proposition [ prop : right d a * d a ] , and we define @xmath1220 .
let @xmath1173 be a @xmath0-bundle and @xmath315 be such that @xmath765 .
we choose a flat smooth connection @xmath1109 on @xmath74 . by the assertion of proposition [ prop : uhlenbeck ] there exists @xmath1176
such that @xmath1221 furthermore , by the assertion of proposition [ prop : right d a * d a ] there exists a right inverse @xmath804 of @xmath1209 satisfying @xmath1222 , where @xmath1223 denotes the operator norm of @xmath804 .
we define the map @xmath1224 by @xmath1225 $ ] . since @xmath1109 is flat , the hypotheses of lemma [ le
: infty 1 p a ] are satisfied . hence by this lemma and ( [ eq : g * a a 0 ] ) , we have @xmath1226 for every @xmath1227 .
hence @xmath1057 is well - defined , and @xmath1228 hence @xmath1229 is invertible , and the von neumann series @xmath1230 converges in the operator norm and equals @xmath1231 .
furthermore , @xmath1232 . we define @xmath1233 .
since @xmath1234 , we have @xmath1235 .
furthermore , for every @xmath1236 , we have @xmath1237\big\vert_p\ ] ] @xmath1238 where @xmath1239|\,\big|\,\al,\be\in \wone(\g_p):\,|\al|,|\be|\leq1\big\}$ ] . here in the last step we used ( [ eq : g * a a 0 ] ) and the fact @xmath765 .
this proves the second statement and concludes the proof of proposition [ prop : right ] .
the next lemma was used in the proof of theorem [ thm : fredholm ] . here for a linear map @xmath1240 we denote @xmath1241 .
the proof of lemma [ le : x y z ] is straight - forward and left to the reader .
the following result was used in the proof of proposition [ prop : triv ] .
we define the map @xmath535 by @xmath536 . for two topological spaces @xmath169 and @xmath1251 we denote by @xmath1256 the set of all continuous maps from @xmath169 to @xmath1251 , and by @xmath1257 $ ] the set of all ( free ) homotopy classes of such maps .
let @xmath880 be a finite dimensional complex vector space .
we denote by @xmath1258 the space of its ( complex ) endomorphisms of @xmath880 , by @xmath1259 the determinant map , and by @xmath1260 the group of automorphisms of @xmath880 .
we choose a hermitian inner product @xmath880 and denote by @xmath1264 the corresponding group of unitary automorphisms of @xmath880 .
the map @xmath1265 induces an isomorphism of fundamental groups , see proposition 2.23 in the book by d. mcduff and d. a. salamon @xcite .
furthermore , the space @xmath1266 strongly deformation retracts onto @xmath1264 .
( this follows from the gram - schmidt orthonormalization procedure . )
this implies that the map @xmath1267 induces an isomorphism of fundamental groups .
let @xmath1268 .
it follows that the map @xmath1261 given by @xmath1262 descends to an isomorphism between @xmath1269 and @xmath1270 .
since this group is abelian , the map @xmath1271 $ ] that forgets the base point @xmath1272 , is a bijection .
the statement of lemma [ le : aut ] follows from this .
the next lemma was used in the proof of theorem [ thm : l w * r ] .
let @xmath169 and @xmath5 be manifolds , @xmath0 a lie group with lie algebra @xmath4 , @xmath70 an invariant inner product on @xmath4 , @xmath1273 a @xmath0-invariant riemannian metric on @xmath5 , and @xmath35 its levi - civita connection . for @xmath54 we denote by @xmath1274 the vector field on @xmath5 generated by @xmath1275 .
we define the tensor @xmath1276 by @xmath1277 a short calculation shows that @xmath1278 is skew - symmetric .
this two - form was introduced in @xcite ( page 181 ) .
the next lemma corresponds to proposition 7.1.3(a , b ) in @xcite .
let @xmath1104 be a principal bundle , @xmath753 , @xmath1279 , @xmath1280 , and @xmath1281 .
we define the connection @xmath211 on @xmath1282 as in section [ sec : back ] .
let @xmath340 and @xmath173 be as in section [ sec : main ] , and @xmath1284 .
the following remark was used in the proofs of theorems [ thm : fredholm ] and [ thm : l w * r ] . recall the definition ( [ eq : m * ] ) of @xmath1285 , and that @xmath369 denotes the orthogonal projection onto @xmath370 .
[ rmk : c]let @xmath1286 be compact .
we define @xmath1287 .
then @xmath1288 . let @xmath1289 .
then @xmath1290 is invertible , and @xmath1291 where the @xmath1292 s denote operator norms .
furthermore , @xmath1293 , for every @xmath1294 .
these assertions follow from short calculations .
it follows from hypothesis ( h ) that there exists @xmath1296 such that @xmath1297 .
we define @xmath1298|\,\big|\,\xi,\eta\in\g:\,|\xi|\leq1,\,|\eta|\leq 1\big\},&\\ \nn&c_0:=\inf\big\{|l_x\xi|/|\xi|\,\big|\,x\in \mu^{-1}(\bar b_{\de_0}),\,0\neq\xi\in\g\big\}.\end{aligned}\ ] ] since the action of @xmath0 on @xmath514 is free , it follows that @xmath27 is injective , for @xmath1299 .
furthermore , by hypothesis ( h ) the set @xmath1300 is compact .
it follows that @xmath1301 .
we choose a positive number @xmath1302 , and we define @xmath1303 .
let @xmath511 and @xmath1304
. then @xmath1305+l_x^*l_x\eta.\ ] ] using the last assertion in ( [ eq : l x * l x c ] ) , we have @xmath1306.\ ] ] by the first assertion in ( [ eq : l x * l x c ] ) , we have @xmath1307 . combining this with ( [ eq : d mu x l x],[eq : pr x l x ] ) and the second assertion in ( [ eq : l x * l x c ] ) , we obtain @xmath1308 inequality ( [ eq : c inf ] ) follows now from our choice of @xmath1309 .
this proves claim [ claim : c inf ] and completes the proof of lemma [ le : u ] .
the following lemma was mentioned in section [ sec : main ] .
recall the definition ( [ eq : bb p lam ] ) of @xmath1310 , and that @xmath232 denotes the group of gauge transformations on @xmath74 of class @xmath1311 .
let @xmath1312 and @xmath1313 be such that @xmath1314 .
it follows from hypothesis ( h ) that there exists @xmath404 such that @xmath405 ( defined as in ( [ eq : m * ] ) ) .
furthermore , lemma [ le : si ] below implies that there exists @xmath407 such that @xmath1315 , for @xmath1316 .
we choose @xmath1317 .
since @xmath1318 and @xmath1319 , it follows that @xmath1320 .
let @xmath1321 .
we choose a smooth path @xmath1322\to p$ ] such that @xmath1323 , for @xmath1324 .
then the map @xmath1325\to g$ ] solves the ordinary differential equation @xmath1326 , @xmath1327 .
it follows that @xmath1328 , and hence @xmath1329 .
this proves lemma [ le : free ] .
[ le : si ] there exists a smooth section @xmath443 of the restriction of the bundle @xmath74 to @xmath1048 and a point @xmath445 , such that @xmath1332 converges to @xmath448 , uniformly in @xmath363 , as @xmath362 , and @xmath1333 .
we define the function @xmath1335 .
it follows from the ad - invariance of @xmath70 that @xmath1336 since @xmath99 is compact , we have @xmath1337 .
furthermore , @xmath1338 . combining this with ( [ eq : d mu u 2 ] ) , it follows that @xmath1339 .
therefore , by proposition [ prop : hardy ] ( hardy - type inequality , applied with @xmath280 replaced by @xmath1133 ) the expression @xmath743 converges to some number @xmath935 , as @xmath362 , uniformly in @xmath363 .
since @xmath1340 , it follows that @xmath1341 .
this proves claim [ claim : r u ] .
it follows from hypothesis ( h ) that there exists a @xmath404 such that @xmath405 ( defined as in ( [ eq : m * ] ) ) .
we choose @xmath407 so big that @xmath1342 if @xmath1343 .
since @xmath0 is compact , the action of it on @xmath5 is proper .
hence the local slice theorem implies that @xmath1344 carries a unique manifold structure such that the canonical projection @xmath1345 is a submersion .
consider the map @xmath1346 defined by @xmath1347 , where @xmath546 is arbitrary .
we choose @xmath219 and an embedding @xmath1350 .
furthermore , we choose a smooth function @xmath1351 that vanishes on @xmath1352 and equals 1 on @xmath1353 .
we define @xmath1354 to be the map given by @xmath1355 on @xmath1356 and by 0 on @xmath1352 .
it follows that @xmath1357 .
furthermore , @xmath1358 our assumption @xmath344 implies that @xmath1359 .
furthermore , @xmath8 is proper by the hypothesis ( h ) , hence the set @xmath1360 is compact .
thus the same holds for the set @xmath1361 .
this set contains the image of @xmath150 .
it follows that @xmath1362 , and therefore @xmath1363 .
hence the hypotheses of proposition [ prop : hardy ] are satisfied .
it follows that the point @xmath743 converges to some point @xmath1364 , uniformly in @xmath363 , as @xmath362 .
claim [ claim : bar u ] follows .
let @xmath1365 be as in claim [ claim : bar u ] .
we choose a local slice around @xmath1365 , a pair @xmath1366 , where @xmath1367 is an open neighborhood of @xmath1365 , and @xmath1368 is a smooth map satisfying @xmath1369 .
then there exists a unique section @xmath1370 of @xmath1371 such that @xmath1372 . by the homotopy lifting property of
@xmath74 we may extend this to a continuous section @xmath1373 of @xmath1374 . smoothing out @xmath1373 on @xmath1375
, we obtain a smooth section @xmath443 of @xmath1374 .
we define @xmath1376 .
it follows from claim [ claim : bar u ] that @xmath447 converges to @xmath448 , uniformly in @xmath69 , for @xmath362 .
furthermore , @xmath1377 @xmath1378 since @xmath1379 and @xmath869 , it follows that @xmath1333 .
this proves lemma [ le : si ] .
[ proof : prop : p ] we choose @xmath443 and @xmath448 as in lemma [ le : si ] , and we define @xmath346 to be the quotient of @xmath1380 under the equivalence relation generated by @xmath1381 , where @xmath510 is determined by @xmath545 , for @xmath321 .
we define @xmath548 by @xmath549):=u(p)$ ] , for @xmath321 , and @xmath550):=g^{-1}x_\infty$ ] , for @xmath510 .
if follows from the statement of lemma [ le : si ] that this map is continuous .
let now @xmath1382 and @xmath1383 be two extensions of @xmath74 , for which the map @xmath280 extends to continuous maps @xmath1384 .
we define @xmath1385 to be the identity on @xmath1386 , and for @xmath1387 in the fiber of @xmath1382 over @xmath28 we define @xmath1388 to be the unique point @xmath1389 such that @xmath1390 .
( it follows from lemma [ le : si ] that this point is unique . )
this map has the required properties .
this proves proposition [ prop : p ] . | consider a hamiltonian action of a compact connected lie group @xmath0 on an aspherical symplectic manifold @xmath1 . under some assumptions on @xmath1 and the action ,
d. a. salamon conjectured that counting gauge equivalence classes of symplectic vortices on the plane @xmath2 gives rise to a quantum deformation @xmath3 of the kirwan map .
this article is the first of three , whose goal is to define @xmath3 rigorously .
its main result is that the vertical differential of the vortex equations over @xmath2 ( at the level of gauge equivalence ) is a fredholm operator of a specified index .
potentially , the map @xmath3 can be used to compute the quantum cohomology of many symplectic quotients .
conjecturally it also gives rise to quantum generalizations of non - abelian localization and abelianization ( see @xcite ) . |
the concept of @xmath2-clustering is essential to understand the structure of light nuclei @xcite . in particular
, @xmath1o is one of the nuclei which have the most extensively been studied with the cluster model as well as with the shell model . while the ground state is well understood to have a doubly - closed - shell structure , the excited state , such as the second @xmath3 state , is known as one of `` the mysterious @xmath3 states '' whose excitation energy , @xmath12 mev , is too low to be explained in terms of the shell model @xcite .
it had also been mysterious that the second @xmath3 state forms a rotational band with low - lying @xmath13 and @xmath14 states . besides those states
, the third @xmath3 state at @xmath15 mev has been difficult to be understood from the shell model aspect as well .
one of the main aims of the cluster model approach has been to describe these states from a different perspective and , indeed , the studies based on the @xmath2-cluster model have succeeded in reproducing many experimental data .
it was clarified that the @xmath16 and @xmath17 states have @xmath18c@xmath19 and @xmath20c@xmath21 cluster structures , where the @xmath2 particle rotates in an @xmath22 and @xmath23 wave around the @xmath24c core being in @xmath3 and @xmath13 states , respectively @xcite .
the great success also comes from having further revealed that almost all of the excited states up to @xmath25 mev have considerable large @xmath26c cluster components @xcite . besides @xmath26c clustering , @xmath0 clustering
can also be expected to exist around the decay threshold into @xmath0 particles at @xmath27 mev , as suggested by ikeda _
et al . _
@xcite . as an intriguing subject investigating the @xmath0 cluster structure
, the presence of a @xmath0 linear - chain state still remains an open question @xcite .
another possibility are dilute gas - like states composed of weakly interacting @xmath0 particles @xcite .
they are analogues to the hoyle state , i.e. the @xmath16 state in @xmath24c , well described by a gas - like @xmath28-particle structure with dilute density @xcite . in good approximation it forms a product state of the @xmath28 particles in the lowest @xmath7 orbit of the corresponding mean field potential @xcite .
a striking fact is that the hoyle state wave function obtained more than 30 years ago by solving the @xmath28 rgm ( resonating group method ) equation of motion @xcite is to very good approximation identical to an antisymmetrized product type wave function @xcite , the so - called thsr ( tohsaki , horiuchi , schuck and rpke ) wave function , introduced in ref .
@xcite , and slightly modification in ref .
thus , the state has been proposed to be related to boson condensation of @xmath2 particles in infinite nuclear matter @xcite .
the new interpretation of the hoyle state as an @xmath2 condensate has provoked many theoretical and experimental works on @xmath2-particle condensation phenomena in light nuclei @xcite .
it should be mentioned that also for @xmath29 nuclei such as @xmath30b and @xmath31c several authors try to find states which are analogous to the hoyle state @xcite .
the first investigation of the @xmath0-particle condensate state was performed via the thsr ansatz in ref .
@xcite , where two excited @xmath3 states were obtained . in particular , the third @xmath3 ( @xmath32 ) state has an excitation energy close to the @xmath0 threshold and a considerably larger r.m.s .
radius , @xmath33 fm , than the one of the ground state , @xmath34 fm , and therefore the @xmath32 state was regarded as a candidate of the @xmath0 condensate state .
based on this theoretical prediction , an experimental search for the @xmath0 condensate state was undertaken via inelastic @xmath2 scattering on @xmath1o .
a new @xmath3 state was observed , as the fourth @xmath3 state , at @xmath35 mev with a relatively large @xmath2 decay width of @xmath36 mev @xcite .
the angular distribution of the cross section of the inelastic @xmath2 scattering to the new state well agrees with the calculated one using the @xmath32 wave function .
the relatively large decay width of @xmath36 mev is also consistently reproduced by the thsr wave function @xcite .
these results lead to the conclusion @xcite that the @xmath32 state may be assigned to the new @xmath3 state at @xmath35 mev .
thus the @xmath35 mev state was considered to be a candidate of the @xmath0 condensate state @xcite .
in order to confirm the above mentioned thsr description of @xmath1o , recently the present authors investigated the @xmath0 condensate state via the @xmath0 ocm ( orthogonality condition model ) @xcite .
the four - body ( @xmath0 ) problem with respect to the @xmath2-@xmath2 relative motion was solved in a model space large enough to describe the dilute @xmath0 gas - like configuration , as well as @xmath26c cluster and shell - model - like ground state structures , using the gaussian expansion method @xcite .
a one - to - one correspondence with the observed energy spectrum up to the @xmath10 state ( see also fig .
[ fig:3 ] ) was observed .
it further was shown that the calculated @xmath10 ( @xmath37 ) state has large occupation probability of @xmath38 % for @xmath0 s in the single @xmath7 orbit and also a large overlap with the @xmath39c@xmath40(hoyle state ) structure .
these results indicate that the @xmath37 state can be regarded to be the @xmath0 condensate state as the analogue to the hoyle state , to be identified with the sixth experimental @xmath3 state at @xmath11 mev .
the well - known @xmath2-cluster structures of the @xmath16 and @xmath17 states were simultaneously reproduced , in agreement with previous ocm calculations @xcite .
it was also shown that the @xmath41 and @xmath42 states both have @xmath39c cluster structures , where the @xmath2 cluster orbits in a higher nodal @xmath22-wave and in a @xmath43-wave around the @xmath24c core with @xmath44 and @xmath45 , respectively .
these states had not been discussed so far in the literature . in this work , we confront our previous analysis with the thsr wave function for @xmath1o [ 11 ] with the above mentioned results from the ocm approach . as we explained above ,
the @xmath0 ocm results tell us that the @xmath10 state at @xmath11 mev is a good candidate for the @xmath0 condensate state . in ref .
@xcite using the thsr wave function only three @xmath3 states were obtained , including the ground state .
this scarcity of @xmath3 states in the thsr approach is easily understood by the fact that it allows only for two limiting configurations , that is a pure slater determinant for @xmath46 and a pure @xmath2-particle gas for @xmath47 @xcite .
asymptotic configurations like @xmath48 are absent . they may be represented in a rough average way by the intermediate @xmath3 state . in any case
the situation is such that it incited us to reanalyze the previous study of ref .
the important question is whether or not the counterpart of the @xmath37 state can also be found with the thsr ansatz , which should be a further state located above the @xmath32 state . in this respect , we need to recall that , in the previous thsr calculation @xcite , a proper resonance boundary condition was not imposed even for the states above the @xmath0 threshold .
thus we could not obtain reliable results for higher @xmath3 states than the @xmath32 state .
the other excited state then only mocks up in an average way the states which have dominant @xmath26c structure .
the purpose of this paper is to report on results of a @xmath0 thsr calculation with proper treatment of the resonances .
we will see that this calculation gives us a new @xmath3 state as the @xmath6 state , slightly above the @xmath0 threshold . by comparing the wave functions of the @xmath6 state and the @xmath37 state in the @xmath0 ocm calculation
, it will be clarified that both states well correspond to each other .
in particular , we will show that the @xmath6 state has a larger r.m.s .
radius , a larger amount of condensate fraction , and a more sharpened peak around zero momentum for the @xmath2-particle momentum distribution , than the @xmath32 state .
those properties are comparable to those of the @xmath37 state .
this indicates that the newly found @xmath6 state should be considered as the @xmath0 condensate state and is likely to be assigned to the @xmath10 state observed at @xmath11 mev .
the outline of the present paper is as follows : in sec .
[ subsec : thsr ] , we introduce the thsr wave function and the microscopic hamiltonian . the present approach to resonances
is briefly explained in sec .
[ subsec : rmsmin ] , which is , together with the thsr ansatz , applied to @xmath49 in sec .
[ subsec:04 + ] .
the obtained wave functions of @xmath49 are analyzed in detail in sects .
[ subsec : wf ] and [ subsec : overlap ] .
the quantities based on @xmath2-particle bosonic degree of freedom , such as the amount of condensation , are discussed in sec .
[ subsec : boson ] .
further discussions are presented in sects .
[ subsec : size ] and [ subsec : interp ] .
[ sec : conc ] is devoted to the conclusion .
in general , the @xmath50 cluster model wave function can be presented in the following form : @xmath51 \label{eq:1}\ ] ] with @xmath52 the antisymmetrizer and @xmath53 the intrinsic wave function of the @xmath54-th @xmath2-particle , which is taken as a gaussian , @xmath55 .
\label{eq:2}\ ] ] the wave function @xmath56 for the c.o.m .
motion of the @xmath2 s with @xmath57 is , of course , also chosen translationally invariant , that is it depends only on the relative coordinates @xmath58 or on the corresponding jacobi coordinates .
the spin - isospin part in eq .
( [ eq:1 ] ) is not written out but supposed to be of scalar - isoscalar form .
we will not mention it henceforth . in 2001 ,
tohsaki , horiuchi , schuck , and rpke ( thsr ) @xcite adopted , as the @xmath50 condensate type wave function , the following ansatz for @xmath56 in eq .
( [ eq:1 ] ) , @xmath59 with the total c.o.m .
coordinate @xmath60 and @xmath61 , that is a gaussian with a large width parameter @xmath62 which is of the nucleus dimension .
the product of @xmath63 identical @xmath7 wave functions reflects the boson condensate character .
this feature is realized as long as the action of the antisymmetrizer in eq .
( [ eq:1 ] ) is sufficiently weak .
this type of condensate wave function has been known , in the meantime , to have considerable success , notably with an accurate description of the hoyle state , proposing it as the first of a series of excited states in @xmath50 nuclei with @xmath2-particle condensate character .
those states can be considered as the precursors of macroscopic @xmath2-particle condensation in nuclear matter at low density [ 19 ] .
the internal part of the hamiltonian we adopt in the present study of @xmath1o is composed of kinetic energy @xmath64 , with nucleon mass @xmath65 , the coulomb force @xmath66 and the effective nuclear interaction : @xmath67 where the c.o.m .
kinetic energy @xmath68 is subtracted . in this calculation
, we adopted the same nucleon - nucleon force with two - body @xmath69 and three - body @xmath70 terms as adopted in ref .
@xcite . both two - body and three - body forces consist of finite - range gaussian functions with
a parameter set named f1 given in ref .
@xcite .
the wave functions of quantum states in @xmath49 can then be expanded using the @xmath0 thsr wave function , like @xmath71 where @xmath72 is the @xmath0 thsr wave function which has the form of @xmath73 .
\label{eq : thsr}\ ] ] the discrete variational parameters @xmath74 represent the generator coordinate of the hill - wheeler ansatz .
the expansion coefficients @xmath75 and the corresponding eigenenergy @xmath76 for the @xmath77-th eigenstate are obtained by solving the following hill - wheeler equation , @xmath78 one of the aims of the present study is to discuss the states obtained by the thsr ansatz with proper treatment of resonances .
the calculation of the resonance state in the bound state approximation is usually done by diagonalizing the hamiltonian with the use of a finite number of square - integrable basis wave functions .
the calculated positive energy eigenstates which are linear combinations of the basis wave functions are divided into resonance states and continuum states .
there are many practical methods to carry out this division .
one of the methods , found in ref .
@xcite by some of the present authors , was shown to give consistent results with the ones obtained with a well known approach to resonances such as the accc ( analytic continuation of the coupling constant ) method @xcite .
we apply the method to the present study , and following the literature @xcite , briefly explain it in this subsection .
let us first consider an attractive pseudo - potential @xmath79 which is added to the original hamiltonian @xmath80 , yielding @xmath81 where @xmath8 is a coupling constant used to vary the strength of the pseudo - potential .
we diagonalize this new hamiltonian @xmath82 by using the same set of basis wave functions as used for @xmath80 . as we increase the coupling constant @xmath8 from the physical value , @xmath83 , the eigenenergy of any resonance state decreases , and eventually the resonance state is transformed into a bound state . by contrast ,
continuum states exhibit almost no change in their eigenenergies as @xmath8 increases . )
are schematically shown as functions of @xmath8 .
there are three crossing points near which the mixing of continuum and resonance components occurs .
( see the main text for a detailed explanation . )
figure adopted from ref .
@xcite . ]
this behavior of the eigenenergies as functions of @xmath8 is shown schematically in fig .
[ fig : scheme ] . in this figure
, the curve labeled ( b ) around @xmath83 represents the eigenenergy of the resonance state denoted as @xmath84 .
it decreases as @xmath8 increases .
this curve ( b ) then crosses the curve ( a ) , which is flat around @xmath83 and is the eigenenergy of a continuum state denoted as @xmath85 .
after the closest approach of the two curves ( b ) and ( a ) , the curve labeled ( b ) , which is the smooth continuation of the curve ( a ) , and hence is disconnected from the curve ( b ) , represents the eigenenergy of the resonance state @xmath84 .
this reminds us of the landau - zener level crossing scenario , see e.g. @xcite .
the eigenenergy of the continuum state @xmath85 , which is first represented by the curve ( a ) , continues with an almost flat shape , except near the crossing points .
eventually it is converted into the curve ( a ) . in the case of the resonance state @xmath84 at @xmath83 , we can consider that this state contains a negligible amount of admixture of continuum state components , since it is located far from the crossing points with the continuum states .
however , in the case of the resonance state @xmath86 corresponding to the resonance state curve ( d ) in fig .
[ fig : scheme ] , the state is at @xmath83 already involved in the crossing with the continuum state curve ( c ) .
therefore , @xmath86 at @xmath83 should contain a sizable admixture of the continuum state .
this is the usual situation for a resonance state with a broad width , and the resonance wave function obtained by using the bound state approximation can not be trusted , because of the sizable admixture of the continuum state . in order to overcome this difficulty , we note the utility of the continuum state wave functions obtained using the bound state approximation in the neighborhood of the resonance state , since they should contain sizable admixtures of the resonance state component . in our present simple example considered in fig .
[ fig : scheme ] , we notice the utility of the continuum state wave function @xmath87
. we can express both states , @xmath86 and @xmath87 , as linear combinations of the pure resonance state @xmath88 and the pure continuum state @xmath89 : @xmath90 the method for the extraction of the pure resonance component @xmath88 which we proposed in ref .
@xcite is obtained by noticing that the extension of the density of a resonance state is far smaller than the one of continuum states .
more precisely , we can determine @xmath2 and @xmath91 from the requirement that the quantity @xmath92 be a minimum , where @xmath93 here @xmath94 is the mass number .
this requirement is satisfied by diagonalizing the operator @xmath95 in the basis states @xmath86 and @xmath87 , and by choosing @xmath96 to be the eigenvector belonging to the smallest eigenvalue .
when the resonance component is distributed among several energy eigenstates , we need to diagonalize the operator @xmath95 in these energy eigenstates . when the energy eigenstates contain two resonance components , the diagonalization of the operator @xmath95 gives two small eigenvalues .
two resonance states are obtained by diagonalizing the hamiltonian @xmath82 with the two wave functions which give the two smallest eigenvalues of the operator @xmath95 .
we are now able to calculate approximate values of resonance energies along the resonance energy curve .
if a point on the resonance state curve is not influenced by any crossing points with continuum state curves , we can regard the energy of the point to be just the resonance energy . on the other hand ,
if a point on the resonance state curve is influenced by some near - by crossing point , we extract the resonance wave function using the procedure explained above and then calculate the expectation value of the hamiltonian @xmath82 with the extracted wave function , which is just the resonance energy at this point .
we apply this method for extracting the @xmath6 state with the use of the thsr wave function in sec .
[ subsec:04 + ] .
in the hamiltonian ( [ eq:3 ] ) , the explicit form of the long - range part of two - body nuclear force @xmath69 is given by @xmath97\exp \big[-\big(\frac{r_{ij}}{2.5[{\rm fm}]}\big)^2\big]\big(1-m_1^{(2)}+m_1^{(2)}p_{ij}\big),\ ] ] with the value of majorana parameter @xmath98 and @xmath99 the exchange operator of spatial coordinates .
suppose that we here artificially vary the parameter @xmath100 around the physical value @xmath101 , or equivalently , vary the coupling constant defined as @xmath102 around @xmath83 .
this corresponds to the situation explained in sec .
[ subsec : rmsmin ] , where a pseudo potential @xmath103 is added to the original hamiltonian , like @xmath104 with @xmath105 \exp \big[-\big(\frac{r_{ij}}{2.5[{\rm fm}]}\big)^2\big](1-p_{ij}).\ ] ] the energy eigenvalues and eigenfunctions can then be obtained as @xmath8-dependent solutions of the following hill - wheeler equation , @xmath106 here the @xmath77-th eigenfunction is given as , @xmath107 in this calculation , for any values of @xmath8 , the following @xmath108 points of the generator coordinate @xmath109 are adopted : @xmath110 with @xmath111 and @xmath112 .
the size parameter of the @xmath2 particle is fixed as @xmath113 fm . )
( solid curves ) which are measured from the @xmath0 threshold ( doted curve ) .
the crosses denote the binding energies of the fourth @xmath3 state @xmath114 in eq .
( [ eq : spps ] ) . ] of fig .
[ fig:2 ] is enlarged .
the adopted eigenstates in the superposition of eq .
( [ eq : spps ] ) are denoted by open circles .
the crosses denote the binding energies of the fourth @xmath3 state @xmath114 in eq .
( [ eq : spps ] ) . ] in fig .
[ fig:2 ] , the binding energies @xmath115 , with @xmath116 the @xmath0 threshold energy , are plotted up to 10 mev in the range @xmath117 .
the @xmath0 threshold energy is calculated as @xmath118 mev , with @xmath119 the binding energy of the @xmath2 particle , which does not depend on values of @xmath8 , since the pseudo potential @xmath103 does not contribute to the binding energy of the @xmath2 particle with the intrinsic wave function @xmath120 in eq .
( [ eq:2 ] ) . with the increase of @xmath8 from left to right
, we see that three curves go down and pass through the @xmath0 threshold one after another .
the curves are considered to be trajectories corresponding to a resonance or bound state , which has a compact structure and is sensitive to the variation of the strength of the pseudo potential , as discussed in sec .
[ subsec : rmsmin ] . on the contrary ,
the other curves , which are relatively straight , i.e. insensitive to the variation , are considered to correspond to continuum states with unrestrictedly large spatial extension .
we should here notice the important fact that the fourth @xmath3 state should be at around @xmath121 mev above threshold and that it has been missed in the previous calculation of ref .
@xcite . with the decrease of @xmath8 , the trajectory for the fourth @xmath3
state passes through the @xmath0 threshold at @xmath122 , and then from @xmath123 the trajectory quickly becomes invisible .
this is because the fourth state has non - negligible decay width as a resonance .
this is the very reason why the fourth @xmath3 state could not be found in ref .
@xcite , while the first , second , and third @xmath3 states can without problem be assigned to the single eigenstates of the hamiltonian , @xmath124 , @xmath125 and @xmath126 , in eq .
( [ eq:5 ] ) , respectively . however , the method explained in sec .
[ subsec : rmsmin ] enables us to extract the fourth state , which must exist in our approach around @xmath121 mev above the @xmath0 threshold .
the fourth @xmath3 wave function can then be expressed as @xmath127 where the coefficients @xmath128 are determined so that the correct resonance wave function @xmath114 should have a smaller radius than the continuum wave functions with spatially extended structures : @xmath114 should satisfy the following relation against the variation of @xmath129 , @xmath130 in fig .
[ fig : enlarge ] , where the region @xmath131 in fig .
[ fig:2 ] is enlarged , the adopted eigenstates are marked by open circles at given @xmath8 values , and the binding energy of the extracted resonance state is denoted by crosses . for example , at @xmath83 , we utilize in the superposition of eq .
( [ eq : spps ] ) the sixth , seventh and eighth states : @xmath132 , @xmath133 and @xmath134 obtained by solving the hill - wheeler equation ( [ eq:4 ] ) , the binding energies of which are marked by three open circles , while the cross at @xmath83 denotes the binding energy of the extracted fourth @xmath3 state .
we can also see in this figure that binding energies of the extracted resonance states denoted by crosses are reasonably on the trajectory for the fourth @xmath3 state in a wide range of @xmath8 values , indicating the validity of this treatment of resonances . in all subsequent sections , as a matter of convenience , we refer to the wave function of the extracted fourth @xmath3 state at @xmath83 as simply @xmath135 instead of @xmath136 .
thus , in the following , we represent the wave functions of @xmath137 , @xmath138 , @xmath32 , and @xmath6 states as @xmath139 , @xmath140 , @xmath141 , and @xmath135 , respectively .
we present the calculated spectrum for the @xmath137@xmath6 states in fig .
[ fig:3 ] , together with the results of experiment and the @xmath0 ocm calculation @xcite .
we can see that the @xmath0 ocm calculation gives a satisfactory one - to - one correspondence with the experimental spectrum .
this can be understood by the fact that in the calculation the relative motions of @xmath2 particles are solved in a huge model space , which is spanned by gaussian basis functions @xcite .
this allows us to represent the @xmath0-particle gas , @xmath39c clustering , and shell - model configurations , which are realized in the @xmath37 state , the @xmath142@xmath143 states , and the ground state @xmath144 , respectively .
thus , the @xmath37 state , which has the @xmath2 condensate character , is to be identified with the experimental @xmath10 state at @xmath11 mev as the @xmath0 condensate state .
we should note that the observed widths of the experimental @xmath41 , @xmath42 , and @xmath10 states also consistently correspond to the calculated ones ( see table [ tab:1 ] ) . in the same table
, we can see that the @xmath145 values for the @xmath142 , @xmath146 , and @xmath143 states are in good agreement with the corresponding experimental values .
energy spectra between experiment , the @xmath0 ocm calculation @xcite and the present calculation .
dotted line denotes the @xmath0 threshold .
experimental data are taken from ref .
@xcite and from ref .
@xcite for the @xmath41 state .
see table [ tab:1 ] for the values . ] on the other hand , the present thsr ansatz gives only four @xmath3 states including the ground state .
let us now recall that in principle the thsr wave function of eq .
( [ eq : thsr ] ) allows for only two limiting configurations , a pure slater determinant for @xmath46 and a pure @xmath2-particle gas for @xmath47 .
thus , as already mentioned , the scarcity of the @xmath3 states comes from the fact that the thsr wave functions can represent the asymptotic configurations like the @xmath39c clustering only in a rough average way .
in particular , the @xmath17 and @xmath42 states which are considered to mainly include @xmath20c@xmath147 and @xmath148c@xmath149 clustering configurations , respectively , can hardly be represented by this wave function , since it does not possess nonzero spin waves with respect to the @xmath2-@xmath2 relative motions .
we should , however , note that the thsr wave functions are appropriate for describing states which dominantly have a gas - like structure of @xmath2 particles weakly interacting in a relative @xmath22-wave , as the @xmath2 condensate .
also the ground state with essentially a slater determinant configuration is within the possibilities of the thsr wave function . [
cols="^,^,^,^,^,<,<,^,^,^,^,^,<,<,<,^,^,^,^ " , ] as mentioned above , we concluded in ref .
@xcite that the @xmath37 state , which might correspond to the experimental @xmath10 state at @xmath11 mev , may be identified as the @xmath0 condensate state .
it is therefore important to see whether or not the thsr ansatz can give a state corresponding to the @xmath37 state and the @xmath10 state at @xmath11 mev state as well . on the other hand , as mentioned in sec .
[ sec : intro ] , it was discussed in ref .
@xcite that the @xmath32 state could be assigned to the @xmath41 state observed at 13.6 mev , which corresponds to the @xmath150 state .
this means that we should now examine the @xmath6 state found in the preceding subsection , as an important candidate for the @xmath0 condensate state . in table
[ tab:1 ] , the binding energies , r.m.s . radii , monopole matrix elements with the ground state @xmath145 , and @xmath2-decay widths of the @xmath137@xmath6 states are shown .
the reason why the present numbers for the @xmath137@xmath32 states differ from those in ref .
@xcite is due to the difference of the treatment of c.o.m .
motion in the thsr wave function .
in the present work the c.o.m .
motion is completely eliminated , while in ref @xcite this is approximately done .
the @xmath32 state has a large r.m.s .
radius of @xmath151 fm and the @xmath6 state has even a larger one of @xmath152 fm , which are comparable to the values for the @xmath150 and @xmath37 states in the @xmath0 ocm calculation , respectively .
the @xmath6 state as well as the @xmath37 state come quite a bit higher in energy than the observed @xmath10 state at 15.1 mev .
this probably entails that these states have the extremely large r.m.s .
radii , @xmath152 fm and @xmath153 fm , respectively .
these two features possibly go together , since a state high up in the coulomb barrier will necessarily have a quite more spatially extended structure .
the @xmath145 values of the @xmath32 and @xmath6 states well agree with those of the @xmath150 and @xmath37 states , respectively .
these indicate that the @xmath6 state corresponds to the @xmath37 state , and hence to the @xmath11 mev state .
it will be discussed in sec .
[ subsec : interp ] , to what extent the @xmath32 state approximates the @xmath150 state .
@xmath6 states defined by eq .
( [ eq : density3 ] ) .
note that they are multiplied by @xmath154 . ] the nucleon density distribution defined by , @xmath155 for the @xmath137@xmath6 states are shown in fig .
[ fig:4 ] .
while the ground state has a compact saturation density , the @xmath156 state has a very dilute density structure .
this is a common feature of hoyle - like states .
since the effect of the antisymmetrizer in the thsr wave function will be negligible in the state with dilute density @xcite , this state is considered to have the product structure of loosely bound @xmath0 particles , contained in the thsr wave function .
the analogy to the hoyle state can be discussed directly in the @xmath1o system by calculating the overlap amplitude with the @xmath26c@xmath157 structure .
the @xmath0 condensate state must have a large overlap with this structure , since it is natural that if one @xmath2 particle is knocked out from the @xmath0 condensate , the remaining @xmath24c nucleus is in the hoyle state with the @xmath28 condensate structure .
this can be seen in figs .
[ fig:5](a)(d ) , where the following reduced width amplitudes ( rwa s ) for the @xmath158@xmath156 states are calculated , @xmath159 \big|\psi_k \big\rangle , \label{eq : rwa}\ ] ] with @xmath160 being in the ground or hoyle state of @xmath24c obtained via thsr ansatz . for the ground state @xmath137 , the rwa in the channel @xmath26c@xmath161 has a two - nodal behavior due to the pauli principle .
accordingly , the @xmath138 , @xmath32 and @xmath6 states have rwa s with three , four and five nodes , respectively .
the hoyle state component is then the most largely included in the @xmath6 state ( see fig .
[ fig:5](d ) ) , though the @xmath32 state also has a relatively large hoyle - state component ( see fig .
[ fig:5](c ) ) .
we show in fig .
[ fig : rwa_ocm ] the rwa s for the @xmath37 state .
we can see that the amplitudes in the ground state @xmath162 and the hoyle state @xmath163 channels , for the @xmath37 and @xmath6 states , well correspond to each other .
this indicates that the @xmath6 state is the counterpart of the @xmath37 state and most appropriate for the @xmath0 condensate state , rather than the @xmath32 state . on the contrary ,
the @xmath137 and @xmath138 states dominantly have a large amplitude in the ground state channel @xmath26c@xmath161 , reflecting rather compact structures for these states .
calculated in ref .
@xcite . ] based on the @xmath164-matrix theory @xcite , we can calculate the @xmath2-decay widths @xmath165 for the @xmath32 and @xmath6 states , which are shown in table [ tab:1 ] , with the following fomulae : @xmath166 where @xmath77 , @xmath167 and @xmath168 are the wave number of the relative motion , the channel radius , and the reduced mass , respectively , and @xmath169 , @xmath170 , and @xmath171 are the regular and irregular coulomb wave functions and the corresponding penetration factor , respectively .
the reduced width of @xmath172 is related to the rwa s as , @xmath173 . in this calculation , the decay energies are all taken as given by the experimental values : the excitation energy of the @xmath6 state is assumed to be the one of the observed @xmath10 state at @xmath11 mev . similarly the excitation energy of the @xmath32 state is assumed to be the one of the observed @xmath41 state at @xmath35 mev .
the total @xmath2-decay width of @xmath174 mev for the @xmath6 state is in good agreement with the corresponding experimental value of @xmath175 mev .
this total width only comes from the decay into the @xmath176 channel . the partial width decaying into the @xmath177 channel is completely suppressed , up to @xmath178 ev , due to the very small value of the penetration factor @xmath179 in eqs .
( [ eq : gamma ] ) .
this is caused by the sufficiently small decay energy of @xmath180 mev into this channel , in spite of the large overlap between this state and @xmath177 wave function with a certain channel radius , as seen in fig .
[ fig:5](d ) .
in contrast , for the decay into the @xmath176 channel , large penetration is caused by the large decay energy of @xmath181 mev .
nevertheless , the small overlap of this state with the @xmath176 wave function suppresses the decay into this channel , and as a result , the relatively small width of @xmath174 mev is realized , indicating that this state is unusually longlived .
the partial decay width into the @xmath176 channel is calculated with a channel radius @xmath182 fm .
this mechanism also explains the small width of the @xmath37 state , and hence the observed small width of the @xmath10 state at @xmath11 mev as well , which is discussed in ref .
@xcite . in the present calculation ,
we neglect the decay into the other open channel @xmath183 , since the thsr wave function of eq .
( [ eq : thsr ] ) practically can not have nonzero spin components with respect to the @xmath2-@xmath2 relative motions . while the thsr wave function represents the physically clear picture as the @xmath0 condensation , the @xmath184 wave functions @xmath185 with @xmath186 of eq .
( [ eq : hwwf ] ) are expressed as the superposed ones of the thsr wave function of eq .
( [ eq : thsr ] ) with different @xmath62 values . one might therefore suspect that the superposed wave functions of the @xmath184 states no longer keep the remarkable character of a single component wave function , retained in eq .
( [ eq : thsr ] ) . in order to clarify this point
, we first construct orthonormal wave functions from the thsr wave functions as follows : @xmath187 where @xmath188 is projection operators defined by , @xmath189 and @xmath190 is a normalization constant , @xmath191 the wave function @xmath192 depends on the parameter @xmath62 and is orthogonal to the wave functions @xmath193 of the @xmath194 states with @xmath195 .
we then calculate the following squared overlap amplitudes between the wave functions @xmath196 and @xmath185 , @xmath197 note that @xmath198 is written in terms of the parameter @xmath199 , which is related to @xmath62 like @xmath200 . between the wave functions @xmath185 and @xmath192 with @xmath186 , as a function of @xmath199 defined by eq .
( [ eq : theta ] ) . ] in fig .
[ fig:6 ] we show @xmath198 of eq .
( [ eq : theta ] ) as a function of @xmath199 . for any wave functions @xmath185 @xmath201 ,
@xmath198 @xmath201 takes values close to 100 % at optimum values of @xmath199 .
in particular , the maximum of @xmath202 amounts to @xmath203 % at @xmath204 fm .
this means that the @xmath6 state is almost represented by a single thsr wave function parametrized by a large @xmath199 value in the space orthogonal to the other @xmath184 @xmath205 states . we should note that this feature is exactly verified for the hoyle state , strongly supporting the picture that the hoyle state is a @xmath28 condensate @xcite
therefore , the @xmath6 state shows the @xmath0 condensate character , as does the thsr wave function @xmath206 with a large @xmath199 value .
the @xmath32 state also gives a large squared overlap at a large @xmath199 value , i.e. as a maximum @xmath207 at @xmath208 fm .
the r.m.s .
radius of @xmath151 fm of this state is comparable to the one of the hoyle state , and therefore this state also shows @xmath0 condensate character .
the @xmath137 and @xmath138 states also have large squared overlaps , with maximal values of @xmath209 and @xmath210 at @xmath211 fm and @xmath212 fm , respectively .
however , as shown in table [ tab:1 ] , the r.m.s .
radii of these states are rather small . in these states
the antisymmetrizer strongly disturbs the product structure of @xmath0 particles , and therefore these rather compact states will not have a dilute gas - like structure of the @xmath0 s .
the quantities such as occupation probability of a single-@xmath2 orbit and momentum distribution of @xmath2 particles are very important information to discuss the nature of the @xmath2 condensate .
however , the wave functions @xmath185 of the @xmath184 states are not based on the degree of freedom of the @xmath2 particles ( boson ) but on that of the nucleon ( fermion ) , and therefore the various bosonic quantities can not directly be given by the wave functions . in order to obtain them , it is necessary to extract the bosonic wave functions from the fermionic wave functions @xmath185 .
the wave functions @xmath185 in eq .
( [ eq : hwwf ] ) are rewritten as follows : @xmath213 , \label{eq : rgmwf}\ ] ] and @xmath214 where @xmath215 is defined in terms of jacobi coordinates : @xmath216 , @xmath217 , @xmath218 .
we here mention a few features of the wave function eq .
( [ eq : rgmwf ] ) .
it can be rewritten in the form @xmath219 \chi_k(\vc{\eta}),\ ] ] where the following abbreviations are used : @xmath220 , @xmath221 , and @xmath222 . by using the norm kernel @xcite defined by @xmath223 \big| { \cal a } \left[\delta^3(\vc{\eta}-\vc{\xi}^\prime ) \phi^4_\alpha \right ] \right\rangle,\ ] ] the orthonormal condition can be expressed as follows : @xmath224 with @xmath225 .
the above relation suggests that , as discussed by several authors @xcite , a normalized wave function in the bosonic space corresponding to the fermionic wave function @xmath185 can be taken as @xmath226 we should note that the above wave function is written in terms of the relative coordinates of the @xmath2 particles , @xmath227 , and does not depend on the internal coordinates of @xmath2 particles , which are integrated out in the norm kernel @xmath228 . for simplicity , in refs .
@xcite the following approximate form has been taken as @xmath229 , @xmath230 we also adopt this approximation in the following discussions as one option . on the other hand , the square root of the norm kernel can be expanded in using the so - called exchange kernel @xmath231 @xcite , @xmath232 in dilute density states such as expressed by the thsr wave function , contributions from the exchange kernel @xmath231 can be expected to be small . we therefore approximate the above quantity , up to first order , as follows : first let us consider a projection operator @xmath233 @xcite onto the pauli allowed space in the @xmath0 system .
since the norm kernel @xmath234 is defined in the pauli allowed space , the following relations are satisfied , @xmath235 , and hence @xmath236 .
these properties should also be kept after the expansion of eq .
( [ eq:1-k ] ) , which therefore should have the following form in the first order approximation : @xmath237 where for the second equality @xmath238 is used . in diluted density states , to good approximation , the operator @xmath233 can only act on any pairs of @xmath2 particles to exclude the pauli - forbidden , @xmath7 , @xmath239 , and @xmath240 relative states .
we thus adapt @xmath233 to the present calculation as follows , @xmath241 where @xmath242 is the harmonic oscillator wave function in a @xmath7 , @xmath239 , or @xmath240 state , denoted by @xmath243 , for the @xmath2-@xmath2 relative motion .
eventually , we can also adopt as another form the following approximate normalized wave function in the bosonic space , @xmath244 we adopt the bosonic wave function @xmath245 , in the forms of eq .
( [ eq:23 ] ) and eq .
( [ eq:27 ] ) , for the present calculations of quantities such as the one - body density matrix in discussing the amount of @xmath2 condensation .
hereafter , we refer to the ways of approximation for @xmath245 of eq .
( [ eq:23 ] ) and eq .
( [ eq:27 ] ) as apr1 and apr2 , respectively . by using the above @xmath2-particle bosonic wave function ,
the one - body density matrix for the @xmath2 particle is defined following refs .
@xcite , like @xmath246 then , by solving the following eigenvalue problem , @xmath247 we can obtain the occupation probability @xmath168 and the corresponding single-@xmath2 orbit @xmath248 as the eigenvalue and eigenfunction , respectively . the occupation probability @xmath168 is labeled with the angular momentum @xmath249 and the quantum number of a positive integer @xmath250 , like @xmath251 : in this work , for a single-@xmath2 orbit with an angular momentum @xmath249 , we denote the largest occupation probability as @xmath252 @xmath253 , the second largest as @xmath254 @xmath255 , the third largest as @xmath256 @xmath257 , etc .
( see also ref .
@xcite ) .
-particle orbits with @xmath258 and @xmath259 @xmath260 for the hoyle state calculated with apr1 ( solid curve ) and apr2 ( dash - dotted curve ) approximations .
see text for the definitions of apr1 and apr2 . ]
-particle orbits with @xmath258 and @xmath259 @xmath260 , for the @xmath137 ( dotted curve ) and @xmath6 ( solid and dash - dotted curves ) states .
see text for the difference between the solid and dash - dotted curves .
the values of @xmath168 in the parentheses are the occupation probabilities into the corresponding orbits . ]
-particle orbits with @xmath258 and @xmath259 @xmath260 , for the ground state @xmath144 ( dotted curve ) and the hoyle state @xmath37 ( solid curve ) states , which are obtained from the @xmath0 ocm calculation @xcite .
the values of @xmath168 in the parentheses are the occupation probabilities into the corresponding orbits . ] in order to check if the above - mentioned two approximations of apr1 and apr2 work well , at least , in dilute density systems , we first apply the above formalism to the hoyle state which has already been well known to have the dilute gas - like @xmath28 structure .
we show in fig .
[ fig:7 ] the single-@xmath2 orbits for the hoyle state , which are labeled as @xmath261 and obtained with the two ways , apr1 ( solid curve ) and apr2 ( dash - dotted curve ) .
the microscopic wave function of the hoyle state is given by solving the hill - wheeler equation with the thsr ansatz , where the same effective nuclear force is adopted as in the present study of @xmath49 .
we can see that both curves are in good agreement with each other .
this fact guarantees that the approximations are realistic for the dilute density state .
we see that the single-@xmath2 @xmath22 orbit has @xmath7 nodal behavior with a very extended tail .
this is the characteristic feature of the @xmath2 condensate state @xcite , where the @xmath2 particles occupy the lowest @xmath7 orbit of a rather soft mean - field - like potential @xcite .
as we discussed above for the @xmath24c case , also for @xmath1o case it is important to calculate the occupation probability and the corresponding single-@xmath2-particle orbit , which are the crucial quantities for judging the amount of the @xmath2 condensation . in fig .
[ fig:12 ] , the single-@xmath2 @xmath22 orbits for the @xmath137 ( dotted curve ) and @xmath6 ( solid and dash - dotted curves ) states are shown .
the solid and dash - dotted curves are obtained with the approximations of apr1 and apr2 , respectively , i.e. the ones of eq .
( [ eq:23 ] ) and eq .
( [ eq:27 ] ) for the bosonic wave function of eq .
( [ eq:21 ] ) . the dotted curve for the ground state
is calculated with apr1 .
the occupation probabilities into the corresponding orbits are also shown in parentheses .
the fact that both curves with apr1 and apr2 for the @xmath6 well agree with each other again gives a support that both approximations are taken correctly .
the approximation apr2 can be justified for the description of dilute density states , so that the effect of the antisymmetrizer is weakened .
we see in fig .
[ fig:12 ] for the @xmath6 state the large single-@xmath2 @xmath22 orbit occupancy of @xmath262 % and @xmath263 % for the approximations apr1 and apr2 , respectively , which has a nodeless behavior with a very extended tail .
we can see in fig .
[ fig:13 ] that the @xmath37 state gives the similar value of @xmath38 % of the occupation probability and the corresponding single-@xmath2 @xmath22 orbit also resembles the one for the @xmath6 state , concerning the @xmath7 nodal behavior and the spatial extension .
these results again give us support that the @xmath6 state is to good approximation the counterpart of the @xmath37 state described in our previous paper @xcite .
we should note that this behavior of the single-@xmath2 orbit for this state is characteristic as an analogue to the hoyle state @xcite .
on the other hand , for the ground state , the single-@xmath2 orbit has @xmath264 nodal behavior with a strongly reduced tail , reflecting a compact density and hence the strong effect of the pauli principle , for both approaches , thsr and ocm .
particle defined by eq .
( [ eq : alpha_mom ] ) , ( a ) @xmath265 and ( b ) @xmath266 , for the @xmath137@xmath6 states .
the approximation of apr1 is adopted to obtain the bosonic wave function.,title="fig : " ] particle defined by eq .
( [ eq : alpha_mom ] ) , ( a ) @xmath265 and ( b ) @xmath266 , for the @xmath137@xmath6 states .
the approximation of apr1 is adopted to obtain the bosonic wave function.,title="fig : " ] ) , ( a ) @xmath267 and ( b ) @xmath268 , for the @xmath137@xmath6 states.,title="fig : " ] ) , ( a ) @xmath267 and ( b ) @xmath268 , for the @xmath137@xmath6 states.,title="fig : " ] the distributions of the occupation probabilities into the single @xmath2 orbits with apr1 are shown for the @xmath137 , @xmath138 , @xmath32 , and @xmath6 states in figs .
[ fig:14](a)(d ) , respectively .
the orbits shown in fig . [ fig:12 ] correspond to the ones labeled as @xmath261 for the @xmath137 state ( fig .
[ fig:14](a ) ) and the @xmath6 state ( fig . [ fig:14](d ) ) .
we can thus see that the occupation probability for the @xmath6 state concentrates on the single @xmath7 orbit shown in fig .
[ fig:12 ] with @xmath262 % .
this also means that the remaining @xmath269 % to @xmath270 % come from the occupations into higher orbits , following the effect of the antisymmetrization . on the other hand
, the @xmath137 state has rather equal distributions of occupation probabilities labeled as @xmath261 , @xmath271 , and @xmath272 .
in particular , the largest contribution comes from the @xmath272 orbit , reflecting more or less the doubly closed shell structure with the su(3 ) configuration of @xmath273(\lambda , \mu)=(0,0)$ ] .
we should mention that the @xmath32 state and even the @xmath138 state also have considerable concentration of occupation into a single-@xmath2 @xmath22 orbit as seen in figs .
[ fig:14](b ) and ( c ) .
the values of the occupation probability are @xmath270 % and @xmath274 % for the @xmath138 and @xmath32 states , respectively , to be compared with the value of @xmath262 % for the @xmath6 state . however , as mentioned in sec .
[ subsec : wf ] , we think that these states , just in a rough average way , correspond to the observed @xmath16 and @xmath41 states with the @xmath275 structures , respectively . it is more likely that the concentrations of the occupation into the single-@xmath2 @xmath22 orbits for these states disappear in a situation where the @xmath2-@xmath2 relative motion is solved in a larger model space so as to include the @xmath276c clustering configurations , as was done in the previous @xmath0 ocm calculation . for the @xmath137@xmath6 states , the sums of the occupation probabilities for the ten orbits ( @xmath261 , @xmath277 , @xmath271 , @xmath278 , @xmath272 , @xmath279 , @xmath280 , @xmath281 , @xmath282 , and @xmath283 ) are @xmath284 , @xmath285 , @xmath285 , and @xmath285 , respectively .
due to antisymmetrization , there are about 10 percent missed from higher orbits other than those shown in fig .
[ fig:14 ] .
the momentum distribution of @xmath2 particle is also an important quantity to judge which state has a @xmath0 condensate nature .
it can be obtained by calculating the following doubly fourier transformation of the one - body density matrix @xmath286 in eq .
( [ eq:37 ] ) , @xmath287 in fig .
[ fig:15 ] the above momentum distributions of the @xmath2 particles are shown for the @xmath137@xmath6 states .
as is consistent with the large occupation into the single @xmath2 @xmath7 orbit , the @xmath6 state has the most prominent @xmath8-function - like peak at around zero momentum , which is one of the typical characters of the @xmath2 condensate state . according to the rather strong concentrations of the occupation probability on the @xmath22 orbits for the @xmath138 and @xmath32 states
, these states also have some enhancement around zero momentum . on the other hand
, the @xmath137 state does not show such an enhancement .
we show in fig .
[ fig:16 ] the nucleon momentum distributions for the @xmath137@xmath6 states , which can be defined , as in the @xmath2-particle case in eqs .
( [ eq:37 ] ) and ( [ eq : alpha_mom ] ) , like @xmath288 with @xmath289 the prominent peak of the @xmath2-particle momentum distribution seen at around zero momentum for the @xmath6 state disappears in the nucleon momentum distribution .
this means that the nucleons , on average , have rather high momentum components even for the @xmath6 state .
the nucleon momentum distribution corresponds to the fourier transformation of the nucleon density distribution , which is defined by eq .
( [ eq : density3 ] ) ( see also fig . [ fig:4 ] ) .
thus , this just reflects the diluteness of the nucleon density and does not give the property of the condensate structure composed of the @xmath2 particles .
it is instructive to discuss the nuclear size dependence of the occupation probabilities of the single-@xmath2 orbits in the @xmath1o@xmath290 state within the framework of the @xmath0 thsr ansatz .
the results are shown in fig .
[ fig:18 ] for the single-@xmath2 orbits labeled as @xmath261 , @xmath277 , @xmath271 @xmath278 , and @xmath272 .
they are calculated based on the single @xmath0 thsr wave function defined by eq .
( [ eq : thsr ] ) , which has the free parameter @xmath62 triggering the size of the nucleus .
starting with the single @xmath0 thsr wave function , we extract the corresponding bosonic wave function with the approximation of apr1 , and then calculate the one - body density matrix of the bosonic state , to obtain the occupation probabilities as the eigenvalues of the one - body density matrix .
the obtained occupation probabilities then depend on the @xmath62-parameter , i.e. the r.m.s .
radius @xmath291 of the nucleus .
-particle orbits labeled as @xmath261 , @xmath277 , @xmath271 , @xmath278 , and @xmath272 in the @xmath292 state on its r.m.s .
radius @xmath291 . the way of apr1 is adopted to obtain the bosonic wave function .
] we can clearly see that with the increase of the r.m.s .
radius @xmath291 the occupation probability comes to concentrate on a single @xmath22 orbit and the ones into the other orbits get rapidly suppressed .
a similar behavior is also seen in the previous analysis for the @xmath24c@xmath290 state based on the @xmath28 ocm @xcite and in infinite nuclear matter case @xcite , where the condensate fraction is enhanced as the density decreases . here
we notice that the r.m.s .
radii of the @xmath137 and @xmath6 states are calculated as @xmath293 fm and @xmath294 fm , respectively .
the distributions of the occupation probability at @xmath293 fm are close to the ones of the full solution of the @xmath137 state using the hill - wheeler equation , which are shown in fig .
[ fig:14 ] .
this is reasonable , since the single component thsr wave function , which coincides with the doubly closed shell wave function in the compact limit , as mentioned in sec .
[ subsec : thsr ] , describes the @xmath137 state very well .
we should recall that the squared overlap between them is @xmath295 , as discussed in sec .
[ subsec : overlap ] . however , the occupation probability of the @xmath261 orbit at @xmath294 fm amounts to about @xmath296 % , which is considerably larger than the value of @xmath262 % for the @xmath6 state .
the reason of this difference exists in the fact that the @xmath6 state can no longer be represented to good approximation by the form of the thsr wave function @xmath297 of eq .
( [ eq : thsr ] ) , but by the form of the wave function @xmath298 , as precisely defined by eq .
( [ eq:38 ] ) .
let us now recall the discussion made in sec .
[ subsec : overlap ] , where the latter wave function is constructed with the orthogonalization to the lower excited states and the ground state .
that is , we can think that the orthogonalization reduces the occupation probability into a single @xmath22 orbit to the value of @xmath262 % , from about @xmath296 % .
in other words , we can say that the reduction is caused by the effect of the antisymmetrization , which remains in the orthogonalization operator @xmath299 defined by eq .
( [ eq:39 ] ) . in the previous sections
, we clarified that the @xmath6 state apparently has a @xmath0 condensate structure , and is likely to be the counterpart of the @xmath37 state with the @xmath0 condensate structure obtained in the @xmath0 ocm calculation .
we therefore consider that the @xmath6 state probably corresponds to the observed @xmath10 state at @xmath11 mev , as does the @xmath37 state .
in this situation , we should keep in mind that the thsr wave function is not adapted to treat @xmath276c configurations , since all @xmath2 s are treated on equal footing .
therefore , the @xmath138 and @xmath32 states obtained from the thsr wave function mock up in an incomplete way the @xmath39c configurations in @xmath1o
. it will be important in future work to include those configurations into the thsr description .
on the other hand , the position of the @xmath32 state is rather close to the @xmath0 threshold , and we see that the state also has a large amount of @xmath2 condensate fraction , though the amount is not so large as that of the @xmath6 state . in this respect ,
the @xmath32 state could also be a candidate for the @xmath0 condensate .
however , as mentioned above , the result of the @xmath0 ocm calculation indicates that a sizable fraction of the @xmath0 condensation is only included in the @xmath37 state and not in the other five @xmath3 states .
the explanation may again reside in the fact that we treat with thsr ansatz all @xmath2 particles equally and then a relatively large r.m.s .
radius value necessarily also leads to an @xmath2-gas - like state whereas in reality a large size also can be obtained from an @xmath26c configuration .
this points to the necessity to include @xmath26c configurations into the thsr wave function .
let us mention that in ref .
@xcite the magnitude of the angular distribution of cross section in the @xmath300 inelastic scattering to the observed @xmath41 state agrees very well with that calculated for the @xmath32 state , in spite of our present conclusion that the @xmath17 state is not very similar to the @xmath6 state .
this agreement , nevertheless , seems rather natural , since as discussed in refs .
@xcite , the magnitude of the angular distribution in the inelastic scattering depends most sensitively on the size of the excited state and not so much on its internal structure .
in this work , we reanalyzed the thsr wave function for the @xmath0 system introduced previously by the last four authors of the present paper . searching for states in the continuum we found an additional @xmath3 state .
this @xmath6 state was obtained at 3 mev above the @xmath0 threshold and was shown to have a large r.m.s .
radius of 6.1 fm .
this finding could be done with the help of the proper treatment of resonances based on the accc method , which was developed by some of the present authors .
we compared the wave function of the @xmath6 state with the one of the @xmath37 state , which was previously obtained with the @xmath0 ocm calculation as the @xmath0 condensate state .
it was clarified that we could reasonably understand that the @xmath6 state corresponds to the @xmath37 state , and then is the most appropriate for the @xmath0 condensate state , rather than the @xmath32 state which had been considered to correspond to the condensate state in @xcite . in the analyses of the thsr wave function
, we showed that the @xmath6 state contains a large component of @xmath26c@xmath157 clustering in a way which is similar to the @xmath37 state .
this indicates that both states correspond to each other and have a gas - like configuration of the @xmath0 particles as the analogue to the hoyle state , from the fact that the hoyle state has a @xmath28-gas - like configuration as the @xmath28 condensate state .
the @xmath6 state is then described by a single thsr wave function with a large value of the @xmath62 parameter in a space orthogonal to the other three @xmath137@xmath32 states , though the @xmath137@xmath6 states are in general expressed as the form of superposition of many thsr wave functions with respect to values of the @xmath62 parameter . in order to further discuss quantitatively the amount of @xmath2 condensate fraction , we extracted the boson degrees of freedom from the thsr wave function , which is microscopically described based on the nucleon degrees of freedom .
it was done via somewhat approximate procedures which should be valid at low densities .
we showed that the occupation probability for the @xmath6 state concentrates by a large amount on a single-@xmath2 @xmath7 orbit , like the @xmath37 state .
the shape of the orbits for both states is similar to each other and characteristic for hoyle - like states .
furthermore , the @xmath6 state is shown to have a strong peak around zero momentum in the @xmath2-particle momentum distribution , indicating that the @xmath2 particles are condensed around zero momentum .
we also discussed the reason why the @xmath32 state , as well as the @xmath138 state , can not be regarded as corresponding to the @xmath0 condensate state , in spite of the fact that the states possess somewhat enhanced condensate - like features .
all these results strongly suggest that the @xmath6 state corresponds to the @xmath37 state and is the most appropriately considered to be the @xmath0 condensate state .
this further gives us a strong support that the @xmath0 condensate state exists around the @xmath0 breakup threshold and is very likely the observed @xmath10 state at @xmath11 mev .
this conclusion will be supplemented with more extended calculation than the present thsr description so as to incorporate @xmath301 clustering configurations . as concluded from comparison with ocm calculations
, the @xmath301 components may then have a substantial contribution so that the condensate fractions will be reduced .
it is nevertheless expected that the @xmath6 state remains an excellent approximation also in such an extended thsr description .
one of the authors ( y. f. ) acknowledges financial assistance from the special postdoctoral researchers program of riken .
this work was partially supported by jsps ( japan society for the promotion of science ) grant - in - aid for young scientists ( b ) ( 21740209 ) and for scientific research ( c ) ( 21540283 ) . v. i. kukulin and v. m. krasnopolsky , j. of phys .
a * 10 * , 33 ( 1977 ) ; v. i. kukulin , v. m. krasnopolsky and m. miselkhi , sov .
* 29 * , 421 ( 1979 ) ; v. i. kukulin , v. m. krasnopolsky , and j. hor@xmath303ek , _ the theory of resonances : principles and applications _
( kluwer academic , dortrecht , 1983 ) , chapter 5 .
y. suzuki , w. horiuchi , m. orabi , k. arai , few - body systems * 42 * , 33 ( 2008 ) .
t. yamada , y. funaki , h. horiuchi , g. rpke , p. schuck and a. tohsaki , phys .
a * 78 * , 035603 ( 2008 ) ; phys .
c * 79 * , 054314 ( 2009 ) .
note that the positive integer @xmath250 is different from the number of node for the radial part of the corresponding single-@xmath2 orbit @xmath248 .
for instance , in fig
. [ fig:12 ] , the single-@xmath2 orbits labeled as @xmath258 and @xmath259 @xmath260 have @xmath264 and @xmath7 nodal behaviour for the ground @xmath137 and the @xmath6 states , respectively . | the @xmath0 condensate state for @xmath1o is discussed with the thsr ( tohsaki - horiuchi - schuck - rpke ) wave function which has @xmath2-particle condensate character . taking into account a proper treatment of resonances , it is found that the @xmath0 thsr wave function yields a fourth @xmath3 state in the continuum above the @xmath0-breakup threshold in addition to the three @xmath3 states obtained in a previous analysis .
it is shown that this fourth @xmath3 @xmath4 state has an analogous structure to the hoyle state , since it has a very dilute density and a large component of @xmath5 configuration .
furthermore , single-@xmath2 motions are extracted from the microscopic 16-nucleon wave function , and the condensate fraction and momentum distribution of @xmath2 particles are quantitatively discussed .
it is found that for the @xmath6 state a large @xmath2-particle occupation probability concentrates on a single-@xmath2 @xmath7 orbit and the @xmath2-particle momentum distribution has a @xmath8-function - like peak at zero momentum , both indicating that the state has a strong @xmath0 condensate character .
it is argued that the @xmath6 state is the counterpart of the @xmath9 state which was obtained as the @xmath0 condensate state in the previous @xmath0 ocm ( orthogonality condition model ) calculation , and therefore is likely to correspond to the @xmath10 state observed at @xmath11 mev . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.